ÅBO AKADEMI UNIVERSITY FACULTY OF SCIENCE AND ENGINEERING By: JORGE FERNÁNDEZ MÉNDEZ Tuthors: Miguel Ladero Galán. Henrik Grénman. Bachelor Thesis Final Work. Presented as mandatory requirement in the obtention chemical engineering bachelors graduate. Turku, June 2020 UNIVERSIDAD COMPLUTENSE DE MADRID CHEMICAL SCIENCES FACULTY CHEMICAL ENGINEERING AND MATERIALS DEPARTMENT By. JORGE FERNÁNDEZ MÉNDEZ Tuthors. Miguel Ladero Galán. Henrik Grénman. Bachelor Thesis Final Work. Presented as mandatory requirement in the obtention chemical engineering bachelors graduate. Madrid, June 2020 UNIVERSIDAD COMPLUTENSE DE MADRID FACULTAD DE CIENCIAS QUÍMICAS DEPARTAMENTO DE INGENIERÍA QUÍMICA Y DE MATERIALES DISEÑO DE LA ZONA DE REACCIÓN PARA LA PRODUCCIÓN ENZIMÁTICA DE HIDROLIZADOS DE ORIGEN AGROALIMENTARIO Por: Jorge Fernández Méndez. Tutores: Miguel Ladero Galán. Henrik Grénman. TRABAJO FIN DE GRADO Presentado como requisito para optar al título de Graduado en Ingeniería Química Madrid, Mayo de 2020 v THANKS & DEDICATORIES. Writing the full stop on this paper represents more than just the completion of an assignment. It is the end of a personal stage. That kind of pause in our way that always comes, usually with a completely different appearance than the one imagined. Finding right now myself writing these lines 3335 km away from what has always been my home is something that won’t be possible without the support and help of many people. First of all I would like to thank my family, for their advice and respect for my own decisions, as well as the permanent support I have always been lucky to live. Special mention to my parents, to whom I owe much more than what is possible to write in this page, being always present in any circumstance. Also to Elvira, who has always known how to have the best details at the most opportune moments. I cannot forget two of the most important people in my life; Pablo and Abraham. Words are unnecessary, but none of us would be the person we are today if we had not met each other. Thank you for our close and sustained friendship over time, to the point of building each other up. Thank you to all our classmates. Both those of us who knew each other from the first days, and those with whom we were able to share an intense but short time. You have always tried to take care of each other and to be sincerely concerned about the rest. Despite the impersonality present in faculties as large as ours, it is these small details that give meaning to attending personally to the lectures rather than from online. In addition, this work would not have been possible without the continuous work, recommendations and comments of my tutors. I would like to thank my tutor in Spain, Miguel, who in spite of the distance has always shown total closeness and sincere interest, virtues that are difficult to find. I must also mention the warm welcome, suggestions and flexibility of my tutor in Finland: Henrik. Thanks to his support, writing a bachelor’s thesis at a foreign university has been almost like doing it at home. Finally, I would like to thank all those people with whom I have shared those little moments that bring the light of meaning to life. Sincere friendship, understanding, philosophizing about the world or simply overlapping presences without specifying anything else. Too many names spread out over time to fit on this page, but none of them of less importance. Thank you all for everything. vi Index of contents. Summary of contents. (Abstract). ............................................................................... 11 CHAPTER 1 .................................................................................................................. 12 1.1 Scope and objectives of the project. Conceptual framework. .................... 12 1.1.1 The Biorefinery as a conceptual framework.............................................. 12 1.1.2 Definition of work......................................................................................... 14 1.2 Product definition. D-Mannose. 2 ................................................................. 15 1.3 Process Feedstock. Galactoglucomannan. ........................................................ 17 1.4 Process Technology ............................................................................................. 19 1.4.1 Extraction of galactoglucomannans ........................................................... 19 1.4.2 Galactoglucomannans hydrolysis. .............................................................. 20 1.4.3 Heterogeneous Catalysts .............................................................................. 22 CHAPTER 2 .................................................................................................................. 25 2.1 Market Study. .................................................................................................. 25 CHAPTER 3 .................................................................................................................. 32 3.1 Process thermodynamics analysis. ................................................................. 32 3.2 Process kinetics and catalysts analysis. ......................................................... 35 CHAPTER 4 .................................................................................................................. 40 4.1 Pretreatment section ....................................................................................... 40 4.2 Reaction Section. ............................................................................................. 42 4.3 Purification and separation. Downstream section. ...................................... 43 4.4 Mass and energy balances of the process. ..................................................... 45 CHAPTER 5 .................................................................................................................. 46 5.1 Pre-treatment section. ..................................................................................... 46 5.2 Separation and purification............................................................................ 47 5.3 Design of the reaction section ......................................................................... 48 5.4 Plant layout. Layout ........................................................................................ 69 CHAPTER 6 .................................................................................................................. 71 6.1 Dow’s fire and explosion index. ..................................................................... 71 6.2 Análisis de riesgo y operabilidad. Estudio HAZOP. .................................... 72 CHAPTER 7 .................................................................................................................. 74 7.1 Plant location and Environmental legislation. .............................................. 74 7.2 Environmental Impacts. ................................................................................. 76 7.3 Corrective measures. ....................................................................................... 81 vii CHAPTER 8 .................................................................................................................. 83 8.1 Investment estimation. ........................................................................................... 83 8.2 Sales estimation ............................................................................................... 85 8.3 Production cost estimation. ............................................................................ 85 8.4 Economic viability studies. ............................................................................. 88 ANNEXES ..................................................................................................................... 97 Table Index. Item Page. Table 1. General properties of D-Mannose 16 Table 2. Solid catalysts for acid hydrolysis of hemicelluloses 22 Table 3. Prices of the main resources involved in the production of pulp by thermo-mechanical processing 28 Table 4. Data used to estimate the cost of obtaining GGM. Price and consumption of services referred to tones of pulp obtained on a dry basis 29 Table 5. Cost estimates of aqueous GGM solutions from thermo-mechanical pulp processing. 29 Reference thermodynamic parameters for the hydrolysis of cellulose 33 Table 7. Linear correlation parameters for Cp as a function of concentration (%wt) 34 Table 8. Kinetic parameters for acid hydrolysis of GGM. Smopex-101 cation exchange resin. 39 Table 9. Material balance for the reaction section 45 Table 10. Properties of AISI-31676 steel 57 Table 11. Main parameters of the reactors in the hydrolysis unit 58 Table 12. Parameters and dimensions calculated for the agitation system 60 Table 13. Overall heat exchange coefficients 62 Table 14. Characteristics of the reaction section exchangers 63 Table 15. Characteristics of the reaction section pumps 67 Table 16. Calculation of the general risk factor of the process 71 Table 17. Process-specific risk factor calculation 72 Table 18. Environmental impact matrix for the GGM hydrolysis process. 81 Table 19. Cost estimation for plant equipment. 84 Table 20. Fixed and working capital cost estimation by percentage method 84 viii Table 20. Annual sales estimation. 85 Table 21. Total production cost for the installation. 87 Figure Index Item Page. Figure 1. Preferential fractionation pyramid within a biorefinery process 13 Figure 2. Main components of lignocellulosic material. 14 Figure 3. Mannose molecule in conformations and 16 Figure 4. Main uses of D-Mannose as a product. 17 Figure 5. Generic molecular structure of an o-acetyl-galactoglucomannan chain. 18 Figure 6. Reaction mechanism for acid hydrolysis of O-glycoside bonds. 21 Figure 7. Development and projection of the global D-mannose market. 26 Figure 8. Specific heat for aqueous solutions of D-Galactose, D-Mannose and D-Glucose. 34 Figure 9. Micrograph of Smopex 101 36 Figure 10. Enlargement of the block diagram. Pretreatment section. 41 Figure 11. Enlargement of the flow diagram. Pretreatment section. 41 Figure 12. Enlargement of the block diagram. Reaction section. 43 Figure 13. Enlargement of the flow diagram. Reaction section. 43 Figure 14. Enlargement of the flow diagram. Purification section. 44 Figure 15. Enlargement of the block diagram. Purification section or Downstream. 44 Figure 16. Conversion - Reactor Volume. Piston flow simulation. 53 Figure 17. Conversion - time for hydrolysis reaction in batch reactor. 54 Figure 18. Temperature - time for the hydrolysis reaction in the batch reactor 54 Figure 19. Pressure drop (20ºC) in beds with DOWEX® HCR-W2. 56 Figure 20. High efficiency pitch blade turbine for solids suspension. 59 Figure 21. Parameters and dimensions of a standard agitation system. 60 Figure 22. Temperature of the reaction medium considering heat dissipation. 64 ix Figure 23. Aerial view of Holmen's plant near Hallstavik. 74 Figure 24. Annual wind distribution in the Hallstavik region. Annual wind hours by direction and wind speed. 75 Figure 25. Geographical location of the mill. 75 Abbreviations. Symbol Index. HL Filling height in the tank E Height of the agitator on the bottom W Width of the agitator blade J Width of baffle plates Cp Specific heat ( ) Molar amount of monosaccharide "i" as free sugar in the hydrolysate. Q Heat flow Mass flow rate Volumetric flow rate of stream "i Coefficient of autocatalysis for the resin. Exponential coefficient of autocatalysis for the resin. U Overall heat exchange coefficient Cw Water concentration. C Sugar concentration (%wt). Ci Concentration of monosaccharide "i in the GGM molecule. Concentration of the monosaccharide "i" in the GGM Initial water concentration. Initial concentration of monosaccharide "i" in the GGM ki Global kinetic constant of the monosaccharide "i". k’i Intrinsic kinetic constant of hydrolysis for the monosaccharide "i". Annual cost of amortization Total annual production cost Man D-Mannose. PVDF Polyvinylidene fluoride Gal D-Galactose. PVP Poly-4-vinyl-pyridine GGM Galactoglucomannan. SMB Simulated Moving Bed Chromatography Glu D-Glucose. TMP Thermomechanical processing of paper pulp. PHW Extraction with pressurized hot water. NPV Net present value x Density of the mixture Fluid Density Agitator Diameter DT Tank diameter S Maximum stress supported CA Permissible corrosion thickness e Reactor thickness Heat exchanger correction factor Correction factor for suspensions in prevented sedimentation Arrhenius pre-exponential factor Volume fraction of catalyst in the system DP Average degree of polymerization. Corporate tax L Agitator blade length Po No. of stirring power Rem Reynolds' number in agitation Mass percentage of solids Pm Total electrical power required P Power requirement. Pd Vessel design pressure Pi Relative proportion of monosaccharide "i" in the GGM molecule Molar ratio of reaction of monosaccharide "i" per volumetric unit of catalyst Molar ratio of reaction of total monosaccharide "i" in the reaction system Net Percentage Profitability Discount rate / Hot fluid inlet/outlet temperature (in tubes) / Cold fluid inlet/outlet temperature (in the housing) ∆ Temperature variation N Stirring speed Nmin Minimum stirring speed Sedimentation terminal velocity prevented. Annual sales ( ) Volume occupied by the liquid phase per reaction cycle Volume occupied by the reaction mixture per cycle 11 Summary of contents. (Abstract). The following document corresponds with the final work presented as mandatory requirement in the obtention of the chemical engineering bachelor’s graduate. This document has been elaborated with the support of two high education institutions; Universidad Complutense de Madrid and Åbo Akademi University, under the supervision of two co-tutors from the former mentioned universities. The redaction of this report has been performed during an Erasmus stay in Turku, Finland. Then, attending to convenience reasons and with the approval from both institutions, the topic of this work has been conducted in a way that makes suitable to exploit the research performed in the host university. The aim of the following report is to study the valorization of agroindustrial residues and side streams. Attending to biorefinery principles, the focus is made on the key industries and materials found on the receiving institution region: forestry and pulping. Because of that after the title of the report, an explanatory text is found, defining the key insight about the contents developed in this document. From all potential valorization strategies and technologies, hemicellulose hydrolysis has been selected. The high potential of implementation in already existing industrial settings, and the current limited usage of hemicelluloses on traditional pulping processes are the key factors for this selection. Main contents of this final bachelor’s thesis are structured in a way similar to the studies and documents used for chemical engineering projects planification and development. First, key theoretical insights and objectives of the project are exposed (chapter 1) as well as a previous study of the materials and technologies that make it feasible (chapter 2). Then the technical memory is presented, showing the thermodynamic and kinetic properties involved in the process (chapter 3) and shaping the basic schemes for process operations (chapter 4). Once the process is defined, basic engineering calculations are exposed, while equipment’s and process’ specifications are provided (chapter 5). Later, safety and environmental studies of the project are also performed and detailed (chapters 6 and 7). Eventually economic viability of the process is studied (chapter 8), determining the potential real implementation of the proposed operations and overall process. 12 CHAPTER 1 Introduction and objectives. 1.1 Scope and objectives of the project. Conceptual framework. Recently, interest in developing a transition from the current fossil-resource based production model to an economy independent of fossil sources has grown significantly. There are two main reasons that fuel this trend; The growing interest in the environmental impact of economic activity, as well as the greater awareness of the need for sustainable production models. It implies, societal and economic development without assuming a depletion of the resources used or damaging the environment. In the chemical industry, concerns about environmental impact and sustainability have been addressed for a couple of decades through the green chemistry and process intensification approaches. Likewise, in recent years, the growing interest in the development of a new production model based on the principles of the circular economy has given rise to the concept of biorefinery within the chemical industry sector. This new production philosophy advocates the use of biomass as starting point for product obtention. 1.1.1 The Biorefinery as a conceptual framework. The concept of biorefinery has evolved significantly during the last decade, establishing itself as a set of principles and philosophies based on the sustainable use of organic matter of biological origin, available on a recurring bias. This concept offers a framework for multiple fractionation technologies for biomass conversion into multiple products of interest, in a manner inspired by the current fractionation and processing technologies carried out with fossil-based materials. 13 The principles of this approach consist in the integral use of all the components present within biomass, starting from the products with the highest added value to the recovery and processing of any valuable compound, eventually ending with the residual components valorization, often as fuels or burnable materials for energy production. Figure 1 depicts the priority order to follow in every biomass fractionation. Biomass processing is develop following the order defined by products market price. Likewise, due to the greater complexity and variability of different biomass sources, the technologies and processes involved within a biorefinery should be very specific to the local conditions of availability and characteristics of the biological resources available in the region. In this way, the concept of biorefineries advocates the creation of installations of moderate size and productive capacity, distributed homogeneously with the aim of limitations and accessibility of the different forms of biomass available. The following work is focused on green biorefinery, using lignocellulosic biomass as feedstock. The aim of the project consists on the design of appropriate technology for the valorization of the hemicellulosic fraction once separated from the starting material. The distribution of the components in the hemicellulosic material is observable in figure 2. Hemicelluloses are heteropolysaccharides composed of different pentoses and hexoses (arabinose, xylose, ribose, galactose, glucose and mannose) besides acidic sugars (galacturonate, glucuronate) linked by O-glycosidic bonds. Its biological role is mainly structural, being part of the cell walls of plant cells in combination with cellulose, lignin and pectin. Figure 1. Preferential biomass fractionation in biorefinery.1 14 Unlike cellulose, hemicelluloses form amorphous structures that help stabilizing the cell walls. Due to its integration with other polymeric molecules, its extraction from the material is limited, in addition to its solubility, by its ability to diffuse through the matrix. Figure 2. Main components of the lignocellulosic material. Likewise, there is a great variety of hemicelluloses, where their chemical composition and properties differ widely between plant species. In general, softwoods are highly abundant in galactoglucomannan, while hardwoods have xylanes as the main hemicellulosic constituent. It should be noted that there are other types of hemicelluloses found in annual plants. As an example glucans are commonly found in cereals. In addition to the different types of hemicellulosic polymers, characterized by a similar distribution of bonds and similar composition, the ratio of each constituent monomer can vary with each species. 1.1.2 Definition of work. In the following work, the design of an unit for the biorefinery processes is described: A hemicellulose catalytic hydrolysis unit for obtaining monosaccharides. The following work is carried out in collaboration with the Finnish university Åbo Akademi during a mobility program, due to this reason the proposed unit is especially oriented to the use and conversion of resources with greater availability in the formerly mentioned region; wood and its derivatives. Norway Pine, Spruce and Birch are the main industrially available tree species. So that, galactoglucomannan (GGM) has been selected as the starting material model, and mannose (Man) as the main product of interest. The objective of the following work is to establish a general procedure for the designing of hemicellulose hydrolysis units. Despite working with two well-defined compounds, the design can be easily extrapolated to a wide range of possible 15 hemicellulosic materials and final products. Thus, it aims to define a design example for hemicellulose hydrolysis units using heterogeneous catalysts. Hemicellulose hydrolysis units are one more element involved into the complex scheme of possible stages involved in a typical green biorefinery process (based on lignocellulosic materials usage). In accordance with the principle of making a full use of the different available materials, the hemicellulose recovery stages consist on an intermediate step in multiple processes. Consequently, the designed unit constitutes a model applicable to different industrial processes, where streams rich in hemicelluloses capable of being valued as hydrolysates rich in monosaccharides of interest are present. As main examples, the unit can be introduced coupled to thermomechanical pulping processes, where the aqueous liquid fraction obtained in the conditioning stages of the lignocellulosic material has a high content of hemicelluloses, where galactoglucomannan is the fundamental constituent for the main woods used in these processes. Another possibility is the direct extraction of hemicelluloses using pressurized hot water, also obtaining streams rich in hemicelluloses, and a material composed of cellulose and lignin that can be subsequently fractionated. In that way, the following work proposes the design of a typical unit within the framework of the green biorefinery, taking as an example those compounds most commonly found in current industrial processes based on lignocellulosic biomass. 1.2 Product definition. D-Mannose. 2 D-Mannose (Man), also known as seminose or carabinose, is a glucose C-2 epimer monosaccharide with the molecular formula C5H11O5CHO. It is a white crystalline solid that is generally obtained in the form of a powder, the presentation in aqueous solutions is also common. Likewise, there are two main forms of the molecule; D-mannopyranose or D-mannofuranose, where the first mentioned is the most common. It is water soluble (248g / 100 mL) and slightly soluble in ethanol. 16 Furthermore, during the crystallization of aqueous solutions, about 60% crystallize as -D- Mannopyranose, while the remaining 40% do so as -D-mannopyranose. The first form presents a soft and sweet flavor, while the second is bitter, due to this, the crystallized D-Mannose presents a sweet flavor with a bitter aftertaste. The sweetening power is 0.6 times less than that of table sugar. Figure 3 shows the two main forms -D-Mannopyranose and -D-Mannopyranose. Table 1 resumes the main physicochemical properties of D-Mannose. Table 1. General D-Mannose physicochemical properties. Property Value Molecular Formula C6H12O6 Molecular Weight 180.16 g/mol Appearance White Crystalline Solid (Powder) Melting Point 132 ºC Optic Activity -17º (forma ) / +34º (forma ) Solubility 248g / 100 mL (17 ºC in H2O) Density 1.54 g / cm3 (Crystalline Solid) Caloric Value 3.75 kcal / g Likewise, D-mannose is a product with a wide range of applications both at industrial level and as a final consumer product. Its main applications are in the food Figure 3. Mannose molecules in and conformation. 17 sector (food additive, texturizer3.), Cosmetics (emollient agent) and pharmaceutical (direct use, synthesis of anti-tumor4 or immunostimulatory agents5). Likewise, there is a growing interest in D-Mannose due to its potential applicability in new sectors, such as livestock (feed additive6), or as a new base compound for the synthesis of drugs and other chemicals. Figure 4 summarizes some of the main current applications of the product. Figure 4. Principal applications of D-Mannose. In addition, as hydrolysis sub-products amount of D-glucose (21%) , D- Galactose (13%) and other monosaccharides (12%) are found.7 These compound should also be considered as potential products for further usage within the biorefinery process or for direct market sell. 1.3 Process Feedstock. Galactoglucomannan. Galactoglucomannan (GGM) is a heteropolysaccharide composed of a linear backbone of β-D-mannopyrose and β-D-glucopyranose with (1→4) bonding, accompanied by monomeric units of (1→6) α-D-galactorpyranose linked in some of the mannose units as chain side group. Likewise, some acetyl groups are present in the C2 and C3 positions of mannose. Acetylated galactoglucomannan has a higher solubility in water, easing its possible extraction from the initial lignocellulosic material. The solubility increases with the degree of acetylation. The average molecular weight of the GGM is from 30,000 to 60,000 Da.8,9,11 18 It should be noted that depending on the origin of the GGM, the degree of acetylation, the ratio of monosaccharides in its composition, as well as its average molecular mass, may differ. In general, the Man: Glu: Gal ratio in the GGM is 0.5-1: 1: 3-4.9 However, some softwoods exploited industrially nowadays have much richer compositions in D-Mannose. In Picea Abies wood (Norway Spruce), the Man: Glu: Gal ratio is 2-4: 1: 0.6, appreciating a majority content of D-Mannose as the main constituent.7,12 Figure 5 shows the GGM molecular structure. Figure 5. Generic molecular structure of a chain of o-acetyl-galactoglucomannan. Galactoglucomannan is the main constituent of hemicellulose in softwoods, reaching 10-20% of the dry weight of the starting material. Likewise, there are different extraction techniques, where the use of pressurized hot water as a solvent has become one of the most promising options, due to its simplicity and saety.7,10 Forestry is one of the main sources of GGM potentially available today. In the case of the pulping industry, process water can be directly used12 as well as new wood pre-treatments for simultaneous hemicelluloses extraction and feedstock conditioning its subsequent processing.7,8,10,11,12 Other residues such as sawdust can be used in similar extraction procedures.13 On another side, hardwoods have a very different composition, their main hemicelluloses are Xylanes, representing 15-30% of the dry weight of the material. These are polysaccharides consisting of a linear skeleton of 30 to 100 units of β-D-xylopyranose with (1→4) bonds. Some Xylans also have side monomers of arabinose (arabinoxylans), glucose (glucoxylanes), galactose (galactoxylanes) or glucuronate acid (glucuronoxylanes) attached to the C3 of xylose.14 These materials could also be used as 19 a potential source of monosaccharides of interest in hydrolytic systems similar to the one proposed. However, the main interest of the work will focus on the use of softwoods with wide industrial use nowadays. Considering the exposed information, the GGM used as starting material should be selected according to its origin, presentation and composition in order to be considered as an acceptable raw material for the majority production of D-Mannose in the hydrolysis unit. Thus, the galactoglucomannan from Picea Abies is considered as the model raw material, due to their availability and high D-Mannose content. To produce other monosaccharides of interest other types of woods could be taken into account though. 1.4 Process Technology The following section sets out the main technological foundations that allow the development of the biomass fractionation and conversion operations previously proposed. The technology for obtaining galactoglucomannan is also superficially considered, while greater emphasis is developed on the technology for the hydrolysis process. 1.4.1 Extraction of galactoglucomannans Most of the galactoglucomannan is partially acetylated, reaching values of 0.28- 0.37 in the material from Picea Abies. Because of that, its solubility in water is high enough to be extracted using water as solvent. The process waters used within the production of thermomechanical pulping processes are aqueous solutions composed by dissolved hemicellulose, which is mainly GGM in the case of softwoods. The use of these process streams can be a direct source of GGM, however, there are new processes with a higher degree of hemicellulose recovery. Hemicellulose extraction methods include water vapor pretreatments14 or microwave irradiation.11,15 These pretreatments are usually subsequently combined with other solvents or different chemical agents in order to dissolve the hemicellulosic fraction. Attending to bibliography, extraction using pressurized hot water (PHW) is the technology with the greatest number of adNPVtages. It is a process based on the direct 20 extraction of GGM and hemicelluloses using liquid water at temperatures of 100-180 ºC. Water usage as a solvent, allows for direct extract usage on its subsequent hydrolysis, besides the use of moderate temperatures. In addition, the operation in a single stage allow to reduce the overall operation cost compared to the technologies mentioned above (a subsequent recovery of solvents or purification of the extract). The main disadNPVtage of PHW processes lies in the tendency of hemicelluloses to hydrolyze under process conditions, reducing their degree of polymerization and average molecular weight. Likewise, due to hydrolysis, the degree of acetylation is also reduced, which can negatively affect the solubility of the hemicellulosic material. Furthermore, acidification of the solution by deacetylation could favor hydrolysis phenomena. For these reasons, the extraction conditions must be optimized in a way for each raw material and specifications to be reached in the extract obtained.10 In the situation where the hemicellulosic extract is going to be used for its subsequent hydrolysis, the reduction of the degree of polymerization by partial hydrolysis is not a very important problem, since it allows to accelerate the hydrolytic process. However, the conditions must be controlled in order to minimize other possible degradation products of the material (5-hydroxymethylfurfural and furfural mostly). 1.4.2 Galactoglucomannans hydrolysis. There are two main routes for the hydrolysis of hemicelluloses; chemical hydrolysis (acid hydrolysis) and enzymatic. Enzymatic processes operate under milder conditions and minimize undesired material degradation. However, they are also slower and require multiple enzymatic hydrolysis steps. Likewise, due to the variety of possible hemicelluloses, the hydrolysis of each material requires an adequate selection of the enzyme cocktail. Because of the complexity of the process and its lower reaction rate, chemical hydrolysis is preferred. Acid hydrolysis processes take place through the protonation of O-glycosidic bonds and subsequent breakdown of the polymeric chain, consuming one molecule of water to regenerate the terminal hydroxyl groups in each molecule produced. Acid hydrolysis processes have traditionally been developed using homogeneous catalysis in 21 acid solutions, however, the possibility of using heterogeneous catalyst is nowadays a viable option. Figure 6 shows the mechanism of the hydrolysis process, where a water molecule is protonated, subsequently reacting with the polymer, splitting it into two polymeric moieties that will continue to react to the basic sugar monomeric units that compose them. The rate limiting step is the last bond cleavage occurring simultaneously to the fast water addition. 1. 2. 3. Figure 6. O-glycosidic bonds hydrolysis mechanism by acid catalysis in polysaccharides. According to their features, using heterogeneous catalysts makes possible to overcome some of the disadNPVtages found in conventional homogeneous acid catalyzed processes. In homogeneous catalytic processes, the concentration of the acid must be controlled in order to avoid reducing the yield, due to the subsequent decomposition of the monosaccharides produced through secondary reactions. Furthermore, the main disadNPVtage of homogeneous catalysts relies on the subsequent need for neutralization. Using organic acids, acid recovery can be achieved by evaporation, whereas with mineral acids (H2SO4 or HCl as the typical acids), they must be precipitated as salts that must be subsequently removed from the reaction media. The use of heterogeneous acid catalysts allows to solve neutralization requirements, reducing operation costs associated with a greater number of processing steps. Furthermore, the use of solid acid catalysts prevents the degradation of sugars to undesired compounds, since the hydrolysis conditions are similar to those of homogeneous catalysis with dilute acids. Apart from the above, the use of solid catalysts allows the development of continuous processes, which possess greater releNPVce and applicability at the industrial scale. The processes based on homogeneous catalysis are usually carried out almost 22 exclusively in batch reactors, increasing operating costs and more difficult system’s scale- up lt. However, it is worth noting the existence of proposed systems that employ homogeneous catalysis in a continuous regimen with residence times close to 100 minutes. The study of which may also offer useful information for further development of heterogeneous processes without the need for subsequent neutralization. Due to these reasons, the use of heterogeneous acid catalysis is considered the most efficient option currently for the hydrolysis of hemicelluloses, by combining acceptable reaction rates, ease of industrial scaling and avoiding neutralization operations.18 1.4.3 Heterogeneous Catalysts Hydrolysis process by means of solid catalysts develops via the mechanism shown in figure 6. The protonation of water takes place through the release of a proton from the solid, where later during the bond cleavage step the solid catalyst accepts a proton again. There are multiple materials that combine high specific areas and a strong acid character or the ability to be functionalized with acid groups. Table 2 shows the main catalysts studied. Table 2. Solid catalysts used for hemicellulose acid hydrolysis. Material. Acid Function. Reference. Zeolites Acidic Zeolites (H form) -SO3H Metals (Ru or Ti) Metal Oxides (WO3/ZrO2) 16, 17,19. Cation Exchange Resins -SO3H 10,17,18. Functionalized Mesoporous Silica -SO3H Metals (Ru or Ti) 19. Functionalized Mesoporous Carbon -SO3H Metals (Ru or Ti) 19. Heteropolyacids Heteropolyacid 20. 23 The main drawbacks of solid catalysts are their possible deactivation and activity. The first, derived mainly from the loss of activity due to leaching or reaction of the acid groups. The second one is mainly controlled by diffusional effects, especially at the beginning of the hydrolysis, when the hemicellulose molecules feature a high degree of polymerization. Likewise, the selectivity to monosaccharides is also affected by the type of acid function, the presence of metals can favor other secondary reactions (oxidations or isomerizations mainly). Zeolitic materials have great versatility of functionalization, high specific areas, as well as a high concentration of acid sites in some cases. However, its possible instability in aqueous media, added to their reduced pore size (great diffusion impediment), represent a disadNPVtage for its application in direct hydrolysis processes of hemicellulose.10,19. On the other hand, cation exchange resins (polymers functionalized with acidic groups), or functionalized mesoporous materials can exhibit more pronounced deactivation phenomena, due to the leaching of functional groups during operation. This requires greater control of operating conditions and proper catalyst selection.17 Considering these aspects, catalysts with sulfonic groups are considered as a preferred option, due to their strong acid character and high immobilization capacity on different supports. Likewise, in order to minimize diffusional and deactivation phenomena, cation exchange resins are selected as the most convenient catalyst. Based on the available literature, sulfonate-group functionalized polymeric resins can feature open fibrous structures, which greatly enhance the diffusion of large molecules into catalyst particles. In this way, fibrous structures guarantee a shorter diffusion length, increasing catalytic activity. Among the various commercially available exchange resins, the Smopex®-101 material is selected as process catalyst. Mainly due to its proven high activity in multiple reactions that take place through acid catalysis, besides the existence of documentation on its previous usage for GGM hydrolysis processes and its good stability in terms of loss of acid function. 10,17,18,21 However, other acid catalysts similar to Smopex®-101 are also available. Some of them feature a higher degree of functionalization (more active centers), which could be used in an similar way by adjusting the catalyst concentration for the same medium 24 pH. These materials include Dowex® HCR-W2, Dowex® M-31 and Amberlyst®-15 resins. The main difference of these resins in comparison with Smopex®-101 is their structure, being meso or macroporous spherical particles with a moderate specific surface (20-50 m2 g-1), and diameters in the range of 0.4 – 1.2 mm. Under the process conditions, the inactivation of the catalyst is not considered a phenomenon, because the feed to the system is free of impurities that negatively affect the catalyst. Likewise, deactivation phenomena have not been documented in the literature. However, there is evidence of acid activity loss in ion exchange resins over time. The catalytic step does not consume the protons associated to the resin by means of sulphonic acid residues, therefore, the reduction in activity over time is mainly due to the deactivation of the acid sites by interaction with other compounds. The exchange with metal cations present in the medium (mainly Ca+2 and Mg+2) is the main cause, as well as the interaction with other molecules as a secondary process that motivates the reduction of the activity of the catalyst. Due to these phenomena, the possibility of regenerating the catalyst by immersion in sulfuric acid is considered if a significant loss of activity is observed. Further discussion of these aspects can be found in chapter 6. 25 CHAPTER 2 Viability Studies. 2.1 Market Study. 2.1.1 Production, consumption, supply and demand studies.22,24,25. D-Mannose as well as other rare sugars is a chemical compound derived from biological sources with a growing and developing market. D-Mannose market is currently divided into four main sectors; The food, nutritional, pharmaceutical and animal feeding markets, other applications still developing though. Within all of them, due to the reduced implementation of industrial processes based on D-Mannose as raw material, the direct sale of D-Mannose as nutritional supplements (capsule format) is the largest sector, covering near the 50% of global market . D-Mannose is mainly sold as a crystalline solid with 99% purity and less than 0.5% humidity. The standard industrial sales format is 25 kg drums. However, the product can also be encapsulated in the production installations, and be sold directly to the consumer market. Within pharmaceutical applications or other sectors, there are higher purity products with a lower market volume. D-Mannose global market is currently growing, featuring an estimated compound annual growth rate (CAGR) close to 3.5% (period 2017-2025). In 2017 the global market value reached US $ 700 (€ 620 Ma). Figure 7 shows the historical development and projections for the global D-Mannosa market. At the supply level, the main industrial players are multinational companies. VWR Corporation, Atrium Innovations Inc, Ferro Pfanstiehl Laboratories, Inc, Now Health Group Incorporation, Hänseler AG, Kyowa Hakko Kirin Co Ltd, EI du Pont de Nemours and Company, Biotech Pharmacal Inc, Jarrow Formulas Inc. and Nutraceutical International Corporation., are the main producers of D-Mannosa worldwide. Anticipating the market growth, Now Health Group Inc., Nutraceutical International Corp., EI du Pont de Nemours Company as well as Jarrow Formulas Inc. have announced a Calculated according to the overall 2017 US $ to € change.23 26 their interest in increasing their production capacities, focusing on their penetration within emerging and developing markets through other local agents involved on them. Regarding demand, the main actors for its development are the European market, the North American market and the Asia-Pacific market. In the first case, the dietary supplements sector is the most , while for the North American and Asia- Pacific markets, the preferred use is the broad introduction of the use of D-Mannose in multiple consumer and personal care products. The estimates represent an increase equivalent to the CAGR of the demand for D-Mannosa until 2025, mainly driven by these three markets. Figura 7. D-Mannose market value. Historical values and future projections. Eventually, it is also remarkable the possibility of a new emerging market development in the medium term; the application of D-Mannose as a starting base product for the industrial production of other formerly mentioned higher-value compounds (Mannitol, Drugs, etc.) Its massive implementation in the livestock industry as a substitute for antibiotics is also a future possibility. The potential development of these markets would imply a significant increase in global demand and production of D- Mannose. 2.1.2 Price Analysis. D-Mannose. 27 The overall price for food grade D-Mannosa (99% purity, 0.5-1% moisture with no significant microbial contamination.26), is highly variable depending on the quantity of product, due to its sale from small format (capsules) to the sale by the tons. However, the medium format (25 kg drums) is the most used currently. The current price of D-Mannosa is estimated by means of the average price offered for the sale of D-Mannosa in medium or high volume (minimum of 25 kg of product). The average obtained through different online purchase portals27,28 of industrial products, gives a value of 51 kg-1. It should be considered that the estimated price refers to a global average, where in the European and American market the final sale price may be increased. This factor should be considered in future studies in greater depth. However, for the following work, the value of 51 EUR kg-1 will be used as a reference for further estimations. Raw Materials. Aqueous solutions of hemicelluloses rich in GGM are the raw material. Picea Abies wood is chosen as the preferred raw material. Since the designed unit operates coupled to a biorefinery platform with greater complexity and diversity of products, the price of the material fed to the unit will be estimated as the price of the raw material plus a fraction of the operating costs required to obtain the hemicellulosic extract. Although there exist several methods for obtaining the hemicellulose extract that are detailed above. For the sake of simplicity and current implementation of the facilities, the thermo-mechanical processing of wood to obtain pulp (Thermomechanical pulping, TMP) will be considered, where the hemicellulose extract is found as the waste stream of process water leaving the initial wood treatment unit. Costs are thus conservatively estimated, considering the price of the wood as well as the fraction of the operating costs of TMP. This fraction is established from the mass ratio of extracted hemicellulose to the total mass of the processed material. The cost estimate is made considering the main items for a TMP plant: Raw material, electricity, heating and process water. The cost of energy together with the raw material are the main determining factors in the cost of production. Labor and other maintenance costs are not considered in this case, as they are accounted as shared with other parts of the installation. 28 The amount of GGM dissolved in the process water is 8% of the total content in the wood;18% on a dry basis. Using this amount as a basis for calculation. The amount of water required is estimated as the process water fed to the system. Tables 3 and 4 show the data used for the estimation. Annex 1 sets out the calculation procedure in a more detailed way. Price estimations are made considering the average inflation in Sweden and considering the value of the average exchange rate in 2020. In the case of the price of process water for industrial use, the highest value found in reference to the Nordic countries is selected, due to its great variability and complexity of obtaining an accurate value. A second cost value for the starting GGM solution is established as the cost of extraction and concentration of the hemicellulosic material coming from thermo- mechanical processing via ultrafiltration and nanofiltration technology based on membranes.32 The estimated final cost of obtaining the GGM by this route is 670 € t-1 Table 5 shows the results of the different cost estimates for feeding the process. Table 3. Prices of the main resources and services involved in TMP operation. Item Value Units References Comments Wood price 240 SEK t-1 30 Picea Abies wood 2005. Sweden Heating price 80 € MWh-1 33 2005 Price. Sweden Electricity price 75 € MWh-1 33,34 2005 Price. Sweden Water price 2 € m-3 35,36,37,38 Average water price 2020. Sweden Inflation 1,3 % 39 2000-2020 Average. Sweden 29 Table 4. Data used to estimate the cost of GGM solutions. Price and consumption of services referring to tones of pulp obtained on a dry basis. Table 5. Cost estimations of aqueous solutions of GGM from TMP process. Item Value Units References Direct process water from TMP 200 €*t-1 -- GGM concentration by membranes 670 €*t-1 32 GGM extract concentration (membranes) 30 g*l-1 32 GGM extract concentration (TMP) 5,26 g*l-1 -- Item Value Units References Comments Wood Price 27,5 € t-1 30 Picea Abies wood. Sweden. Updated 2020 Wood Processability 60 % 12 Usable fraction of wood material Wood-to-pulp ratio 88.2 % 31 Accepted pulp yield from usable wood Wood GGM content 18 %wt 7,11,12,32 Average content on Picea Abies wood. Extracted GGM during TMP 8 % 31,33 Total GGM extracted from Wood. Water consumption 3 m3 t-1 33 Electric consumption 2,5 MWh t-1 33 Heat consumption 0,03 MWh t-1 33 Heating price 97 € MWh-1 33 Updated 2020. Sweden. Electricity price 91 € MWh-1 33,34 Updated 2020. Sweden. Water Price 2 € m-3 35,36,37,38 Average water price. 2020 Sweden 30 It should be considered that the estimated concentration is an approximation based on the documented ratio of extracted GGM in TMP processes and available process water. The amount of recoverable process water is equivalent to the water fed to the facilities. However, according to bibliographic sources it can be considered lower, close 1.15 g l-1 of dissolved hemicellulose.40 Changes on operating conditions can enhance de extracted GGM within the process water. 2.1.3 Location. The proposed hemicellulose hydrolysis unit should always be coupled to a larger process. In the case of this work, a TMP installation is chosen as a model. The location of the installation is projected in Sweden, as a processing unit integrated within a thermo-mechanical-pulp mill, both planned and pre-existing. As a reference, Hallsta paper mill, which belongs to the Swedish company Holmen, is selected. Hallsta Paper Mill is found within the north of Upland Province, Sweden, 10 km away from the Baltic Sea. As the mill is currently in operation, the introduction of an additional processing unit would involve changes to other aspects of the plant.33,42 The main reasons for selecting this site are the high availability of Picea Abies wood in the region and the growing interest in and political and economic support for the development of integrated biorefinery processes. The availability of information and the high concentration of the timber industry in the region are also determining factors in this choice. 2.1.4 Design Capacity. The design capacity of the unit should be established in accordance with two main factors; processing capacity of the plant, and maximum demand for the final product. Attending the data showed in this section, the demand for D-Mannose, as well as other sugars (Mainly Gal and secondly Glu), may be considered sufficiently high not to limit production capacity. Consequently, hydrolysis unit design capacity is established as the excedent process water used in the thermo-mechanical processing of the pulp. The capacity is set according to the volume of excedent process water available in the plant after TMP wood processing, which depends on the amount of wood 31 processed. From this information, an estimation for available GGM and D-Mannose can be calculated. The calculation is presented in Annex 1. A design capacity of approximately 2400 t year-1 of D-Mannose is defined. 32 CHAPTER 3 Process thermodynamics and kinetic studies. 3.1 Process thermodynamics analysis. Galactoglucomannan hydrolysis thermodynamics have not been studied in depth. However, there is available information on the hydrolysis of other polysaccharides of similar nature to GGM (Xylans, starches or celluloses). The precise measurement of thermodynamic parameters of polysaccharide hydrolysis is complex and most of the available information is based on complex mixtures of lignocellulosic materials. The most thermodynamic parameters considered are reaction enthalpy (ΔHR), reaction entropy (ΔSR) and Gibbs free energy (ΔGR). The enthalpic term allows to know the energy released or required during the reaction. The other two terms allow to determine the spontaneity of the process, determining its tendency to occur and the characteristics of the chemical equilibrium. Given the observed similarity between the main thermodynamic parameters of polysaccharide-based materials, as well as their similar chemical nature, an average value obtained from different bibliographic sources will be considered as a reference for calculations. Available information for celluloses and starches is used, due to the fact that they are also made of hexoses and have a density of O-glycosidic bonds close to that found in GGM. In addition, crystallinity of the material is considered as a key factorwith great impact on thermodynamic parameters. Given the amorphous nature of GGM, only the thermodynamic parameters of hydrolysis of non- crystalline materials are studied. According to the available information, a reduced value for hydrolysis enthalpy is observed. This reduced value, allows explaining the variability found between different polysaccharides (finding values in the range of 5-100 J g-1). In any case, within this range, the reaction enthalpy for the hydrolysis of hemicelluloses at low to moderate concentrations (1 - 100 g L-1) would not have a great impact on the final temperature of the system. (Increase lower of 1 ºC at complete conversion). According to the literature43,44,45, both the reaction enthalpy and the Gr of the hydrolysis process show some insensitivity to temperature. Table 6 shows the thermodynamic information, taken as a reference for the hydrolysis process. Due to the 33 different degree of polymerization of the polysaccharide to be hydrolyzed, the parameters are expressed on a mass basis and per molar unit of the produced monosaccharide. Table 6. Reference thermodynamic parameters selected for GGM hydrolisis.45 Likewise, the reaction is considered to be an irreversible non-equilibrium reaction, where the hydrolysis process is not limited by the temperature increase in the studied range (80-130ºC), nor by monosaccharide re-polymerization phenomena.46 Thus, ΔSR and ΔGR are not studied in depth, since the reaction can be considered as irreversible. 3.1.1 Specific heat of mixtures. The specific heat of the GGM solutions and products obtained is considered similar, allowing their grouped study with the same thermal behavior. Experimental correlations47 are available for glucose solutions as a function of temperature and concentration (Equation [3.1]), as well as bibliographical information for the specific heat of aqueous solutions of D-Mannose, D-Glucose and D-Galactose48. The influence of temperature on the heat capacity is studied through the previously mentioned correlation, where the parameter a1 can be modified for the reference value of each sugar. For the studied temperature range (25-120 ºC), the dependence of the Cp. of the solutions with the temperature is not (< 0.1 J g-1 ºC). Consequently, the Cp value of the mixtures is considered an exclusive function of the concentration and type of monosaccharide. Figure 8 shows how the Cp of the solutions decreases with the concentration of sugar. The bibliographic values47 are adjusted to a linear model in order to obtain a correlation of Cp as a function of concentration. Table 7 shows the parameters of these adjustments. Parameter Value (J g-1) HºR (J G-1) -100 HºR (J MOL-1) -16.4 34 The specific heat for the GGM solution or mixture is finally estimated considering the contribution of each sugar in the reference ratio 3:1:0.6, Man:Glu:Gal. The weighted mean is made with the parameters obtained by linear regression of the experimental data. The correlation obtained will be used in the calculation of the Cp of the sugar solutions. [3.1] = ( ∙ ∙ , ∙ ) , = 4.144 − 1.673 ∙ ( ∙ º ) [3.2] ai (J g-1 ºC-1) bi (J g-1 ºC-1) R2 Man 4.144 -1.665 0.995 Glu 4.149 -1.744 0.995 Gal 4.136 -1.600 0.991 Figure 8. Specific heat for D-Galactose, D-Mannose and D-Glucose aqueous solutions.46,47 Table 7. Lineal correlation parameters as a function of sugar concentration (%wt). 35 3.1.2 Viscosity of mixtures. The viscosity of the solutions used during the process is also studied. Due to their different behavior, solutions with high GGM content and those rich in monosaccharides are studied separately. Aqueous GGM solutions have a higher viscosity than the sugars ones. In high concentration (> 10%wt), they display a non-Newtonian viscoelastic rheopectic behavior, reducing their apparent viscosity in the when the applied force is increased. The viscosity also depends on the state of the suspended polymer, as well as its degree of polymerization. At reduced concentrations the non-Newtonian behavior becomes less . Under the premise of working at moderate concentrations (< 100 g L-1 of GGM), documented data of viscosity49 at different concentrations of GGM are used, obtaining an exponential correlation as a function of the GGM content, useful in the range 0.5 % - 10 %. (Equation [3.3]). ( ∙ ) = 22.9 ∙ . ∙ ² = 0.97 [3.3] Likewise, monosaccharide solutions are studied as a group. Considering a similar nature for all monosaccharides, with a behavior comparable to glucose solutions. Viscosity values at different concentrations and temperatures are obtained from bibliographic sources50 and are detailed in Annex 2. 3.2 Process kinetics and catalysts analysis. 3.2.1 Heterogeneous acid hydrolysis process. GGM hydrolysis using acid heterogeneous catalysts relies on the surface reaction of GGM molecules with the acid groups of the cation exchange resin used as a catalyst. The chosen exchange resin (Smopex-101) is a non-porous material, where the sulfonic acid residues act as active centers, being exposed to the reagent on the external surface of the resin. In addition, due to the fibrous structure of the catalyst, the mass transfer phenomena can be considered negligible, operating under non internal diffusion 36 limiting conditions. Likewise, it has been verified that a minimum pH of 1.5 is required for the hydrolysis kinetics to be appreciable. In the case of heterogeneous catalysis, this value is easy to achieve. However, in heterogeneous catalysis, only some materials with a high concentration of acid sites meet these requirements. Acid hydrolysis of GGM and other polysaccharides shows first-order kinetics depending on the concentration of protons in the medium as well as the concentration of non-hydrolyzed GGM. In the case of using heterogeneous catalysts, the process occurs showing an autocatalytic behavior. There is an initial latency period where the reaction rate is low. After this period, hydrolysis takes place according to the pseudo-first order kinetics observed during homogeneous catalysis. The autocatalytic behavior of the reaction is determined by the interaction of the GGM molecules with the catalyst. In the initial stages, the molecules are spheroidal polymeric chains coiled up on themselves. The random collisions of these molecules with the polymer surface are what allow the initial hydrolysis of monosaccharides, reducing the size and length of the polysaccharide with every collision. The decrease in the degree of polymerization of the GGM leads to a more open and unfolded conformation of the polymer, allowing a greater interaction with the surface of the catalyst. It has been proven that hydrolysis of monosaccharide residues takes place in a random way. In the case of heterogeneous catalysis, during the initial stages there is a greater tendency to lateral galactose residues preferential hydrolysis, as they are more easily accessible. However, as the degree of polymerization of GGM is reduced and the polymer structure becomes more open, hydrolysis develops in a random manner, where the ratio of released Man:Glu:Gal is constant. In addition, no appreciable content of oligomers is observed, as these are rapidly degraded to the final monosaccharides. Likewise, the hydrolysis process is highly temperature-dependent, where activation energies for monomer formation have been documented at 100-140 kJ mol-1 of sugar obtained.16 The calculated activation energy depends on the monosaccharide whose production kinetics have been studied. An average value of 120 kJ mol-1 is Figure 9. Smopex 101 fibers micrography51. 37 selected as the activation energy of the reaction, where the kinetic constant is temperature- dependent on an Arrhenius-like fashion. According to the manufacturer52 the selected catalyst is designed for operation at temperatures close to 100 ºC. However, the fibers hold at temperatures up to 200 ºC without showing significant degradation. Excessively high temperatures could mean the need to operate under pressure to avoid water evaporation, however, the kinetics of the process would be greatly favored. Furthermore, the possible degradation of monosaccharides to other by-products should be considered. Hexose has a higher thermal stability than pentoses often found in other hemicellulosic materials, allowing to consider negligible the degradation of obtained monosaccharides over a wider temperature range. According to the available documentation, it is observed that at temperatures below 130 ºC there is no significant degradation of the monosaccharides obtained.53,54 Because of this, an operating temperature range of 90 - 125 ºC is selected as a design condition. 3.2.2 Kinetic model. The design of the hydrolysis unit requires a kinetic model capable of predicting the hydrolysis of each monosaccharide, as well as including the documented autocatalytic phenomena. Thus, the kinetic model must allow the calculation of the concentration of GGM and monosaccharides and their integration within the mass balance on the reaction unit. A kinetic model developed specifically for the hydrolysis of GGM is selected.18 In the proposed model, the reaction mechanism outlined in Chapter 1, Figure 6, is considered. The first two stages are considered to be in quasi equilibrium, while stage 3 is the control stage. Thus, the hydrolysis rate is defined individually for each monomer, considering instead of the total GGM only the fraction of O-glycoside bonds to be hydrolyzed belonging to each specific type of sugar. The quantity of total O-glycoside bonds in the GGM can be estimated from the average degree of polymerization of the GGM (DP) as well as the ratio Man:Glu:Gal. The hydrolysis ratio for each monosaccharide is defined by equation [3.4]. Considering the initial concentrations of monosaccharides forming part of the polymer, the amount of unhydrolyzed 38 monosaccharide units per material balance is expressed, applying the same reasoning for water consumption during the reaction (equation [3.5]). ∙ = ′ ∙ ∙ ∙ [3.4] ∙ = ′ ∙ ∙ − ∙ ( − ) [3.5] Also, considering the nature of the catalyst, the kinetic constant is defined as a grouping of the intrinsic constant (k'i) and the concentration of interchangeable protons in the resin (CH). In addition, the reaction rate is referred to the volume of catalyst to be used, considering the density of the material (b). This defines the general kinetic constant of the process (ki). = ′ ∙ ∙ . The autocatalytic behavior is modeled considering the evolution of the hydrolysis kinetic constant with the conversion of the reaction (X), which can be followed according to the reduction of the degree of polymerization of the GGM (DP), since the number of monomers forming each molecule can be defined as DP+1. (Equation 3.6). According to a model of potential increase mediated by the exponent, equations 3.7 and 3.8 can be developed, considering the limit situations with null conversion (beginning) and total conversion (end of the reaction). It is considered that in X=0 , ki = ki0 and in X=1 , ki = . So that, two possible models dependent only on two parameters are proposed to define the autocatalytic behavior. equations 3.9 and 3.10 show these models. [3.6] = − ∙ // = ′ ∙ [3.7] [3.8] = + ( − ) ∙ ( − − 1 ) ≡ 1 + ′(1 − (1 − ) ) [3.9] = ∙ (1 + ) [3.10] 39 Because of its simplicity, equation 3.10 is chosen as the expression for kinetic constant calculation. Both the parameter and are considered exclusive functions of the catalyst nature. Table 8 shows the kinetic information for the selected catalyst. Likewise, the activation energy of the process and the temperature are used for calculation of the Arrhenius pre-exponential factor for each monosaccharide ( ), the Arrhenius equation is used (Equation 3.11). = ∙ [3.11] Parámetro Valor ( °) (L mol-1 min-1) 0,694 ∙10-6 ( °) (L mol-1 min-1) 0,600 ∙10-6 ( °) (L mol-1 min-1) 0,721 ∙10-6 35,5 0,7 a (kJ mol-1) 120 (L mol-1 min-1) 1,26 ∙1011 (L mol-1 min-1) 1,09∙1011 (L mol-1 min-1) 1,31∙1011 Table 8. Kinetic parameters for the acid hydrolysis of GGM over Smopex-101 cation exchange resin. Values determined at 90 ºC. It is considered a single Ea for all hydrolysis 40 CHAPTER 4 Process description. The following chapter gives brief insight within the basic units of the proposed process for thermo-mechanical pulp processing plant excedent process water. According to the block diagram attached in annex 3, the process is divided into three main sections: pre-treatment, reaction and downstream. Annex 4 also shows the process flow diagram. Given the peculiarities of the described process, D-Mannose production is considered as the main objective. However, some basic unitarian operations could be included or modified in the different sections of the designed system, in order to increase the adaptation of the process to different contexts and demands of different products. The proposed design is based on the characteristics considered optimal for the TMP facility used as an example for the integration with an hemicellulose hydrolysis unit. Other possibilities are briefly discussed in the following sub-sections. 4.1 Pretreatment section Within TMP facilities, excess process water is usually stored in tanks for further treatment or discharged in a controlled manner to the environment. Although the process water can be sent to the hydrolysis unit directly, the content of suspended solids, impurities, as well as the low GGM concentration, establish the requirements for a pretreatment stage. An adequate pre-treatment can drastically reduce the demands of the reaction and purification stages, lowering the final cost of the process. Pretreatment is a crucial stage for the viability of the unit. The selected pretreatment consists of two consecutive stages. The first consists on a microfiltration unit to purify the process water, reducing the content of suspended solids. Then the filtered solution is concentrated by membranes usage from an initial average value of 1 g L-1. to a value of 60 g L-1. In this way, a GGM solution is obtained with a reduced content of organic matter and impurities. Based on the composition documented using this type of technology, the GGM solution obtained has a reduced content of impurities and a high concentration of GGM that will favor the kinetics of the hydrolysis reaction. 41 Figure 10 shows an extension of the block diagram of the process. In the same way in figure 11 the main equipment of the unit is detailed. In summary, the excess water from the TMP process is fed to the pre-treatment section, eliminating suspended solids during a first stage, and subsequently concentrating the GGM by means of membrane technology. A main stream rich in GGM is obtained, as well as two waste streams. In the microfiltration stage, the solids consisting mainly of lignocellulosic fines and microscopic solid suspended particles constitute a waste flow. In the ultrafiltration stage, a solution mainly composed by low molecular weight solutes from the previous thermo-mechanical treatment of the wood is obtained also as a side stream. Figure 10. Block diagram of the pretreatment section. Figure 11. Process flow diagram of the pretreatment section. 42 It should be noted that, for the proposed process, a specific pre-treatment section is designed to take adNPVtage of a excedent process water stream from the TMP processes, due to the fact that these are industrial facilities currently in operation. However, depending on the integration of the hydrolysis unit in different contexts, the pre-treatment section is the most susceptible to severe modifications. As an example, starting from a hypothetical biorefinery installation, based on the total fractionation of the lignocellulosic material A first stage of hemicellulose extraction from chipped wood, using a PHWE process, would allow a stream to be obtained with a very high GGM content (60 - 70 g L-1)9 and reduced impurities, without requiring previous concentration and purification stages. At the same time, this type of pre- treatment increases the processability of the remaining solid material, allowing the subsequent use of the lignin and cellulose fractions. The main adNPVtages of this approach rely on the reduction of both costs and the energy and organic solvents usage. On the other hand, the lack of implementation nowadays of integral lignocellulosic biomass fractionation installations makes the current implementation of the process in that context difficult. 4.2 Reaction Section. Reaction section consists mainly on the hemicellulose hydrolysis reactor. A fixed- bed reactor design, operating in adiabatic regime is selected. The enthalpy of the hydrolysis reaction is moderate, and the processed solutions have a reduced GGM concentration. As a result, the reaction develops close to isothermal conditions. Figures 12 and 13 show the reaction section in the block and flow diagrams respectively. The exit stream from the pretreatment section is preheated before it is fed into the reactor bringing it up to reaction temperature. After the reaction, a solution with a negligible GGM content and a mixture of D-Man, D-Glu and D-Gal sugars, as well as other minority impurities, is obtained. This stream is directed to the purification section in order to recover the sugars released during the reaction. 43 4.3 Purification and separation. Downstream section. The obtained aqueous solution of D-mannose, D-glucose and D-galactose at the outlet of the reactor is then pumped into the downstream section. In the developed design, the purification of the hexoses by chromatography is proposed, obtaining a high purity solution of D-Mannose, D-Galactose and D-Glucose (> 99.9%). Likewise, an effluent stream with other impurities found in the reaction mixture is also obtained, mainly acetic acid and non-hemicellulose organic matter. The hexose solutions obtained can be subsequently crystallized, obtaining the product in its final presentation for direct sale in the consumer market. D-Mannose has a high market price and a growing demand, as does D-Galactose, making it viable to sell them directly in the market under they crystallized form. On the other hand, due to the low price of D-Glucose in comparison with the other obtained sugars, D-glucose stream can be derived to other units that allow its valorization. As it is Figure 12. Reaction section in the block diagram. Figure 13. Reaction section in the flow process diagram. 44 a purified product, hydrogenation by heterogeneous catalysis to obtain sorbitol is considered an economically viable possibility. Likewise, another alternative for the pure sugar solutions obtained is evaporation. This allows them to be marketed as syrups or other preparations, avoiding the costs of crystallization operations. Figure 14 shows the block diagram of the section, in the same way that figure 15 shows an extension of the proposed flow diagram. Figure 14. Downstream section in the block diagram. Figure 15. Downstream section in the process flow diagram. 45 4.4 Mass and energy balances of the process. The design of the process is centered on the reaction section, consequently the balance of matter around the hydrolysis section shown in Figures 12 and 13 is made. Since the section has a battery of 5 tanks in discontinuous operation, the results are presented for the loading and unloading streams of the unit, both at the beginning and at the end of the operation. Table 9 shows the results of the material and energy balances for the system, referring to the different streams. The details of the calculation are detailed in chapter 6 of this report. Annex C includes the extension of the material balance sheet with mole fractions. C N m COMPOSITION xH2O xGGM xMan xGlu xGal T P nº kmol min-1 kg min-1 (g L-1) ºC atm 8 19,804 377,94 995,7 60 0 0 0 30 1 9 A-I 19,804 377,94 995,7 60 0 0 0 30 2,1 10 A-I 19,804 377,94 943 60 0 0 0 120 2,1 11 A-I 19,885 377,94 943 5,99 35,48 11,36 7,16 120 2,1 12 A-I 19,885 377,94 943 5,99 35,48 11,36 7,16 120 2,1 13 A-I 19,885 377,94 992,4 5,99 35,48 11,36 7,16 40 2,1 Table 9. Mass and energy balance for reaction section 46 CHAPTER 5 Engineering and process design. The chapter on detailed engineering covers those aspects related to the demands and technical specifications for the construction of the process installation. The main content of this work consists of the exhaustive design of the hydrolysis reactor (R-101), including its dimensioning, construction specifications, fluid dynamic and heat exchange aspects. Also included are basic definitions of the operating principles and equipment required in each section of the process proposed in the previous chapters. The possibilities for energy integration of the process are also considered. The proposal of an instrumentation and control system for the whole process is included. Finally, the layout of the different sections and equipment of the process is considered in an implementation or layout plan according to design rules. The design proposal for the pre-treatment and downstream sections of the process is consolidated around different publications and bibliography. Documented operating configurations are selected as effective for processes with a high degree of similarity to those required in each section. 5.1 Pre-treatment section. The proposed pre-treatment is mainly based on the publications of T. Persson et al. 32,55. The possibility of isolating the hemicellulose fraction dissolved in the excess water of a thermo-mechanical pulp processing plant is demonstrated. In the same publication, a techno-economic evaluation is also carried out, demonstrating the viability of the process. For the microfiltration stage, the use of a rotating drum filter is proposed. The filter medium consists of a 30 m polystyrene film. After microfiltration, the solution is purified and concentrated by diafiltration and ultrafiltration. Polyvinylidene fluoride (PVDF) ultrafiltration membranes with a molecular weight cut-off of 10 kDa are used for this purpose. 47 The configuration of the membrane unit is considered as a multi-stage waterfall with recirculation. The membranes are of the coiled spiral type, presenting the same area in all stages of the process. Likewise, the operating temperature limit of the membranes is 60 ºC, since the average temperature of the excess water from the TMP processes is close to 80 ºC, heat exchange equipment is also required to lower the temperature of the feed. A shell-and- tube heat exchanger is the proposed solution for this purpose. It should be noted that in the above-mentioned publication, the concentration of the solutions is selected up to a value of 30 g L-1, however, the proposed process could be used to obtain solutions with a concentration of up to 66 g L-1. The increase in the cost of the operation could be considered as a compensable fact with the higher concentration of GGM and the consequent reduction in the flow of solution to be treated in the hydrolysis reactor and in the subsequent downstream stages. Likewise, the content of low molecular weight impurities is reduced as the concentration of hemicellulose increases. Membrane technology is selected in preference to other concentration routes because of its ability to simultaneously remove impurities present in the feed, as well as to present a sufficiently low operating cost to allow the process to be viable. 5.2 Separation and purification. The separation and purification stage of the products, or downstream of the process, consists of two main operations; chromatographic separation of the sugars in the mixture and conditioning as a final product by continuous crystallization or evaporation. The sequence of operations considered is based on the bibliographic evidence. The reactor effluent does not present a high content of impurities or degradation products and does not require pre-treatment before being fed to the separation unit. Furthermore, the concentration of sugars present is in the usual range for chromatography operations documented in the literature. Chromatographic purification is proposed under the simulated moving bed (SMB) operation configuration. It consists of a set of interconnected chromatographic filler columns, using poly-4-vinyl-pyridine (PVP) as the stationary phase. Chromatography purification is selected due to the high degrees of purity required for the commercialization of the final product. Likewise, the study of new chromatographic separation processes is one of the fields with the greatest development during the last 48 decades, existing at present a great range of technologies with potential to be implemented industrially. According to the literature, it is possible to separate mixtures of hexose in solution, achieving the fractionation of the sugars of interest, as well as the separation of impurities during the same operation. The estimated cost is 0.012 $ L-1* of processed food, being within the limits of viability of the process. Based on the bibliographical references, SMB chromatography processes allow the resolution of ternary mixtures of compounds in continuous with a high performance, reduced consumption of resources (energy and eluent), besides presenting a much lower cost than the classic chromatography. The resolution of ternary mixtures can be performed in SMB systems based on five zones, which represent a modification of the usual processes based on 4 chromatographic zones. There is evidence of the process design for the separation of 5 different monosaccharides which can be adapted for the resolution of the 3 hexoses of interest in the process. 56 Likewise, there is enough study of this type of SMB chromatography for the resolution of ternary mixtures to be able to consider them a real possibility in their industrial implementation. 57,58,59,60 The reactor effluent is cooled down to the optimum temperature for chromatographic separation (60 - 70 ºC) and fed into the chromatographic separation unit. On leaving the unit, three streams are obtained with a concentration of D-Man, D-Glu and D-Gal of nearly 99% purity. After chromatography, the D-mannose and D-galactose streams are directed to the crystallization unit. It consists of two continuous crystallization units. This type of operation is selected due to the need to obtain a high-purity solid product ready for marketing, as well as the high degree of development of this technology. 61,62 Likewise, the D-glucose stream is directed to an evaporator where it is concentrated into a concentrated solution (> 100 g L-1). This glucose rich stream can later be used in other sections of the process. Since it is a purified sugar, it is proposed to hydrogenate it by means of heterogeneous catalysis. 63,64 5.3 Design of the reaction section The design of the reaction section involves the definition of multiple interrelated operating variables. The objective is to define the characteristics and operability of the unit so that the operation carried out develops in an optimal way. * Price updated to 2020 considering an average inflation of 2% per year, corresponding to the region of the study. 49 Among the possible design methodologies, it was decided to follow a holistic approach, considering those factors influenced by the performance of the reactor, in addition to those with influence on the hydrolysis reaction. To this end, the interaction of all preliminary designs with the rest of the factors is studied, using this information for the optimization of the initial design. In the following chapter the final design of the reactor is detailed, explaining pertinently those modifications on preliminary designs in view of the results. 5.3.1 Reaction conditions. Based on the above information, the reaction can be considered as virtually thermo-neutral at the operating concentrations. The reaction kinetics is relatively slow, with an initial latency period and a very high activation energy. Likewise, the concentration of reagent in the food is reduced and the process takes place by heterogeneous acid catalysis. The pH of the reaction medium must be between 0.5-1 to allow for hydrolysis of the material without apparent degradation of the products. The catalyst used is an ion exchange polymer resin, functionalized with sulfonic groups responsible for the acid activity. Given its nature, the catalyst has a low density and a high tendency to swell. Likewise, the particles offer pressure dropes that can exceed 5 bar m-1 according to the manufacturer. 65 This set of characteristics motivates the following design decisions for the reaction section. Considering the operating limit of the catalyst and the kinetics of the process, an operating temperature of 120 ºC is selected, finding a compromise between the resistance of the catalyst and the significant increase in the kinetics of the reaction with the temperature. Operation in a liquid medium at 120 ºC implies the need to pressurize the reactor. Therefore, the operating pressure will be 2.1 atm, corresponding to the saturation point of the water and considering a safety margin of 0.1 atm. The catalyst concentration is defined by the required pH of the medium. According to the literature a pH between 0.5 and 1 is optimal for hydrolysis. A pH value of 0.75 is selected. The kinetic model used has been obtained at pH = 1, however, taking into account the reaction ratios with homogeneous catalysis at lower pH the kinetic increase is moderate. Operating at a slightly higher catalyst concentration allows to avoid conversion reductions below specifications even after moderate deactivation of the catalyst. The catalyst concentration is calculated considering the ion exchange capacity of the resin. 50 Given the low concentration in the feed and the slow kinetics, the concentration gradient operation is favorable. Therefore, the possibilities of operation in fixed beds, tank batteries and batch reactors have been evaluated. Based on simulations carried out for fixed-bed reactors with piston flow, high residence times (order of hours) are required and the required bed length offers an unacceptable pressure drop. Pumping the fluid at high pressures implies the compression of the catalyst. In addition, the acidity of the medium must be adjusted. In fixed-bed reactors, mixing the catalyst with an inert filler is required, increasing the volume of the reactor. The use of a tank battery can solve the problem of load loss, being a configuration usually used in the operation with this type of resins in metal recovery. However, the exchange kinetics in these processes is fast. Simulations in stirred tank batteries with catalyst retention offer similar results to piston flow conditions when the number of tanks is greater than 15. This, added to the difficulties of separating the catalyst during continuous operation, makes it an option with a high investment and operating cost (pumping requirements and loss of charge in each stage of catalyst recovery). However, the data from the tank battery simulation allows an approximation of the amount of catalyst required for a given residence time. This value is used as an initial iteration in the sizing of the reactor. The batch operation is proposed as a solution for hydrolysis. The storage of the concentrated GGM and the hydrolysate produced in buffer tanks allows the transition from the discontinuous operation (hydrolysis section) to continuous operation (rest of the plant). There are multiple reasons for selecting this type of reactor. Considering the deactivation observed during the use of acid catalysts in similar reactions 66,67,68 (mainly esterification), a significant loss of activity (10-20%) is observed after 5 reaction cycles. Given the high life span of these catalysts (15-20 years), the regeneration of the catalyst is necessary. Batch operation offers the advantage of operating with multiple reactors, optimizing the times to regenerate the catalyst in one of them while the rest remain in operation, loading or unloading. Because of this, a "slurry" reactor is selected where an agitated tank is used to contact the solid catalyst and the liquid. The reduced density of the catalyst allows for easy dispersion. The high degree of mixing, added to the slow kinetics of the reaction, allows operation under kinetic control conditions, where the transfer of matter is not 51 significant in fibrous or macroporous materials . The discontinuous operation provides a greater flexibility to the process, where variations in the composition of the hemicellulosic fraction to be hydrolyzed can be quickly adapted by adjusting the reaction times and operation cycles. In addition, the discontinuous operation allows to avoid problems derived from the pressure drop. A crucial consideration for the design of the equipment should also be mentioned in this section. Despite the advantages offered by fibrous catalysts such as Smopex®-101, documented evidence of swelling phenomena and lack of information on their physical characteristics during hydrolysis (density, water content, porosity...) prevent reliable design of process equipment. This fact adds to the problems observed with long filtration times due to the possible agglomeration of fibers under pressure. Consequently, based on studies at Åbo Akademi University , a different catalyst with similar properties to Smopex®-101 is selected where the parameters of the catalyst are available, and both the activity and the matter transfer phenomena are comparable. The catalytic activity will be extrapolated to that studied in Smopex®-101 considering the acidic capacity of the new resin, defining the volumetric catalyst load to obtain the same concentration of acid centers. The new catalyst is evaluated between Dowex® 50W-X8 resins with a particle size of 0.15-0.3 mm in diameter and Dowex® HCR-W2 with a larger capacity and sizes of 0.42-1.2 mm (diameter). The slightly higher capacity of the second resin, its high thermal resistance (allowing operation at 120 ºC) and its recommended use for ester hydrolysis processes lead to the selection of this second material as a catalyst. The main disadvantage compared to Smopex®-101 is the higher stirring power required for dispersion in the reactor. the use of small-sized catalysts, the recovery of the catalyst is planned in three stages. After the reaction, agitation is stopped, allowing solid-liquid separation by density difference. Subsequently, there are two discharge intakes. A first tangential outlet at a enough height above the deposited catalyst. Most of the liquid in the tank is pumped through this. A second discharge point is located at the base of the tank, where a filtering medium is placed. The filtration of the remaining liquid allows the recovery of most of the hydrolysate, where a marginal amount remains inside the tank together with the Based on unpublished studies at the time of submission of this paper. 52 catalyst. After this process, the system is ready for a new reaction cycle or regeneration of the catalyst. The filtration of the hydrolysate is aided by the operating pressure, acting as a driving force in conjunction with gravity. Due to the future implementation of materials from the Smopex®-101 family with reduced pressure drop68 this mode of operation is considered a possibility for future operations. After the modification of the catalyst used in the process, the high particle size allows the total recovery of the liquid by means of particle retention with perforated metal filter. The regeneration of the catalyst is carried out with 2% sulphuric acid and 14%v/v solid concentration according to the manufacturer's recommendations. 68 After a regeneration time of 10 hours, it is considered that the resin activity has been fully recovered. Published studies observe increases even over the initial activity. After regeneration, the excess acid is removed by rinsing the equipment with two partial water loads, up to In addition, objective conversion should be considered. Based on previous simulations with continuous stirred tank batteries (CSTR) and in piston flow beds, an initial latency period, a region of rapid hydrolysis (pseudo-first order) and a region limited by low reagent concentration are observed. Figure 16 shows the conversion against reactor volume for an arbitrary flow fed to the process concentration. Given the similarity between batch operation and continuous piston flow operation, this simulation is used to determine the target conversion. Based on the results obtained, a conversion for the GGM of 90% is set. The slight increase in conversion is not compensated for by the significant increase in operating time (or residence time). Likewise, during the subsequent catalyst-hydrolyzed separation the reaction will continue to take place, allowing a slight increase in the final conversion. Unreacted GGM is not a hazardous component. Despite its possible recovery by means of membrane technology, the recovery costs are higher than the benefits derived from its recirculation to the process, therefore it will be emitted together with the process wastewater stream. Volumetric ratio 1:6 catalyst - regeneration solution Based on unpublished studies at the time of submission of this paper. 53 5.3.2 Material balance and dimensioning of the reactor The reactor can be sized on the basis of the reaction time, which can be calculated from the material balance in the reactor. As the reaction is in a liquid phase (constant volume), the reaction time will depend on the concentration of the feed and the amount of catalyst. The volume of the reactor will be defined by the amount of liquid to be processed per batch. Apart from the reaction time, the loading and unloading times of the product are calculated. For catalyst conditioning, a contribution of 20% of the regeneration time (equal distribution of the regeneration time between the number of cycles between each regeneration stage) and flushing time is considered for each operating cycle. The general material balance equation for a batch reactor (equation 5.1) is considered, where r represents the total molar generation ratio in the system. The production rate of each compound from the available kinetic model (combination of equations 3.5, 3.10 and 3.11) and the total volume of catalyst in the reactor. The detailed calculations of the material balance are given in Annex D, where the other equations used in the resolution are expressed below. Material balance in a batch reaction system = − [5.1] Production ratio of monosaccharide "i = ∙ ∙ (1 + ) ∙ ∙ ∙ ∙ [5.2] Figure 16. X-V curve of the reactor. Piston flow simulation. 54 Initial concentration of non- hydrolyzed monosaccharide fractions in the GGM = ( ∙ ) ( ∙ ) ∙ ∙ [5.3] Degree of polymerization as a function of the extent of reaction = − ∑∆ ∑∆ [5.4] Catalyst volume as a function of reaction volume per batch = (1 − ) ∙ [5.5] Molar consumption of non- hydrolyzed monosaccharide fractions = − ∙ ∆ [5.6] Molar production of monosaccharides in the hydrolysate ( ) = ( ) + ∙ ∆ [5.7] Molar water consumption during the hydrolysis reaction = − ∙ ∆ [5.8] Enthalpy balance = + ∆ ∙ (∑∆ ) ∙ (180,16 − 18) [5.9] Figures 17 and 18 show the results of the material balance in the reactor under the above-mentioned reaction conditions. With an initial concentration of 60 g L-1 of GGM, a volumetric catalyst load of 8.1%, at an initial temperature of 120 ºC and 2.1 atm of pressure, gives a result of 1660 minutes as reaction time for a 90% conversion. The reaction shows an increase of 1.23 ºC during the hydrolysis reaction under these conditions. Figure 17. X-t curve for the hydrolysis reaction in a batch reactor. Figure 18. T-t curve for the hydrolysis reaction in a batch reactor. 55 Based on the result obtained, the number of annual operating cycles is calculated as 290, with a volume of liquid per batch of 215 m3 (234 m3 considering the catalyst). Given the high volume to be processed per batch and the reduced cost of the batch-agitated tank type reactors, parallel operation with 4 agitated tanks is possible. Consequently, the reaction volume in each tank will be 58.5 m3. This decision is justified due to the lower cost associated with conventionally sized equipment. Likewise, operating at lower reaction volumes allows the effect of possible failures in one of the equipment to be mitigated. The number of cycles used in the calculation considers the loading, unloading and conditioning times of the reactors to be negligible, allowing the sizing of the equipment to be carried out. The above calculations define the final information obtained after the simulation and optimization of the system. The reactor can be sized based on the volume. It was decided to use a reactor with a Klopper (torispherical) bottom, in accordance with DIN-28011. The use of rounded bottoms facilitates the dispersion of the solid and reduces the probability of the formation of dead volumes in the system. Likewise, a relationship between the filling height and the diameter defined in equation 5.10 is considered, in accordance with the typical dimensions for stirred tank type reactors72,73. To calculate the total volume of the reactor, a safety factor of 25% is taken into account for the tank height (DIN-28110 standard). The filling volume will correspond to 75% of the tank height. Equation 5.11 allows the calculation of the tank volume with the previously defined specifications. = 1 [5.10] = ∙ 0,0314 ∙ + ∙ 4 ∙ 0,8065 [5.11] By solving equation 5.11 for a fill volume of 58.5 and applying the 25% safety height correction, the tank height and diameter are obtained. Annex D contains the sizing calculations. The equipment is finally designed with the measures set out below, where the filling design capacity is 77% of the total volume of the equipment. The remaining volume serves as a margin for the accommodation of auxiliary equipment (agitation, sensors, etc.), as well as safety volume. = 7,65 = 6,1 = 78,3 = 58,5 56 The reactor loading and unloading times are then defined according to equations 5.12 and 5.13 For the reactor discharge, the phenomena of sedimentation of the catalyst and external flow of the hydrolysate in the bed formed by them are considered. Knowing the dimensions of the reactor, it is possible to calculate the dimensions of the catalyst bed formed when the system is stopped. The calculated height is 0.71 m. The pressure drop data provided by the manufacturer70 are studied to determine the influence on the discharge. For the discharge flow established at 0.358 m3 min-1 , the pressure drop is practically nil, due to the width of the bed and the reduced linear speed of emptying the equipment (2 m h-1 ). The result of the calculation (see Annex D) gives the following results. The catalyst settling time is calculated considering the prevented settling of the catalyst, such as spherical particles with an average diameter of 0.0008 m. The terminal velocity is 0,36 m min-1, with a consequent settling time of 17 minutes. Based on both results, the reactor discharge time will be considered as the maximum possible, corresponding to the sum of the total sedimentation time and the fluid discharge time through the catalyst bed. However, the discharge operation shall begin immediately after the agitator has stopped. The fluid height shall be reduced over time, and the initial pressure drop during fluid discharge shall be less. Therefore, the actual discharge time of the product will be slightly less than estimated. Thus, the total discharge time of the reactor is estimated at 167 minutes. The load flow will be 0.385 m3 min-1. The loading operation will take 150 minutes. = ( ) [5.12] = ( ) + [5.13] Figure 19. Pressure drop (20°C) in beds operated with DOWEX® HCR-W2 Resin. 70 57 Although the reaction time is high, the loading and unloading times of the product are not negligible for the initial value of the number of operating cycles. Considering the loading and unloading times, as well as the kinetic aspects derived from heat dissipation, the real number of annual operating cycles is 233, increasing the total volume to be processed per cycle to 268.4 m3. This value can be achieved by increasing the number of battery tanks to 5, while maintaining the previously exposed dimensions for the system. The regeneration time is also not negligible (with a contribution of almost 3h extra per reaction cycle). For this reason, the possibility of operating with two 5-tank batteries is being considered. Where one of the batteries remains in regeneration, while the second is used for the reaction. The dead time present after regeneration allows a certain flexibility to carry out maintenance and cleaning operations of the equipment. The advantages of operating with two batteries are also discussed in the section on energy integration. The material selected for the construction of the reactor is AISI-316 stainless steel, due to its resistance to corrosion. The reagents are not corrosive, however, when working with aqueous solutions at temperatures above 100 ºC and with an acid pH in the reaction medium, a steel with high resistance to corrosion is required. The properties are shown in table 10. Also, since the reaction is carried out in an adiabatic regime, the reactor must be isolated from the environment, a layer of glass fiber is selected as the system's insulator, due to its inert nature and reduced thermal conductivity. Composition Maximum Effort (MPa) Limit Elastic (MPa) Elongation Maximum (%) Hardness Maximum (HB) Thermal conductivity (W m-1 K-1) C (%) Cr (%) Mo (%) Ni (%) < 0,07 16,5 - 18,5 2 - 2,5 12 565 248 55 149 16,3 Finally, the reactor thickness is calculated, following the ASME standard, Section VIII, Division 1, for the design of closed process vessels under pressure. 77 The calculated thickness (0.85 mm) is below the minimum threshold recommended by the standard, After recalculating the material and energy balance considering the heat loss (section 6.2.6), the reaction time is 80 minutes longer than calculated in the initial material balance. Table 10. Properties of AISI-31676 steel 58 consequently and due to the industrial availability of the material, a thickness of 1/16 inch (1.6 mm) is selected. Table 11 shows the calculated times for loading, unloading, regeneration and reaction operations of the reactor. Detailed calculations and considerations are given in Annex D. Parameter Value Parameter Value HL (m) 6,1 (min) 1740 (m) 7,65 (min) 150 (m) 6,1 (min) 167 (m3) 78,3 (min) 800 = (m3) 58,5 Reactivity ratio. - Regenerating. 5:1 = ( ) (m3) 53,75 Reactors per cycle 5 (kg) 5.781 Annual Cycles 233 e (mm) 1,6 Isolation Fiberglass 5.3.3 Fluid dynamics. Agitation and transport of matter. Given the kinetic characteristics of the process, catalysis is not limited by matter transfer phenomena even under low turbulence conditions. The relatively slow kinetics of the process control the reaction, where catalysis takes place almost exclusively on the surface of the catalyst, without encountering material transport limitations. Studies carried out at Åbo Akademi University, continuing the line observed in references [16, 17], for the selected catalyst, have proven this phenomenon. Due to this, the fluid dynamic study of the system is limited to the design and dimensioning of a suitable agitation system, which allows the homogeneous distribution of the catalyst particles in the reactor. The design of an adequate agitation system implies making a high number of choices of the elements of the system, where any change in the type of elements, their Considering the catalyst once hydrated under reaction conditions. Table 11. Main parameters of the reactors in the hydrolysis unit 59 arrangement or dimensions will produce a different flow pattern in the system. In addition to the flow pattern, the viscosity of the working solution will define the type of agitator and its speed range. Since the working solution reduces its viscosity during hydrolysis, but when operating at high temperatures it is reduced and remains in the same order of magnitude, the initial viscosity will be used for the selection and design of the agitation system. In solid-liquid systems where the objective is the dispersion of the catalyst particles, axial flow is the preferred component. Also, the formation of vortices that dissipate the stirring energy without contributing to the mixing must be avoided. For this reason, an agitator type with a high axial component is selected; a high efficiency pitch blade turbine, with the blades inclined with respect to the plane of the agitator (see figure 20). The installation of 4 deflector walls is also considered, which increase the flow turbulence and prevent the formation of vortices in the system. The typical dimensions for an agitation system are shown in figure 21. These ratios will be used to size the agitation system. Table 12 shows these parameters for the defined system. It should be noted that given the unitary relationship between the diameter and height of the tank, only one agitation element will be required. Likewise, the suspension of solids is favoured when the agitator is close to the bottom of the reactor72 . Figure 20. High efficiency pitch blade turbine for solids suspension. 60 DT = HL (m) 6.10 E (m) 0,87 DA (m) 2.03 L (m) 0.51 J (m) 0.51 W (m) 0.41 In addition, a fundamental parameter in the suspension of solids is the agitation speed. Depending on the type of agitator and characteristics of the system, there will be a minimum speed that allows the dispersion of the particles without any of them remaining static on the bottom. The calculation of the minimum speed is made in accordance with correlation 5.19.73 where the available empirical parameters "" and "" correspond to the use of simple turbine agitators, with a lower efficiency in the dispersion of solids. Therefore, this minimum agitation speed is considered to be the design speed of the agitator that guarantees the homogeneous dispersion of the solid. ( . ) = ∙ ∙ , ∙ − , ∙ , ∙ , ∙ , , ∙ , α = 1.4 ;β = 1.5 (S. I. units) [5.19] Figure 21. Parameters and dimensions of a standard agitation system.73,79 Table 12. Parameters and dimensions calculated for the agitation system 61 The minimum calculated speed is 38 rpm, since standard gearboxes for stirring offer stirring speeds of 37 and 45 rpm78 , a stirring speed of 37 rpm is selected, since the suspension efficiency of the selected type of stirrer is considerably higher than those of turbines with flat blades defined by correlation 5.19. Later, the agitation power is calculated from the power number (Po) (equation 5.20) , which is obtained through empirical correlations of this against Reynolds' numbers (Re) (equation 5.21) and Froude (Fr). This second term is considered negligible, as it measures the relationship between gravitational and inertial force of the fluid, considered due to the use of deflector partitions and the elimination of the vortex. After calculating the stirring Reynolds, the stirring of the system in turbulent regime is confirmed (Re = 272,475), since in values of Re > 10,000 the value of Po is constant, the number of Po is calculated from the values documented for the type of stirrer selected73 . In this way Po = 0.21 for the designed agitation system. From the system Po and equation 5.20 the maximum stirring power required is calculated. Finally, a stirring motor efficiency of 70% is considered, calculating the electrical power consumed by the stirring system. = 0,21 ∙ 965 ∙ 37 60 ∙ 2,03 0,7 = 2.340 The type of agitator is crucial in the exposed agitation system. The initial design of the system considered a standard pitch blade stirrer with 4 blades at 45º, where Po = 1.27 and a final agitation power of 14.4 kW. Due to the inefficiency of the system, it was considered to revise the initial selection of the agitator type. 5.3.4 Thermal exchange and energy integration The heat exchange of the process is considered from two sides. Firstly, the design of the auxiliary heat exchange equipment is studied. Both for the preheating of the feedstock and the medium pressure steam heater. Secondly, the heating needs of the Page 577, figure 7-15, for a high efficiency agitator. = ∙ ∙ [5.20] = ∙ ∙ [5.21] 62 reactor are studied. For this purpose, the heat loss of the reactor during the reaction time is estimated, evaluating the insulation needs. The detailed calculations are set out in Annex D. For the preheating and heating stages the use of shell-and-tube heat exchangers is selected because of their efficiency and moderate size. The supply temperature is defined as 30ºC, considering that the fluid stored in the buffer tank, despite its isolation, loses part of its heat to the outside environment. Equations 5.22, for the calculation of the required heat and 5.23, 5.2479 for the design of the exchange area in a shell-and-tube type exchanger operating in countercurrent are used. According to the usual design heuristics, the hot fluid will in any case flow through the tubes and the cold fluid through the casing. 79 Medium pressure steam (10 atm, 185 ºC) is used for the final heating. Finally, a minimum effective heat exchange temperature variation between the hot and cold fluid of 10 ºC is considered. The overall heat exchange coefficients (U) are obtained by considering the values documented in the literature for shell-and-tube type exchangers as a function of the fluids. 80 The values shown in Table 13 are considered. Fluids in heat exchange Value (W m-2 K-1) Water - Water 800 Steam condensing - Water 1200 After carrying out the relevant operations, an exchange area is obtained in the exchangers of 241.75 m2 for the casing-tube preheater and 20.84 m2 for the steam heater. The required steam flow will be 0.115 kg s-1. = ∆ [5.22] = ∙ − − − ln − − ∙ = ∙ ∆ ∙ [5.23] F = ( )∙ ( )∙ with R = = ; B = [5.24] Table 13. Global heat exchange coefficients 63 Also, based on the standard dimensions according to the TEMA standard for the selected exchanger type (type E, single flow through casing and tubes), the use of tubes with an external diameter of ¾ in (19.05 mm), baffles with a cut-off section of 25% of the casing area is considered. The tube arrangement shall correspond to a triangular pitch of 1.25 in (31.75 mm). 83 The material shall be AISI 316 steel. The pressure drop is also defined on the basis of the maximum values commonly found in this type of equipment. Due to the small size of the steam heater, the pressure drop is considered as an overall value for both heat exchangers. 84 Table 14 shows the characteristics of the E-202, E-203 exchangers. Parameter E-202 (A-I) E-203 (A-I) Exchange rate83 T -E T - E Exchange area (m2) 242 21 Cold fluid flow (kg s-1) 5,97 5,97 Hot fluid flow (kg s-1) 5,97 0,115 Heat flow (kW) 1934 242 Pressure drop (kPa)84 70 Heat losses from the reactor are studied considering only the heat loss across the cylindrical section of the equipment. Internal convection and conduction to the reactor wall are considered negligible in relation to the resistance offered by insulation and the natural convention with the outside air. First, the critical radius of the insulator is determined (equation 5.25), checking its possible use (rc = 3.5 mm). The heat loss is evaluated by considering the transverse conduction through the reactor walls to the outside. The combination of heat transport resistance (conduction in the insulation and natural convection in the air) is evaluated by equation 5.26.81 The ambient temperature is considered as the annual average (6 ºC, see Chapter 8). The coefficient of natural convection in air is defined as 9.4 W m-2 K-1, according to the empirical values obtained in vertical cylinders under natural convection circumstances. 82 Table 14. Characteristics of the reaction section exchangers 64 After the simulation of the heat loss, without insulation, a sharp drop in temperature of up to 86 ºC is observed during the operation time. Consequently, due to the thermal sensitivity of the reaction, the need to insulate the reactor or provide heat is assessed. After the simulation of the process, the use of a 20 cm thick insulation jacket is considered, which allows operation in a practically isothermal regime. This unusually high thickness is decided upon due to the low cost of the insulation compared to the implementation and operating costs of the installation of a heating system in the reactor. Figure 22 shows the thermal profile in reaction mixture time for an insulation thickness of 0.2 m. In view of the results obtained, the temperature profile obtained is used to correct the reaction kinetics in the reactor model, obtaining greater precision in the operating time required. Finally, it is worth mentioning the energy integration aspects of the process. Firstly, the integration of the unit in the TMP processing plant allows the use of the surplus heating steam generated in the facility. Likewise, within the designed section, the need to recover the heat from the mixture after the hydrolysate is discharged is considered. In spite of being at a low temperature (120 ºC), the volume of solution is very high and carries a large amount of energy. Energy integration is implemented by = ℎ [5.25] . = ( − ) ln 2 ∙ ∙ ∙ + 1 2 ∙ ∙ 2 + + ∙ ∙ ∙ = 2 + + ; = 2 + [5.26] Figure 22. Temperature of the reaction medium with heat dissipation 65 crossing the output stream with the new supply to the reactors, allowing the recovery of a large part of the energy used in heating. As mentioned in this chapter, two 5-reactor batteries are used, so that while one opera is being discharged, the second operates in charge, thus allowing the energy integration of both streams. During plant start up and shut down, the use of accessory heat exchangers should be considered to supply/remove the heat normally exchanged at this integration stage. The regeneration of the catalyst is carried out at lower temperature and pressure. Therefore, after the regeneration of the catalyst, the heating of the food must be done with an auxiliary system. Due to the shorter time required for regeneration, the use of the reactor half-shell or an additional auxiliary heater is considered. 5.3.5 Pumping equipment. The fluids used in the reaction section are controlled by the P-201 A-I and P-205 A-I pumps. Due to the handling of high flows of fluids with a reduced viscosity (< 100 cP), the use of centrifugal pumps is considered, also considering their low cost and standardization. Each pump in the P-201 / P-205 sets can work independently, however, in the usual operation they act synchronously, operating 5 at a time for the loading / unloading of the reactors. The pumping power required by both pump assemblies for reactor loading and unloading shall be calculated. The fluid is admitted by means of a pipework in the upper region of the tank (above the filling height). Discharge is from the bottom of the tank, under conditions of external flow over the catalyst particles settled at the bottom of the tank. In this case the maximum possible pressure drop at the point of tank discharge shall be taken into account. The loading pump is placed prior to the heat exchange, pressurising the flow to 2.1 atm, allowing operation in the liquid phase at the reaction temperature. The pumping power calculations are made by applying Bernoulli's equation (5.27) and approximating the friction losses by means of the Fanning (5.28) and Colebrook-White (5.29) equations. The length of the pipes is estimated considering the safety distance between the process equipment. However, after the detailed design of the plant layout, it will be necessary to recalculate the power of the pumps based on the length and dimensions of the pipes used. 66 Pressure drop in heat exchangers is included as an extra term. To determine the electrical power consumed in pumping, a total efficiency of 65% is considered. The diameter of the pipes is defined according to the design heuristics set out in equation 5.3079. According to the above equations, a diameter of 4.28 in. A diameter of 4 in is selected according to the nearest industry standard. According to the flow rate and the diameter of the pipes, the velocity of the fluid will be 0.64 m s-1 , within the usual range (0.5 - 5 m s-1 ) for liquid impulsion. According to the calculated value of Re at the initial (30º) and final (120ºC) viscosity (Re of 2320-7225), the flow regime in the tube develops in laminar and transition flow (Re < 10,000 therefore, a1/2 = 0.5). The length of the pipes is estimated as the sum of the safety distances between the equipment involved. Due to the heat exchange between the discharge and output streams, both are considered equal. For reasons of simplicity, the main line section is considered as part of the individual feed streams to each reactor. The total length of pipe is estimated at 120 m. The roughness is considered to be 0.046 mm. 79 Likewise, accidents in the pipes are estimated as equivalent length referred to the diameter79. The flow division is considered to be the contribution of 8 T-joints, in addition to 20 elbows at 90º necessary to redirect the fluid. Both the length and the accidents are considered as half for the reactor loading section, and the remaining half for the unloading. The pressure drop in both sections is established as symmetrical (however, the feed will have a somewhat lower loss in the heat exchanger due to its flow through the casing). 2 · − 2 · + ( − ) + ( − ) + ∆ + = [5.27] = 2 · · · [5.28] = 1 −2 log 3,7 + 2,51 [5.29] ( ∙ ) = 4 ∙ = 1 + 12 ∙ 6 ; in. [5.30] 67 The discharge/loading into the buffer tanks is considered at the level of the base of the tank. The hydrostatic pressure due to the fluid column in the tanks is modeled considering the filling level, simplifying it as an average of 5 meters, however, the driving power will vary slightly as the tank level is reduced. Equation 5.27 applies between the upper surface of the fluid from the tank (T-200) or reactor (R-204) and the final discharge point in the reactor (R-204) or tank (T-206). The results of the pumping power calculations are shown in table 15. Details of the calculation are given in Annex D. Parameter P-201 (A-I) P-205 (A-I) Pump type Centrifuge Centrifuge Flow rate Fluid (kg s-1) 5,97 5,97 Theoretical Power (kW) 2,94 1,9 Electrical Power (kW) 4,52 2,92 Suction Pressure (kPa) 150,3 262,4* Pressure Discharge 202,6 150,3* 5.3.6 Instrumentation and control. The adequate operation of any chemical plant depends to a great extent on the design of the control and instrumentation of its equipment, allowing it to operate within the established operating margins. The following chapter describes the control and instrumentation study for the system reaction unit. Due to the parallel operation with a battery of tanks, the study of instrumentation and control is carried out on a single reactor, considering the implementation of the same system in the rest of the equipment in parallel. Reactor control and instrumentation. The reactor operates in discontinuous mode, therefore once the load is done it is considered a watertight system where the temperature and pressure of the system must be controlled. Likewise, the implementation of instrumentation for the measurement of glucose concentration is considered Pressure relative to the height of fluid in the discharge tank. It is considered an average value of fluid height. The pumping power shall be controlled to maintain a constant flow rate. Table 15. Characteristics of the reaction section pumps 68 The temperature of the reactor must be controlled due to its high impact on the reaction kinetics. Although initially the isolation of the reactor allows the reaction to take place in an isothermal regime, an alarm system for excessively low or excessively high temperature must be included. Both situations would compromise the reaction (slow kinetics or catalyst degradation). In the case of low temperatures, the implementation of a heating jacket of half a cane is considered, which would allow the punctual heating of the fluid, being controlled by a control loop. This control loop would regulate the heating steam flow in the half pipe as a function of the reactor temperature, acting for this purpose on an automatic valve. The pressure of the reactor is a fundamental parameter, since it allows operation in the liquid phase at temperatures above 100ºC. The installation of alarms in the event of excessively low or high pressure is considered essential. Likewise, the pressure inside the reactor is regulated by injecting air from an auxiliary compressor. A control loop is designed, where in case of pressure reduction, the compressor is acted upon to maintain the design value. No pressure increases are expected in the system, however, in the event of excessive pressure the relief valves would allow the system to be returned to operating pressure. Control and instrumentation of auxiliary equipment. Apart from the reactor, the pumps P-201 A-I, P-205 A-I and the heat exchangers E-202 A-I, E-203 A-I, require control systems to ensure that loading and unloading operations are carried out in accordance with the established values. With respect to the pumps, closed flow control loops will be implemented, where, based on the measured flow rate feeding the reactor, a control valve present after the pump discharge will be actuated. Each pump will be installed in duplicate, allowing operation in the event of failure of one of the units through a bypass. Each line of the bypass will have a flow valve connected to a flow control loop downstream of the bypass (reactor feed). In addition, maximum and minimum flow alarms shall be included. Finally, the pump discharge line shall have a pressure indicator, connected to a high pressure alarm, and a check valve to prevent possible reverse flow The pumps will be installed in duplicate on a bypass, since, in the event of failure of one of the pumps, it will be possible to continue operating through the bypass. The flow control loop has a flow indicator, connected to a flow transmitter that measures and sends an electrical signal to the flow controller, which compares the signal with the 69 setpoint and sends an output signal to the flow valve. Each line of the bypass will have a flow valve connected to a flow controller loop located downstream of the bypass. In addition, two high and low flow alarms are connected to the flow indicator. As for the heat exchangers, maintaining the temperature of the feed to the reactor is a crucial factor for the process. Because of this, a control loop of the feed output temperature is established. This loop measures the food discharge temperature in the reactor, acting on an automatic valve that controls the flow of heating steam fed to the E- 203 A-I heater. In addition, maximum and minimum temperature alarms will be included in the fluid discharge, allowing operators to be alerted of possible deviations in operation. Also, an indicator for the admission of the hydrolysate to the E-202 A-I exchangers will be included, allowing the temperature of the discharged hydrolysate to be monitored. The implementation of these systems is reflected in Annex E. 5.4 Plant layout. Layout In this section, the distribution of the different process equipment is determined. Although this report mainly deals with the design of the reaction section, the layout of the pre-treatment and product purification sections will also be outlined. The objective is to determine the adequate distribution of the different process areas, as well as the optimal location of equipment and auxiliary services. The spatial arrangement and relative distances between the equipment obey passive safety, economic and operational criteria. A correct distribution of the equipment increases the safety of the plant, prevents damage in the event of an accident and optimizes the use of space while guaranteeing the operability of the installations. 84,85 The proposed process is designed taking as an example its implementation in a previously existing industrial facility. The distribution diagram considers the implementation of the process section on the ground in a clear area of land, close to the mentioned facility. However, typical areas in any facility, such as parking area, laboratories, control rooms, offices and common facilities, in addition to the electrical services, heating and effluent treatment sections, are considered as part of the previously existing facility. Because of this, the previously mentioned areas are not included in the plant layout. The distribution of the process areas and equipment is carried out according to implementation heuristics, considering safety criteria and spatial layout of the equipment. 70 Likewise, a representation of the access roads and evacuation routes in case of emergency is included, arranged appropriately. The main elements of the layout include storage section for reagents and products, the pre-treatment, reaction and purification sections. Also included are pipe racks, fire safety and fire defense equipment, access routes, and the layout of bulky equipment in each of the process areas. The list of such equipment is included in the table below the diagram. Annex F contains the implementation plan or layout of the plant. 71 CHAPTER 6 Safety Studies. In the following chapter, the process safety analysis is carried out, studying the dangerousness of the compounds used and the risk associated with the different equipment in the reaction section. 6.1 Dow’s fire and explosion index. Given the nature of the compounds handled, there is no high degree of hazard associated with the substances and mixtures in the process. Likewise, the operation takes place at low temperatures (maximum 120º) and moderate pressures (2.1 atm). The compound with the greatest risk in the installation is the acid used during the regeneration of the catalyst, where the appropriate measures will be taken to reduce the possible risk associated with its storage and handling. In order to assess the possible risks associated with the compounds and mixtures during their handling in the process, the fire and explosion rate is calculated for the critical unit of the process, the reactor. The calculation is made in accordance with the 5th edition of the DOW Chemical Company calculation manual.86 Tables 16 and 17 show the calculation of the general (F1) and specific (F2) factors of the process, allowing the subsequent calculation of the fire and explosion rate of the unit. Item Interval Assigned Value Exothermic chemical reactions 0.3 – 1.25 0.3 Endothermic processes 0.2 – 0.4 0 Material handling and transport 0.25 – 1.05 0.5 Confined processing units 0.25 – 0.9 0.5 Access 0.2 – 0.35 0.2 Drainage and leakage control 0.25 – 0.5 0.25 F1 1.75 Table 16. Calculation of general risk factor for the process. 72 Item Interval Assigned Value Toxic substances 0.2 – 0.8 0.2 Below-atmospheric pressure 0.5 0 Operation at or near flammable limits 0.3 – 0.8 0 Dust explosion 0.25 – 0.8 0.25 Very low temperature 0.2 – 0.3 0.2 Pressure relief devices 0.1 -1.13 0.15 Amount of flammable or unstable material 0.15 – 4 0 Corrosion and erosion 0.1 – 0.75 0.1 Leaking seals and gaskets 0.1 – 1.5 0.3 Presence of ovens 0.1 – 1 0.1 Thermal oil heat transmission system 0.15 -1.15 0 Rotary equipment 0.5 0.5 F2 1.75 The material factor of the reaction mixture is also determined. The adiabatic decomposition point and auto-ignition temperature of glucose and epimers are considered, as well as their presence in dilute aqueous solution (60 g L-1). The material factor assigned is 1 (MF = 1), representing low flammability and reactivity of the mixture. Eventually, equation 6.1 is applied to calculate the fire and explosion rate of the unit. The value obtained for the hazard and explosion index is 3.06, reflecting the reduced degree of danger of the operations and substances handled during the operation. 6.2 Análisis de riesgo y operabilidad. Estudio HAZOP. As a main component of the plant safety study, the following chapter presents the HAZOP (Hazard and Operability Study) analysis of the reaction unit. This methodology makes it possible to identify potential risks to the equipment and the process, and to take corrective and preventive measures. Those streams where a significant variation in their characteristic variables (Pressure, Temperature, Flow, Composition...) is found will be studied by means of the use of key words (more, less and not). Those meaningless deviations for a stream will not be considered.87 = ( 1 ∙ 2) ∙ [6.1] Table 17. Calculation of the specific risk factor for the process. 73 HAZOP analysis is performed on the process reactor feed and discharge streams. As it is a discontinuous operation, possible deviations in each stage of the process will be considered. It should be noted that the regeneration stream with dilute sulfuric acid (1%) and wash water is considered as a specific case where the problems derived will be the same as those found for the reactor loading and unloading streams. However, there is an added risk of acid spillage in the event of pipe breakage, which must be considered in order to adopt adequate safety measures for operation with acid solutions during the regeneration stage. Finally, information on risks derived from possible deviations from normal operation will be used for the implementation of the process control and instrumentation systems. The detailed analysis by streams is included in Annex G. The results are reflected in the measures implemented in the instrumentation and control of the process. 74 CHAPTER 7 Environmental study. All industrial activity has an impact on the environment where it takes place. The study of the environmental impact is the ideal tool to identify the different components of this impact, as well as its magnitude. In this way, the environmental impact study is an indispensable tool in the decision-making process when faced with the different possible alternatives for the development of the target activities. The aim of this chapter is to make a first general evaluation of the different sources of pollution and environmental impact of the process. In this way, adequate knowledge is obtained for the selection of those options that safeguard the general interest and purpose of the installation. 7.1 Plant location and Environmental legislation. The selected location is the municipality of Hallstavik, within the municipality of Norrtälje, County Stockholm, Sweden. The installation is planned as an integrated unit within the paper mill belonging to the company HOLMEN, located in the region. Figure 23 shows the above-mentioned facilities. Figure 23. Helicopter view of Holmen's plant near Hallstavik. 75 The Hallstavik region is mostly flat, with an area of 4.69 km2 . The average height of the terrain is 31 meters above sea level.88 The region's climate is continental wet Dfb, according to Köppen's classification. Rainfall in Hallstavik is significant, with an annual average of 536 mm and rainfall of over 30 mm even during the driest month (February - March). The average annual temperature is 5.7 °C.89 In addition, there is a moderate wind all year round, mostly from the south. The average wind speed is 10 m s-1. The predominant soil type is glacial till.90 The plant is located 10 km from the Åland Sea and is situated on one of the shores of the sea inlets on the land. Figure 24 indicates the wind in the region, Figure 25 shows the exact location of the plant. Figure 24. Annual wind distribution in the Hallstavik region Annual wind hours by direction and wind speed.89 Figure 25. Geographical location of the TMP facilities. 76 On a legislative level, the main environmental directive is the Swedish environmental code (SFS 1998:808), which has been in force since January 1999. The code unifies 15 central acts for environmental regulation, constituting the guiding document in this respect. Each act in the code details the legal provisions and rules applicable to each area. Likewise, multiple modifications of the code have taken place until now, with the objective of adapting to European objectives and those of the country itself, becoming more restrictive and precise in terms of environmental regulation.91,92,93 For aspects related to waste treatment, the European Waste Framework Directive (2008/98/EC) has been adapted within the Swedish environmental code. Likewise, the waste ordinance (2011:927) covers the legal aspects of waste management and disposal in Sweden. 94 7.2 Environmental Impacts. The impact of the process on the environment can be studied by distinguishing according to the different environments likely to be affected by the activities carried out. The identification of potential environmental risks and corrective measures is based on an integrated analysis of all the elements involved in the proposed process, from its construction to its normal operation. This analysis is based on previously published literature.95,96,97,98,99 7.2.1 Soil and land impact. Aspects related to land use are dealt with in the Swedish environmental code in a fragmented way among the different acts that form it. The Agriculture and Land Management Act and the Natural Resources and Environmental Damage Act are the main documents. The waste management ordinance, as well as the new adaptations to the environmental code adopted in 2020, also set out the directives governing the environmental impact on soil. The main impact on soil will take place during the construction phase. Excavation work, foundations and the continuous passage of heavy machinery will lead to soil compaction and the destruction of the soil's surface horizon, altering its properties. 77 During the normal operation of the plant, the main impacts on the soil will be the occupation of the land by the facility itself, as well as the potential risk of spillage and infiltration of working solutions and products stored in the process tanks. However, due to the location of the process line within a previous installation, the pre-existing infrastructures can be used for the transport of materials. In addition, the site of the installation has been previously conditioned. Likewise, the nature of the compounds used in the plant does not present a potential risk to the soil in any of the cases. No obstante, debido a la ubicación de la línea de proceso dentro de una instalación previa, pueden emplearse las infraestructuras preexistentes para el transporte de materiales. Además, el terreno de asentamiento de la instalación ha sido previamente acondicionado. Así mismo, la naturaleza de los compuestos empleados en la planta no presenta un riesgo potencial para el suelo en ninguno de los casos. 7.2.2 Impact on water. The water act, as well as the waste to water act and the environmental damage act, of the Swedish environmental code consists of the main legislation regarding water pollution. It states that any operation requiring water operations may only be carried out when the benefits of the operation from a unified public-private interest point of view are greater than the potential damage associated with the operation. Likewise, access to and use of water in Sweden is controlled by different public institutions, where the SwAM (Swedish Agency for Marine and Water Management) is the main body. The potential alterations to the aquatic environment will take place on the groundwater and surface water courses up to its mouth in the sea. The environmental impact is due to the contribution of allochthonous materials that can modify the quality of the water, that is, the contribution of materials that are not naturally present in the waters of the region. The greatest impact will take place during the construction works of the facility. The main sources of pollution are the dispersion of dust, fine particles in suspension or even construction waste that eventually end up in the surface water altering its solid content and related properties. Special attention should also be paid to the possible dragging of cement and oils into watercourses. The strongly alkaline nature of cement 78 can increase the pH of the aquatic environment, severely affecting the organisms present and potentially making it unsuitable for life. On the other hand, oils are a widely studied potential source of pollution, reducing the transfer of oxygen between water and air, and even causing the drowning of small animals. During the normal operation of the plant, a minimum impact on the water environment is expected. Huge amounts of water are used during the process where, almost exclusively, the water comes from a secondary stream from the adjacent pulp mill. In addition, much of the water is subsequently recovered in the form of steam, which can be used for various purposes within the process and finally after being partially condensed and recirculated as process water. The main aspect to be considered is the thermal load of the process flows. Given that prior to emission into the environment, the temperature must be reduced to avoid an impact on the surroundings of the plant. Likewise, the waste flows of the process are mainly composed of water, unreacted galactoglucomannan and other compounds from the wood. Small amounts of bases (sodium hydroxide) and acids (mainly acetic) can also be found in the process effluent. However, given their nature and degree of dilution these substances do not pose a potential environmental risk and are practically harmless. Any solids removed during the process water filtration and conditioning stages can be used in the plant's existing energy recovery facilities. 7.2.3 Atmospheric impact The potential alterations to the aquatic environment will take place on the groundwater and surface water courses up to its mouth in the sea. The environmental impact is due to the contribution of allochthonous materials that can modify the quality of the water, that is, the contribution of materials that are not naturally present in the waters of the region. The greatest impact will take place during the construction works of the facility. The main sources of pollution are the dispersion of dust, fine particles in suspension or even construction waste that eventually end up in the surface water altering its solid content and related properties. Special attention should also be paid to the possible dragging of cement and oils into watercourses. The strongly alkaline nature of cement can increase the pH of the aquatic environment, severely affecting the organisms present and potentially making it unsuitable for life. On the other hand, oils are a widely studied 79 potential source of pollution, reducing the transfer of oxygen between water and air, and even causing the drowning of small animals. During the normal operation of the plant, a minimum impact on the water environment is expected. Huge amounts of water are used during the process where, almost exclusively, the water comes from a secondary stream from the adjacent pulp mill. In addition, much of the water is subsequently recovered in the form of steam, which can be used for various purposes within the process and finally after being partially condensed and recirculated as process water. The main aspect to be considered is the thermal load of the process flows. Given that prior to emission into the environment, the temperature must be reduced to avoid an impact on the surroundings of the plant. Likewise, the waste flows of the process are mainly composed of water, unreacted galactoglucomannan and other compounds from the wood. Small amounts of bases (sodium hydroxide) and acids (mainly acetic) can also be found in the process effluent. However, given their nature and degree of dilution these substances do not pose a potential environmental risk and are practically harmless. Any solids removed during the process water filtration and conditioning stages can be used in the plant's existing energy recovery facilities.101 7.2.4 Noise levels and noise pollution. Due to the effects that high noise levels can have, it is to consider the possible noise pollution associated with the process.97 The regulation of noise levels is covered by the adaptation of the European environmental noise directive (2002/49/EC), in the environmental noise ordinance, which is included in the Swedish environmental code. During the construction phase, a high noise impact is expected, where traffic, machinery operation and construction work will be a constant source of noise. However, these noise sources are of a one-off nature, limited to the duration of the construction work. Likewise, there are no dwellings nearby on the site, so the impact of noise pollution is limited to the natural environment surrounding the plant. In the normal operation of the plant, the process itself is a source of noise. The operation of the process equipment, especially pumps, evaporators, heat exchangers and reactor agitation The transport of the final products of the process must also be considered 80 as a source of noise. It is important to note that, given the installation of the process line within the boundaries of the paper plant site, the process noise levels are lower than those generated by the refiners and mechanical disruptors in the main plant. 7.2.5 Visual Impact. The visual impact of the installation is fundamentally determined by the introduction of external elements to the natural landscape, altering its morphology and appearance. It should be noted that, given the construction of the unit in the surroundings of a previously existing installation, the visual impact on the surroundings is very reduced compared to the creation of an installation in a new location. During the construction stages, the presence of vehicles and machinery will be increased, as well as the vegetation and shape of the land surrounding the plant will be altered. Once the hydrolysis unit has been built, quantifying the visual impact is complex, as it involves a wide range of subjective considerations. However, considering the layout in the landscape within the limits of an existing facility, the visual impact of the facility can be considered as a minimum. 7.2.6 Environmental impact matrix. Likewise, after the construction of the facilities, Table 18 summarizes the impact of each of the process units on the different elements of the environment. In general, the environmental impact of the process can be considered very low. However, in the initial stage of construction, special care must be taken with regard to the impact and good practices during the works. For the installation in operation, the main environmental aspect to be considered will be energy consumption from non-renewable sources, with the consequent indirect emission of greenhouse gases. The emission of high quantities of water vapor into the atmosphere, as well as the possible accidental spillage of working solutions, are not due to the nature of the compounds used in the process. 81 Table 18. Environmental impact matrix for the GGM hydrolysis process. Equipments Impact Soil Water Air Acoustic Visual Ultrafiltration Unit + + - - ++ Tanks + + - - +++ Hydrolysis Unit + + + + ++ SMB Unit + + - - ++ Crystallization / Evaporation Unit + - ++ ++ +++ Pumps + + + +++ + Heat Exchangers + + ++ ++ + (-) No impact (+) Low Impact (++) Moderate Impact (+++) High Impact The matrix shown details the relative releNPVce of each of the main elements of the hydrolysis process. The impact on the soil is mainly determined by the needs of foundation and possible vibration of the equipment. The impact on the water environment is defined by the consumption of water in the units and the generation of waste flows. The impact on the atmosphere is measured on the basis of indirect or direct emissions of greenhouse gases, considering the consumption of electricity or heating in the process to be responsible for this emission. The acoustic impact is determined based on the usual noise levels of the equipment used. The visual impact is established according to the size of the equipment. 7.3 Corrective measures. Considering the different environmental impacts mentioned in the previous section, a series of corrective measures are considered to allow the adequate management of the environmental impact of all activities involved in the process. Firstly, special attention must be paid to the environmental impact produced during the construction phase. In order to mitigate the excessive raising of dust from the land, the application of periodic watering of the land is considered in the event of reduced rainfall. However, the climatic conditions of the region (moderate wind and abundant precipitation) are a natural factor that help to mitigate the release of dust and particles, as well as reduce the concentration by dispersion of those particles projected. 82 The acoustic impact of the facility must be addressed through various noise reduction measures. The location of the pumping section and those process units with the highest noise generation should be located, if possible, in the same hall or in an adjacent hall to the main TMP process facility. The concentration of the main noise sources in the same acoustic enclosure will significantly reduce the noise emitted to the outside. The potential risk of discharges and consequent infiltration into the subsoil or their arrival in watercourses must be considered. Thus, preventive measures must be implemented. During the construction phase, rainwater from the work area will be collected and pumped to the wastewater treatment plant (WWTP) already existing in the main facility. Also, if temporary storage of wastewater in waste ponds is required, these must be adequately waterproofed and have adequate drainage to prevent leakage into the ground. Likewise, during the normal operation of the unit, the installation of equipment in confined areas with paved floors is considered. The development of the industrial process inside a building increases the durability of the sensitive process equipment, reducing the risks of breakage or failure, as well as minimizing the possibilities of accidental spillage or leakage of the process compounds. Buffer and product tanks should be located outside the building due to their size. However, they should be properly paved and waterproofed to facilitate the collection of any spillage. 83 CHAPTER 8 Economic Analysis. This chapter evaluates the economic aspects of the process described in this report. The needs and economic viability of the reaction section will be evaluated. However, due to the integration of the process in larger installations, and the need for further treatment before the sale of the product, calculations of economic viability are also presented considering the proposed stages of pre-treatment and purification. 8.1 Investment estimation. Due to the novelty of the process developed, the use of methods for estimating fixed capital based on references is not possible. Consequently, the value of the fixed capital and the required investment are determined using the percentage method. Thus, the investment of the plant is estimated on the basis of the item of equipment of the plant. The cost of the equipment is estimated using two online cost estimators, using the design and material parameters established in chapter 6.102,103 The material selected for the equipment not involved in the rectification section will be AISI-304 steel, for the rest (exchangers, pumps, piping and reactor) will be AISI-316. The volume will be considered sufficient to feed two reaction cycles (540 m3). Table 19 shows the total cost of the process equipment, once its value has been updated according to the CEPCI (Chemical Engineering Plant Cost Index) and equation 8.1. The reference of the estimates is in 2002 and 2014, with a CEPCI of 390.4 and 576.1 while in 2020 the preliminary value available is 596.2104 The change of currency from $ to Euros is made considering the factor corresponding to the year of reference (1,33 $ €-1 in 2014 and 0,95 $ €-1 in 2002).105 = [8.1] Subsequently, the percentages method79,106 is applied to calculate the total value of the fixed assets. The selection of percentages considers the complexity or relative importance of each item in the process. Table 20 shows the value of the different items, as well as the final investment required. 84 Investment Component Item Applied % Cost DIRECT FIXED CAPITAL Equipment 100% 12,191,764 Installation 76% 9,265,741 Piping 33% 4,023,282 Instrumentation & control 13% 1,584,929 Electric installation 9% 1,097,258 Buildings 18% 2,194,517 Land conditioning 10% 1,219,176 Service installation 45% 5,486,294 Land 6% 731,505 TOTAL 25,602,706 INDIRECT FIXED CAPITAL Engineering & supervision. 33% 4,023,282 Building expense. 41% 4,998,623 Legal expenses. 4% 487,670 Industrial benefit. 22% 2,682,188 Contingency expenses. 36% 4,389,035 TOTAL 16,580,800 FIXED CAPITAL 42,183,506 WORKING CAPITAL 89% 10,850,670 2 Average industrial Price found for large-scale acid catalytical resins.107 Updated to 2020 according to CEPCI in 2007 (CEPCI = 595,4). * Updated to 2020 according to CEPCI in 2005 (CEPCI = 468,2). Equipment ID Number Unitary Cost (€ ud-1) Total Cost (€) Loading Pumps P-201 A-I 10 5140.18 50,747 Discharge Pumps P-205 A-I 10 5074.70 50,747 Heat Exchanger E-202 10 223755.17 2,237,552 Steam Heater E-203 10 12491.57 124,916 Reaction Intermediu m Tanks T-201 T-206 2 268646.73 53,7293 Reactors R-204 A-I 10 619893.95 6,198,939 Catalysts2 -- 57,810 kg 2.7 € kg-1 156,087 TOTAL (Reaction) 9,356,281 TOTAL (Pretreatment) 32 1,562,096 TOTAL (Purification)60*3 1,273,387 TOTAL 12,191,764 Table 19. Cost estimation for plant equipment. Table 20. Fixed and working capital cost estimation by percentage method. 85 8.2 Sales estimation The income from the process is calculated from the sales of the products obtained after purification. The price of Man, Gal as crystalline solids of food purity is considered, as well as the standard price of glucose syrups. Likewise, the sale of the total estimated production is considered. Table 21 shows the sales of each product and the total annual income. Product Production (kg yr-1) Sale Price (€ kg-1) Income (€ yr-1) D-mannose 99.5 % 2.202.000 51.00 112,302,000 D-galactose 99.5% 441.000 35.78 15,778,980 D-glucose syrup (90%wt) 815.555 0.73108 595,355 Total 128,676,335 8.3 Production cost estimation. The production costs of the process are estimated in two stages. Initially the main costs are calculated; Raw materials, direct labor, general services and depreciation. The remaining production costs are then estimated as a percentage of the former. 8.3.1 Raw Materials (M1). The main raw material in the process is the GGM solution from the effluent water from obtaining pulp by TMP. Although this flow is an effluent of the main process, the price corresponding to the extraction of the GGM present (see chapter 2 and Annex A), added to the cost of its extraction during pre-treatment, shall be considered. The energy invested in the pre-treatment has been considered within the calculations made. Thus, the cost of the reaction section feed will be 870 Considering the production of 3752 t year-1, the cost of the raw materials (M1) will be 3,264,240 The cost of the regeneration acid and washing water is considered negligible in comparison. 8.3.2 Direct labor costs (M2). Direct labor (M2) is all those plant employees directly involved in the operation of the process. Among the multiple correlations available, Wessel's equation109 (equation 8.2) is used to determine the hourly workforce requirements as men h-1 t-1. (ℎ − ∙ ) = 23 ∙ º ( ∙ ) [8.2] Table 21. Annual sales estimation. 86 The calculation will be estimated considering 2 stages in the production; reaction and purification, given that the labor costs and operability of the pre-treatment are already considered in the previous section, besides being of lesser magnitude. The value of T is 8.32, so considering the annual production of 3,458.6 t of products per year, a total of 28,776 man-hours per year-1 will be required. On this value, a 20% increase is applied considering the complexities derived from the discontinuous batch operation (34531 man-hours year-1). Finally, considering the average annual hours worked in Sweden110 (1740 man-hours year-1), the amount of direct labor required is calculated, being 20 employees. Based on the average salary of the chemical engineer in Sweden111 (55,872 euros year-1), the cost of direct labor is calculated as 1,117,440 euros year-1. 8.3.3 General process services. (M3). The proposed process is included within a larger facility with a positive energy balance. Likewise, by including the proposed unit, the cost of effluent treatment is reduced. Because of this, the cost of general services (M3) will not consider the needs for heating, waste treatment or water. The calculation is based on the main component, represented by the electricity consumption of the agitation systems and pumping equipment. Considering the operating time established at 8,000 h year-1 and the total power required for agitation in the reactors (11.7 kW) and fluid pumping (7.45 kW). The reactors operate in agitation for 85% of the total operating time, while the pumps do so for 15%. Considering the price of electricity (91 MWh-1 , chapter 1), the cost of general services is estimated at 533,506 year -1. 8.3.4 Depreciation costs (M4). The depreciation costs are defined by applying a fixed annual quota based on the total of the fixed assets, considering a minimum period of 15 years of useful life for the installation. The residual value of the equipment is considered to be zero. Consequently, depreciation is estimated at 6.7 % of the total fixed assets. Based on this calculation sequence, the depreciation cost will be 2,826,295 euros per year-1. 87 8.3.5 Total production cost calculation. The calculation of the rest of the items for the production costs is calculated in accordance with their proportionality with the main items, considering the relative importance of each of them in the process used.79,106 Likewise, the operating cost of the purification section is considered as an additional item, using the flow of fluid processed and the cost60 per kg of purified solution (0.099 ' kg-1 , updated in 2020 values by CEPCI). The annual flow of fluid processed is established on the basis of the reactor sizing calculations, being 62534 m3 year-1. Table 22 shows the results of the calculation of the other items, as well as the sum of the final production costs. Item Estimation Method Cost (€ year-1) Raw materials (M1) Section 9.3.1 3,264,240 Direct workforce (M2) Section 9.3.2 1,117,440 Process services (M3) Section 9.3.3 533,506 Depreciation (M4) Section 9.3.4 2,826,295 Indirect workforce (M5) 15 % de M2 167,616 Patents (M6) 5% Sales 6,433,817 Supplies (M7) 0,5% Fixed capital 210,918 Maintenance (M8) 5% Fixed capital 2,109,175 Laboratory (M9) 15% M2 167,616 Technicians & staff (M10) 15% M2 167,616 Rentals (M11)4 0% 0 Packaging (M12)*** 10% Fixed capital 4,218,351 Taxes and charges (M13) 0,5% Fixed capital 210,918 Insurance (M14) 1% Fixed capital 421,836 Downstream (M15) Reference 60 6,190,866 TOTAL COST 28,040,210 It should be noted that sales and administration costs have not been considered in this estimate, as the process is considered to be part of a larger facility. Considering the probable usage of patents during downstream/pretreatment processes. 4 Considered as neglectable since the process develops in a bigger facility. *** Considering that the product is sold in small-medium format. Table 22. Total production cost for the installation. 88 8.4 Economic viability studies. Estimates of the economic viability of the proposed process are complex. Since the proposed process constitutes a unit within a larger plant, with the objective of taking advantage of secondary streams of the main process. Despite this, the economic viability of the process is analyzed taking into account two main indicators; the net percentage profitability (Rnp) (equation 8.3) and the net present value (NPV) (equation 8.4). = ( − ) 0,01 ∙ ∙ (1 − ) [8.3] ; = ( [( − ) ∙ (1 − ) + ] (1 + ) ) − [8.4] Where V, C, and A represent sales, annual and depreciation costs respectively. Ui is the value of the total tax burden, considered as 21.4%112. I represent the total investment of the project; the sum of fixed and working capital. Likewise, the discount rate applied to the estimates is defined as i. The calculation of the Rnp of the installation, is calculated as 149% in 2021, which implies the total recovery of the investment and the obtaining of benefits even in the first year of sales. Likewise, the net present value of the installation is calculated considering a period of 15 years, with a discount rate of 35%, taking into account the risk associated with the novelty and risks of the process. Conservatively, an annual increase of 2% in costs is considered, while sales are considered constant. Under these conditions, after 15 years the NPV of the installation calculated is 9,849,915 Considering both economic viability indicators, the economic viability and profitability of the proposed unit can be confirmed. However, this indicator is very sensitive to the selling price of the products, which is subject to a high degree of uncertainty, as has been explained in chapter 2. According to the sensitivity analysis, the limit selling price for D-mannose as a product is 7.71 € kg-1. Lower prices would result in negative NPV. In any case, greater access to sales price information for D-Mannose will allow the estimation of the viability of the process with a greater degree of certainty. 89 References and Bibliography. [1] Barbier M. & Charrier B. «PEGASUS - Phycomorph European Guidelines for a SUstainable Aquaculture of Seaweeds. », Barbier M. & Charrier B. 2019, doi: 10.21411/2c3w-yc73 [2] Hu, X., Shi, Y., Zhang, P., Miao, M., Zhang, T. and Jiang, B., 2016. D-Mannose: Properties, Production, and Applications: An Overview. Comprehensive Reviews in Food Science and Food Safety, 15(4), pp.773-785. [3] S. A. Wittinger y D. E. Smith, «Effect of Sweeteners and Stabilizers on selected Sensory Attributes and Shelf Life of Ice Cream», J Food Science, vol. 51, n.o 6, pp. 1463-1466, nov. 1986, doi: 10.1111/j.1365-2621.1986.tb13835.x. [4] M. M. Kamel, H. I. Ali, M. M. Anwar, N. A. Mohamed, y A. M. Soliman, «Synthesis, antitumor activity and molecular docking study of novel Sulfonamide- Schiff’s bases, thiazolidinones, benzothiazinones and their C-nucleoside derivatives», European Journal of Medicinal Chemistry, vol. 45, n.o 2, pp. 572-580, feb. 2010, doi: 10.1016/j.ejmech.2009.10.044. [5] K. Ranta et al., «Evaluation of Immunostimulatory Activities of Synthetic Mannose- Containing Structures Mimicking the β-(1→2)-Linked Cell Wall Mannans of Candida albicans», Clin. Vaccine Immunol., vol. 19, n.o 11, pp. 1889-1893, nov. 2012, doi: 10.1128/CVI.00298-12. [6] B. A. Oyofo, J. R. DeLOACH, D. E. Corrier, J. O. Norman, R. L. Ziprin, y H. H. Mollenhauer, «Prevention of Salmonella typhimurium Colonization of Broilers with D- Mannose», Poultry Science, vol. 68, n.o 10, pp. 1357-1360, oct. 1989, doi: 10.3382/ps.0681357. [7] A. Pérez Nebreda, "Valuable monomers and oligomers from hemicelluloses". Master Thesis at Åbo Akademi University, Turku, 2019. Available at: http://www.doria.fi/handle/10024/168436. [8] S. Willfor y B. Holmbom, «Isolation and characterisation of water soluble polysaccharides from Norway spruce and Scots pine», Wood Science and Technology, vol. 38, n.o 3, pp. 173-179, jun. 2004, doi: 10.1007/s00226-003-0200-x. [9]E. Sjöström, Wood chemistry. San Diego: Academic Press, 1993. [10] T. Song, A. Pranovich, I. Sumerskiy, y B. Holmbom, «Extraction of galactoglucomannan from spruce wood with pressurised hot water», Holzforschung, vol. 62, n.o 6, nov. 2008, doi: 10.1515/HF.2008.131. [11] S. Azhar, "Extraction of Polymeric Hemicelluloses from SpruceWood", Wallenberg Wood Science CenterDepartment of Fiber and Polymer TechnologySchool of Chemical Science and EngineeringKTH Royal Institute of Technology, Estocolmo, 2015. [12] S. WILLFÖR, P. REHN, A. SUNDBERG, K. SUNDBERG, y B. HOLMBOM, «Recovery of water-soluble acetylgalactoglucomannans from mechanical pulp of spruce», Recovery of water-soluble acetylgalactoglucomannans from mechanical pulp of spruce, vol. 2, n.o 11, pp. 27-32, 2003. [13] Morad Chadni, Nabil Grimi, Olivier Bals, Isabelle Ziegler-Devin, Nicolas Brosse. «Steam explosion process for the selective extraction of hemicelluloses polymers from 90 spruce sawdust.» Industrial Cropsand Products, Elsevier, 2019, 141, pp.111757. 10.1016/j.indcrop.2019.111757. hal-02289732 [14] L. Michalak, S. H. Knutsen, I. Aarum, y B. Westereng, «Effects of pH on steam explosion extraction of acetylated galactoglucomannan from Norway spruce», Biotechnol Biofuels, vol. 11, n.o 1, p. 311, dic. 2018, doi: 10.1186/s13068-018-1300-z. [15] Ueki, J. Azuma, T. Koshijima, Microwave irradiation of lignocellulosic materials IX. «Conversion of microwave-irradiated lignocellulose into ethanol». Journal of the Japan Wood Research Society. 1988, pp. 462-468. [16] A. P. Nebreda et al., «Acid hydrolysis of O-acetyl-galactoglucomannan in a continuous tube reactor: a new approach to sugar monomer production», Holzforschung, vol. 70, n.o 3, pp. 187-194, mar. 2016, doi: 10.1515/hf-2014-0314. [17] P. D. Carà et al., «Hemicellulose hydrolysis catalysed by solid acids», Catal. Sci. Technol., vol. 3, n.o 8, p. 2057, 2013, doi: 10.1039/c3cy20838a. [18] B. T. Kusema et al. , "Acid hydrolysis of O-acetyl-galactoglucomannan", Catal. Sci. Technol. Vol. 3, No. 1, pp. 116-122, 2013, doi: 10.1039/C2CY20314F. [19] L. Cheng-Yu, "Mesoporous Materials-Based Catalysts for Chemical Hydrolysis of Polysaccharides", Journal of Thermodynamics & Catalysis, vol. 63, 2015. Available: https://www.longdom.org/open-access/mesoporous-materialsbased-catalysts-for- chemical-hydrolysis-ofpolysaccharides-2157-7544-1000e134.pdf [20] A. A. Shatalov, «Highly efficient hydrolysis of plant hemicelluloses by mixed- addenda Keggin-type (Mo-V-P)-heteropolyacids in diluted aqueous solution», Carbohydrate Polymers, vol. 206, pp. 80-85, feb. 2019, doi: 10.1016/j.carbpol.2018.10.106. [21] Shagufta, I. Ahmad, y R. Dhar, «Sulfonic Acid-Functionalized Solid Acid Catalyst in Esterification and Transesterification Reactions», Catal Surv Asia, vol. 21, n.o 2, pp. 53-69, jun. 2017, doi: 10.1007/s10563-017-9226-1. [22] "D-Mannose Market To Reach US$1 bn by 2025 | CAGR 3.5%", Transparencymarketresearch.com. [Online]. Available: https://www.transparencymarketresearch.com/mannose-dmannose-market.html. [Accessed: 19- Mar- 2020]. [23]"Euro to U.S. dollar exchange rate 1999-2018 | Statista", Statista. [Online]. Available: https://www.statista.com/statistics/412794/euro-to-u-s-dollar-annual- average-exchange-rate/. [Accessed: 21- Mar- 2020]. [24] "Rare Sugar Market Statistics 2026 | Growth Projections Report", Global Market Insights, Inc.. [Online]. Available: https://www.gminsights.com/industry-analysis/rare- sugar-market. [Accessed: 20- Mar- 2020]. [25] "D-Mannose Market - Global Industry Growth Analysis, Size, Share, Trends, and Forecast 2020 - 2026", news.marketsizeforecasters.com. [Online]. Available: http://news.marketsizeforecasters.com/d-mannose-market. [Accessed: 21- Mar- 2020]. 26 "D-Mannose_Shijiazhuang Huaxu Pharmaceutical Co.,Ltd", Huaxu-pharma.com Online. Available: https://www.huaxu- pharma.com/product/362.html?gclid=Cj0KCQjwmdzzBRC7ARIsANdqRRk3EshJXrvp rtV-U_UTJ7T9OCo3t32sP5y0gX_ lffFk46sPSgDUwqoaAtD6EALw_wcB. Accessed: 22-Mar-2020. 91 [27] "D-mannose - Indiamart", Dir.indiamart.com. [Online]. Available: https://dir.indiamart.com/search.mp?ss=D-mannose. Accessed: 22- Mar- 2020]. [28] "Food Grade D-mannose-Food Grade D-mannose Manufacturers, Suppliers and Exporters on Alibaba.comSweeteners", Alibaba.com. [Online]. Available: https://www.alibaba.com/trade/search?fsb=y&IndexArea=product_en&CatId=&Search Text=food+grade+D-Mannose+&viewtype=&tab=. [Accessed: 22- Mar- 2020]. [29]J. Laurijssen, Energy use in the paper industry. Eindhoven: Lecturis BV, 2013. [30] M. Jonsson, T. Ranius, H. Ekvall, y G. Bostedt, «Cost-effectiveness of silvicultural measures to increase substrate availability for wood-dwelling species: A comparison among boreal tree species», Scandinavian Journal of Forest Research, vol. 25, n.o 1, pp. 46-60, feb. 2010, doi: 10.1080/02827581003620347. [31] J. He, P. Engstrand, O. Björkqvist, y W. Zhang, «Techno-economic evaluation of a mechanical pulp mill with gasification», Nordic Pulp & Paper Research Journal, vol. 28, n.o 3, pp. 349-357, ago. 2013, doi: 10.3183/npprj-2013-28-03-p349-357. [32] T. Persson, A.-K. Nordin, G. Zacchi, y A.-S. Jönsson, «Economic Evaluation of Isolation of Hemicelluloses From Process Streams From Thermomechanical Pulping of Spruce», en Applied Biochemistry and Biotecnology, J. R. Mielenz, K. T. Klasson, W. S. Adney, y J. D. McMillan, Eds. Totowa, NJ: Humana Press, 2007, pp. 741-752. 33] J. Pettersson, "Energikartläggning för energiledningssystem - Holmen Paper, Hallsta", p. 166. [34] "Industrial prices for electricity in Sweden 1996-2018 | Statista", Statista, 2020. [Online]. Available: https://www.statista.com/statistics/596262/electricity-industry- price-sweden/. [Accesed: 26- Mar- 2020]. [35] Facts on Water Supply and Sanitation in Sweden. [Online]. Available: https://www.mynewsdesk.com/material/document/4989/download?resource_type=resou rce_document. [Accessed: 26- Mar- 2020]. [36] Sweden & Finland Water. Online]. Available: https://eau3e.hypotheses.org/files/2010/11/Sweden_Finland.pdf [Accessed: 26- Mar- 2020]. [37] "The Cost Of Water In The Process Industry", Process Industry Informer. [Online]. Available: https://www.processindustryinformer.com/process-industry-water-cost. [Accessed: 26- Mar- 2020]. [38] E. Roth, Water pricing in the EU: a review. Bruselles: European Environment Bureau, 2001. [Online]. Available: http://www.semide.net/media_server/files/g/r/ReviewWaterPricing2001.pdf [Accesed: 26- Mar- 2020]. [39] "Sweden Inflation Rate | 1980-2020 Data | 2021-2022 Forecast | Calendar | Historical". [Online]. Available at: https://tradingeconomics.com/sweden/inflation-cpi. [Accessed: 24-Mar-2020] [40] T. Persson, H. Krawczyk, A.-K. Nordin, y A.-S. Jönsson, «Fractionation of process water in thermomechanical pulp mills», Bioresource Technology, vol. 101, n.o 11, pp. 3884-3892, jun. 2010, doi: 10.1016/j.biortech.2009.12.142. [41] M. Illikainen, «Mechanisms of thermomechanical pulp refining», p. 78. 42] Holmen.com. [Online]. 92 Available: https://www.holmen.com/globalassets/holmen/sustainability/environment/miljoarbetet- vid-holmens-enheter/environmental-work-at-holmen-paper-hallsta-2018.pdf/. [Accessed: 28-Mar-2020]. [43] Z. Hameed, Z. Aman, S. R. Naqvi, R. Tariq, I. Ali, y A. A. Makki, «Kinetic and Thermodynamic Analyses of Sugar Cane Bagasse and Sewage Sludge Co-pyrolysis Process», Energy Fuels, vol. 32, n.o 9, pp. 9551-9558, sep. 2018, doi: 10.1021/acs.energyfuels.8b01972. [44] R. N. Goldberg et al., «A thermodynamic investigation of the cellulose allomorphs: Cellulose(am), cellulose Iβ(cr), cellulose II(cr), and cellulose III(cr)», The Journal of Chemical Thermodynamics, vol. 81, pp. 184-226, feb. 2015, doi: 10.1016/j.jct.2014.09.006. [45] N. V. Grishatova, V. F. Ur’yash, A. E. Gruzdeva, y N. V. Novoselova, «A calorimetric study of the enthalpy of enzymatic hydrolysis of components of a product from vegetable raw materials for enteral feeding», Russ J Appl Chem, vol. 79, n.o 12, pp. 1930-1934, dic. 2006, doi: 10.1134/S1070427206120044. [46] C. Wyman, S. Decker, M. Himmel, J. Brady, C. Skopec, y L. Viikari, «Hydrolysis of Cellulose and Hemicellulose», en Polysaccharides, S. Dumitriu, Ed. CRC Press, 2004. [47] R. Darros‐Barbosa, M. O. Balaban, y A. A. Teixeira, «Temperature and Concentration Dependence of Heat Capacity of Model Aqueous Solutions», International Journal of Food Properties, vol. 6, n.o 2, pp. 239-258, ene. 2003, doi: 10.1081/JFP-120017845. [48] F. Kawaizumi, N. Nishio, H. Nomura, y Y. Miyahara, «Heat-capacity measurements of aqueous solutions of mono-, di-, and tri-saccharides using an isoperibol twin calorimeter», The Journal of Chemical Thermodynamics, vol. 13, n.o 1, pp. 89-98, ene. 1981, doi: 10.1016/S0021-9614(81)80013-4. [49] C. Xu, S. Willför, P. Holmlund, y B. Holmbom, «Rheological properties of water- soluble spruce O-acetyl galactoglucomannans», Carbohydrate Polymers, vol. 75, n.o 3, pp. 498-504, feb. 2009, doi: 10.1016/j.carbpol.2008.08.016. [50] V. R. N. Telis, J. Telis-Romero, H. B. Mazzotti, y A. L. Gabas, «Viscosity of Aqueous Carbohydrate Solutions at Different Temperatures and Concentrations», International Journal of Food Properties, vol. 10, n.o 1, pp. 185-195, ene. 2007, doi: 10.1080/10942910600673636. [51] B. Kusema et al., "Selective Hydrolysis of Arabinogalactan into Arabinose and Galactose Over Heterogeneous Catalysts", Catalysis Letters, vol. 141, no. 3, pp. 408- 412, 2010. Available: 10.1007/s10562-010-0530-x [Accesed 6 April 2020]. [52] Alfaaesar.com. [Online]. Available at: http://www.alfaaesar.com/EN/DOCS/PROD_INSTRUCT_GEN/Smopex.pdf. Accessed: 09- Apr- 2020]. [53] K. S. Woo, H. Y. Kim, I. G. Hwang, S. H. Lee, y H. S. Jeong, «Characteristics of the Thermal Degradation of Glucose and Maltose Solutions», JFN, vol. 20, n.o 2, pp. 102-109, jun. 2015, doi: 10.3746/pnf.2015.20.2.102. [54] R. Alén, E. Kuoppala and P. Oesch, "Formation of the main degradation compound groups from wood and its components during pyrolysis", Journal of Analytical and 93 Applied Pyrolysis, vol. 36, no. 2, pp. 137-148, 1996. doi: https://doi.org/10.1016/0165- 2370(96)00932-1 [55] Persson, T., Jönsson, A. and Zacchi, G., 2005. 14Th Europan Biomass Conference & Exhibilion. [S.l.]: ETA-Renewable Energies and WIP-Renewable Energies, pp.1693- 1696. [56] Y. Xie et al., «A Five-Zone Simulated Moving Bed for the Isolation of Six Sugars from Biomass Hydrolyzate», Ind. Eng. Chem. Res., vol. 44, n.o 26, pp. 9904-9920, dic. 2005, doi: 10.1021/ie050403d. [57] F. Wei, B. Shen, M. Chen, y Y. Zhao, «Novel Simulated Moving-Bed Cascades with a Total of Five Zones for Ternary Separations», Ind. Eng. Chem. Res., vol. 51, n.o 16, pp. 5805-5812, abr. 2012, doi: 10.1021/ie2024189. [58] H. Khan y M. Younas, «Theoritical analysis and simulation of five-zone simulating moving bed for ternary mixture separation», Can. J. Chem. Eng., vol. 89, n.o 6, pp. 1480-1491, dic. 2011, doi: 10.1002/cjce.20485. [59] J. S. Hur y P. C. Wankat, «New Design of Simulated Moving Bed (SMB) for Ternary Separations», Ind. Eng. Chem. Res., vol. 44, n.o 6, pp. 1906-1913, mar. 2005, doi: 10.1021/ie040164e. [60] B. Pynnonen, «Simulated moving bed processing: escape from the high-cost box», Journal of Chromatography A, vol. 827, n.o 2, pp. 143-160, dic. 1998, doi: 10.1016/S0021-9673(98)00732-8. [61] R. C. Bennett, «Crystallizer selection and design», en Handbook of Industrial Crystallization, Elsevier, 2002, pp. 115-140. [62]S. Diab and D. Gerogiorgis, "Technoeconomic Optimization of Continuous Crystallization for Three Active Pharmaceutical Ingredients: Cyclosporine, Paracetamol, and Aliskiren", Industrial & Engineering Chemistry Research, vol. 57, no. 29, pp. 9489-9499, 2018. Available: 10.1021/acs.iecr.8b00679 [Accessed 23 April 2020]. [63] B. Kusserow, S. Schimpf, y P. Claus, «Hydrogenation of Glucose to Sorbitol over Nickel and Ruthenium Catalysts», Advanced Synthesis & Catalysis, vol. 345, n.o 12, pp. 289-299, ene. 2003, doi: 10.1002/adsc.200390024. [64] A. Aho, S. Roggan, K. Eränen, T. Salmi, y D. Yu. Murzin, «Continuous hydrogenation of glucose with ruthenium on carbon nanotube catalysts», Catal. Sci. Technol., vol. 5, n.o 2, pp. 953-959, 2015, doi: 10.1039/C4CY01088D. [65] C. M. Namara, J. Torroba, y A. Deacon, «New Smopex® Ion Exchange Materials for the Removal of Selenium from Industrial Effluent Streams», johnson matthey technol rev, vol. 59, n.o 4, pp. 334-352, oct. 2015, doi: 10.1595/205651315X689694. [66] S. Leveneur et al., «Synthesis of peroxypropionic acid from propionic acid and hydrogen peroxide over heterogeneous catalysts», Chemical Engineering Journal, vol. 147, n.o 2-3, pp. 323-329, abr. 2009, doi: 10.1016/j.cej.2008.11.045. [67] X. Hu et al., «Investigation of deactivation mechanisms of a solid acid catalyst during esterification of the bio-oils from mallee biomass», Applied Energy, vol. 111, pp. 94-103, nov. 2013, doi: 10.1016/j.apenergy.2013.04.078. [68] V. M. Bhandari, L. G. Sorokhaibam, y V. V. Ranade, «Ion Exchange Resin Catalyzed Reactions—An Overview», en Industrial Catalytic Processes for Fine and 94 Specialty Chemicals, Elsevier, 2016, pp. 393-426 [Online]. Available en: https://linkinghub.elsevier.com/retrieve/pii/B9780128014578000094. [Accesed: 12- mayo -2020] [69] "ION EXCHANGE RESINS". [Online]. Available: http://www2.udec.cl/~analitic/Interc-iones.pdf. Accessed: 08- May- 2020]. [70] "Dowex® HCR-W2 Datasheet", Dowex® HCR-W2 Datasheet. [Online]. Available: https://www.lenntech.com/Data-sheets/Dowex-HCR-W2-L.pdf. Accessed: May 12 - 2020. 71] "Properties of saturated water", University of Cantabria. Online]. Available: https://personales.unican.es/renedoc/Trasparencias%20WEB/Trasp%20Termo%20y%2 0MF/00%20GRADOS/Tablas%20Termo.pdf. Accessed: 14- May- 2020]. [72] J. Couper, Chemical process equipment. Amsterdam: Gulf, 2005, Capítulo 10. Mixing and Agitation páginas. 277-328 . [73] A. Coker, Modeling of chemical kinetics and reactor design. Boston, Mass.: Gulf Professional Pub., 2001, páginas. 554-642. [74] "Fact Sheet for Slawinski's Background." [Online]. Available: https://www.slawinski.es/cms/upload/downloads/Slawinski-Datasheet-EN.pdf. [Accessed: 16- May- 2020]. [75] "Sedimentation Notes". [Online]. Available: http://sgpwe.izt.uam.mx/files/users/uami/favela/Notas_Sedimentacion.pdf. Accessed: 19- May- 2020]. [76] J. DAMN FLOWERS, STEEL AND ITS APPLICATIONS. Mexico. Universidad Autónoma de Nuevo León, 1996, p. 95. Online]. Available: http://eprints.uanl.mx/421/1/1020118272.PDF [Accessed: 19- May- 2020]. 77] "Rules for the Construction of Pressure Vessels", Asme.org, 2013. Online]. Available: https://www.asme.org/getmedia/443f6b0a-4e45-4a12-8f83- 3230fa92306a/35832.pdf. Accessed: 23- May- 2020]. [78] V. Atiemo-Obeng, S. Kresta and E. Paul, Handbook of industrial mixing. Hoboken: Wiley-Interscience, 2011, pág. 352. [79] R. Turton, Analysis, synthesis, and design of chemical processes, fifth edition. Boston [etc.]: Prentice Hall, 2018. [80] A. Dimian and C. Bildea, Chemical process design. 2008, Appendix B: Heat‐ Exchanger Design (Págs. 474-482). [81] C. Kothandaraman, Fundamentals of heat and mass transfer. New Delhi: New Age International (P) Limited, Publishers, 2012. [82] N. B.Totala, "Natural Convection Characteristicsin Vertical Cylinder", International Journal Of Engineering And Science, vol. 3, 8, pp. 27-31, 2013. Available: http://www.researchinventy.com/papers/v3i8/E038027031.pdf. [Accessed 27 May 2020]. [83]Standards of the Tubular Exchanger Manufacturers, 9th ed. Tarrytown, New York: TEMA, 2013. [84]W. Seider, D. Lewin, J. Seader, S. Widagdo, R. Gani and K. Ng, Product and process design principles. 95 [85] Process Unit and Offsites Layout Guide. PIP, 2013. [Online]. Available: https://www.pip.org/docs/default-source/practices- documents/pne000035c13a80395a262f789edff00008ddc6a.pdf [Accesed: 23- May- 2020]. [86] Dow Chemical Company., "Fire and Explosion Index. Hazard Classification Guide.” 5ª Edicción., 1986. [87] M. Rausand and S. Haugen. “Risk assessment: Theory, methods and applications”, 1st ed. Wiley, 2011. 88] "Geographical coordinates of Hallstavik, Sweden", Dateandtime.info. Online. Available at: https://dateandtime.info/es/citycoordinates.php?id=2708426 [Accessed: 02- May- 2020]. 89] "Hallstavik Climate: Temperature, Weather and Water Temperature of Hallstavik - Climate-Data.org". Online]. Available at: https://es.climate- data.org/europe/suecia/stockholms-laen/hallstavik-228769/. Accessed: 02-May-2020 [90] "Climate Hallstavik", meteoblue. [Online]. Available at: https://www.meteoblue.com/en/weather/historyclimate/climatemodelled/hallstavik_swe den_2708426. [Accessed: 02-May-2020] [91] «Environment & Climate Change Law 2020 | Sweden | ICLG». [Online]. Available en: https://iclg.com/practice-areas/environment-and-climate-change-laws-and- regulations/sweden. [Accesed: 02-may-2020] 92] R. och Regeringskansliet, "The Swedish Environmental Code", Regeringskansliet, 08-Jan-2000 Online]. Available at: https://www.government.se/legal- documents/2000/08/ds-200061/. Accessed: 03-May-2020 [93] "Swedish Environmental Law", Naturvårdsverket. [Online]. Available at: http://www.naturvardsverket.se/Om-Naturvardsverket/Publikationer/ISBN/6700/978- 91-620-6790-8/ [Accessed: 02-May-2020] [94] "Sweddish Waste Management" Avfallsverige.se, 2018. [Online]. Available: https://www.avfallsverige.se/fileadmin/user_upload/Publikationer/Avfallshantering_201 8_EN.pdf [Accessed: 02- May- 2020]. 95] Act (2004:1199) on emissions trade. "FAO.org ": [Online]. Available at: http://www.fao.org/faolex/results/details/en/c/LEX-FAOC131813. Accessed: 02-May- 2020 [96] C. Eriksson, M. E. Nilsson, G. Pershagen, Sverige, y Naturvårdsverket, Environmental noise and health: current knowledge and research needs. Estocolmo: Agencia Sueca de Protección Medioambiental, 2013. [97] "ESTUDIO DE IMPACTO AMBIENTAL DE UNA ESTACIÓN DEPURADORA DE AGUAS RESIDUALES EN PGZ", Euskadi.eus, 2005. Online]. Available: https://www.euskadi.eus/contenidos/documentacion/esia_papelera_guipuzcoana/es_doc /adjuntos/esia_estacion_depuradora.pdf. Accessed: 03- May- 2020]. [98] "ESTUDIO DE IMPACTO AMBIENTAL DE UNA PLANTA DE COGENERACIÓN EN PAPELERA GUIPUZCOANA DE ZICUÑAGA S.A.", Euskadi.eus, 2007. [Online]. Available: https://www.euskadi.eus/contenidos/documentacion/esia_papelera_guipuzcoana/es_doc /adjuntos/esia_estacion_depuradora.pdf. Accessed: 03- May- 2020]. 96 [99] M. Söderman, "Environmental Management in the Construction Industry - A Comparative Analysis of Skanska’s Environmental Risk Assessment", 2006. [Online]. Available: https://stud.epsilon.slu.se/11028/1/soderman_m_170918.pdf. [Accessed: 03- May- 2020]. [100]R. Fairman, C. Mead and W. Williams, Environmental risk assessment. Copenhagen, Denmark: European Environment Agency, 1998. [101] L. Freitas & A. Magrini, «Waste Management in Industrial Construction: Investigating Contributions from Industrial Ecology», Sustainability, vol. 9, n.o 7, p. 1251, jul. 2017, doi: 10.3390/su9071251. [102] M. Peters , K. Timmerhaus. and R. West., "Plant Design for chemical engineers 5th edition", Mhhe.com. [Online]. Available: http://www.mhhe.com/engcs/chemical/peters/data/ce.html. [Accesed: 02- Jun- 2020]. [103]"Matches' 275 Equipment Cost Estimates.", Matche.com, 2020. [Online]. Available: https://www.matche.com/equipcost/Default.html. [Accesed: 02- Jun- 2020]. [104] "2020 CEPCI Updates: February (prelim.) and January (final) - Chemical Engineering | Page 1", Chemengonline.com, 2020. [Online]. Available: https://www.chemengonline.com/2020-cepci-updates-february-prelim-and-january- final/. [Accesed: 02- Jun- 2020]. [105]"Euro to U.S. dollar exchange rate 1999-2018 | Statista", Statista. [Online]. Available: https://www.statista.com/statistics/412794/euro-to-u-s-dollar-annual- average-exchange-rate/. [Accesed: 02- Jun- 2020]. [106] Perry, Robert H., and Chilton, Cecil H., Perry’s Chemical Engineers’ Handbook (8ª Edition), McGraw-Hill, Nueva York, 2008, Capítulo 9. Tabla 9-51. [107] "Macroporous Cation Exchange Catalyst Resin - Buy Macroporous Cation Exchange Catalyst Resin,Tulsion T-56,Tulsion T-56 Mp Product on Alibaba.com", www.alibaba.com, 2020. [Online]. Available: https://www.alibaba.com/product- detail/Macroporous-Cation-Exchange-Catalyst- Resin_62064674831.html?spm=a2700.7724857.normalList.23.52ff254aQPloVs&s=p. [Accesed: 02- jun- 2020] [108] "USDA ERS - Sugar and Sweeteners Yearbook Tables", Ers.usda.gov, 2020. [Online]. Available: https://www.ers.usda.gov/data-products/sugar-and-sweeteners- yearbook-tables/sugar-and-sweeteners-yearbook- tables/#U.S.%20Imports%20of%20Sugar%20Syrups. [Accesed: 02- Jun- 2020]. [109] J. McKetta, Encyclopedia of chemical processing and design. Vol. 12. Corrosion to cottonseed. New York: Dekker, 1981. [110] "Average annual hours actually worked per worker", Stats.oecd.org, 2020. [Online]. Available: https://stats.oecd.org/Index.aspx?DataSetCode=ANHRS. [Accessed: 02- jun- 2020]. [111] E. Institute, "Chemical Engineer Salary in Sweden", Erieri.com, 2020. [Online]. Available: https://www.erieri.com/salary/job/chemical-engineer/sweden. [Accesed: 02- Jun- 2020]. [112] "Corporation tax in Sweden", 2020. Online]. Available: https://santandertrade.com/es/portal/establecerse-extranjero/suecia/fiscalidad. Accessed: 02- Jun- 2020] 97 ANNEXES ANNEX A. Cost estimate for the extraction of GGM Design capacity. Obtaining 1 t of GGM is considered as a basis for calculation. From the total GGM content in the wood, as well as the percentage of which is solubilized by thermo- mechanical processing in the process waters, the amount of wood to be fed into the process, after debarking and size reduction, is calculated. Using the wood utilization ratio, the required amount of wood to be purchased is calculated. 1 (0,18 ∙ 0,08) = 69,4 − 69,4 ∙ 1 0,6 = 116 The amount of pulp obtained after processing the wood is obtained from the overall wood-pulp ratio. This data is then used to calculate the service requirements of the process. 69,44 ∙ 0,882 = 61,3 61,3 ∙ 3 = 184 61,3 ∙ 2,5 = 150,8 ℎ 61,3 ∙ 0,03 = 1,9 ℎ ℎ Water 184 ∙ 2 = 368 € Electricity 150,8 ∙ 75 = 11310 € Heating 1,9 ∙ 80 = 152 € Wood 116∙ 27,5 = 3190 € Thus the final cost per ton of GGM extracted is 15,020 euros / t GGM. However, this estimate shows the total cost of thermo-mechanical pulp processing based on the majority items. For the purposes of considering the double use of the operation, for the production of pulp, and the extraction of LGM, the mass fraction of LGM extracted from the total available in the wood fed into the process is considered as an approximation of the cost. 13.740 ∙ 0,18 ∙ 0,08 ≈ 220 € ∗ t GGM . Finally, the amount of process water required is assumed to be equivalent to that used in the process, thus calculating the maximum theoretical concentration of GGM in the process water. 10 188 = 5,3 ∙ It should be noted that the amount of water taken as a reference is based on a thermo-mechanical pulp processing facility (Hallsta, Sweden). However, the amount of 98 water fed into the process could be increased, in order to obtain an additional stream of aqueous hemicellulose extract with a lower concentration. According to bibliographical references, lower concentrations are usually found in process waters40. Similarly, the initial water content in the wood and the amount of process water that cannot be recovered are considered to be comparable for the purpose of determining the maximum theoretical concentration as a guideline. The calculation of the design capacity is made from the maximum amount of processable GGM in the installation taken as a reference. 33 With an average capacity of 230,000 t / year of pulp obtained, and a ratio of 61.3 t pulp / t GGM, the maximum amount of GGM processable is estimated. . , = 3752 t GGM / year Considering an average composition for the GGM of Abies Man:Glu:Gal of 3:1:0.6 and the conversion process, the maximum amount of D-mannose obtainable by hydrolysis is estimated. A maximum design capacity of approximately 2400 t year-1 of D-mannose as the main product is considered. Other sugars such as Gal and Glu could be considered. , ∙ 3752 ∙ 0,9 = 2202 t D − Mannose / year , , ∙ 3752 ∙ 0,9 = 441t D − Gal / year , ∙ 3752 ∙ 0,9 = 734 t D − Glu / year ANNEX B. Data used for the calculation of linear correlations = 4,144 − 1,673 ∙ Conc. (%wt) mPa s 0.5 30 5 70 10 400 Exponential Adjustment Viscosity (0 - 10 % wt) GGM = 22,9 ∙ , ∙ ² = 0,97 ai bi R2 Man 4,144 -1,665 0,995 Glu 4,149 -1,744 0,995 Gal 4,136 -1,600 0,991 Weighted Average (GGM) 4,144 -1,673 Viscosity in GGM solutions Linear Regression for Cp Cp = a + b*c 99 Viscosity in glucose solutions. 50 To estimate the viscosity at 90ºC, linear extrapolation is used, considering a temperature range close enough to make the approximation. 10 0 A N N E X C . E xp an si on o f p ro ce ss a nd m at er ia l b al an ce d ia gr am s C .1 B L O C K D IA G R A M . 10 1 C .2 P R O C E SS F L O W D IA G R A M (P FD ). 10 2 C .3 E xt en de d m at er ia l b al an ce . St re am N m C O M PO SI T IO N T P xH 2O xG G M xM an xG lu xG al xH 2O xG G M xM an xG lu xG al N o. km ol m in -1 kg m in -1 m ol e fr ac tio n (g L -1 ) ºC at m 8 19 ,8 04 37 7, 94 1, 0E +0 0 3, 7E -0 5 0 0 0 99 5, 7 60 ,0 0, 0 0, 0 0, 0 30 1 9 A -I 19 ,8 04 37 7, 94 1, 0E +0 0 3, 7E -0 5 0 0 0 99 5, 7 60 ,0 0, 0 0, 0 0, 0 30 2, 1 10 A -I 19 ,8 04 37 7, 94 1, 0E +0 0 3, 7E -0 5 0 0 0 94 3, 0 60 ,0 0, 0 0, 0 0, 0 12 0 2, 1 11 A -I 19 ,8 85 37 7, 94 9, 9E -0 1 3, 9E -0 6 3, 6E -0 3 1, 2E -0 3 7, 3E -0 4 94 3, 0 6, 0 35 ,5 11 ,4 7, 2 12 0 2, 1 12 A -I 19 ,8 85 37 7, 94 9, 9E -0 1 3, 9E -0 6 3, 6E -0 3 1, 2E -0 3 7, 3E -0 4 94 3, 0 6, 0 35 ,5 11 ,4 7, 2 12 0 2, 1 13 A -I 19 ,8 85 37 7, 94 9, 9E -0 1 3, 9E -0 6 3, 6E -0 3 1, 2E -0 3 7, 3E -0 4 99 2, 4 6, 0 35 ,5 11 ,4 7, 2 40 1 103 ANNEX D. Calculations relating to reactor design and operation D.1 Design considerations. First, the following design considerations are made. The reactor system is properly stirred, the mixing conditions with homogeneous throughout the fluid. The reaction pH (pH=0.75) defines the concentration of the catalyst. The catalyst is considered solid spheres of negligible porosity once they have been previously hydrated. The specifications of the catalyst are obtained from the manufacturer's data sheet. 70 The catalyst charge is defined as 8.1 %v/v. Consequently, the volume of the mixture of the liquid phase is defined as 91.9 % of the total reaction volume. Capacity = 1.8 mmol (H ) ∙ g ; ρ = 1.22 g ∙ l = 10 . 1.8 ∙ 10 = 98.8 ∙ = 8.1 ∙ It is considered an annual time of operation in plant of 8,000 h year-1. The total volume of solution to be processed is calculated from the concentration of GGM in the reactor feed and the total amount obtained from the process water. ( % ) = 3752000 ( ∙ ) 0.06 ( ∙ ) ∙ 10 ( ∙ ) = 62533 Fluid catalyst separation is modeled as the flow of the hydrolysate through a bed of catalyst particles. Consequently, the pressure drop of a bed with the height defined by the reactor dimensions is considered. The calculation of the volume per batch is defined by the operation time in addition to the loading, unloading and conditioning times. Since the discharge time is a function of the dimensions of the reactor, it is initially considered negligible in comparison with the rest. 104 D.2 Material and energy balances in the reactor. From equations 5.1 and 5.2 the overall material balance to the reactor can be derived. The reaction model considered considers three different reactive fractions in the starting material; D-man, D-glu and D-gal bonded by O-glycoside. Thus, the initial concentration of each of the reaction fractions as well as the conversion of the process must be defined. Based on the documented average molecular weight of 29 kDa for GGM dissolved in TMP12 processing water (see Chapter 1), its degree of polymerization (DP) is calculated using the molecular weight of glucose and its epimers after dehydration when forming the O-glycosidic bond. = 29.000 180,16 − 18 = 179 Knowing the degree of polymerization, it is possible to define the number of monosaccharide units per initial GGM molecule. Thus, the molar concentration of non- hydrolysed sugar units can be defined according to equation 5.3. The degree of polymerization, initial hemicellulose concentration, and relative ratio of each monosaccharide (Pi) are considered.3 According to the ratio 3:1:0.6 (D-man:D-glu:D-gal) defined for the MGL, the relative ratios are respectively 65.3%, 21.7%, and 13%. = ( ∙ ) ( ∙ ) ∙ ∙ [5.3] Besides, the evolution of the reaction can still be considered as the reduction of the PD with the reaction. Based on the initial amount of moles of GGM in the system, the number of monosaccharide units that must react to reduce a PD unit is defined. Equation 5.4 presents in the numerator the moles of hydrolyzed monosaccharide, while the denominator defines the moles corresponding to each unit of degree of polymerization average of the GGM in the system. The conversion of the GGM is defined according to equation 3.6 (see chapter 3, section 2). The volume of catalyst in the system defines the reaction rate, which can be calculated according to equation 5.5, where ε represents the volumetric fraction of the catalyst in the system. = − ∑∆ ∑∆ [5.4] = (1 − ) ∙ [5.5] 105 From the above considerations, it is possible to calculate the initial number of moles of each monosaccharide fraction by multiplying the concentration by the volume of the liquid phase. The reaction rate for each fraction is calculated from the total volume of catalyst in the system and the concentrations of each unreacted monosaccharide. Equation 5.1 can be re-expressed in finite time terms, allowing the system to be solved using time differentials, allowing the calculation of the evolution of hydrolysis (equation 5.6). Likewise, the concentrations of free monosaccharides can be defined in a similar way (equation 5.7). N( ) represents the amount of the monosaccharide "i" as free sugar in the hydrolysate. The water consumption with each hydrolysis reaction must also be considered (equation 5.8). The initial amount of water is calculated considering the molarity of 1 L of water (55.56 M) and the volume of liquid. = − ∙ ∆ [5.6] ( ) = ( ) + ∙ ∆ [5.7] = − ∙ ∆ [5.8] Finally, the concentrations of each unhydrolyzed monosaccharide fraction can be recalculated from the number of moles present in the system and the volume of the liquid phase. Using finite 10-minute increments, the material balance is resolved step by step, recalculating in each interval the degree of polymerization, conversion, kinetic constants, reaction ratios, unreacted monosaccharide concentrations and water concentration. At the same time, the energy balance can be made on the system. Considering the enthalpy of the reaction (100 J gGGM-1 , Table 6, Chapter 3) and the correlation for the specific heat of the hydrolysate (equation 3.2), the enthalpy balance (equation 5.9) is derived. The amount of reacted GGM is defined as the sum of the mass corresponding to the three hydrolysed monosaccharide fractions. Their contribution to the mass of the GGM is considered as the molecular weight of glucose after the formation of the O- glycosidic bond. Using equation 3.2, the specific heat is Cp = 4005 J kg-1 K-1 and is considered constant throughout the hydrolysis process. The density of the solution is considered to be 1000 kg m3 , the reduction in density due to the increase in temperature 120ºC71 = 942 kg m3 ) is compensated by the solutes dissolved in it (60 kg m3 ). The contribution of the catalyst to the specific heat is considered negligible. The term ∆C represents the increased concentration of monosaccharide "i" in the hydrolysate. 106 T = T + ∆ ∙(∑∆ )∙( , ) [5.9] Finally, the coupling of the energy balance with the material balance should be considered, correcting the value of the kinetic constants of hydrolysis with the new temperature. The resolution of the material balance is done numerically by means of the calculation software Microsoft Excel 2016. For an initial GGM concentration of 60 g L- 1, with a volumetric catalyst load of 8.1%, at an initial temperature of 120 ºC and 2.1 atm pressure, it gives a result of 1660 minutes as reaction time. Likewise, the enthalpic balance of the system shows a temperature increase of 1.34 ºC, which, in spite of being reduced, is significant in the increase of the kinetic reaction. D. 3 Reactor dimensioning. The reactor loading time is estimated from the volume of the reactor and the flow rate fed into it. Similarly, the time taken for the fluid to be discharged into the tank will depend on the size of the reactor and the volume of fluid inside it. Consequently, these times are initially considered as negligible compared to the reaction time, making a first estimate of the number of operation cycles. From these, the volume of hemicellulose extract to be processed per cycle is defined. Considering the volumetric fraction of the catalyst, the total reaction volume can be defined. = 8000 (ℎ ∙ ) ( ) ∙ 60 1 (min∙ ℎ ) ≈ 290 V ( ) = 62533 = 215 m ∙ cycle ; V = ( , ) = 234 m ∙ cycl Given the high volume to be processed per batch and the reduced cost of the batch agitated tank type reactors, it was decided to operate in parallel with 4 agitated tanks. Consequently, the reaction volume in each tank will be 58.5 m3. The dimensioning of the reactor is carried out by combining the volume of a torispherical bottom (Klopper type DIN-28011) with the volume of the cylinder that constitutes the body of the tank. Defining the diameter of the tank as the main design parameter, the rest of the dimensions refer to it. Thus, the tank body volume is defined as V = π ∙ ∙ H . Considering equation 5.10 the filling height can refer to the diameter of the tank. The dimensions according to DIN-28011 are used to calculate the volume of the bottom, 107 ignoring the value of the wall thickness in relation to the diameter magnitude. 74 By replacing these relations in the volume equation of a torispherical cap75 the volume expression for the tank bottom can be obtained. = 3 ∙ (2ℎ − (2 + + 2 ) ∙ ( − ℎ) + 3 ∙ ( − ℎ − ) R = D ; a = 0.1 ∙ D h = 0,1935 ∙ D ; c = (R − a) − (R − h) By making the corresponding substitution in the volume equation, the expression for the volume of the tank bottom can be obtained according to DIN-28011. Combining the volume expressions for the tank body and the tank bottom, equation 5.11 is obtained, allowing the calculation of the tank volume as a function of D. It should be considered that the height of the cylindrical body of the reactor will be the total filling height (HL = D) discounting the depth of the bottom (h). V = π ∙ 0,0314 ∙ D ; H = D − h = D − 0.1935D = 0.8065D = = ∙ 0,0314 ∙ + ∙ 4 ∙ 0,8065 [5.11] The resolution of equation 5.11 for the tank volume allows for a diameter of 6.11 m, considering the filling height of 6.11 m, the correction of 25% of the tank height according to DIN-28110 applies). Finally, the dimensions of the reactor are rounded off and the following characteristics are considered. D = 6,11 m ; H = 6,11 ∙ 1,25 = 7,64 m = 7,65 = 6,1 = 78,3 = 58,5 D. 4 Reactor Operability. Loading and unloading times. First, the reactor discharge time is estimated according to equation 5.12 = ( ) + [5.12] Figure A3.1 Dimensions of a torispherical cap according to DIN-28011 108 From the dimensions of the reactor, it will be possible to calculate the depth of the catalyst "bed" at the bottom of the equipment when stopping the stirring of the system. Considering the volumetric loading of the catalyst and the filling volume of the system, the volume of catalyst deposited at the bottom can be obtained. = ∙ = 58,5 ∙ 0,081 = 4,74 Consequently the catalyst will mainly be deposited on the bottom of the tank. The volume equation of a Klopper-type bottom is used to clear the height of the bed. Similarly, the diameter of the bottom passage section where the catalyst is deposited is calculated. 4,74 ( ) = 3 ∙ (2ℎ − (2 0,1 + ( − 0,1 ) − ( − ℎ) + 0,2 ) ∙ ( − ℎ) + 3 (0,9 ) − ( − ℎ) ∙ − ℎ 0,9 → ℎ = 0,712 4,74 = ∙ 0,0314 ∙ → = 3,64 Knowing the bed height, the pressure drop that it will offer to the fluid passage can be known. Since the reactor operates under pressure, the driving force of the fluid through the bed is the internal pressure of the reactor (2.1 atm) added to the hydrostatic pressure of the fluid column. The data provided by the manufacturer are used. The hydrolysate is considered to have a fluid dynamic behavior comparable to that of water, as the viscosity data corroborate with the concentrations of water. The temperature-related pressure drop correction found in the catalyst data sheet shows a reduction in pressure drop with temperature, however, the operating temperature used is outside the temperature range. The value at 20°C is used as a conservative approximation to the actual pressure drop in the bed. A fluid discharge speed of 2.03 m h-1 is selected, corresponding to a negligible pressure drop < 0.1 atm m-1 (see figure19). Thus, considering the fluid passage section equivalent to the diameter of the upper region of the deposited catalyst bed, the discharge flow rate is calculated in this way. Once the flow rate is known, the time required for hydrolysate removal through the catalyst bed can be determined. = 0,0338 ( ∙ ) ∙ ∙ ( 4 ) = 0,358 ∙ 109 The sedimentation time is estimated considering the terminal velocity in prevented sedimentation of the catalyst particles. 75 The viscosity of the hydrolysate will be slightly higher than that of water, due mainly to the residual GGM content (0.6%wt). Using equation 3.3, the viscosity at 25ºC is obtained. Viscosity at 120 ºC is estimated as the reduction in water viscosity from 25 to 120ºC (IAPWS standard). = 22,9 ∙ , ∙% ∙ º º = 22,9 ∙ , ∙ , ∙ 0,215 0,900 = 6,44 ∙ The flow rate is approximated by calculating Reynolds' number analytically. The viscosity of the mixture (μ) and mixture density (ρ )The following are estimated as functions of solid content. An empirical correction function is calculated for this purpose. 75 Equations 5.13-5.16 are used = 1 10 , ∙ , = 0,71 ; = 9,8 ∙ (0,0008) ∙ 965 ∙ (1220 − 965) 18 ∙ 0,009 = 0,85 The Reynolds number obtained shows how the flow regime is clearly laminar around the particles, and equation 5.17 can be applied to the calculation of the terminal velocity in impeded sedimentation. = (0,919) ∙ 0,7112 ∙ 9,81 ∙ (0,0008) (1220 − 965) 18 ∙ 0,009 = 0,00594 ∙ Subsequently, the settling time can be calculated by dividing the height of the liquid level in the reactor by the impeded terminal settling velocity. In this way, the maximum discharge time is considered as the sum of both = = 6,11 0,00594 = 1028 → 17 min , 8 = + = 53,75 0,358 = 150 + 17 = 167 [5.13] [5.14] [5.15] = ∙ ∙ ∙ − 18 ∙ [5.16] = ∙ ∙ ∙ − 18 ∙ [5.17] 110 The reactor loading time is defined from the load flow, establishing this as an identical value to the discharge flow, thus allowing energy recovery between the discharge and feed streams. = = 53,75 0,358 = 150 Finally, the contribution of the regeneration time is considered. Given that during regeneration the reactor is filled with a smaller quantity of fluid (14%v/v), the loading and unloading times will be 57.8% less than the usual operation. In addition to the high regeneration time (10h), loading and unloading are considered to be 60% of the usual reaction time. Thus the contribution of regeneration to conditioning operations is defined in accordance with the provisions of Chapter 5. = (600 + (150 + 167) ∙ 0,6) ∙ 0,2 = 158 D. 5 Reactor thickness. In accordance with the ASME standard, the thickness of the reactor is defined by equation 5.18. Where Rt is the tank radius in inches, E is the welding efficiency (considered 0.85), S is the maximum allowable stress and CA is the maximum allowable corrosion, being 0.083 in for stainless steels. Replacing in the equation, considering P as the design pressure, calculated as the reactor pressure added to the hydrostatic pressure due to the fluid column at the point of maximum height and corrected according to equation 5.18. ( ) = (202600 + 6,1 ∙ 965 ∙ 9,81) + 6895 = 267242 → 38,76 ( ℎ) = 38,76 ∙ 120,07 2 ∙ 0,85 ∙ 81946 − 0,2 ∙ 38,76 + 0,083 = 0,0334 ℎ → 0,85 D. 6 Agitation System. By applying equation 5.19, the minimum suspension speed of the catalyst can be obtained. Previously, the catalyst mass percentage (WL-S) should be calculated, using ( ) = ( + ) + 6895 [5.18] ( ℎ) = ∙ ( ℎ) 2 ∙ ∙ ( ) − 0,2 ∙ + [5.17] 111 the density of water at 120ºC (943 kg m-3 , referred to earlier in this annex), corrected with the relative density of the suspension due to the solid content (calculated during the prevented sedimentation). Subsequently, by applying equation 5.19 the minimum stirring rate can be calculated. = 0,081 ∙ 1,22 1 ∙ 0,943 ∙ (1 − 0,081) ∙ 100 = 11,4 % N = 1,4 ∙ (3) , ∙ , , ∙( ) , ∙ , , ∙ , , ∙ , , , , ∙ , ∙ , ( ∙ ) = 38 rpm D. 7 Reactor heat exchange The design of the heat exchangers is done in a basic way, considering the required area of heat exchange. Initially, the heat flow to be exchanged and the heat flow required to reach the reaction temperature are defined. A constant Cp is considered for the solution of 4.05 J g-1 K-1 calculated by equation 3.2. The mass flow rate is obtained by considering the volume flow rate and the mass concentration of GGM (60 g L-1). The water density is considered constant as 1000 kg m-3, considering that the dissolved solutes compensate the reduction of the fluid density with the temperature. Thus, the mass flow rate per reactor is calculated as 5.97 kg s-1; the heat available in the preheating is calculated considering the 10º temperature difference between the supply and the output. = ∆ = 0,358 ∙ 1000 60 ∙ 4,05 ∙ (120 − 40) = 1934 Thus the output temperature of the preheated power supply will be 110 ºC. = ∆ = 0,358 ∙ 1000 60 ∙ 4,05 ∙ (110 − 120) = 241,75 From the necessary heating heat, the heating steam flow to be supplied is obtained, assuming total condensation in the pipes. The latent heat of vaporization (Lv) of the water at 10 atm, 180 ºC, Lv = 2108 kJ kg-1 is used. = 241,75 2108 = 0,115 In the preheater, since it is the same fluid and with equal mass flow in both streams, equation 5.23 is simplified, where the term ∆T is constant and equal to 10°C. Furthermore, for both equipment a factor Fc = 1 is considered, due to the temperature gradient established in both situations (R = 0,00876 B= 0 in the case of the heater, leading to an inconsistent calculation of the Correction Factor). Knowing the temperatures at the 112 exit and entrance of the exchanger, equation 5.23 is used to calculate the exchange area. Thus APreheater = 241.75 m2 For the shell and tube exchanger. Aheater = 20.84 m2 For the steam heater. The application of equation 5.22 allows to determine the critical radius of the insulator in 3.5 mm, being a suitable insulator for the system at higher thicknesses as the one selected. D. 8 Fluid power equipment. The calculation of the fluid velocity is made on the basis of the diameters of the pipes used. For both loading and unloading the flow rate will be 5.97 L s-1. The diameter is then set at the nearest industry standard; 4 in. Equation 5.30 is used to calculate the diameter of the pipe. 0,2013 ( ∙ ) 4 ∙ = 1 + 12 ∙ 6 → = 4,28 ; = 0,00928 = 4 ∙ = 0,00597 0,00928 = 0,64 ∙ The safety distances according to the references are considered over the minimum value, establishing 50 meters between the storage tanks and the reactor and 5 meters between the heat exchanger and the reactor. The pumps are considered to be arranged with sufficient distance to allow access, considering an additional length of 10m due to the arrangement of the pumps. In total 120 m of conduction. The above-mentioned accidents (20 90° elbows and 8 T-joints) will provide a pressure drop equivalent to x m of driving. º = 35 → 20 ∙ 35 ∙ 0,1016 = 71,2 = 50 → 8 ∙ 50 ∙ 0,1016 = 40,6 The calculation of the friction factor is estimated for the average Reynolds value in the system. 113 = → ( º) = 2320 ; ( º) = 7225; = 4773 In the case of a turbulent regime with D > 4000, the friction factor of the colebrook-white equation is cleared (Eq. 5.29). Calculating an f = 0.03839 Using the Fanning equation (5.28) the friction losses in the pipes are calculated. Half the length and load loss due to accidents are considered, as only the losses in the loading/unloading section are calculated. = 2 ∙ 0.3839 ∙ 120 + 71,2 + 40,6 2 ∙ 0,64 0,1016 = 359 Finally, equation 5.27 is used to calculate the pumping power. The initial velocity on the surface of the fluid in both the tank and the reactor is considered negligible due to its high cross-sectional area. The reactor will operate at a pressure of 2.1 atm, while the buffer tanks are at atmospheric pressure. Bernoulli's equation is applied twice, for the T-200 section up to the R-204 A-E / F-I reactors and in the R-204 A-E / F-I section up to the T-206 tank. The discharge pressure in the buffer tank will be considered as the atmospheric pressure plus the hydrostatic pressure due to the height of the tank. = 1000 ∙ 9,8 ∙ 5 = 49.000 For charging operation (P-201 A-I) 0,64 2 − 0 2 + (6,1 − 5) ∙ 9,81 + (202600 − 101300 − 49000) + 70000 1000 + 359 = 492 → = 0,00597 ∙ 1000 ∙ 492 0,65 = 4520 For unloading operation (P-205 A-I). 0,64 2 − 0 2 + (0 − 6,1) ∙ + (101300 + 49000 − 202600) + 70000 1000 + 359 = 317 → = 0,00597 ∙ 1000 ∙ 317 0,65 = 2912 11 4 A N N E X E . P ro ce ss in st ru m en ta tio n an d co nt ro l ( P& ID ) d ia gr am 11 5 A N N E X F . P ro ce ss a re a di st ri bu tio n pl an . L ay ou t. 11 6 Id en tif ie r R ep re se nt ed E qu ip m en t Fu nc tio n D es ig n (a tm ) T de si gn (m in -m ax ) (º C ) D (m ) H (m ) D is po si tio n In te rn s T -A 1 A tm os ph er ic ta nk Pr oc es s w at er st or ag e TM P 1 80 º -- -- V er tic al N o T -A 2 A tm os ph er ic ta nk R eg en er at iv e A ci d St or ag e 1 80 º -- -- V er tic al N o T -A 3 A tm os ph er ic ta nk Pr oc es s w at er st or ag e an d w as hi ng . 1 80 º -- -- V er tic al N o T -A 4 So lid s S ilo Pr od uc t s to ra ge 1 -- -- -- V er tic al N o T -A 5 So lid s S ilo Pr od uc t s to ra ge 1 -- -- -- V er tic al N o T -A 6 A tm os ph er ic ta nk Pr od uc t s to ra ge 1 -- -- -- V er tic al N o T -2 00 A tm os ph er ic ta nk Pu lm on ar y ta nk fo r r ea ct or lo ad in g 1 30 -- -- V er tic al N o E -2 02 H ea t e xc ha ng er Sh el l-t ub e ty pe E h ea t ex ch an ge r. 2, 1 10 30 12 0 -- -- H or iz on ta l B af fle s In st ru m en ta tio n E -2 03 A -I H ea t e xc ha ng er M ed iu m P re ss ur e St ea m H ea te r 2, 1 18 5 11 0 -- -- H or iz on ta l D ef le ct or s Tu be b un dl e R -2 04 A -I R ea ct or St irr ed ta nk ty pe h yd ro ly sis re ac to rs w ith fi lte r a t di sc ha rg e 2, 1 12 0 6, 1 7, 65 V er tic al St irr er B af fle s In st ru m en ta tio n T -2 06 A tm os ph er ic ta nk Pu lm on ar y ta nk fo r r ea ct or di sc ha rg e 1 -- V er tic al N o 11 7 A N N E X G . R is k an al ys is a nd p ro ce ss o pe ra bi lit y H A Z O P T ab le s. G G M so lu tio n su pp ly st re am s t o th e R -2 04 A -I r ea ct or s K ey w or d D ev ia tio n C au se (s ) C on se qu en ce (s ) M ea su re m en t( s) M or e T em pe ra tu re In co rr ec t o pe ra tio n in th e E- 20 2 / E -2 03 e xc ha ng er s. R ed uc ed re ac tio n tim e. A cc el er at io n of c at al ys t de ac tiv at io n Te m pe ra tu re c on tro l l oo p, re gu la tin g th e he at su pp lie d to th e E - 20 3 he at er . Pr es su re In co rr ec t o pe ra tio n of th e P- 20 1 pu m p. E xc es si ve p um p po w er . In cr ea se d st re ss o n pi pe lin es an d th e re ac to r. C on tro l o f p re ss ur e in si de th e re ac to r an d pr es su re a la rm s a t t he p um p di sc ha rg e. F lo w c on tro l. Fl ow In cr ea se d flo w in u ps tre am eq ui pm en t, ov er -d riv in g of pu m ps . R ed uc ed re ac to r l oa di ng tim e. P os si bl e ov er fil lin g. In cl ud e flo w re gu la tio n co nt ro l l oo p. In cl ud e le ve l a la rm . M in us T em pe ra tu re In co rr ec t o pe ra tio n in th e E- 20 2 / E -2 03 e xc ha ng er s. In cr ea se d re ac tio n tim e. Te m pe ra tu re c on tro l l oo p. Im pl em en ta tio n of a n au xi lia ry h al f- sh el l i n th e re ac to r. Pr es su re In co rr ec t o pe ra tio n of th e P- 20 1 pu m p. E xc es si ve p um p po w er . Pa rti al e va po ra tio n an d tw o- ph as e flo w in p ip es . L ow pr es su re in th e re ac to r. Pr es su re a la rm s o n pu m ps . R ea ct or pr es su re c on tro l l oo p. Fl ow In cr ea se d flo w in u ps tre am eq ui pm en t, ov er -d riv in g of pu m ps . Lo ng er lo ad in g tim e, re du ce d pr od uc tio n. In cl ud e flo w re gu la tio n co nt ro l l oo p. In cl ud e le ve l a la rm . N o Fl ow B re ak ag e of p re vi ou s e qu ip m en t or o bs tru ct io n in th e pi pe s. Pa us e in p ro du ct io n. Fl ow c on tro lle r. Le ve l a la rm s i n bu ff er ta nk s. U se o f f ilt er s i n pu m p ad m is si on . 11 8 G G M so lu tio n di sc ha rg e st re am s t o th e R -2 04 A -I r ea ct or s K ey w or d D ev ia tio n C au se (s ) C on se qu en ce (s ) M ea su re m en t( s) M or e T em pe ra tu re U ne xp ec te d re ac tio ns in th e re ac to r. Po ss ib le c on ta m in at io n of fo od . M or e ex ch an ge o n pr eh ea te r E- 20 2. P os si bl e de ac tiv at io n of th e ca ta ly st . Te m pe ra tu re m ea su re m en t. H ea te r E- 20 3 co nt ro l l oo p. Pr es su re Ex ce ss iv e re ac to r pr es su riz at io n. U ne xp ec te d ga s pr od uc tio n. In cr ea se d m ec ha ni ca l s tre ss on th e re ac to r a nd in te rn al el em en ts . C on tro l o f p re ss ur e in si de th e re ac to r an d pr es su re a la rm s a t t he p um p di sc ha rg e. F lo w c on tro l. R el ie f va lv e. Fl ow O ve r- dr iv in g of th e P- 20 5 pu m ps . R ed uc ed re ac to r d is ch ar ge tim e. L ow er e ne rg y re co ve ry . In cl ud e co nt ro l l oo p to re gu la te th e di sc ha rg e flo w . I nc lu de le ve l a la rm s. M in us T em pe ra tu re D ra st ic re du ct io n of th e am bi en t t em pe ra tu re . I ns ul at io n fa ilu re . Le ss e ne rg y re co ve ry . R ed uc ed c on ve rs io n in th e pr ev io us st ag e. Te m pe ra tu re c on tro l l oo p. Im pl em en ta tio n of a ux ili ar y ha lf- sh el l i n th e re ac to r. Pr es su re In co rr ec t o pe ra tio n of th e P- 20 5 pu m p. P os si bl e br ea ks in th e pi pe s. Pa rti al e va po ra tio n an d tw o- ph as e flo w in p ip es . L ow er en er gy re co ve ry . Pr es su re a la rm s o n pu m ps . R ea ct or pr es su re c on tro l l oo p. Fl ow In cr ea se d flo w in u ps tre am eq ui pm en t, ov er -d riv in g of pu m ps . Lo ng er lo ad in g tim e, re du ce d pr od uc tio n. In cl ud e flo w re gu la tio n co nt ro l l oo p. In cl ud e le ve l a la rm . N o Fl ow R ea ct or b re ac h. B lo ck ag e of th e di sc ha rg e fil te r. B re ak ag e or o bs tru ct io n in th e co nd uc tio n. Pa us e in p ro du ct io n. Fl ow c on tro lle r. Le ve l a la rm s i n bu ff er ta nk s. U se o f f ilt er s i n pu m p ad m is si on . R ep la ce m en t o f d am ag ed ca ta ly st . 11 9 R ea ct or P re ss ur iz at io n St re am R -2 04 A -I K ey w or d D ev ia tio n C au se (s ) C on se qu en ce (s ) M ea su re m en t(s ) M or e T em pe ra tu re H ot a ir in ta ke . H ig h am bi en t te m pe ra tu re V er y sl ig ht in cr ea se in re ac tio n ki ne tic s d ue to te m pe ra tu re . It is n ot c on si de re d a pr ob le m at ic ph en om en on . Pr es su re R ea ct or p re ss ur e co nt ro l sy st em fa ilu re . In cr ea se d m ec ha ni ca l s tre ss in th e re ac to r. In cr ea se d di sc ha rg e flo w . In cl ud e co nt ro l l oo p to re gu la te th e di sc ha rg e flo w . I nc lu de p re ss ur e al ar m a nd re lie f v al ve s M in us T em pe ra tu re C ol d ai r i nt ak e. L ow a m bi en t te m pe ra tu re . V er y sl ig ht re du ct io n of th e re ac tio n te m pe ra tu re . Im pl em en ta tio n of a n au xi lia ry h al f- sh el l i n th e re ac to r. Pr es su re A ux ili ar y co m pr es so r f ai lu re . Im po ss ib le to c on tro l p re ss ur e in th e re ac to r. R ea ct or p re ss ur e al ar m s. H ea tin g st re am o f t he h al f p ip e ja ck et o f t he R -2 04 A -I r ea ct or s, by m ea ns o f m ed iu m p re ss ur e st ea m . ( M PS ) K ey w or d D ev ia tio n C au se (s ) C on se qu en ce (s ) M ea su re m en t(s ) M or e Fl ow V al ve fa ilu re o r s te am su pp ly Ex ce ss iv e he at in g of th e re ac tio n m ix tu re . C on tro l l oo p fo r t he a ux ili ar y he at in g of th e re ac to r a cc or di ng to th e te m pe ra tu re o f t he re ac tio n m ix tu re Im pl em en ta tio n of a n au xi lia ry h al f- sh el l i n th e re ac to r. R ea ct or p re ss ur e al ar m s. Pr es su re In cr ea se in th e te m pe ra tu re o f th e st ea m . In cr ea se d he at e xc ha ng e. M in us Fl ow V al ve fa ilu re o r s te am su pp ly Ex ce ss iv e he at in g of th e re ac tio n m ix tu re . Pr es su re Lo w er st ea m te m pe ra tu re . Po ss ib le e xc es si ve co nd en sa tio n on th e sh irt . 120 ANNEX H. Economic studies. Calculation of the NPV of the installation. (Ui = 21,4 %; i = 35%). Year V (€ year-1) C (€ year-1) A (€ year-1) Investment (€) Cash Flow (€ year-1) NPV (€) 0 0 0 0 42183506 -42183506 9864230 1 128676335 28040210 2812234 0 81310410 2 128676335 28601014 2812234 0 80869618 3 128676335 29173034 2812234 0 80420010 4 128676335 29756495 2812234 0 79961410 5 128676335 30351625 2812234 0 79493638 6 128676335 30958658 2812234 0 79016510 7 128676335 31577831 2812234 0 78529840 8 128676335 32209387 2812234 0 78033437 9 128676335 32853575 2812234 0 77527105 10 128676335 33510647 2812234 0 77010647 11 128676335 34180860 2812234 0 76483859 12 128676335 34864477 2812234 0 75946536 13 128676335 35561766 2812234 0 75398467 14 128676335 36273002 2812234 0 74839436 15 128676335 36998462 2812234 0 74269224 16 128676335 37738431 2812234 0 73687608