CJ.Banks et al/Bioresource Technology 102 (2011)612-620 613 at high ammonia concentrations can give stable biogas production and type of waste was recorded.Water usage was monitored by at alkaline pH over extended periods of time under continuous separate meters,one for the industrial process water and another loading conditions.In digesters treating food waste these condi- for staff facilities (e.g.toilets and washrooms). tions can also lead to operation at elevated levels of volatile fatty acids in the digestate (Banks et al.,2008:Neiva Correia et al., 2.2.2.Biogas sampling analysis and quantification 2008).Similar conditions have been reported in thermophilic cattle A biogas sample was taken daily from the gas holder feeding the slurry digesters(Neilsen and Angelidaki,2008),and in other nitro- CHP and analysed for methane and carbon dioxide content using a gen rich substrates such as slaughterhouse waste (Banks and GA2000 portable infrared gas analyser(Geotechnical instruments. Wang.1999;Wang and Banks 2003). Leamington Spa,UK).Biogas volumes were recorded on an indus- The current work presents the results of a mass and energy bal- trial gas flow meter,and readings were manually adjusted for ance over a 14-month period for a full-scale food waste digester water vapour content and expressed at standard temperature operating at high ammonia and VFA concentrations.During the and pressure(STP)of 273.15 K and 101.325 kPa. study period the digester was fed mainly on food waste collected from domestic properties mixed with small amounts of commer- 2.2.3.Waste input sampling and analysis cial food waste and municipal green waste.Since the study was Daily composite samples of the shredded feedstock were taken completed the plant has continued to operate successfully as a for analysis of the total solids (TS)and volatile solids(VS)content commercial facility processing food waste. according to Standard Method 2540 G (APHA,2005).Further com- posites were prepared from the daily composites over two-week 2.Methods periods for determination of Total Kjeldahl Nitrogen(N),phospho- rus(P),and potassium(K).Total Kjeldahl N was determined using a 2.1.Digestion plant Kjeltech block digestion and steam distillation unit according to the manufacturer's instructions (Foss Ltd.,Warrington,UK).Sam- The plant was commissioned in March 2006 and for the first ples for Potassium and Phosphorus were extracted using concen- 9 months of operation was fed on mixed kitchen and garden waste trated HNOs in a CEM Microwave Accelerated Reaction System collected from domestic properties.From January 2007 the feed for Extraction (MARSX)(CEM Corporation,North Carolina,USA). was gradually switched to source segregated food waste only. Potassium was quantified using a Varian Spectra AA-200 atomic The study period began on 1 June,2007(day 0).and data for the absorption spectrophotometer (Varian,Australia)according to mass and energy balances was collected for 426 days.During the the manufacturer's instructions.Phosphorus was measured spec- study 3936 tonnes of waste were processed of which 95.5%was trophotometrically by the ammonium molybdate method (ISO source-segregated domestic food waste,with the remainder con- 6878:2004) sisting of commercial food waste from restaurants and local busi- nesses (2.9%)including a small amount of whey,and grass 2.2.4.Digester and digestate sampling and analysis cuttings (1.6%).The food waste received at the plant was first Samples of digestate were taken on a regular basis for analysis. shredded in a rotary counter-shear shredder to reduce the particle Total and volatile solids were measured as above.Ammonia was size,then passed to a feed preparation vessel where it was mixed determined using a Kjeltech steam distillation unit according to with recirculated whole digestate and macerated to give a particle the manufacturer's instructions (Foss Ltd.,Warrington,UK).VFA size less than 12 mm.The feed to the digester was via a buffer stor- were quantified in a Shimazdu GC-2010 gas chromatograph,using age tank providing 3 days storage,to allow continuous feeding over a flame ionization detector and a capillary column type SGE BP-21 weekends and public holidays.The digester itself was a 900 m with helium as the carrier gas at a flow of 190.8 ml min-,with a tank that was completely mixed by continuous gas recirculation split ratio of 100 giving a flow rate of 1.86 ml min-in the column and maintained at 42C by external heat exchangers:the choice and a 3.0 ml min-purge.The GC oven temperature was pro- of temperature was based on the previous experience and prefer- grammed to increase from 60 to 210C in 15 min,with a final hold ence of the plant operator.The digestate was passed batch-wise time of 3 min.The temperatures of injector and detector were 200 to a pasteurisation tank(60 m)where it was heated to 70C for and 250 C,respectively.Samples were prepared by acidification in a minimum of 1 h.Pasteurised digestate was transferred to the dig- 2%formic acid.A standard solution containing acetic,propionic, estate storage tank(900 m3),where it was kept until being ex- iso-butyric,n-butyric,iso-valeric,valeric,hexanoic and heptanoic ported to local farms for use on agricultural land as either acids,at three dilutions giving individual acid concentrations of separated fibre,liquor or whole digestate.The biogas generated 50.250 and 500 mgI-1,respectively,was used for calibration. was used to produce electricity using a 195 kW MAN Combined Alkalinity was measured by titration using 0.25 N H2SO4 to end- Heat and Power(CHP)unit with an assumed electrical conversion points of 5.7 and 4.3(Ripley et al.,1986).Digestate pH was mea- efficiency of 32%at full load and a potential for 53%recovery of heat sured using a combination glass electrode and meter calibrated via the jacket and exhaust cooling water streams.Electricity pro- in buffers at pH 4,7 and 9. duced by the CHP and imports and exports to the grid were all me- tered.The power requirements of the plant were calculated from 3.Results and discussion (CHP generator meter grid import meter-grid export meter). Some of the heat produced by the CHP was fed back into the pro- cess.Temperatures in all tanks were recorded continuously using 3.1.Feedstock characteristics,organic loading rate and retention time a SCADA.More detailed descriptions of individual components of Fig.1 shows values for TS and VS throughout the study period the plant are given in Chesshire(2007)and Arnold et al.(2010). for the domestic food waste and the commercial food waste (not including whey)components of the feedstock.The average solids 2.2.Sampling,measurement and analysis content was similar for domestic food waste(TS 27.7%,VS 24.4%) and commercial food waste (TS 27.8%,VS 24.3%).As can be seen 2.2.1.Quantification of input waste and other materials in Fig.1a and b,there was some variation in the TS and VS content All vehicles delivering waste to the plant were weighed on a of individual samples of domestic food waste but no strong evi- weighbridge before and after discharging their load.The origin dence of seasonal variation and the VS:TS ratio remained fairlyat high ammonia concentrations can give stable biogas production at alkaline pH over extended periods of time under continuous loading conditions. In digesters treating food waste these condi￾tions can also lead to operation at elevated levels of volatile fatty acids in the digestate (Banks et al., 2008; Neiva Correia et al., 2008). Similar conditions have been reported in thermophilic cattle slurry digesters (Neilsen and Angelidaki, 2008), and in other nitro￾gen rich substrates such as slaughterhouse waste (Banks and Wang, 1999; Wang and Banks 2003). The current work presents the results of a mass and energy bal￾ance over a 14-month period for a full-scale food waste digester operating at high ammonia and VFA concentrations. During the study period the digester was fed mainly on food waste collected from domestic properties mixed with small amounts of commer￾cial food waste and municipal green waste. Since the study was completed the plant has continued to operate successfully as a commercial facility processing food waste. 2. Methods 2.1. Digestion plant The plant was commissioned in March 2006 and for the first 9 months of operation was fed on mixed kitchen and garden waste collected from domestic properties. From January 2007 the feed was gradually switched to source segregated food waste only. The study period began on 1 June, 2007 (day 0), and data for the mass and energy balances was collected for 426 days. During the study 3936 tonnes of waste were processed of which 95.5% was source-segregated domestic food waste, with the remainder con￾sisting of commercial food waste from restaurants and local busi￾nesses (2.9%) including a small amount of whey, and grass cuttings (1.6%). The food waste received at the plant was first shredded in a rotary counter-shear shredder to reduce the particle size, then passed to a feed preparation vessel where it was mixed with recirculated whole digestate and macerated to give a particle size less than 12 mm. The feed to the digester was via a buffer stor￾age tank providing 3 days storage, to allow continuous feeding over weekends and public holidays. The digester itself was a 900 m3 tank that was completely mixed by continuous gas recirculation and maintained at 42 C by external heat exchangers: the choice of temperature was based on the previous experience and prefer￾ence of the plant operator. The digestate was passed batch-wise to a pasteurisation tank (60 m3 ) where it was heated to 70 C for a minimum of 1 h. Pasteurised digestate was transferred to the dig￾estate storage tank (900 m3 ), where it was kept until being ex￾ported to local farms for use on agricultural land as either separated fibre, liquor or whole digestate. The biogas generated was used to produce electricity using a 195 kW MAN Combined Heat and Power (CHP) unit with an assumed electrical conversion efficiency of 32% at full load and a potential for 53% recovery of heat via the jacket and exhaust cooling water streams. Electricity pro￾duced by the CHP and imports and exports to the grid were all me￾tered. The power requirements of the plant were calculated from (CHP generator meter + grid import meter grid export meter). Some of the heat produced by the CHP was fed back into the pro￾cess. Temperatures in all tanks were recorded continuously using a SCADA. More detailed descriptions of individual components of the plant are given in Chesshire (2007) and Arnold et al. (2010). 2.2. Sampling, measurement and analysis 2.2.1. Quantification of input waste and other materials All vehicles delivering waste to the plant were weighed on a weighbridge before and after discharging their load. The origin and type of waste was recorded. Water usage was monitored by separate meters, one for the industrial process water and another for staff facilities (e.g. toilets and washrooms). 2.2.2. Biogas sampling analysis and quantification A biogas sample was taken daily from the gas holder feeding the CHP and analysed for methane and carbon dioxide content using a GA2000 portable infrared gas analyser (Geotechnical instruments, Leamington Spa, UK). Biogas volumes were recorded on an indus￾trial gas flow meter, and readings were manually adjusted for water vapour content and expressed at standard temperature and pressure (STP) of 273.15 K and 101.325 kPa. 2.2.3. Waste input sampling and analysis Daily composite samples of the shredded feedstock were taken for analysis of the total solids (TS) and volatile solids (VS) content according to Standard Method 2540 G (APHA, 2005). Further com￾posites were prepared from the daily composites over two-week periods for determination of Total Kjeldahl Nitrogen (N), phospho￾rus (P), and potassium (K). Total Kjeldahl N was determined using a Kjeltech block digestion and steam distillation unit according to the manufacturer’s instructions (Foss Ltd., Warrington, UK). Sam￾ples for Potassium and Phosphorus were extracted using concen￾trated HNO3 in a CEM Microwave Accelerated Reaction System for Extraction (MARSX) (CEM Corporation, North Carolina, USA). Potassium was quantified using a Varian Spectra AA-200 atomic absorption spectrophotometer (Varian, Australia) according to the manufacturer’s instructions. Phosphorus was measured spec￾trophotometrically by the ammonium molybdate method (ISO 6878: 2004). 2.2.4. Digester and digestate sampling and analysis Samples of digestate were taken on a regular basis for analysis. Total and volatile solids were measured as above. Ammonia was determined using a Kjeltech steam distillation unit according to the manufacturer’s instructions (Foss Ltd., Warrington, UK). VFA were quantified in a Shimazdu GC-2010 gas chromatograph, using a flame ionization detector and a capillary column type SGE BP-21 with helium as the carrier gas at a flow of 190.8 ml min1 , with a split ratio of 100 giving a flow rate of 1.86 ml min1 in the column and a 3.0 ml min1 purge. The GC oven temperature was pro￾grammed to increase from 60 to 210 C in 15 min, with a final hold time of 3 min. The temperatures of injector and detector were 200 and 250 C, respectively. Samples were prepared by acidification in 2% formic acid. A standard solution containing acetic, propionic, iso-butyric, n-butyric, iso-valeric, valeric, hexanoic and heptanoic acids, at three dilutions giving individual acid concentrations of 50, 250 and 500 mg l1 , respectively, was used for calibration. Alkalinity was measured by titration using 0.25 N H2SO4 to end￾points of 5.7 and 4.3 (Ripley et al., 1986). Digestate pH was mea￾sured using a combination glass electrode and meter calibrated in buffers at pH 4, 7 and 9. 3. Results and discussion 3.1. Feedstock characteristics, organic loading rate and retention time Fig. 1 shows values for TS and VS throughout the study period for the domestic food waste and the commercial food waste (not including whey) components of the feedstock. The average solids content was similar for domestic food waste (TS 27.7%, VS 24.4%) and commercial food waste (TS 27.8%, VS 24.3%). As can be seen in Fig. 1a and b, there was some variation in the TS and VS content of individual samples of domestic food waste but no strong evi￾dence of seasonal variation and the VS:TS ratio remained fairly C.J. Banks et al. / Bioresource Technology 102 (2011) 612–620 613