Resource-efficient treatment of organic industrial waste: Optimization of different treatment options using reMIND

The aim of this work was to find the optimal resource-efficient treatment alternative for organic waste from the food industry. For that, four different treatment methods for thirteen feedstocks were studied: animal fodder production, incineration, biological treatment and biogas production. An optimization model was used to find which treatment alternative is the optimal from a variety of perspectives. The studied systems were evaluated from three different evaluation perspectives: economy, energy and environment. The energy evaluation included two different electricity systems: coal condensing power and wind energy. The results show that there is no single optimum feedstock treatment method taking all the perspectives studied into account. Instead, it is important to consider all different perspectives when evaluating the resource efficiency of the treatment method for a feedstock. However, both incineration and anaerobic digestion of the food waste can be considered as resource-efficient treatment options for the studied feedstocks.


Introduction
The food industry is efficient in terms of energy and primary resources, but the produced organic waste needs some sort of treatment.Both uncontrolled disposal and landfilling of organic waste could lead to emissions of the greenhouse gases carbon dioxide (CO 2 ) and methane (CH 4 ).CO 2 is produced when the waste is degraded in an aerobic environment and the more potent greenhouse gas CH 4, when the waste is degraded in an anaerobic one.The European Union has a goal to reduce the amount of biodegradable waste sent to landfill (European Parliament, 1999) and to ensure that all waste suitable for recycling or recovery should not be sent to landfill by 2030 (European Parliament, 2018).It is possible to recover both energy and nutrients from food waste, depending on the treatment method used (Lindkvist et al., 2019a).The use of waste as a resource also means that use of alternative products can be reduced, leading to both economic and environmental benefits.If food waste is used as animal fodder or composted, for example, nutrients in the waste are recovered and used in place of other animal fodder or fertilizer.If the waste is incinerated, energy is recovered and used to replace heat and electricity produced by alternative methods (Scarlat et al., 2018).A fourth option for treatment of the food waste is anaerobic digestion in a biogas plant to produce biogas and fertilizer.The produced biogas could be used either for generating heat and electricity or as vehicle fuel (Scarlat et al., 2018), and the fertilizer could be used on farmland instead of mineral fertilizer (Drosg et al., 2015).Hence, with biogas production, both energy and nutrients could be recovered from the food waste.
There are previous studies on different treatment options for food waste.Ascher et al. (2020) performed a life-cycle assessment (LCA) of a food waste treatment system in Glasgow, UK.The results showed that, for every ton of food waste treated for the production of biogas, almost 100 tons of CO 2 -eq were saved.The produced biogas was used in a combined heat and power (CHP) plant, and the produced heat and electricity was assumed to replace heat and electricity generated by natural gas.The study also took into account the emissions avoided due to the use of the digestate as biofertilizer.According to Awasthi et al. (2020), the composting of food waste is an eco-friendly, efficient treatment option, which transforms the waste into a nutritional biofertilizer.Salemdeeb et al. (2017) conducted an LCA of different food waste treatment systems, including production of animal feed, anaerobic digestion and composting.The results showed that of the treatment systems studied, the production of animal feed from the food waste had the best outcome in terms of environmental and health impacts.The good results were mainly due to the replacement of conventional feed, E-mail address: emma.lindkvist@liu.se.production of which has large environmental and health impacts.The produced biogas was used for electricity production and assumed to replace produced power in line with the British electricity mix.The digestate was assumed to replace artificial fertilizers, the same was assumed for the composted food waste.
There have also been different studies on the utilization of produced biogas.Hakawati et al. (2017) evaluated different biogas-to-energy routes, focusing on energy efficiency.Patrizio et al. (2015) performed a study in northern Italy comparing the upgrading of biogas to biomethane for transport or injection into the natural gas grid with the generation of heat and electricity in CHP plants using biogas.The analysis was focused on environmental and economic benefits.There are also some studies that have included an optimization tool in their studies of different waste management options.Shahid and Hittinger (2021) performed a techno-economic optimization of food waste and compared anaerobic digestion with composting.The produced biogas was assumed to be used for electricity generation.Santibañez-Aguilar et al. ( 2013) presented a methodology based on formulating a mathematical problem concerned with maximizing the economic benefits of municipal solid waste management.The study did not consider the separation of organic waste.Even though different waste management options have been compared earlier, a comprehensive analysis considering energy, environmental and economic evaluations using an optimization tool is lacking.
In this study the aim was to find the optimum resource-efficient treatment alternative for food waste from the food industry.The term 'resource efficient' includes energy, environmental and economic perspectives.Four different treatment methods for food waste were considered: animal fodder production, incineration, biological treatment (composting) and anaerobic digestion with biogas production.For the produced biogas, two different pathways were considered, namely the generation of heat and electricity in a CHP plant and upgrading to biomethane for transport.Thirteen different feedstocks were included in the analysis: flour/bread, sugar, mixed food waste, vegetable oil, fruit and vegetable waste, potato waste, yeast, dairy waste, grains, dry food, draff, slaughterhouse waste and fishery waste.An optimization model was used to find which treatment alternative was the optimal for each studied feedstock from the different perspectives.

Method
The research design presented in Lindkvist et al. (2019b) was applied in this study.The research design includes identification of cases, defining scenarios, system development, evaluation perspectives and system analysis.In this study, an optimization model was used to find the optimum treatment method for each feedstock from different perspectives, and therefore no scenarios were included in the study.

Identification of cases
Food manufacturing companies in Sweden were contacted and asked about the amount of organic waste they produced as well as their current waste treatment methods.In total, 123 companies were interviewed, and based on these interviews 13 different organic feedstocks were identified.These are flour/bread, sugar, mixed food waste, vegetable oil, fruit and vegetable waste, potato waste, yeast, dairy waste, grains, dry food, draff, slaughterhouse waste and fishery waste.Additionally, five different treatment methods for the feedstocks were also identified as incineration, composting, animal fodder production, anaerobic digestion where the biogas produced is used in a CHP plant, and anaerobic digestion where the produced biogas is used for vehicle fuel.Based on the interviews, possible treatment methods were identified for each

Table 1
The available treatment methods for the studied organic feedstocks, derived from the interviews.Biogas production is assumed to be available for all feedstocks.
E. Lindkvist feedstock included, as can be seen in Table 1.green tick indicates that the treatment method is used by at least one of the interviewed companies for that specific feedstock.A red cross indicates that the treatment method was not used by any of the interviewed companies for the specific feedstock.All feedstocks were assumed that can be used for biogas production.The theoretical biogas yield from the different feedstocks can be found in Table S1 in the Supplementary material.Based on the interviews, the theoretical amount of feedstock was set to 10 000 tons per year for all feedstocks to enable comparison between the different feedstocks.

System development
This study was based on a systems analysis which is shown in Fig. 1.The feedstock can be treated by five different treatment methods (blue boxes): animal fodder production, incineration, biological treatment (composting) and biogas production where the biogas is either incinerated or upgraded to vehicle fuel.There are four different end products in the system (orange boxes): fertilizer/digestate, animal feed, heat and electricity and vehicle fuel.The demands for these end products were assumed to be constant in the system, no matter what treatment method was used for the feedstock.This means that an alternative production method for the end products must be included in the system.The alternative production methods in this study are other production of heat and electricity, diesel, mineral fertilizer and soya or grain.
As mentioned, the total amount of all feedstocks was set to 10 000 tons per year.In Fig. 2, the total amount of each end product that can be produced for each of the studied feedstock are shown.The amounts were calculated with Table S1 in the Supplementary material, and the numerical results can be found in Table S2 in the Supplementary material.Each feedstock has different share of total solids (TS) and methane potential, which affects the output from each feedstock, and therefore the values for the demands of the end products are different for different feedstocks.The calculations were conducted for each of the end products individually, and hence the different treatment methods did not influence each other.As can be seen in Fig. 1, some of the end products can be produced from more than one treatment method.For example, heat and electricity is produced from both direct incineration of the feedstocks and from biogas production with incineration of the produced biogas.Fertilizer is the end product from both composting and biogas production (digestate).For these two, the treatment method that leads to the largest amount of end product was used for the calculations, e.g.direct incineration of the feedstock for heat and electricity and composting for the end product fertilizer.

Evaluation perspectives
The studied systems were evaluated from three different evaluation Fig. 1.The system studied.
E. Lindkvist perspectives: economy, energy and the environment, in accordance with Lindkvist et al. (2019b).The system boundaries for the energy and environmental perspectives are represented by the dotted line in Fig. 1.

Economy
The economic evaluation was performed by calculating the net profit of the system, and, as suggested by Lindkvist et al. (2019b), production, operation and maintenance costs for the different treatment methods have been included, and also revenues from the end products.Investment costs for biogas production and upgrading have been included in the economic evaluation.The other treatment methods were assumed to already exist along with the alternative production methods.In accordance with Lindkvist et al. (2019a), the system boundary for the economic perspective is narrower, including the treatment method of the organic by-product and the revenues from the end products, but not the costs associated with the alternative production methods.The costs associated with production can be found in Table S3 in the Supplementary material.The costs and revenues for the different products in the system are given in Table S4 in the Supplementary material.

Energy
The energy evaluation of the system was studied.Primary energy was used for the energy evaluation to enable a comparison of highquality energy with low-quality energy.High quality energy is for example electricity and low-quality energy is for example low temperature heat (Jensen, 1980).High quality energy can be used to generate low quality energy, but the opposite is more difficult and even impossible in some cases (Jensen, 1980).Primary energy factors for different energy conversion processes can be found in Table S5 in the Supplementary material.The energy demanded by the different processes in the study is given in Table S6 in the Supplementary material.In the energy evaluation both treatment methods and alternative production methods were considered.The energy content in the end products was not considered in the energy evaluation.However, the energy content of the end products was considered in terms of fulfilment of demand for the various end products.For example, to fulfil the demand for vehicle fuel, expressed in MWh, the energy content of upgraded biogas was considered as well as the energy content of diesel.
The energy evaluation includes two different electricity systems, one where all electricity is produced by coal condensing power and one where all electricity is produced by wind energy.This is done to understand how the electricity system can influence the results.The change of electricity system was not assumed to have an influence on the economic evaluation since the cost of electricity is assumed to be constant regardless of the electricity system.The environmental evaluation was, however, affected by the change in production of electricity and, hence, two different results are presented for both the energy evaluation and the environmental evaluation, one result for an electricity system based on wind energy and one result for an electricity system based on coal condensing power.

Environment
For the environmental evaluation, three impact categories were considered: global warming potential (GWP), acidification potential  S1 in the supplementary material and the assumptions that the total amount for all feedstocks was set to 10 000 tons per year.The numerical results for the calculations can be found in Table S2 in the supplementary material.

E. Lindkvist
(AP) and eutrophication potential (EP).All emissions in the studied system have been translated to CO 2 -equivalents (GWP), SO 2 -equivalents (AP) or PO 4 3− -equivalents (EP), using the conversion factors presented in Table 2. Specific emissions connected to energy conversion can be found in Table S5 in the Supplementary material, and other specific emissions used in the study are given in Table S7 in the Supplementary material.Emissions connected to the construction of treatment or production facilities were not included in the system.During production of biogas, the assumed methane slip was 2% of the produced biogas for anaerobic digestion and 1% for the upgrading process (Lantz et al., 2009).This was included in the calculations, and therefore the GWP for incineration of biogas is set to zero in Table S5 in the Supplementary material.

Systems analysis (reMIND)
An optimization tool can be used to model this kind of complex system.The optimization tool reMIND is based on the Method for analysis of INDustrial energy systems (MIND), which in turn is based on mixed integer linear programming (MILP) (Karlsson, 2011).The reMIND tool can be used for modelling any kind of system.The system cost of the modelled system is minimized or maximized, based on the constraints included.However, it is possible to accomplish any kind of minimization or maximization in the tool, e.g., CO 2 emissions, energy use (Wolf and Karlsson, 2008).The reMIND tool have earlier been used to study a variety of industries and energy systems, eg.automobile manufacturing (Thollander et al., 2009), pulp and paper (Karlsson and Söderström, 2002), a district heating system in a small city (Svensson and Moshfegh, 2011), the potential for excess heat use in an energy system in a Swedish county (Viklund and Karlsson, 2015) and waste minimization from a steel plant (Riesback et al. 2015).Based on the possibility to model large systems and include a variety of different flows, whilst also be able to minimize the system cost as well as the total energy use and the environmental impacts of the system, reMIND was considered a good choice for this study.
The system is represented by nodes (boxes in Fig. 3) and branches (arrows in Fig. 3).Each branch represents a flow in the system.This could be, for example, electricity, biogas or digestate.Each node represents a process of the system and includes numerous functions describing the functionality of the node.The relations between the nodes and branches in the system are included in a standardized file, which is optimized using an optimization solver.The optimization solver used by reMIND is CPLEX.
The system was optimized in terms of minimizing cost, energy use, GWP, AP and EP.Both costs and revenues were included in the system cost, revenues being counted as negative costs.For the energy balance, the primary energy use was optimized, and hence both energy inputs and energy outputs were considered in the model.For GWP, AP and EP, only direct emissions were considered in the model.The optimized system consists of 19 nodes and ten different kinds of flows.The different nodes and flows used in the model, and the connections between them, can be seen in Table S8 in the Supplementary material.For all feedstocks, the optimization was performed for both an electricity system based on wind power and an electricity system based on coal condensing power.

Results and analysis
In Table 3, the overall results are shown.The numerical results for this study are presented in Table S9 in the Supplementary material.As can be seen, there is no one treatment method which can be optimum for a feedstock for all perspectives.
Starting with the economic evaluation, as presented in the method section, the same economic results are valid for both electricity systems included, since the cost and prices connected to electricity was assumed to not be affected by the switch in electricity system.Hence, only one set of results is presented for the economic evaluation in Table 3.The results show that for most of the substrates, the optimal option from an economic perspective would be to produce vehicle fuel.For four of the feedstocks, animal fodder would be the optimal, and for dry food, direct incineration of the substrate would be optimal from an economic perspective.
For the energy evaluation, the results vary depending on the electricity system in place.With an electricity system based on coal condensing power, the optimal option is to use the feedstock to produce heat and electricity.For the feedstocks where incineration of the feedstock is an available treatment option, this is the optimal option and for the other feedstocks, the optimal option is to produce biogas from the feedstocks and then use that biogas for heat and electricity production.With an electricity system based on wind energy, the optimal option for all feedstocks from an energy perspective would be to produce biogas and upgrade it to vehicle fuel.
The environmental evaluation included global warming potential (GWP), acidification potential (AP) and eutrophication potential (EP).The results differ somewhat according to the electricity system used.As can be seen in Table 3, with an electricity system based on coal condensing power, incineration is the optimum treatment option when available.However, there is one exception from this, namely fishery waste when minimizing EP, where instead the vehicle fuel option yields the optimum result.However, the numerical results in Table S9 in the Supplementary material, show that vehicle fuel would result in 1.15 tons PO 4 -eq., and incineration in 1.23 tons PO 4 -eq.For the feedstocks where the biogas option is optimal, Table 3 shows that for EP vehicle fuel has better results than the CHP option for all feedstocks concerned.However, the numerical results in Table S9 in the Supplementary material show that the results for CHP and vehicle fuel for EP, for these feedstocks, are in the same magnitude.
When the electricity system is assumed to be based on wind energy, the results for the environmental evaluation are similar to the results for an electricity system based on coal condensing power.The difference is that instead of using the biogas for production of heat and electricity, the biogas should be used for vehicle fuel with an electricity system based on wind energy.However, for the feedstocks grains, slaughterhouse waste and fishery waste, the most optimum treatment option is to use the feedstock for vehicle fuel when minimizing both AP and EP.
Looking closer at the numerical results in Table S9 in the Supplementary material, one can see that some of the treatment options have results close to each other.For example, for the feedstock flour/bread, VF and incineration have similar results for economy, energy use (wind) and EP (wind).The same is true for mixed food waste for energy use (wind) and EP (wind).

Discussion
It is important to understand that the results in this study are for the whole system studied and not for a single treatment option.If, for example, the optimal treatment option for the feedstock studied is animal fodder production, the results also include alternative production of

Table 2
Conversion factors for GWP (global warming potential) (Solomon et al., 2007), AP (acidification potential) and EP (eutrophication potential) (Guinée, 2002) heat and electricity as well as alternative production of vehicle fuel and fertilizer.The overall trends in the results show that with an electricity system based on coal condensing power, the optimum treatment option for the feedstocks studied are either direct incineration of the feedstock or biogas production with CHP utilization.The reason why incineration, when available, is a better alternative than biogas production could be that the feedstock in question does not need to be pre-treated in any way, which would require heat and electricity.In the calculations there are losses included in the biogas treatments, both of feedstock but also of biogas.In the incineration option, the losses have been assumed to be Fig. 3.A simplified biogas system in reMIND.

Table 3
The results for the study, indicating the optimum treatment option for each feedstock for the different perspectives and the different electricity systems. negligible.
As can be seen in Table S9 in the Supplementary material, the numerical results for the incineration treatment option show the same results regardless of the electricity system.This is because incineration is assumed to not have any electricity use from external electricity sources.For the alternative treatment options, the energy demand is consistent no matter of the applied electricity system.This is because the energy demand does not only include electricity, both also other types of energy.This is also the reason why the environmental impacts of the alternative production options are independent from the electricity system in place.
For the economy perspective, vehicle fuel is the best treatment option for most of the feedstocks.This could be because vehicle fuel results in two different end products, namely vehicle fuel and digestate, and hence results in two revenues.The revenue for vehicle fuel is higher than the revenue for heat and electricity generated from biogas.For the feedstock dry food, incineration is the most profitable treatment option.This could be due to the low methane yield from this feedstock, as can be seen in Table S1 in the Supplementary material.
Table S5 in the Supplementary material sheds light on the results for GWP.The GWP for coal condensing power is much higher than any of the other ones.Hence, electricity production from the feedstocks will be prioritized when optimizing the model for minimizing GWP.For diesel, only production and distribution are included in the GWP since the use of diesel is outside the scope of this study.Awasthi et al. (2020) claim that composting of food waste is the most effective strategy for producing a biofertilizer from food waste.However, the results in this study show that composting is not the optimal treatment for the feedstocks studied under any of the perspectives when considering a larger system.In a study performed by Hakawati et al. (2017), it was concluded that the most energy efficient way to utilize biogas is through CHP.This is somewhat in line with the results from this study, when considering an electricity system based on coal condensing power.However, with an electricity system based on mainly renewable energy, the most energy efficient way to utilize biogas (and treat the studied feedstocks) is to produce vehicle fuel from biogas.Laso et al. (2018) combined LCA with linear programming to minimize water and energy use as well as greenhouse gas emissions related to the waste management of an anchovy canning industry.The studied treatment options were landfill, incineration and valorisation.The results showed that incineration was the better option.Pierie et al. (2017) used energy and environmental system analysis, LCA and modelling to analyse and optimize symbiosis of a farm-scale anaerobic digestion system, utilizing locally available biomass waste streams.Four indicators of sustainability were analysed: energy efficiency, carbon footprint, environmental impacts and costs.One result from the study was that replacing mineral fertilizers with bio-fertilizers (digestate) significantly reduces the carbon footprint of the system studied.Both these studies showed results that are in line with the findings in this study.Edwards et al. (2017) performed a life-cycle inventory for seven different waste management systems for organic waste, including landfill, anaerobic digestion and composting.The analysis included energy and water balance, together with air, water and soil emissions.The conclusions of this study stated that it is important to divert organic waste from landfill to reduce emissions of CH 4 as well as leakage of organic nitrogen and phosphate.Further, where the electricity system is based on coal, electricity generation from organic waste could significantly decrease emissions of CO 2 , SO x and NO x .Salemdeeb et al. (2017) compared the environmental impact of four different treatment options for food waste, including animal feed, anaerobic digestion and composting.They concluded that animal feed would lead to lower environmental impacts than the other options studied.
The overall results in this study show that incineration of the food waste is the most efficient alternative when available, for the perspectives studied, except for economy, and energy when the electricity system is based on wind power.For the feedstocks where incineration is not an available option, the most efficient treatment option is anaerobic digestion.The most efficient utilization path for the produced biogas depends on the electricity system.With coal condensing power as electricity source, the biogas should be used for heat and electricity production to replace the coal.However, with wind power, the biogas should instead be used for vehicle fuel and replace diesel.

Conclusion
The aim of this study was to find the most resource-efficient treatment alternative for food waste from the food industry.However, the results show that the best treatment option depends on both the composition of the food waste and the perspective in question (e.g., energy, economy, or environment).It is also evident that the results depend on what is being replaced in the model, which explains the differences in the results for the two studied electricity systems.This shows the importance of systems analysis when performing these kinds of studies.To evaluate the resource efficiency of a waste treatment method, it is important to consider all perspectives.In this study, one conclusion is that composting of the food waste is not a resourceefficient treatment for any of the feedstocks studied.Additionally, both incineration and anaerobic digestion of the food waste can be considered as resource-efficient treatment options for the feedstocks studied here.From an economic perspective, anaerobic digestion with production of vehicle fuel is the most profitable option for most of the feedstocks, followed by animal fodder production.However, incineration of the food waste is the most efficient alternative when available, for all feedstocks, for the energy and environmental perspectives, when replacing coal condensing power.For the other feedstocks, anaerobic digestion with heat and electricity production is the most effective alternative.With an electricity system based in wind power, anaerobic digestion with vehicle fuel production is the best alternative from an energy perspective.From an environmental perspective, incineration and vehicle fuel are the most preferred options.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Total available production of heat and electricity, animal feed, fertilizer and vehicle fuel for the feedstocks studied.The Figure is based on calculations with TableS1in the supplementary material and the assumptions that the total amount for all feedstocks was set to 10 000 tons per year.The numerical results for the calculations can be found in TableS2in the supplementary material. .