- Email: [email protected]
Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation
Waste Management xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Waste Management journal homepage: www.elsevier.com/locate/wasman
Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation A. Bernstad Saraiva a,⇑, Å. Davidsson b, M. Bissmont c,d a
SAGE/COPPE, UFRJ, Rio de Janeiro, Brazil Water and Environmental Engineering, LTH, Lund University, Lund, Sweden c VA SYD, Malmö, Sweden d Urban Studies, Malmö University, Malmö, Sweden b
a r t i c l e
i n f o
Article history: Received 8 December 2015 Revised 12 April 2016 Accepted 30 April 2016 Available online xxxx Keywords: Food waste disposers Food waste Collection system Lifecycle assessment Waste segregation Energy balance
a b s t r a c t An increased interest for separate collection of household food waste in Sweden has led to development of a number of different collection-systems – each with their particular benefits and drawbacks. In the present study, two systems for collection of food waste in households were compared; (a) use of food waste disposers (FWD) in kitchen sinks and (b) collection of food waste in paper bags for further treatment. The comparison was made in relation to greenhouse gas emissions as well as primary energy utilization. In both cases, collected food waste was treated through anaerobic digestion and digestate was used as fertilizer on farmland. Systems emissions of greenhouse gases from collection and treatment of 1 ton of food waste (dry matter), are according to the performed assessment lower from the FWD-system compared to the reference system ( 990 and 770 kg CO2-eq./ton food waste dry matter respectively). The main reasons are a higher substitution of mineral nitrogen fertilizer followed by a higher substitution of diesel. Performed uncertainty analyses state that results are robust, but that decreasing losses of organic matter in pre-treatment of food waste collected in paper bags, as well as increased losses of organic matter and nutrients from the FWD-system could change the hierarchy in relation to greenhouse gas emissions. Owing to a higher use of electricity in the FWD-system, the paper bag collection system was preferable in relation to primary energy utilization. Due to the many questions still remaining regarding the impacts of an increased amount of nutrients and organic matter to the sewage system through an increased use of FWD, the later treatment of effluent from the FWD-system, as well as treatment of wastewater from kitchen sinks in the reference system, was not included in the assessment. In future work, these aspects would be of relevance to monitor. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Collection of food waste from households for subsequent biogas production is becoming increasingly common in Sweden. The most utilized scheme for collection is use of paper bags for collection of food waste in households and later disposal in single or multicompartment waste bins (Waste Management Sweden, 2014). However, in several other countries, use of food waste disposers (FWD) is a common method for separate collection of food waste from households. Previous studies have suggested that FWDs can present a practical alternative for source-separation of food waste, without ⇑ Corresponding author at: Centro de Gestão Tecnológica – CT2, Rua Moniz de Aragão, No. 360 – Bloco 2, Ilha do Fundão – Cidade Universitária, Rio de Janeiro, RJ 21.941-972, Brazil. E-mail address: [email protected] (A. Bernstad Saraiva).
increasing transports, and through avoidance of problems related to odor and increased need for waste bins (Marashlian and ElFadel, 2005). However, several questions can be raised regarding the effects of FWDs connected to conventional sewer systems. Several potential adverse effects have previously been described. Bolzonella et al. (2003) stated that FWD could cause an increased organic load in the biological step at the wastewater treatment plant (WWTP) and thereby increase energy demand for wastewater treatment. Nilsson et al. (1990) raised problems with increased oil and grease load at WWTPs as well as risks of increased production of H2S in sewerage systems, potentially resulting in corrosion of cement pipes. On the other hand, Evans (2012) and Galil and Yaacov (2001) presented measurements of significant increases in biogas production in WWTP-sludge digestion when 50% of connected households introduced FWD. At the same time, Raunkjaer et al. (1995) showed that the removal of dissolved organic matter and proteins in
http://dx.doi.org/10.1016/j.wasman.2016.04.036 0956-053X/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
2
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
wastewater during sewage transport to WWTP could be considerable, which implies that effects on subsequent WWTP-processes will be of lesser significance. In addition, the effects on WWTP processes due to FWDs will to a large extent depend on the WWTP design. Bolzonella et al. (2003) state that an increased carbon concentration in incoming wastewater gained through use of FWDs can improve the C/N and C/Pratio in WWTPs and, depending on the process design, result in an improved nutrient removal and reduced requirement for external carbon sources. In summary, our knowledge of the quantity and quality of food waste disposed of in FWDs that actually reaches the WWTP as well as net-effects on sewerage systems and WWTP-processes from FWD installation is still limited, and probably influenced by several different factors. Thus, this is certainly an area for further investigations. With the potential problems stated above as a background, and yet a constant search for methods which can facilitate household food waste collection both for users as well as for solid waste and wastewater management organizations, a novel FWD-system was developed in Sweden in 2001. With the aim of combining the benefits from use of FWD for users, without increasing risks for problems in later transport and treatment of household wastewater, a system with tank-connected FWDs was implemented and tested. The system is graphically illustrated in Fig. 1. The system included around 60 households in low-rise buildings in Malmö, a city with around 320,000 habitants in southern Sweden. A similar system was installed in 2007, linked to more than 140 apartments in a high rise building in the same area. The evaluation showed that the design of these systems has not been optimal, with the consequence that the fraction of organic matter in collected material is low, and that a significant amount of organic material is lost from the tank between each emptying (Davidsson et al., 2011). However, the evaluations of these earlier systems were partly based on estimations, due to an absence of flow measurements at the outlet from the tank. As the system can provide large benefits from various perspectives (working environment for waste collectors, maximizing use of high value urban land, reduction of potential risks for clogging of sewage pipes, etc.), a system similar to the ones described above was installed in a newly constructed quarter in the same city in 2010. However,
some changes were made with the intention to optimize the system. 1.1. Aim and scope The purpose of the present study is to evaluate this new system from an environmental perspective, using lifecycle assessment (LCA) methodology, and compare this system to a reference collection system for household food waste. In addition, the aim is also to identify processes with a large impact to overall results, as this constitutes an important basis for further improvements and optimizations. The aim is thus to make a statement about which of the compared systems is more advantageous from an environmental perspective, and the conditions under which this is true. The functional unit in the study is defined as: ‘‘Management and treatment of 1 ton TS source separated food waste from households.” The proposal means that we in the functional unit takes no account of the recycling rate may be different in the different systems. This is because we, in the context of this study, not will be able to establish any general differences between use of food waste disposers and the reference system with regard to separation behavior. In the case of the food waste disposer system, production of biogas and nutrient content in digestate is assessed based on experimental data, while literature data is used for these parameters in the assessment of the reference system. Collection of food waste in paper bags is selected as the reference system, as this system currently is the most common system for food waste collection in Sweden (Waste Management Sweden, 2013). Produced biogas is assumed to substitute diesel as fuel in busses, as upgrading of biogas for use as vehicle fuel was the most common application of biogas in Sweden in 2014 (Swedish Energy Authority, 2015). 1.2. System boundaries Choices of system boundaries should be based on the following principles (ILCD, 2010):
Collecon of grinded FW
Kitchen sink with disposer
Plasc holder for separaon FW in paper bag and disposal in waste bin
Separaon tank
Collecon
Treatment of effluent in WWTP
Pre-treatment
AD of grinded FW
AD of WWTP Sludge
Incineraon of pre-treatment residue and AD of FW
Fig. 1. Graphical representation of the character of the tank-connected food waste disposer system (top) and the reference system (bottom) investigated in the study. FW = Food waste, AD = Anaerobic digestion. Processes with labels in italics are not included in the assessment.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
All parameters of relevance to the outcome of the LCA should be considered. The systems being compared are equivalent. Parameters that are the same for the systems being compared may be excluded. These principles are many times difficult to follow in practice. It is often difficult to know if a process is relevant or not until it is included in the LCA, and one also can see the effects of excluding the same process. Assuming that processes that are seen as relevant in previous studies examining similar systems can be a good indicator but also a risk of repeating mistakes made in previous studies. In addition, systems that are of interest to compare are in many cases different in aspects which challenge exact comparisons. Thus, exact equivalence is therefore often difficult to achieve, but should always be sought. Based on this, system boundaries demonstrated in Fig. 2 were used in the present study. With the selected functional unit and system boundary setting, organic matter found in effluent from the collection tank are excluded from the assessment. The main reason for this choice is the many times contradictory statements of the affect from FWDs on sewage systems and WWTP processes presented above (see e.g. Raunkjaer et al., 1995; Evans, 2012). The potential environmental benefits from reducing these losses are however estimated for an assessment of optimization potentials. As seen in Fig. 2, the present study considers the possible substitution of mineral fertilizers by nutrients in food waste when applied on farmland, while use of digestate was omitted in previous studies of FWD-systems (Lundie and Peters, 2005; Diggelman and Ham, 2003). Differently from several other studies of food waste collection systems (Assefa et al., 2005; Khoo et al., 2010; Kong et al., 2012), the present study also includes production of the material needed for collection of food waste. Previous studies of food waste disposal behavior in the UK show that as much as 24% of all domestic food and drink waste in the
3
country is disposed of in the sewer, although only a very small part of the population have kitchen disposers (WRAP, 2009). In other studies, the actual content of organic matter and nutrients in wastewater from kitchen sinks was measured (Friedler and Butler, 1996). Results vary vastly between studies, to some extent explained by variations in water consumption and inclusion or not of washing machines. However, Friedler and Butler (1996) also shows that the composition of kitchen sink wastewater varies largely between households and over time. Thus, the selection of households and timing of measurements can effect results gained. Recent studies in Sweden performed as questionnaires amongst 2050 households, suggest that around 26 kg food and beverage per person and year are wasted through the kitchen sink (Störme et al., 2014). This equals 23% of total household food wastage, which is similar to results gained in the UK (WRAP, 2009). Coffee and tea, followed by dairy products were responsible for around 65% of the total wastage through kitchen sinks. This fraction of food wastage is of course considered in the FWD-system, while excluded in the reference system. The main reason is a current lack of knowledge of the destiny of this food waste fraction along the sewer system and in the later wastewater treatment plant. Distribution of produced biogas to fuel stations is not addressed, since distribution not is included in data used for description of diesel replaced by biogas. 1.3. Delimitations The study is limited to an investigation of the two collection systems from a climate change perspective, i.e. only those processes generating emissions that contribute to global warming are taken into account in the study. CO2, CH4 and N2O are considered with characterization factors according to IPCC 2007 (Ecoinvent Center, 2013). The greenhouse effect is studied in a 100-year perspective and emissions of biogenic carbon are
Fig. 2. System boundaries used in the assessment of the two collection systems; disposal of food waste in tank-connected FWD (above) and in paper bags (below).
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
4
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
Table 1 Conversion factors used in the study, according to Börjesson and Berglund (2007). Type of energy carrier
Primary energy factor
Electricity Heat Diesel
2.2 1 1.18
disconsidered. The energy balance presented in the study is made on the basis of primary energy conversion factors listed in Table 1. 2. Life cycle inventory 2.1. FWD system The content of dry matter (DM) and volatile solids (VS) in ground food waste collected from the FWD-system was determined to 3–5% and 95% (as % of DS) respectively (Bissmont et al., 2015). Electricity consumption in disposers was estimated based on installed power (373 W/s) and the assumption that the mills used 60 s per household and day. Assuming that 0.24 kg dry food waste is ground per household and week (Bissmont et al., 2015), results in an energy consumption of 8.7 kW h/kg DS. Extra water use in kitchen sinks was estimated to 4.2 L/kg DS, based on Davidsson et al. (2011). However, several previous studies have not been able to detect any increase in household water-use due to installation of FWD (Evans, 2012). Collection of ground food wastes is done by a biogas-powered tank-vehicle with a volume of 12 m3. The vehicle makes between 15 and 25 collection routes per day, giving an average fuel consumption of 118 MJ per collection route or 9.8 MJ/m3 collected sludge volume (Jönsson, 2014). The tank is emptied on a monthly basis. Measurements carried out in the framework of this study show that a certain amount of N, P, K and COD is lost as effluent from the tank between collection-occasions (Bissmont et al., 2015). Effluent from the tank is passed on to the sewage treatment plant without further treatment.
collects food waste and residual waste at the same time, but in different compartments. The fraction of separately collected food waste was estimated to 16% mass (Bissmont, 2014), and thus 16% of emissions and energy use from collection was allocated to food waste. The transport distance was estimated to 15 km (both ways), and the energy use to 33.8 MJ/ton DM (Rehnlund, 2010). Food waste collected in paper or plastic bags must be pretreated prior to digestion. This is done with the following process: bag-opener, conveyors, screw press and dilution, based on the concept used at the food waste-treatment plant in the case study area. The screw press separates paper bags, impurities and large size food waste (i.e. larger meat bones, etc.) to a reject fraction. This fraction is transported on a conveyor belt to the incineration chamber in the same facility. Electricity consumption in pretreatment is 19.4 kW h/ton incoming food waste (Bernstad et al., 2012). The amount of reject in the pre-treatment plant has previously been determined to 27%-mass (Bernstad et al., 2012). Very few studies have examined how mechanical pretreatment affects food waste composition and thus little knowledge of how this affects the biogas yield and the amount of nutrients in food waste biomass reaching the digester. According to Bernstad et al. (2012) the distribution of N-tot between slurry and reject after pretreatment with screw press is in line with total mass losses (on wet weight). Thus, about 27% of N-tot-content of incoming food waste contained in the reject, and approximately 73% in slurry. This is also the assumption made in this study. Due to a lack of data, the same assumption is made for phosphorus (P-tot) and potassium. It is assumed that 100% of the paper bags are separated as reject in the pre-treatment process. Energy recovery from incineration of reject is calculated separately for paper bags and food waste, based on the calorific value of 13.7 MJ/kg for paper bags and 10.8 MJ/kg reject (Truedsson, 2010). The contents of a number of elements in food waste is presented in Table 1, based on previous studies. DM and VS was assumed to 35% and 86% of DM respectively, based on Waste Management Sweden (2011).
2.3. Biogas production 2.2. Reference system In this system, food waste is separated by households and discarded in paper bags. Bags used for collection of food waste in Malmö are made from virgin paper. Each bag weighs 15.6 g. Here it is assumed that 3 kg (wet weight) of food waste is disposed of in each paper bag, giving a consumption of 5.2 kg paper bags per ton of food waste (wet weight), or 14.9 kg paper bag per ton DM. Plastic holders used for collection of food waste in households (see Fig. 1) are made of HDPE. Each holder weighs 0.22 kg. The lifetime of the holder is assumed to 5 years. Assuming a food waste generation of 0.33 kg DM/household and week (Bissmont et al., 2015) gives a consumption of 2.56 kg plastic holder per ton DM source separated food waste. Distribution of paper bags and plastic holders from production site to the users in Malmö is not included in the analysis. Household waste from multi-family dwellings is, in Sweden, commonly collected in ‘‘recycling buildings”, i.e. separate, small buildings containing waste bins for separate collection of household waste into several different fractions, such as paper, metal, plastics, and glass, as well as bins for residual waste. As separate collection of food waste does not necessarily mean an increased need for space in such buildings (as waste merely is transferred from bins for residual waste to bins for food waste), electricity and resource use related for maintenance of these buildings was not included in the study. Food waste segregated by households is collected in a 2-compartment truck (6 ton capacity). The truck
Separately collected food waste is treated in a mesophilic, onestage anaerobic digestion reactor, with an average residence time of 20 days. A wet process is assumed for both systems, as this is the, by far, most common technology for anaerobic digestion of household food waste in Sweden (Waste Management Sweden, 2011). In both cases, it was assumed that food waste was to be co-digested with other flows (such as food waste from the industry or manure), and thus mixed to an average DM of around 10%. In this, the difference in water content between the two flows was not reflected in the study. Energy use in the digester is assumed to 17 kW h electricity and 57 kW h heat per ton of wet weight (Waste Refinery, 2013). Electricity use for upgrading of biogas with water scrubber was assumed to 0.351 kW h/Nm3 upgraded gas (Näslund, 2013). Methane emissions during digestion and upgrading are estimated as 1% and 0.8% respectively. As the composition of food waste collected in the two systems differ, also overall methane production will differ. It is assumed that methane production mainly is related to the VS content of the material (Davidsson et al., 2007). Methane production from the FWDsystem amounts to 561 Nm3/ton VS food waste (89% decomposition) based on batch methane potential tests on substrate collected from the investigated tank-system presented in Bissmont et al. (2015) and Magnusson (2015). Methane production from pretreated food waste in the reference system was assumed to 589 Nm3/ton VS based on batch methane potential tests on substrate collected from the pre-treatment plant used as reference in
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
anaerobic digestion of food waste has previously been assumed to between 30% (la Cour Jansen et al., 2007) and 100% (Hirai et al., 2000; Aye and Widjaya, 2005; Lantz et al., 2009; Khoo et al., 2010). How much of the nutrients in the digestate that is plant available and thus can be considered to substitute chemical fertilizers depend on the ratio of nitrogen in organic respective mineral form, timing of application of digestate on farmland and type of crops grown in the field (Jönsson, 2013). In the present study, we assume that digestate is applied on crop with an extensive growth period, which increases the utilization rate of nitrogen in digestate, according to previous studies up to 122–144% compared to mineral fertilizers (Christensson and Blohmé, 2002). A substitution rate of 100% was presented by Waste Management Sweden (2010), and is used in the present study. The replacement rate of mineral phosphorus and potassium fertilizers is, based on previous studies, assumed to 100% (Aye and Widjaya, 2005; Lantz et al., 2009; Khoo et al., 2010; Börjesson and Berglund, 2007). Production of mineral fertilizers can lead to large emissions of greenhouse gases (Jensen and Kongshaug, 2003). This applies mainly to the production of synthetic nitrogen, which is energyintensive and where nitrous oxide is produced as a waste product. Catalytic cleaning of nitrous oxide can however reduce nitrous oxide emissions by 70–90% (Jensen and Kongshaug, 2003). The Swedish fertilizer market is dominated by Yara AB, which in 2013 accounted for 65% of the mineral fertilizers used in Sweden (Nihlén, 2013). Yara has in recent years introduced catalytic reduction of nitrous oxide in a large proportion of their production, and since 2011, the company assures that all nitrogen-containing products sold on the Swedish market have a climate impact below 3.6 kg CO2-eq./kg product (Yara, 2015). However, the Swedish import of fertilizers from Eastern Europe, where catalytic reduction still is rare, has increased in recent years (Nihlén, 2013). As the actual origin of substituted fertilizers is unknown, a worst-case scenario, without catalytic reduction was assumed, based on data for production of mineral fertilizers presented by Börjesson and Berglund (2007). Effects from use of the values presented by Yara (2015) are investigated in a sensitivity analysis. Emissions occurring during storage of digestate will partly relate to the fertilizer content, but previous studies show that factors such as coverage, temperature and duration are more critical for emissions of greenhouse gases during storage of digestate (Hansen et al., 2004). Thus, data used for modelling of storage were collected from a previous case study, assuming typical Swedish conditions. Emissions of methane and ammonia from storage of digestate were assumed to 10% of residual CH4-production and
the study (Truedsson et al., 2010) (80% decomposition, based on Davidsson et al., 2007). 2.4. Use and replacement of energy carriers As the present study has the aim to identify effects of a potential change in current waste management system, a consequential perspective, using marginal data, was seen as relevant. The marginal perspective can be defined as the last energy unit to be used in every moment. The average perspective is instead a way of describing how the production of each energy unit is distributed within a given system and a given period. The system can be regional, national or multi-national, i.e. nations with an interconnected electricity network, or nations under common energy policy regimes, such as through emissions trading schemes. The choice of an average or marginal electricity can have a profound effect on the results of a life cycle assessment (Mathiesen et al., 2010). A common assumption is that the marginal electricity is the one produced to the highest cost. In a long-term perspective, however, one must also take into account initial investment costs. Mattsson et al. (2003) assume that marginal electricity in the EU in the long term will be based on natural gas. As the Nordic electricity system is connected to the central European, it is assumed that the same supposition could be made in the present study. When it comes to heat, the situation is different, as heat often is used in the immediate closeness of the production facility. Even in areas with district heating, these are relatively limited geographically. However, in order to make the study relevant in a broader perspective, a national average of heat production is used in the study. Gode et al. (2011) have compiled data for all fuels currently used for district heating in Sweden, and concluded that the average greenhouse gas impact of 1 kW h of heat equal to 0.089 kg CO2-eq. A marginal perspective is used also on the transport side. This is based on the goal of achieving a fossil free transport sector by 2030, adopted in Sweden (Swedish government, 2008), and use of taxes to decrease the demand for fossil fuels. Thus, it is assumed that produced biogas is used in the transport sector to replace diesel. The environmental impact from the production and combustion of diesel is based on Gode et al. (2011). 2.5. Replacement of nutrients The nutritional content of digestate produced from food waste replaces production and use of mineral fertilizers on farmland. The amount of mineral fertilizers substituted by digestate from
400
GWP (kg CO2-ekv./ton DM food waste)
5
Substituted fertilizers
200
Combustion of refuse
0 -200
FWD
FWD total
RS
RS total
Substituted diesel Emissions from farmland
-400 -600 -800
Biogas production Pre-treatment
-1000
Collection
-1200
Collection material
-1400 Fig. 3. Emissions of greenhouse gases from compared systems, by phase and total. ‘‘Collection” refers to fuel consumption in collection vehicles, while ‘‘Collection material” refers to all impacts related to processes needed for source-separation, including electricity use in food waste disposers. ‘‘Emissions from farmland” refers to storage of digestate and net emissions from substitution of chemical fertilizers. FWD = Food waste disposer system, RS = Reference system (collection of food waste in paper bags).
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
6
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
7% of N-tot content respectively, based on Lantz et al. (2009). Emissions of ammonia from digestate and chemical fertilizers on farmland were assumed to 5 and 1% of total ammonia content respectively (Lantz et al., 2009). Direct and indirect emissions of nitrous oxide were assumed to 1.2% of N-tot content (IPCC, 2006) and 1% of ammonia evaporation (Lantz et al., 2009) respectively, both for digestate and chemical fertilizers.
collected in the reference system is compensated by the electricity generation in later incineration of refuse. These are the main reasons behind the energy balance being advantageous for the reference system ( 13,037 MJ/ton DM compared to 12,337 MJ/ton DM) (Fig. 4).
3. Results
A number of sensitivity analyses were performed to assess the robustness of gained results. Results from these are summarized in Fig. 5.
Overall systems emissions of greenhouse gases are higher from the FWD-system compared to the reference system. The main reasons are use of electricity in disposers as well as emissions from biogas production plant. However, due to a higher energy and nutrient recovery per ton source-separated food waste in the FWD-system, avoided emissions from this system is higher compared to the reference system (Fig. 3). The higher decomposition ratio of VS in anaerobic digestion of food waste in the FWDsystem reduces subsequent emissions from storage of digestate in the same system, which also can be seen in Fig. 3. In the FWD-system, the electricity needed for grinding of food waste (included in the category ‘‘collection material” in Fig. 4) causes a high need for input of primary energy to the system. At the same time, the amount of primary energy needed for production of paper bags and plastic bins in the reference system is low. In addition, electricity use in pre-treatment of food waste
4. Sensitivity analyses
A. The amount of electricity used in the grinding of food waste is largely dependent on user behavior. The use vary between households, as well as between individuals within the same household. To examine the effect of uncertainties related to this process, a sensitivity analysis in which usage increases or decreases by 50% compared with the assumption in the base case was performed. According to the results, this uncertainty would affect overall GHG-emissions from the system by less than 8%. B. In results presented above, losses of food waste from tank system are estimated to 30% of TS, 56% of N-tot, 59% of Ptot and as much as 94% of K. However, large uncertainties were attached to these results, due to difficulties in the measuring of the effluent flow. Thus, in a sensitivity analysis, the
Fig. 4. Primary energy use (PEU) from compared systems, by phase and total.
GWP (kg CO2-eq/ton DM food waste)
A
B
C
Base case (+50%) (-50%) (+20%) (-20%) (+50%) (-50%)
D
E
F
G (-20%) (+20%)
0 -200 -400 -600 -800 -1000 -1200
FWD
RS
Fig. 5. Results from sensitivity analyses. Dotted lines represents results from base cases. FWD = Food waste disposer, RS = Reference system.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
C.
D.
E.
F.
amount lost to the effluent were increased respectively decreased by 20% in all cases. This change would increase respectively decrease overall GHG-emissions by 23%. In the base case, losses of food waste in the pretreatment were estimated to 27% by weight. In a sensitivity analysis, the material loss was increased respectively decreased by 50%, without changing the assumption that paper bags to 100% separated as refuse, and that VS, N-tot, P tot, and K are distributed in the same proportions as in the base case. Results show that this parameter has a large effect on overall results from the reference system. A reduction of losses by 40% already, or from 270 kg to 160 kg per ton collected food waste) would make the reference system preferable in relation to system GHG-emissions. Data used in the base case to describe the environmental impact of chemical fertilizers replaced is in a sensitivity analysis replaced by data from Yara (2013), stating that a maximum of 3.6 kg CO2-eq./kg N-fertilizer product on the Nordic market. Although overall systems emissions of GHGs are vastly impacts by this change, the hierarchy between the two alternatives is unaltered. As seen in Table 2, the content of nitrogen in food waste collected in the FWD-system is higher compared to the content in food waste collected in paper bags. The main reason for this is probably a different composition of the food waste discarded in sinks compared to paper bags. Liquid/semiliquid food such as milk and yoghurt are potentially more commonly disposed of in sinks rather than paper bags. A many times higher nutrient content in this type of waste is likely to be a part of the explanation to the large differences in nutrient composition from food waste collected in the two system. The importance of this difference between the two systems was investigated, assuming the same nitrogen content in food waste collected from the FWD-system as from the food waste system. Results show a clear reduction in avoided GHG-emissions, around 14%, but no change in the hierarchy between compared collection alternatives. The vehicles used for waste collection in Malmo run on biogas. In many other municipalities collection vehicles run on diesel, however. The effect of changing fuel is investigated in a sensitivity analysis, assuming the same energy use per ton
Table 2 Carbon and nutrient balance in FWD and reference system, based on Bissmont et al. (2015). System
Parameter
N
P
K
C
FWD
g/kg DM ground food waste collected from tank g/hh, w (in ground food waste collected from tank) g/hh, w (in outlet) g/hh, w (total) Fraction lost to sewer system (%) Fraction lost in storage/ spreading of digestate (%) Substituted fertilizers (kg/ton TS food waste)
33
2.9
2.4
570
11
0.94
0.79
190
330
14 24 56
1.6 2.6 52
11 12 93
150 340 44
140 470 30
9.0
–
–
0.8
30
2.4
8.3
–
g/kg DM pretreated food waste g/kg DM mechanically separated food waste g/kg DM (total) Fraction lost to incineration in pre-treatment (%) Fraction lost in storage/ spreading of digestate (%) Substituted fertilizers (kg/ton TS food waste)
18 6.5
2.7 1.0
8.6 3.2
18 6.5
24 27
3.7 27
12 27
24 27
9.9
–
–
1.4
13
2.8
7.3
–
Reference system
DM
7
collected food waste as in the base case, and emission factors for production and use of diesel in heavy vehicles from Gode et al. (2011). Such a change would have an insignificant effect on overall GHG-emissions; 1.0% in the FWD-system and 0.3% in the reference system. G. Methane batch tests of food waste collected in the FWDsystem showed that yields varied between 484 and 632 Nm3 CH4/ton VS, representing at most 15% deviation from the average yield used in the study. A sensitivity analysis was performed, decreasing the methane yield by 20% in the FWD-system. Results show that overall avoidance of GWP would decrease by 12%, but not change the hierarchy between compared alternatives. A reduction of the methane yield by 37% in the FWD-system or an increase by 41% in the reference system is needed to change the hierarchy between compared alternatives. 5. Discussion The processes most relevant for gained results were identified through a contribution analysis (Fig. 3). Substitution of diesel and mineral fertilizers are the processes contributing the most to avoided GHG-emissions, while energy use and fugitive emissions from biogas production is the process with the highest emissioncontribution, followed by emissions from storage and spreading of digestate on farmland. Processes more apparent for the actual users of the systems, such as electricity use in disposers, use of collection material and fuel in collection vehicles are of lesser importance to overall systems emissions. However, the primary energy balance shows that use of disposers in households results in a significant contribution to overall primary energy input, while electricity generated in the incineration process reduce the same. Thus, results from the carbon footprint analysis and the primary energy balance are contradictory. As the amount of biomass and thereby potential production of methane and biofertilizers from the system is of key importance for the environmental balance in both systems, losses from tank prior to collection as well as losses in pre-treatment of food waste in the reference system are of major concern. Even with the changes made in the tank system in the area under study in the present work compared to the system studied by Bernstad et al. (2012), it is clear that a large amount of nutrients from grinded food waste still are transported from the collection tank with effluent. Weakly measurements of the nutrient concentration in effluent presented in Bissmont et al. (2015) show no correlation between concentrations and the number of days after collection of tank content, and thus no trade-off between increased transports and reduced losses from the tank. Decreasing losses of nutrients from the tanks would, with the system boundaries chosen in the present work, increase environmental benefits vastly. Thus, how to reduce these losses are an area of further investigation. As the composition of nutrients collected in the two systems differ (N-content being substantially higher in food waste collected from tanks in the FWD-system), it could be stated that the two systems are not comparable. At the same time, it is of relevance to assess also this type of effects of the different types of collection systems. In fact, a key issue in the comparison between the systems is the amount of food waste actually collected in the two systems, i.e. any impact on household waste sorting behavior. Unfortunately, few studies have been performed yet which can provide more insights to this question. It is obvious that some types of food waste, such as larger bones and nutshells, commonly not will be ground in FWDs. However, it is also plausible that the same type of food waste commonly is separated as residues in the subsequent pre-treatment, when screw press is used. At the same time, it is plausible that other types of food waste, such as
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
8
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
liquid dairy products and soups, not are separately collected in the paper bag system, but enter the food waste flow when FWDs are used. 26 kg per capita of liquid food waste is yearly flushed down the drain by Swedish households, while the amount of solid food waste yearly generated by households is 81 kg per capita (Störme et al., 2014). As the present study was based on results from batch methane potential tests as well as analyses of nutrient content, using materials collected from full-scale applications of the two compared systems, any differences in food waste composition could be regarded as adjusted for. System limits chosen in the present study implies that processes that occur in, and on the way down to the sewage treatment plant are disregarded. Including this part of the system would increase uncertainties vastly due to a currently very limited knowledge on degradation processes in the sewage system. Degradation processes will affect parameters such as electricity use and potential changes in the need for external carbon source in wastewater treatment, as well as potential recovery of nutrients and energy through sludge digestion and on-land application of digested sludge. Previous studies on systems with sewage-connected kitchen disposers suggest that incoming organic material largely is separated in primary tanks as particulate matter, and will thus not affect energy use and need for external carbon source, but increase biogas yield from the treatment plant due to increased generation of primary sludge (Evans, 2012). However, these conclusions might not be relevant for tank-connected systems, as carbon and nutrients in effluent to a majority are encountered in soluble forms, and suggest that effects will be dependent on the design and process configuration of the wastewater treatment plant in question. Increased knowledge of these processes are needed to increase our knowledge of the environmental impacts from the investigated system. However, even with this limitation, the evaluation of the two systems can be seen as relevant. The current system boundary setting captures the main difference between the two systems, as it is likely that a large fraction of the food waste lost from the collection tank consists of liquids that will be poured down the drain also in the reference system. As the food waste grinding system investigated in the present study is unique in its kind, the possibilities to compare gained results to previous studies is restricted. In Bernstad and la Cour Jansen (2012), the GWP from a food waste grinding system similar to the one investigated in the present study was estimated to around 650 kg CO2-eq./ton DM. The difference compared to results gained in the present study is explained by the assumed higher biogas production and nutrient recovery from collected waste in the present study compared to the prior, as well as lower fuel use per ton collected food waste due to a higher DM-content in collected grinded food waste. In addition, losses from food waste settling tanks were only estimate in the prior study, while the data used in the present study was based on measurements. Lundie and Peters (2005) compared a sewage connected FWD-system to home-composting, central composting and incineration, showing that the FWD-system, after a well-functioning home-composting system, was preferable in relation to GWP and energy use, but performed worse than other alternatives in relation to water usage. No considerations were taken to potential energy recovery through anaerobic digestion of grinded food waste, and overall GHGemission per ton DM food waste were estimated to 204 kg CO2eq. Diggelman and Ham (2003) state that use of sewage connected FWD-systems and later anaerobic digestion is preferable to central composting, incineration and landfilling of food waste in relation to GWP as well as systems energy input, but also in this study, benefits from energy and nutrient recovery were excluded from the GHG-balance. Comparing results presented here to other studies of more conventional systems for separate collection and anaerobic digestion of household food waste show that avoided
GHG-emissions could be around 12–34% higher in the system presented here (Kirkeby et al., 2006; Fruergaard and Astrup, 2011). However, cross-study comparisons should be handled with care, due to the many factors influencing gained results (type of energy carriers substituted, fertilizer substitution ratio, etc.).
6. Conclusions A system for collection of household food waste through the use of food waste disposers (FWD) in kitchen sinks was evaluated and compared with a reference system were food waste is collected in paper bags for further treatment. The assessment was made using lifecycle assessment methodology. In both cases, collected food waste was treated through anaerobic digestion and digestate was used as fertilizer on farmland. The comparison was made in relation to greenhouse gas emissions as well as use of primary energy.. System emissions of greenhouse gases from collection and treatment of 1 ton of food waste (dry matter), are according to the performed assessment lower from the FWD-system compared to the reference system. The main reasons for the benefits seen in the FWD-system are a higher substitution of mineral nitrogen fertilizer followed by a higher substitution of diesel. Performed sensitivity analyses state that results are robust in relation to several parametric uncertainties assessed in compared systems. However, decreasing losses of organic matter in pre-treatment of food waste collected in paper bags, as well as increased losses of organic matter and nutrients from the FWD-system could change the hierarchy between the systems in relation to greenhouse gas emissions. Due to the many questions still remaining regarding the impacts of an increased amount of nutrients and organic matter to the sewage system, the later treatment of effluent from the FWD-system, as well as treatment of wastewater from kitchen sinks in the reference system, was not included in the assessment. In future work, it would be of relevance to monitor these aspects in order to present a more holistic representation of the systems. Acknowledgements The project was partly financed by the Swedish Energy Agency and the Swedish Gas Technology Centre. Thanks to Roland Svensson (VA SYD), Gertrud Persson (Lund University) and Tomas Wolf (VA SYD) for their contributions. References Assefa, G., Eriksson, O., Frostell, B., 2005. Technology assessment of thermal treatment technologies using ORWARE. Energy Convers. Manage. 46, 797–819. Aye, L., Widjaya, E.R., 2005. Environmental and economic analyses of waste disposal options for traditional markets in Indonesia. Waste Manage. 26, 1180–1191. Bernstad, A., la Cour Jansen, J., 2012. Separate collection of household food waste for anaerobic degradation – comparison of different techniques from a systems perspective. Waste Manage. 32 (5), 806–815. Bernstad, A., Malmquist, L., Truedsson, C., la Cour Jansen, J., 2012. Need for improvements in physical pretreatment of source-separated household food waste. Waste Manage. 33 (3), 746–754 (March 2013). Bissmont, M., 2014. Mimmi Bissmont, VA SYD. Personal Communication. Bissmont, M., Davidsson, Å., Bernstad Saraiva, A., 2015. New collection system for food waste to biogas. Energiforsk Report 2015:100. Bolzonella, D., Pavan, P., Battistoni, P., Cecchi, F., 2003. The under sink garbage grinder: a friendly technology for the environment. Environ. Technol. 24 (3), 349–359. Börjesson, P., Berglund, M., 2007. Environmental systems analysis of biogas systems—Part II: the environmental impact of replacing various reference systems. Biomass Bioenergy 31, 326–344. Christensson, K., Blohmé, H., 2002. Slutrapport – Filborna Biogödsel/Final Report – Biodigestate from Filborna. NRS, Agellus Miljökonsulter and HS Malmöhus, Sweden (in Swedish). Davidsson, Å., Pettersson, F., Bernstad, A., 2011. Förstudie av olika system för matavfallsutsortering medavfallskvarnar. SGC/Svenskt Vatten/Avfall Sverige, Malmö, Sweden.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx Davidsson, Å., Gruvberger, C., Christensen, Thomas Højlund, Hansen, Trine Lund, Jansen, J.l.C., 2007. Methane yield in source-sorted organic fraction of municipal solid waste. Waste Manage. 27, 406–414. Diggelman, C., Ham, R., 2003. Household food waste to wastewater or to solid waste? That is the question. Waste Manage. Res. 21 (6), 501–514. Ecoinvent 3.0. LCI Database, Ecoinvent Center, 2013. Swiss Center for Life Cycle Inventory.. Hansen, M.H., Birkmose, T., Mortensen, B., Skaaning, K., 2004. Effects of separation and anaerobic digestion of slurry on odour and ammonia emission during subsequent storage and land application. In: Proceedings of the 11th International Conference of the FAO ESCORENA Network on the Recycling of Agricultural, Municipal and Industrial Residues in Agriculture, Murcia, Spain, 6-9 October 2004, vol. I. Hirai, Y., Murata, M., Sakai, S., Takatsuki, H., 2000. Life cycle assessment for food waste recycling and management. In: Proceedings from the 4th International Conference on EcoBalance Methodologies for Decision Making in a Sustainable 21st Century. October 31st–November 2nd, 2000, Tsukuba, Japan. ILCD, 2010. International Reference Life Cycle Data System Handbook. General guide for Life Cycle Assessment – Detailed Guidance. Joint Research Centre Institute for Environment and Sustainability, European Commission, Ispra, Italy. IPCC, 2006. Guidelines for National Greenhouse Gas Inventories. The Intergovernmental Panel on Climate Change, IPCC, Geneva, Switzerland. Jensen, T.K., Kongshaug, G., 2003. Energy consumption and greenhouse gas emissions in fertiliser production. In: Proceedings No. 509, Paper Presented to the International Fertiliser Society at a Meeting in London, on 3rd of April 2003. The International Fertilizer Society, York, United Kingdom, ISBN: 0 85310 145 0. Jönsson, G., 2014. Ragnsells – upphandlad entreprenör för körning av slam i Malmö. Jönsson, H., 2013. Personal Communication. Khoo, H.H., Lim, T.Z., Tan, R.B.H., 2010. Food waste conversion options in Singapore: environmental impacts based on an LCA perspective. Sci. Total Environ. 408, 1367–1373. Kirkeby, J.T., Birgisdottir, H., Lund Hansen, T., Christensen, T.H., Bhander, G.S., Hauschild, M., 2006. Evaluation of environmental impacts from municipal solid waste management in the municipality of Aarhus, Denmark (EASEWASTE). Waste Manage. Res. 24, 16–26. Kong, D., Shan, J., Iacoboni, M., Maguin, S.R., 2012. Evaluating greenhouse gas impacts of organic waste management options using life cycle assessment. Waste Manage. Res. 30 (8), 800–812. la Cour Jansen, J., Christensen, T., Davidsson, Å., Lund Hansen, T., Jönsson, H., Kirkeby, J., 2007. Biowaste – Decision Support Tool for Collection and Treatment of Source-Sorted Organic Municipal Solid Waste. TemaNord 2007:602. Nordic Council of Ministers, Copenhagen. Lantz, M., Ekman, A., Börjesson, P., 2009. Systemoptimerad produktion av fordonsgas. En miljöoch energisystemanalys av Söderåsens biogasanläggning. Miljö- och energisystem, LTH, Lund University. Report 69.
9
. Lundie, S., Peters, G.M., 2005. Life cycle assessment of food waste management options. J. Clean. Prod. 13 (3), 275–286. Magnusson, N., 2015. Utvärdering av matavfallskvarnssystem i kv. Fullriggaren, Västra Hamnen i Malmö/Evaluation of food waste disposer systems in Fullriggaren, western harbor in Malmö Master Thesis. LTH, Lund University, VA-teknik, Kemiteknik, Sweden (in Swedish). Marashlian, N., El-Fadel, M., 2005. The effect of food waste disposers on municipal waste and wastewater management. Waste Manage. Res. 23 (20), 20–31. Mathiesen, B., Münster, M., Fruergaard, T., 2010. Energy system analyses of the marginal energy technology in life cycle assessments. In: SETAC Europe 14th Case Studies Symposium – Extended Abstracts, pp. 15–18. Mattsson, N., Unger, T., Ekvall, T., 2003. Effects of perturbations in a dynamic system – the case of Nordic power production in Unger, T. (2003). In: Common Energy and Climate Strategies for the Nordic Countries – A Model Analysis. Chalmers University of Technology, Gothenburg, Sweden. Näslund, D., 2013. Electricity Use in Upgrading of Biogas with Malmber Water Scrubber. Personal Communication. Nihlén, A., 2013. Utrymme för billigare gödsel. (accessed 2016-04-05, in Swedish). Truedsson, C., 2010. Utvärdering av förbehandlingsanläggning för matavfall. Positiva och negativa miljöaspekter och områden för optimering för förbehandlingsanläggningen på Sysav Biotec Master Thesis. Lund University, LTH. Waste Management Sweden, 2010. Biogödselhandbok – Biogödsel frånstorskaliga bio-gasanläggningar. RAPPORT U2010:11. ISSN 1103-4092. Avfall Sverige, Malmö Sweden. Waste Management Sweden, 2011. Hjälpmedel för introduktion av system för insamling av källsorterat matavfall. Rapport U2011:19, Avfall Sverige, Malmö, Sweden. Waste Management Sweden, 2014. Waste Management in Sweden 2013. Avfall Sverige, Malmö. Waste Refinery, 2013. Energi- och kostnadseffektiv biogasproduktion från avfall kartläggning och jämförande av nyckeltal. WRAP, 2009. Household Food and Drink Waste in the UK. ISBN: 1-84405-430-6. . Yara, 2013. Klimatavtryck. . Yara, 2015. Garanti klimatavtryck. .
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
Contents lists available at ScienceDirect
Waste Management journal homepage: www.elsevier.com/locate/wasman
Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation A. Bernstad Saraiva a,⇑, Å. Davidsson b, M. Bissmont c,d a
SAGE/COPPE, UFRJ, Rio de Janeiro, Brazil Water and Environmental Engineering, LTH, Lund University, Lund, Sweden c VA SYD, Malmö, Sweden d Urban Studies, Malmö University, Malmö, Sweden b
a r t i c l e
i n f o
Article history: Received 8 December 2015 Revised 12 April 2016 Accepted 30 April 2016 Available online xxxx Keywords: Food waste disposers Food waste Collection system Lifecycle assessment Waste segregation Energy balance
a b s t r a c t An increased interest for separate collection of household food waste in Sweden has led to development of a number of different collection-systems – each with their particular benefits and drawbacks. In the present study, two systems for collection of food waste in households were compared; (a) use of food waste disposers (FWD) in kitchen sinks and (b) collection of food waste in paper bags for further treatment. The comparison was made in relation to greenhouse gas emissions as well as primary energy utilization. In both cases, collected food waste was treated through anaerobic digestion and digestate was used as fertilizer on farmland. Systems emissions of greenhouse gases from collection and treatment of 1 ton of food waste (dry matter), are according to the performed assessment lower from the FWD-system compared to the reference system ( 990 and 770 kg CO2-eq./ton food waste dry matter respectively). The main reasons are a higher substitution of mineral nitrogen fertilizer followed by a higher substitution of diesel. Performed uncertainty analyses state that results are robust, but that decreasing losses of organic matter in pre-treatment of food waste collected in paper bags, as well as increased losses of organic matter and nutrients from the FWD-system could change the hierarchy in relation to greenhouse gas emissions. Owing to a higher use of electricity in the FWD-system, the paper bag collection system was preferable in relation to primary energy utilization. Due to the many questions still remaining regarding the impacts of an increased amount of nutrients and organic matter to the sewage system through an increased use of FWD, the later treatment of effluent from the FWD-system, as well as treatment of wastewater from kitchen sinks in the reference system, was not included in the assessment. In future work, these aspects would be of relevance to monitor. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Collection of food waste from households for subsequent biogas production is becoming increasingly common in Sweden. The most utilized scheme for collection is use of paper bags for collection of food waste in households and later disposal in single or multicompartment waste bins (Waste Management Sweden, 2014). However, in several other countries, use of food waste disposers (FWD) is a common method for separate collection of food waste from households. Previous studies have suggested that FWDs can present a practical alternative for source-separation of food waste, without ⇑ Corresponding author at: Centro de Gestão Tecnológica – CT2, Rua Moniz de Aragão, No. 360 – Bloco 2, Ilha do Fundão – Cidade Universitária, Rio de Janeiro, RJ 21.941-972, Brazil. E-mail address: [email protected] (A. Bernstad Saraiva).
increasing transports, and through avoidance of problems related to odor and increased need for waste bins (Marashlian and ElFadel, 2005). However, several questions can be raised regarding the effects of FWDs connected to conventional sewer systems. Several potential adverse effects have previously been described. Bolzonella et al. (2003) stated that FWD could cause an increased organic load in the biological step at the wastewater treatment plant (WWTP) and thereby increase energy demand for wastewater treatment. Nilsson et al. (1990) raised problems with increased oil and grease load at WWTPs as well as risks of increased production of H2S in sewerage systems, potentially resulting in corrosion of cement pipes. On the other hand, Evans (2012) and Galil and Yaacov (2001) presented measurements of significant increases in biogas production in WWTP-sludge digestion when 50% of connected households introduced FWD. At the same time, Raunkjaer et al. (1995) showed that the removal of dissolved organic matter and proteins in
http://dx.doi.org/10.1016/j.wasman.2016.04.036 0956-053X/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
2
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
wastewater during sewage transport to WWTP could be considerable, which implies that effects on subsequent WWTP-processes will be of lesser significance. In addition, the effects on WWTP processes due to FWDs will to a large extent depend on the WWTP design. Bolzonella et al. (2003) state that an increased carbon concentration in incoming wastewater gained through use of FWDs can improve the C/N and C/Pratio in WWTPs and, depending on the process design, result in an improved nutrient removal and reduced requirement for external carbon sources. In summary, our knowledge of the quantity and quality of food waste disposed of in FWDs that actually reaches the WWTP as well as net-effects on sewerage systems and WWTP-processes from FWD installation is still limited, and probably influenced by several different factors. Thus, this is certainly an area for further investigations. With the potential problems stated above as a background, and yet a constant search for methods which can facilitate household food waste collection both for users as well as for solid waste and wastewater management organizations, a novel FWD-system was developed in Sweden in 2001. With the aim of combining the benefits from use of FWD for users, without increasing risks for problems in later transport and treatment of household wastewater, a system with tank-connected FWDs was implemented and tested. The system is graphically illustrated in Fig. 1. The system included around 60 households in low-rise buildings in Malmö, a city with around 320,000 habitants in southern Sweden. A similar system was installed in 2007, linked to more than 140 apartments in a high rise building in the same area. The evaluation showed that the design of these systems has not been optimal, with the consequence that the fraction of organic matter in collected material is low, and that a significant amount of organic material is lost from the tank between each emptying (Davidsson et al., 2011). However, the evaluations of these earlier systems were partly based on estimations, due to an absence of flow measurements at the outlet from the tank. As the system can provide large benefits from various perspectives (working environment for waste collectors, maximizing use of high value urban land, reduction of potential risks for clogging of sewage pipes, etc.), a system similar to the ones described above was installed in a newly constructed quarter in the same city in 2010. However,
some changes were made with the intention to optimize the system. 1.1. Aim and scope The purpose of the present study is to evaluate this new system from an environmental perspective, using lifecycle assessment (LCA) methodology, and compare this system to a reference collection system for household food waste. In addition, the aim is also to identify processes with a large impact to overall results, as this constitutes an important basis for further improvements and optimizations. The aim is thus to make a statement about which of the compared systems is more advantageous from an environmental perspective, and the conditions under which this is true. The functional unit in the study is defined as: ‘‘Management and treatment of 1 ton TS source separated food waste from households.” The proposal means that we in the functional unit takes no account of the recycling rate may be different in the different systems. This is because we, in the context of this study, not will be able to establish any general differences between use of food waste disposers and the reference system with regard to separation behavior. In the case of the food waste disposer system, production of biogas and nutrient content in digestate is assessed based on experimental data, while literature data is used for these parameters in the assessment of the reference system. Collection of food waste in paper bags is selected as the reference system, as this system currently is the most common system for food waste collection in Sweden (Waste Management Sweden, 2013). Produced biogas is assumed to substitute diesel as fuel in busses, as upgrading of biogas for use as vehicle fuel was the most common application of biogas in Sweden in 2014 (Swedish Energy Authority, 2015). 1.2. System boundaries Choices of system boundaries should be based on the following principles (ILCD, 2010):
Collecon of grinded FW
Kitchen sink with disposer
Plasc holder for separaon FW in paper bag and disposal in waste bin
Separaon tank
Collecon
Treatment of effluent in WWTP
Pre-treatment
AD of grinded FW
AD of WWTP Sludge
Incineraon of pre-treatment residue and AD of FW
Fig. 1. Graphical representation of the character of the tank-connected food waste disposer system (top) and the reference system (bottom) investigated in the study. FW = Food waste, AD = Anaerobic digestion. Processes with labels in italics are not included in the assessment.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
All parameters of relevance to the outcome of the LCA should be considered. The systems being compared are equivalent. Parameters that are the same for the systems being compared may be excluded. These principles are many times difficult to follow in practice. It is often difficult to know if a process is relevant or not until it is included in the LCA, and one also can see the effects of excluding the same process. Assuming that processes that are seen as relevant in previous studies examining similar systems can be a good indicator but also a risk of repeating mistakes made in previous studies. In addition, systems that are of interest to compare are in many cases different in aspects which challenge exact comparisons. Thus, exact equivalence is therefore often difficult to achieve, but should always be sought. Based on this, system boundaries demonstrated in Fig. 2 were used in the present study. With the selected functional unit and system boundary setting, organic matter found in effluent from the collection tank are excluded from the assessment. The main reason for this choice is the many times contradictory statements of the affect from FWDs on sewage systems and WWTP processes presented above (see e.g. Raunkjaer et al., 1995; Evans, 2012). The potential environmental benefits from reducing these losses are however estimated for an assessment of optimization potentials. As seen in Fig. 2, the present study considers the possible substitution of mineral fertilizers by nutrients in food waste when applied on farmland, while use of digestate was omitted in previous studies of FWD-systems (Lundie and Peters, 2005; Diggelman and Ham, 2003). Differently from several other studies of food waste collection systems (Assefa et al., 2005; Khoo et al., 2010; Kong et al., 2012), the present study also includes production of the material needed for collection of food waste. Previous studies of food waste disposal behavior in the UK show that as much as 24% of all domestic food and drink waste in the
3
country is disposed of in the sewer, although only a very small part of the population have kitchen disposers (WRAP, 2009). In other studies, the actual content of organic matter and nutrients in wastewater from kitchen sinks was measured (Friedler and Butler, 1996). Results vary vastly between studies, to some extent explained by variations in water consumption and inclusion or not of washing machines. However, Friedler and Butler (1996) also shows that the composition of kitchen sink wastewater varies largely between households and over time. Thus, the selection of households and timing of measurements can effect results gained. Recent studies in Sweden performed as questionnaires amongst 2050 households, suggest that around 26 kg food and beverage per person and year are wasted through the kitchen sink (Störme et al., 2014). This equals 23% of total household food wastage, which is similar to results gained in the UK (WRAP, 2009). Coffee and tea, followed by dairy products were responsible for around 65% of the total wastage through kitchen sinks. This fraction of food wastage is of course considered in the FWD-system, while excluded in the reference system. The main reason is a current lack of knowledge of the destiny of this food waste fraction along the sewer system and in the later wastewater treatment plant. Distribution of produced biogas to fuel stations is not addressed, since distribution not is included in data used for description of diesel replaced by biogas. 1.3. Delimitations The study is limited to an investigation of the two collection systems from a climate change perspective, i.e. only those processes generating emissions that contribute to global warming are taken into account in the study. CO2, CH4 and N2O are considered with characterization factors according to IPCC 2007 (Ecoinvent Center, 2013). The greenhouse effect is studied in a 100-year perspective and emissions of biogenic carbon are
Fig. 2. System boundaries used in the assessment of the two collection systems; disposal of food waste in tank-connected FWD (above) and in paper bags (below).
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
4
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
Table 1 Conversion factors used in the study, according to Börjesson and Berglund (2007). Type of energy carrier
Primary energy factor
Electricity Heat Diesel
2.2 1 1.18
disconsidered. The energy balance presented in the study is made on the basis of primary energy conversion factors listed in Table 1. 2. Life cycle inventory 2.1. FWD system The content of dry matter (DM) and volatile solids (VS) in ground food waste collected from the FWD-system was determined to 3–5% and 95% (as % of DS) respectively (Bissmont et al., 2015). Electricity consumption in disposers was estimated based on installed power (373 W/s) and the assumption that the mills used 60 s per household and day. Assuming that 0.24 kg dry food waste is ground per household and week (Bissmont et al., 2015), results in an energy consumption of 8.7 kW h/kg DS. Extra water use in kitchen sinks was estimated to 4.2 L/kg DS, based on Davidsson et al. (2011). However, several previous studies have not been able to detect any increase in household water-use due to installation of FWD (Evans, 2012). Collection of ground food wastes is done by a biogas-powered tank-vehicle with a volume of 12 m3. The vehicle makes between 15 and 25 collection routes per day, giving an average fuel consumption of 118 MJ per collection route or 9.8 MJ/m3 collected sludge volume (Jönsson, 2014). The tank is emptied on a monthly basis. Measurements carried out in the framework of this study show that a certain amount of N, P, K and COD is lost as effluent from the tank between collection-occasions (Bissmont et al., 2015). Effluent from the tank is passed on to the sewage treatment plant without further treatment.
collects food waste and residual waste at the same time, but in different compartments. The fraction of separately collected food waste was estimated to 16% mass (Bissmont, 2014), and thus 16% of emissions and energy use from collection was allocated to food waste. The transport distance was estimated to 15 km (both ways), and the energy use to 33.8 MJ/ton DM (Rehnlund, 2010). Food waste collected in paper or plastic bags must be pretreated prior to digestion. This is done with the following process: bag-opener, conveyors, screw press and dilution, based on the concept used at the food waste-treatment plant in the case study area. The screw press separates paper bags, impurities and large size food waste (i.e. larger meat bones, etc.) to a reject fraction. This fraction is transported on a conveyor belt to the incineration chamber in the same facility. Electricity consumption in pretreatment is 19.4 kW h/ton incoming food waste (Bernstad et al., 2012). The amount of reject in the pre-treatment plant has previously been determined to 27%-mass (Bernstad et al., 2012). Very few studies have examined how mechanical pretreatment affects food waste composition and thus little knowledge of how this affects the biogas yield and the amount of nutrients in food waste biomass reaching the digester. According to Bernstad et al. (2012) the distribution of N-tot between slurry and reject after pretreatment with screw press is in line with total mass losses (on wet weight). Thus, about 27% of N-tot-content of incoming food waste contained in the reject, and approximately 73% in slurry. This is also the assumption made in this study. Due to a lack of data, the same assumption is made for phosphorus (P-tot) and potassium. It is assumed that 100% of the paper bags are separated as reject in the pre-treatment process. Energy recovery from incineration of reject is calculated separately for paper bags and food waste, based on the calorific value of 13.7 MJ/kg for paper bags and 10.8 MJ/kg reject (Truedsson, 2010). The contents of a number of elements in food waste is presented in Table 1, based on previous studies. DM and VS was assumed to 35% and 86% of DM respectively, based on Waste Management Sweden (2011).
2.3. Biogas production 2.2. Reference system In this system, food waste is separated by households and discarded in paper bags. Bags used for collection of food waste in Malmö are made from virgin paper. Each bag weighs 15.6 g. Here it is assumed that 3 kg (wet weight) of food waste is disposed of in each paper bag, giving a consumption of 5.2 kg paper bags per ton of food waste (wet weight), or 14.9 kg paper bag per ton DM. Plastic holders used for collection of food waste in households (see Fig. 1) are made of HDPE. Each holder weighs 0.22 kg. The lifetime of the holder is assumed to 5 years. Assuming a food waste generation of 0.33 kg DM/household and week (Bissmont et al., 2015) gives a consumption of 2.56 kg plastic holder per ton DM source separated food waste. Distribution of paper bags and plastic holders from production site to the users in Malmö is not included in the analysis. Household waste from multi-family dwellings is, in Sweden, commonly collected in ‘‘recycling buildings”, i.e. separate, small buildings containing waste bins for separate collection of household waste into several different fractions, such as paper, metal, plastics, and glass, as well as bins for residual waste. As separate collection of food waste does not necessarily mean an increased need for space in such buildings (as waste merely is transferred from bins for residual waste to bins for food waste), electricity and resource use related for maintenance of these buildings was not included in the study. Food waste segregated by households is collected in a 2-compartment truck (6 ton capacity). The truck
Separately collected food waste is treated in a mesophilic, onestage anaerobic digestion reactor, with an average residence time of 20 days. A wet process is assumed for both systems, as this is the, by far, most common technology for anaerobic digestion of household food waste in Sweden (Waste Management Sweden, 2011). In both cases, it was assumed that food waste was to be co-digested with other flows (such as food waste from the industry or manure), and thus mixed to an average DM of around 10%. In this, the difference in water content between the two flows was not reflected in the study. Energy use in the digester is assumed to 17 kW h electricity and 57 kW h heat per ton of wet weight (Waste Refinery, 2013). Electricity use for upgrading of biogas with water scrubber was assumed to 0.351 kW h/Nm3 upgraded gas (Näslund, 2013). Methane emissions during digestion and upgrading are estimated as 1% and 0.8% respectively. As the composition of food waste collected in the two systems differ, also overall methane production will differ. It is assumed that methane production mainly is related to the VS content of the material (Davidsson et al., 2007). Methane production from the FWDsystem amounts to 561 Nm3/ton VS food waste (89% decomposition) based on batch methane potential tests on substrate collected from the investigated tank-system presented in Bissmont et al. (2015) and Magnusson (2015). Methane production from pretreated food waste in the reference system was assumed to 589 Nm3/ton VS based on batch methane potential tests on substrate collected from the pre-treatment plant used as reference in
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
anaerobic digestion of food waste has previously been assumed to between 30% (la Cour Jansen et al., 2007) and 100% (Hirai et al., 2000; Aye and Widjaya, 2005; Lantz et al., 2009; Khoo et al., 2010). How much of the nutrients in the digestate that is plant available and thus can be considered to substitute chemical fertilizers depend on the ratio of nitrogen in organic respective mineral form, timing of application of digestate on farmland and type of crops grown in the field (Jönsson, 2013). In the present study, we assume that digestate is applied on crop with an extensive growth period, which increases the utilization rate of nitrogen in digestate, according to previous studies up to 122–144% compared to mineral fertilizers (Christensson and Blohmé, 2002). A substitution rate of 100% was presented by Waste Management Sweden (2010), and is used in the present study. The replacement rate of mineral phosphorus and potassium fertilizers is, based on previous studies, assumed to 100% (Aye and Widjaya, 2005; Lantz et al., 2009; Khoo et al., 2010; Börjesson and Berglund, 2007). Production of mineral fertilizers can lead to large emissions of greenhouse gases (Jensen and Kongshaug, 2003). This applies mainly to the production of synthetic nitrogen, which is energyintensive and where nitrous oxide is produced as a waste product. Catalytic cleaning of nitrous oxide can however reduce nitrous oxide emissions by 70–90% (Jensen and Kongshaug, 2003). The Swedish fertilizer market is dominated by Yara AB, which in 2013 accounted for 65% of the mineral fertilizers used in Sweden (Nihlén, 2013). Yara has in recent years introduced catalytic reduction of nitrous oxide in a large proportion of their production, and since 2011, the company assures that all nitrogen-containing products sold on the Swedish market have a climate impact below 3.6 kg CO2-eq./kg product (Yara, 2015). However, the Swedish import of fertilizers from Eastern Europe, where catalytic reduction still is rare, has increased in recent years (Nihlén, 2013). As the actual origin of substituted fertilizers is unknown, a worst-case scenario, without catalytic reduction was assumed, based on data for production of mineral fertilizers presented by Börjesson and Berglund (2007). Effects from use of the values presented by Yara (2015) are investigated in a sensitivity analysis. Emissions occurring during storage of digestate will partly relate to the fertilizer content, but previous studies show that factors such as coverage, temperature and duration are more critical for emissions of greenhouse gases during storage of digestate (Hansen et al., 2004). Thus, data used for modelling of storage were collected from a previous case study, assuming typical Swedish conditions. Emissions of methane and ammonia from storage of digestate were assumed to 10% of residual CH4-production and
the study (Truedsson et al., 2010) (80% decomposition, based on Davidsson et al., 2007). 2.4. Use and replacement of energy carriers As the present study has the aim to identify effects of a potential change in current waste management system, a consequential perspective, using marginal data, was seen as relevant. The marginal perspective can be defined as the last energy unit to be used in every moment. The average perspective is instead a way of describing how the production of each energy unit is distributed within a given system and a given period. The system can be regional, national or multi-national, i.e. nations with an interconnected electricity network, or nations under common energy policy regimes, such as through emissions trading schemes. The choice of an average or marginal electricity can have a profound effect on the results of a life cycle assessment (Mathiesen et al., 2010). A common assumption is that the marginal electricity is the one produced to the highest cost. In a long-term perspective, however, one must also take into account initial investment costs. Mattsson et al. (2003) assume that marginal electricity in the EU in the long term will be based on natural gas. As the Nordic electricity system is connected to the central European, it is assumed that the same supposition could be made in the present study. When it comes to heat, the situation is different, as heat often is used in the immediate closeness of the production facility. Even in areas with district heating, these are relatively limited geographically. However, in order to make the study relevant in a broader perspective, a national average of heat production is used in the study. Gode et al. (2011) have compiled data for all fuels currently used for district heating in Sweden, and concluded that the average greenhouse gas impact of 1 kW h of heat equal to 0.089 kg CO2-eq. A marginal perspective is used also on the transport side. This is based on the goal of achieving a fossil free transport sector by 2030, adopted in Sweden (Swedish government, 2008), and use of taxes to decrease the demand for fossil fuels. Thus, it is assumed that produced biogas is used in the transport sector to replace diesel. The environmental impact from the production and combustion of diesel is based on Gode et al. (2011). 2.5. Replacement of nutrients The nutritional content of digestate produced from food waste replaces production and use of mineral fertilizers on farmland. The amount of mineral fertilizers substituted by digestate from
400
GWP (kg CO2-ekv./ton DM food waste)
5
Substituted fertilizers
200
Combustion of refuse
0 -200
FWD
FWD total
RS
RS total
Substituted diesel Emissions from farmland
-400 -600 -800
Biogas production Pre-treatment
-1000
Collection
-1200
Collection material
-1400 Fig. 3. Emissions of greenhouse gases from compared systems, by phase and total. ‘‘Collection” refers to fuel consumption in collection vehicles, while ‘‘Collection material” refers to all impacts related to processes needed for source-separation, including electricity use in food waste disposers. ‘‘Emissions from farmland” refers to storage of digestate and net emissions from substitution of chemical fertilizers. FWD = Food waste disposer system, RS = Reference system (collection of food waste in paper bags).
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
6
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
7% of N-tot content respectively, based on Lantz et al. (2009). Emissions of ammonia from digestate and chemical fertilizers on farmland were assumed to 5 and 1% of total ammonia content respectively (Lantz et al., 2009). Direct and indirect emissions of nitrous oxide were assumed to 1.2% of N-tot content (IPCC, 2006) and 1% of ammonia evaporation (Lantz et al., 2009) respectively, both for digestate and chemical fertilizers.
collected in the reference system is compensated by the electricity generation in later incineration of refuse. These are the main reasons behind the energy balance being advantageous for the reference system ( 13,037 MJ/ton DM compared to 12,337 MJ/ton DM) (Fig. 4).
3. Results
A number of sensitivity analyses were performed to assess the robustness of gained results. Results from these are summarized in Fig. 5.
Overall systems emissions of greenhouse gases are higher from the FWD-system compared to the reference system. The main reasons are use of electricity in disposers as well as emissions from biogas production plant. However, due to a higher energy and nutrient recovery per ton source-separated food waste in the FWD-system, avoided emissions from this system is higher compared to the reference system (Fig. 3). The higher decomposition ratio of VS in anaerobic digestion of food waste in the FWDsystem reduces subsequent emissions from storage of digestate in the same system, which also can be seen in Fig. 3. In the FWD-system, the electricity needed for grinding of food waste (included in the category ‘‘collection material” in Fig. 4) causes a high need for input of primary energy to the system. At the same time, the amount of primary energy needed for production of paper bags and plastic bins in the reference system is low. In addition, electricity use in pre-treatment of food waste
4. Sensitivity analyses
A. The amount of electricity used in the grinding of food waste is largely dependent on user behavior. The use vary between households, as well as between individuals within the same household. To examine the effect of uncertainties related to this process, a sensitivity analysis in which usage increases or decreases by 50% compared with the assumption in the base case was performed. According to the results, this uncertainty would affect overall GHG-emissions from the system by less than 8%. B. In results presented above, losses of food waste from tank system are estimated to 30% of TS, 56% of N-tot, 59% of Ptot and as much as 94% of K. However, large uncertainties were attached to these results, due to difficulties in the measuring of the effluent flow. Thus, in a sensitivity analysis, the
Fig. 4. Primary energy use (PEU) from compared systems, by phase and total.
GWP (kg CO2-eq/ton DM food waste)
A
B
C
Base case (+50%) (-50%) (+20%) (-20%) (+50%) (-50%)
D
E
F
G (-20%) (+20%)
0 -200 -400 -600 -800 -1000 -1200
FWD
RS
Fig. 5. Results from sensitivity analyses. Dotted lines represents results from base cases. FWD = Food waste disposer, RS = Reference system.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
C.
D.
E.
F.
amount lost to the effluent were increased respectively decreased by 20% in all cases. This change would increase respectively decrease overall GHG-emissions by 23%. In the base case, losses of food waste in the pretreatment were estimated to 27% by weight. In a sensitivity analysis, the material loss was increased respectively decreased by 50%, without changing the assumption that paper bags to 100% separated as refuse, and that VS, N-tot, P tot, and K are distributed in the same proportions as in the base case. Results show that this parameter has a large effect on overall results from the reference system. A reduction of losses by 40% already, or from 270 kg to 160 kg per ton collected food waste) would make the reference system preferable in relation to system GHG-emissions. Data used in the base case to describe the environmental impact of chemical fertilizers replaced is in a sensitivity analysis replaced by data from Yara (2013), stating that a maximum of 3.6 kg CO2-eq./kg N-fertilizer product on the Nordic market. Although overall systems emissions of GHGs are vastly impacts by this change, the hierarchy between the two alternatives is unaltered. As seen in Table 2, the content of nitrogen in food waste collected in the FWD-system is higher compared to the content in food waste collected in paper bags. The main reason for this is probably a different composition of the food waste discarded in sinks compared to paper bags. Liquid/semiliquid food such as milk and yoghurt are potentially more commonly disposed of in sinks rather than paper bags. A many times higher nutrient content in this type of waste is likely to be a part of the explanation to the large differences in nutrient composition from food waste collected in the two system. The importance of this difference between the two systems was investigated, assuming the same nitrogen content in food waste collected from the FWD-system as from the food waste system. Results show a clear reduction in avoided GHG-emissions, around 14%, but no change in the hierarchy between compared collection alternatives. The vehicles used for waste collection in Malmo run on biogas. In many other municipalities collection vehicles run on diesel, however. The effect of changing fuel is investigated in a sensitivity analysis, assuming the same energy use per ton
Table 2 Carbon and nutrient balance in FWD and reference system, based on Bissmont et al. (2015). System
Parameter
N
P
K
C
FWD
g/kg DM ground food waste collected from tank g/hh, w (in ground food waste collected from tank) g/hh, w (in outlet) g/hh, w (total) Fraction lost to sewer system (%) Fraction lost in storage/ spreading of digestate (%) Substituted fertilizers (kg/ton TS food waste)
33
2.9
2.4
570
11
0.94
0.79
190
330
14 24 56
1.6 2.6 52
11 12 93
150 340 44
140 470 30
9.0
–
–
0.8
30
2.4
8.3
–
g/kg DM pretreated food waste g/kg DM mechanically separated food waste g/kg DM (total) Fraction lost to incineration in pre-treatment (%) Fraction lost in storage/ spreading of digestate (%) Substituted fertilizers (kg/ton TS food waste)
18 6.5
2.7 1.0
8.6 3.2
18 6.5
24 27
3.7 27
12 27
24 27
9.9
–
–
1.4
13
2.8
7.3
–
Reference system
DM
7
collected food waste as in the base case, and emission factors for production and use of diesel in heavy vehicles from Gode et al. (2011). Such a change would have an insignificant effect on overall GHG-emissions; 1.0% in the FWD-system and 0.3% in the reference system. G. Methane batch tests of food waste collected in the FWDsystem showed that yields varied between 484 and 632 Nm3 CH4/ton VS, representing at most 15% deviation from the average yield used in the study. A sensitivity analysis was performed, decreasing the methane yield by 20% in the FWD-system. Results show that overall avoidance of GWP would decrease by 12%, but not change the hierarchy between compared alternatives. A reduction of the methane yield by 37% in the FWD-system or an increase by 41% in the reference system is needed to change the hierarchy between compared alternatives. 5. Discussion The processes most relevant for gained results were identified through a contribution analysis (Fig. 3). Substitution of diesel and mineral fertilizers are the processes contributing the most to avoided GHG-emissions, while energy use and fugitive emissions from biogas production is the process with the highest emissioncontribution, followed by emissions from storage and spreading of digestate on farmland. Processes more apparent for the actual users of the systems, such as electricity use in disposers, use of collection material and fuel in collection vehicles are of lesser importance to overall systems emissions. However, the primary energy balance shows that use of disposers in households results in a significant contribution to overall primary energy input, while electricity generated in the incineration process reduce the same. Thus, results from the carbon footprint analysis and the primary energy balance are contradictory. As the amount of biomass and thereby potential production of methane and biofertilizers from the system is of key importance for the environmental balance in both systems, losses from tank prior to collection as well as losses in pre-treatment of food waste in the reference system are of major concern. Even with the changes made in the tank system in the area under study in the present work compared to the system studied by Bernstad et al. (2012), it is clear that a large amount of nutrients from grinded food waste still are transported from the collection tank with effluent. Weakly measurements of the nutrient concentration in effluent presented in Bissmont et al. (2015) show no correlation between concentrations and the number of days after collection of tank content, and thus no trade-off between increased transports and reduced losses from the tank. Decreasing losses of nutrients from the tanks would, with the system boundaries chosen in the present work, increase environmental benefits vastly. Thus, how to reduce these losses are an area of further investigation. As the composition of nutrients collected in the two systems differ (N-content being substantially higher in food waste collected from tanks in the FWD-system), it could be stated that the two systems are not comparable. At the same time, it is of relevance to assess also this type of effects of the different types of collection systems. In fact, a key issue in the comparison between the systems is the amount of food waste actually collected in the two systems, i.e. any impact on household waste sorting behavior. Unfortunately, few studies have been performed yet which can provide more insights to this question. It is obvious that some types of food waste, such as larger bones and nutshells, commonly not will be ground in FWDs. However, it is also plausible that the same type of food waste commonly is separated as residues in the subsequent pre-treatment, when screw press is used. At the same time, it is plausible that other types of food waste, such as
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
8
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx
liquid dairy products and soups, not are separately collected in the paper bag system, but enter the food waste flow when FWDs are used. 26 kg per capita of liquid food waste is yearly flushed down the drain by Swedish households, while the amount of solid food waste yearly generated by households is 81 kg per capita (Störme et al., 2014). As the present study was based on results from batch methane potential tests as well as analyses of nutrient content, using materials collected from full-scale applications of the two compared systems, any differences in food waste composition could be regarded as adjusted for. System limits chosen in the present study implies that processes that occur in, and on the way down to the sewage treatment plant are disregarded. Including this part of the system would increase uncertainties vastly due to a currently very limited knowledge on degradation processes in the sewage system. Degradation processes will affect parameters such as electricity use and potential changes in the need for external carbon source in wastewater treatment, as well as potential recovery of nutrients and energy through sludge digestion and on-land application of digested sludge. Previous studies on systems with sewage-connected kitchen disposers suggest that incoming organic material largely is separated in primary tanks as particulate matter, and will thus not affect energy use and need for external carbon source, but increase biogas yield from the treatment plant due to increased generation of primary sludge (Evans, 2012). However, these conclusions might not be relevant for tank-connected systems, as carbon and nutrients in effluent to a majority are encountered in soluble forms, and suggest that effects will be dependent on the design and process configuration of the wastewater treatment plant in question. Increased knowledge of these processes are needed to increase our knowledge of the environmental impacts from the investigated system. However, even with this limitation, the evaluation of the two systems can be seen as relevant. The current system boundary setting captures the main difference between the two systems, as it is likely that a large fraction of the food waste lost from the collection tank consists of liquids that will be poured down the drain also in the reference system. As the food waste grinding system investigated in the present study is unique in its kind, the possibilities to compare gained results to previous studies is restricted. In Bernstad and la Cour Jansen (2012), the GWP from a food waste grinding system similar to the one investigated in the present study was estimated to around 650 kg CO2-eq./ton DM. The difference compared to results gained in the present study is explained by the assumed higher biogas production and nutrient recovery from collected waste in the present study compared to the prior, as well as lower fuel use per ton collected food waste due to a higher DM-content in collected grinded food waste. In addition, losses from food waste settling tanks were only estimate in the prior study, while the data used in the present study was based on measurements. Lundie and Peters (2005) compared a sewage connected FWD-system to home-composting, central composting and incineration, showing that the FWD-system, after a well-functioning home-composting system, was preferable in relation to GWP and energy use, but performed worse than other alternatives in relation to water usage. No considerations were taken to potential energy recovery through anaerobic digestion of grinded food waste, and overall GHGemission per ton DM food waste were estimated to 204 kg CO2eq. Diggelman and Ham (2003) state that use of sewage connected FWD-systems and later anaerobic digestion is preferable to central composting, incineration and landfilling of food waste in relation to GWP as well as systems energy input, but also in this study, benefits from energy and nutrient recovery were excluded from the GHG-balance. Comparing results presented here to other studies of more conventional systems for separate collection and anaerobic digestion of household food waste show that avoided
GHG-emissions could be around 12–34% higher in the system presented here (Kirkeby et al., 2006; Fruergaard and Astrup, 2011). However, cross-study comparisons should be handled with care, due to the many factors influencing gained results (type of energy carriers substituted, fertilizer substitution ratio, etc.).
6. Conclusions A system for collection of household food waste through the use of food waste disposers (FWD) in kitchen sinks was evaluated and compared with a reference system were food waste is collected in paper bags for further treatment. The assessment was made using lifecycle assessment methodology. In both cases, collected food waste was treated through anaerobic digestion and digestate was used as fertilizer on farmland. The comparison was made in relation to greenhouse gas emissions as well as use of primary energy.. System emissions of greenhouse gases from collection and treatment of 1 ton of food waste (dry matter), are according to the performed assessment lower from the FWD-system compared to the reference system. The main reasons for the benefits seen in the FWD-system are a higher substitution of mineral nitrogen fertilizer followed by a higher substitution of diesel. Performed sensitivity analyses state that results are robust in relation to several parametric uncertainties assessed in compared systems. However, decreasing losses of organic matter in pre-treatment of food waste collected in paper bags, as well as increased losses of organic matter and nutrients from the FWD-system could change the hierarchy between the systems in relation to greenhouse gas emissions. Due to the many questions still remaining regarding the impacts of an increased amount of nutrients and organic matter to the sewage system, the later treatment of effluent from the FWD-system, as well as treatment of wastewater from kitchen sinks in the reference system, was not included in the assessment. In future work, it would be of relevance to monitor these aspects in order to present a more holistic representation of the systems. Acknowledgements The project was partly financed by the Swedish Energy Agency and the Swedish Gas Technology Centre. Thanks to Roland Svensson (VA SYD), Gertrud Persson (Lund University) and Tomas Wolf (VA SYD) for their contributions. References Assefa, G., Eriksson, O., Frostell, B., 2005. Technology assessment of thermal treatment technologies using ORWARE. Energy Convers. Manage. 46, 797–819. Aye, L., Widjaya, E.R., 2005. Environmental and economic analyses of waste disposal options for traditional markets in Indonesia. Waste Manage. 26, 1180–1191. Bernstad, A., la Cour Jansen, J., 2012. Separate collection of household food waste for anaerobic degradation – comparison of different techniques from a systems perspective. Waste Manage. 32 (5), 806–815. Bernstad, A., Malmquist, L., Truedsson, C., la Cour Jansen, J., 2012. Need for improvements in physical pretreatment of source-separated household food waste. Waste Manage. 33 (3), 746–754 (March 2013). Bissmont, M., 2014. Mimmi Bissmont, VA SYD. Personal Communication. Bissmont, M., Davidsson, Å., Bernstad Saraiva, A., 2015. New collection system for food waste to biogas. Energiforsk Report 2015:100. Bolzonella, D., Pavan, P., Battistoni, P., Cecchi, F., 2003. The under sink garbage grinder: a friendly technology for the environment. Environ. Technol. 24 (3), 349–359. Börjesson, P., Berglund, M., 2007. Environmental systems analysis of biogas systems—Part II: the environmental impact of replacing various reference systems. Biomass Bioenergy 31, 326–344. Christensson, K., Blohmé, H., 2002. Slutrapport – Filborna Biogödsel/Final Report – Biodigestate from Filborna. NRS, Agellus Miljökonsulter and HS Malmöhus, Sweden (in Swedish). Davidsson, Å., Pettersson, F., Bernstad, A., 2011. Förstudie av olika system för matavfallsutsortering medavfallskvarnar. SGC/Svenskt Vatten/Avfall Sverige, Malmö, Sweden.
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036
A. Bernstad Saraiva et al. / Waste Management xxx (2016) xxx–xxx Davidsson, Å., Gruvberger, C., Christensen, Thomas Højlund, Hansen, Trine Lund, Jansen, J.l.C., 2007. Methane yield in source-sorted organic fraction of municipal solid waste. Waste Manage. 27, 406–414. Diggelman, C., Ham, R., 2003. Household food waste to wastewater or to solid waste? That is the question. Waste Manage. Res. 21 (6), 501–514. Ecoinvent 3.0. LCI Database, Ecoinvent Center, 2013. Swiss Center for Life Cycle Inventory.
9
Please cite this article in press as: Bernstad Saraiva, A., et al. Lifecycle assessment of a system for food waste disposers to tank – A full-scale system evaluation. Waste Management (2016), http://dx.doi.org/10.1016/j.wasman.2016.04.036