Waste Management 75 (2018) 236–250

Contents lists available at ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

Experimental and feasibility assessment of biogas production by

anaerobic digestion of fruit and vegetable waste from Joburg Market
S.O. Masebinu a,⇑, E.T. Akinlabi a, E. Muzenda b, A.O. Aboyade c,d, C. Mbohwa c
a
Department of Mechanical Engineering Science, University of Johannesburg, South Africa
b
Department of Chemical Engineering Technology, University of Johannesburg, South Africa
c
Department of Quality and Operations Management, University of Johannesburg, South Africa
d
United States Agency for International Development/Southern Africa, Pretoria, South Africa

a r t i c l e i n f o a b s t r a c t

Article history: Substrate-induced instability of anaerobic digestion from fruit and vegetable waste (FVW) results in low
Received 11 October 2017 biogas yield. In this study, substrate management through fruit to vegetable mix ratio in a two-stage
Revised 31 January 2018 semi-continuous digester was investigated as a pathway for optimality of yield. The experiment con-
Accepted 6 February 2018
ducted over 105 days with 62.52 kg of FVWs sourced from Joburg Market, South Africa showed that a
Available online 2 March 2018
stable process was achieved at a fruit to vegetable waste mix ratio of 2.2:2.8. At this ratio, optimal organic
loading rate ranged between 2.68 and 2.97 kg VS/m3-d which resulted in a specific biogas yield of 0.87
Keywords:
Nm3/kg VS with 57.58% methane on average. The results of the experimental study were used as a fea-
Anaerobic digestion
Techno-economics
sibility assessment for a full-scale 45 tonnes/d plant for Joburg Market considering three energy path-
Fruit waste ways. The plant will produce 1,605,455 Nm3/y of biogas with the potential for offsetting 15.2% of the
Mono- digestion Joburg Market energy demand. Conversion of all biogas to biomethane was the most economically attrac-
Waste management tive energy pathway with a net present value of $2,428,021, an internal rate of return of 16.90% and a
Biogas simple payback period of 6.17 years. This route avoided the greenhouse gas emission of 12,393 tonnes
CO2, eq. The study shows that the anaerobic digestion of FVWs as sole substrate is possible with financial
and environmental attractiveness.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction
An amalgamation of several factors related to rapid urbanisa-
tion within the City of Johannesburg (CoJ) in South Africa has made
the efficient management of its municipal solid waste (MSW) chal-
lenging. The challenge is expected to be exacerbated since MSW
generation is projected to rise from 1.6 to 3.6 million tonnes/year
should its population reach 9.2 million by 2040 (Aurecon South
Africa (Pty) Ltd, 2015). A study in 2016 revealed that the organic
percentage of MSW collected from households, restaurants and
Joburg Market -a fruit and vegetable market wholly owned and
operated by Johannesburg Metropolitan Municipality- was 34%,
14% and 93% by weight respectively (Masebinu et al., 2017b). The
current management system for MSW disposal is landfilling. How-
ever, CoJ’s landfills are currently estimated to have between 5 and
11 years lifespan. As the MSW generation rate increases (e.g. from
between 0.35 and 1.6 kg/capita/day in 2003 to between 0.92 and
1.91 kg/capita/day in 2015 (Aurecon South Africa (Pty) Ltd, 2015;
⇑ Corresponding author.
E-mail address: masebinity@gmail.com (S.O. Masebinu).
Ogola et al., 2011)), it can be reasonably expected that the pro-
jected lifespan may be impacted. The establishment of new land-
fills within and around CoJ’s metropolitan area is an unlikely
option because of a lack of suitable and available land (City of
Johannesburg, 2011), which suggest the urgent need to consider
other options for its MSW disposal.
As part of an integrated MSW management plan, the CoJ is
interested in anaerobic digestion (AD) technology for the treat-
ment of highly biodegradable wet organic MSW, for example, fruit
and vegetable waste (FVW) from Joburg Market (JM). JM receives
farm produce from about 5000 farmers across South Africa gener-
ating an average of 47 tonnes/d of FVWs which are sent to the
landfill. The disposal of FVWs in landfills come with specific chal-
lenges, most of which are related to high moisture content and a
high degree of perishability (Scano et al., 2014). The fast degrada-
tion rate of FVW leads to the generation of greenhouse gases
(GHG), specifically methane (CH4) and carbon dioxide (CO2)
(Asquer et al., 2013). Aside from the climate change impact, dis-
posal cost and sale losses due to the quantity of FVWs generated
directly impacts the operating cost of the wholesale market. On
the other hand, the physicochemical characteristics of FVW make

https://doi.org/10.1016/j.wasman.2018.02.011
0956-053X/Ó 2018 Elsevier Ltd. All rights reserved.
 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
Nomenclature
Acronyms
Meaning
V, V v volume of biogas [Nl/d],[Nm3/h]
T BG ambient temperature [°C]
PBG ambient pressure [N/m2]
Nc number of tips per day [count/d]
Vc volume of biogas per tip [l/count]
V pH5 titration volume to pH 5 [ml]
V pH4:4 titration volume to pH 4.4 [ml]
V dig volume of digestate [ml]
NH2 SO4 acid normality [M]
M CaCO3 molar mass of CaCO3
Q dig total thermal energy losses [kW]
r radius of digester [m]
T o;min minimum outside temperature [°C]
hair;forced forced air convention [W=m2
C
t insulation thickness of insulation [mm]
K ps heat conduction coefficient of insulation [W=m2
C
it particularly suitable as feedstock for AD (Bouallagui et al., 2003;
Wu et al., 2016). The choice of AD inherently introduces a multi-
objective approach to energy generation, creation of secondary
raw material and reduction of environmental burden.
AD is a biological method that can be used to reduce and sta-
bilise organic matter of various types in an oxygen-starved envi-
ronment by a consortium of bacteria and archaea while
simultaneously producing an energy carrier called biogas
(Schnürer et al., 2017; Wang et al., 2017). Biogas is mainly com-
posed of CO2 and CH4 and finds application in heat, electricity gen-
eration and as a vehicular fuel (Parajuli et al., 2014; Shen et al.,
2015). Biogas generated from FVWs of JM can offset part of the
22,431 MW h/year of electricity consumed at the cost of $2.5 mil-
lion/year (Market, 2016). Aside biogas, a sludge-like by-product
called digestate, that is potentially useful for agronomic applica-
tions, is also produced (Mangwandi et al., 2013). It can, therefore,
be summarised that AD technology proffers energy, environmen-
tal, waste management and socio-economic benefits (Amoo and
Fagbenle, 2013; Fallde and Eklund, 2015; Mangwandi et al.,
2013; Olsson and Fallde, 2015; Tan et al., 2015). Although AD offers
many benefits, the process presents some limitations mainly due
to the duration for stabilisation of organic waste, slow degradation
of volatile solids, fluctuating biogas yield rate and substrate-
induced process instability (Khalid et al., 2011; Pavi et al., 2017).
AD process stability and biogas yield rate are strongly related to
the source and composition of the substrate, process and opera-
tional conditions and partly, the type of digestion technology used
(Muhammad Nasir et al., 2012; Raposo et al., 2011; Ratanatamskul
et al., 2015). The extent of stability of an AD process can be
adversely affected by the excessive accumulation of volatile fatty
acid (VFA) and ammonia. The accumulation of VFA results in acid-
ification of the AD process thereby inhibiting the methanogenic
activity (Asquer et al., 2013; Scano et al., 2014; Wu et al., 2016).
AD of FVW as the sole substrate is challenging due to the tendency
to accumulate VFA caused by the degradation of highly soluble
simple sugar (Lebuhn et al., 2014; Wang et al., 2014).
Several studies have reported AD instability resulting from
digesting FVWs as sole substrate (Alkanok et al., 2014;
Bouallagui et al., 2003; Bouallagui et al., 2005; Sitorus et al.,
2013). Maile et al. (2016) terminated the digestion of JM’s FVW
after four days due to acidification problem encountered during
the batch experiment. A conventional approach for managing the
237
hwater;free free water convection [W=m2
C]
t pv c thickness of digester dome cover [mm]
K pv c heat conduction coefficient of dome cover [W=m2
C]
hair;free free air convection [W=m2
C]
K soil heat conduction of soil [W=m2
C]
Q sub thermal energy to raise substrate to digester tempera-
ture [kW]
msub mass of substrate [kg/d]


C specific specific heat capacity of substrate [kJ=kg C]


T dig digester operating temp [ C]
T sub substrate delivery temperature [
C]
CF t annual after-tax cash flow [$]
i discount rate [%]
t time [years]
CC CHP capital cost of the CHP [$]
CC BM capital cost of biogas scrubber [$]
Pp installed power of CHP [kW]
n project life [years]
instability caused by rapid acidification is the application of alka-
line chemicals to adjust the AD pH to levels between 6.8 and 7.2,
thereby neutralising the acidification process (Gao et al., 2015).
Another approach is to make use of two-stage digestion that pro-
motes different types of oxidation and reduction reactions, pH
optima and growth rate of acidogens and methanogens to increase
process efficiency (Ganesh et al., 2014; Ward et al., 2008; Wu et al.,
2016). The two-stage approach has the advantage of hydrolysing
the substrate and buffering organic loading rate (OLR) in the first
stage. In the second stage, the hydrolysed and homogenised sub-
strate can be easily digested by the methanogens. Co-digestion of
different substrates provides buffering capacity to the digester
(Fonoll et al., 2015; Ganesh et al., 2013; Wang et al., 2014). It also
increases the biodiversity of the microbial community that facili-
tates the AD process (Wang et al., 2018). The microbial population
in the digester and the biogas yield is partly dependent on the ratio
of the different types of waste within the substrate mix (Lin et al.,
2012). Pavi et al. (2017) reported higher biogas yield and process
stability during co-digestion of food waste with FVWs at a ratio
of 1:3 compared to mono-digestion. Huang et al. (2016) found
optimal mix ratio at 3:1 of Aloe vera waste to dairy manure. Li
et al. (2015) found a ratio of rice husk to pig manure of 1:1 as opti-
mum. Molinuevo-Salces et al. (2012) reported a 1:1 mix ratio as
optimal for AD of vegetable waste with pig manure. Furthermore,
Wang et al. (2018) observed that increasing fraction of FVWs by
5% during co-digestion increased CH4 yield by 22.4%. Cabbai et al.
(2013) and Koch et al. (2016) showed that co-digestion could
increase CH4 yield rate by 18–48%. Aside co-digestion, manage-
ment of total solids (TS) concentration in the substrate has been
reported to improve the performance of digesters treating FVWs.
Bouallagui et al. (2003) found that a substrate with TS concentra-
tion <6% wet weight favours process stability but at the expense
of high biogas yield. Also, Alkanok et al. (2014) digested FVWs with
a 5% TS concentration and reported high volatile solid (VS)
degradation.
Thermal and chemical pre-treatment is another viable way to
reduce tendencies of instability when digesting FVWs as reported
by Dahunsi et al. (2016). Other stabilisation approach includes
hydraulic retention time management (Bouallagui et al., 2003; Di
Maria et al., 2015), organic loading rate (Aboudi et al., 2015; Di
Maria et al., 2015), digestate recycling (Zuo et al., 2015), elemental
composition adjustment (Wang et al., 2012; Wang et al., 2013),
238 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
substrate management with artificial intelligence (Jacob and
Banerjee, 2016), microbes addition (Yan et al., 2014), biochar addi-
tion (Shen et al., 2017) and dosing of trace elements (Jiang et al.,
2012; Zhang et al., 2012).
The conventional process related routes for managing AD insta-
bility have their disadvantages as well. The depletion of the neu-
tralising alkali chemical results in the increase of process
acidification. The use of two-stage digestion does not guarantee
process stability as increasing OLR to above 2.5 gVS/l-d for food
waste has been reported to lead to process instability (Shen
et al., 2013). The practical implementation of co-digestion requires
that all needed substrates are available at all times and also the
right mixture must be fed into the digester (García-Gen et al.,
2014; Rao and Baral, 2011). Though co-digestion and two-stage
digestion are known for performance improvement (Di Maria and
Barratta, 2015), it is possible that optimal substrate management
can produce enough buffering capacity that reduces or eliminate
the need for alkali chemical and co-digestion. Scano et al. (2014)
emphasised that feeding a balanced blend of FVW can guarantee
efficient control of the acidification processes that often lead to
process instability. In this paper, therefore, the impact of digesting
FVWs as a single substrate at different ratios of fruit to vegetable
waste was investigated with a view toward finding the optimal
mix that will limit the need for other additives while still guaran-
teeing process stability and high biodegradability. The potential of
utilising JM’s FVWs for energy generation through continuous AD
has not been investigated. Studies on the potential energy that
can be recovered from FVW in South Africa is limited. Therefore,
the recoverable energy from digesting JM’s FVWs was investigated
through a techno-economic feasibility assessment to provide pri-
mary data on a regional level.
2. Material and methods
Designing and operating a full-scale AD plant with FVW as a
sole substrate will require extensive knowledge of the perfor-
mance of such plant. Particularly, how the continuous loading of
a highly heterogeneous mixture of FVW impacts on the stability
of the process, biogas yield and specific methane yield rate.
According to Karellas et al. (2010), such information is difficult to
capture by a batch biochemical methane potential (BMP) test. To
obtain the information of interest, a two-stage laboratory-scale
AD process was designed, fabricated and operated in a semi-
continuous mode. Scano et al. (2014) also designed and fabricated
a micro-pilot plant digesting solely FVWs.
To further investigate the potentially recoverable energy from
the sole digestion of FVWs at a full-scale plant, assess the cost
implication for the construction and economic viability for the util-
isation of the biogas and digestate, a techno-economic study was
conducted. The OLR, biogas yield rate, HRT and specific methane
yield rate experimental data and other data source from literature
and equipment suppliers were used as input data for the full-scale
AD plant techno-economic assessment.
2.1. Feedstock sampling, preparation and characterisation
Source separated FVWs were collected once in two weeks over
the first two months due to logistic reasons and subsequently on a
weekly basis from JM over the duration of the experiment. Random
sampling was used to collect FVW from each waste skip at JM. Ran-
dom sampling is the simplest method of sampling, and it underlies
many of the more complex method (Lebersorger and Schneider,
2011). The random sampling approach was applied because there
was no prior knowledge of each type of fruit and vegetable waste
generated on a daily basis. Also, it is difficult to collect every single
type of fruit and vegetable waste generated at JM. However, during
the experimentation, attempts were made to capture the percent-
age fruit to vegetable waste generated at the Market using the
weekly sale of fruits and vegetables as published on JM website.
Though this approach did not cater for the actual type of each fruit
or vegetable waste generated, it provided a reasonable basis for a
fraction of fruit to vegetable in the substrate mix. The experimental
period was for 105 days.
The FVWs collected were separately macerated without the
addition of water in an industrial food grinder as shown in Fig. 1.
A homogenised sample was taken for characterisation. Samples
not immediately used were packed into labelled plastic containers
and stored at 18 °C. Before use, the frozen substrate was thawed
and stored in a cold room operating at about 1 °C.
Proximate and ultimate analyses were conducted. Proximate
analysis for each substrate was performed after homogenisation
to determine the TS, VS and ash content of substrate as received
using a Barnstead Thermolyne 6000 Furnace. Details of the proce-
dure followed is as provided by Hart and Fisher (1971). The diges-
tate was also analysed according to procedures presented in Hart
and Fisher (1971). Ultimate analysis for each substrate was con-
ducted after homogenisation on TS basis to determine the carbon
(C), hydrogen (H), Nitrogen (N), and Sulphur (S) present in the sub-
strate using a Thermo Scientific Flash 2000 Organic Element
Analyser.
2.2. Micro-pilot scale system
The experimental setup of the micro-pilot plant shown in Fig. 2
includes; an industrial food grinder, a 6-litre hydrolysis unit oper-
ating at 35 ± 1 °C, and a vertical continuously stirred 35-litre diges-
ter operating at 35 ± 1 °C. The temperature of the digester is
supplied by a 75 W submersible heating element and connected
to a temperature controller. The operating volume of the hydroly-
sis unit was 3 L and the digester was 29.69 L. The hydrolysis was
manually agitated by shaking twice a day while the digester was
fitted with an IKA agitator that operates at 100 RPM over 10 min/
h. The biogas produced from the digester passes through a wet tip-
ping bucket flow meter that tips at every 100 ml of accumulated
biogas. The flowmeter was filled with acidified sodium chloride
as a barrier solution. Each tip of the flow meter was recorded as
one count of a digital counter. Daily biogas production was nor-
malised to standard temperature and pressure of 0 °C, 1 atm and
dry gas according to Eq. (1) (Bioprocess Control, 2014). The first
term of Eq. (1) converts the biogas volume to standard temperature
Fig. 1. Separately macerated vegetable (left) and fruit (right) waste.
 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
Fig. 2. CAD drawing of the micro-pilot 35-L digester (Below: fabricated digester).
and pressure. The second term corrects the biogas volume to only
account for dry gas volume using the Antonine equation, and the
third term calculates the total volume of biogas produced per
day. The average daily ambient temperature and pressure data
were retrieved from a WH2303 HP weather station and data log-
ger. The biogas produced is stored in a 10-litre Tedlar bag. The bio-
gas composition (CH4, CO2, O2) was analysed at least four times
daily with a Geotech Biogas 5000. After each biogas composition
test, the Tedlar bag is emptied by burning the biogas and squeezing
the Tedlar bag to force out most of the stored biogas.

 " 1730:63
ð8:1962ð233:426þT Þ
# source of inoculum can affect the kinetic rate of degradation. To
273:15 ðPBG þ 0:6Þ 10 BG Þ
V¼   1
ðT BG þ 273:15Þ 101:325 10  ðPBG þ 0:6Þ
 ½Nc  V c 
2.3. Start-up of the digester and monitoring
The inoculum used for starting the AD process was provided by
IBERT PTYs, a biogas company in Johannesburg, South Africa. The
inoculum obtained from a plant treating abattoir waste. The chem-
ical properties of the inoculum are presented in Table 1. The plant
operates at 35 °C, and the inoculum had a pH of 7.75. Since FVW
will be digested, it would have been most suitable to use inoculum
239
from AD plants treating such waste. However, most biogas plants in
CoJ and its environs operate on non-FVWs as substrate. Therefore,
only inoculum from such plants is available for use. Christof et al.
(2012) indicated that no significant difference was found among
BMP results for the same substrate using different sources of inocu-
lum. They concluded that the degradation would proceed provided
the inoculum contains sufficiently diverse microbial communities
to cope with the disintegration of the substrate. Also, Raposo
et al. (2011) reported the influence of inoculum sources was
insignificant regarding the extent of biodegradation. However, the
adapt the microbial community to the new substrate as against
its source, a low OLR was used as the initial feeding and gradually
increased until the starting inoculum was completely replaced,
ð1Þ
and the microbial population acclimated. The starting OLR was
set to 0.5 gVS/l-d for the first week and was iteratively increased
as shown in Fig. 3. Subsequent feeding and adjustment of OLR that
deviate from Fig. 3 were a response to the different composition of
FVWs. Parameters monitored during this experimental period were
the fruit to vegetable mix ratio, mass flowrate, and proximate char-
acteristics of substrate and digestate. Also monitored were OLR, pH,
VFA to total alkalinity (TA) ratio (also referred to as FOS/TAC in Ger-
man literature), CH4 content in biogas, biogas yield, and hydraulic
retention time. VFA/TA ratio and pH are critical substrate evolution
240 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250

Table 1
Characteristics of the inoculum used.

Inoculum TS (%ww) VS (%ww) C%TS H%TS N%TS

Abattoir digestate 2.33 2.01 42.97 5.97 2.33

Fig. 3. Iterative procedure for starting and monitoring the biodigester.

 NH SO

monitoring parameters in the digester while OLR is a control ðV pH4:4  V pH5 Þ  V20  0:1
2 4
 1:66  0:15  500  V dig
VFA=TA ¼
dig
response parameter for management of the evolutions. The
VFA/TA parameter is the ratio of the volatile fatty (organic) acid
0:5  NH2 SO4  V pH5  MCaCO3  1000
measured in equivalent mg of acetic acid per litre to the total alka- ð2Þ
linity (inorganic) that measures the alkaline buffer capacity of the
digestate in equivalent mg of calcium carbonate (CaCO3) per litre
(Bioprocess Control, 2015). According to Bensmann et al. (2016), 2.4. Full-scale plant assessment
VFA/TA ratio provides an immediate response to variation in the
stability of the digester than pH especially for a well-buffered sys- The full-scale plant is expected to receive about 47 tonnes of
tem and has been widely accepted as a better indicator than pH FVWs daily (approximately 17,000 tonnes/y). With consideration
only. In this study, both VFA/TA and pH were used as indicators given to lost FVW and pre-treatment inefficiency, it is expected
for the process evolution and the results guided on the adjustment that the plant will digest 45 tonnes/day (16,424 tonnes/y). The
of the OLR. The VFA/TA ratio was determined using titration of fil- waste stream of interest can be deemed clean as the FVWs are less
tered digestate with sulphuric acid to specific pH values of 5 and contaminated with other waste, although, sorting and cleaning to
4.4 as proposed by Rieger and Weiland (2006). The volume of acid achieve a higher purity is still required. For the proposed full-
consumed to attain both pH values were recorded and used to cal- scale plant, a consideration for mid-level mechanisation with
culate the VFA/TA ratio according to Eq. (2). human support at the front end of the sorting process was
 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
assumed. The sorted FVW is cleaned up, crushed and transported
by appropriate equipment. The proposed plant was divided into
six functional units. The major components of each functional unit
were sized following procedures of (Akbulut, 2012; Samer, 2010;
University of Southampton, 2016). The procedure for energy
demand and costing of each functional unit is presented in the
sub-sections below.
2.4.1. Reception, sorting and storage unit
This functional unit is described according to a modified
approach of Sobacken biogas plant (Eriksson, 2009). FVWs will be
discharged in a holding pit of the reception hall after being weighed.
FVWs are loaded onto a conveyor belt using a front wheel loader for
manual sorting. The sorted waste is transported into a bio-separator
using a screw conveyor. At the bio-separator, heavy matter like
undetected metal and grits settled at the bottom and expelled, while
non-organic materials like plastics floats at the top of the bio-
separator and are scooped off. The FVWs suspended between the
heavy matter, and non-organic waste is transported into a crusher.
Spillages from the bio-separator are retained in a pump pit for recir-
culation. The clean and sorted FVW is crushed into a buffer tank. In
the buffer tank, mixing with any additional slurry and substrate
amendment can take place before being transported to the sterilisa-
tion and hydrolysis unit. The specific electrical energy demand of
this functional unit can range between 8 kW /tonne for a PURAC
EFWA system to 50.2 kW h/tonne (Eriksson, 2009; PURAC, 2016).
The lower value has been used due to the scale of the plant.
2.4.2. Pre-treatment unit
The pre-treatment unit is intended to transform the clean but
heterogeneous FVWs into a homogenised single-feed stream with
properties within the specified range of TS, temperature, pH, C:N
ratio, and particle size. This unit will include a mixer, heating,
pumping, internal transport and ventilation systems. The specific
electrical energy demand is taken to be 18.97 kW h/tonne
(Lindkvist et al., 2017).
2.4.3. Digestion unit
The digestion unit is the key component of any AD process. The
digester for the FVWs will consist of a hydrolysis unit, the primary
digester where about 90% of the biogas is produced and a sec-
ondary digester that serves as a digestate holder and gas storage.
Internal agitators for mixing and heating system to maintain meso-
philic conditions were assumed to be installed in the hydrolysis
unit and primary digester. The specific electrical energy demand
for a biogas production process ranges between 7.32 and 10 kW h/-
tonne for a commercial-scale plant (Akbulut, 2012; Scano et al.,
2014). The average of 8.66 kW h/tonne is assumed in this study.
This is because part of the energy demand required for pre-
treatment and conditioning of the substrate has been specified in
Sections 2.4.1 and 2.4.2. Thermal energy was calculated according
to Eqs. (3) and (4) resulting in 46 kW h/tonne of the substrate. Eq.
(3) calculates total thermal energy losses through the walls of the
digester and digester dome cover (Q dig ) while Eq. (4) calculates the
thermal energy required to bring the fresh substrate up to the
digester operating condition (Q sub ).
ðT dig  T o;min Þ
Q dig ¼ 2  p  r    þ p  r2
1
þ tinsulation
hair;forced K ps
þh 1
water;free
 
T dig  T 0;min
  þ p  r2
þ Kppvvcc þ h 1
1 t
hair;forced air;free
 
T dig  T 0;min
  ð3Þ
K
0:5
þ tinsulation
K pv c
þh 1
soil water;free
241
Q sub ¼ msub  C specific  ðT dig  T sub Þ ð4Þ
2.4.4. Gas treatment unit
The end use of the biogas determines the type of gas treatment
unit required. The use of biogas in open flame application does not
need any form of cleaning. A minimum biogas cleaning that
involves the removal of condensate and hydrogen sulphide is
required before use in an internal combustion engine for heat
and power generation. For high-value utilisation such as in an
automobile, the removal of carbon dioxide and compression to a
high pressure increases the energy density of the gas. For all com-
mercial biogas plants, a gas flare unit is mandatory for combustion
of biogas during the period of scheduled maintenance, equipment
failure or unavailability of utilisation equipment. A compressor is
the main electrical component facilitating the transport of the bio-
gas through the series of cleaning equipment. The specific electri-
cal energy demand facilitating the transportation of the biogas
through the desulphurisation unit was taken as 0.1 kW h/Nm3 of
biogas.
2.4.5. Digestate management unit
This unit separates the solid fraction of the digestate from the
whole digestate. A separator with a capacity to achieve a volume
reduction of 41% and solid fraction of 56% as presented by
Breitenbeck and Schellinger (2004) was considered. The specific
electrical energy of the digestate solid/liquid fraction separation
was taken as 0.7 kW h/tonne of the whole digestate according to
University of Southampton (2016).
2.4.6. Biogas utilisation unit
After the biogas treatment, the conversion of the biogas into
useful energy is done using specialised equipment. In this study,
three energy utilisation pathways were considered as presented
in Table 2. The energy demand of the major biogas conversion to
energy equipment was taken from literature. The parasitic specific
electrical energy demand for the combined heat and power (CHP)
engine was assumed as 3% of the installed power. Biogas upgrading
inclusive of cooling duty and compression to 200 bars were 0.25
kW h/Nm3 and 0.11 kW h/Nm3 respectively (Budzianowski et al.,
2017; Cozma et al., 2013; Sun et al., 2015). The heat duty for biogas
upgrading was taken as 0.05 kW h/Nm3 of biogas (Cozma et al.,
2013; Cozma et al., 2015).
2.5. Economic considerations and assumptions
The economic performance of the three energy pathways for the
digestion of FVWs is examined in this section. Data for estimating
the specific cost of the major functional unit towards determining
the total plant cost (TPC), capital expenditure (CAPEX) and operat-
ing cost (OPEX) is presented. The data consisted of both literature
study of similar plant capacity, real data from engineering, pro-
curement and construction companies and specific equipment
suppliers.
The decision to invest in a biogas project is based on some pri-
mary values of merit. These are the net present value (NPV), the
internal rate of return (IRR) and payback period. The NPV is the
Table 2
Biogas utilisation pathways.
Pathways Biogas use
1 All biogas to electricity and exported to the grid
2 All biogas to compressed biomethane
3 Part of biogas for plant-parasitic energy demand and the rest
upgraded to biomethane
242 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250

sum of the discounted after-tax annual cashflows calculated Table 3

according to Eq. (5). The time value of money consideration makes TPC and CAPEX cost component.

the NPV a factor for quantifying an investment, but the financial Functional level TPC component Specific cost ($/tonne)
attractiveness depends on the extent of positive value recorded. 1. Civil and infrastructure 30.04
The accuracy of the NPV depends on the validity of the discounting 2. Reception and pre-treatment 25.55
rate which is often difficult to estimate accurately. The IRR is the 3. Digesters and ancillaries 37.72
discount rate that equates the net present revenue of the plant to 4. Digestate management 12.57
5. Biogas cleaning 12.15
its present total cost (Gebrezgabher et al., 2010) as calculated in 6. Electrical, SCADA and switchboard 21.72
Eq. (6). An IRR higher than the prevailing discount rate indicates 7. Other subsystems 12.92
economic viability for biogas project. IRR gives the measure of 8. Major energy conversion unit See Equation (1) and (2)
profitability of a project per unit of investment but does not 9. Total Plant Cost (TPC) =sum(1:8)
10. Project development cost 0.75 * TPC
account for both the scale of investment and measurement of abso-
11. Project contingency cost 0.05 * TPC
lute profitability. Payback compares the annual revenue with the 12. Total CAPEX =sum(9:11)
total investment of the plant and determines the duration of time
required to recoup the initial investment. The payback approach
does not account for the benefits incurred after the break-even Table 4
point (Masebinu et al. (2017)). Total operating cost factors.

X
n
CF t Total operating cost component Factors
NPV ¼ t ð5Þ
t¼0 ð1 þ iÞ Fixed cost
1. Labour overhead 10% of labour cost
2. Insurance 1% of TPC
X
n
CF t 3. General Plant overhead 0.5% of TPC
0¼ ð6Þ 4. Laboratory overhead 0.5% of turnover
t¼0 ð1 þ IRRÞt
2. Plant Management and marketing 1% of turnover

The three econometric indicators presented were investigated in Variable cost

this study. The determination of the econometric indicators
requires the calculation of the TPC, CAPEX and OPEX. Cost compo-
nent that makes up a capital cost quotation from technology ven-
dors might include in addition to the main equipment cost, cost
of erection, piping, instrumentation, electrical and SCADA system,
civil works, building, engineering, design, supervision, manage-
ment, commissioning, contingency, and interest during construc-
tion. In this study, all elements of the capital cost have been
incorporated into the TPC. The TPC was calculated by grouping
equipment into functional unit cluster discussed in Section 2.4 at
a specific cost per tonne of FVWs basis or cubic meter of biogas.
The capital cost for the CHP and biomethane unit are calculated
based on Eqs. (7) and (8) respectively. Eq. (7) was a modified equa-
tion from data presented Fig. S1 (Supplementary Material) by Lantz
(2012) and updated with recent cost data from CHP suppliers. The
equation was also validated with a maximum standard deviation
of ±36 $/kW (Table S1 Supplementary material). The equation ade-
quately predicts for CHP capacities <400 kW. Eq. (8) for the biogas
scrubber was derived through a non-linear regression of cost data
for real plant obtained from literature (Sun et al., 2013) and equip-
ment suppliers. The data showed tendencies of asymptote and was
therefore modelled using a concave asymptotic regression function
in Minitab 17 (Fig. S2, Supplementary material). Table 3 summaries
the specific cost at functional unit level. Factors considered for the
operational cost of the plant are presented in Tables 4 and S1–S3
Supplementary material
CC CHP ¼ 5904  P0:667
p ð7Þ
CC BM ¼ 66639:2  V 0:5165
v ð8Þ
3. Results and discussion
3.1. Experimental results
3.1.1. Proximate and ultimate analysis
During the duration of the experiment, a total of 62.52 kg of
mixed FVWs was processed and digested. Variation in FVWs com-
position and characteristics were observed. This was partly due to
the dynamic weather conditions and seasonality of fruits and veg-
1. Maintenance 2.5% of TPC
2. Consumables 0.5% of TPC
3. Utilities 0.1% of TPC
4. Electricity cost Calculated per pathway
5. Liquid digestate disposal cost 0.3% of TPC
6. Additional amendment substrate cost 5000 $/y
7. Contingency 3% of sum(1:6)
etables delivered to the JM over the duration of the experiment.
Table 5 shows the average fruit fraction during each experimental
cycle and the main composition of the substrate mixes used. The
whole experimental period can be divided into three cycles
namely; the start-up cycle, HRT 1 and 2 cycles. Each cycle after
start-up was approximately 27 days. During the experimental per-
iod, C:N ratio was between 18.44 and 42.24 depending on the FVW
substrate mixes. The average C:N ratio was 28.77.
3.1.2. Micro-Pilot plant Start-up and operation
Fig. 4 shows the fraction of fruit in the substrate mix and TS
concentration fed to the digester. Regression analysis with the frac-
tion of fruit in the substrate mix as the predictor variable and pH,
CH4 concentration and VFA/TA as response variable was con-
ducted. A unit increase in the fruit fraction caused a 1.12 and
7.38-unit reduction in pH and CH4 concentration respectively
while a 0.51-unit increase was observed for VFA/TA. This shows
that as fruit fraction increases in the substrate mix, the pH of the
digester gradually reduces towards acidity, the conversion effi-
ciency of the methanogens reduces, and the stability of the digester
reduces with increasing VFA/TA ratio. It was also observed that
increasing TS up to 15% wet weight resulted in an initial increase
in biogas yield and then a decline followed. A detailed discussion
of these observations and results of other monitored parameters
during the experimental period are summarised in Fig. 5A–D. The
monitored parameters are daily loading rate and volatile solid in
the substrate mix (Fig. 5A), OLR and biogas yield (Fig. 5B), pH
trends and VFA/TA ratio (Fig. 5C), biogas yield rate and average
daily CH4 concentration in the biogas (Fig. 5D).
In Fig. 5A, during the start-up cycle, the daily loading (kg/d) was
gradually increased to 0.6 kg/d. During the start-up cycle, continu-
ous increase in daily loading rate was to bring the digesters into
optimal loading conditions and completely replace the starting
 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250 243

Table 5
Mixed FVWs composition range.

Cycles Fruit [%] Total solid (%ww) Volatile solid (%ww) C (%TS) H (%TS) N (%TS)
Start-up 36.89 8.02–11.95 (Avg: 9.71) 7.15–11.52 (Avg: 9.19) 42.97–46.9 5.97–5.99 1.63–2.33
HRT 1 52.83 9.06–15.74 (Avg: 12.05) 8.53–14.99 (Avg: 11.51) 44.35–48.25 5.88–6.03 1.05–1.56
HRT 2 44.62 9.20–15.36 (Avg: 11.68) 8.64–14.97 (Avg: 11.12) 43.56–46.39 5.68–6.03 1.56–2.66

Fig. 4. Percentage Fruit and TS of substrate.

Fig. 5. Experimental results of the micro-pilot system. [A] Loading rate and volatile solid of substrate. [B] Organic loading rate and biogas yield. [C] pH and VFA/TA ratio. [D]
Biogas yield rate and Methane content.
244 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
inoculum. The complete replacement of the starting inoculum was
achieved in 52 days. Subsequently, fluctuation in the daily loading
rate during HRT cycle 1 and 2 was caused by corrective response
needed to stabilise the system while attempting to increase the
biogas yield rate. The maximum loading of the digester was 1.25
kg/d, but the digester was only stable when the daily loading
was between 0.7 and 0.9 kg/d with a VS ranging between 9 and
11% of the daily loading. As the TS and VS content of the substrate
changes, the daily loading rate was adjusted to maintain the
desired OLR in Fig. 5B. Higher VS requires a lower loading rate
and vice versa. A corrective measure with the adjustment of the
OLR was carried out while monitoring the pH, VFA/TA, biogas yield
(Nm3biogas/m3reactor-d), biogas yield rate (Nm3biogas/kgVS-d) and CH4
concentration.
In Fig. 5B, the biogas yield (Nm3biogas/m3reactor-d) increases propor-
tionally with the OLR up to the maximum stable limit of the digester
before a decline in the biogas yield. It was observed that as the OLR
exceed 3.4 kgVS/m3-d, there was a rapid decline in biogas yield. The
statistical regression with OLR as the predictor variable and biogas
yield as the response variable indicated that a unit increase in OLR
increases biogas yield by a factor of 1.12. The decline in biogas yield
was also noted on the statistical regression. Therefore, the increase
by a factor of 1.12 only holds true up to a loading rate less than
3.4 kgVS/m3-day. During HRT cycle 1, around day 68, a rapid decline
in biogas yield was recorded. This was due to the large quantity of
citrus waste that constituted a significant fraction of the fruit waste
mix. The FVWs mix was fed to the digester at an OLR of 3.4 kgVS/m3-
day. A situation of this nature is possible in the full-scale plant
around June and July when high citrus wastes are generated and ran-
domly when the rejection of large quantities farm produce can occur
due to damages during transportation. With the plant at a high load-
ing rate, a sudden change in the composition of substrates may occur
and will affect the performance of the plant as observed in this
study. High OLR makes the plant highly sensitive to any fluctuation
both in feed composition and operating parameters. It is therefore
pertinent for the plant manager to operate below critical loading
conditions and conduct a physical assessment of waste delivered
to the plant regularly before pre-treatment or a pre-mixing unit that
reduces the possibility of loading the digester with a substrate that
contains high inhibitory chemicals that can be released under anaer-
obic condition. Literature has reported challenges of digesting citrus
waste due to peel oil containing 32–98% of an alkylated aromatic
compound called limonene (Droby et al., 2008; Fagbohungbe et al.,
2016; Mizuki et al., 1990). Recommended pre-treatment methods
suggested in literature are steam distillation and aeration (Mizuki
et al., 1990). To recover the process stability, the OLR was immedi-
ately reduced, and the substrate mix was diluted with more veg-
etable waste.
As part of the monitoring parameters, the pH and VFA/TA ratio
provided insight into the degradation efficiency and stability
within the digester. Fig. 5C shows the pH trends and VFA/TA ratio
in response to the OLR. Statistical regression indicated that for
every unit increase in OLR, the pH decreases by a factor of 0.16
while the VFA/TA increases by a factor of 0.07. The adjustment of
OLR in response to pH and more particularly VFA/TA was guided
by acceptable range suggested in literature. As a general rule, the
VFA/TA ratio of anaerobic digestion processes should be in the
range of 0.3–0.4 (Lossie and Putz, 2008). Lossie and Putz (2008)
reported that VFA/TA ratios above 0.4 indicate an overloading of
the digester, therefore, indigestion and instability might occur.
Although some few processes treating energy crops requires VFA/
TA within 0.4–0.6, such systems require long-term observation.
In this experiment, once the VFA/TA exceeded 0.4, the OLR was
reduced with guidance from Fig. 3. Two significant instabilities
were observed over the experimental period during HRT 1 and
HRT 2 cycles. It was observed that as the OLR reaches 3.42 and
4.06 kgVS/m3-d, the VFA/TA reached peaks of 0.45 and 0.53 for
HRT 1 and 2 cycles respectively. It was noted that a reduction in
OLR from 3.42 to 1.44 kgVS/m3-d over a period of 5 days, resulted
in a corresponding decrease in VFA/TA ratio from 0.45 to 0.23 for
HRT 1 cycle with a pH adjustment from 7.12 to 7.28. While for
HRT 2 cycle, reduction of OLR from 4.06 to 1.92 kgVS/m3-d reduces
the VFA/TA to 0.29 over a period of 4 days with a pH increase from
7.06 to 7.27. It was, however, observed that the change in pH was
not as instantaneous as that of VFA/TA. However, at exceptional
high loading rate, as experimented with an OLR of 4.06 kgVS/m3-
d with a high concentration of fruits, both pH and VFA/TA gave
an immediate indication of instability. This indicates that at lower
OLR with a highly buffered system as in the start-up cycle, VFA/TA
ratio should have a higher monitoring priority than pH while at
high loading, both pH and VFA/TA are good indicators of the extent
of degradation and stability of the digester. Aside pH of the main
digester, the pH of the 6-litres hydrolysis unit was monitored.
The pH ranged between 3.81 and 4.94. The hydrolysis unit pro-
vided buffering capacity for the main digester and allowed dilution
of substrate mix to increase the pH level. Throughout the experi-
ment and despite the high loading, the pH of the digester was
not below 7. The lowest pH recorded was 7.06 on day 99.
In Fig. 5D, the effect of the OLR and substrate mix ratios on bio-
gas yield rate and CH4 concentration was presented. During the
start-up cycle, the biogas yield rate was unstable with high peak
and sudden valleys. This can be attributed to the initialisation of
the acclimation process of the microbial community. The average
biogas yield rate was 0.6 Nm3/kgVS with specific methane yield rate
of 0.35 Nm3/kgVS. Although biogas yield rate during the start-up
cycle was low compared to 0.91 Nm3/kgVS and 0.98 Nm3/kgVS for
HRT 1 and 2 cycles respectively, the CH4 concentration reached a
peak of 72.4%. The high CH4 concentration achieved during the
start-up cycle was due to low OLR that facilitated complete degra-
dation of the substrate. This was an indication of the active culture
of the microbial community in the starting inoculum. After the
start-up cycle, the average CH4 concentration was in the range of
55–57%. Fig. 5D also shows that variation in FVW mix composition
led to a CH4 concentration higher than 64% or lower than 49% only
for a short period. The depth of valley in CH4 concentration was
also observed to also relate to the frequency of feeding. Multiple
feeding in a day yielded a more stable CH4 concentration than
feeding once a day.
A summarised weekly average of fruit fraction in the substrate
mix and its effects on the monitored parameters in presented in
Figs. S4 and S5 (Supplementary material). Average values of biogas
yield, biogas yield rate and composition during the three experi-
mental cycles are presented in Table 6. It was observed that when
OLR was less than 2 kgVS/m3-d and VFA/TA ratio less than 0.21, the
reactor was at suboptimal conditions producing below 1 Nm3biogas/
(m3reactor-d). The intermediate optimal operating parameters are
loading rate that ranges between 2 and 2.68 kgVS/m3-d, VFA/TA ratio
of 0.23–0.29 resulting in biogas yield that is greater than 1 but less
than 2.4 Nm3biogas/(m3reactor-d). Optimal conditions during this exper-
iment were found to be an OLR that was between 2.68 and 2.97 kgVS/
m3-d, VFA/TA ratio that range between 0.29 and 0.36 resulting in a
biogas yield between 2.4 and 2.8 Nm3biogas/(m3reactor-d). The average
specific biogas at the optimal conditions was 0.87 Nm3/kgVS with a
specific methane 0.49 Nm3/kgVS corresponding to 57.58% CH4 con-
centration in the biogas. The results obtained from this experiment
are high compared to some result from literature at mesophilic con-
ditions but are in line with other studies (Di Maria and Barratta,
2015; Gunaseelan, 2004; Raposo et al., 2011; Scano et al., 2014) that
digested FVWs solely. Scano et al. (2014) digested FVWs solely in a
0.95 m3 digester over 180 days and reported a specific CH4 yield rate
of 0.43 Nm3/kgVS. Bohn et al. (2007) digested sugar beet tops in a
1.83 m3 digester and achieved a specific CH4 yield rate of
 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250 245

Table 6
Average values of biogas production and composition.

Experimental Cycles Biogas yield (Nm3biogas/m3reactor-d) Biogas yield rate (Nm3/kgvs) CH4 (%vol.) CO2 (%vol.) O2 (%vol.) Others (% vol.)
Start-up 0.95 0.60 58.95 39.86 0.43 0.76
HRT 1 2.47 0.90 56.72 42.77 0.32 0.19
HRT 2 2.89 0.98 55.85 43.61 0.20 0.34

Table 7 Table 9
Parameters considered for the full-scale plant. Economic comparison of the energy pathways.

Parameters Factors used Pathway 1 Pathway 2 Pathway 3

Daily waste received/digested 47 tonnes/day/45 tonnes/day TPC $2,928,223 $3,574,337 $3,557,594

FVW composition characteristics TS = 11.84% ww; VS = 11.29%ww CAPEX $3,294,250 $4,021,129 $4,002,293
Organic loading rate 2.68 kgVS/(m3d) NPV $1,262,549 $2,428,021 $1,208,509
Operating Temperature 35C IRR 14.30% 16.90% 13.61%
Specific biogas production 0.8658 Nm3/kg VS Simple payback 7.42 y 6.17 y 7.44 y
Specific methane production 0.4985 Nm3/kg VS Discounted payback 11.76 y 8.87 y 11.96 y
Specific Energy Potential of Biogas 5.5 kW h/Nm3
Density of biogas 1.18 kg/m3
HRT hydrolysis/methanogenesis 5/30
CHP electrical/thermal Efficiency 40/45% 3.2. Full-scale plant
CHP utilisation factor/replacement year 92%/7 years
Electrical/Thermal losses 1%/10%
Methane slip 1.5% Results obtained from the experimental phase provided the
Digestate storage duration 3 months basis for the feasibility assessment of the full-scale plant. More
CO2 emissions for coal sourced electricity 0.99 CO2/kW h specifically, the specific biogas production, CH4 concentration, TS
and VS were considered. Plant operational data and other assump-
tions are summarised in Table 7. The envisaged process layout at
0.39 /kgVS. Raposo et al. (2011) obtained 0.37 Nm3/kgVS for mung functional unit level is presented in Fig. 6. The major tanks that
bean from an interlaboratory study. In this study, the high perfor- make-up AD system are the feed preparation (item 6 of Fig. 6),
mance obtained can be due to the long start-up period of 52 days hydrolysis (item 8 of Fig. 6), main digester (item 9 of Fig. 6) and
that allowed for an excellent acclimation of the microbial popula- digestate storage (item 10 of Fig. 6) and their respective capacity
tion. It can also be attributed to the high biodegradability of an aver- are 99 m3, 292 m3, 1755 m3 and 3996 m3. The digestate storage
age value of 78.74% achieved by analysing the digestate once a week of 3,996 m3 was sized for winter, a duration of three months, in
throughout the experimental period. This indicated that the bio- South Africa. The annual biogas production will be 1,605,455
digestion process is highly efficient in converting the volatile solids Nm3/y which is equivalent to 8,830 MW h/y. The AD plant will pro-
into biogas. Aside from the high biodegradation, the hydrolysis dura- cess 16,424 tonnes/y of FVWs and will produce 14,530 tonnes/y
tion ranged between 4 and 6 days on average that allowed for com- whole digestate. With the solid fraction of the digestate consid-
plete hydrolysis of the fresh substrate. Also, an average of 38 days ered, the annual compost production will be 5,914 tonnes. The
HRT as shown in Fig. S3 (Supplementary material) for all cycles fur- energy demand and annual products depending on the energy
ther improved the methanogenic activity which favours biogas yield. pathway are summarised in Table 8.
At the optimal condition, the pH varied between 7.32 and 7.44 with Pathway 1 will require a 430 kW CHP to convert all biogas into a
the fruit to vegetable ratio of 2.2:2.8. At conditions where fruit frac- total of 3.4 GW h/y of electricity. This is equivalent to 15.2% of the
tion is greatly higher than vegetable, the initial biogas rate was high electrical energy demand of JM. Net excess heat available for other
but with lower CH4 concentration. The reduction in CH4 is due to the industrial use is 2.7 GW h/y. There are two possibilities considering
rapid degradation of simple sugars that increase the initial biogas GHG emissions for AD systems: CO2 emission that would have
yield but simultaneously liberates carbon dioxide. Maintaining a occurred should the biogas escape to the atmosphere and CO2
high fraction of fruit in the substrate mix caused a decline in the emissions from combusting biogas equivalent to, respectively,
pH, an increase in VFA/TA ratio above recommended stable region 8.5 kg CO2/Nm3 biogas and 1.9 kg CO2/Nm3 biogas (Akbulut,
and an eventual reduction in biogas yield. These conditions were 2012). If the AD process had not occurred, 13,588 tCO2eq/y would
observed on days 62–68 and 96–100. However, the reduction of have been emitted. However, with the diversion of FVWs into an
OLR lowered the VFA/TA ratio, improved the pH and resulted in a AD and producing electricity from it, 10,143 tonnes CO2eq/y was
return to stability with biogas yield improvement. avoided.

Table 8
Energy demand and product potential.

Pathway 1 Pathway 2 Pathway 3

Biogas Production 1,605,455 1,605,455 1,605,455
Electrical Demand Energy 614,901 kW h/y 1,207,421 kW h/y 1,207,421 kW h/y
Thermal Demand Energy 780,735 kW h/y 861,008 kW /y 833,566 kW /y
Installed CHP 430 kW – 160 kW
Product potential
Electricity 3,369,529 kW /y – 1,280,000 kW /y
Biomethane – 900,288 Nm3/y 599,246 Nm3/y
Compost 5914 tonnes/y 5914tonnes/y 5914 tonnes/y
Excess Thermal Energy (kWh/yr) 2,715,143 kW /y – 388,948 kW /y
CO2 avoided (tCO2eq) 9833 12,393 12,513
246 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250

Fig. 6. Schematic of the proposed 45 tonnes/d full-scale biogas plant.

 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250 247

Fig. 7. Impact of variable variation on energy pathway 1.

Fig. 8. Impact of variable variation on energy pathway 2.

Pathway 2 will produce 900,288 Nm3/y of biomethane. The pur-
chase of electricity from the grid and upgrading all biogas to bio-
methane will avoid 12,393 tonnes CO2eq/y. Considering the CoJ
metro bus that uses on average 110 L of diesel per day
(University of Johannesburg, 2016), the biomethane produced
can fuel 18 dedicated natural gas metro bus per year.
Pathway 3 produces both its electrical and thermal energy need
and the remaining biogas is upgraded to biomethane. The designed
system requires a 160 kW CHP that will produce 1,280,000 kW h/y
electricity with the excess thermal energy of 388,948 kW h/y after
meeting the thermal energy needs of the plant. Also, 599,246 Nm3/
y of biomethane will be produced to fuel 12 dedicated metro buses.
The production of parasitic electricity and conversion of biogas to
biomethane avoided the on-site GHG emissions by 12,513 ton-
nes CO2eq/y. This pathway is the most environmentally friendly.
Table 9 presents the economic performance of the energy path-
ways investigated. The energy pathway 1 had the least TPC of
$2,928,223 while pathway 2 had the highest. The TPC for pathways
2 and 3 were identical due to a marginal 50 m3/h difference in
flowrate capacity of the upgrading unit. The NPV for each energy
pathway was calculated after considering the unit cost and selling
price of raw material, utilities and products among other variables
as specified in Table 4. The unit cost of electricity and selling price
were 0.0825 and 0.0847 $/kW h respectively. Landfill charges and
compost selling prices were 17.58 and 17.76 $/tonne respectively.
The price of biomethane was calculated on gasoline energy equiv-
alent basis as 0.68 $/Nm3 of biomethane. The NPV was calculated
at discounting rate of 10%. Of the three energy pathways, pathway
2 involving the purchase of all electrical needs from the grid and
harnessing the excess thermal energy from a LFG plant on site
while upgrading all biogas to biomethane was the most economi-
cally viable pathway. This pathway had NPV of $2,428,021, IRR of
16.90% and a simple payback period of 6.17 years. Energy pathway
2 leveraged on the low cost of grid electricity and the high cost of
vehicular fuel for profitability. Pathway 3 economic performance
was also favourable with a positive NPV though the least among
the three pathways investigated as shown in Table 9. The low eco-
nomic performance of pathway 3 was because 34.19% of the biogas
generated was required to meet parasitic energy demand. Also, the
capital cost of biogas upgrading unit was almost identical to that of
pathway 2 and therefore reducing any potential benefit that could
have been accrued from self-generation of energy. Pathway 1 had a
good economic performance over pathway 3 but its performance
was lower than pathway 2 due to the marginal price differential
248 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
Fig. 9. Impact of variable variation on energy pathway 3.
between electricity purchase and selling price. Using the renew-
able energy feeding tariff (REFIT) for biogas (0.11 $/kW h) the
NPV would be $2,209,223 with a payback period of 6.22. This will
make pathway 1 very competitive with pathway 2. However, an
industry expert has described the process of obtaining the REFIT
licence as cumbersome and costly. Therefore, in this paper, a more
pessimistic approach was taken in consultation with biogas plant
operators in South Africa.
The impact of the discounting rate on payback period was con-
sidered in this study. The study showed that if the assumed dis-
counting rate remains valid through the project lifetime, the
payback period could range between 9 years for the most econom-
ically viable pathway to 12 years for the least economically viable
pathway as shown in Table 9. Discounted payback period is not
often used as an econometric indicator but it is important to indi-
cate the expected long-term economic viability of the project
should the discounting rate assumed remains valid through the
project life. Also, since NPV already considers the time value of
money, a simple payback period could be deceptive as a negative
NPV can still have simple payback period that is lower than the
project life.
A sensitivity analysis was conducted on some primary variables
for each energy pathway and the effect on IRR was studied (Figs. 7-
9). Biogas volume, CAPEX, product selling price and landfill gate
charges had the highest impact on IRR. A ±10% variation on pri-
mary variables on IRR for energy pathway 1 is presented in
Fig. 7. The IRR increased by 1.74%, 0.95%, 0.83% and 1.62% with
10% increase in biogas volume, electricity selling price, landfill gate
price and CAPEX respectively.
On the most economically viable energy pathway, pathway 2
shown in Fig. 8, biogas volume is the variable with the greatest
impact on IRR. A 10% increase in biogas volume improved the
IRR by 2.31% and an equivalent reduction reduces IRR by 2.38%.
A 10% increase in CAPEX reduces the IRR by 1.33% while a 10%
decrease increased IRR by 1.47%. The sensitivity analysis shows
that reduction in CAPEX without improvement in plant productiv-
ity has a small impact on the economic viability of the plant com-
pared to the biogas volume.
In pathway 3, 10% reduction in CAPEX increased IRR by 1.78%
and biogas volume reduces IRR by 2.83% as shown in Fig. 9. A
±10% variation in primary variable had a proportional impact of
±0.72%, ±0.45%, ±0.35%, and ±0.25% landfill gate price, fixed cost,
variable cost and compost price respectively. The sensitivity anal-
ysis showed that variation in biogas production has the highest
impact on the econometric indicators of the plant. Therefore,
improving biogas yield will be favourable for plant economic
returns.
4. Conclusion
AD is an effective technology for management of FVWs, gener-
ation of energy and compost and mitigation of GHG emission. The
high specific biogas yield showed that the digestion of FVWs is
viable as a mono-substrate when the fraction of fruit waste in
the substrate mix is controlled and below that of vegetable. Where
fruit fractions are higher than vegetable, the AD process becomes
unstable with low biogas yield and low methane concentration.
This study further supports existing proposition that the extent
of instability in a digester could be studied sufficiently through
the VFA/TA and pH while the OLR is an effective control response
strategy. The feasibility assessment for the full-scale plant showed
that FVWs generated at JM could offset 15.2% of the Market annual
energy consumption and reduce GHG emission. Despite the esca-
lating cost of electricity, the study showed that it was economically
viable to purchase electricity from the grid to meet parasitic load
as proposed by pathway 2 than a self-generation and consumption
as suggested by pathway 3. The result of this study could be
applied to other fruit and vegetable markets in South Africa as well
as other cities in Africa with similar climatic condition and socio-
economic status as to what is obtainable in the CoJ.
Acknowledgement
This work was supported by the University of Johannesburg-
University Research Committee [UJ-URC-201339837]. Also, the
authors are grateful to the City of Johannesburg, Environmental
Infrastructural Service Department, Joburg Market and Process
Energy and Environmental Technology Station at the University
of Johannesburg for providing support for the research conducted.
The opinion expressed in this paper are those of the authors and
are not necessarily to be attributed to the sponsors.
 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250 249

Appendix A. Supplementary material Fonoll, X., Astals, S., Dosta, J., Mata-Alvarez, J., 2015. Anaerobic co-digestion of
sewage sludge and fruit wastes: Evaluation of the transitory states when the co-
substrate is changed. Chem. Eng. J. 262, 1268–1274.
Supplementary data associated with this article can be found, in Ganesh, R., Torrijos, M., Sousbie, P., Lugardon, A., Steyer, J.P., Delgenes, J.P., 2014.
the online version, at https://doi.org/10.1016/j.wasman.2018.02. Single-phase and two-phase anaerobic digestion of fruit and vegetable waste:
Comparison of start-up, reactor stability and process performance. Waste
011.
Manage. (Oxford) 34, 875–885.
Ganesh, R., Torrijos, M., Sousbie, P., Steyer, J.P., Lugardon, A., Delgenes, J.P., 2013.
References Anaerobic co-digestion of solid waste: effect of increasing organic loading rates
and characterization of the solubilised organic matter. Bioresour. Technol. 130,
559–569.
Aboudi, K., Álvarez-Gallego, C.J., Romero-García, L.I., 2015. Semi-continuous
Gao, S., Huang, Y., Yang, L., Wang, H., Zhao, M., Xu, Z., Huang, Z., Ruan, W., 2015.
anaerobic co-digestion of sugar beet byproduct and pig manure: effect of the
Evaluation the anaerobic digestion performance of solid residual kitchen waste
organic loading rate (OLR) on process performance. Bioresour. Technol. 194,
by NaHCO3 buffering. Energy Convers. Manage. 93, 166–174.
283–290.
García-Gen, S., Rodríguez, J., Lema, J.M., 2014. Optimisation of substrate blends in
Akbulut, A., 2012. Techno-economic analysis of electricity and heat generation from
anaerobic co-digestion using adaptive linear programming. Bioresour. Technol.
farm-scale biogas plant: Cicekdagi case study. Energy 44, 10.
173, 159–167.
Alkanok, G., Demirel, B., Onay, T.T., 2014. Determination of biogas generation
Gebrezgabher, S.A., Meuwissen, M.P.M., Prins, B.A.M., Lansink, A.G.J.M.O., 2010.
potential as a renewable energy source from supermarket wastes. Waste
Economic analysis of anaerobic digestion—à case of Green power biogas plant in
Manage. 34, 134–140.
The Netherlands. NJAS – Wageningen J. Life Sci. 57, 109–115.
Amoo, O.M., Fagbenle, R.L., 2013. Renewable municipal solid waste pathways for
Gunaseelan, V.N., 2004. Biochemical methane potential of fruits and vegetable solid
energy generation and sustainable development in the Nigerian context. Int. J.
waste feedstocks. Biomass Bioenergy 26, 389–399.
Energy Environ. Eng. 4, 42.
Hart, F.L., Fisher, H.J., 1971. In: Introduction—General Methods for Proximate and
Asquer, C., Pistis, A., Scano, E.A., 2013. Characterisation of fruit and vegetable wastes
Mineral Analysis, Modern Food Analysis. Springer, Berlin, Heidelberg, pp. 1–27.
as a single substrate for anaerobic digestion. Environ. Eng. Manage. J. 12, 4.
Huang, X., Yun, S., Zhu, J., Du, T., Zhang, C., Li, X., 2016. Mesophilic anaerobic co-
Aurecon South Africa (Pty) Ltd, 2015. Appendix A: Waste characterisation study for
digestion of aloe peel waste with dairy manure in the batch digester: focusing
September 2014, November 2014 and June 2015, Feasibility study for
on mixing ratios and digestate stability. Bioresour. Technol. 218, 62–68.
alternative waste treatment technology, Johannesburg, p. 88.
Jacob, S., Banerjee, R., 2016. Modeling and optimization of anaerobic codigestion of
Bensmann, A., Hanke-Rauschenbach, R., Heyer, R., Kohrs, F., Benndorf, D.,
potato waste and aquatic weed by response surface methodology and artificial
Kausmann, R., Plöchl, M., Heiermann, M., Reichl, U., Sundmacher, K., 2016.
neural network coupled genetic algorithm. Bioresour. Technol. 214, 386–395.
Diagnostic concept for dynamically operated biogas production plants. Renew.
Jiang, Y., Heaven, S., Banks, C.J., 2012. Strategies for stable anaerobic digestion of
Energy 96, 479–489.
vegetable waste. Renew. Energy 44, 206–214.
Bioprocess Control, 2014. AMPTS II-Automatic Methane Potential Test – Operation
Market, Joburg, 2016. Energy Management Assessment and Energy Supply Strategy
and Maintenance Manual. Bioprocess Control, Lund, Sweden, p. 86.
for: Joburg Market. Johannesburg, Joburg Market, p. 10.
Bioprocess Control, 2015. Redimensioning the Importance of the VFA/TA (FOS/TAC)
Karellas, S., Boukic, I., Kontopoulos, G., 2010. Development of an industrial decision
Method. In: Rosato, M. (Ed.).
tool for biogas production from agricultural waste. Renew. Sustain. Energy Rev.
Bohn, I., Björnsson, L., Mattiasson, B., 2007. The energy balance in farm scale
14, 1273–1282.
anaerobic digestion of crop residues at 11–37 °C. Process Biochem. 42, 57–64.
Khalid, A., Arshad, M., Anjum, M., Mahmood, T., Dawson, L., 2011. The anaerobic
Bouallagui, H., Ben Cheikh, R., Marouani, L., Hamdi, M., 2003. Mesophilic biogas
digestion of solid organic waste. Waste Manage. 31, 1737–1744.
production from fruit and vegetable waste in a tubular digester. Bioresour.
Koch, K., Plabst, M., Schmidt, A., Helmreich, B., Drewes, J.E., 2016. Co-digestion of
Technol. 86, 85–89.
food waste in a municipal wastewater treatment plant: comparison of batch
Bouallagui, H., Touhami, Y., Ben Cheikh, R., Hamdi, M., 2005. Bioreactor performance
tests and full-scale experiences. Waste Manage 47, 28–33.
in anaerobic digestion of fruit and vegetable wastes. Process Biochem. 40, 989–
Lantz, M., 2012. The economic performance of combined heat and power from
995.
biogas produced from manure in Sweden – A comparison of different CHP
Breitenbeck, G.A., Schellinger, D., 2004. Calculating the reduction in material mass
technologies. Appl. Energy 98, 502–511.
and volume during composting. Compost Science Utilization 12, 365–371.
Lebersorger, S., Schneider, F., 2011. Discussion on the methodology for determining
Budzianowski, W.M., Wylock, C.E., Marciniak, P.A., 2017. Power requirements of
food waste in household waste composition studies. Waste Manage. 31, 1924–
biogas upgrading by water scrubbing and biomethane compression:
1933.
comparative analysis of various plant configurations. Energy Convers.
Lebuhn, M., Munk, B., Effenberger, M., 2014. Agricultural biogas production in
Manage. 141, 2–19.
Germany – from practice to microbiology basics. Energy, Sustain. Soc. 4, 10.
Cabbai, V., Ballico, M., Aneggi, E., Goi, D., 2013. BMP tests of source selected OFMSW
Li, D., Liu, S., Mi, L., Li, Z., Yuan, Y., Yan, Z., Liu, X., 2015. Effects of feedstock ratio and
to evaluate anaerobic codigestion with sewage sludge. Waste Manage. (Oxford)
organic loading rate on the anaerobic mesophilic co-digestion of rice straw and
33, 1626–1632.
pig manure. Bioresour. Technol. 187, 120–127.
Christof, H., Nathalie, B., Helene, F.D.L., Marc, D., Yves, M., Arthur, W., 2012. Inoculum
Lin, J., Zuo, J., Ji, R., Chen, X., Liu, F., Wang, K., Yang, Y., 2012. Methanogenic
origin and adaptation had no influence on biomethane potential results with
community dynamics in anaerobic co-digestion of fruit and vegetable waste
different substrates. IV International Symposium on Energy from Biomass and
and food waste. J. Environ. Sci. 24, 1288–1294.
Waste. International Waste Water Group, San Servolo, Venice, Italy.
Lindkvist, E., Johansson, M.T., Rosenqvist, J., 2017. Methodology for analysing
City of Johannesburg, 2011. City of Johannesburg integrated waste management
energy demand in biogas production plants-A comparative study of two biogas
plan, Annual Report. City of Johannesburg, Johannesburg, p. 140.
plants. Energies, 20.
Cozma, P., Ghinea, C., Mămăligă, I., Wukovits, W., Friedl, A., Gavrilescu, M., 2013.
Lossie, U., Putz, P., 2008. Targeted control of biogas plants with the help of FOS/TAC.
Environmental impact assessment of high pressure water scrubbing biogas
In: Hack-Lange (Ed.), Practise Report: Laboratory Analysis Titration.
upgrading technology. CLEAN – Soil Air Water 41, 917–927.
Maile, I., Muzenda, E., Mbohwa, C., 2016. Biogas production from anaerobic digestion
Cozma, P., Wukovits, W., Mămăligă, I., Friedl, A., Gavrilescu, M., 2015. Modeling and
of fruit and vegetable waste from Johannesburg market. In: 17th International
simulation of high pressure water scrubbing technology applied for biogas
Conference on Biology, Environment, and Chemistry. IPCBEE, pp. 100–104.
upgrading. Clean Technol. Environ. Policy 17, 373–391.
Mangwandi, C., JiangTao, L., Albadarin, A.B., Allen, S.J., Walker, G.M., 2013. The
Dahunsi, S.O., Oranusi, S., Owolabi, J.B., Efeovbokhan, V.E., 2016. Mesophilic
variability in nutrient composition of Anaerobic Digestate granules produced
anaerobic co-digestion of poultry dropping and Carica papaya peels:
from high shear granulation. Waste Manage. 33, 33–42.
modelling and process parameter optimization study. Bioresour. Technol. 216,
Masebinu, S.O., Akinlabi, E.T., Muzenda, E., Aboyade, A.O., 2017. Techno-economics
587–600.
and environmental analysis of energy storage for a student residence under a
Di Maria, F., Barratta, M., 2015. Boosting methane generation by co-digestion of
South African time-of-use tariff rate. Energy 135, 413–429.
sludge with fruit and vegetable waste: Internal environment of digester and
Masebinu, S.O., Akinlabi, E.T., Muzenda, E., Aboyade, A.O., Mbohwa, C., 2017b.
methanogenic pathway. Waste Manage. 43, 7.
Assessing the morphological composition and energy potential of MSW, the
Di Maria, F., Sordi, A., Cirulli, G., Micale, C., 2015. Amount of energy recoverable
case of the City of Johannesburg. In: International Conference on Industrial
from an existing sludge digester with the co-digestion with fruit and vegetable
Engineering and Operations Management. IEOM, Rabat, Morocco, p. 7.
waste at reduced retention time. Appl. Energy 150, 9–14.
Mizuki, E., Akao, T., Saruwatari, T., 1990. Inhibitory effect of Citrus unshu peel on
Droby, S., Eick, A., Macarisin, D., Cohen, L., Rafael, G., Stange, R., McColum, G., Dudai,
anaerobic digestion. Biol. Wastes 33, 161–168.
N., Nasser, A., Wisniewski, M., Shapira, R., 2008. Role of citrus volatiles in host
Molinuevo-Salces, B., González-Fernández, C., Gómez, X., García-González, M.C.,
recognition, germination and growth of Penicillium digitatum and Penicillium
Morán, A., 2012. Vegetable processing wastes addition to improve swine
italicum. Postharvest Biol. Technol. 49, 386–396.
manure anaerobic digestion: evaluation in terms of methane yield and SEM
Eriksson, M., 2009. In: Energy optimisation, Sobacken biogas plant. Uppsala
characterization. Appl. Energy 91, 36–42.
Universitet, Sweden, p. 57.
Muhammad Nasir, I., Mohd Ghazi, T.I., Omar, R., 2012. Production of biogas from
Fagbohungbe, M.O., Herbert, B.M.J., Hurst, L., Li, H., Usmani, S.Q., Semple, K.T., 2016.
solid organic wastes through anaerobic digestion: a review. Appl. Microbiol.
Impact of biochar on the anaerobic digestion of citrus peel waste. Bioresour.
Biotechnol. 95, 321–329.
Technol. 216, 142–149.
Ogola, J., Chimuka, L., Tshivhase, S., 2011. Management of Municipal Solid Wastes: A
Fallde, M., Eklund, M., 2015. Towards a sustainable socio-technical system of biogas
case study in Limpopo Province, South Africa, Integrated Waste Management-
for transport: the case of the city of Linköping in Sweden. J. Cleaner Prod. 98,
Volume I. InTech, p. 24.
17–28.
250 S.O. Masebinu et al. / Waste Management 75 (2018) 236–250
Olsson, L., Fallde, M., 2015. Waste(d) potential: a socio-technical analysis of biogas
production and use in Sweden. J. Cleaner Prod. 98, 107–115.
Parajuli, R., Østergaard, P.A., Dalgaard, T., Pokharel, G.R., 2014. Energy consumption
projection of Nepal: An econometric approach. Renew. Energy 63, 432–444.
Pavi, S., Kramer, L.E., Gomes, L.P., Miranda, L.A.S., 2017. Biogas production from co-
digestion of organic fraction of municipal solid waste and fruit and vegetable
waste. Bioresour. Technol. 228, 362–367.
PURAC, 2016. Effective pre-treatment of organic waste into biogas, PURAC efficient
food waste application. PURAC, Sweden, p. 4.
Rao, P.V., Baral, S.S., 2011. Experimental design of mixture for the anaerobic co-
digestion of sewage sludge. Chem. Eng. J. 172, 977–986.
Raposo, F., Fernández-Cegrí, V., De la Rubia, M.A., Borja, R., Béline, F., Cavinato, C.,
Demirer, G., Fernández, B., Fernández-Polanco, M., Frigon, J.C., Ganesh, R.,
Kaparaju, P., Koubova, J., Méndez, R., Menin, G., Peene, A., Scherer, P., Torrijos,
M., Uellendahl, H., Wierinck, I., de Wilde, V., 2011. Biochemical methane
potential (BMP) of solid organic substrates: evaluation of anaerobic
biodegradability using data from an international interlaboratory study. J.
Chem. Technol. Biotechnol. 86, 1088–1098.
Ratanatamskul, C., Wattanayommanaporn, O., Yamamoto, K., 2015. An on-site
prototype two-stage anaerobic digester for co-digestion of food waste and
sewage sludge for biogas production from high-rise building. Int. Biodeterior.
Biodegrad. 102, 143–148.
Rieger, C., Weiland, P., 2006. Prozessstorungenfruhzeitig erkennen. Biogas J. 4, 3.
Samer, M., 2010. A software program for planning and design of biogas plant. Trans.
ASABE 53, 9.
Scano, E.A., Asquer, C., Pistis, A., Ortu, L., Demontis, V., Cocco, D., 2014. Biogas from
anaerobic digestion of fruit and vegetable wastes: experimental results on
pilot-scale and preliminary performance evaluation of a full-scale power plant.
Energy Convers. Manage. 77, 22–30.
Schnürer, A., Bohn, I., Moestedt, J., 2017. Protocol for Start-Up and Operation of CSTR
Biogas Processes. In: McGenity, T.J., Timmis, K.N., Nogales, B. (Eds.),
Hydrocarbon and Lipid Microbiology Protocols: Bioproducts, Biofuels,
Biocatalysts and Facilitating Tools. Springer, Berlin, Heidelberg, pp. 171–200.
Shen, F., Yuan, H., Pang, Y., Chen, S., Zhu, B., Zou, D., Liu, Y., Ma, J., Yu, L., Li, X., 2013.
Performances of anaerobic co-digestion of fruit & vegetable waste (FVW)
and food waste (FW): Single-phase vs. two-phase. Bioresour. Technol. 144,
80–85.
Shen, Y., Forrester, S., Koval, J., Urgun-Demirtas, M., 2017. Yearlong semi-continuous
operation of thermophilic two-stage anaerobic digesters amended with biochar
for enhanced biomethane production. J. Cleaner Prod.
Shen, Y., Linville, J.L., Urgun-Demirtas, M., Mintz, M.M., Snyder, S.W., 2015. An
overview of biogas production and utilization at full-scale wastewater
treatment plants (WWTPs) in the United States: challenges and opportunities
towards energy-neutral WWTPs. Renew. Sustain. Energy Rev. 50, 346–362.
Sitorus, B., Sukandar, Panjaitan, S.D., 2013. Biogas recovery from anaerobic digestion
process of mixed fruit -vegetable wastes. Energy Proc. 32, 176–182.
Sun, L., Wan, S., Luo, W., 2013. Biochars prepared from anaerobic digestion residue, palm
bark, and eucalyptus for adsorption of cationic methylene blue dye:
characterization, equilibrium, and kinetic studies. Bioresour. Technol. 140, 406–413.
Sun, Q., Li, H., Yan, J., Liu, L., Yu, Z., Yu, X., 2015. Selection of appropriate biogas
upgrading technology – a review of biogas cleaning, upgrading and utilisation.
Renew. Sustain. Energy Rev. 51, 521–532.
Tan, S.T., Ho, W.S., Hashim, H., Lee, C.T., Taib, M.R., Ho, C.S., 2015. Energy, economic
and environmental (3E) analysis of waste-to-energy (WTE) strategies for
municipal solid waste (MSW) management in Malaysia. Energy Convers.
Manage. 102, 111–120.
University of Johannesburg, 2016. Feasibility study: Implementation of a pilot
biogas plant at Robinson deep landfill. In: Masebinu, S.O. (Ed.), Waste to Energy.
University of Johannesburg, Johannesburg, p. 178.
University of Southampton, 2016. Anaerobic Digestion Assessment Tool. University
of Southampton, London, p. 68.
Wang, D., Ai, J., Shen, F., Yang, G., Zhang, Y., Deng, S., Zhang, J., Zeng, Y., Song, C.,
2017. Improving anaerobic digestion of easy-acidification substrates by
promoting buffering capacity using biochar derived from vermicompost.
Bioresour. Technol. 227, 286–296.
Wang, L., Shen, F., Yuan, H., Zou, D., Liu, Y., Zhu, B., Li, X., 2014. Anaerobic co-
digestion of kitchen waste and fruit/vegetable waste: Lab-scale and pilot-scale
studies. Waste Manage. 34, 2627–2633.
Wang, X., Li, Z., Bai, X., Zhou, X., Cheng, S., Gao, R., Sun, J., 2018. Study on improving
anaerobic co-digestion of cow manure and corn straw by fruit and vegetable
waste: Methane production and microbial community in CSTR process.
Bioresour. Technol. 249, 290–297.
Wang, X., Yang, G., Feng, Y., Ren, G., Han, X., 2012. Optimizing feeding composition
and carbon–nitrogen ratios for improved methane yield during anaerobic co-
digestion of dairy, chicken manure and wheat straw. Bioresour. Technol. 120,
78–83.
Wang, X., Yang, G., Li, F., Feng, Y., Ren, G., Han, X., 2013. Evaluation of two statistical
methods for optimizing the feeding composition in anaerobic co-digestion:
mixture design and central composite design. Bioresour. Technol. 131, 172–178.
Ward, A.J., Hobbs, P.J., Holliman, P.J., Jones, D.L., 2008. Optimisation of the anaerobic
digestion of agricultural resources. Bioresour. Technol. 99, 7928–7940.
Wu, Y., Wang, C., Liu, X., Ma, H., Wu, J., Zuo, J., Wang, K., 2016. A new method of two-
phase anaerobic digestion for fruit and vegetable waste treatment. Bioresour.
Technol. 211, 16–23.
Yan, B.H., Selvam, A., Xu, S.Y., Wong, J.W.C., 2014. A novel way to utilize hydrogen
and carbon dioxide in acidogenic reactor through homoacetogenesis. Bioresour.
Technol. 159, 249–257.
Zhang, Y., Banks, C.J., Heaven, S., 2012. Co-digestion of source segregated domestic
food waste to improve process stability. Bioresour. Technol. 114, 168–178.
Zuo, Z., Wu, S., Qi, X., Dong, R., 2015. Performance enhancement of leaf vegetable
waste in two-stage anaerobic systems under high organic loading rate: role of
recirculation and hydraulic retention time. Appl. Energy 147, 279–286.