Dairy Environmental Systems Program

Transcription

1 Anaerobic Digestion Monitoring at Synergy Biogas, LLC Cornell University Agreement No FINAL REPORT EVALUATION OF THE CONTINUOUSLY-MIXED ANAEROBIC DIGESTER SYSTEM AT SYNERGY BIOGAS FOLLOWING THE PROTOCOL FOR QUANTIFYING AND REPORTING THE PERFORMANCE OF ANAEROBIC DIGESTION SYSTEMS FOR LIVESTOCK MANURES (JUN 2012 TO MAY 2014) Prepared for Wyoming County Industrial Development Agency (WCIDA) Perry, NY Mr. James Pierce Executive Director/CEO/CFO, WCIDA Prepared by PRO-DAIRY Program Department of Biological and Environmental Engineering Cornell University Ithaca, NY Curt A. Gooch, P.E. Principal Investigator Rodrigo Labatut, Ph.D. Postdoctoral Associate November 20, 2014

2 NOTICE This report was prepared by Rodrigo Labatut and Curt Gooch P.E. 1 in the course of performing work contracted for and sponsored by the Wyoming County Industrial Development Agency (hereafter WCIDA). The opinions expressed in this report do not necessarily reflect those of WCIDA, and reference to any specific product, service, process, or method does not constitute an implied or expressed recommendation or endorsement of it. Anaerobic digestion (AD) systems are dynamic in nature, and tend to have elements that are subject to change according to management decisions made on the plant. The farm and digester information in this report was specific to the time period the performance evaluation took place, and could be substantially different at other points in time. 1Licensed in the State of Maryland, No PRO-DAIRY Program - Cornell University - 2 -

3 TABLE OF CONTENTS Notice Table of Contents Executive Summary Introduction Overview Farmstead Anaerobic Digestion (AD) System Methods and Materials Data Collection Sample Collection Instrumentation and Equipment Results Digester Influent: Dairy Cattle Manure Digester Influent: Co-substrates Physicochemical characterization and nutrients analysis Effect of Anaerobic Digestion on Constituents: Waste Stabilization Biogas characterization and production Biogas composition: anaerobic digester Biogas composition: biological scrubber Biogas production: normalized by number of cows Biogas production: normalized by influent biomass Biogas energy Biogas utilization Electrical energy PRO-DAIRY Program - Cornell University - 3 -

4 Biogas to electricity and thermal conversion efficiency Capacity factor and online efficiency Thermal energy AD System Economics Capital cost and grants received Source of revenues Summary and Conclusions References PRO-DAIRY Program - Cornell University - 4 -

5 List of Tables Table 1. Monitoring sampling and data collection period schedule over the entire project (June 2012 to May 2014) Table 2. General information about Synergy Dairy livestock operation Table 3. AD system design assumptions and characteristics Table 4. Data collected/measured on-site at each sampling date Table 5. Data obtained from the SCADA system for each period Table 6. Standard analytical methods used by CES laboratory for sample analyses Table 7. Ranges and precision of the portable, on-site gas measuring device used for monitoring (Multitec 540) Table 8. Specifications of equipment used in performance evaluation data collection Table 9. Monitoring summary data Table 10. Monthly animal management group average populations and calculated manure produced Table 11. Basic description and classification of wastes imported by Synergy since the start of the monitoring Table 12. Monthly volume (gal) of co-substrates received per source during the entire monitoring project (June 2012 May 2014) Table 13. Digester influent mixed characterization (average of triplicates) Table 14. Digester effluent characterization (average of triplicates) Table 15. Raw manure characterization (average of triplicates) Table 16. Imported waste characterization (average of triplicates) Table 17. Overall statistics of the digester influent mixed and digester effluent characteristics over the entire monitoring project Table 18. Estimated mass and ratio of nutrients in the digester influent by monitoring period Table 19. Estimated mass and ratio of nutrients in the digester effluent by monitoring period Table 20. Estimated mass and ratio of nutrients contributed by raw manure Table 21. Estimated mass and ratio of nutrients contributed by imported waste Table 22. Calculated water vapor density of biogas for each sampling date Table 23. Initial capital costs for the Synergy Biogas AD system Table 24. Net energy sold and potential revenues from electricity and REC s sales by month PRO-DAIRY Program - Cornell University - 5 -

6 Table 25. Statistics of main parameters obtained during the 24 months of monitoring project; values are per sampling period PRO-DAIRY Program - Cornell University - 6 -

7 List of Figures Figure 1. Aerial view of the Synergy Biogas AD system and Synergy Dairy farmstead. The lactating.. cow barns are on the left side of the picture, while the anaerobic digestion system is located on the right side Figure 2. Synergy Biogas AD system flow diagram (simplified); meters and sampling locations are identified in circular labels; segmented biogas line between CHP and boiler denotes previous configuration (see text) Figure 3. Continuously-stirred tank reactor (background) and pasteurization units (foreground) Figure 4. Daily loading rate, co-digestion ratio (food waste to influent), biogas production, and CHP instant power at the Synergy AD system during the two years of monitoring (May 2012 May 2014) Figure 5. Percent contribution (volume basis) of each imported waste to the Synergy digester based on industry sector and waste characteristics Figure 6. Percent change in chemical parameters concentrations in the digester effluent relative to the influent for each sampling date; a positive change represents a decrease, while a negative change represents an increase Figure 7. Top: average mass of constituent in the digester influent and effluent for the entire monitoring project; Bottom: average percent change in constituent mass in the digester effluent relative to the influent for the entire monitoring project; error bars represent the standard deviation Figure 8. Percent contribution of total volatile solids, VS (blue bar), and fixed solids, FS (red bar) in the AD influent (top) and effluent (bottom) for sampling dates; total solids represent 100% of the dry influent material (i.e., VS + FS) Figure 9. Methane and carbon dioxide content in biogas measured at each sampling date from the digester vessel headspace on a dry biogas basis (moisture corrected) Figure 10. Concentration of hydrogen sulfide before and after the biological scrubber and its corresponding efficiency at each sampling date Figure 11. Average daily biogas production per lactating cow (LC) contributing to the digester observed for each period; the number shown on top of the bars is the average lactating cows for the corresponding period PRO-DAIRY Program - Cornell University - 7 -

8 Figure 12. Average specific biogas production as a function of the total influent biomass observed for each period; the number on top of the bars indicates the ft 3 per lb of biomass Figure 13. Total energy (MMBtu) produced by the anaerobic digester for each period Figure 14. Average daily electrical energy production by CHP observed for each period Figure 15. Volume of biogas required to generate a unit of electrical energy (kwh) and thermal conversion efficiency of the engine-generator set calculated for each period Figure 16. Capacity factor and online efficiency calculated for each period Figure 17. Breakdown of problems generating downtime at Synergy Figure 18. Thermal energy recovered from CHP (MMBtu) and percent of recovered energy to the total energy produced by digester (see Figure 13) for each period. No data available from Periods 1 8 and for Period Figure 19. Monthly electrical energy generated/sold and parasitic load of the Synergy AD system throughout the monitoring; No parasitic load data were available from Jun to Sep 2012 and for April PRO-DAIRY Program - Cornell University - 8 -

9 EXECUTIVE SUMMARY The Synergy Biogas, LLC anaerobic digester system was monitored for two years, from June 2012 to May 2014, using the EPA Protocol for Quantifying and Reporting the Performance of Anaerobic Digestion Systems for Livestock (EPA, 2011), as a guideline. The Synergy Biogas anaerobic digester, supplied by CH4 Biogas, LLC, was a continuously-stirred tank reactor operated in the mesophilic temperature range, designed to co-digest multiple off-farm streams with dairy manure and produce the necessary biogas to power a 1.4-MW internal combustion engine-generator set. During the monitoring, effluent solids were separated and used by the Synergy farm for cow stall bedding, while the separated liquid was stored in the farm s uncovered earthen long-term storage for use as cropland fertilizer. For the entire monitoring project, an average of 1,891±62 lactating cows per day from Synergy Dairy contributed manure to the digester. The average daily loading rate of the digester was 80,408±19,266 gal, where the average percent of imported waste (mostly food-grade residues) codigested with manure was 25±6% on a volume-to-volume (v/v) basis. The average reduction of organic matter thru the monitoring project was 42% with respect to the influent, while 75% of the odor-causing volatile fatty acids were reduced. In comparison, a previous monitoring study reported by the authors in five manure-based co-digestion operations showed a reduction in organic matter and volatile acids between 36% and 53% and 85% and 91%, respectively (Gooch et al., 2011). The average daily digester biogas production for the entire monitoring project was 495±78 ft 3 per 1,000 lbs of total influent added to the digester, or 173±34 ft 3 per cow contributing to the digester. The engine-generator set produced an average of 23±7 MWh of electricity per day, from which the average daily parasitic load of the AD system was 3±1 MWh, accounting for approximately 14% of the electricity generated by the plant. Overall, the average capacity factor and online efficiency of the Synergy AD system during the entire monitoring project were 0.66±0.22 and 80±23%, respectively. The electrical energy generated translated into an overall thermal conversion efficiency of 42±4%. Also, an additional 13±5% of the total energy in the biogas was recovered by the engine as hot water. Thus, an overall 55% (electrical + thermal) of the total energy contained in the input biogas was recovered by the engine-generator set during the monitoring project. PRO-DAIRY Program - Cornell University - 9 -

10 The majority of the challenges experienced by the Synergy AD system were of mechanical origin, whereas 20% were related to the biological process; only 8% of the downtime was due to scheduled systems maintenance. Some of the problems were related to the extreme cold conditions experienced in the Northeast during the period from December 2013 to February According to NOAA s National Climatic Data Center, this period was the 34th coldest for the contiguous 48 states since modern records began in 1895, with an average temperature of 31.3F, 1.0F below the 20th century average (NOAA, 2014). A summary of the most relevant data produced during the 24 months of this project is shown in the table below. Project Summary. Statistics of main parameters obtained during the monitoring project per month (i.e., sampling period). System parameters Average Standard deviation Lactating cows contributing to digester 1, Influent biomass (gal) 2,377, ,410 Digestate volume (gal) 1,852, ,644 Co-digestion (% imported waste) 25 6 Hydraulic retention time (d) 29 7 Electrical energy generated (MWh) Instantaneous power (MW) Plant capacity factor (decimal) Plant online efficiency (%) CHP thermal efficiency (%) 42 4 PRO-DAIRY Program - Cornell University

11 1. Introduction The U.S. EPA Protocol for Quantifying and Reporting the Performance of Anaerobic Digestion Systems for Livestock Manures (EPA, 2011) was used as a guide for the collection and compilation of performance data for the Synergy Biogas anaerobic digester (AD) system. As stated in this document, adherence to the protocol will provide an objective and unbiased assessment of individual system performance and provide vendors with the ability to demonstrate the validity of performance claims. Furthermore, it will provide a common basis for comparing the performance of different designs approaches and a standard for acceptance of performance evaluation reports if a central repository is created in the future, and it can serve as an appropriate basis for developing technology-specific certification programs. The performance evaluation of the Synergy Biogas AD system took place from June 2012 to May In accordance with the EPA protocol, once-per-month samples were taken of the digester influent, consisted of a mixture of cow manure and food-grade wastes, and the digester effluent. However, during the first three months additional samples were collected to better assess influent biomass variability, and samples from the food waste and raw manure receiving tanks were collected to evaluate their characteristics previous to mixing. In addition, during each visit, on-site biogas composition measurements were performed and animal population information from the farm was collected. Finally, continuous data collected by the supervisory, control, and data acquisition (SCADA) system at Synergy Biogas were processed, compiled and analyzed periodically. For purposes of clarity in this report, each on-site visit is referred to as a sampling date, even though much more was completed than sample collection. Also, the interval period between sampling dates, where data from the SCADA system was collected, is simply referred to as period. The two years of monitoring period is referred to as the entire monitoring project. Several offfarm substrates were co-digested with Synergy Farm s cow manure at the Synergy Biogas plant for this report, these are referred to as food wastes or imported feedstocks. The monitoring schedule for sampling and data collection is shown in Table 1. PRO-DAIRY Program - Cornell University

12 Table 1. Monitoring sampling and data collection period schedule over the entire project (June 2012 to May 2014). Month Period On-site sampling date Data collection period July 1 7/12/2012 6/7/2012 7/12/2012 July 2 7/31/2012 7/13/2012 7/31/2012 August 3 8/16/2012 8/1/2012 8/16/2012 September 4 9/17/2012 8/17/2012 9/17/2012 October 5 10/15/2012 9/18/ /15/2012 November 6 11/16/ /16/ /16/2012 December 7 12/13/ /17/ /13/2012 January 8 1/17/ /14/2012 1/17/2013 February 9 2/15/2013 1/18/2013 2/15/2013 March 10 3/14/2013 2/16/2013 3/14/2013 April 11 4/15/2013 3/15/2013 4/15/2013 May 12 5/15/2013 4/16/2013 5/15/2013 June 13 6/21/2013 5/16/2013 6/21/2013 July 14 7/17/2013 6/22/2013 7/17/2013 August 15 8/16/2013 7/18/2013 8/16/2013 September 16 9/16/2013 8/17/2013 9/16/2013 October 17 10/17/2013 9/17/ /17/2013 November 18 11/15/ /18/ /15/2013 December 19 12/19/ /16/ /19/2013 January 20 1/14/ /20/2013 1/14/2014 February 21 2/20/2014 1/15/2014 2/20/2014 March 22 3/14/2014 2/21/2014 3/14/2014 April 23 4/15/2014 3/15/2014 4/15/2014 May 24 5/15/2014 4/16/2014 5/15/2014 PRO-DAIRY Program - Cornell University

13 1.1. Overview CH4 Biogas, LLC formed Synergy Biogas, LLC to own and operate an anaerobic digester at the Synergy Dairy in Covington, New York. Synergy Dairy leased the site to Synergy Biogas for the facility and supplied manure to the digester. Synergy Dairy was owned by Synergy, LLC, a partnership of large-scale dairy and field crop producers in Western New York State organized to develop and manage projects that benefit from their complementary strengths. Mr. John Noble was the President and CEO of Synergy, LLC. CH4 Biogas, LLC was formed by Mr. Paul Toretta, Mr. Bob Blythe and Bigadan A/S in 2008 in response to market demand for anaerobic digestion and alternative energy in the United States. Ms. Lauren Toretta was the current president of CH4 Biogas at the time this report was released. The anaerobic digester co-digested manure from over 1,890 milking cows at the dairy location with imported food-grade organic waste, which was transported to the site. Biogas from the digester was used to fuel an engine-generator set with capacity to generate up to 1.4 MW of electrical power. In addition, a screw-press solid-liquid separator was used to separate solids for bedding and liquid for land application Farmstead Basic information about Synergy Dairy is provided in Table 2. Also, an aerial view of Synergy Biogas AD system and Synergy Dairy farmstead is shown in Figure 1. Table 2. General information about Synergy Dairy livestock operation. Name of operation Synergy Dairy, LLC Address 6540 Lemley Road, Covington, NY Type of operation Dairy Farm Land base 750 acres Dairy Information 1 a. Breed Holstein b. Average no. lactating cows 1,800 c. Average no. dry cows 200 d. Average no. replacements - e. Fraction of manure collected for AD 100% f. Type of manure collection system Alley scrapers g. Frequency of manure collection Constant 1 The animal numbers shown represent average values for those animal units on the farm that contributed manure to the AD (data collected during the performance monitoring project). PRO-DAIRY Program - Cornell University

14 Figure 1. Aerial view of the Synergy Biogas AD system and Synergy Dairy farmstead. The lactating cow barns are on the left side of the picture, while the anaerobic digestion system is located on the right side Anaerobic Digestion (AD) System The Synergy Biogas AD system was designed and built by the Danish contractor Bigadan A/S. Manure was supplied by lactating cows from the Synergy farm, and supply contracts were continuously being negotiated for the food-grade substrates incorporated to the anaerobic digester. The effective treatment volume of the anaerobic digester was 2.2 million gallons. Based on the monitoring project data, the average hydraulic retention time (HRT) was calculated to be 27 days this was based on an average total influent of 81,000 gal per day, which was composed of 60,000 gal of manure and 21,000 gal of off-farm wastes. A flow diagram of the Synergy Biogas AD system is shown in Figure 2. The diagram shows the meter locations were data were acquired continuously by the SCADA system, as well as the locations were biogas content was measured and biomass samples were collected on a monthly basis. PRO-DAIRY Program - Cornell University

15 In Figure 2, two receiving tanks are shown, each 100,000 gal in total volume; one to store imported food waste substrates and one to store the manure supplied by the farm. Manure was collected via skid steer three times per day during milking shifts and scraped to a below-grade, in-barn reception pit center. Manure flowed by gravity to a central collection pit and was pumped to one of the 100,000-gal receiving tanks. Manure and contents of the food waste receiving tank were pumped one at a time (sequentially) via a macerator pump through a heat exchange system to be combined in one of three 8,000-gal batch process pasteurization tanks. The influent blend was heated to 158F and maintained for one hour for pasteurization. Pasteurized biomass was pumped to the digester through a heat exchange system to bring the temperature of the biomass to a mesophilic range, approximately 100F. Digested effluent (i.e., digestate) was transferred to a biomass/gas storage tank with an approximate 4-day HRT, and then pumped to two screw-press units (Integrity Ag Systems) for liquid-solid separation. Liquid effluent was diverted to the farm s long-term earthen storage, which was then applied on cropland. The separated solid portion was used for cow bedding at the dairy farm. According to the Synergy Dairy manager, enough bedding was produced for the 10 freestall barn cow pens located in three separate barns. PRO-DAIRY Program - Cornell University

16 Synergy Dairy Synergy Biogas SF SF Heat exchangers FT Farm barn effluent Milking center wastewater Imported food-grade waste Manure receiving tank Food waste receiving tank FT Pasteurization tanks Nomenclature Gas flow Biomass flow Water/coolant flow Electricity flow SFTP Heat storage S: Sampling location F: Flow meter T: Temperature meter P: Pressure meter STP H 2 S biofilter SF T1 E: Electricity meter T2 Q: Heat meter Condensate trap SF Anaerobic digester QFT FT Exhaust Liquid digestate to lagoons Blower FTP CHP E To AD system/grid Bedding to barns F Propane Boiler Solid-liquid separators Gas and digestate storage Flare Exhaust Figure 2. Synergy Biogas AD system flow diagram (simplified); meters and sampling locations are identified in circular labels; segmented biogas line between CHP and boiler denotes previous configuration (see text). PRO-DAIRY Program - Cornell University

17 The anaerobic digester was a cylindrical upright continuously-stirred tank reactor (74 feet diameter, 72 feet high) made of bolted carbon steel construction (Figure 3). The top ring and roof were stainless steel, to resist corrosion. Mixing was provided by a single roof-mounted mixer with two large impellers driven by a central vertical shaft. The tank was insulated with 3-½ inch fiberglass batt insulation between the exterior siding and carbon steel wall. Figure 3. Continuously-stirred tank reactor (background) and pasteurization units (foreground). PRO-DAIRY Program - Cornell University

18 Biogas produced by the digester was treated by a biological filter to reduce biogas hydrogen sulfide (H 2S) concentrations. The unit had a 21,000-gal packed-media column. According to the manufacturer, the scrubber was designed to reduce H 2S concentration to less than 100 ppm when the microbiological community was operating at full capacity. Scrubbed biogas was contained in a 396,000-gal biomass/gas storage tank with a flexible membrane cover. Biogas flowed by pressure through a condensate well, and blowers were used to pressurize the biogas for the CHP, boiler or flare. A 2.9-MBtu/hr capacity gas boiler (Columbia, MPH-70) was included in the plant. For the first half of the monitoring the boiler could be fueled by either natural gas or biogas, but by the beginning of the second half it was changed to be used with natural gas exclusively. The boiler provided heat for biomass pasteurization and anaerobic digestion when the electrical generator was idle. Excess biogas was combusted by a flare (C-DEG, GmbH) with a 3.5 MW capacity. After treated via the biological scrubber, the biogas flowed through a condensate well to reduce moisture and via a blower was supplied to a GE Jenbacher model 420 combined heat and power (CHP) unit. The containerized CHP unit, which was installed by North East Energy Systems, had a rated capacity of 1,426-kW of electric power output and 5,411 MBtu/hr of thermal output. After meeting the AD system demands, all the electricity generated was sold through NYISO grid to counterparties in the Northeast. The project received a NYSERDA grant under the Anaerobic Digester Gas to Electricity Program. Engine oil changes were be performed every 2,000 hours, and a complete rebuild of the engine was scheduled at 60,000 hr as required by the manufacturer. A summary of the main characteristics of the Synergy Biogas anaerobic digestion system is presented on Table 3. PRO-DAIRY Program - Cornell University

19 Table 3. AD system design assumptions and characteristics. Type of digester Continuously-Stirred Tank Reactor System representative company CH4 Biogas, LLC a. Address 30 Lakewood Circle N, Greenwich, CT b. Contact information System designer company Bigadan A/S a. Address Vroldvej 168, DK-8660 Skanderborg, Denmark b. Contact information Digester design assumptions a. Number and type of animals 1,860 mature Holstein equivalents b. Manure volume (gal/day) 55,000 c. Off-farm waste volume (gal/day) 40,000 d. Wastewater volume (gal/day) 5,000 e. Bedding type Post-digested separated manure solids f. Pre-treatment for AD Pasteurization Solid-liquid separation with solids for bedding and g. Treatment of AD effluent liquid for long-term earthen storage for land application h. Method of AD effluent storage Gas storage tank (396,000 gal) and long-term earthen storage Physical description a. General description Above-grade; Steel body and stainless steel top ring and roof b. Dimensions h = 72 feet, D = 74 feet c. Insulation 3-½ inch fiberglass batt insulation between siding and carbon steel wall d. Operating volume 2.2 million gallons e. Design HRT 22 days f. Design operating temperature 100F g. Compliance with Farm s CAFO permit Yes Biogas utilization Generation of electricity a. Type of electric generation unit Internal combustion engine; GE Jenbacher model 420; 1,426-kWe; 5,411 MBtu/hr (thermal) b. Component integration North East Energy Systems c. Origin of equipment controller Bigadan d. System installer North East Energy Systems e. Stand-alone capacity Yes f. Pre-treatment of biogas Biological hydrogen sulfide reduction system g. Exhaust gas emission control None h. Utility interconnect NYISO Grid i. Engine-generator set waste heat utilization Total system cost $7,750,000 Cost basis As built Itemized list of component capital costs Shown in Table 23, pg. 69 Heat source: reclaimed engine heat of combustion; Waste heat utilization: pasteurization tanks PRO-DAIRY Program - Cornell University

20 2. Methods and Materials This section explains in detail the collection of project data, sampling of digester influent and effluent, equipment and instrumentation used and installed, and dairy cattle population. Details about each topic are discussed in depth throughout this section Data Collection In addition to the digester influent and effluent samples taken during each monthly sampling date, on-site measurements were taken and data were manually recorded from equipment and plant logs. A particularly important log, were the imported feedstocks brought on-site for inclusion to the AD. This log recorded the date and time feedstock was delivered, the type of feedstock, and the volume delivered. The specific data collected/measured are presented on Table 4. Table 4. Data collected/measured on-site at each sampling date. Item 1. Date and time of readings 2. CH 4, CO 2, O 2, and H 2S concentrations in biogas after digester 3. CH 4, CO 2, O 2, and H 2S concentrations in biogas after bio-scrubber 4. Engine-generator set run time 5. Cumulative electricity purchased and sold 6. Daily animal populations since previous sampling event 7. Logs of imported feedstocks 8. Problems occurred during period Further, the following data from the supervisory, control, and data acquisition (SCADA) system at the Synergy Biogas plant were downloaded, compiled and analyzed for each period (Table 5). SCADA data were generated from an array of sensors and meters (see Table 8) originally installed by the company that designed and built the digester, i.e., Bigadan A/S. Sensors and meters were factory calibrated, but no further calibration took place during the monitoring period. PRO-DAIRY Program - Cornell University

21 Table 5. Data obtained from the SCADA system for each period. Parameter 1. Total influent to pasteurization 2. Food waste to pasteurization 3. Manure to pasteurization 4. Biomass from pasteurization to digester 5. Effluent digester to storage tank 6. Biogas production digester 7. Biogas to generator 8. Generator electrical energy output 9. Generator thermal energy recovered 10. Digester vessel upper temperature 11. Digester vessel lower temperature 2.2. Sample Collection Overall, digester influent and effluent samples were collected with the goal of obtaining representative samples. To do this, grab samples were collected directly from both the digester influent and effluent lines over a period of approximately 30 min during a pumping sequence, to develop a 5-gallon composite, master-sample. The entire volume of this sample was then agitated using a paint mixer powered by a portable electric drill until visibly determined to be homogenized. A 1-liter composite sample was immediately taken and stored on ice, and subsequently frozen before being sent for laboratory analysis. Samples were taken in this fashion approximately every 30 days over the 24-month monitoring period. Additionally, samples coming from the raw manure receiving tank and from the combined imported feedstocks tank were also obtained for two sampling dates at the beginning of the monitoring project to characterize the individual influent streams to the digester. All samples collected during the 24-month monitoring period were sent for analysis to Certified Environmental Services (CES) laboratory in Syracuse, NY, approved by the New York State Department of Health, Environmental Laboratory Approval Program (NYSDOH-ELAP #11246). All samples were analyzed in triplicate for: total solids (TS), total volatile solids (VS), chemical oxygen PRO-DAIRY Program - Cornell University

22 demand (COD), ph, and total volatile acids as acetic acid (TVFA). In addition, the following nutrients were determined in triplicate: total phosphorus (TP), ortho-phosphorus (OP), total Kjeldahl nitrogen (TKN), ammonia-nitrogen (NH 3-N) and potassium (K). CES completely blended each sample in a food processor and subsequently diluted the sample with de-ionized water; the amount of dilution depended on the concentration history of the sample. CES followed the appropriate testing methods outlined in Table 6 for each parameter measured. Table 6. Standard analytical methods used by CES laboratory for sample analyses. Parameter Standard Total Solids (TS) EPA Total Volatile Solids (VS) EPA Fixed Solids (FS) EPA Volatile Acid as Acetic Acid (TVFA) SM C Chemical Oxygen Demand (COD) SM B ph SW Total Kjeldahl Nitrogen (TKN) EPA Ammonia-Nitrogen (NH 3-N) SM F Organic-Nitrogen (ON) By subtraction: TKN - NH 3-N Total Phosphorous (TP) EPA Ortho Phosphorous (OP) EPA Total Potassium (K) EPA SW Instrumentation and Equipment Refer to Figure 2 for details on the locations of meters and sensors within the plant. The volume of biogas produced by the digester was measured using a temperature and pressure compensated Esters gas flow meter, model Fluidistor Gas Flowmeter GD 100/LRM (Esters Elektronik GmbH, Germany), installed in the biogas line on top of the digester vessel. The volume of biogas to the engine-generator (CHP) system was measured using a separate Esters gas flow meter installed in the biogas line after the gas storage tank and the blower. During the first year of the monitoring project (i.e., until Period 13), the boiler and CHP system were in the same biogas line, and therefore shared the same Esters meter (shown by a segmented line in Figure 2). However, since neither the PRO-DAIRY Program - Cornell University

23 engine-generator nor the boiler was operated at any given time, it was possible to easily determine the flow of biogas to each component for analysis. For the second half of the monitoring, the boiler was taken out of the biogas line, and converted to propane; thus, it was decided that the biogas to boiler data would not be relevant for this project, and therefore it was not included. A Sage gas meter (model Prime) was installed by Synergy Biogas to measure the CHP biogas flow exclusively; however, according to the AD plant manager its accuracy was questioned since readings were highly variable. Therefore, it was decided not to include these data. Biomass from the food waste receiving tank (#1) and from the raw manure receiving tank (#2) to the pasteurization tanks was pumped sequentially, each biomass stream at a time, through the same piping line. Thus, a single Siemens mag flow transmitter (model SITRANS F M MAG 3100) located in the pipe was able to measure the flow rate for both streams separately. Likewise, the effluent flow from the digester vessel was measured with a Siemens mag flow transmitter, as was the effluent flow from the gas storage tank. Thermal energy in the form of hot water was recovered from the engine section of the CHP unit. The water flow rate and temperatures before and after the CHP engine recovery heat exchanger were continuously measured and reported by the GE Jenbacher control panel (model DIA.NE XT), which in turn exported these data to the SCADA system. The Esters biogas flow meters and Siemens biomass flow meters, installed as part of the system s standard components, were new and factory-calibrated six months prior to the beginning of the monitoring; however, no opportunity existed to check the calibrations or re-calibrate the meters during the project. The plant had a single bi-directional electrical meter installed by the utility company, National Grid. Bills from National Grid were used to determine the plant s electricity usage. Methane (CH 4), carbon dioxide (CO 2), hydrogen sulfide (H 2S), and oxygen (O 2) concentration in biogas, were measured on-site during monthly visits using a Multitec 540 (Sewerin GmbH, Germany), a portable hand-held gas measuring device equipped with infra-red/electrochemical sensors. Measuring ranges and precision of the Multitec 540 are presented on Table 7. PRO-DAIRY Program - Cornell University

24 Table 7. Ranges and precision of the portable, on-site gas measuring device used for monitoring (Multitec 540). Gas component Measuring range Resolution Sensor Methane % (v/v) 0.1 % (v/v) Infra-red Carbon dioxide % (v/v) 1 % (v/v) Infra-red Hydrogen Sulfide 0 2,000 ppm 1 ppm Electrochemical Oxygen % (v/v) 0.1 % (v/v) Electrochemical A list of the equipment used and corresponding specifications is provided in Table 8. Table 8. Specifications of equipment used in performance evaluation data collection. Equipment Parameter measured Details Instantaneous power output (kw) GE Jenbacher engine-generator Cumulative energy produced (kwh) set control panel Engine run time (hrs) Model DIA.NEXT Esters gas flow meter (Digester) Volume of biogas (m 3 ) from digester Model 016GD150 Temperature and pressure compensated Esters gas flow meter (CHP) Volume of biogas (m 3 ) to CHP Model 016GD150 Temperature and pressure compensated Siemens magnetic mass flow transmitters (food waste) Volume of biomass (m 3 ) from food waste tank Model SITRANS F M MAG 3100 Siemens magnetic mass flow transmitters (manure) Siemens magnetic mass flow transmitters (effluent) Multitec 540 Volume of biomass (m 3 ) from manure tank tank Volume of effluent (m 3 ) out of digester Methane, carbon dioxide, hydrogen sulfide, and oxygen content in biogas Model SITRANS F M MAG 3100 Model SITRANS F M MAG 3100 See Table 7 PRO-DAIRY Program - Cornell University

25 3. Results In this section, the results of all analyses performed in accordance with the protocol are presented, as well as some analyses performed that were not included in the protocol. Where data were not available or could not be determined, the abbreviation ND appears in the tables. A summary of the most relevant data collected is presented in Table 9. Also, the daily values for the most important operational parameters presented in Table 9 were plotted in Figure 4. The plot includes the digester s influent loading rate, food-waste-to-influent co-digestion, biogas production, and CHP instant power for the entire monitoring project. The objective of this figure is to show the variability of the AD system during this period, and to serve as support information to explain the results discussed in this report. PRO-DAIRY Program - Cornell University

26 Table 9. Monitoring summary data. Month Jul-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 Period Biomass Influent biomass (gal/period) 2,439,589 1,462,096 1,071,908 1,903,617 1,367,203 2,133,814 1,866,645 Influent food waste (gal/period) 394, , , , , , ,145 Influent raw manure (gal/period) 2,044,783 1,276, ,069 1,572,684 1,081,394 1,539,551 1,246,500 Co-digestion ratio - food waste to influent (gal/gal) Effluent (gal/period) 2,180,828 1,335, ,363 1,702,547 1,232,854 1,844,739 1,615,090 Temperature Average digester temperature (F) Biogas Biogas produced by digester (ft 3 /period) 10,630,133 5,694,951 4,233,761 9,138,592 6,710,980 9,570,202 8,007,873 Average biogas production per day (ft 3 /day) 303, , , , , , ,588 CHP Cumulative operating time (h) 2,264 2,700 3,011 3,555 3,709 4,449 5,047 Energy generated (MWh/period) Average daily energy generated (MWh/day) Average daily instant power (MW) Plant capacity factor (decimal) Dairy Farm Average lactating cows contributing to digester 1,970 2,017 2,006 1,978 1,905 1,931 1,944 PRO-DAIRY Program - Cornell University

27 Table 9 (continued). Monitoring summary data. Month Jan-13 Feb-13 Mar-13 Apr-13 May-13 Jun-13 Jul-13 Period Biomass Influent biomass (gal/period) 1,789,315 1,515,053 1,796,384 2,160,602 2,346,075 3,118,507 2,668,168 Influent food waste (gal/period) 374, , , , , , ,984 Influent raw manure (gal/period) 1,415,308 1,060,511 1,402,201 1,499,092 1,732,439 2,247,895 2,060,184 Co-digestion ratio - food waste to influent (gal/gal) Effluent (gal/period) 1,569,158 1,300,327 1,607,578 1,813,075 1,993,243 2,332,326 1,836,774 Temperature Average digester temperature (F) Biogas Biogas produced by digester (ft 3 /period) 8,868,460 7,976,473 8,612,820 10,776,758 10,844,936 14,330,681 10,225,707 Average biogas production per day (ft 3 /day) 253, , , , , , ,296 CHP Cumulative operating time (h) 5,262 5,609 6,114 6,803 7,507 8,324 8,935 Energy generated (MWh/period) , Average daily energy generated (MWh/day) Average daily instant power (MW) Plant capacity factor (decimal) Dairy Farm Average lactating cows contributing to digester 1,934 1,912 1,900 1,887 1,896 1,879 1,879 PRO-DAIRY Program - Cornell University

28 Table 9 (continued). Monitoring summary data. Month Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Jan-14 Feb-14 Period Biomass Influent biomass (gal/period) 3,152,645 3,517,967 3,237,216 3,135,991 3,334,404 2,444,506 3,237,703 Influent food waste (gal/period) 816, , , , , ,950 1,015,859 Influent raw manure (gal/period) 2,335,652 2,659,532 2,241,692 2,241,832 2,416,116 1,852,556 2,221,844 Co-digestion ratio - food waste to influent (gal/gal) Effluent (gal/period) 2,217,268 2,408,716 2,166,136 2,109,616 2,600,420 1,967,380 2,309,644 Temperature Average digester temperature (F) Biogas Biogas produced by digester (ft 3 /period) 11,872,661 11,687,947 13,013,467 11,915,127 11,492,657 7,093,028 12,819,925 Average biogas production per day (ft 3 /day) 395, , , , , , ,484 CHP Cumulative operating time (h) 9,629 10,299 11,033 11,692 12,402 12,980 13,850 Energy generated (MWh/period) , Average daily energy generated (MWh/day) Average daily instant power (MW) Plant capacity factor (decimal) Dairy Farm Average lactating cows contributing to digester 1,885 1,866 1,859 1,846 1,825 1,829 1,834 PRO-DAIRY Program - Cornell University

29 Table 9 (continued). Monitoring summary data. Month Mar-14 Apr-14 May-14 Average Value Period (Total Value) Biomass Influent biomass (gal/period) 1,613,576 2,895,836 2,847,064 2,377,329 Influent food waste (gal/period) 489, , , ,950 Influent raw manure (gal/period) 1,124,300 1,997,140 2,028,132 1,759,379 Co-digestion ratio - food waste to influent (gal/gal) Effluent (gal/period) 1,189,520 2,176,148 2,027,852 1,852,882 Temperature Average digester temperature (F) Biogas Biogas produced by digester (ft 3 /period) 5,787,585 8,283,610 12,570,603 9,673,296 Average biogas production per day (ft 3 /day) 263, , , ,709 CHP Cumulative operating time (h) 14,116 14,595 15,269 (15,269) Energy generated (MWh/period) Average daily energy generated (MWh/day) Average daily instant power (MW) Plant capacity factor (decimal) Dairy Farm Average lactating cows contributing to digester 1,825 1,806 1,783 1,891 PRO-DAIRY Program - Cornell University

30 Digester loading rate (gal/day x 1,000) May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May Food waste-to-influent co-digestion ratio (decimal) May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May Biogas production rate (ft 3 /min) May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May 1.6 Average daily CHP instant power (MW) Average daily CHP instant power (MW) CHP rated capacity (MW) May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May 10-Jun 10-Jul 10-Aug 10-Sep 10-Oct 10-Nov 10-Dec 10-Jan 10-Feb 10-Mar 10-Apr 10-May Figure 4. Daily loading rate, co-digestion ratio (food waste to influent), biogas production, and CHP instant power at the Synergy AD system during the two years of monitoring (May 2012 May 2014). PRO-DAIRY Program - Cornell University

31 3.1. Digester Influent: Dairy Cattle Manure The daily number of animals contributing manure (urine and feces) to the AD system was obtained from the farm monthly. The daily population values over each period were averaged to calculate a period average for the lactating cows, which was the only group contributing manure to the AD, as detailed on Table 10. The actual amount of manure pumped from the receiving tank to the digester is also included for comparison. Table 10. Monthly animal management group average populations and calculated manure produced. Month and year Period Average number of lactating cows Theoretical amount of manure produced (gal/day) 1 Measured volume of manure pumped to digester (gal/day) Calculated Differential (gal/day) July, ,970 35,170 58,422 23,252 July, ,017 36,021 67,194 31,173 August, ,006 35,821 57,942 22,120 September, ,978 35,327 49,146 13,819 October, ,905 34,017 38,621 4,604 November, ,931 34,474 48,111 13,637 December, ,944 34,708 46,167 11,458 January, ,934 34,538 40,437 5,900 February, ,912 34,147 36,569 2,423 March, ,900 33,921 51,933 18,012 April, ,887 33,700 46,847 13,146 May, ,896 33,854 57,748 23,894 June, ,879 33,548 60,754 27,206 July, ,879 33,562 79,238 45,676 August, ,885 33,652 77,855 44,203 September, ,866 33,330 85,791 52,461 October, ,859 33,191 72,313 39,121 November, ,846 32,956 77,305 44,349 December, ,825 32,587 71,062 38,476 January, ,829 32,657 71,252 38,595 February, ,834 32,744 60,050 27,306 March, ,825 32,582 51,105 18,523 April, ,806 32,250 62,411 30,161 May, ,783 31,837 67,604 35,767 Average 1,891 33,775 59,828 26,053 Maximum 2,017 36,021 85,791 52,461 Minimum 1,783 31,837 36,569 2,423 Standard deviation 62 1,109 13,733 14,183 n Based on 150 lbs of manure/day-cow (ASABE, 2006) PRO-DAIRY Program - Cornell University

32 The volume of manure produced was calculated from the average number of lactating cows contributing to the digester per period, the standard value of 150 lb manure/day-cow for an average sized lactating cow (ASABE, 2006), and the conversion factor of 8.40 lb manure/gal manure. The average daily volume of manure produced for the entire monitoring project was predicted to be 33,775 gal/day. The measured amount pumped from the manure receiving tank was on average 59,828 gal/day, which was almost double the predicted amount. There are many possible reasons to explain such significant discrepancy. First, the biomass meter was factory calibrated, but an accuracy check, although intended, could not be performed during the monitoring because of the disruption this could have caused to the operation of the plant; second, the manure receiving tank also received milking center wastewater; third, cows were cooled in the summer with in-barn sprinkler systems that can contribute meaningful amounts of excess water to barn effluent; and fourth Synergy biogas plant operators reported that during the summer months there were many times when it was necessary to add additional water to dilute the manure when it was too thick to pump and/or to help regulate digester temperature Digester Influent: Co-substrates A basic description of the composition of each waste received by the Synergy plant is shown on Table 11. Synergy maintained detailed logs of imported wastes delivered to the plant. The monthly and total amounts of food wastes and other co-substrates imported by Synergy during the entire monitoring project are detailed on Table 12. The volume of food waste received by Synergy consistently increased since March 2013, reaching its peak in October A decrease in the volumes of food waste received in November and December 2013 was expected due to the closing of some industries and decrease in business activity during the holidays. Overall, 16,188,704 gallons of substrate for co-digestion were imported to the site. If all delivery trucks were 8,000 gallons each, the number of loads delivered was over 2,000 for two years time. The percent contributions of each waste in terms of industry and material characteristics, is shown in Figure 5. The largest portion of the imported wastes corresponded to dairy waste, mainly derived from yogurt production process. These wastes consist of easily digestible sugars and proteins but are usually highly diluted. Also, both dense and diluted fats, oils, and greases (FOG), mostly coming from the prepared-foods industry, constituted a significant portion of the imported waste added to the digester. These type of wastes are high energy yielding; however, they represent a higher risk for the operation as substrate over loading and process upset are more likely. PRO-DAIRY Program - Cornell University

33 Table 11. Basic description and classification of wastes imported by Synergy since the start of the monitoring. Co-substrate Description Classification Industry Classification Waste type Source 1 Acid whey Dairy Sugars, proteins Source 2 Grease trap cleanings Prepared foods Diluted FOG Source 3 Rendering solids Prepared foods Diluted FOG Source 4 Processed foods Prepared foods Dense FOG Source 5 Wine pulp Alcohol Sugars, alcohol Source 6 Hog processing waste - DAF Animal processing Protein, diluted FOG Source 7 Cream waste (frostings, cool whip) and wash water from line cleaning Prepared foods Dense FOG Source 8 Tomato paste Prepared foods Sugars, starches Source 9 Grease trap cleanings Prepared foods Diluted FOG Source 10 Source 11 Yogurt waste/whey and wash water from line cleaning Dairy waste and wash water from line cleaning Dairy Dairy Sugars, proteins Sugars, proteins Source 12 Alcohol-mix drink waste Alcohol Sugars, alcohol Source 13 Vegetable waste Restaurants/grocery stores Fibers, sugars, starches Source 14 Grease trap cleanings (cook oil, soup Restaurants/grocery stores Diluted FOG waste) PRO-DAIRY Program - Cornell University

34 Table 12. Monthly volume (gal) of co-substrates received per source during the entire monitoring project (June 2012 May 2014). Month Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12 Period Source 1 64,000 Source 2 17,300 21,100 6,900 16,180 7,800 13,000 19,600 Source 3 36,414 26,334 18,909 24,172 19,607 30,754 39,215 Source 4 Source 5 Source 6 85,000 77,375 79,055 62,672 81,569 61,489 52,711 Source 7 49,455 43,214 56,761 98,373 40,350 58,715 27,872 Source 8 5,730 15,600 5,800 6,200 11,380 Source 9 2,000 6,000 2,000 10,000 2,000 Source 10 Source 11 Source 12 Source 13 Source 14 52,000 70,000 65,500 52,000 46,000 54,000 53,500 Total 240, , , , , , ,278 PRO-DAIRY Program - Cornell University

35 Table 12 (continued). Monthly volume (gal) of co-substrates received during the entire monitoring project (June 2012 May 2014). Month Jan-13 Feb-13 Mar-13 Apr-13 May-13 Jun-13 Period Source 1 145, , , , , ,770 Source 2 16,200 2,400 9,800 26,300 17,600 17,600 Source 3 19,194 11,337 31,061 29,438 36,266 31,917 Source 4 Source 5 Source 6 47,483 53,618 23,577 Source 7 91,179 65, ,532 72, , ,944 Source 8 4,500 13,300 11,000 4,500 3,000 3,300 Source 9 10,000 4,000 6,000 6,000 Source 10 Source , , , ,700 Source 12 Source 13 Source 14 46,500 52,500 50,000 63,500 50,500 50,500 Total 370, , , , , ,731 PRO-DAIRY Program - Cornell University

36 Table 12 (continued). Monthly volume (gal) of co-substrates received during the entire monitoring project (June 2012 May 2014). Month Jul-13 Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Period Source 1 256, , , , , ,000 Source 2 15,800 15,900 21,000 19,100 12,900 17,100 Source 3 36,534 33,077 33,173 55,860 28,932 24,914 Source 4 3,245 Source 5 48,000 8,000 Source 6 Source 7 95,782 93, , , ,626 35,618 Source 8 3,300 19,700 6,550 4,400 11,100 3,200 Source 9 2,000 2,000 4,000 Source 10 97, , , , ,000 Source , , , , , ,200 Source 12 1,100 1,000 4,800 3,258 2,800 Source 13 Source 14 43,500 58,000 35,000 47,000 36,500 42,500 Total 802, , ,922 1,139, , ,332 PRO-DAIRY Program - Cornell University

37 Table 12 (continued). Monthly volume (gal) of co-substrates received during the entire monitoring project (June 2012 May 2014). Month Jan-14 Feb-14 Mar-14 Apr-14 May-14 Period Total Source 1 263, , , , ,690 3,954,980 Source 2 19,600 11,900 23,200 16,200 16, ,980 Source 3 29,820 12,297 31,078 61,205 46, ,456 Source 4 20,075 19,525 42,845 Source 5 8,000 64,000 Source 6 624,549 Source 7 72,221 69,594 99,511 97,744 72,671 1,918,556 Source 8 4,469 3, ,582 Source 9 2,000 2,000 7,500 15,100 82,600 Source , , , , ,500 1,688,800 Source , , , , ,000 4,417,280 Source ,700 15,158 Source Source 14 35,500 39,000 33,000 46,000 39,000 1,161,500 Total 889, , , ,643 1,011,487 15,240,952 PRO-DAIRY Program - Cornell University

38 By industry sector Prepared foods 21.7% Restaurants/grocery stores 7.6% Animal processing 4.1% Dairy 66.0% Alcohol 0.5% By waste characteristics Diluted FOG 15.6% Dense FOG 12.9% Protein, diluted FOG 4.1% Sugars, starches 0.9% Sugars, proteins 66.0% Sugars, alcohol 0.5% Figure 5. Percent contribution (volume basis) of each imported waste to the Synergy digester based on industry sector and waste characteristics. PRO-DAIRY Program - Cornell University

39 3.3. Physicochemical characterization and nutrients analysis The laboratory results for the analysis of the individual feedstocks, i.e., manure and food waste, in addition to the mixed influent and effluent of the digester are presented in Table 13 Table 16. Table 17 presents a summary of the characteristics of the digester influent and effluent for Periods 1 to 20 of the monitoring (i.e., sampling date 1/14/2014); laboratory analyses were terminated for the last few months of the project due to budget limitations. No laboratory analyses were conducted for the last four periods due to funds budgeted for this work being fully expended; the monitoring project was extended 12 months beyond its original ending date. In terms of nutrients, total nitrogen and potassium had the highest concentrations. Nitrogen was fairly equally contributed by imported waste and raw manure; however, potassium came primarily from raw manure. The relative concentration of total phosphorous was low in both food waste and raw manure. PRO-DAIRY Program - Cornell University

40 Table 13. Digester influent mixed characterization (average of triplicates). Sampling date Jun-12 Jul-12 Jul-12 Aug-12 Sep-12 Dec-12 Jan-13 Feb-13 Mar-13 Apr-13 May-13 Jun-13 Period* Chemical parameters Alkalinity (mg/kg as CaCO3) ND 9,333 7,733 6,533 7,333 6,667 8,167 10,583 4,500 4,467 5,667 4,600 Volatile fatty acids (as acetic acid) (mg/kg) 4,444 4,183 4,103 3,976 5,165 4,250 4,167 3,408 3,742 2,170 2,319 2,964 COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) Nutrient characterization Ammonia-N (mg/kg) 1,727 1,293 1,385 1,049 1,471 1,794 1,498 1, Total Kjeldahl nitrogen (mg/kg) 3,937 3,673 3,112 2,641 3,682 3,106 3,421 3,620 1,874 2,600 2,917 2,613 Total organic nitrogen (mg/kg) 2,210 2,380 1,727 1,593 2,211 1,312 1,923 2,186 1,101 1,946 2,200 1,887 Total phosphorus (mg/kg) Ortho phosphorus (mg/kg) ND ND ND Total potassium (mg/kg) 1,766 1,369 2,543 2,475 3,583 1,677 1,730 1, ,192 1,578 1,235 *No data for periods 5 and 6 due to samples lost by external laboratory PRO-DAIRY Program - Cornell University

41 Table 13 (continued). Digester influent mixed characterization (average of triplicates). Sampling date Jul-13 Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Jan-14 Period Chemical parameters Alkalinity (mg/kg as CaCO3) 5,067 6,400 5,867 6,533 6,000 7,067 5,733 Volatile fatty acids (as acetic acid) 3,165 2,939 3,019 3,170 3,467 2,602 2,602 (mg/kg) COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) Nutrient characterization Ammonia-N (mg/kg) , Total Kjeldahl nitrogen (mg/kg) 2,103 2,330 2,652 2,631 3,952 3,211 2,966 Total organic nitrogen (mg/kg) 1,504 1,809 2,098 1,723 2,586 2,525 2,522 Total phosphorus (mg/kg) Ortho phosphorus (mg/kg) NA NA NA NA NA NA NA Total potassium (mg/kg) 1,203 1,306 1,713 1,402 1,251 1,302 1,431 PRO-DAIRY Program - Cornell University

42 Table 14. Digester effluent characterization (average of triplicates). Sampling date Jun-12 Jul-12 Jul-12 Aug-12 Sep-12 Dec-12 Jan-13 Feb-13 Mar-13 Apr-13 May-13 Jun-13 Period* Chemical parameters Alkalinity (mg/kg as CaCO3) ND 13,133 11,733 12,133 12,267 10,400 14,433 10,400 13,067 10,000 10,133 9,067 Volatile fatty acids (as acetic acid) (mg/kg) 1, , ,300 1, COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) Nutrient characterization Ammonia-N (mg/kg) 1,990 1,810 1,617 1,703 1,694 2,049 2,254 1,267 1,748 1,464 1,619 1,262 Total Kjeldahl nitrogen (mg/kg) 4,243 3,707 3,137 3,190 3,183 2,479 3,455 3,314 3,205 3,246 3,218 2,924 Total organic nitrogen (mg/kg) 2,253 1,897 1,520 1,487 1, ,200 2,047 1,457 1,781 1,599 1,662 Total phosphorus (mg/kg) Ortho phosphorus (mg/kg) ND ND ND Total potassium (mg/kg) 1,738 1,405 3,157 3,094 3,309 1,699 2,013 1,595 1,676 1,405 1,183 1,215 *No data for periods 5 and 6 due to samples lost by external laboratory PRO-DAIRY Program - Cornell University

43 Table 14 (continued). Digester effluent characterization (average of triplicates). Sampling date Jul-13 Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Jan-14 Period Chemical parameters Alkalinity (mg/kg as CaCO3) 10,533 9,200 9,600 10,133 10,533 9,733 10,267 Volatile fatty acids (as acetic acid) (mg/kg) 1, , , COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) Nutrient characterization Ammonia-N (mg/kg) 1,257 1, ,747 1,138 1,530 1,430 Total Kjeldahl nitrogen (mg/kg) 2,292 2,010 2,575 3,959 3,490 3,852 3,478 Total organic nitrogen (mg/kg) 1, ,779 2,212 2,352 2,322 2,048 Total phosphorus (mg/kg) , Ortho phosphorus (mg/kg) NA NA NA NA NA NA NA Total potassium (mg/kg) 1,378 1,415 1,539 1,309 1,176 1,455 1,242 PRO-DAIRY Program - Cornell University

44 Table 15. Raw manure characterization (average of triplicates). Sampling date Jul-12 Sep-12 Period 2 4 Chemical parameters Alkalinity (mg/kg as CaCO 3) 9,867 11,867 Volatile fatty acids (as acetic acid) (mg/kg) 3,289 3,879 COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) Nutrient characterization Ammonia-N (mg/kg) 1,505 1,593 Total Kjeldahl nitrogen (mg/kg) 2,739 3,840 Total organic nitrogen (mg/kg) 1,235 2,247 Total phosphorus (mg/kg) Ortho phosphorus (mg/kg) ND ND Total potassium (mg/kg) 3,256 4,210 PRO-DAIRY Program - Cornell University

45 Table 16. Imported waste characterization (average of triplicates). Sampling date Jul-12 Sep-12 Period 2 4 Chemical parameters Alkalinity (mg/kg as CaCO 3) UR UR Volatile fatty acids (as acetic acid) (mg/kg) 5,033 8,139 COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) Nutrient characterization Ammonia-N (mg/kg) 780 1,182 Total Kjeldahl nitrogen (mg/kg) 3,307 4,582 Total organic nitrogen (mg/kg) 2,527 3,400 Total phosphorus (mg/kg) Ortho phosphorus (mg/kg) ND ND Total potassium (mg/kg) UR: Under range PRO-DAIRY Program - Cornell University

46 Table 17. Overall statistics of the digester influent mixed and digester effluent characteristics over the entire monitoring project. Influent mixed Average Maximum Minimum Standard deviation Effluent n Average Maximum Minimum Standard deviation Chemical parameters Alkalinity (mg/kg as CaCO3) 6,569 10,583 4,467 1, ,931 14,433 9,067 1, Volatile fatty acids (as acetic acid) (mg/kg) 3,466 5,165 2, , COD (g/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) n Nutrient characterization Ammonia-N (mg/kg) 1,032 1, ,550 2, Total Kjeldahl nitrogen (mg/kg) 3,002 3,952 1, ,208 4,243 2, Total organic nitrogen (mg/kg) 1,971 2,586 1, ,659 2, Total phosphorus (mg/kg) , Ortho phosphorus (mg/kg) Total potassium (mg/kg) 1,647 3, ,737 3,309 1, PRO-DAIRY Program - Cornell University

47 Using the digester influent and effluent volumes (Table 9) and the nutrients concentration data (Table 13 to Table 16), an estimation of the cumulative mass and average daily mass of nutrients for each period was made for the digester influent and effluent by multiplying the nutrient concentrations by the respective volume flow rates (Table 18 and Table 19, respectively). Further, Table 20 and Table 21 present an estimate of the cumulative mass and daily mass of nutrients contributed by the raw manure and food waste streams, respectively. As expected, most of the estimated total nutrient mass in the digester influent came from raw manure as imported waste only constituted between 13 to 33% of the influent mix on a volume/volume basis. Also, with exception of Period 1, the average estimated daily mass of nutrients was fairly constant in the influent, effluent, and raw manure. After Period 4 and until Period 16, the N/P/K ratio in the effluent changed from 5/1/5 to a fairly constant 6-7/1/3. For the last four sampling dates, after Period 16, the nutrient ratio changed to about 6/1/2. The values above need to be taken with caution. First, because the accuracy of the digester effluent biomass meter was unable to be verified during the monitoring period, and the large discrepancy between the measured and predicted values; also, because digestate samples for laboratory analyses were collected only once a month, thus were not likely to represent adequately the physicochemical makeup of the digestate over the roughly 25 days of digestion. As for the first factor, our group is developing a scientific paper showcasing theoretical methods to estimate biomass volume reduction due to digestion, which should help to determine the accuracy of the meter. The published paper will be posted on the Dairy Environmental Systems website. PRO-DAIRY Program - Cornell University

48 Table 18. Estimated mass and ratio of nutrients in the digester influent by monitoring period. Sampling date Jul-12 Jul-12 Aug-12 Sep-12 Dec-12 Jan-13 Feb-13 Mar-13 Apr-13 May-13 Jun-13 Period* Total Estimated Mass for Period Total Nitrogen (lb) 74,884 38,026 23,659 58,570 48,443 51,151 45,834 28,136 46,936 57,180 68,084 Total Phosphorous (lb) 8,114 5,278 3,547 7,487 6,031 7,845 7,318 5,559 7,812 10,175 15,036 Potassium (lb) 27,902 31,070 22,172 56,995 26,163 25,862 21,548 12,629 21,515 30,936 32,174 Average Estimated Daily Mass for Period Total Nitrogen (lb/d) 2,140 2,001 1,479 1,830 1,794 1,461 1,580 1,042 1,467 1,906 1,840 Total Phosphorous (lb/d) Potassium (lb/d) 797 1,635 1,386 1, , Ratio Total Nitrogen Total Phosphorous Total Potassium *No data for periods 5 and 6 due to lost samples by laboratory PRO-DAIRY Program - Cornell University

49 Table 18 (continued). Estimated mass and ratio of nutrients in the digester influent by monitoring period. Sampling date Jul-13 Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Jan-14 Period Total Estimated Mass for Period Total Nitrogen (lb) 46,881 61,391 77,961 71, ,556 89,464 60,583 Total Phosphorous (lb) 8,197 9,897 13,719 11,857 15,715 13,123 7,864 Potassium (lb) 26,822 34,415 50,367 37,916 32,791 36,287 29,231 Average Daily Estimated Mass for Period Total Nitrogen (lb/d) 1,803 2,046 2,515 2,296 3,571 2,631 2,330 Total Phosphorous (lb/d) Potassium (lb/d) 1,032 1,147 1,625 1,223 1,131 1,067 1,124 Ratio Total Nitrogen Total Phosphorous Total Potassium PRO-DAIRY Program - Cornell University

50 Table 19. Estimated mass and ratio of nutrients in the digester effluent by monitoring period. Sampling date Jul-12 Jul-12 Aug-12 Sep-12 Dec-12 Jan-13 Feb-13 Mar-13 Apr-13 May-13 Jun-13 Period* Total Estimated Mass for Period Total Nitrogen (lb) 67,549 35,010 24,854 45,289 33,452 45,299 36,010 43,054 49,174 53,599 56,994 Total Phosphorous (lb) 7,557 4,561 2,670 5,890 1,066 3,191 4,571 3,658 4,515 3,659 5,613 Potassium (lb) 25,604 35,234 24,106 47,077 22,925 26,400 17,328 22,514 21,292 19,699 23,686 Average Estimated Daily Mass for Period Total Nitrogen (lb/d) 1,930 1,843 1,553 1,415 1,239 1,294 1,242 1,595 1,537 1,787 1,540 Total Phosphorous (lb/d) Potassium (lb/d) 732 1,854 1,507 1, Ratio Total Nitrogen Total Phosphorous Total Potassium *No data for periods 5 and 6 due to lost samples by laboratory PRO-DAIRY Program - Cornell University

51 Table 19 (continued). Estimated mass and ratio of nutrients in the digester effluent by monitoring period. Sampling date Jul-13 Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Jan-14 Period Total Estimated Mass for Period Total Nitrogen (lb) 35,184 37,235 51,836 71,656 61,524 83,707 57,177 Total Phosphorous (lb) 5,229 5,169 7,830 18,683 11,255 13,637 8,648 Potassium (lb) 21,150 26,217 30,984 23,694 20,737 31,617 20,413 Average Estimated Daily Mass for Period Total Nitrogen (lb/d) 1,353 1,241 1,672 2,311 2,122 2,462 2,199 Total Phosphorous (lb/d) Potassium (lb/d) Ratio Total Nitrogen Total Phosphorous Total Potassium PRO-DAIRY Program - Cornell University

52 Table 20. Estimated mass and ratio of nutrients contributed by raw manure. Sampling date Jul-12 Sep-12 Period 2 4 Total Estimated Mass for Period Total Nitrogen (lb) 29,224 50,467 Total Phosphorous (lb) 4,310 6,116 Potassium (lb) 34,732 55,325 Average Estimated Daily Mass for Period Total Nitrogen (lb/d) 1,538 1,577 Total Phosphorous (lb/d) Potassium (lb/d) 1,828 1,729 Ratio Total Nitrogen 4 5 Total Phosphorous 1 1 Total Potassium 5 5 Table 21. Estimated mass and ratio of nutrients contributed by imported waste. Sampling date Jul-12 Sep-12 Period 2 4 Total Estimated Mass for Period Total Nitrogen (lb) 5,124 12,677 Total Phosphorous (lb) Total Potassium (lb) 590 1,041 Average Estimated Daily Mass for Period Total Nitrogen (lb/d) Total Phosphorous (lb/d) Potassium (lb/d) Ratio Total Nitrogen 8 9 Total Phosphorous 1 1 Total Potassium 1 1 PRO-DAIRY Program - Cornell University

53 3.4. Effect of Anaerobic Digestion on Constituents: Waste Stabilization In the protocol, waste stabilization is considered to represent the difference, or percent change, of selected constituents/parameters between the influent and effluent of the digester. The chemical constituent concentration parameters determined for the digester influent and effluent presented in Table 13 and Table 14, respectively, were used in Equation 1 to calculate the average percent change within the AD system for each period. Equation 1: AAAAAAAAAAAAAA pppppppppppppp cchaaaaaaaa [aaaaaaaaaaaaaa iiiiiiiiiiiiiiii cccccccccccccccccccccccccc] [aaaaaaaaaaaaaa eeeeeeeeeeeeeeee cccccccccccccccccccccccccc] = [aaaaaaaaaaaaaa iiiiiiiiiiiiiiii cccccccccccccccccccccccccc] 100 [ ] = Concentration The calculated results for each sampling date of the project are shown in Figure 6. A positive change represents a decrease of the parameter s concentration in the digester effluent with respect to the influent, while a negatives change represents an increase of the parameter s concentration in the digester effluent with respect to the influent. Overall, a slight decrease in the organic material (i.e., volatile fatty acids (VFA), volatile solids (VS) and chemical oxygen demand (COD)) destruction due to the digestion process was observed up to March However, from April 2013 and on, organic matter destruction was back to previously observed values, i.e., >50% destruction of influent VS, COD. The high VS destruction calculated from the July 2013 sampling (>86%) was the result of lower than normal solid concentration in the effluent (<0.5% VS) which could be related to the type of co-substrates and/or dilution of the raw manure influent stream. This is likely given the lower co-digestion ratio in July, the increase in digestate alkalinity, and the low influent COD observed (i.e., 70 g/l vs. 100 g/l in average). Because conversion of organic matter into biogas produces alkalinity, alkalinity increases, and so does the ph. The high increment of alkalinity shown in March 2013 was due to the significantly lower influent alkalinity compared to previous PRO-DAIRY Program - Cornell University

54 sampling dates. This suggests that the food waste co-digested at the time of sampling may have been whey, which was highly acidic (ph ~ 3.5) and does not contribute alkalinity. Acid whey has a relatively low solids content compared to other materials co-digested, which explains why the calculation of VS destruction was positive in March (suggesting that VS increased in the digester). It is important to consider that sampling only represented the influent characteristics at the moment, and thus it was just a snapshot of the entire digestion process, which likely varies considerably throughout the approximately 30 days of treatment. The overall waste stabilization results for the entire monitoring project in terms of the total influent and effluent masses are shown in Figure 7. Overall, 57% of the influent organic matter mass in terms of VS was stabilized. This compares favorably with respect to a previous monitoring study of five AD systems, where the authors found that manure-based co-digestion operations achieved between 36% and 53% VS destruction (Gooch et al., 2011). Additionally, at the Synergy digester, 54% of the influent COD mass was destroyed, and only 20% of the influent VFA mass was found in the effluent (Figure 7). The percent contribution of volatile solids (VS) and fixed solids (FS) in the AD influent and effluent streams is depicted in Figure 8. The graphs show the proportion of VS and FS within the total solids (TS) portion of the influent, i.e., TS = VS + FS. With the exception of March and July, the percent contribution of VS in the influent was about 80%, with the remaining 20% being non-digestible inert solids. In the effluent, the VS concentration decreases to less than 60%, which is expected, as the biodegradable organic matter contained in the VS was stabilized through the AD process. The 20% decrease in VS with respect to the influent was higher than that seen in other co-digestion operations, i.e., 9 12% (Gooch et al., 2011). PRO-DAIRY Program - Cornell University

55 Alkalinity (mg/kg as CaCO3) Volatile fatty acids (as HAc) (mg/kg) COD (mg/kg) ph Total solids (g/kg) Volatile solids (g/kg) Fixed solids (g/kg) 100% Change in constituent concentration (%) 40% -20% -80% -140% -200% Sampling date Figure 6. Percent change in chemical parameters concentrations in the digester effluent relative to the influent for each sampling date; a positive change represents a decrease, while a negative change represents an increase. PRO-DAIRY Program - Cornell University

56 Constituent mass (x1000 lbs.) 2,500 2,000 1,500 1, Total lbs influent Total lbs effluent 0 Alkalinity (as CaCO3) Volatile fatty acids (as Acetic acid) Chemical oxygen demand, COD Total solids, TS Volatile solids, VS Fixed solids, FS Change in constituent mass (%) 120% 80% 40% 0% -40% -80% Average project Alkalinity (as CaCO3) Volatile fatty acids (as Acetic acid) Chemical oxygen demand, COD Total solids, TS Volatile solids, VS Fixed solids, FS Figure 7. Top: average mass of constituent in the digester influent and effluent for the entire monitoring project; Bottom: average percent change in constituent mass in the digester effluent relative to the influent for the entire monitoring project; error bars represent the standard deviation. PRO-DAIRY Program - Cornell University

57 100% Fixed solids, FS (g/kg) Volatile solids, VS (g/kg) Percent in AD inflluent 80% 60% 40% 20% 0% 100% Percent in AD effluent 80% 60% 40% 20% 0% Sampling date Figure 8. Percent contribution of total volatile solids, VS (blue bar), and fixed solids, FS (red bar) in the AD influent (top) and effluent (bottom) for sampling dates; total solids represent 100% of the dry influent material (i.e., VS + FS). PRO-DAIRY Program - Cornell University

58 3.5. Biogas characterization and production Biogas composition: anaerobic digester Under normal anaerobic conditions, raw biogas as produced by the digester, is mainly composed of methane, carbon dioxide, and water vapor. Also, traces amounts of ammonia and hydrogen sulfide are usually present. No oxygen is expected, unless air or pure oxygen is purposely added to the digester s gas head space to scrub hydrogen sulfide. Using the gas analyzer described in Section 2.3, water vapor and ammonia cannot be measured; however, water vapor concentration, which depends on gas temperature, can be calculated using Equation 2 on a mass-by-volume basis, assuming that biogas was vapor saturated (Richards et al., 1991). In turn, the percent of water vapor in biogas can be calculated using the Ideal Gas Law, as shown in Equation 3, assuming that the pressure inside the digester vessel was equal to 1 atm. Results for each sampling date are presented in Table 22. Equation 2: CC HH2 OO(m) = ee ( (TT 32) 5 9 Where, CC HH2 OO(m) = Concentration of water vapor on a mass-by-volume basis (g/l) T = Biogas temperature ( F). Equation 3: CC HH2 OO(v) = CC HH2OO(m) MMMM(HH 2 OO) RRTT aaaaaa PP Where, CC HH2 OO(v) = Concentration of water vapor on a volume-by-volume basis (L/L) MMMM(HH 2 OO) = Molecular weight of water (18 g/mol) R = Universal gas constant ( L-atm/mol-K) T abs = Absolute temperature, K = (T ( o C) ) P = Absolute pressure (1 atm) PRO-DAIRY Program - Cornell University

59 As shown in Table 22, the average biogas moisture content was calculated to be 8.5% (v/v basis), assuming that biogas temperature is equal to the digestate temperature (actual biogas temperature was not measured). It was apparent that the moisture content increased during the warmer months accordingly with the increase in digester temperature (see Table 9). Table 22. Calculated water vapor density of biogas for each sampling date. Sampling date Water vapor density of biogas Mass basis, g/l Volume basis (%) June, July, August, September, October, November, December, January, February, March, April, May, June, July, August, September, October, November, December, January, February, March, April, May, Average The percent of methane and carbon dioxide measured in the biogas are shown in Figure 9. The measurements were taken at a sampling port located in the digester headspace. As expected, no oxygen was detected in the digester biogas throughout the monitoring project. A decrease in the PRO-DAIRY Program - Cornell University

60 methane content was observed from Periods 3-5 (August 2012 to October 2012). There are several reasons that may cause this, including an increase in the digestate alkalinity; however, it is apparent that a less stable organic loading rate and highly variable food waste-to-influent codigestion ratio observed in September, October and December could have upset the digester decreasing the methane content and biogas production (see Figure 4, pg. 26). With the exception of Period 13, a sustained increase in methane content (>60%) was observed after Period 11 (April 2013), which correlates with the increase in imported waste volume and co-digestion ratio increase. Overall, the methane and carbon dioxide content in dry biogas for the entire monitoring project was 60% and 37% on average (water vapor concentration is not taken into account). Most of the remaining gas was likely constituted by ammonia, as hydrogen sulfide was less than 1%, as discussed below. 100% CO2 (%) CH4 (%) CH4 and CO2 in biogas (%) 80% 60% 40% 20% 0% 36% 35% 40% 37% 38% 37% 39% 38% 38% 38% 36% 36% 38% 36% 35% 35% 35% 35% 36% 37% 35% 38% 37% 39% 61% 62% 57% 57% 58% 61% 57% 59% 58% 59% 61% 61% 59% 60% 63% 62% 60% 64% 61% 60% 61% 59% 61% 59% Sampling date Figure 9. Methane and carbon dioxide content in biogas measured at each sampling date from the digester vessel headspace on a dry biogas basis (moisture corrected) Biogas composition: biological scrubber The efficiency of the biological hydrogen sulfide (H 2S) scrubber is shown in Figure 10. No influent data are available for Period 3 (August). With the exception of Period 5 (October), the scrubber s efficiency was found to be between 80 and 100% over the monitoring project. For the sampling on Period 5, the flow of biogas to the bio-scrubber had been on and off due to the mechanical problems mentioned above. An intermittent flow of biogas will starve and overload the sulfide-degrading PRO-DAIRY Program - Cornell University

61 bacteria in the bio-scrubber, which will upset the process and decrease the efficiency of the system for several days after biogas production returns to relatively a steady state condition. From Periods 8 11 (January to April 2013), the efficiency of the H 2S scrubber steadily decreased from over 90% to around 60%. This decrease in efficiency can be attributed to similar mechanical problems as mentioned above; however, the main reason was the result of the normal accumulation of by-product material (e.g., sulfur) on the surface of the media which decreases biofilm activity and thus efficiency of microbial hydrogen sulfide oxidation. This was a problem observed by the plant operators that occurred consistently during the monitoring, and required opening the bio-scrubber for thorough clean up. In fact, after the biological scrubber was cleaned, H 2S removal efficiency returned to expected values (e.g., >80%) during Periods 12 and 13. However, efficiency during Periods decreased again, this time likely attributed to the inability to maintain optimal operational conditions for the scrubber, namely O 2 content and ph. By the end of the monitoring, the bio-scrubber was still not working as expected (i.e., H 2S concentrations < 100 ppm), showing very unstable performance since July 2013 until May 2014, and allowing high concentrations of H 2S ahead of the CHP unit. Operators reported that sulfur was again accumulating inside the bio-scrubber and needed through cleaning. Overall, for the entire monitoring project, H 2S concentration was reduced in average from 1,209 to 488 ppm, resulting in a bio-scrubber efficiency of 62±29%. 100% % Efficiency H2Sin H2Sout 2,100 H 2 S Removal efficiency (%) 80% 60% 40% 20% 0% 1,800 1,500 1, H 2 S Concentration (ppm) Sampling date Figure 10. Concentration of hydrogen sulfide before and after the biological scrubber and its corresponding efficiency at each sampling date. PRO-DAIRY Program - Cornell University

62 Biogas production: normalized by number of cows Figure 11 shows the average daily biogas production for each period as normalized by the number of lactating cows contributing to the digester, which are shown on top of the bars. Normalization was done by dividing the total biogas produced for each period by the average daily number of cows contributing manure to the anaerobic digester for that period. Overall, this was done to generally express how much biogas was produced when cow manure was co-digested with substrates. The decrease observed in Periods 3 5 (August thru October 2012) is explained by the lower biogas production and increased number of cows contributing to the digester during this period. The decreased observed in Period 8 (January 2013) is expected due to the low and unstable biogas production during the two previous periods. From Period 10 and on, biogas production improved due to the drastic increase in the digester loading rates, i.e., 20% with respect to the previous period (see Table 9, pg. 22), and particularly in the influent co-digestion ratio (i.e., percent of imported waste in the influent). From Periods 19 to 20, the digester loading rate and imported waste co-digestion decreased affecting biogas and energy production. However, the decrease in biogas production during Periods 22 and 23 was only apparent, as mechanical problems with the flare increased the gas pressure in the digester vessel forcing unintended gas release through pressure relief valve, and preventing the biogas to reach the gas meter and thus be accounted for. On average for the project, 173±34 ft 3 per lactating cow contributing to the digester. 300 Daily biogas production per cow (ft 3 /LC) ,880 1,916 1,911 1,889 1,809 1,843 1,854 1,845 1,813 1,792 1,887 1,896 1,879 1,879 1,885 1,866 1,859 1,846 1,825 1,829 1,834 1,825 1,806 1, Period Figure 11. Average daily biogas production per lactating cow (LC) contributing to the digester observed for each period; the number shown on top of the bars is the average lactating cows for the corresponding period. PRO-DAIRY Program - Cornell University

63 Biogas production: normalized by influent biomass The average biogas production rate normalized by the digester s influent biomass (i.e., specific biogas production) is shown in Figure 12. The figure shows that for each unit of biomass added to the digester (wet basis), between 0.4 and 0.6 ft 3 of biogas were produced. The observed differences in these values are primarily determined by the type of the material and the co-digestion ratio. The figure also suggests that the decrease in biogas production observed in Period 7 was probably due to the lower loading rate rather than a digester upset (see Figure 4, pg. 26). The high specific biogas production rate seen in Period 9 is likely explained by the higher co-digestion rate (i.e., 30%) of such month (see Table 9, pg. 22). An apparent decrease in the digester biomass conversion efficiency is observed after the increase in loading rates and co-digestion ratio in Period 10. This is expected, as the digester HRT decreased from an average 36 days from Periods 1 9, to an HRT of 25 from Periods The lower biogas production observed during Periods and Periods was caused by a decrease in digester loading rates and mechanical problems, respectively, as detailed in the previous section. Overall, the average biogas volume production per unit of mass of total influent added to the digester was 0.495±0.078 ft ft 3 per lb influent biomass Period Figure 12. Average specific biogas production as a function of the total influent biomass observed for each period; the number on top of the bars indicates the ft 3 per lb of biomass. PRO-DAIRY Program - Cornell University

64 Biogas energy The thermal value of biogas is based on its methane content. Since it is generally assumed that biogas is saturated with water vapor, to calculate the total energy produced by the digester, the volume of biogas needs to be converted to a dry basis. Specifically for this project, biogas had a calculated average water vapor content of 8.5% (v/v) for the entire monitoring period (see Table 22, pg. 55). The lower heating value (LHV) of methane suggested by the EPA protocol is 960 Btu/ft 3 (EPA, 2011); however, other sources indicate that the LHV of methane is 896 Btu/ft 3 (at 68F and 14.7 psia) (Marks, 1978). To be conservative, the latter value was used in the calculations used to develop values in this report. Thus, the total energy produced by the digester was determined by subtracting the calculated volume of water vapor from the measured volume of biogas produced by the digester, and multiplying this difference by the measured methane content and standard LHV value for each period. Figure 13 shows the total calculated energy produced by the digester for each period. For the entire monitoring project, the average energy produced by the digester was 4,766±1,340 MMBtu per period. 8,000 7,000 Biogas energy (MMBtu) 6,000 5,000 4,000 3,000 2,000 1, Period Figure 13. Total energy (MMBtu) produced by the anaerobic digester for each period. PRO-DAIRY Program - Cornell University

65 3.6. Biogas utilization The biogas produced was used to fuel a combined heat and power (CHP) system to produce electrical energy and recover thermal energy. The SCADA system continuously recorded the electrical and thermal energy produced by the system Electrical energy The average daily electrical energy generated by the CHP for each period is shown in Figure 14. Energy generation tended to steadily decrease after its initial peak in Period 2 (July 2012). However, a more drastic decrease (> 60%) is observed between Periods 4 and 5 (September and October 2012). This is explained by the decreased biogas production observed during this period and other mechanical (i.e., blower, engine) problems that occurred concurrently. The decrease in electricity generation observed in Period 8 (January 2013) was due to a 2-week downtime of the CHP system in the previous period, which was forced by an ignition rail component failure. The low energy output in the Period 9 is attributed to low and unstable biogas production (see Figure 4, pg. 26). As expected due to the increase in biogas production and quality (i.e., higher methane content) since Period 10, electricity generation increased to reach a somewhat stable production up to date. During Periods 22 and 23 (particularly March 2014), the AD plant experienced several mechanical problems, first with the CHP unit, which in three separate occasions broke down each time lasting several days (See Figure 4, pg. 26). The second major problem occurred with the flare, which malfunctioned and was down for several days and prevented the biogas to reach the CHP system (more details in Section 3.5.3). PRO-DAIRY Program - Cornell University

66 Average daily energy generated (MWh) Period Figure 14. Average daily electrical energy production by CHP observed for each period Biogas to electricity and thermal conversion efficiency The volume of biogas needed to generate one unit of electrical energy (kwh) was calculated using Equation 4. In addition to the CHP s thermal conversion efficiency, this value is dependent on the thermal energy of biogas (i.e., methane content) and its water vapor density at any given time. Equation 4 takes only into account the volume of biogas delivered to the engine (i.e., not necessarily the total volume of biogas produced by the digester), and thus, it was assumed that all the water vapor contained in biogas had already been removed by the biogas condensation system (see Section 1.1.2), and therefore no water vapor correction was performed. Equation 4: BBBBBBBBBBBB vvvvvvvvvvvv pppppp kkkkh gggggggggggggggggg = vvvvvvvvvvvv bbbbbbbbbbbb tttt eeeeeeeeeeee ( ffff3 dddddd ) eeeeeeeeeeeeeeeeeeee eeeeeeeeeeee pppppppppppppppp ( kkkkh dddddd ) The thermal conversion efficiency of the engine to produce electrical energy from the combustion of biogas was calculated using Equation 5 and the LHV value of 896 Btu/ft 3 (Marks, 1978). PRO-DAIRY Program - Cornell University

67 Equation 5: TTheeeeeeeeee cccccccccccccccccccc eeeeeeeeeeeeeeeeeeee (%) kkkkh gggggggggggggggggg pppppp dddddd 3,412 BBBBBB = kkkkh 100 vvvvvvvvvvvv bbbbbbbbbbbb tttt eeeeeeeeeeee ffff3 BBBBBB mmmmmmhaaaaaa cccccccceennnn 896 dddddd ffff 3 The calculated volume of biogas needed to generate a unit of electrical energy (kwh) and the calculated thermal conversion efficiency of the engine-generator from methane gas to electricity are shown in Figure 15. A significantly higher thermal conversion occurred in Period 5 likely due to be an artifact of the data, therefore it was excluded from the graph. The thermal conversion efficiency after Period 10 maintained an average of 44%. For the entire monitoring project, the average thermal conversion efficiency was 42±4%, which was overall slightly higher than the reported value stated in the GE Jenbacher 420 gas engine performance specifications, i.e., 39.4% (General Electric, 2007). Part of the difference can be explained by the fact that, although the biogas volume to the engine-generator set was measured continuously, only one measurement of methane content was performed per period. This value was extrapolated to the entire period to calculate the thermal conversion efficiency using Equation 5. Thermal conversion efficiency (%) 60% 50% 40% 30% 20% 10% 0% Thermal conversion efficiency (%) ft3/kwh Biogas per electricity generated (ft 3 /kwh) Period Figure 15. Volume of biogas required to generate a unit of electrical energy (kwh) and thermal conversion efficiency of the engine-generator set calculated for each period. PRO-DAIRY Program - Cornell University

68 Capacity factor and online efficiency The capacity factor compares the electrical energy that was produced in a given time frame by the CHP system to the electrical energy that could have been produced over that same period of time. A capacity factor of less than 1.0 indicates that the engine-generator set did not run at full capacity and/or did not run continuously. The capacity factor was calculated for each period, using Equation 6. The capacity rating on the engine-generator set at Synergy Biogas was 1,426-kW. Equation 6: eeeeeeeeeeeeeeeeeeee eeeeeeeeeeee ppppoooooooooooo dddddddddddd oooooo ssssssssssssssss pppppppppppp (kkkkh) CCCCCCCCCCCCCCCC ffffffffffff = dddddddd pppppp pppppppppppp 24 hoooooooo eeeeeeeeeeee gggggggggggggggggg ssssss cccccccccccccccc rrrrrrrrrrrr (kkkk) dddddd The online efficiency indicates the percent of time the engine-generator was operating over the same period of time as the capacity factor. The engine-generator operating hours were recorded during each sampling date, and used to calculate the online efficiency of the system using Equation 7. Equation 7: OOOOOOOOOOOO eeeeeeeeeeeeeeeeeeee (%) = eeeeeeeeeeee gggggggggggggggggg ssssss hoooooooo rrrrrr pppppp ssssssssssssssss pppppppppppp dddddddd iiii ssssssssssssssss pppppppppppp 24 hoooooooo dddddd 100 The capacity factor and online efficiency of the Synergy plant are shown combined in Figure 16. In Periods 5 and 8 (October 2012 and January 2013), the capacity factor and online efficiency decreased to almost 20%, meaning that the plant produced 20% of the electricity it could have generated during the same period of time if running continuously and at full capacity. After Period 10, the capacity factor and online efficiency of the plant significantly improved reaching an average value of over 0.8 and 90%, respectively. As mentioned before, during Periods 22 and 23 the AD PRO-DAIRY Program - Cornell University

69 system experienced several problems of a mechanical origin, which decreased the overall energy production of the system significantly. Online efficiency (%) Capacity factor 100% 1.0 Online Efficiency (%) 80% 60% 40% 20% Capacity Factor (out of 1.0) 0% Period 0.0 Figure 16. Capacity factor and online efficiency calculated for each period. Overall, the average capacity factor and online efficiency of the plant during the entire monitoring project were 0.66±0.22 and 80±23%, respectively. In general, the great majority of the problems experienced by the Synergy AD system that generated down time were of mechanical origin. Only 20% were related to the process and only 8% were part of normal systems maintenance (Figure 17). Maintenance 8% Process 20% Mechanical 72% Figure 17. Breakdown of problems generating downtime at Synergy. PRO-DAIRY Program - Cornell University