International O.R. Insights: Alternative energy powers Sweden

By Patrik Flisberg, Mikael Frisk and Mikael Rönnqvist

Countries around the world are looking for new, alternative energy resources for production, transport and heating. The current main energy sources fueling these activities – oil, coal and natural gas – raise concerns regarding CO2 emissions, air quality and renewability. In contrast, more than 47 percent of the total energy used in Sweden in 2009, according to the Swedish Energy Agency [1], came from renewable resources such as water, forest biomass, wind and solar sources. For example, Sweden developed a well-functioning district heating system that provides heating and electricity for more than 50 percent of the households in the country using energy mostly derived from waste and forest biomass, the latter of which is renewable and CO2 neutral.

This article is based on two large case studies done together with two of the largest suppliers of forest biomass for energy. The article focuses on the logistics of using forest biomass as an energy source for district heating and how operations research serves as a vital tool for the efficient utilization of the system.

District Heating System

The idea behind district heating is simple (description by Swedish District Heating Association [2]). By using inexpensive, locally available energy resources, local heating needs are met in an efficient and environmentally friendly way. District heating is an infrastructure that connects the energy resources available in the region or city with those who need the warmth. Instead of each building having its own boiler, heat to many buildings is supplied from a central plant with advanced treatment that significantly reduces the pollutants released into the air. A central heating plant produces hot water that heats the entire community from well-insulated pipes. The hot water to the building’s heat exchanger ensures that the elements heat up and that there is hot water from the tap. The water is between 70 and 120 degrees Celsius, depending on season and weather.

Many companies use a mix of fuels, depending on local availability, cost, climate and environment. Resources that would otherwise be lost – such as residue from forestry operations (tops and branches and other biomass – are put to use to the benefit of entire communities. The system also takes advantage of heat produced by electricity generation, waste incineration and waste heat from local industries. When electricity is produced simultaneously with district heating, it’s known as “cogeneration.” Cogeneration (Figure 1) is both more efficient and more environmentally friendly than other forms of energy generation since it provides both electricity and heat through better utilization of the fuel source.

Figure 1: Illustration of cogeneration of district heating and electricity. The biomass-based energy supplies hot water for heating and drinking, as well as electricity, to the house. (Illustration by Per Thorneus)

District heating is the most common form of heating in Sweden. All major cities use district heating systems. Of Sweden’s 290 municipalities, 270 are using district heating. More than half of all residential and commercial buildings are heated with district heating. Among the apartment block, the proportion is about 90 percent, while it is slightly lower for the premises and lowest for single-family homes. However, many of the houses being built today will be connected to district heating systems, and the number is growing at 18,000-20,000 units per year.

The development of district heating started in the 1940s, but the real breakthrough came after the first big oil crisis of 1973. During that time, many new apartment homes could be directly adapted and connected to district heating. A slight decline occurred in the 1980s when several nuclear reactors came online and the price of electricity was relatively low. But the situation changed. Today, Swedish district heating companies produce almost 50 TWh of heat. The development is illustrated in Figure 2.

Figure 2: Production of energy in district heating between 1955 and 2010.

Different fuels are used in order to heat the water to a sufficient temperature. In 1980, oil accounted for more than 90 percent of district heat production, but by 1988 it was pushed down to 14 percent. In the 1990s, environmental and climate policies began to affect people’s thinking. Climate scientists warned of the so-called greenhouse effect that emissions of carbon dioxide induced. Politicians wanted to influence the development of district heating and imposed special taxes on carbon dioxide and sulfur fuels. That meant large increases in costs for thermal plants, whose activities were powered by carbon-based fuel, compared to heating plants that utilized biomass and did not have to pay any carbon tax. So began a comprehensive transition of the entire industry from fossil fuels to biofuels. Today, the majority of district heating comes from renewable and recycled energy (Figure 3).

Figure 3: Proportion of fuel for district heating in 2010.

Logistic System for Forest Biomass

Many articles in the literature address planning of forest operations, i.e., harvesting and transportation for round-wood (e.g., D’Amours et al. [3]). Round-wood is the common name for saw logs and pulp logs, and many planning systems focus on either harvesting and/or transportation. Optimizing the supply chains for round-wood and for forest biomass is similar but implies two major differences. First, forest biomass (for example, tree parts, low quality logs and branches) normally has to be converted into chips before delivery to heating plants for burning. Second, the demand for forest biomass varies over the year as it is related to the outside temperature. Figure 4 gives an example of the energy demand (left) based on forest biomass and an example of the temperature distribution at one heating plant during one year. Clearly, the high demand appears in the cold months. Also, the demand difference is a factor of four between summer (June-August) and winter (December-March). To balance the conversion and transportation capacities over time, it is very important to manage the inventory levels of forest fuel products at terminals.

Figure 4: Left: Demand for forest biomass (in MWh) for one supplier. Right: Daily temperatures during one year (Linköping, 2008-2009, Data from Swedish Meteorological and Hydrological Institute).

Another difference is that more machine systems are used in the forest fuel operations depending on their capacity. The typical process is to forward the logging residues to a landing site where they are left to dry for some months during which most of the needles drop off. The chipping can then take place at the harvest site or the logging residues can be moved to a terminal for chipping and/or storage. For large customers with their own chipping system, the residues can also be transported directly to the heating plants.

There is a critical trade-off between the chipping location and transportation capacity. In Sweden, the weight limit of a truck (including the truck itself) is 60 metric tons and the length limit of the truck (including trailers) is 24 meters. If logging residues such as branches are transported, it is not possible to load 60 tons, so the weight capacity cannot be fully used. If a chip truck is used, it is possible to load more with higher transportation efficiency. However, the chipping must take place at the harvesting location with higher relative cost. Generally speaking, the aim is to transport logging residues over as short a distance as possible. If the heating plants are far away, it is better to transport residue to a terminal where it is chipped, and then the chips are transported over the longer distance with chip trucks. Balancing the chipping and transportation decisions is important in order to reduce costs and maximize efficiency.

The growing use of forest biomass for energy production increases the need for effective planning systems to manage and plan the forest fuel supply chain. This is particularly important as the values of the assortments [tree parts, low quality logs and branches] are very low. It is critical to establish a low-cost system in order to remain profitable. A general description of forest biomass can be found in Björheden et al. [4]. Some key logistic questions are:

• Which customers (heating plants) are profitable?
• Which terminals should be used?
• What assortments, where and what volumes should be kept in inventory?
• Which machine system for chip conversion should be used and where?
• How should the assortments be transported, with truck systems or with a combination of trucks and train systems?
• Which forest biomass areas should be used and which should be purchased?

O.R. Methodology and the FuelOpt System

The operations research (O.R.) model needs to make decisions on use of machine systems, transportation and inventory levels. Each machine system must be described with its activities (chipping, crushing, bundling, etc.) and their cost and performance. Each transportation system has its capacity and cost. The supply (forest biomass assortments) are described by volume (or energy contents), location and purchase price. The demand at each heating plant is described by energy content (GWh) for each assortment and time period and a given price. The objective is to maximize the profit (revenue from sales minus costs) while considering aspects like availability of mobile chippers and trucks. The model can be formulated in a linear programming model. Manual planning is the dominating approach for the forest biomass logistic planning. However, as this is very time consuming and there are many possible machine systems and transportation decisions to make, the use of a decision support system (DSS) is desired. The Forestry Research Institute of Sweden (Skogforsk) has in cooperation with some major Swedish forest companies developed the DSS FuelOpt (Frisk et al., 2012 [5]) that has been used in several case analyses as noted below.

Experiences and Results

Case Sveaskog: Sveaskog AB is the largest forest owner in Sweden with 3.2 million hectares of productive forest land. The company, owned by the Swedish state, sells about 12.5 million meter3 (superscript) forest products on a yearly basis, delivering to approximately 70 saw log customers, 30 fiber customers and 40 biofuel customers. The objective of the study was to analyze the company’s forest fuel logistics in a specific region in the middle of Sweden. Some case specific data is reported in Table 1.

Table 1.

The optimization results indicated a theoretical profit for Sveaskog of 10.5 million SEK or 15 SEK/MWh. The costs (in millions SEK) for each activity were computed; Purchase – 56.6, Chipping – 20.9, Transport – 23.6 and Storage – 0.3. The average transportation cost was 34 SEK/MWh. The transportation distances for trucks varied between 15 and 120 km with a volume weighted average distance of 40 km. Figure 5 illustrates the flows of different assortments in the results from the optimization. To get a geographical figure of the flows provide in itself a great support for the understanding and analysis.

Figure 5: Flows from the optimization results. Red and yellow lines are train transportations; other colors are truck transportation of different assortments.

A majority of the residue slash (55 percent) was chipped directly at the forest site. Only 1 percent was bundled and the rest was transported as slash directly to heating plants. Of the fuel wood, 60 percent was transported directly to customers, 30 percent was chipped on terminals, and the remaining 10 percent was kept in inventory at terminals. A total of 72 percent of logging residues was chipped directly at the forest site, and the rest was transported to heating plants or kept at terminals. The optimization determined the optimal use of the different machine types for chipping and crushing. For example, the crushers (large mobile chippers) at terminals were used mainly to crush tree slash. The use of the forward-mounted chipper was spread over the entire region, both close and far away from heating plants. The combined chipper and chip truck was used mainly for forest sites further away from the heating plants. The contribution of each customer to the profit was also computed, which turned out to be a very good basis for further contract negotiations.

Case SEBAB: The second case study was performed together with the company Stora Enso Bioenergi AB (SEBAB). SEBAB is a company that purchases, produces and sells forest biomass, mostly to energy companies and forest industries. The company is one of Sweden’s biggest forest biomass companies and has activities in southern and central Sweden (Figure 6). The turnover is approximately 800 million SEK for a total of 5TWh forest biomass each year.

Figure 6: Geographical distribution of supply of forest fuel (left) and demand for energy and terminal locations (right) for Stora Enso Bioenergi AB.

The planning horizon for this case study was one year divided into twelve months and covered the entire company’s forest biomass flows (southern and mid Sweden). It played an important role in the company’s work to increase effectiveness and to improve the operations. Some case specific data are reported in table 2.

Table 2.

The optimization included costs for purchasing, chipping, inventory and transportation costs that varied depending on assortments, transportation modes, etc. The price that the customer would pay for different assortments was also added.

In the different optimization alternatives, the capacity on the truck-mounted chippers, the possibility to transport chip via train and the customers’ demand varied in order to find answers to the important logistic questions.

One of the most important results was the comparison between the reality and the optimization when the same conditions were used. This provided information not only about the potential of improving the logistics, but also a description of how the logistic system should be improved. The difference in total costs between reality and the optimization was 6.3 percent. However, the profit surplus for forest fuel activities could increase by 21 percent. A major difference between reality and optimization was that in the optimal solution the volume chipped in the forest decreases by 20 percent while the chipping at terminals increases. On the transport side, in the optimal solution the use of trains increases dramatically, and the supply was more likely to be allocated to the nearest customer in the optimization. This decreased the average transportation distance from 58.8 km in the reality to 41.7 in the optimization.

SEBAB also wanted to study whether there was any potential to increase the proportion of their transports by train from terminals directly to customers. Compared to the reality, more train transportation would decrease transport costs by 4.5 percent when the amount of forest biomass transported by rail increased from 34 GWh to 212 GWh.

Further Challenges

Future plans for DSS FuelOpt include boosting robustness to incorporate the uncertainty in the monthly demands due to the weather situations. This is interesting since many contracts include minimum and maximum levels that are determined on a monthly basis throughout the year. Some initial tests are reported in Flisberg et al. [6]. Among the result interpretations, the developers will also add a module to measure environmental effects such as emissions of CO2, CO, SO2 and NOx. Another future research topic is to consider an integration of the round-wood and forest biomass supply chains. Today, pulp logs are already being converted into chips and used at heating plants instead of at pulp and paper mills. By an integration of the chains, it would be possible to analyze the impact of different taxation systems, for example, and investigate how the chains can complement each other in different scenarios; some early studies are done in Kong et al. [7].

Patrik Flisberg and Mikael Frisk work as researchers at the Forestry Research Institute of Sweden (Uppsala, Sweden). Mikael Rönnqvist is a professor in industrial engineering at Université Laval (Québec, Canada). The authors acknowledge the contributions of Håkan Alexandersson and Maria Gabrielsson at StoraEnso Bioenergi AB and Anna Ahlin and Mattias Carlsson at Sveaskog AB for valuable discussions and input data for the case studies.

References

1. http://www.energimyndigheten.se/en/Facts-and-figures1/Publications/ pdf document: http://www.energimyndigheten.se/Global/Engelska/Facts%20and%20figures/Energy%20in%20Sweden%20facts%20and%20figures%202011.pdf (Figure 10 last withdrawn information on March 9).
2. Swedish District Heating Association, http://www.svenskfjarrvarme.se/In-English/District-Heating-in-Sweden/ (last withdrawn information on March 9, 2012).
3. D’Amours, S, Rönnqvist, M. and Weintraub, A., 2008, “Using Operational Research for supply chain planning in the forest product industry,” INFOR, Vol. 46, No. 4.
4. Björheden, R, Eliasson, L. and Thorsén, Å. (Eds.), 2010, “Efficient forest fuel supply systems,” composite report from a four-year R&D program, 2007-2010, Skogforsk, Sweden.
5. Flisberg, P., Frisk, M. and Rönnqvist, M., 2012, “FuelOpt – A decision support system for forest fuel logistics,” Journal of the Operational Research Society, published online, doi: 10.1057/jors.2011.157.
6. Bredström, D., Flisberg, P. and Rönnqvist. M., 2012, “A new method for robustness in rolling horizon planning,” International Journal of Production Economics, published online, doi: 10.1016/j.ijpe.2011.02.008.
7. Kong, J., Frisk, M. and Rönnqvist, M., 2011, “Modeling an integrated market for sawlog, pulpwood and forest bioenergy,” Canadian Journal of Forest Research, Vol. 42, No. 2, pp. 315-332.