PRE-TREATMENT TECHNOLOGIES, AND THEIR EFFECTS ON THE INTERNATIONAL BIOENERGY SUPPLY CHAIN LOGISTICS

Transcription

1 PRE-TREATMENT TECHNOLOGIES, AND THEIR EFFECTS ON THE INTERNATIONAL BIOENERGY SUPPLY CHAIN LOGISTICS Techno-economic evaluation of torrefaction, fast pyrolysis and pelletisation Ayla Uslu Report number: NWS-I December 2005

2 PRE-TREATMENT TECHNOLOGIES AND THEIR EFFECTS ON THE INTERNATIONAL BIOENERGY SUPPLY CHAIN LOGISTICS Techno-economic evaluation of torrefaction, fast pyrolysis and pelletisation Supervision: Dr. Andre Faaij P.C.A. Bergman Ayla Uslu This study was concluded in partial fulfilment of the Master of Science program in Sustainable Development- Energy and Resources. Department of Science, Technology & Society Utrecht University, the Netherlands Part of the research was conducted in the Energy research Centre of the Netherlands (ECN). Biomass Department, Petten, the Netherlands 2

3 ACKNOWLEDGEMENT...6 ABSTRACT INTRODUCTION GENERAL BACKGROUND PROBLEM FORMULATION OBJECTIVES METHODOLOGY AND EVALUATION CRITERIA REPORT STRUCTURE TECHNOLOGIES TORREFACTION PYROLYSIS PELLETISATION COMPARISON OF PROCESSES SENSITIVITY ANALYSES FOR PRE-TREATMENT TECHNOLOGIES FINAL CONVERSION CHAIN ANALYSIS APPROACH AND METHODOLOGY LOGISTIC OPERATIONS DESIGNED CHAINS CHAIN ANALYSIS SENSITIVITY ANALYSIS DISCUSSION AND CONCLUSION PRE-TREATMENT TECHNOLOGIES CHAIN ANALYSIS RECOMMENDATIONS REFERENCE LIST APPENDICES

4 LIST OF TABLES Table 1 Net calorific values (LHV) of untreated wood, torrefied biomass, charcoal and coal Table 2 Comparison of the three evaluated reactor types for 150 MWth output torrefaction process Table 3 Wood, torrefied biomass, wood pellets and torrefied pellets property comparison Table 4 Technical performance of torrefied wood pelletisation for sawdust and green wood chips Table 5 Overall energy balance of the 150 MWth torrefaction process Table 6 Local parameters assumed for different world regions Table 7 Total capital investment of a MW th torrefaction plant Table 8 Specific investment cost calculations of various capacities Table 9 Typical properties of wood derived crude bio-oil Table MWth-input capacity pyrolysis process investment cost calculations (sawdust as feedstock)37 Table 11 Economic parameters for a 25 MWth biomass pyrolysis plant Table 12 Production cost of a 25 MWth rotating cone pyrolysis plant Table 13 Characteristics of wood pellets (Sawdust, cutter shavings, and wood-grinding dusts as raw materials Table 14 Cost data of pelletisation process Table 15 Technical comparison of torrefaction, TOP, pyrolysis and pelletisation pre-treatment processes Table 16 Economic comparison of torrefaction, TOP, pyrolysis and pelletisation pre-treatment processes Table 17 Main parameters used and ranges for sensitivity analysis Table 18 Effects of scale on pellet production costs Table 19 Cost figures for final conversion step Table 20 Characteristics of Latin American energy crops considered in this study (Ranges indicate short and long term,) Table 21 Logistics of biomass from harvesting to the pre-treatment process point Table 22 Designed chains Table 23 Parameter used for sensitivity analysis and ranges Table 24 Techno-economic comparison of torrefaction, TOP, pelletisation and pyrolysis Table 25 Costs of chains delivering fuel and power LIST OF FIGURES Figure 1 Main physico-chemical phenomena during heating of lignocellulosic materials at pre-pyrolytic conditions (torrefaction). Main decomposition regimes are based on Koukios et al. (1982)...13 Figure 2 Stages in the heating of moist biomass from ambient temperature to the desired torrefaction temperature and the subsequent cooling of the torrefied product (Bergman et al, 2005)...14 Figure 3 Schematic representation of the Pechiney process (Berman et al, 2005)...17 Figure 4 General flow diagram of ECN torrefaction process (Bergman at al, 2005)...18 Figure 5 Torrefaction process authothermal operation equations...19 Figure 6 General flow diagram of torrefaction in combination with pelletisation...21 Figure 7 Net mass flows corresponding with torrefaction of cuttings at 280 o C and 17.5 min reaction time (HE: Heat exchanger) Figure 8 Net energy flows (in MW th ) corresponding with torrefaction of woodcuttings at 280 C and 17.5 min reaction time (HE: heat exchanger)...25 Figure 9 Total capital investment cost breakdown Figure 10 Equipment cost breakdown...29 Figure 11 State of the art Ensyn RTP plant flow diagram (Source)...34 Figure 12 A schematic presentation of the mass and energy balance of a 5-ton dry feed per hour pyrolysis plant...35 Figure 13 Specific Pyrolysis Plant Investments (Solantausta, 2001)

5 Figure 14 Specific Pyrolysis Plant Investments (Historic data from1987 to 2003 (Solantausta, 2001)...40 Figure 15 Fast pyrolysis total plant costs versus feedstock (dry) input (Bridgwater et al, 2002) Figure 15 Specific investment cost versus capacity (MWth LHV ) data of 5 different pyrolysis plants...42 Figure 17 Total capital investment cost versus feedstock (ton dry /hr) of 5 different pyrolysis plants (Graph uses the same data as Figure 15)...43 Figure 17 State-of-the-art pellet production process diagram...45 Figure 18 Mass balance of the pelletising process...48 Figure 19 Energy balance for a pellet production process (Cost data is obtained from Thek and Obernberg (2004)) Figure 20 Cost breakdown of the total capital investment cost of a ton pellet process (data from Thek and Obernberg, (2004) is used)...52 Figure 21 Sensitivity of torrefied biomass production...55 Figure 22 Sensitivity analysis of bio-oil production costs...56 Figure 23 Pellet production sensitivity analysis...56 Figure 24 Effects of scale on torrefaction investment costs...57 Figure 25 Effect of scale on the bio-oil production costs...58 Figure 26 Biomass transport overview...63 Figure 27 Schematically presentation of transportation distances...66 Figure 28 Modelled bio-energy chains from Latin America to West Europe...68 Figure 29 Cost data of chains delivering pellets in / ton dry delivered Figure 30 Cost of bio-oil delivered to West Europe in /GJ HHV...71 Figure 31 Costs of FT liquid for different pre-treated feedstock (Conversion in the graph comprises pretreatment and FT processes) Figure 32 Cost of the chains delivering electricity by means of BIGCC...73 Figure 33 Cost of power obtained by combustion for various bio-energy chains...74 Figure 34 Cost of power obtained by co-firing for various bio-energy chains...74 Figure 35 Energy use of chains delivering biomass to the Rotterdam harbour in GJ/ ton dry delivered...75 Figure 36 Energy use of bio-oil delivered to West Europe in GJ/GJ HHVbio-oil...76 Figure 37 Energy use of the chains delivering FT liquid...76 Figure 38 Primary energy use of chains delivering energy...76 Figure 39 Primary energy use for chains delivering electricity...77 Figure 40 CO 2 emission of chains delivering biomass to West Europe in kg CO 2 / ton dry delivered...78 Figure 41 CO 2 emission caused by bio-oil delivery from Latin America to West Europe Figure 42 CO2 emissions of chains delivering FT liquid...79 Figure 43 Comparison of the power cost delivery figures for every chain Figure 44 CO 2 emissions of the chains delivering electricity (conversion in this graph means the pretreatment steps and emissions count for the electricity utilised during pre-treatment) Figure 45 Sensitivity analysis for TOP process (OW: harvest operation period)...82 Figure 46 Sensitivity analysis of TOP delivery versus pre-treatment unit scale...83 Figure 47: Relation between the MCA and the energy yield of torrefaction. Values are taken from the main design matrix. The error bars represent the possible inaccuracy in the HHV measurement Figure 48: Size reduction results of various torrefied biomass and feed biomass...95 Figure 49 Bench scale pyrolysis unit flow diagram Figure 50 BioTherm pyrolysis process flow diagram Figure 51 RTI process for production of bio-oil, bubbling fluidised bed type of reactor Figure 52 Circulating fluid bed reactor Figure 53 schematic diagram of the CFB unit for biomass flash pyrolysis Figure 54 Rotating cone reactor Figure 55 Rotating cone reactor pyrolysis process flow diagram

6 Acknowledgement This research has been performed at Utrecht University and at the Energy Research Centre of the Netherlands (ECN). First, I would like to thank my supervisor Andre Faaij, for his support and guidance. I also thank to Patrick Bergman who has never hesitated to share his valuable knowledge, advice and who has supervised my work during my internship at ECN. Many thanks to Jaap Kiel, who made the internship at ECN possible. Special thanks go to my friends and my family. They give a special colour to my life, which makes it more valuable. 6

7 Abstract The need for supplying sustainable energy resources raises the urgency of finding optimised bioenergy chains. Pre-treatment step is one of the factors, that has a significant influence on the overall chain. Torrefaction, pelletisation and pyrolysis technologies convert biomass into dens energy carriers that ease the transportation and handling. Besides, they influence the final conversion step. In this study, first the technical and economic performances of the pre-treatment technologies are assessed. Next, the economy of scale of each technology is analysed and the impacts of the pre-treated energy carriers on the final conversion stage is evaluated. Finally, several scenarios are produced, where Latin American energy crops are pre-treated and intermediate energy carriers are delivered to Western Europe and converted into power and syn-fuel. In this chain analysis step, not only the technologies but also the different scales are compared with each other. The technology assessment part indicates that torrefaction is a very promising technology due to its high process efficiency compared to pelletising and pyrolysis technologies. Moreover, when torrefaction is combined with pelletisation, the product (TOP pellets) energy content is as high as GJ/ton. For comparison, conventional pellet energy content is 17.7 GJ/ton while pyrolysis oil energy content is 17 GJ/ton. Another important point is that above 40 MWth torrefaction plant capacities, there is no economy of scale. When the economics of the three pre-treatment technologies are compared, pelletisation has the lowest specific capital investment, followed by torrefaction. However, there is a significant variation between the cost figures found in scientific literature for pyrolysis technology. The bioenergy chain analysis indicates that 89% of the biomass initial content can be delivered as cheap as 73.4 /ton (3.34 /GJ) in the form of torrefied pellets (TOP). In fact, TOP process increases the bulk density 15% compared to conventional pellets, which lowers the first truck transport. When the biomass is converted to pellets and delivered to Rotterdam harbour, the cost is 3.94 /GJ HHV, which is very close to TOP pellet delivery. On the other hand when the biomass is delivered to the Rotterdam harbour in the form of pyrolsyis oil, the cost is in the range of /GJ HHV. Furthermore, this study indicates that electricity can be produced as little as 4.4 cent/kwhe from an existing co-firing plant, while the electricity cost from a BIGCC facility is 4.6 cent/kwhe even though it includes high amount of investment costs. Fisher Tropish fuel produced in Europe costs 6.44 /GJ HHV for TOP pellets and 6.97 /GJ for conventional pellets delivered from Latin America. On the other hand, pyrolysis liquid can be converted into FT liquid with a cost of 9.5 /GJ HHV. In addition, the energy requirement indicates how sustainable the designed chains are. For TOP delivery, the primary energy requirement is as low as 0.05 GJ/GJ delivery, while it is 0.12 GJ/GJ delivery for pellets and 0.08 GJ/GJ HHV for pyrolysis oil. 7

8 1 Introduction 1.1 General Background Sustainable development is defined by the Brundtland Commission as a process of change where the exploitation of resources, the direction of investments, the orientation of technological development and institutional changes are all in harmony and enhance both current and future potential to meet human needs and aspirations. In this respect sustained energy supply is one of the essential targets to be achieved and depends on secure and reliable energy sources. The current global energy development pattern brings, however, significant danger due to the reliance on fossil fuels. Most of the world s oil reserves are located in certain countries, which make energy supply vulnerable. Another problem that has emerged is that fossil fuel consumption causes substantial environmental harm. In the IPCC 2 nd report, it is mentioned that human activity has caused a rise in atmospheric temperature in the recent years. In addressing those threats, developed countries increasingly shift towards renewable energy sources. Biomass is one of the renewable energy sources that are available worldwide. It is abundant throughout the planet and can be used as an energy source with no CO 2 addition to the environment. Currently, bio-energy contributes 35% of the primary energy consumption in developing countries, which accounts for 9-14% of global energy demand (Hamelinck, 2004). According to some IPCC SRES 1 future market scenarios, 30% of the total energy supply is ascribed to biomass in 2100 (Nakicenovic and Swart, 2000). Those figures are the indicators for a future biomass based energy market potential. In fact, in some countries such as Brazil, Sweden and Finland bio-energy markets have already emerged. Moreover, large scale biomass trade can be the centre of attention due to European legislations concerning climate policies. For example, in the agreement of the Dutch government with Dutch energy sector, it is stated that 6 Mton/year of fossil fuel CO 2 reduction needs to be accomplished in by the coal-fired power stations in the Netherlands. Half of this reduction is planned to be accomplished through replacing coal by biomass (Bergman et al, 2005). Another phenomenon that triggers the international biomass trade is that some countries have larger biomass resources compared to other countries. For example; Latin America holds high biomass energy production potential (Hoogwijk et al. 2003; Damen and Faaij 2003; Agterberg and Faaij 1998). Several researches have given indications that intercontinental trade of biofuel could be economically feasible (Hamelinck, 2004; Wasser and Brown 1995; Agbert and Faaij, 1998). Hamelinck et al (2003) has conducted a study to assess whether large-scale long distance transport of bio fuel from certain world regions, such as Latin America and Eastern Europe, is economically and energetically feasible and attractive in terms of greenhouse gas (GHG) reduction. International biomass logistics were analysed by conducting several bio energy chains. These studies exposed that the pre-treatment step plays an eminent role in the whole chain since it affects the storage, transport and final conversion steps. Broadly, feedstock costs contribute 1 IPCC SRES; International Panel on Climate Change (IPCC) Special Report on Emissions Scenario (SRES) 8

9 20%-65% of the total delivery cost whereas pre-treatment and transport contribute 20%-25% and 25%-40%, respectively, depending on the location of the biomass resource. When the overall chain is considered, the final conversion contributes roughly to more than 50% of electricity or fuel delivered. Existing research has been conducted mostly to obtain design data, process performance, product properties and the by product composition of the pre-treatment technologies. However, their techno-economic performances on the total bioenergy supply chain require detailed study. In this context, the key technologies, pyrolysis, torrefaction and pelletising need to be analysed in terms of technical performances and economy of scale. Pyrolysis and torrefaction are the thermochemical conversion technologies where bio-oil and torrefied biomass are produced in different temperature ranges respectively. On the other hand in pelletisation biomass is dried and compressed to produce cylindrical pieces. Various bio-energy chains can be designed where these technologies are considered and depending on the economy and impacts on environment in terms of GHG emissions, the optimal chain can be obtained. 1.2 Problem formulation The studies concerning long distance bio-energy transport analysed several cases to perform the biomass delivery and final energy production costs. Hamelinck developed a tool with which different bioenergy chains can be analysed. This tool enables to assess the influence of different pre-treatment technologies on the technical and economic performance of the whole chain. It is clear from the work of Hamelinck (2004) that energy densification of the biomass is crucial. Converting biomass into a densified intermediate can save transport and handling costs. In addition, it can improve the efficiency of the final conversion stage. Subsequently, pre-treatment methods deserve more attention in the positioning of the chain and techno-economic analysis of the treatment processes themselves. Torrefaction, pyrolysis and pelletising are the pre-treatment technologies considered in this study. Currently, the state-of-the-art biomass-to-energy chains are mostly based on energy densification by means of pelletisation. However, commercially available bio pellets are expensive and cannot be produced economically from a wide variety of biomass resources (only sawdust and planer shavings) when the small (smaller than 30 mm) particle sizes are required. New pre-treatment technologies are currently under development. Fast pyrolysis, charcoal production and torrefaction may improve economics of the overall production chain. However, these technologies are still under development and their economic and technical performances are unclear. Moreover, there is no normalised data set, which can give a clear picture of the economic performances. Available information, however, mainly discusses the technology and the intermediate products they produce, rather than their influence on the performance of the whole production chain. Therefore, the key research question in this study is: Which pre-treatment method(s), at what point of the chain, with which conversion technology (ies) would give the optimal power and fuel (syngas) delivery costs for international biomass supply chains? 9

10 1.3 Objectives The main objective of this study is to identify the optimal bioenergy chain design and provide insight in the differences between the torrefaction, pyrolysis and pelletisation pre-treatment technologies. This can be done by performing a techno-economic analysis of the biomass-toenergy production chain that is based on pelletisation, torrefaction and pyrolysis. Moreover, this study focuses on the influence of the pre-treated biomass (intermediates) on transportation. The sub-objectives are: -Analysing the potential and future technical and economic performances of torrefaction, pyrolysis and pelletising in relation to scale. -Evaluating the influence of intermediate products on transportation. -Analysing the impacts of intermediate products on power and syngas conversion technologies. As pelletisation is commercially applied, this technology is considered the state-of-the-art reference (SOTA-system). Hence, the introduction of alternative pre-treatment technologies as pyrolysis and torrefaction are only interesting when they are comparable or better than the economics of this reference. However, the improvement options on pelletisation still need to be investigated even though it is commercially applied. The bio-energy techno-economic analysis is based on biomass resources located in South America (Brazil) and the final conversion to power and fuels is situated in North-West of Europe. 1.4 Methodology and evaluation criteria A technology review is performed where the design data of pre-treatment technologies are collected to determine the current technology status of pyrolysis, torrefaction and pelletisation. The state-of-the-art processes are identified according to their commercial or demonstrated applications. Following the literature survey, mass yields, energy yields and process efficiencies of each technology are presented. Economic evaluation is conducted by calculating the required capital investments and total production costs. The capital costs are based on component level cost data, which are obtained from literature and personal communication. Since the capacities of the components affect the specific cost of a plant, economy of scale is analysed. This is done by identifying the base scales, the base costs and the maximum scales of the equipments. Next, the actual equipment costs are calculated using the scale factor R. R-values per component are obtained from literature (Faaij et al., 1998; Bergman et al, 2005). Finally, the total capital investment requirements are calculated (See Appendix 1). This is followed by the sensitivity analysis of the parameters that influence the production costs The cost data are normalised using the OECD deflator and exchange rates of national currencies per US$(See Appendix 2). In addition, the impacts of the pre-treated intermediates on the final conversion step are assessed. Entrained flow gasification for Fischer Tropsch liquid production, biomass integrated gasification combined cycle (BIGCC), combustion and co firing for power production are considered as final conversion technologies. Since, the final conversion technologies have got specific requirements like the feedstock particle size, shape; compressibility, bulk density, and moisture content, 10

11 intermediate energy carriers can alter the feedstock preparation and feeding system. Moreover, they can influence the gas-cleaning step that takes part in gasification. Therefore, impacts of pretreated intermediates on conversion technologies are investigated. Following the final conversion stage, several biomass to-energy chains are designed. In this part, Hamelinck s (2003) biomass logistic tool is used. This tool gives the possibility to set up harvesting, transport, storage, handling, pre-treatment and final conversion steps in many ways. It also enables to visualize the different combinations and scales of operations and carries out techno-economic analysis. Furthermore it allows conducting sensitivity analyses to test the robustness of the study results and assess the variation in fuel/power costs. South American bio-energy crops (eucalyptus) are considered as the source of biomass in this study. The final conversion stage is assumed to be applied in Rotterdam by means of Fischer- Tropsch (FT) and conventional power plants. The transport mediums are chosen to be trucks for local transport and ships for international transport. 1.5 Report structure The report begins with the review of the pre-treatment technologies. Chapter 2 presents the techno-economic analysis and the improvement options for torrefaction, pyrolysis and pelletisation processes. In addition to this, it provides the sensitivity analyses, scale effects and a brief techno-economic data comparison of pre-treatment processes. The influence of the pretreated intermediate on gasification and combustion in terms of efficiency and economy are also presented in this chapter Chapter 3 describes the approach and methodology used for biomass chain analysis. Moreover, it presents the designed chains and provides the results. Finally, Chapter 4 summarises the main conclusions of this study and presents the discussion and recommendations. 11

12 2 Technologies 2.1 Torrefaction Process definition Torrefaction is a thermal pre-treatment technology carried out at atmospheric pressure in the absence of oxygen. It occurs at temperatures between oc where a solid uniform product is produced. This product has very low moisture content and a high calorific value compared to fresh biomass. The origin of torrefaction comes from roasting of coffee beans. This process however has been done in lower temperatures in the presence of oxygen. Since the 1930 s on, torrefaction has been used in relation to woody biomass. The research outcomes in those years appeared to be favourable at the technical stage; however, torrefied wood did not find the market outlets. In France, torrefaction has been applied to produce a wood that is used as a building material (Bioenergy, 2002). In the 1980s the focus was on using torrefaction technology to produce wood that can be used as a reducing agent in metallurgic applications. A torrefaction plant demonstration with a production capacity of ton/year was constructed and operated (Bergman et al, 2005). In recent years, torrefaction has gained significant importance for energy applications. Even though it is still in its infancy, several studies have shown that torrefaction upgrades the energy density, hydrophobic nature and grindability properties of biomass (Bergman et al, 2005; Bioenergy 2000; Prins, 2004). Through torrefaction biomass is converted into torrefied biomass which is typically 70% of its initial weight and contains 90% of the original energy content (Bioenergy, 2000; Bergman et al, 2005). 30% of the biomass is converted into torrefaction gas that contains 10% of the initial energy content. The moisture uptake of torrefied biomass is very limited varying from1-6%. Destruction of OH groups in the biomass by dehydration reactions causes the loss of capacity to form hydrogen bonds with water. In addition, non-polar unsaturated structures are formed which makes the torrefied biomass hydrophobic (Bergman et al, 2005) Torrefaction decomposition mechanism The polymeric structure of woody and herbaceous biomass comprises mainly cellulose, hemicellulose and lignin. The most reactive polymer is hemicellulose whereas the cellulose is the thermo stable part. At low torrefaction temperatures decomposition occurs in the hemicellulose structure by means of a limited devolatilisation and carbonisation. In the lignin and cellulose structure however a minor decomposition is expected. When the temperature is raised up to o C, hemicellulose extensively decomposes into volatiles and char-like solid products, whereas limited devolatilisation and carbonisation occur in the lignin and cellulose structure (See Figure 1, Bergman et al, 2005) 12

13 Hemicellulose Lignin Cellulose 300 Extensive Devolatilisation and E E 300 carbonisation (E) D Temperature ( C) Limited devolatilisation and carbonisation (D) D D C TORREFACTION Temperature ( C) C 150 depolymerisation and recondensation (C) 150 glass transition/ softening (B) A 100 drying (A) Hemicellulose Lignin Cellulose A Figure 1 Main physico-chemical phenomena during heating of lignocellulosic materials at prepyrolytic conditions (torrefaction). Main decomposition regimes are based on Koukios et al. (1982) Depending on the torrefaction conditions and the biomass properties, torrefaction products can be classified as solid, liquid and gas at room temperature. The solid phase, so called torrefied biomass, consists of original sugar structures and the reaction products, which are modified sugar structures, newly formed polymeric structures, char and ash fractions. The gas phase consists of mainly CO, CO 2, and traces of H 2, CH 4 and light aromatic components. And in the liquid phase; H 2 O from biomass thermal decomposition, organics from devolatilisation and carbonisation and lipids consist. Nonetheless, torrefaction conditions; reaction time and torrefaction temperature are fairly important in product specification (Berman et al, 2005) Torrefaction conditions Reaction time and reactor residence time is defined clearly before clarifying the necessary torrefaction temperature and the significance of moisture content in the torrefaction process. In fact, biomass has to be heated through several stages before the real torrefaction regime is reached. In practice, the solid residence time in a reactor is never equal to the time which biomass particles are exposed to torrefaction (Bergman et al, 2005). In Figure 2, temperature-time stages

14 for a batch-wise operated torrefaction reactor is illustrated; however, those stages are the same for continuous operations. 300 T tor Biomass temperature ( C) t h t dry t tor,h t tor t tor,c t c t h,int post drying and intermediate initial heating pre-drying heating torrefaction solids cooling Moisture content time Heat demand (cummulative) Mass yield (dry) Figure 2 Stages in the heating of moist biomass from ambient temperature to the desired torrefaction temperature and the subsequent cooling of the torrefied product (Bergman et al, 2005). Temperature-time profile is considered to be typical for a batch-wise operated reactor. Explanation: t h = heating time to drying, t dry = drying time, t h,int = intermediate heating time from drying to torrefaction, t tor = reaction time at desired torrefaction temperature, t tor,h = heating time torrefaction from 200 C to desired torrefaction temperature (T tor ), t tor,c = cooling time from the desired T tor to 200 C, t c = cooling time to ambient temperature. In the initial heating stage, biomass moisture content evaporation is very slow; nonetheless, the biomass temperature increases. In the pre-drying stage, moisture content decreases dramatically while the biomass temperature stays constant. Following this stage, post-drying and intermediate heating occurs. The temperature increases up to 200 oc and the physically bounded water is time 14

15 released. Above 200 o C torrefaction reaction occurs. Devolatilisation takes part in this stage. And finally, solid product is cooled to below 200 o C Product quality During torrefaction biomass loses relatively more oxygen and hydrogen compared to carbon, subsequently the calorific value of the product increases. The net calorific value (LHV dry ) of torrefied biomass is in the range of MJ/kg or MJ/kg when the HHV (dry) is concerned. In Table 1, the comparison of torrified biomass with raw wood, charcoal and coal is presented. Table 1 Net calorific values (LHV) of untreated wood, torrefied biomass, charcoal and coal Untreated Torrefied Charco Coal wood biomass al LHV dry (MJ/kg) Source: Bergman et al, 2005 The moisture uptake of torrefied biomass is very limited due to the dehydration reactions that take place during the torrefaction reaction. Those reactions prevent the bonding of biomass hydrogen with water. Another change in the biomass occurs is the volumetric density. The torrefied biomass becomes more porous with a volumetric density of 180 to 300 kg/m 3 depending on the initial biomass density and torrefaction conditions. It becomes more fragile as it looses its mechanical strength. As a result, torrefied biomass size reduction becomes easier, which is advantageous in conversion applications since they require biomass in very small sizes such as in the form of powder for entrained flow gasification Production technology Current status Torrefaction technology is not commercially available yet. However, torrefaction history dates back to the late 1980 s when the upgrading of wood and briquettes by means of torrefaction was investigated at the Asian Institute of Technology, Bangkok, Thailand. Another study was held in Brazil. The Grupo Combustíveis Alternativos (GCA) at the University of Campinas in Brazil used a bench unit for biomass torrefaction (Bioenergy, 2002). Those studies provided information about the product quality. The recent studies on torrefaction technology are based on the airless drying concept (Bioenergy, 2002) and the rotating drum concept (Duijn, 2004). In addition to this, at ECN extensive research has been carried out. They designed an optimum integrated process, and studied the impacts of torrefied biomass on co-firing and entrained flow gasification. In addition to this, they have developed a TOP process (Torrefaction and Pelletisation) which aims to bring TOP pellets to the energy market. The only commercially applied torrefaction plant, PECHINEY, was built up in 1980s in France and was operated for a few years. It was a continuous process which was applied by Le Bois 15

16 Torrefie du Lot, a subsidiary of Pechiney Electrometallurgie in order to produce roasted wood for use in manufacturing silicon steel (Bioenergy, 2002) State-of the- art-system (SOTA) description PECHINEY has built the first demonstration unit of a torrefaction process that was in operation since early 1987s with a capacity of ton/year of torrefied wood as a reduction agent for silicon metal (Girard and Shah, 2005). Since this is the only commercially built process, it is accepted as the state-of-the-art system. However, it should not be forgotten that the process conditions applied were different than it should be to yield a product that can be converted into energy in a later stage. The process applied at Pechniey mainly consisted of a chopper where grinding is done, a drying kiln, and a torrefaction reactor (roaster) (see Figure 3). The received wood was cut into 50 to 80 mm long and 15 mm thick chips at a rate of 20 ton/hr by a drum chopper. The pieces larger than 80 mm were removed at the outlet of the screen to be recycled and the fines that were smaller than 15 mm were sent to the boiler to be combusted. Following this operation the wood chips were transported to a tunnel kiln for drying. The chips, dried to 10%mc, were transported to the roaster (torrefaction reactor) which was a hot mixing device with a double sheath and a rotating shaft with disc sections perpendicular to the axis of the shafts. The reactor was heated by conduction and thermal oil was used as a heat transfer fluid. This fluid was recycled between boiler and the reactor. The gases generated in the roaster were combusted and the fumes were returned to the kiln after being de-dusted. The temperature at the end of the roaster was reduced (Bergman et al, 2005). 16

17 large rejects Wood Storage Chopper Wood Screen wood fines to boiler flue gas incinerator combustibles TW leisure TW Drying kiln Roaster Screen TW metallurgy thermal liquid TW fines wood fines from wood screen Boiler Figure 3 Schematic representation of the Pechiney process (Berman et al, 2005) The Pechiney process aimed at a product that has got a fixed carbon content with homogenized moisture content. The operation temperature was in the range of 240 to 280 o C with a residence time of 60 to 90 minutes. Various problems existed with this reactor design when it was considered for bigger scales. The heat exchange area in this reactor was a limiting factor. The feed moisture content was limited to 15% while the reactor through put was limited to 2 ton/hr. Higher moisture contents would drop the reactor throughput. Another disadvantage was that the reactor required free-flowing feed particles (Bergman et al, 2005). The energy efficiency of Pechiney was calculated as 65-75%. The low process efficiency was caused mainly by the feedstock losses during the chipping and sieving steps Assessment of the Pechiney process The investment cost of Pechiney process was approximately 2.9 M in 1985 with a specific investment of 25 /ton product (Bergman et al, 2005). More than 80% of the total investment cost was arising from the reactor. The torrefied wood production was roughly 100 /ton when the feedstock costs were excluded. When the feedstock costs were included, the production cost would approximately be around /ton. Process scaling up could reduce the production costs. However, the reactor used in the Pechiney process was poor in scaling up properties and expensive (Bergman et al, 2005). These properties imply the need to search further for a better 17

18 process technology. In fact, a torrefaction process with a capacity of 150 MWth is designed by ECN ECN Torrefaction technology The ECN torrefaction process is based on direct heating of the biomass during torrefaction by using the hot gas that is recycled. The torrefaction gas is re-pressurised and heated before it is recycled to the reactor. The necessary heat for drying and torrefaction is produced by combustion of liberated torrefaction gas. In fact, the energy content of the torrefaction gas plays a significant role in providing the energy demand of the dryer and the reactor. When the energy obtained from the torrefaction gas is equal to the energy demand of the dryer and the reactor, the system operates authothermal. On the other hand when the energy content of the gas is not enough, utility energy needs to be used. The moisture content of the feedstock is extremely important since the feedstock property determines the required heat demand. The relationship between the feedstock moisture content and energy yield in ECN process can be summarised by; the wetter the biomass feedstock is the lower the energy yield to be allowed in torrefaction in order to perform authothermal operation. In Appendix 3 the relationship between authotermal operation of the process in relation to the moisture content and energy yield of torrefaction is presented. The energy content of the torrefaction gas is directly related to the solid and energy yield of torrefaction. Therefore, when torrefaction is operated in a different condition that produces less energy yield, there will be surplus of energy in the form of torrefaction gas, in contrast when the energy yield of the process is higher than the optimised conditions this time the energy produced from the gas will not be enough to meet the dryer and reactor demand. This phenomenon is explained in Figure 5. TheECN torrefaction process consists of dryer, reactor, heat exchanger, combustion and cooling (Figure 4). Air utillity Fuel Fluegas biomass Torrefaction gases gas recycle Combustion Drying Torrefaction Cooling Torrefied biomass DP Fluegas Heat exchange Fluegas Figure 4 General flow diagram of ECN torrefaction process (Bergman at al, 2005) 18

19 (DP: Pressure Drop recovery) Qgas Dryer Torref.reactor Qdryer Qreact. Qgas = Qdryer + Q react. Figure 5 Torrefaction process authothermal operation equations In this design the torrefaction gas is expected to be combusted so that the energy demand of the dryer can be met without or with a little utility fuel consumption. Thus, this would result in a self supporting, high efficient system. The calorific value of the gas ranges from 5.3 MJ/Nm 3 to 16.2 MJ/Nm 3 at the temperatures of 265 o C and 290 o C, respectively. 19

20 Components of the system Dryer: Directly heated rotating drum technology was selected to dry the biomass from 50% moisture content to the desired 15%. Dryer design was based on flue gas re-circulation and modelled by ECN (Bergman, 2004 a ). Reactor: During the process design at ECN, three different reactor types are considered. Screw reactor was one of the options since it was used in the SOTA process (Pechiney). The second option was the directly heated rotating drum technology due to its many applications as a dryer and the third option was the directly heated moving bed. Directly heated moving bed gains from its compactness (high fill percentage), simplicity in construction, high heat transfer rates and small reactor size requirement. Another advantage of the moving bed reactor is that there is not a specific feedstock shape requirement. It can handle non-free flowing materials. In fact, design studies in ECN resulted in favour of the moving bed reactor as the most promising technology. The three reactor comparison is summarised in the following table (Table 2). Table 2 Comparison of the three evaluated reactor types for 150 MWth output torrefaction process Heat transfer Total Fill Shape of Costs 2 coefficient 1 residence time material input W/m 2 /K minutes % Indirectly Free-moving expensive heated screw Directly Free-moving medium heated rotating drum Directly flexible cheap heated moving bed 1 Estimated heat transfer coefficient is based on m 2 exchange area for the screw reactor, on m3 reactor volume for the rotating drum and on m 2 particle surface area for moving bed. 2 Equipment purchase costs are compared with each other; moving bed reactor is roughly 6 times cheaper than the screw reactor. Source: Bergman et al, Cooler The torrefied biomass was cooled down to 50 o C by an indirectly heated rotating drum technology. Water was chosen as a coolant Torrefied biomass densification In the section torrefied biomass is defined as a porous product, with a low density. However, torrefied biomass is fragile which makes it relatively easy to grind. On the other hand decreased mechanical strength and increased dust formation capacity in addition to low volumetric density makes a densification stage necessary. Besides, when the long distance 20

21 transport is considered, especially the shipment, pelletising the torrefied biomass is inevitable. This subject is discussed further in the biomass supply chain chapter. The mass density of torrefied biomass pellets is about 22 MJ/kg whereas the energy density reaches up to 18 GJ/ m 3. Although this energy density is less than that of coal (20.4 GJ/m 3 ), it is still 20% higher than commercial wood pellets (Lipinsky et al, 2002; Bergman et al, 2005). So, torrefaction in combination with pelletising offers significant advantages when the biomass logistics are considered. Densification following torrefaction is considered in several studies (Lipinsky et al, 2002; Reed and Bryant, 1978; Koukios, 1993; Bergman et al, 2005). Those studies indicate that the pressure required for densification could be reduced with a factor of 2 at 225 o C, while the energy consumption of densification could be reduced by a factor of 2 compared to biomass pelletisation. Biomass pelletising consists of drying and size reduction prior to the densification. Steam conditioning is applied to soften the biomass fibres. Following densification bio-pellets are cooled down. However, when torrefaction is considered, steam pre-conditioning is not required since torrefied biomass is fragile. Following torrefaction, size reduction, densification and cooling can be achieved (Figure 6). Biomass Drying Torrefaction Size Reductio Densification Cooling Torrefied pellets Figure 6 General flow diagram of torrefaction in combination with pelletisation Size reduction: The power consumption for size reduction following torrefaction is reduces around % compared to biomass pelletisation (Bergman, 2005). Besides, production capacity of a chipper increases with a factor 7.5 to 15 compared to biomass. A simpler type of size reduction such as cutting mills, and jaw crushers can be deployed instead of hammer mills, which are used for the conventional pelletising process (Bergman, 2005). The experimental results of size reduction carried out at ECN (Bergman et al, 2005) are presented in Appendix 4. Densification: AT ECN, a piston press has carried out torrefied biomass densification experiments. This press was modified to press in different diameters of various biomass products under different torrefaction conditions. The comparison of torrified pellets with torrefied biomass, wood pellets and fresh wood is shown in Table 3 (Bergman, 2005). 21

22 Table 3 Wood, torrefied biomass, wood pellets and torrefied pellets property comparison. Properties Unit Wood Torref.biomass Wood pellets Torref. pellets Low High Low High Moisture content %wt 35% 3% 10% 7% 5% 1% Calorific value (LHV) Dry MJ/kg As received MJ/kg Mass density (bulk) Kg/m Energy density GJ/m Pellet strength - - Good Very good Dust formation Moderate High Limited Limited Hydroscopic nature Biological degradation Water uptake Hydro phobic Swelling/ water uptake Poor swelling/ Hydro phobic Possible Impossible Possible Impossible Pelletising the torrefied biomass not only increases the mass density but also the energy density. Besides, mechanical strength improves. According to the experiments at ECN, torrefied pellets can withstand 1.5 to 2 times the force exerted on conventionally produced pellets before breakage. The water intake capacity of the product was determined by immersing them into water for 15 hours. A gravimetric measurement device was employed to measure the water intake contents. The results showed that torrefied pellet water intake was very limited (up to 10-20% on mass basis) whereas wood pellets swell rapidly (Bergman, 2005). Technical performance characteristics of torrefied wood pelletising for sawdust and greenwood chips are given below (Table 4). Table 4 Technical performance of torrefied wood pelletisation for sawdust and green wood chips Item Unit Torrefied wood pelletisation (sawdust) Torrefied wood pelletisation (green wood chips) Feedstock capacity Kton/y Moisture content Wt. 57% 57% LHV ar feed MJ/kg Production capacity Kton/y MWth fuel Product Moisture content Wt. 3% 3% LHV ar product MJ/kg Cooling water m 3 /ton product Utility fuel MW th

23 Electricity consumption MW e Thermal efficiency 98.5% 96.5% Net efficiency 93.7% 90.8% a.r: as received 23

24 2.1.3 Technology evaluation Objectives & Methodology The focus in this section is on the technical feasibility of torrefaction process. This technology is not commercialised yet and most of the demonstrations have been done under laboratory condition. As a newly emerging technology, it requires technology analysis. In addition, the exploration of future development options is aimed at. In the analysis, the data obtained from the ASPEN flow sheet simulation package, which was created at ECN, is used. Overall mass and energy balances derived from those studies are used to calculate overall performance of the system. The process conditions are assumed to be the same as in the ECN torrefaction study (Bergman et al, 2005) Mass and energy balances According to the ECN study (Bergman et al, 2005) done under 280 o C temperature and 17.5 min reaction time conditions, the torrefaction reaction mass yield was determined as around 70% (Figure 7) whereas the mass yield of drying was around 60%. This number corresponds to the moisture content loss. In fact the biomass input moisture content was 50% and the biomass moisture content leaving the dryer was 15% kg/s Combustion flue gas(combustables +non combust.) kg/s 0.01 kg/s ash 3.62 kg/s torref. gas biomass kg/s kg/s 7.54 kg/s 20 kg/s Drying Torrefaction torrefied biomass HE kg/s flue gas Figure 7 Net mass flows corresponding with torrefaction of cuttings at 280 o C and 17.5 min reaction time (HE: Heat exchanger). The thermal efficiency of the whole process was calculated as 96% while the net efficiency was approximately 92 %( includes the utility consumption) (Figure 8 and Table 5). The thermal efficiency of the whole process is mainly determined by the efficiency of the drying. The energy flow in dryer increases from MWth to MWth (See Figure 7). The high torrefaction efficiency takes its roots from the authothermal operation where the produced torrefaction gas is combusted and the energy demand of the dryer and reactor is met. 24

25 21.4 Combustion 27.7 Torrefaction gas 7.97 Torrefied biomass biomass Drying Torrefaction Cooling HE 5.47 Fluegas Figure 8 Net energy flows (in MW th ) corresponding with torrefaction of woodcuttings at 280 C and 17.5 min reaction time (HE: heat exchanger). Fluegas Table 5 Overall energy balance of the 150 MWth torrefaction process Utility Unit Value Thermal output MWth 150 Thermal input MWth Electricity input Dryer Reactor Fired heater/heat exchanger MWth MWth MWth Axial flow fan MWe 1.28 Product cooler MWe 1 Air turbo blower MWe 0.4 Electricity input MWe 2.68 Thermal equivalent 1 MWth 6.7 Thermal efficiency 2 % 0.96 Net efficiency 3 % The efficiency of energy conversion is accepted as Thermal efficiency is calculated as: thermal output/thermal input 3 Net efficiency is calculated as: thermal output/ (thermal input + utility) 25

26 2.1.4 Economic evaluation Methodology Economic evaluation is based on the estimations of required capital investment and total production costs. Factorial method described by Peters and Timmerhaus (1991) is used as the reference source in the calculation assumptions (See Appendix 1). Direct costs consist of equipments and their installation, buildings, process and auxiliary, service facilities, yard improvements and the cost of land. Indirect costs include engineering & supervision, construction expense/constructer fee and contingencies. Both direct and indirect costs compose the fixed capital investment. The torrefied biomass production costs are calculated by dividing the total annual costs of a system by the produced amount of torrefied biomass. The total annual costs are; Annual investment Operation and maintenance Biomass feedstock Electricity demand The annual investment cost is calculated according to the below equation: I annul = α * It r α = L 1 (1 + r) where: I annual = annual investment cost α = the capital recovery factor It = total investment r = the discount rate L = the life time or depreciation period of the equipment The depreciation period was set to 10 years due to the expected lifetime of dryer and reactor. 8000h/year is assumed to keep the operation continuous. The chain analysis in this thesis is based on the assumption that biomass is imported from Latin America to West Europe and the biomass pre-treatment step is supposed to takes part in the Latin American region. Therefore, the parameters assumed for different world regions and used in the torrefied wood production cost calculation are presented in Table 6. The production cost breakdown estimations are presented in Appendix 5. 26