- Email: [email protected]
Biogas production from brown grease using a pilot-scale high-rate anaerobic digester
Renewable Energy 68 (2014) 304e313
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Biogas production from brown grease using a pilot-scale high-rate anaerobic digester Pengchong Zhang a, b,1, Che-Jen Lin a, c, *, James Liu c, Pruek Pongprueksa d, Simon A. Evers e, Peter Hart f, ** a
College of Environment and Energy, South China University of Technology, Guangzhou, 510006, China Department of Chemical Engineering, Lamar University, Beaumont, TX 77710-10053, USA Department of Civil Engineering, Lamar University, Beaumont, TX 77710-10024, USA d Department of Mechanical Engineering, Lamar University, Beaumont, TX 77710-10028, USA e Meridian Bioenergy, Inc., The Woodlands, TX 77380, USA f MeadWestvaco Corporation, Richmond, VA 23219-0501, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 May 2013 Accepted 30 January 2014 Available online 3 March 2014
Food wastes are typically disposed of in landfills for convenience and economic reasons. However, landfilling food wastes increases the organic content of leachate and the risk of soil contamination. A sound alternative for managing food wastes is anaerobic digestion, which reduces organic pollution and produces biogas for energy recovery. In this study, anaerobic digestion of a common food waste, brown grease, was investigated using a pilot-scale, high-rate, completely-mixed digester (5.8 m3). The digestibility, biogas production and the impact of blending of liquid waste streams from a nearby pulp and paper mill were assessed. The 343-day evaluation was divided into 5 intensive evaluation stages. The organic removal efficiency was found to be 58 9% in terms of COD and 55 8% in terms of VS at a hydraulic retention time (HRT) of 11.6 3.8 days. The removal was comparable to those found in organic solid digesters (45e60%), but at a much shorter HRT. Methane yield was estimated to be 0.40 1 , higher than the typical range of other food wastes (0.11e0.42 m3e0.77 m3-CH4@STP kg-VSremoved 1 CH4@STP kg-VSremoved), with a mean methane content of 75% and <200 ppm of hydrogen sulfide in the biogas. The blending of selected liquid wastes from a paper mill at 10 vol% of brown grease slurry did not cause significant reduction in digester performance. Using a pseudo-first-order rate law, the observed degradation constant was estimated to be 0.10e0.19 d1 compared to 0.03e0.40 d1 for other organic solids. These results demonstrate that brown grease is a readily digestible substrate that has excellent potential for energy recovery through anaerobic digestion. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Brown grease Biogas Anaerobic digestion Renewable energy
1. Introduction Using biogas as an alternative source of energy is gaining more attention globally in recent decades [1,2]. There has been an increasing number of studies performed to evaluate the conversion of waste streams such as animal manure, municipal solid wastes, energy crops, municipal bio-solids and food wastes to biogas [2e5]. According to International Energy Association (IEA), there are over 9000 anaerobic digesters in operation using these materials to * Corresponding author. Department of Civil Engineering, Lamar University, Beaumont, TX 77710-10024, USA. Tel.: þ1 409 880 8761. ** Corresponding author. E-mail addresses: [email protected] (C.-J. Lin), [email protected] (P. Hart). 1 Current address: Oil & Gas Division, Conceptual & Engineering Services, Siemens Energy Inc., Houston, TX, USA. http://dx.doi.org/10.1016/j.renene.2014.01.046 0960-1481/Ó 2014 Elsevier Ltd. All rights reserved.
produce biogas [6]. For instance, w15% of organic wastes are being converted annually in Germany [7,8]. The practice of converting wastes to energy provides a two-fold benefit of environmental protection and energy recovery. Brown grease (BG) is a mixture consisting of trapped grease, sewage grease, and black grease collected in grease interceptors (traps) of restaurants and food industries [9]. In the United States, there are 1.84 million tons of BG produced every year [10]. Most collected BG eventually ends up in landfills. In the US, the landfill cost for BG varies by region and was up to US$110 per metric tonne in 2002 [10]. This results in a very high direct disposal cost. In addition, the moisture content in BG can lead to soil and water pollution, making the soil sterile and unable to support plant life [11]. Because of these drawbacks, the European Union enacted a general ban on landfilling organic waste in 2005 [11]. An earlier
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Nomenclature AD ALK BAL BG BMP COD CSTR EPA FAC FC GPD HRT
anaerobic digester total alkalinity balance tank brown grease biochemical methane potential chemical oxygen demand continuous stirred-tank reactor Environmental Protection Agency facultative tank foul condensate, a liquid waste stream from kraft mill process gallons per day hydraulic retention time
study suggested that 14 106 m3 of methane could be produced in the US annually by converting the generated BG into biogas [12]. This is a substantial amount of renewable bio-energy. Recovering the energy while eliminating the waste input to landfills yields both economic and environmental benefits [13]. Anaerobic digestion is a treatment process capable of producing biogas from organic wastes. The benefits of anaerobic digestion include smaller reactor size in terms of organic loading, lower air emissions and a smaller amount of generated sludge compared to aerobic biological treatment [14]. Greasy wastes such as BG have been added as a lipid-rich co-substrate in earlier anaerobic digestion studies for sewage sludge [15,16], municipal wastewater [17e19], and the digestible fraction of municipal solid wastes [20,21]. Typically it is blended at 2e50% of the primary substrate’s organic loading to improve the biogas yield and methane content [15e21]. However, higher lipid loading (>50% of the substrate) can cause long chain fatty acid (LCFA) inhibitions [15,16,22], scum and foam formation [23] and fat clogging problems [17,23]. To our knowledge, there are few studies devoted to investigating the degradability and biogas production using BG alone. Biochemical methane potential (BMP) data provide an estimate of biogas production from digesting an organic substrate anaerobically. Combining the BMP data and the associated kinetics of substrate degradation, the waste treatment efficiency and the cost benefits of an anaerobic digestion process can be optimized [24,25]. BMP measurements are typically carried out in lab-scale experiments. Although effective and easy-to-control, the estimated BMP can be up to 50% higher than the biogas production in full-scale digesters where operational constrains exists [26]. This leads to a large uncertainty in estimating the biogas production from commercial-scale operation. Investigating methane production in a pilot-scale operation can provide a more practical estimate of the energy values of digested substrates. The objective of this study is to employ a pilot-scale anaerobic digestion system to assess the methane yield and treatment efficacy of BG. Kinetic evaluation was performed to estimate the substrate utilization rate constant. Various liquid waste streams from an adjacent paper mill including foul condensate (FC) and screw press liquor (SPL) were blended as an effort to minimize the water use in the feed. The effects of process parameters including substrate composition, hydraulic retention time (HRT), organic loading rate (OLR) and operational changes on the system performance are reported.
LCFA OLR ORP SPL ST STP TN TP TS TSS VFA VS VSS
305
long chain fatty acid organic loading rate oxidation reduction potential screw press liquor, a liquid waste stream from paper mill sedimentation tank standard temperature and pressure (0 C, 1 atmosphere) total nitrogen total phosphorous total solids total suspended solids volatile fatty acids volatile solids volatile suspended solids
2. Materials and method 2.1. Pilot-scale system setup The high-rate anaerobic digestion system employed in this work comprises of three CSTRs and a clarifier: a balance tank (BAL), a facultative tank (FAC), an anaerobic digester (AD), and a final sedimentation tank (ST). The system schematic and the corresponding mixing and transferring pumps are presented in Fig. 1. The BAL and FAC are rectangular shaped tanks having an adjustable volume of 0.2e1.0 m3. AD is a cylindrical tank with a total volume of 7.6 m3 (1.6 m in diameter and 3.8 m in height) with adjustable reaction volumes of 4.3, 5.8 and 7.6 m3. The nominal reactor volume in this study was 5.8 m3. It has a Plexiglas window at the top for observing the mixed liquor in the digester. BAL and FAC were mixed with submerged mixing pumps. Mixing in AD was accomplished by re-circulating the digester liquor. The mixing was checked regularly by taking suspended solid measurements at different reactor heights to ensure a completely mixing condition. The sedimentation tank (1.5 m3) has a cylindrical shape with a conical bottom at 1:1 slope. The BG feedstock is a solid phase substrate with a high chemical oxygen demand (COD). Before feeding to BAL, the BG feedstock was re-suspended in tap water to form BG slurry (water:BG ¼ 1:1 in volume) using a hand-held mixing pump. The re-suspended BG slurry was then mixed with tap water and/or co-substrates in the BAL yielding a substrate concentration of 25,000e50,000 mg L1 COD. After this homogenization process, the substrate was pumped to FAC for pre-digestion. A commercial facultative culture powder (Meridian Bioenergy) was introduced (at 1:5000 mass ratio of bioculture to COD) to FAC for bio-augmentation at 12-h HRT. The purpose of this process is to initially introduce a series of microbial strains, degrade possible LCFA and to eliminate potential process inhibitors as reported earlier [15]. Afterwards, the pre-digested substrates were pumped continuously into AD for digestion. The AD effluent was transferred to ST gravitationally for sedimentation. 2.2. Evaluation schedule The evaluation period lasted for 343 days. Excluding the system start-up, maintenance and feeding transition periods, process data were collected for 238 days. The evaluation was divided into five intensive evaluation periods (IeV). During each operating period, a steady stage (S1 to S5, defined as a state with relatively consistent biogas production and organic removal) was selected for intensive
306
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 1. Schematic diagram of the pilot-scale high-rate anaerobic digester.
Table 1 Feeding characteristics and reactor configuration during the evaluation. Data were collected in five different periods for analysis.
Date Days of operation Days of intensive evaluationa Sedimentation tank Feeding Influent COD (mg L1) Influent VS (mg L1) OLRb (kg-VS m3d1) HRT (days) Activity a b
System start-up
I
II
III
IV
V
4/13/11e7/26/11 / /
7/27/11e8/7/11 1e12 1e12 (S1)
8/8/11e10/24/11 13e90 34e45 (S2)
10/25/11e12/7/11 91e135 107e133 (S3)
12/8/11e2/29/12 136e217 184e217 (S4)
3/1/12e3/21/12 218e238 218e238 (S5)
No BG / / / / Seeding and initiating
No BG 34,510 2557 13,965 1262 2.0 0.2 7.3 0.6 Establish BG steady state
Yes BG 56,570 3894 23,937 1625 2.7 0.3 (2.0 0.3) 8.9 0.9 Add ST
Yes BG þ FC 26,570 6264 10,139 754 1.0 0.2 (0.8 0.2) 15.2 1.1 Establish BG þ FC steady state
Yes BG 33,881 9176 13,224 3236 0.8 0.2 (0.6 0.2) 15.8 1.9 Back to BG steady state
Yes BG þ SPL 30,200 1503 13,225 1891 1.2 0.3 (0.9 0.3) 11.0 0.1 Establish BG þ SPL steady state
S1eS5 stands for five selected stages with intensive evaluation and stable data consistency. OLR and HRT were calculated based on AD only. The OLR in the parentheses for S2eS5 were based on AD þ ST.
measurement and data analysis. Table 1 summarizes the evaluation schedule and the corresponding operating parameters in each stage. During S1 (day 1e12, Table 1), the system was operated without ST and the AD effluent was considered as the final effluent. To understand the effect of solid recycle, ST was added to the system during S2 (day 13e90). During S3eS5 (day 91e238), the system flow follows the schematics in Fig. 1. The supernatant (overflow) of ST was the final effluent and the underflow was returned to AD. In S3 and S5, FC and SPL was fed respectively as co-substrates to reduce tap water use and to investigate their impact on the system performance. The S4 period was used to re-establish baseline conditions between the FC and SPL co-substrate additions. 2.3. Chemical analysis and biogas measurement Daily liquid samples were taken from five sampling points: BAL effluent, FAC effluent, AD effluent, ST underflow (return sludge) and ST overflow (final effluent). Gas samples were drawn from a gas sampling outlet at top of AD at least twice a day. Table 2 shows the analytical measurements and frequencies for each sample.
The pH, dissolved oxygen (DO), oxygen reduction potential (ORP) and temperature (T) were measured on site using a calibrated portable meter with appropriate probes (Hach HQ40D with pH101, LBOD101, and IntelliCAL ORP probes) after samples were drawn from the reactors. All laboratory measurements were performed within 2 h after sampling, therefore no sample preservation was performed. The total solids (TS), volatile solids (VS), total suspended solids (TSS) and volatile suspended solids (VSS) were measured according to EPA Standard Methods 2540 [27]. Dissolved parameters including dissolved COD (dCOD), total nitrogen (TN), total phosphorous (TP), volatile fatty acids (VFA), total alkalinity (ALK), sulfide and sulfate were measured after the samples were centrifuged (AccSpinÔ 400) for 30 min at 6000 rpm. These analyses were performed using EPA approved methods (Hach Company DOC 316.5, EPA 310.2, 350.1, 353.2, and 365.3, CFR 136.3, 141.5) with a time-lapse heating reactor (Hach DRB200) and a spectrophotometer (Hach DR3800). The cumulative volume of produced biogas was measured continuously using a digital gas flow meter (FILLRITE Model 820, TTS Corp., IN, USA). The composition of biogas was measured using a portable gas analyzer with an infrared detector (Model GEM 2000, Landtec Inc., Colton, CA, USA). Standard gases
Table 2 Analytical parameters and frequency for each sampling point. Sampling point
After BAL
After FAC
After AD
ST recycling sludge
ST up-flow (final effluent)
Biogas (gas point on AD)
Analytical frequencies Measurements
Daily pH, T, OLR, COD, VFA
Daily pH, T, DO, ORP, TN, TP, VFA, ALK, TS, VS, TSS, VSS, COD
Daily pH, T, DO, ORP, TN, TP, VFA, ALK, TS, VS, TSS, VSS, COD
Daily pH, T, TS, VS, TSS, VSS, COD
Daily pH, T, DO, COD
At least twice a day Flowrate, content of CH4, CO2, H2S, and O2
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
307
Table 3 Substrate characteristics. Brown grease was used as the primary substrate and the other two liquid wastes were used as co-substrates in part of the evaluation. Parameter
Brown Grease (BG)a (m s, n ¼ 17)
Foul condensate (FC) (m s, n ¼ 11)
Screw press liquor (SPL) (m s, n ¼ 13)
COD (mg L1) dCOD (mg L1) TS (mg L1) VS (mg L1) VS/TS ratio TSS (mg L1) VSS (mg L1) VSS/TSS ratio Alkalinity (mg L1 as CaCO3) pHb TN (mg L1)c TP (mg L1)c Sulfide (mg L1) Sulfate (mg L1) Moisture content (wt%)
910,634 229,993 / 437,778 91,348 372,111 77,646 0.85 0.06 / / / / 6.51 0.77 / / / / 56 9
2973 142 2740 125 406 104 210 14 0.53 0.1 357 58 339 46 0.83 0.25 205 50 9.28 0.18 52.2 4 0.24 0.09 52.2 20.5 <40 /
4498 2020 609 189 8768 7957 3742 1666 0.5 0.1 4048 1750 1997 875 0.49 0.06 / 8.44 0.83 2.3 0.1 0.41 0.04 / / /
a b c
Here BG stands for pre-treated brown grease in solid phase, thus the unit of COD, TS, and VS is mg kg1. pH of brown grease was measured by suspending 100 g brown grease in 1 L tap water that has a pH of 8.05 and alkalinity of 55 mg L1 as CaCO3. In aqueous phase.
(50% CO2, 50% CH4, dry air and 1000 ppm H2S) were used for the calibration of the gas analyzer daily. The methane yield of the substrates was estimated as the ratio of the volume of CH4 (at the standard temperature and pressure, STP) produced in the digester to the VS of the substrate digested (m3-CH4@STP kg-VS1 removed). During the five intensive evaluation periods (S1eS5, Table 1), analyses of four replicates of randomly selected samples for all the above-mentioned parameters typically had a relative standard deviation less than 5 % except for solids, which had larger analytical errors (<10%). The completeness of data used for process performance evaluation was over 90%.
2.4. Substrate characteristics The BG feedstock of this study was obtained daily from a waste management plant in Houston, TX, where the raw BG was screened to remove large particles, stabilized by lime addition, and then flocculated and dewatered. The processed BG is a light-brown, sticky, greasy solid with a perished food smell. In different stages of evaluation, it was emulsified into water at different concentrations (Table 1). After re-suspension in tap water, the slurry has a neutral pH range (6.5e7.5), suitable for biological treatment such as anaerobic digestion. No nutrients were added because of the
balanced nutrient composition of BG was deemed sufficient for biological treatment [28]. The characteristics of BG feedstock, FC, and SPL are shown in Table 3. Since the BG has an extremely high organic content (w1 kg-COD kg-BG1, Table 3), the feeding stream was diluted to the range of 25,000e50,000 mg L1 COD. FC and SPL have a relatively low COD level and solids content compared with BG (Table 3). In addition, their mild alkalinity (Table 3) effectively offset the mild acidity in BG. FC is a liquid substrate with relatively low solid content (TS ¼ 400 mg L1), its major organic content is in the dissolved phase (dCOD is >90% of total COD, Table 3). SPL has a TS content less than 1.0 wt%. Its dCOD concentration is <20% of total COD concentration (Table 3), indicated the major organic content is in the solid phase.
3. Results and discussion 3.1. Performance of anaerobic digester The operating parameters of AD and system performance during the five intensive study periods (S1eS5, Table 1) are summarized in Table 4. For comparison, the typical values for anaerobic digestion systems are also included. During the evaluation, the pH in AD was
Table 4 Anaerobic digestion operating parameters and system performance in five selected stages. Stages
I
II
III
IV
V
Typical rangea
pH T ( C) DO (mg L1) ORP (mV) TN (mg L1) TP (mg L 1) Alkalinity (mg L1 as CaCO3) VFA (mg L1 as HAc) COD removal efficiency (%) VS removal efficiency (%) CH4 content (%) CO2 content (%) H2S content (ppm) CH4 yield (m3-CH4 kg-VS1 removed)
7.34 0.05 36.0 0.7 0.01 0.00 209 14 591 83 3.4 2.4 3087 282 274 97 42.1 6.7 26.8 7.9b 74.3 2.0 22.3 1.3 38.2 4.1 0.40e0.49
/ 36.3 0.7 / 228 24 409 37 1.5 0.4 / / 50.6 5.8 37.1 4.3b 74.6 1.0 / / 0.58e0.77
7.12 0.08 34.3 1.8 0.06 0.04 243 40 237 74 0.9 0.4 1455 457 199 76 73.8 11.0 72.7 7.4 75.9 1.9 23.9 1.9 147.2 34.8 0.49
7.10 0.07 34.3 2.1 0.15 0.05 247 37 314 50 2.3 1.1 2478 291 394 84 61.7 12.3 57.9 13.2 74.6 1.8 25.2 1.8 185.2 28.1 0.48
7.01 0.17 37.9 1.0 0.10 0.03 263 23 306 46 2.2 0.4 2204 222 469 378 53.5 8.7 56.4 9.9 75.4 1.0 24.2 1.0 371.7 127.6 0.45
6.5e8.5 [52] 35e40 [52] / m400w150 [53] 60e1000 [54] 6e50 [54] 1500e5000 [55] <1800 [53] / / / / / 0.11e0.42 [3,24,44e50]
a Typical value of operating parameters including pH, T, ORP, TN, TP, VFA and alkalinity were based on the description of typical anaerobic digestion systems. Typical values of methane yield were based on earlier literature. b For comparison purpose, VS removal efficiency in S1 and S2 has not been corrected by biomass calculation.
308
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 2. VFA concentrations before and after AD. The alkalinity in AD is also shown.
controlled at neutral range (7.01e7.34, optimal pH 6.9e7.6 for anaerobic digestion) and the temperature was maintained in the mesophilic range (34.3e37.9 C). The low DO concentration (0.05e 0.20 mg L1) and ORP value (<200 mV) indicated strictly
anaerobic conditions were maintained during the evaluation. TN and TP concentrations in the system were 230e600 mg L1 as nitrogen and 1e4 mg L1 as phosphorous, indicating adequate nitrogen but possible phosphorus deficiency. VFA concentration in AD was generally lower than 500 mg L1 as acetic acid (HAc) except during the maintenance period (Fig. 2). The average total alkalinity was 2122 mg L1 as CaCO3, suggesting absence of in vitro VFA toxicity [29,30]. These operational parameters were controlled in the typical range of healthy digesters (Table 4). The pre-digestion process in FAC effectively eliminated the negative effects of the lipid-rich substrate. Earlier studies have reported that fat scum foaming by greasy substrates may be a concern [31,32]. However, no significant scum formation was observed in the AD during the evaluation. The surface fat scum was rapidly diminished within an hour in FAC [33], possibly because of the effectiveness of the pre-digestion process breaking down LCFAs that form scum. Perle et al. (1995) conducted a lab-scale experiment of dairy wastewater and noted a similar result [34]. The pre-digestion process also produced VFAs that accelerated the anaerobic digestion process. To verify the VFA production through the bio-augmentation process, a facultative experiment using diluted BG slurry (10,000 mg L1 COD) with the commercial facultative culture was conducted in a lab-scale reactor (0.5 L). Experimental results show that after 12-h of pre-digestion at room temperature, the pH was decreased from 7.03 to 5.55, and VFA concentration was increased from 148 to 293 mg L1 as HAc. In the pilot-scale system, the facultative pre-digestion efficiently
Fig. 3. Concentrations of (a) COD, and (b) VS before and after AD. Their removal efficiencies are shown on the secondary axis. The five selected stages (S1eS5) are marked.
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 4. Volatile fractions of (a) VS/TS and (b) VSS/TSS in FAC and AD.
augmented VFA generation for methanogenesis in AD. Fig. 2 shows the VFA level in FAC and AD as well as the total alkalinity in AD. The FAC effluent has about 2 times higher VFA concentrations (600e 1800 mg L1 as HAc, Fig. 2) compared to the FAC influent. The average VFA concentration decreased from w800 mg L1 in FAC to w413 mg L1 as HAc in AD (Fig. 2). Fig. 3 shows the COD and VS variation and removal efficiency during the five intensive evaluation periods (S1 to S5, Table 1). In S1, only BG was fed to the AD at a mean daily flow rate of 0.568 m3 d1 (150 gallon per day, GPD) without recycling of anaerobic sludge. The COD in the influent and effluent was w33,000 mg L1 and w20,000 mg L1 respectively. Since the organic content of effluent is too high for the subsequent aerobic treatment, a ST was added on day 13 to remove the solid content of the digester effluent and to increase the anaerobic biomass in AD with sludge recycling. In S2, 0.151 m3 d1 (40 GPD) of the anaerobic sludge was recycled to AD. The total daily flow was slightly reduced to 0.503 m3 d1 (133 GPD), i.e., the feeding stream was reduced to 0.341-0.378 m3 d1 (90e100 GPD). After the modification, the influent and effluent organic contents in S1 and S2 were compared. In S2, the overall feed concentration to AD was higher (45,000e 60,000 mg L1 COD) than S1 (30,000e40,000 mg L1 COD) because of the recycled sludge (Fig. 3). The mean influent and effluent VS concentrations were 13,965 and 9448 mg L1 in S1. In S2, due to the added organic load of recycled sludge, the mean influent VS
309
increased to 23,937 mg L1 and the effluent VS increased to 12,078 mg L1. During S1 and S2, the OLR and HRT were calculated based on the flow into and out of AD. Stage I used a lower concentration (w14 g-VS/L) and a higher flowrate (150 GPD), while Stage II used a higher concentration (w24 g-VS/L) and a lower flowrate (90e100 GPD). This resulted in w30% increased of OLR (from 2.0 to 2.7 kg VS m3 d1) in Stage II (Table 1). Since the feeding flow rate in S2 was reduced from 0.568 m3 d1 to 0.503 m3 d1, The HRT of S2 increased to 8.9 days compared with S1 (7.3 days, Table 1). The COD and VS removal of S2 was slightly higher than S1 (Table 4), possibly because of the increased amount of biomass in AD which improved the digester performance. Starting from day 91 (S3), the data analysis included both AD and ST as a digestion system. During S3eS5 (day 91e238), the influent COD concentration was kept at w26,500 mg L1 (Table 1) and the influent VS was in the range of w11,000 mg L1. The flow rate in S3 and S4 was reduced to 0.246e0.322 m3 d1 (65e85 GPD, HRT 15e16 days). The flow rate in S5 was maintained at 0.397 m3 d1 (105 GPD, HRT 11 days, Table 1). During these periods, the ST overflow had a relatively consistent COD and VS concentration (COD w 10,000 mg L1 and VS w 3500 mg L1, Fig. 3). This implies that, the effluent reached a stable level to be treated aerobically after ST. Similar results for palm oil mill effluent were reported by Basri et al. (2010) [35]. FC and SPL were introduced as co-substrates at 10 vol% during S3 and S5, respectively. The addition of the co-substrates did not impose significant impact upon the methane yield (Table 4). Compared to the periods feeding BG only (S4), the COD and VS removal efficiency of S3 increased from w60% to w74% (COD, Table 4 and Fig. 3) and w55% to w73% (VS, Table 4 and Fig. 3) respectively. This is probably because the organic content of FC is relatively easy to degrade (>80%) [36e38]. SPL did not appear to significantly affect the COD and VS removal efficiency (S5, Fig. 3). To understand the organic removal in the system at different stages, the volatile ratios of solids (VS/TS and VSS/TSS) before and after AD were monitored. The volatile ratio is an indicator of organic degradation and of inorganic accumulation since the inorganic mass fraction of solids is conserved during digestion [39]. Fig. 4 shows the VS/TS ratio (Fig. 4a) and VSS/TSS ratio (Fig. 4b) in FAC and AD, respectively. Before AD, the volatile ratio for BG was in the range of 0.70e0.90. After AD, the volatile ratio was reduced to the range of 0.60e0.70, suggesting the organic content of the feed has been effectively removed in the digester. During S3eS5, the addition of co-substrate also affected the volatile ratio, and the varying tendency was in accordance with COD removal efficiency (Figs. 3 and 4). During the entire operating period, the average COD and VS removal efficiency were 58% and 55% (Table 4), lower than those previous reported for lipid-rich wastewaters (w90%) [40] but comparable to those of typical solid digesters (45e60%) [41]. Fig. 5 shows the scatter plot between the organic removal and organic loading rate in terms of VS (Fig. 5a) and COD (Fig. 5b). During S1 and S2, the COD and VS removal efficiency were not significantly affected by OLR variation, resulting in a linear increase of VS and COD mass digestion with respect to the applied OLR (Fig. 5). However, the organic removal efficiency in S2 was greater than that in S1, as reflected by the change of COD and VS in the feed and digester effluent (Fig. 3). S3 to S5 had the similar trend of linear increase of digested COD and VS mass with respect to the OLR, indicating that the system should have a higher treatment capacity than the OLR range applied during the evaluation. The daily biogas production during the evaluation is summarized in Fig. 6. In S1, the biogas production was 5e6 m3 d1. The biogas production was higher in S2 (w7 m3 d1) because of the
310
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 5. Scatter plots between organic loading rate (OLR) and organic removal in terms of (a) VS and (b) COD. The points in the frames show the data during S1 and S2 when the operational parameters were different from those in S3-S4.
higher organic load and recycled biomass (Fig. 5a). During S3eS5, the average daily biogas production was lower than S1 and S2 since the system OLR was reduced (Table 1). In S3, the COD removal efficiency was higher than S4 and S5 (Table 4), leading to higher biogas production (w5.6 m3 d1) compared to that in S4 and S5 (w3.5 m3 d1). The easily digested dCOD in FC accounted for the higher biogas production in S3. The lower biogas production in S4 and S5 (as compared to that of S3) was caused by the lower OLR applied (S4, Table 1) and the slightly lower organic removal (S5, Table 4), respectively. Generally, the biogas production trend in S3e S5 was consistent with the organic removal shown in Fig. 3, suggesting that the biogas production was not significantly affected by the addition of co-substrate. 3.2. Methane yield and kinetic analysis The produced biogas had a consistently high CH4 content (w75%, Table 4). The other major gas (CO2) consisted of the other w25% by volume (Table 4) and trace gases (e.g., H2S). From day 160e175, the system was recovered from system maintenance and the methane content built up from 40% to 75% quickly (Fig. 6). During the entire evaluation, the average H2S concentration was 189 ppm, significantly lower than the level that may cause H2S toxicity (w1500 ppm) [42]. The cumulative CH4 production and digested VS in S3 to S5 are shown in Fig. 7. The methane yields of S3 to S5 were calculated as the ratio of the two slopes. The value was reported based on VS
removal because the organic content of the BG feed was mainly in the suspended solid phase. The methane yield of BG in S3 to S5 was consistently in the range of 0.45e0.49 m3-CH4 kg-VS1 removed (All gas volumes mentioned hereafter have been normalized to STP). For the first two stages (S1 and S2), the apparent VS removal efficiency (27e37%, Table 4) was significantly lower than in S3eS5 (55e75%, Table 4) because ST had not been introduced to the system. Under this condition, the effluent VS during S1 and S2 contains substantial amount of biomass produced from the anaerobic digestion of brown grease. During S3 to S5, ST was used to collect and recycle most of the generated biomass back to the AD, resulting in the higher apparent BG removal efficiency. To estimate the BG conversion into biogas during S1 and S2, a mass balance on solids before and after the AD was performed as followed:
ð1 f ÞF ¼ ð1 aÞX þ ð1 bÞY
(1)
aX þ bY ¼ M
(2)
Eq. (1) represents the mass balance of the fixed (inorganic) solids. Where f is the volatile fraction of influent BG substrate obtained from measurement (0.808 in S1 and 0.816 in S2), F is the mass flow of influent total solid (kg d1). a and b are the volatile fraction (VS/TS) of biomass and undigested BG substrate, respectively (a ¼ w0.80 [43]). X and Y are the mass flow of biomass and undigested BG (kg d1). Eq. (2) represents the VS composition in
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
311
the two lines in the respective period. The estimated methane yield in S1 (0.40e0.49 m3-CH4 kg-VS1 removed) was comparable to the yield in S3 to S5 (0.45e0.49 m3-CH4 kg-VS1 removed). S2 has a higher methane yield (0.58e0.77 m3-CH4 kg-VS1 removed) because of the shorter hydraulic retention time and higher organic loading (Table 1) in the presence of recycled biomass. Under such condition, only the readily degradable fraction of the substrate in the feed was digested, leading to a higher methane yield and lower organic removal efficiency (Table 4). In S3eS5, the lowered organic loading and higher HRT improved the organic removal at the cost of reduced methane production. In practice, the mode of process operation will depend on the treatment objective (better organic removal or higher methane yield). Also, the added co-substrate (FC and SPL) did not adversely affect the methane yield during S3eS5 (Table 4). A pseudo-first-order kinetic model was applied to analyze the substrate utilization. Similar approaches have been used earlier [3,24,44e50]. The substrate concentration was calculated based on VS. For a CSTR at steady state, the effluent concentration (C) can be estimated as:
Fig. 6. Measured daily biogas production and CH4/CO2 content.
C ¼
Fig. 7. Cumulative CH4 production at STP and cumulative VS digested during five selected stages (S1eS5). The slopes of each linear stage were used to calculate corresponding CH4 yield. In S1 and S2, the mass of digested VS was corrected by biomass calculation.
the effluent, where M is the mass flow of VS in the effluent (kg d1). Using the solid measurements, we estimated that the generated biomass constitutes 25e40 wt% in the effluent. Based on the mass balance results, the cumulative CH4 production and the digested VS during S1eS5 were plotted in Fig. 7. The methane yield was then calculated as the ratio of the slopes of
C0 1 þ kq
(3)
where C0 is the influent substrate concentration (mg L1 VS), k is the first-order substrate utilization rate constant (day1) and q is the HRT (d). The estimated k value was in a relatively consistent range of 0.10e0.19 d1 throughout the evaluation process. For comparison, previously reported methane yields of food wastes and their first-order kinetic parameters are shown in Table 5 [3,24,44e50]. Generally, the degradation rate constants were in the range of 0.03e0.4 d1. The rate constant obtained in this study (0.10e0.19 d1) had probably been adversely affected by the greater difficulty of controlling the digestion conditions (temperature and mixing) in a pilot-scale system due to the ambient temperature variation (>15 C diurnal change). It was slightly lower than that of municipal solid sludge in batch reactors (0.2e0.4 d1, [51]), comparable to that of municipal solid sludge in CSTR (0.175 d1, [44]), and higher than that of canary grass in CSTR (0.03e0.04 d1, [3]). The measured methane yield of BG in this pilot study ranged from 0.40 to 0.77 m3-CH4 kg-VS1 removed. Because of the higher lipid content of BG, the value is substantially higher than the methane yields of typical sugary and sludge substrates (0.11e0.42 m3-CH4 kgVS1 removed, Table 5). The biogas quality produced by BG is excellent (w75%, Table 4), possibly also due to the high lipid content of BG. These pilot-plant data suggest that BG can be effectively digested anaerobically for high quality biogas production.
Table 5 Comparison of reported and calculated first order degradation rate constants and methane yields. Substrate
Reactor type
Pseudo-first-order rate constant (d1)
Methane yield (m3-CH4 kg-VS1 removed)
Reference
Brown grease Municipal sludge Municipal sludge Corn stover Rice straw Canary grass Sunflower oil cake Winter wheat Waste activated sludge Waste activated sludge þ fatty wastewater Synthetic kitchen waste Synthetic kitchen waste þ municipal grease waste
CSTR CSTR Batch Batch
0.10e0.19 0.175 0.2e0.4 / / 0.03e0.04 / / / / / /
0.40e0.77 0.309 / 0.239 0.225 0.19e0.33 0.107e0.227 0.311e0.360 0.116 0.362 0.117 0.324e0.418
This study [45] [44] [49] [50] [3] [46] [47] [48] [48] [24] [24]
CSTR Batch Batch Batch Batch Batch Batch
312
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
4. Conclusions In this study, the anaerobic digestion of waste brown grease (BG) in a pilot-scale system was investigated. A mean COD removal of 58% and a mean VS removal of 55% were achieved. The organic removal efficiency was comparable with those found in typical solid digesters. Kinetic analysis showed that the pseudo-first-order degradation rate constant of BG was in the rage of 0.10e0.19 d1. After anaerobic treatment process, the effluent had a consistent effluent organic strength (COD w 10,000 mg L1) that can be treated aerobically. It was concluded that BG was a readily digestible substrate as a sole substrate. The pilot-scale system produced biogas of excellent quality (75% CH4 content), with a methane yield in the range of 0.40e0.77 m3CH4 kg-VS1 removed. The addition of paper mill waste streams (fouls condensate and screw press liquor) as the co-substrates did not adversely affect the methane yield. BG has the industrial potential to be anaerobically treated as an energy feedstock and there has been ongoing commercial effort to build large-scale digesters using BG as the primary substrate. Using BG for biogas production could serve as a profitable model for converting waste to renewable energy.
Acknowledgment This study was supported by MeadWestvaco Evadale TX facility (Project No: MWV0001). The authors would like to thank Dipendra Wagle, Sophia Yang, Yolanda Wang, Brandon Corace and Erik Corace for their assistance in the field work and laboratory analysis. The assistance of Gary Colson, Michael Clapper and Robert Sasser in supporting this work and in obtaining and supplying paper mill samples is greatly appreciated. The administrative support of Stewart Cairns, Reid Sweet and Thomas Sitton are also acknowledged.
References [1] Hilkiah Igoni A, Ayotamuno MJ, Eze CL, Ogaji SOT, Probert SD. Designs of anaerobic digesters for producing biogas from municipal solid-waste. Appl Energy 2008;85:430e8. [2] Nges IA, Escobar F, Fu X, Bjornsson L. Benefits of supplementing an industrial waste anaerobic digester with energy crops for increased biogas production. Waste Manag 2011;32:53e9. [3] Kacprzak A, Krzystek L, Pa zdzior K, Ledakowicz S. Investigation of kinetics of anaerobic digestion of Canary grass. Chem Papers 2012;66:550e5. [4] Sivakumar P, Bhagiyalakshmi M, Anbarasu K. Anaerobic treatment of spoiled milk from milk processing industry for energy recovery e a laboratory to pilot scale study. Fuel 2012;96:482e6. [5] Lay J-J, Lee Y-J, Noike T. Feasibility of biological hydrogen production from organic fraction of municipal solid waste. Water Res 1999;33:2579e86. [6] Internation Energy Association. Country Reports of Task 37: Energy from Biogas. http://www.iea-biogas.net/country-reports.html; November 2013 [accessed December 2013]. [7] Gallert C, Henning A, Winter J. Scale-up of anaerobic digestion of the biowaste fraction from domestic wastes. Water Res 2003;37:1433e41. [8] Weiland P. Anaerobic waste digestion in Germany e status and recent developments. Biodegradation 2000;11:415e21. [9] Burgess MN. What to do with brown grease, plumbing systems & design; 2010. [10] Tyson KS. Brown grease feedstocks for biodiesel. National Renewable Energy Laboratory; 2002. [11] Murto M, Bjornsson L, Mattiasson B. Impact of food industrial waste on anaerobic co-digestion of sewage sludge and pig manure. J Environ Manage 2004;70:101e7. [12] Dichtl N. Thermophilic and mesophilic (two stage)anaerobic digestion. J Charter Inst Water Environ Manage 1997;11:98e104. [13] Dearman B, Bentham RH. Anaerobic digestion of food waste: comparing leachate exchange rates in sequential batch systems digesting food waste and biosolids. Waste Manage 2007;27:1792e9. [14] Gujer W, Zehnder AJB. Conversion processes in anaerobic digestion. Water Sci Technol 1983;15:127e67.
[15] Noutsopoulos C, Mamais D, Antoniou K, Avramides C. Increase of biogas production through co-digestion of lipids and sewage sludge. Global Nest J 2012;14:133e40. [16] Neczaj E, Bien J, Grosser A, Worwag M, Kacprzak M. Anaerobic treatment of sewage sludge and grease trap sludge in continuous co-digestion. Global Nest J 2012;14:141e8. [17] Long JH, Aziz TN, Reyes DL. Anaerobic co-digestion of fat, oil and grease (FOG): a review of gas production and process limitations. Process Safety Environ Protect 2012;90:231e45. [18] Bouchy L, Perez A, Camacho P, Rubio P, Silvestre G, Fernandez B, et al. Optimization of municipal sludge and grease co-digestion using disintegration technologies. Water Sci Technol 2012;65:214e20. [19] Wan C, Zhou Q, Fu G, Li Y. Semi-continuous anaerobic co-digestion of thickened waste activated sludge and fat, oil and grease. Waste Manag 2011;31: 1752e8. [20] Martin-Gonzalez L, Castro R, Pereira MA, Alves MM, Font X, Vicent T. Thermophilic co-digestion of organic fraction of municipal solid wastes with FOG wastes from a sewage treatment plant: reactor performance and microbial community monitoring. Bioresour Technol 2011;102:4734e41. [21] Martin-Gonzalez L, Colturato LF, Font X, Vicent T. Anaerobic co-digestion of the organic fraction of municipal solid waste with FOG waste from a sewage treatment plant: recovering a wasted methane potential and enhancing the biogas yield. Waste Manag 2010;30:1854e9. [22] Zhu Z, He Q. Enhancing biomethanation of municipal waste sludge with grease trap waste as a co-substrate. Renew Energy 2011;36:1802e7. [23] Valladao ABG, Cammarota C, Freire DMG. Performance of an anaerobic reactor treating poultry abattoir wastewater with high fat content after enzymatic hydrolysis. Environ Eng Sci 2011;28:299e307. [24] Li C, Champagne P, Anderson BC. Evaluating and modeling biogas production from municipal fat, oil, and grease and synthetic kitchen waste in anaerobic co-digestions. Bioresour Technol 2011;102:9471e80. [25] Angelidaki I, Alves M, Bolzonella D, Borzacconi L, Campos JL, Guwy AJ, et al. Defining the biomethane potential (BMP) of solid organic wastes and energy crops: a proposed protocol for batch assays. Water Sci Technol 2009;59:927e 34. [26] Bishop GC, Burns. RT, Shepherd. TA, Moody. LB, Gooch. CA. Evaluation of laboratory biochemical methane potentials as a predictor of anaerobic dairy manure digester biogas and methane production. In: ASABE International Meeting, Reno, NV; 2009. [27] APHA, AWWA, WEF. Standard methods for the examination of water and wastewater; 1998. Washington, DC, USA. [28] Han S-K, Shin H-S. Biohydrogen production by anaerobic fermentation of food waste. Int J Hydrogen Energy 2004;29:569e77. [29] Dong J, Zhao YS, Hong M, Zhang WH. Influence of alkalinity on the stabilization of municipal solid waste in anaerobic simulated bioreactor. J Hazard Mater 2009;163:717e22. [30] Vanhaandel AC. Influence of the digested cod concentration on the alkalinity requirement in anaerobic digesters. Water Sci Technol 1994;30:23e34. [31] de Nardi IR, Fuzi TP, Del Nery V. Performance evaluation and operating strategies of dissolved-air flotation system treating poultry slaughterhouse wastewater. Resour Conserv Recycl 2008;52:533e44. [32] Alexandre VMF, Valente AM, Cammarota MC, Freire DMG. Performance of anaerobic bioreactor treating fish-processing plant wastewater prehydrolyzed with a solid enzyme pool. Renew Energy 2011;36:3439e44. [33] Santos RB, Hart PW, Colson GW, Evers S, Evers D. Commissioning of a biogas pilot plant: from brown grease to paper mill-generated organic wastes. Tappi J 2012;11:73e8. [34] Perle M, Kimchie S, Shelef G. Some biochemical aspects of the anaerobic degradation of dairy waste-water. Water Res 1995;29:1549e54. [35] Basri MF, Yacob S, Hassan MA, Shirai Y, Wakisaka M, Zakaria MR, et al. Improved biogas production from palm oil mill effluent by a scaled-down anaerobic treatment process. World J Microbiol Biotechnol 2010;26:505e14. [36] Zhou W, Imai T, Ukita M, Li F, Yuasa A. Effect of limited aeration on the anaerobic treatment of evaporator condensate from a sulfite pulp mill. Chemosphere 2007;66:924e9. [37] Norrman J, Narbuvold R, Nystrom L. Anaerobic treatability of waste water from pulp and paper industries. Biotechnol Adv 1984;2:329e45. [38] Yang MI, Edwards EA, Allen DG. Anaerobic treatability and biogas production potential of selected in-mill streams. Water Sci Technol 2010;62:2427e34. [39] Zhang R, El-Mashad HM, Hartman K, Wang F, Liu G, Choate C, et al. Characterization of food waste as feedstock for anaerobic digestion. Bioresour Technol 2007;98:929e35. [40] Lansing S, Martin JF, Botero RB, da Silva TN, da Silva ED. Wastewater transformations and fertilizer value when co-digesting differing ratios of swine manure and used cooking grease in low-cost digesters. Biomass Bioenergy 2010;34:1711e20. [41] Gunaseelan VN. Anaerobic digestion of biomass for methane production: a review. Biomass Bioenergy 1997;13:83e114. [42] Chen Y, Cheng JJ, Creamer KS. Inhibition of anaerobic digestion process: a review. Bioresour Technol 2008;99:4044e64. [43] Eddy MA. Wastewater engineering treatment and reuse. Mcgraw-Hill Higher Education; 2003. [44] Bolzonella D, Fatone F, Pavan P, Cecchi F. Anaerobic fermentation of organic municipal solid wastes for the production of soluble organic compounds. Indust Eng Chem Res 2005;44:3412e8.
P. Zhang et al. / Renewable Energy 68 (2014) 304e313 [45] de la Rubia Á, Pérez M, Sales D, Romero LI. Municipal sludge degradation kinetic in thermophilic CSTR. AIChE J 2006;52:4200e6. [46] Raposo F, Borja R, Rincon B, Jimenez AM. Assessment of process control parameters in the biochemical methane potential of sunflower oil cake. Biomass Bioenergy 2008;32:1235e44. [47] Rincon B, Banks CJ, Heaven S. Biochemical methane potential of winter wheat (Triticum aestivum L.): influence of growth stage and storage practice. Bioresour Technol 2010;101:8179e84. [48] Carrere H, Rafrafi Y, Battimelli A, Torrijos M, Delgenes JP, Motte C. Improving methane production during the codigestion of waste-activated sludge and fatty wastewater: impact of thermo-alkaline pretreatment on batch and semicontinuous processes. Chem Eng J 2012;210:404e9. [49] Xu F, Shi J, Lv W, Yu Z, Li Y. Comparison of different liquid anaerobic digestion effluents as inocula and nitrogen sources for solid-state batch anaerobic digestion of corn stover. Waste Manage 2013;33:26e32.
313
[50] Song Z, Yang G, Han X, Feng Y, Ren G. Optimization of the alkaline pretreatment of rice straw for enhanced methane yield. Biomed Res Int; 2013. [51] Rincon B, Sanchez E, Raposo F, Borja R, Travieso L, Martin MA, et al. Effect of the organic loading rate on the performance of anaerobic acidogenic fermentation of two-phase olive mill solid residue. Waste Manage 2008;28: 870e7. [52] Khanal SK. Anaerobic biotechnology for bioenergy production principles and applications. 1st ed. Hoboken, NJ: Wiley-Blackwell; 2008. [53] Gray NF. Water technology: an introduction for environmental scientists and engineers. 3rd ed.. Oxford: Elsevier; 2010. [54] Alley ER. Water quality control handbook. 2nd ed.. Colombus, OH: WEFþMcGraw-Hill; 2007. [55] Tchobanoglous G, Burton F, Stensel H. Wastewater engineering: treatment and reuse. 4th ed. Columbus, OH: McGraw-Hill; 2002.
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Biogas production from brown grease using a pilot-scale high-rate anaerobic digester Pengchong Zhang a, b,1, Che-Jen Lin a, c, *, James Liu c, Pruek Pongprueksa d, Simon A. Evers e, Peter Hart f, ** a
College of Environment and Energy, South China University of Technology, Guangzhou, 510006, China Department of Chemical Engineering, Lamar University, Beaumont, TX 77710-10053, USA Department of Civil Engineering, Lamar University, Beaumont, TX 77710-10024, USA d Department of Mechanical Engineering, Lamar University, Beaumont, TX 77710-10028, USA e Meridian Bioenergy, Inc., The Woodlands, TX 77380, USA f MeadWestvaco Corporation, Richmond, VA 23219-0501, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 May 2013 Accepted 30 January 2014 Available online 3 March 2014
Food wastes are typically disposed of in landfills for convenience and economic reasons. However, landfilling food wastes increases the organic content of leachate and the risk of soil contamination. A sound alternative for managing food wastes is anaerobic digestion, which reduces organic pollution and produces biogas for energy recovery. In this study, anaerobic digestion of a common food waste, brown grease, was investigated using a pilot-scale, high-rate, completely-mixed digester (5.8 m3). The digestibility, biogas production and the impact of blending of liquid waste streams from a nearby pulp and paper mill were assessed. The 343-day evaluation was divided into 5 intensive evaluation stages. The organic removal efficiency was found to be 58 9% in terms of COD and 55 8% in terms of VS at a hydraulic retention time (HRT) of 11.6 3.8 days. The removal was comparable to those found in organic solid digesters (45e60%), but at a much shorter HRT. Methane yield was estimated to be 0.40 1 , higher than the typical range of other food wastes (0.11e0.42 m3e0.77 m3-CH4@STP kg-VSremoved 1 CH4@STP kg-VSremoved), with a mean methane content of 75% and <200 ppm of hydrogen sulfide in the biogas. The blending of selected liquid wastes from a paper mill at 10 vol% of brown grease slurry did not cause significant reduction in digester performance. Using a pseudo-first-order rate law, the observed degradation constant was estimated to be 0.10e0.19 d1 compared to 0.03e0.40 d1 for other organic solids. These results demonstrate that brown grease is a readily digestible substrate that has excellent potential for energy recovery through anaerobic digestion. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Brown grease Biogas Anaerobic digestion Renewable energy
1. Introduction Using biogas as an alternative source of energy is gaining more attention globally in recent decades [1,2]. There has been an increasing number of studies performed to evaluate the conversion of waste streams such as animal manure, municipal solid wastes, energy crops, municipal bio-solids and food wastes to biogas [2e5]. According to International Energy Association (IEA), there are over 9000 anaerobic digesters in operation using these materials to * Corresponding author. Department of Civil Engineering, Lamar University, Beaumont, TX 77710-10024, USA. Tel.: þ1 409 880 8761. ** Corresponding author. E-mail addresses: [email protected] (C.-J. Lin), [email protected] (P. Hart). 1 Current address: Oil & Gas Division, Conceptual & Engineering Services, Siemens Energy Inc., Houston, TX, USA. http://dx.doi.org/10.1016/j.renene.2014.01.046 0960-1481/Ó 2014 Elsevier Ltd. All rights reserved.
produce biogas [6]. For instance, w15% of organic wastes are being converted annually in Germany [7,8]. The practice of converting wastes to energy provides a two-fold benefit of environmental protection and energy recovery. Brown grease (BG) is a mixture consisting of trapped grease, sewage grease, and black grease collected in grease interceptors (traps) of restaurants and food industries [9]. In the United States, there are 1.84 million tons of BG produced every year [10]. Most collected BG eventually ends up in landfills. In the US, the landfill cost for BG varies by region and was up to US$110 per metric tonne in 2002 [10]. This results in a very high direct disposal cost. In addition, the moisture content in BG can lead to soil and water pollution, making the soil sterile and unable to support plant life [11]. Because of these drawbacks, the European Union enacted a general ban on landfilling organic waste in 2005 [11]. An earlier
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Nomenclature AD ALK BAL BG BMP COD CSTR EPA FAC FC GPD HRT
anaerobic digester total alkalinity balance tank brown grease biochemical methane potential chemical oxygen demand continuous stirred-tank reactor Environmental Protection Agency facultative tank foul condensate, a liquid waste stream from kraft mill process gallons per day hydraulic retention time
study suggested that 14 106 m3 of methane could be produced in the US annually by converting the generated BG into biogas [12]. This is a substantial amount of renewable bio-energy. Recovering the energy while eliminating the waste input to landfills yields both economic and environmental benefits [13]. Anaerobic digestion is a treatment process capable of producing biogas from organic wastes. The benefits of anaerobic digestion include smaller reactor size in terms of organic loading, lower air emissions and a smaller amount of generated sludge compared to aerobic biological treatment [14]. Greasy wastes such as BG have been added as a lipid-rich co-substrate in earlier anaerobic digestion studies for sewage sludge [15,16], municipal wastewater [17e19], and the digestible fraction of municipal solid wastes [20,21]. Typically it is blended at 2e50% of the primary substrate’s organic loading to improve the biogas yield and methane content [15e21]. However, higher lipid loading (>50% of the substrate) can cause long chain fatty acid (LCFA) inhibitions [15,16,22], scum and foam formation [23] and fat clogging problems [17,23]. To our knowledge, there are few studies devoted to investigating the degradability and biogas production using BG alone. Biochemical methane potential (BMP) data provide an estimate of biogas production from digesting an organic substrate anaerobically. Combining the BMP data and the associated kinetics of substrate degradation, the waste treatment efficiency and the cost benefits of an anaerobic digestion process can be optimized [24,25]. BMP measurements are typically carried out in lab-scale experiments. Although effective and easy-to-control, the estimated BMP can be up to 50% higher than the biogas production in full-scale digesters where operational constrains exists [26]. This leads to a large uncertainty in estimating the biogas production from commercial-scale operation. Investigating methane production in a pilot-scale operation can provide a more practical estimate of the energy values of digested substrates. The objective of this study is to employ a pilot-scale anaerobic digestion system to assess the methane yield and treatment efficacy of BG. Kinetic evaluation was performed to estimate the substrate utilization rate constant. Various liquid waste streams from an adjacent paper mill including foul condensate (FC) and screw press liquor (SPL) were blended as an effort to minimize the water use in the feed. The effects of process parameters including substrate composition, hydraulic retention time (HRT), organic loading rate (OLR) and operational changes on the system performance are reported.
LCFA OLR ORP SPL ST STP TN TP TS TSS VFA VS VSS
305
long chain fatty acid organic loading rate oxidation reduction potential screw press liquor, a liquid waste stream from paper mill sedimentation tank standard temperature and pressure (0 C, 1 atmosphere) total nitrogen total phosphorous total solids total suspended solids volatile fatty acids volatile solids volatile suspended solids
2. Materials and method 2.1. Pilot-scale system setup The high-rate anaerobic digestion system employed in this work comprises of three CSTRs and a clarifier: a balance tank (BAL), a facultative tank (FAC), an anaerobic digester (AD), and a final sedimentation tank (ST). The system schematic and the corresponding mixing and transferring pumps are presented in Fig. 1. The BAL and FAC are rectangular shaped tanks having an adjustable volume of 0.2e1.0 m3. AD is a cylindrical tank with a total volume of 7.6 m3 (1.6 m in diameter and 3.8 m in height) with adjustable reaction volumes of 4.3, 5.8 and 7.6 m3. The nominal reactor volume in this study was 5.8 m3. It has a Plexiglas window at the top for observing the mixed liquor in the digester. BAL and FAC were mixed with submerged mixing pumps. Mixing in AD was accomplished by re-circulating the digester liquor. The mixing was checked regularly by taking suspended solid measurements at different reactor heights to ensure a completely mixing condition. The sedimentation tank (1.5 m3) has a cylindrical shape with a conical bottom at 1:1 slope. The BG feedstock is a solid phase substrate with a high chemical oxygen demand (COD). Before feeding to BAL, the BG feedstock was re-suspended in tap water to form BG slurry (water:BG ¼ 1:1 in volume) using a hand-held mixing pump. The re-suspended BG slurry was then mixed with tap water and/or co-substrates in the BAL yielding a substrate concentration of 25,000e50,000 mg L1 COD. After this homogenization process, the substrate was pumped to FAC for pre-digestion. A commercial facultative culture powder (Meridian Bioenergy) was introduced (at 1:5000 mass ratio of bioculture to COD) to FAC for bio-augmentation at 12-h HRT. The purpose of this process is to initially introduce a series of microbial strains, degrade possible LCFA and to eliminate potential process inhibitors as reported earlier [15]. Afterwards, the pre-digested substrates were pumped continuously into AD for digestion. The AD effluent was transferred to ST gravitationally for sedimentation. 2.2. Evaluation schedule The evaluation period lasted for 343 days. Excluding the system start-up, maintenance and feeding transition periods, process data were collected for 238 days. The evaluation was divided into five intensive evaluation periods (IeV). During each operating period, a steady stage (S1 to S5, defined as a state with relatively consistent biogas production and organic removal) was selected for intensive
306
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 1. Schematic diagram of the pilot-scale high-rate anaerobic digester.
Table 1 Feeding characteristics and reactor configuration during the evaluation. Data were collected in five different periods for analysis.
Date Days of operation Days of intensive evaluationa Sedimentation tank Feeding Influent COD (mg L1) Influent VS (mg L1) OLRb (kg-VS m3d1) HRT (days) Activity a b
System start-up
I
II
III
IV
V
4/13/11e7/26/11 / /
7/27/11e8/7/11 1e12 1e12 (S1)
8/8/11e10/24/11 13e90 34e45 (S2)
10/25/11e12/7/11 91e135 107e133 (S3)
12/8/11e2/29/12 136e217 184e217 (S4)
3/1/12e3/21/12 218e238 218e238 (S5)
No BG / / / / Seeding and initiating
No BG 34,510 2557 13,965 1262 2.0 0.2 7.3 0.6 Establish BG steady state
Yes BG 56,570 3894 23,937 1625 2.7 0.3 (2.0 0.3) 8.9 0.9 Add ST
Yes BG þ FC 26,570 6264 10,139 754 1.0 0.2 (0.8 0.2) 15.2 1.1 Establish BG þ FC steady state
Yes BG 33,881 9176 13,224 3236 0.8 0.2 (0.6 0.2) 15.8 1.9 Back to BG steady state
Yes BG þ SPL 30,200 1503 13,225 1891 1.2 0.3 (0.9 0.3) 11.0 0.1 Establish BG þ SPL steady state
S1eS5 stands for five selected stages with intensive evaluation and stable data consistency. OLR and HRT were calculated based on AD only. The OLR in the parentheses for S2eS5 were based on AD þ ST.
measurement and data analysis. Table 1 summarizes the evaluation schedule and the corresponding operating parameters in each stage. During S1 (day 1e12, Table 1), the system was operated without ST and the AD effluent was considered as the final effluent. To understand the effect of solid recycle, ST was added to the system during S2 (day 13e90). During S3eS5 (day 91e238), the system flow follows the schematics in Fig. 1. The supernatant (overflow) of ST was the final effluent and the underflow was returned to AD. In S3 and S5, FC and SPL was fed respectively as co-substrates to reduce tap water use and to investigate their impact on the system performance. The S4 period was used to re-establish baseline conditions between the FC and SPL co-substrate additions. 2.3. Chemical analysis and biogas measurement Daily liquid samples were taken from five sampling points: BAL effluent, FAC effluent, AD effluent, ST underflow (return sludge) and ST overflow (final effluent). Gas samples were drawn from a gas sampling outlet at top of AD at least twice a day. Table 2 shows the analytical measurements and frequencies for each sample.
The pH, dissolved oxygen (DO), oxygen reduction potential (ORP) and temperature (T) were measured on site using a calibrated portable meter with appropriate probes (Hach HQ40D with pH101, LBOD101, and IntelliCAL ORP probes) after samples were drawn from the reactors. All laboratory measurements were performed within 2 h after sampling, therefore no sample preservation was performed. The total solids (TS), volatile solids (VS), total suspended solids (TSS) and volatile suspended solids (VSS) were measured according to EPA Standard Methods 2540 [27]. Dissolved parameters including dissolved COD (dCOD), total nitrogen (TN), total phosphorous (TP), volatile fatty acids (VFA), total alkalinity (ALK), sulfide and sulfate were measured after the samples were centrifuged (AccSpinÔ 400) for 30 min at 6000 rpm. These analyses were performed using EPA approved methods (Hach Company DOC 316.5, EPA 310.2, 350.1, 353.2, and 365.3, CFR 136.3, 141.5) with a time-lapse heating reactor (Hach DRB200) and a spectrophotometer (Hach DR3800). The cumulative volume of produced biogas was measured continuously using a digital gas flow meter (FILLRITE Model 820, TTS Corp., IN, USA). The composition of biogas was measured using a portable gas analyzer with an infrared detector (Model GEM 2000, Landtec Inc., Colton, CA, USA). Standard gases
Table 2 Analytical parameters and frequency for each sampling point. Sampling point
After BAL
After FAC
After AD
ST recycling sludge
ST up-flow (final effluent)
Biogas (gas point on AD)
Analytical frequencies Measurements
Daily pH, T, OLR, COD, VFA
Daily pH, T, DO, ORP, TN, TP, VFA, ALK, TS, VS, TSS, VSS, COD
Daily pH, T, DO, ORP, TN, TP, VFA, ALK, TS, VS, TSS, VSS, COD
Daily pH, T, TS, VS, TSS, VSS, COD
Daily pH, T, DO, COD
At least twice a day Flowrate, content of CH4, CO2, H2S, and O2
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
307
Table 3 Substrate characteristics. Brown grease was used as the primary substrate and the other two liquid wastes were used as co-substrates in part of the evaluation. Parameter
Brown Grease (BG)a (m s, n ¼ 17)
Foul condensate (FC) (m s, n ¼ 11)
Screw press liquor (SPL) (m s, n ¼ 13)
COD (mg L1) dCOD (mg L1) TS (mg L1) VS (mg L1) VS/TS ratio TSS (mg L1) VSS (mg L1) VSS/TSS ratio Alkalinity (mg L1 as CaCO3) pHb TN (mg L1)c TP (mg L1)c Sulfide (mg L1) Sulfate (mg L1) Moisture content (wt%)
910,634 229,993 / 437,778 91,348 372,111 77,646 0.85 0.06 / / / / 6.51 0.77 / / / / 56 9
2973 142 2740 125 406 104 210 14 0.53 0.1 357 58 339 46 0.83 0.25 205 50 9.28 0.18 52.2 4 0.24 0.09 52.2 20.5 <40 /
4498 2020 609 189 8768 7957 3742 1666 0.5 0.1 4048 1750 1997 875 0.49 0.06 / 8.44 0.83 2.3 0.1 0.41 0.04 / / /
a b c
Here BG stands for pre-treated brown grease in solid phase, thus the unit of COD, TS, and VS is mg kg1. pH of brown grease was measured by suspending 100 g brown grease in 1 L tap water that has a pH of 8.05 and alkalinity of 55 mg L1 as CaCO3. In aqueous phase.
(50% CO2, 50% CH4, dry air and 1000 ppm H2S) were used for the calibration of the gas analyzer daily. The methane yield of the substrates was estimated as the ratio of the volume of CH4 (at the standard temperature and pressure, STP) produced in the digester to the VS of the substrate digested (m3-CH4@STP kg-VS1 removed). During the five intensive evaluation periods (S1eS5, Table 1), analyses of four replicates of randomly selected samples for all the above-mentioned parameters typically had a relative standard deviation less than 5 % except for solids, which had larger analytical errors (<10%). The completeness of data used for process performance evaluation was over 90%.
2.4. Substrate characteristics The BG feedstock of this study was obtained daily from a waste management plant in Houston, TX, where the raw BG was screened to remove large particles, stabilized by lime addition, and then flocculated and dewatered. The processed BG is a light-brown, sticky, greasy solid with a perished food smell. In different stages of evaluation, it was emulsified into water at different concentrations (Table 1). After re-suspension in tap water, the slurry has a neutral pH range (6.5e7.5), suitable for biological treatment such as anaerobic digestion. No nutrients were added because of the
balanced nutrient composition of BG was deemed sufficient for biological treatment [28]. The characteristics of BG feedstock, FC, and SPL are shown in Table 3. Since the BG has an extremely high organic content (w1 kg-COD kg-BG1, Table 3), the feeding stream was diluted to the range of 25,000e50,000 mg L1 COD. FC and SPL have a relatively low COD level and solids content compared with BG (Table 3). In addition, their mild alkalinity (Table 3) effectively offset the mild acidity in BG. FC is a liquid substrate with relatively low solid content (TS ¼ 400 mg L1), its major organic content is in the dissolved phase (dCOD is >90% of total COD, Table 3). SPL has a TS content less than 1.0 wt%. Its dCOD concentration is <20% of total COD concentration (Table 3), indicated the major organic content is in the solid phase.
3. Results and discussion 3.1. Performance of anaerobic digester The operating parameters of AD and system performance during the five intensive study periods (S1eS5, Table 1) are summarized in Table 4. For comparison, the typical values for anaerobic digestion systems are also included. During the evaluation, the pH in AD was
Table 4 Anaerobic digestion operating parameters and system performance in five selected stages. Stages
I
II
III
IV
V
Typical rangea
pH T ( C) DO (mg L1) ORP (mV) TN (mg L1) TP (mg L 1) Alkalinity (mg L1 as CaCO3) VFA (mg L1 as HAc) COD removal efficiency (%) VS removal efficiency (%) CH4 content (%) CO2 content (%) H2S content (ppm) CH4 yield (m3-CH4 kg-VS1 removed)
7.34 0.05 36.0 0.7 0.01 0.00 209 14 591 83 3.4 2.4 3087 282 274 97 42.1 6.7 26.8 7.9b 74.3 2.0 22.3 1.3 38.2 4.1 0.40e0.49
/ 36.3 0.7 / 228 24 409 37 1.5 0.4 / / 50.6 5.8 37.1 4.3b 74.6 1.0 / / 0.58e0.77
7.12 0.08 34.3 1.8 0.06 0.04 243 40 237 74 0.9 0.4 1455 457 199 76 73.8 11.0 72.7 7.4 75.9 1.9 23.9 1.9 147.2 34.8 0.49
7.10 0.07 34.3 2.1 0.15 0.05 247 37 314 50 2.3 1.1 2478 291 394 84 61.7 12.3 57.9 13.2 74.6 1.8 25.2 1.8 185.2 28.1 0.48
7.01 0.17 37.9 1.0 0.10 0.03 263 23 306 46 2.2 0.4 2204 222 469 378 53.5 8.7 56.4 9.9 75.4 1.0 24.2 1.0 371.7 127.6 0.45
6.5e8.5 [52] 35e40 [52] / m400w150 [53] 60e1000 [54] 6e50 [54] 1500e5000 [55] <1800 [53] / / / / / 0.11e0.42 [3,24,44e50]
a Typical value of operating parameters including pH, T, ORP, TN, TP, VFA and alkalinity were based on the description of typical anaerobic digestion systems. Typical values of methane yield were based on earlier literature. b For comparison purpose, VS removal efficiency in S1 and S2 has not been corrected by biomass calculation.
308
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 2. VFA concentrations before and after AD. The alkalinity in AD is also shown.
controlled at neutral range (7.01e7.34, optimal pH 6.9e7.6 for anaerobic digestion) and the temperature was maintained in the mesophilic range (34.3e37.9 C). The low DO concentration (0.05e 0.20 mg L1) and ORP value (<200 mV) indicated strictly
anaerobic conditions were maintained during the evaluation. TN and TP concentrations in the system were 230e600 mg L1 as nitrogen and 1e4 mg L1 as phosphorous, indicating adequate nitrogen but possible phosphorus deficiency. VFA concentration in AD was generally lower than 500 mg L1 as acetic acid (HAc) except during the maintenance period (Fig. 2). The average total alkalinity was 2122 mg L1 as CaCO3, suggesting absence of in vitro VFA toxicity [29,30]. These operational parameters were controlled in the typical range of healthy digesters (Table 4). The pre-digestion process in FAC effectively eliminated the negative effects of the lipid-rich substrate. Earlier studies have reported that fat scum foaming by greasy substrates may be a concern [31,32]. However, no significant scum formation was observed in the AD during the evaluation. The surface fat scum was rapidly diminished within an hour in FAC [33], possibly because of the effectiveness of the pre-digestion process breaking down LCFAs that form scum. Perle et al. (1995) conducted a lab-scale experiment of dairy wastewater and noted a similar result [34]. The pre-digestion process also produced VFAs that accelerated the anaerobic digestion process. To verify the VFA production through the bio-augmentation process, a facultative experiment using diluted BG slurry (10,000 mg L1 COD) with the commercial facultative culture was conducted in a lab-scale reactor (0.5 L). Experimental results show that after 12-h of pre-digestion at room temperature, the pH was decreased from 7.03 to 5.55, and VFA concentration was increased from 148 to 293 mg L1 as HAc. In the pilot-scale system, the facultative pre-digestion efficiently
Fig. 3. Concentrations of (a) COD, and (b) VS before and after AD. Their removal efficiencies are shown on the secondary axis. The five selected stages (S1eS5) are marked.
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 4. Volatile fractions of (a) VS/TS and (b) VSS/TSS in FAC and AD.
augmented VFA generation for methanogenesis in AD. Fig. 2 shows the VFA level in FAC and AD as well as the total alkalinity in AD. The FAC effluent has about 2 times higher VFA concentrations (600e 1800 mg L1 as HAc, Fig. 2) compared to the FAC influent. The average VFA concentration decreased from w800 mg L1 in FAC to w413 mg L1 as HAc in AD (Fig. 2). Fig. 3 shows the COD and VS variation and removal efficiency during the five intensive evaluation periods (S1 to S5, Table 1). In S1, only BG was fed to the AD at a mean daily flow rate of 0.568 m3 d1 (150 gallon per day, GPD) without recycling of anaerobic sludge. The COD in the influent and effluent was w33,000 mg L1 and w20,000 mg L1 respectively. Since the organic content of effluent is too high for the subsequent aerobic treatment, a ST was added on day 13 to remove the solid content of the digester effluent and to increase the anaerobic biomass in AD with sludge recycling. In S2, 0.151 m3 d1 (40 GPD) of the anaerobic sludge was recycled to AD. The total daily flow was slightly reduced to 0.503 m3 d1 (133 GPD), i.e., the feeding stream was reduced to 0.341-0.378 m3 d1 (90e100 GPD). After the modification, the influent and effluent organic contents in S1 and S2 were compared. In S2, the overall feed concentration to AD was higher (45,000e 60,000 mg L1 COD) than S1 (30,000e40,000 mg L1 COD) because of the recycled sludge (Fig. 3). The mean influent and effluent VS concentrations were 13,965 and 9448 mg L1 in S1. In S2, due to the added organic load of recycled sludge, the mean influent VS
309
increased to 23,937 mg L1 and the effluent VS increased to 12,078 mg L1. During S1 and S2, the OLR and HRT were calculated based on the flow into and out of AD. Stage I used a lower concentration (w14 g-VS/L) and a higher flowrate (150 GPD), while Stage II used a higher concentration (w24 g-VS/L) and a lower flowrate (90e100 GPD). This resulted in w30% increased of OLR (from 2.0 to 2.7 kg VS m3 d1) in Stage II (Table 1). Since the feeding flow rate in S2 was reduced from 0.568 m3 d1 to 0.503 m3 d1, The HRT of S2 increased to 8.9 days compared with S1 (7.3 days, Table 1). The COD and VS removal of S2 was slightly higher than S1 (Table 4), possibly because of the increased amount of biomass in AD which improved the digester performance. Starting from day 91 (S3), the data analysis included both AD and ST as a digestion system. During S3eS5 (day 91e238), the influent COD concentration was kept at w26,500 mg L1 (Table 1) and the influent VS was in the range of w11,000 mg L1. The flow rate in S3 and S4 was reduced to 0.246e0.322 m3 d1 (65e85 GPD, HRT 15e16 days). The flow rate in S5 was maintained at 0.397 m3 d1 (105 GPD, HRT 11 days, Table 1). During these periods, the ST overflow had a relatively consistent COD and VS concentration (COD w 10,000 mg L1 and VS w 3500 mg L1, Fig. 3). This implies that, the effluent reached a stable level to be treated aerobically after ST. Similar results for palm oil mill effluent were reported by Basri et al. (2010) [35]. FC and SPL were introduced as co-substrates at 10 vol% during S3 and S5, respectively. The addition of the co-substrates did not impose significant impact upon the methane yield (Table 4). Compared to the periods feeding BG only (S4), the COD and VS removal efficiency of S3 increased from w60% to w74% (COD, Table 4 and Fig. 3) and w55% to w73% (VS, Table 4 and Fig. 3) respectively. This is probably because the organic content of FC is relatively easy to degrade (>80%) [36e38]. SPL did not appear to significantly affect the COD and VS removal efficiency (S5, Fig. 3). To understand the organic removal in the system at different stages, the volatile ratios of solids (VS/TS and VSS/TSS) before and after AD were monitored. The volatile ratio is an indicator of organic degradation and of inorganic accumulation since the inorganic mass fraction of solids is conserved during digestion [39]. Fig. 4 shows the VS/TS ratio (Fig. 4a) and VSS/TSS ratio (Fig. 4b) in FAC and AD, respectively. Before AD, the volatile ratio for BG was in the range of 0.70e0.90. After AD, the volatile ratio was reduced to the range of 0.60e0.70, suggesting the organic content of the feed has been effectively removed in the digester. During S3eS5, the addition of co-substrate also affected the volatile ratio, and the varying tendency was in accordance with COD removal efficiency (Figs. 3 and 4). During the entire operating period, the average COD and VS removal efficiency were 58% and 55% (Table 4), lower than those previous reported for lipid-rich wastewaters (w90%) [40] but comparable to those of typical solid digesters (45e60%) [41]. Fig. 5 shows the scatter plot between the organic removal and organic loading rate in terms of VS (Fig. 5a) and COD (Fig. 5b). During S1 and S2, the COD and VS removal efficiency were not significantly affected by OLR variation, resulting in a linear increase of VS and COD mass digestion with respect to the applied OLR (Fig. 5). However, the organic removal efficiency in S2 was greater than that in S1, as reflected by the change of COD and VS in the feed and digester effluent (Fig. 3). S3 to S5 had the similar trend of linear increase of digested COD and VS mass with respect to the OLR, indicating that the system should have a higher treatment capacity than the OLR range applied during the evaluation. The daily biogas production during the evaluation is summarized in Fig. 6. In S1, the biogas production was 5e6 m3 d1. The biogas production was higher in S2 (w7 m3 d1) because of the
310
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
Fig. 5. Scatter plots between organic loading rate (OLR) and organic removal in terms of (a) VS and (b) COD. The points in the frames show the data during S1 and S2 when the operational parameters were different from those in S3-S4.
higher organic load and recycled biomass (Fig. 5a). During S3eS5, the average daily biogas production was lower than S1 and S2 since the system OLR was reduced (Table 1). In S3, the COD removal efficiency was higher than S4 and S5 (Table 4), leading to higher biogas production (w5.6 m3 d1) compared to that in S4 and S5 (w3.5 m3 d1). The easily digested dCOD in FC accounted for the higher biogas production in S3. The lower biogas production in S4 and S5 (as compared to that of S3) was caused by the lower OLR applied (S4, Table 1) and the slightly lower organic removal (S5, Table 4), respectively. Generally, the biogas production trend in S3e S5 was consistent with the organic removal shown in Fig. 3, suggesting that the biogas production was not significantly affected by the addition of co-substrate. 3.2. Methane yield and kinetic analysis The produced biogas had a consistently high CH4 content (w75%, Table 4). The other major gas (CO2) consisted of the other w25% by volume (Table 4) and trace gases (e.g., H2S). From day 160e175, the system was recovered from system maintenance and the methane content built up from 40% to 75% quickly (Fig. 6). During the entire evaluation, the average H2S concentration was 189 ppm, significantly lower than the level that may cause H2S toxicity (w1500 ppm) [42]. The cumulative CH4 production and digested VS in S3 to S5 are shown in Fig. 7. The methane yields of S3 to S5 were calculated as the ratio of the two slopes. The value was reported based on VS
removal because the organic content of the BG feed was mainly in the suspended solid phase. The methane yield of BG in S3 to S5 was consistently in the range of 0.45e0.49 m3-CH4 kg-VS1 removed (All gas volumes mentioned hereafter have been normalized to STP). For the first two stages (S1 and S2), the apparent VS removal efficiency (27e37%, Table 4) was significantly lower than in S3eS5 (55e75%, Table 4) because ST had not been introduced to the system. Under this condition, the effluent VS during S1 and S2 contains substantial amount of biomass produced from the anaerobic digestion of brown grease. During S3 to S5, ST was used to collect and recycle most of the generated biomass back to the AD, resulting in the higher apparent BG removal efficiency. To estimate the BG conversion into biogas during S1 and S2, a mass balance on solids before and after the AD was performed as followed:
ð1 f ÞF ¼ ð1 aÞX þ ð1 bÞY
(1)
aX þ bY ¼ M
(2)
Eq. (1) represents the mass balance of the fixed (inorganic) solids. Where f is the volatile fraction of influent BG substrate obtained from measurement (0.808 in S1 and 0.816 in S2), F is the mass flow of influent total solid (kg d1). a and b are the volatile fraction (VS/TS) of biomass and undigested BG substrate, respectively (a ¼ w0.80 [43]). X and Y are the mass flow of biomass and undigested BG (kg d1). Eq. (2) represents the VS composition in
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
311
the two lines in the respective period. The estimated methane yield in S1 (0.40e0.49 m3-CH4 kg-VS1 removed) was comparable to the yield in S3 to S5 (0.45e0.49 m3-CH4 kg-VS1 removed). S2 has a higher methane yield (0.58e0.77 m3-CH4 kg-VS1 removed) because of the shorter hydraulic retention time and higher organic loading (Table 1) in the presence of recycled biomass. Under such condition, only the readily degradable fraction of the substrate in the feed was digested, leading to a higher methane yield and lower organic removal efficiency (Table 4). In S3eS5, the lowered organic loading and higher HRT improved the organic removal at the cost of reduced methane production. In practice, the mode of process operation will depend on the treatment objective (better organic removal or higher methane yield). Also, the added co-substrate (FC and SPL) did not adversely affect the methane yield during S3eS5 (Table 4). A pseudo-first-order kinetic model was applied to analyze the substrate utilization. Similar approaches have been used earlier [3,24,44e50]. The substrate concentration was calculated based on VS. For a CSTR at steady state, the effluent concentration (C) can be estimated as:
Fig. 6. Measured daily biogas production and CH4/CO2 content.
C ¼
Fig. 7. Cumulative CH4 production at STP and cumulative VS digested during five selected stages (S1eS5). The slopes of each linear stage were used to calculate corresponding CH4 yield. In S1 and S2, the mass of digested VS was corrected by biomass calculation.
the effluent, where M is the mass flow of VS in the effluent (kg d1). Using the solid measurements, we estimated that the generated biomass constitutes 25e40 wt% in the effluent. Based on the mass balance results, the cumulative CH4 production and the digested VS during S1eS5 were plotted in Fig. 7. The methane yield was then calculated as the ratio of the slopes of
C0 1 þ kq
(3)
where C0 is the influent substrate concentration (mg L1 VS), k is the first-order substrate utilization rate constant (day1) and q is the HRT (d). The estimated k value was in a relatively consistent range of 0.10e0.19 d1 throughout the evaluation process. For comparison, previously reported methane yields of food wastes and their first-order kinetic parameters are shown in Table 5 [3,24,44e50]. Generally, the degradation rate constants were in the range of 0.03e0.4 d1. The rate constant obtained in this study (0.10e0.19 d1) had probably been adversely affected by the greater difficulty of controlling the digestion conditions (temperature and mixing) in a pilot-scale system due to the ambient temperature variation (>15 C diurnal change). It was slightly lower than that of municipal solid sludge in batch reactors (0.2e0.4 d1, [51]), comparable to that of municipal solid sludge in CSTR (0.175 d1, [44]), and higher than that of canary grass in CSTR (0.03e0.04 d1, [3]). The measured methane yield of BG in this pilot study ranged from 0.40 to 0.77 m3-CH4 kg-VS1 removed. Because of the higher lipid content of BG, the value is substantially higher than the methane yields of typical sugary and sludge substrates (0.11e0.42 m3-CH4 kgVS1 removed, Table 5). The biogas quality produced by BG is excellent (w75%, Table 4), possibly also due to the high lipid content of BG. These pilot-plant data suggest that BG can be effectively digested anaerobically for high quality biogas production.
Table 5 Comparison of reported and calculated first order degradation rate constants and methane yields. Substrate
Reactor type
Pseudo-first-order rate constant (d1)
Methane yield (m3-CH4 kg-VS1 removed)
Reference
Brown grease Municipal sludge Municipal sludge Corn stover Rice straw Canary grass Sunflower oil cake Winter wheat Waste activated sludge Waste activated sludge þ fatty wastewater Synthetic kitchen waste Synthetic kitchen waste þ municipal grease waste
CSTR CSTR Batch Batch
0.10e0.19 0.175 0.2e0.4 / / 0.03e0.04 / / / / / /
0.40e0.77 0.309 / 0.239 0.225 0.19e0.33 0.107e0.227 0.311e0.360 0.116 0.362 0.117 0.324e0.418
This study [45] [44] [49] [50] [3] [46] [47] [48] [48] [24] [24]
CSTR Batch Batch Batch Batch Batch Batch
312
P. Zhang et al. / Renewable Energy 68 (2014) 304e313
4. Conclusions In this study, the anaerobic digestion of waste brown grease (BG) in a pilot-scale system was investigated. A mean COD removal of 58% and a mean VS removal of 55% were achieved. The organic removal efficiency was comparable with those found in typical solid digesters. Kinetic analysis showed that the pseudo-first-order degradation rate constant of BG was in the rage of 0.10e0.19 d1. After anaerobic treatment process, the effluent had a consistent effluent organic strength (COD w 10,000 mg L1) that can be treated aerobically. It was concluded that BG was a readily digestible substrate as a sole substrate. The pilot-scale system produced biogas of excellent quality (75% CH4 content), with a methane yield in the range of 0.40e0.77 m3CH4 kg-VS1 removed. The addition of paper mill waste streams (fouls condensate and screw press liquor) as the co-substrates did not adversely affect the methane yield. BG has the industrial potential to be anaerobically treated as an energy feedstock and there has been ongoing commercial effort to build large-scale digesters using BG as the primary substrate. Using BG for biogas production could serve as a profitable model for converting waste to renewable energy.
Acknowledgment This study was supported by MeadWestvaco Evadale TX facility (Project No: MWV0001). The authors would like to thank Dipendra Wagle, Sophia Yang, Yolanda Wang, Brandon Corace and Erik Corace for their assistance in the field work and laboratory analysis. The assistance of Gary Colson, Michael Clapper and Robert Sasser in supporting this work and in obtaining and supplying paper mill samples is greatly appreciated. The administrative support of Stewart Cairns, Reid Sweet and Thomas Sitton are also acknowledged.
References [1] Hilkiah Igoni A, Ayotamuno MJ, Eze CL, Ogaji SOT, Probert SD. Designs of anaerobic digesters for producing biogas from municipal solid-waste. Appl Energy 2008;85:430e8. [2] Nges IA, Escobar F, Fu X, Bjornsson L. Benefits of supplementing an industrial waste anaerobic digester with energy crops for increased biogas production. Waste Manag 2011;32:53e9. [3] Kacprzak A, Krzystek L, Pa zdzior K, Ledakowicz S. Investigation of kinetics of anaerobic digestion of Canary grass. Chem Papers 2012;66:550e5. [4] Sivakumar P, Bhagiyalakshmi M, Anbarasu K. Anaerobic treatment of spoiled milk from milk processing industry for energy recovery e a laboratory to pilot scale study. Fuel 2012;96:482e6. [5] Lay J-J, Lee Y-J, Noike T. Feasibility of biological hydrogen production from organic fraction of municipal solid waste. Water Res 1999;33:2579e86. [6] Internation Energy Association. Country Reports of Task 37: Energy from Biogas. http://www.iea-biogas.net/country-reports.html; November 2013 [accessed December 2013]. [7] Gallert C, Henning A, Winter J. Scale-up of anaerobic digestion of the biowaste fraction from domestic wastes. Water Res 2003;37:1433e41. [8] Weiland P. Anaerobic waste digestion in Germany e status and recent developments. Biodegradation 2000;11:415e21. [9] Burgess MN. What to do with brown grease, plumbing systems & design; 2010. [10] Tyson KS. Brown grease feedstocks for biodiesel. National Renewable Energy Laboratory; 2002. [11] Murto M, Bjornsson L, Mattiasson B. Impact of food industrial waste on anaerobic co-digestion of sewage sludge and pig manure. J Environ Manage 2004;70:101e7. [12] Dichtl N. Thermophilic and mesophilic (two stage)anaerobic digestion. J Charter Inst Water Environ Manage 1997;11:98e104. [13] Dearman B, Bentham RH. Anaerobic digestion of food waste: comparing leachate exchange rates in sequential batch systems digesting food waste and biosolids. Waste Manage 2007;27:1792e9. [14] Gujer W, Zehnder AJB. Conversion processes in anaerobic digestion. Water Sci Technol 1983;15:127e67.
[15] Noutsopoulos C, Mamais D, Antoniou K, Avramides C. Increase of biogas production through co-digestion of lipids and sewage sludge. Global Nest J 2012;14:133e40. [16] Neczaj E, Bien J, Grosser A, Worwag M, Kacprzak M. Anaerobic treatment of sewage sludge and grease trap sludge in continuous co-digestion. Global Nest J 2012;14:141e8. [17] Long JH, Aziz TN, Reyes DL. Anaerobic co-digestion of fat, oil and grease (FOG): a review of gas production and process limitations. Process Safety Environ Protect 2012;90:231e45. [18] Bouchy L, Perez A, Camacho P, Rubio P, Silvestre G, Fernandez B, et al. Optimization of municipal sludge and grease co-digestion using disintegration technologies. Water Sci Technol 2012;65:214e20. [19] Wan C, Zhou Q, Fu G, Li Y. Semi-continuous anaerobic co-digestion of thickened waste activated sludge and fat, oil and grease. Waste Manag 2011;31: 1752e8. [20] Martin-Gonzalez L, Castro R, Pereira MA, Alves MM, Font X, Vicent T. Thermophilic co-digestion of organic fraction of municipal solid wastes with FOG wastes from a sewage treatment plant: reactor performance and microbial community monitoring. Bioresour Technol 2011;102:4734e41. [21] Martin-Gonzalez L, Colturato LF, Font X, Vicent T. Anaerobic co-digestion of the organic fraction of municipal solid waste with FOG waste from a sewage treatment plant: recovering a wasted methane potential and enhancing the biogas yield. Waste Manag 2010;30:1854e9. [22] Zhu Z, He Q. Enhancing biomethanation of municipal waste sludge with grease trap waste as a co-substrate. Renew Energy 2011;36:1802e7. [23] Valladao ABG, Cammarota C, Freire DMG. Performance of an anaerobic reactor treating poultry abattoir wastewater with high fat content after enzymatic hydrolysis. Environ Eng Sci 2011;28:299e307. [24] Li C, Champagne P, Anderson BC. Evaluating and modeling biogas production from municipal fat, oil, and grease and synthetic kitchen waste in anaerobic co-digestions. Bioresour Technol 2011;102:9471e80. [25] Angelidaki I, Alves M, Bolzonella D, Borzacconi L, Campos JL, Guwy AJ, et al. Defining the biomethane potential (BMP) of solid organic wastes and energy crops: a proposed protocol for batch assays. Water Sci Technol 2009;59:927e 34. [26] Bishop GC, Burns. RT, Shepherd. TA, Moody. LB, Gooch. CA. Evaluation of laboratory biochemical methane potentials as a predictor of anaerobic dairy manure digester biogas and methane production. In: ASABE International Meeting, Reno, NV; 2009. [27] APHA, AWWA, WEF. Standard methods for the examination of water and wastewater; 1998. Washington, DC, USA. [28] Han S-K, Shin H-S. Biohydrogen production by anaerobic fermentation of food waste. Int J Hydrogen Energy 2004;29:569e77. [29] Dong J, Zhao YS, Hong M, Zhang WH. Influence of alkalinity on the stabilization of municipal solid waste in anaerobic simulated bioreactor. J Hazard Mater 2009;163:717e22. [30] Vanhaandel AC. Influence of the digested cod concentration on the alkalinity requirement in anaerobic digesters. Water Sci Technol 1994;30:23e34. [31] de Nardi IR, Fuzi TP, Del Nery V. Performance evaluation and operating strategies of dissolved-air flotation system treating poultry slaughterhouse wastewater. Resour Conserv Recycl 2008;52:533e44. [32] Alexandre VMF, Valente AM, Cammarota MC, Freire DMG. Performance of anaerobic bioreactor treating fish-processing plant wastewater prehydrolyzed with a solid enzyme pool. Renew Energy 2011;36:3439e44. [33] Santos RB, Hart PW, Colson GW, Evers S, Evers D. Commissioning of a biogas pilot plant: from brown grease to paper mill-generated organic wastes. Tappi J 2012;11:73e8. [34] Perle M, Kimchie S, Shelef G. Some biochemical aspects of the anaerobic degradation of dairy waste-water. Water Res 1995;29:1549e54. [35] Basri MF, Yacob S, Hassan MA, Shirai Y, Wakisaka M, Zakaria MR, et al. Improved biogas production from palm oil mill effluent by a scaled-down anaerobic treatment process. World J Microbiol Biotechnol 2010;26:505e14. [36] Zhou W, Imai T, Ukita M, Li F, Yuasa A. Effect of limited aeration on the anaerobic treatment of evaporator condensate from a sulfite pulp mill. Chemosphere 2007;66:924e9. [37] Norrman J, Narbuvold R, Nystrom L. Anaerobic treatability of waste water from pulp and paper industries. Biotechnol Adv 1984;2:329e45. [38] Yang MI, Edwards EA, Allen DG. Anaerobic treatability and biogas production potential of selected in-mill streams. Water Sci Technol 2010;62:2427e34. [39] Zhang R, El-Mashad HM, Hartman K, Wang F, Liu G, Choate C, et al. Characterization of food waste as feedstock for anaerobic digestion. Bioresour Technol 2007;98:929e35. [40] Lansing S, Martin JF, Botero RB, da Silva TN, da Silva ED. Wastewater transformations and fertilizer value when co-digesting differing ratios of swine manure and used cooking grease in low-cost digesters. Biomass Bioenergy 2010;34:1711e20. [41] Gunaseelan VN. Anaerobic digestion of biomass for methane production: a review. Biomass Bioenergy 1997;13:83e114. [42] Chen Y, Cheng JJ, Creamer KS. Inhibition of anaerobic digestion process: a review. Bioresour Technol 2008;99:4044e64. [43] Eddy MA. Wastewater engineering treatment and reuse. Mcgraw-Hill Higher Education; 2003. [44] Bolzonella D, Fatone F, Pavan P, Cecchi F. Anaerobic fermentation of organic municipal solid wastes for the production of soluble organic compounds. Indust Eng Chem Res 2005;44:3412e8.
P. Zhang et al. / Renewable Energy 68 (2014) 304e313 [45] de la Rubia Á, Pérez M, Sales D, Romero LI. Municipal sludge degradation kinetic in thermophilic CSTR. AIChE J 2006;52:4200e6. [46] Raposo F, Borja R, Rincon B, Jimenez AM. Assessment of process control parameters in the biochemical methane potential of sunflower oil cake. Biomass Bioenergy 2008;32:1235e44. [47] Rincon B, Banks CJ, Heaven S. Biochemical methane potential of winter wheat (Triticum aestivum L.): influence of growth stage and storage practice. Bioresour Technol 2010;101:8179e84. [48] Carrere H, Rafrafi Y, Battimelli A, Torrijos M, Delgenes JP, Motte C. Improving methane production during the codigestion of waste-activated sludge and fatty wastewater: impact of thermo-alkaline pretreatment on batch and semicontinuous processes. Chem Eng J 2012;210:404e9. [49] Xu F, Shi J, Lv W, Yu Z, Li Y. Comparison of different liquid anaerobic digestion effluents as inocula and nitrogen sources for solid-state batch anaerobic digestion of corn stover. Waste Manage 2013;33:26e32.
313
[50] Song Z, Yang G, Han X, Feng Y, Ren G. Optimization of the alkaline pretreatment of rice straw for enhanced methane yield. Biomed Res Int; 2013. [51] Rincon B, Sanchez E, Raposo F, Borja R, Travieso L, Martin MA, et al. Effect of the organic loading rate on the performance of anaerobic acidogenic fermentation of two-phase olive mill solid residue. Waste Manage 2008;28: 870e7. [52] Khanal SK. Anaerobic biotechnology for bioenergy production principles and applications. 1st ed. Hoboken, NJ: Wiley-Blackwell; 2008. [53] Gray NF. Water technology: an introduction for environmental scientists and engineers. 3rd ed.. Oxford: Elsevier; 2010. [54] Alley ER. Water quality control handbook. 2nd ed.. Colombus, OH: WEFþMcGraw-Hill; 2007. [55] Tchobanoglous G, Burton F, Stensel H. Wastewater engineering: treatment and reuse. 4th ed. Columbus, OH: McGraw-Hill; 2002.