With a high availability of lignocellulosic biomass and various types of cellulosic by-products, as well as a large number of industries, Sweden is a country of great interest for future large scale production of sustainable, next generation biofuels. This is most likely also a necessity as Sweden has the ambition to be independent of fossil fuels in the transport sector by the year 2030 and completely fossil free by 2050. In order to reach competitive biofuel production costs, plants with large production capacities are likely to be required. Feedstock intake capacities in the range of about 1-2 million tonnes per year, corresponding to a biomass feed of 300-600 MW, can be expected, which may lead to major logistical challenges. To enable expansion of biofuel production in such large plants, as well as provide for associated distribution requirements, it is clear that substantial infrastructure planning will be needed. The geographical location of the production plant facilities is therefore of crucial importance and must be strategic to minimise the transports of raw material as well as of final product. Competition for the available feedstock, from for example forest industries and CHP plants (combined heat and power) further complicates the localisation problem. Since the potential for an increased biomass utilisation is limited, high overall resource efficiency is of great importance. Integration of biofuel production processes in existing industries or in district heating systems may be beneficial from several aspects, such as opportunities for efficient heat integration, feedstock and equipment integration, as well as access to existing experience and know-how.
This report describes the development of BeWhere Sweden, a geographically explicit optimisation model for localisation of next generation biofuel production plants in Sweden. The main objective of developing such a model is to be able to assess production plant locations that are robust to varying boundary conditions, in particular regarding energy market prices, policy instruments, investment costs, feedstock competition and integration possibilities with existing energy systems. This report also presents current and future Swedish biomass resources as well as a compilation of three consistent future energy scenarios.
BeWhere is based on Mixed Integer Linear Programming (MILP) and is written in the commercial software GAMS, using CPLEX as a solver. The model minimises the cost of the entire studied system, including costs and revenues for biomass harvest and transportation, production plants, transportation and delivery of biofuels, sales of co-products, and economic policy instruments. The system cost is minimised subject to constraints regarding, for example, biomass supply, biomass demand, import/export of biomass, production plant operation and biofuel demand. The model will thus choose the least costly pathways from one set of feedstock supply points to a specific biofuel production plant and further to a set of biofuel demand points, while meeting the demand for biomass in other sectors.
BeWhere has previously been developed by the International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria and Luleå University of Technology and has been used in several studies on regional and national levels, as well as on the European level. However, none of the previous model versions has included site-specific conditions in existing industries as potential locations for industrially integrated next generation biofuel production. Furthermore, they also usually only consider relatively few different production routes. In this project, bottom-up studies of integrated biofuel production have been introduced into a top-down model and taken to a higher system level, and detailed, site-specific input data of potential locations for integrated biofuel production has been included in the model.
This report covers the first stages of model development of BeWhere Sweden. The integration possibilities have been limited to the forest industry and a few district heating networks, and the feedstocks to biomass originating from the forest. The number of biofuel production technologies has also been limited to three gasification-based concepts producing DME, and two hydrolysis- and fermentation-based concepts producing ethanol. None of the concepts considered is yet commercial on the scale envisioned here.
Preliminary model runs have been performed, with the main purpose to identify factors with large influence on the results, and to detect areas in need of further development and refinement. Those runs have been made using a future technology perspective but with current energy market conditions and biomass supply and demand. In the next stage of model development different roadmap scenarios will be modelled and analysed. Three different roadmap scenarios that describe consistent assessments of the future development concerning population, transport and motor fuel demands, biomass resources, biomass demand in other industry sectors, energy and biomass market prices etc. have been constructed within this project and are presented in this report. As basis for the scenarios the report “Roadmap 2050” by the Swedish Environmental Protection Agency (EPA) has been used, using 2030 as a target year for the scenarios. Roadmap scenario 1 is composed to resemble “Roadmap 2050” Scenario 1. Roadmap scenario 2 represents an alternative development with more protected forest and less available biomass resources, but a larger amount of biofuels in the transport system, partly due to a higher transport demand compared to Roadmap scenario 1. Finally Roadmap scenario 3 represents a more “business as usual” scenario with more restrictive assumptions compared to the other two scenarios.
In total 55 potential biofuel plant sites have been included at this stage of model development. Of this 32 sites are pulp/paper mills, of which 24 have chemical pulp production (kraft process) while eight produce only mechanical pulp and/or paper. Seven of the pulp mills are integrated with a sawmill, and 18 additional stand-alone sawmills are also included, as are five district heating systems. The pulp and paper mills and sawmills are included both as potential biofuel plant sites, as biomass demand sites regarding wood and bioenergy, and as biomass supply sites regarding surplus by-products. District heating systems are considered both regarding bioenergy demand and as potential plant sites.
In the preliminary model runs, biofuel production integrated in chemical pulp mills via black liquor gasification (BLG) was heavily favoured. The resulting total number of required production plants and the total biomass feedstock volumes to reach a certain biofuel share target are considerably lower when BLG is considered. District heating systems did not constitute optimal plant locations with the plant positions and heat revenue levels assumed in this study. With higher heat revenues, solid biomass gasification (BMG) with DME production was shown to be potentially interesting. With BLG considered as a production alternative, however, extremely high heat revenues would be needed to make BMG in district heating systems competitive.
The model allows for definition of biofuel share targets for Sweden overall, or to be fulfilled in each county. With targets set for Sweden overall, plant locations in the northern parts of Sweden were typically favoured, which resulted in saturation of local biofuel markets and no biofuel use in the southern parts. When biofuels needed to be distributed to all parts of Sweden, the model selected a more even distribution of production plants, with plants also in the southern parts. Due to longer total transport distances and non-optimal integration possibilities, the total resulting system cost was higher when all counties must fulfil the biofuel share target. The total annual cost to fulfil a certain biofuel target would also be considerably higher without BLG in the system, as would the total capital requirement. This however presumes that alternative investments would otherwise be undertaken, such as investments in new recovery boilers. Without alternative investments the difference between a system with BLG and a system without BLG would be less pronounced.
In several cases the model located two production plants very close to each other, which would create a high biomass demand on a limited geographic area. The reason is that no restrictions on transport volumes have yet been implemented in the model. Further, existing onsite co-operations between for example sawmills and pulp mills have not always been captured by the input data used for this report, which can cause the consideration of certain locations as two separate plant sites, when in reality they are already integrated. It is also important to point out that some of the mill specific data (obtained from the Swedish Forest Industries Federation’s environmental database) was identified to contain significant errors, which could affect the results related to the plant allocations suggested in this report.
Due to the early model development stage and the exclusion of for example many potential production routes and feedstock types, the model results presented in this report must be considered as highly preliminary. A number of areas in need of supplementing have been identified during the work with this report. Examples are addition of more industries and plant sites (e.g. oil refineries), increasing the number of other production technologies and biofuels (e.g. SNG, biogas, methanol and synthetic diesel), inclusion of gas distribution infrastructures, and explicit consideration of import and export of biomass and biofuel. Agricultural residues and energy crops for biogas production are also considered to be a very important and interesting completion to the model. Furthermore, inclusion of intermediate products such as torrefied biomass, pyrolysis oil and lignin extracted from chemical pulp mills would make it possible to include new production chains that are currently of significant interest for technology developers. As indicated above, the quality of some input data also needs to be improved before any definite conclusions regarding next generation biofuel plant localisations can be drawn.Due to the early model development stage and the exclusion of for example many potential production routes and feedstock types, the model results presented in this report must be considered as highly preliminary. A number of areas in need of supplementing have been identified during the work with this report. Examples are addition of more industries and plant sites (e.g. oil refineries), increasing the number of other production technologies and biofuels (e.g. SNG, biogas, methanol and synthetic diesel), inclusion of gas distribution infrastructures, and explicit consideration of import and export of biomass and biofuel. Agricultural residues and energy crops for biogas production are also considered to be a very important and interesting completion to the model. Furthermore, inclusion of intermediate products such as torrefied biomass, pyrolysis oil and lignin extracted from chemical pulp mills would make it possible to include new production chains that are currently of significant interest for technology developers. As indicated above, the quality of some input data also needs to be improved before any definite conclusions regarding next generation biofuel plant localisations can be drawn.
A further developed BeWhere Sweden model has the potential for being a valuable tool for simulation and analysis of the Swedish energy system, including the industry and transport sectors. The model can for example be used to analyse different biofuel scenarios and estimate cost effective biofuel production plant locations, required investments and costs to meet a certain biofuel demand. Today, concerned ministries and agencies base their analyses primary on results from the models MARKAL and EMEC, but none of these consider the spatial distribution of feedstock, facilities and energy demands. Sweden is a widespread country with long transport distances, and where logistics and localisation of production plants are crucial for the overall efficiency. BeWhere Sweden considers this and may contribute with valuable input that can be used to complement and validate results from MARKAL and EMEC; thus testing the feasibility of these model results. This can be of value for different biofuel production stakeholders as well as for government and policy makers. Further, Sweden is also of considerable interest for future next generation biofuel production from a European perspective. By introducing a link to existing models that operate on a European level, such as BeWhere Europe and the related IIASA model GLOBIOM, BeWhere Sweden could also be used to provide results of value for EU policies and strategies.