Waste Management 35 (2015) 105110

Contents lists available at ScienceDirect

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

Continuous mesophilic anaerobic digestion of manure and rape

oilcake Experimental and modelling study
Sawomir J. Jabonski a, Piotr Biernacki b, Sven Steinigeweg b, Marcin ukaszewicz a,
a
University of Wroclaw, Faculty of Biotechnology, ul. Fryderyka Joliot-Currie 14a, 50-383 Wroclaw, Poland
b
Fachbereich Technik EUTEC Institut, Hochschule Emden/Leer, Constantiaplatz 4, 26723 Emden, Germany

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

Article history: Rape oilcake is a by-product formed after the removal of oil from rapeseed. Due to the high content of
Received 30 April 2014 organic matter rape oilcake seems a good substrate for anaerobic digestion when it cannot be used as fod-
Accepted 13 September 2014 der. The aim of this work was to optimise the parameters used in a mathematical model of anaerobic
Available online 13 October 2014
digestion for rapeseed oilcake and cattle manure. The composition of these substrates was determined
in order to estimate model inputs. Optimised kinetic constants of hydrolysis and decomposition for oil-
Keywords: cake (Kdis = 0.77, KhydCH = 0.55, khydPr = 0.57, khydLi = 0.30) were estimated based on batch fermentation.
Rapeseed
The accuracy of the model with improved input parameters was conrmed by continuous fermentation.
Hydrolysis constants
Agricultural biogas plats
The average concentration of methane in biogas was about 50%. The biogas production efciency from
Anaerobic Digestion Model No. 1 organic matter (dened as volatile solids) was 0.42 m3 kg1 with an organic substrate loading rate equal
Animal faeces to 3.18 kg m3 d1. The fermentation process demonstrated good stability and efciency. The accuracy of
the optimised model seems sufcient for use in modelling of a full scale process.
2014 Published by Elsevier Ltd.

1. Introduction accumulated in the organic matter () but in addition it also allows

Rape (Brassica napus) is the main feedstock used for oil produc-
tion in the European Union (the rapeseed harvest reached 20 Mt in
2009/2010) (Coyatte and Schenk, 2011). After the pressing process
a by-product called oilcake accounts for 65% of seed mass. The
majority of rape oilcake is used as fodder. About 49% of rape oil-
cake is not suitable for this application (because it is microbiolog-
ically contaminated during storage and transport) (Brzska et al.,
2010). Based on this value one can estimate that about 0.51.2
Mt of rapeseed oilcake in Europe currently remains unutilized
annually.
Since the amount of organic matter in oilcakes is high, the use of
this substrate for the production of renewable energy represents a
potential solution. So far three possible ways of energy recovery
from oilcakes have been reported: anaerobic digestion
(Fernndez-Cegr et al., 2012), pyrolysis (Sen and Kar, 2011; Ucar
and Ozkan, 2008) and combustion with fossil fuels.
Anaerobic digestion is a process in which the organic matter is
converted to biogas (a mixture of methane, carbon dioxide and
other trace compounds), biomass and mineral salts. Like the other
two methods this method enables obtaining the energy
Corresponding author. Tel.: +48 71 375 62 50.
E-mail address: marcin.lukaszewicz@uni.wroc.pl (M. ukaszewicz).
the recovery of nitrogen and mineral salts. Furthermore the use of
efuent as a fertilizer reduces the energy consumption required for
the production of fertilizers (especially those containing nitrogen).
The composition of oilcakes seems to make them appropriate
for use as a substrate for anaerobic digestion. The high mass frac-
tion of relatively easily degradable polymers should ensure good
biogas production efciency. Several authors have reported the
use of seed oilcakes as a substrate in anaerobic digestion
(Bohdziewicz et al., 2012; Fernndez-Cegr et al., 2012;
Staubmann et al., 1997). However the appropriate modelling of
the process which is helpful during the introduction of new sub-
strate on an industrial scale has not previously been presented.
An estimation of the possible impact of the addition of rapeseed
oilcake to the other substrates traditionally used in biogas plants
may be obtained from mathematical models such as the Anaerobic
Digestion Model No. 1 (ADM1) (Batstone et al., 2002), developed by
the International Water Association Task Group in 2002. To obtain
a reliable simulation of the anaerobic digestion of a particular sub-
strate some input variables and model parameters have to be
optimised.
The parameters describing the composition of composite sub-
strates (such as plant biomass) may be obtained using the
approach presented by Koch et al. (2010). In this method the pro-
portion of degradable biopolymers (proteins, carbohydrates and

http://dx.doi.org/10.1016/j.wasman.2014.09.011
0956-053X/ 2014 Published by Elsevier Ltd.
106 S.J. Jabonski et al. / Waste Management 35 (2015) 105110
Nomenclature
TS1 total solids
VS volatile solids
CP crude protein
CL crude lipids
CF crude bre
ADF acid detergent bre
NDF neutral detergent bre
lipids) in composite fraction are estimated on the basis of
Weender analysis/van Soest extension (Naumann and Bassler,
1993; van Soest and Wine, 1967), and corrected by an appropriate
factor to match the chemical oxygen demand (COD) unit used in
ADM1.
The decomposition and hydrolysis of composite substrate is
often a rate limiting step during anaerobic digestion (particularly
in the case of plant material) (Vavilin et al., 2008), thus the kinetic
rates of this reaction have a signicant impact on the accuracy of
simulation. Since the rate of disintegration and hydrolysis varies
signicantly depending on the substrates used in anaerobic diges-
tion it is proposed to estimate the kinetic constant for individual
substrates on the basis of batch experiments (Biernacki et al.,
2013). However the batch test does not reproduce completely
the fermentation conditions present in full scale installations and
the model parameters obtained in this way should be compared
with the results obtained from continuous fermentation
(Deublein and Steinhauser, 2008).
The aim of this work was to investigate the possibility of using
rape oilcakes as a substrate for anaerobic digestion and the deter-
mination of selected model inputs and parameters required for
modelling with ADM1. To achieve this goal the composition of
rapeseed oilcake was evaluated. In the second step the kinetic con-
stants describing the disintegration and hydrolysis phases were
optimised on the basis of a batch experiment, as proposed by
Biernacki et al. (2013). In the last stage the accuracy of the param-
eters obtained was veried by the comparison of data from a con-
tinuous fermentation experiment with simulation.
2. Materials and methods
2.1. Substrates and inoculum
Rape plants (cultivar Sherlock and Digger) were cultured on a
farm in Gubczyce 50120 N; 17480 E (Poland). Seeds were har-
vested in July 2011. Rape oilcake was obtained by cold pressing
in NAPUS-OIL S.C. Kietrz 50000 N; 18000 E (Poland). The oil cake
was stored at room temperature before use. The cow manure was
obtained from a farm near Emden 53180 N; 7120 E (Germany) in
April 2012. For fermentation experiments solid particles were
removed from the manure with a 3.6 mm sieve, to prevent the
clogging of pumps and pipes connected to the reactor. Then the
manure was mixed with an appropriate amount of water (in such
a way that the manure mass was 50% of the total mass of substrate)
and ground oilcake, and then stored at 4 C until use.
As the inoculum the sludge from an anaerobic bioreactor was
used. It was obtained from the agricultural biogas plant in Witt-
mund 53300 N; 7540 E (EWE Biogas GmbH, Wittmund, Lower
Saxony, Germany). This facility uses cow manure and food waste
as a substrate. The inoculum was obtained in March 2012 and
the following parameters were measured: pH, ammonium ion con-
centration, alkalinity, total volatile fatty acids, total solids and vol-
atile solids (Table 1).
ADL acid detergent lignin
FOS/TAC volatile organic acids/total inorganic carbonate
COD chemical oxygen demand
NfE nitrogen free extracts
HRT hydraulic retention time
2.2. Batch experiments
The biogas batch experiments were prepared in accordance to
VDI 4630 norm (Verein Deutscher Ingenieure, 2006), as it is
described by Biernacki et al. (2013). The gas production from fer-
mentation batch experiments was measured by the ANKOM Gas
Production System (form ANKOM TECHNOLOGY) (ANKOM
Technology, 2010). The following substrates were tested: rapeseed
oilcake and sieved cow manure. Batch trials were prepared in 1.1 L
borosilicate glass bottles. To 250 cm3 of inoculum 2.5 g of substrate
was added, then the bottles were preincubated at 37 C in a water
bath for 1 h. After incubation the bottles were closed with ANKOM
modules. The anaerobic digestion was wet tape performed under
mesophilic conditions (temperature 37 C). The sludge was manu-
ally mixed for 15 s every 24 h. All fermentation tests were pre-
pared in triplicate. The border pressure in the system was set to
55 kPa. These experiments lasted for 21 days.
2.3. Continuous fermentation
For continuous fermentation a 25 L glass reactor with a water
jacket was used. The temperature in the system was set to 37 C
and was maintained by a water bath with external circulation
model E306 (Lauda Dr. R. Wobser GmbH & CO KG). The prepared
substrate was stored at 4 C and pumped into the reactor with a
cavity pump model I-ID Type 0.03;10 (from Delasco PCM GmbH).
A similar pump was used to remove the digested medium from
the reactor. The uid inside the reactor was continuously mixed
at 0.833 Hz with a laboratory stirrer Eurostar power-B (from IKA
Werke GmbH & Co. KG). The production of biogas was measured
with a gas counter model GT05/5 (from Dr.-Ing. Ritter Apparatebau
GmbH & Co. KG) connected to a computer. The methane concentra-
tion (dened as a volume fraction) was measured with an UNOR
6 N infrared methane detector (SICK MAIHAK GmbH). Dosing
pumps were turned on every 72 min. The volume of dosed sub-
strate per day was about 0.66 L, which corresponded to a hydraulic
retention time (HRT) of 30 days. The reactor was loaded with 15 L
of inoculum and lled up to 20 L with manure; then after 2 days
continuous feeding with manure was started. During the start-up
period the reactor was fed with cow manure diluted with tap water
(1:1) for 21 days. After this period rapeseed oilcake was added to
the feeding substrate, while the concentration of cattle manure
Table 1
Inoculum parameters.
pH 8.2
FOS/TAC 0.243
Alkalinity (CaCO3) 9677 mg L1
Org. acids (CH3COOH) 2352 mg L1
TS 4.08 %
Ash 1.47 %
VS 26.0 g L1
Ammonium N 1440 mg L1
 S.J. Jabonski et al. / Waste Management 35 (2015) 105110 107

remained constant. The concentration of rapeseed oilcake in sub- Table 3

strate suspension (measured as a fresh weight) was increased from Parameters used in modelling (concentration is dened as COD).

20 g L1 up to 80 g L1 over a period of 35 days. Rape seedcake Sieved manure Raw manure
Every day samples were taken for the following analysis: pH, Xc (g kg 1
) 1269.84 84.45 109.47
alkalinity, and total volatile fatty acids. Additionally every three fpr 0.323 0.260 0.192
days the concentration of ammonium ions and the content of vol- i 0.120 0.061 0.052
atile solids were measured. fch 0.425 0.198 0.312
fxi 0.132 0.481 0.444
Degradability 89.50% 49.50% 49.50%
2.4. Analytical methods
Abbreviation description: Xc initial value of variable describing the concentration
of composites (substrate); fpr, i, fch, fxi stoichiometric coefcients describing the
The analysis of substrate composition was carried out according share of individual fractions (proteins, lipids, carbohydrates and inert particles) in
to Wender analysis/van Soest extension (Clesceri et al., 1998; Koch the substrate.
et al., 2010; Naumann and Bassler, 1993; van Soest and Wine,
1967). In all substrates the following parameters were analysed:
total solids (TS), volatile solids (VS), crude protein (CP), crude lipids
(CL), crude bre (CF), acid detergent bre (ADF), neutral detergent purposes of modelling of continuous fermentation of mixed sub-
bre (NDF), acid detergent lignin (ADL) (Clesceri et al., 1998; Koch strates the model parameters (stoichiometric coefcients and
et al., 2010; Naumann and Bassler, 1993; van Soest and Wine, decomposition and hydrolysis kinetic rates) were calculated as
1967; Wichern et al., 2008). The bre bags and detergent solutions weighted arithmetic means. The weights were based on the sub-
used for bre analysis were obtained from Ankom technology. For strate share in the mixture.
drying and incineration an oven Memmert 100800 and mufe
furnace LE4/11/R6 SN 218841; 30-3000 C (Nabertherm) were used
3. Results and discussion
respectively. The determination of Kieldahl nitrogen was done
with the Basic Kjeldahl distillation system VAPODEST 10sn (Ger-
3.1. Substrates composition
hardt GmbH). The reagents (sulphuric acid, K2SO4, CuSO4, TiO2)
were obtained from (Grssing GmbH). The extraction of lipids
All used substrates were investigated with Weender analysis.
was done in standard 100 cm3 Soxlet apparatus with petroleum
Table 2 presents the results of the analysis of the composition of
ether fraction 40/60 obtained from Grssing GmbH.
manure and rapeseed oilcake. The rapeseed oilcake contained a
Alkalinity (dened as CaCO3 mg L1), total volatile fatty acids
low mass fraction of water (total solids 90.48%). The major part
(dened as CH3COOH mg L1) (VFA) and alkalinity to VFA mass
of the dry weight fraction were the organic compounds (volatile
ratio (FOS/TAC in German: Flchtige Organische Suren/Totales
solids 93.23%). The two main components of organic fraction were
Anorganisches Carbonat) were measured with a Biogas Titration
proteins (30.14%) and bre compounds (NDF 39.67%). The lignin
Manager TitraLab (Hach, Germany). pH was measured with a
fraction in rapeseed oilcake is low (ADL 9.1%) ensuring good gas
pH meter pH 538 (WTW GmbH, Germany).
production efciency, since lignin is not digested in anaerobic con-
All data are presented as mean standard deviation (SD). Statis-
ditions (Koch et al., 2010). The composition of analysed rapeseed
tical signicance was determined using a Students t test. The sig-
oilcakes was similar to the results obtained by other authors
nicance level was set at P < 0.05.
(Ramachandran et al., 2007).
The solid fraction in manure was much lower than in the rape-
2.5. Mathematical modelling seed oilcake (8.06% in sieved and 10.79% in raw manure respec-
tively). Also the concentration of organic compounds in solid
The composition of substrate used in the model was calculated fraction was lower in comparison to rapeseed oilcake (65.5% TS
according to the method described by Koch et al. (2010), and mod- in sieved and 71.49% TS in raw manure). The expected methane
ied by Biernacki et al. (2013). The modelling of the fermentation yields from cow manure are lower than those from oilcake due
process was done with SIBMA software version 4.5. The applied to the lower content of lipids and protein, and the higher content
model was the ADM 1 with nitrogen circulation improved accord- of lignin.
ing to Wett et al. (2006). The determination of kinetic constants The analysis of the manure revealed that the sieving process
was done in accordance with Biernacki et al. (2013), and the down- reduced the content of organic matter in the substrate (volatile sol-
hill simplex methods algorithm from Nelder and Mead was applied ids were reduced from 71.49% to 65.50%). One can also observe
(Nelder and Mead, 1965). The degradability parameter (Table 3) that the proportions between individual biopolymers present in
was taken from Kuratorium fr Technik und Bauwesen in der the samples were changed. In sieved manure, proteins and lignin
Landwirtschaft (Wirtschaftlichkeitsrechner Biogas, 2012). For the made up higher part of the volatile solids (proteins 17.05%, lignin

Table 2
Mass fraction composition of substrates used in the fermentation.

Rape seedcake Sieved manure Manure

Total solids 90.48% (0.17%) 8.60% (0.04%) 10.79% (0.04%)
Volatile solids (%TS) 93.23% (0.03%) 65.50% (0.02%) 71.49% (0.08%)
Total lipids (%TS) 11.17% (0.31%) 3.99% (0.30%) 3.72% (0.21%)
Protein (%TS) 30.14% (0.29%) 17.05% (0.79%) 13.73% (0.07%)
Crude bre (%TS) 15.46% (0.05%) 14.61% (0.95%) 21.78% (0.98%)
ADF (%TS) 25.11% (0.28%) 43.02% (0.67%) 40.89% (0.78%)
ADL (%TS) 9.10% (0.51%) 25.87% (0.78%) 18.43% (0.49%)
NDF (%TS) 39.67% (0.29%) 37.04% (1.26%) 44.80% (0.76%)
NfE (%TS) 36.46% 29.86% 32.25%

The numbers in the parenthesis represents the standard deviation of the measurements (n = 3). Differences between tested samples were important with p = 0.05.
108
25.9%) when compared with raw manure (proteins 13.73%, lignin
18.43%). The observed changes are probably caused by the better
availability of substrates in the smaller particles which passed
through the sieve.
The COD to nitrogen ratio in the case of all tested substrates
(29.1 for oilcake and 36.4 for sieved manure) was much lower than
the recommended value (160) (Deublein and Steinhauser, 2008).
Since the nitrogen availability exceeds the demand of biomass syn-
thesis, part of the nitrogen will be released to the medium in the
form of ammonium ions. The increase in the concentration of
ammonium nitrogen may cause inhibition related to the appear-
ance of free ammonia.
The model input data calculation was based on the method
described by Koch et al. (2010), with modications by Biernacki
et al. (2013). The parameters used in the modelling are presented
in Table 3.
3.2. Batch fermentation
For an estimation of disintegration and hydrolysis kinetic con-
stants batch fermentation experiments were prepared. Biogas pro-
duction from individual substrates is illustrated in Fig. 1. The
digestion of rapeseed oilcake was completed in 6 days (Fig. 1A).
The biogas volume reached 1218 cm3, corresponding to a produc-
tion efciency (referring to organic matter measured as VS) of
0.570 m3 kg1. The biogas production efciency obtained during
the batch experiments was comparable to the results obtained by
Bohdziewicz et al. (2012)) (fermentation time 8 days, biogas pro-
duction efciency 0.545 m3 kg1). The production of biogas from
rape oil cake was higher when compared with Jaropha curcas oil-
cake (0.272 m3 kg1) and sunower oilcake (around 0.4 m3 kg1)
(Monlau et al., 2013; Staubmann et al., 1997) digested without
any pre-treatment. The presented results suggest that the anaero-
bic digestion of rape oil cake could be done with low hydraulic
retention times. The digestion process during the batch experiment
was completed in six days, which is four times faster than reten-
tion times in agricultural biogas plants.
Total gas production from sieved manure was 395 cm3 with an
efciency (referring to organic matter measured as VS) of
0.281 m3 kg1. Gas production rates reached the highest level
between the second and fourth days of the experiment and
Total biogas production [L]
1.5
(A)
1
0.5
0
0 5
Total biogas production [L]
1.5
(B)
1 prepared and the results were compared with the simulation.
0.5
0 tation experiments. The composition of substrate and fermentation
0 5
Fig. 1. Total biogas production in batch fermentation trials. Substrates used: (A)
rape seedcake; (B) sieved manure.
S.J. Jabonski et al. / Waste Management 35 (2015) 105110
Table 4
Optimised kinetic constants for hydrolysis and decomposition of selected substrates.
kdis (d1) khydCh (d1) khydPr (d1) khydLi (d1)
Rapeseed cake 0.7716 0.5478 0.5695 0.3036
Sieved manure 0.4691 0.7498 0.0646 0.001
Abbreviation description: kdis kinetic constant of the decomposition process,
khydCh kinetic constant of the hydrolysis of carbohydrates, khydPr kinetic constant
of the hydrolysis of proteins, khydLi kinetic constant of the hydrolysis of lipids.
gradually decreased until the end of the batch trials. Optimised
kinetic constants for hydrolysis and decomposition for rapeseed
oilcake and sieved manure are presented in Table 4.
The decomposition rate optimised for sieved cattle manure was
similar to the results obtained by Wett et al. (0.5 d1) (Wett et al.,
2007). The parameters obtained by the authors of ADM1 were
about two times lower, but these parameters were determined in
other temperature conditions (55 C) and the manure was not
sieved (Batstone et al., 2002). The decomposition constant
obtained by Wichern et al. (0.05 d1) (Wichern et al., 2008) is
one order of magnitude lower in comparison to the presented
results. In experiments presented by Wichern et al. (2008) the sub-
strate was a mixture of manure and fodder, and the parameter was
optimised for this blend. Lower constants could be a result of the
presence of unscrambled plant material.
Hydrolysis constants obtained for both examined substrates are
in the range between 0.3 d1 and 0.8 d1 (only the hydrolysis of
lipids from sieved manure was lower Table 4). Protein and carbo-
hydrate hydrolysis constants optimised for rapeseed oilcake were
around 0.5 d1. These values are usually higher than those
obtained for other popular substrates (range from 0.1 d1 to
0.69 d1) (Batstone et al., 2002). The carbohydrate hydrolysis con-
stant optimised for sieved manure is 0.7498 d1. This value is in
the lower part of the range 0.52 d1 proposed by Garcia-Heras
(2003), but it is higher than the values obtained for other sub-
strates (0.040.58 d1) (Batstone et al., 2002; Lbken et al., 2007).
The decomposition and hydrolysis constants optimised for
rapeseed oilcake were higher in comparison to the results obtained
for cattle manure. The digestion of rapeseed oilcake batch trials
was faster than the digestion of sieved manure. Better bioavailabil-
ity of organic matter probably arises from differences in the struc-
ture of the plant elements present in the substrates used. The
majority of biomass in the case of rapeseed oilcake originates from
cotyledons (Leubner, 2000). In case of B. napus the reserve materi-
als for the embryo are stored in the cotyledons thus the easily bio-
degradable biopolymers (proteins and lipids) represent large mass
fraction of biomass in oilcake. On the other hand biopolymers,
present in stems and leaves (which represent the main position
in the cattle diet) form a scaffold and protective structure in a
plant. The mass fraction of polymers which are hard to biodegrade
Experiment in those parts of a plant is higher than in seeds. Furthermore the
Simulation
fast digestion of oilcake may be caused or enhanced by the break-
10 15 20 25 down of plant tissue during the pressing process.
Time [d]
3.3. Continuous fermentation
To verify the parameters obtained during the batch experiments
Experiment
a continuous fermentation of rapeseed oilcake and manure was
Simulation
The continuous fermentation was also carried out to exclude the
possibility of ammonium inhibition. The input parameters, disinte-
gration and hydrolysis constants were optimised for batch fermen-
10 15 20 25
conditions was similar to that used in agricultural biogas facilities.
Time [d]
Selected fermentation conditions (i.e. temperature, hydraulic
retention time, mixing) reected the conditions in biogas facilities
which are the most probable receivers of this waste oilcake.
 S.J. Jabonski et al. / Waste Management 35 (2015) 105110 109

The data gathered during the process of continuous fermenta- optimal pH range for anaerobic digestion, which is assumed to
tion of cow manure and rapeseed oilcake are presented in Fig. 2. be between 6.8 and 7.5. The FOS/TAC ratio dropped from 0.427
During continuous fermentation the concentration of rapeseed oil- to 0.273 during rst six days of fermentation and remained below
cake (measured as a fresh weight) in substrate was increased from this value until the end of the experiment. Low values of the FOS/
20 to 80 g L1, which corresponds to an increase in the organic load- TAC ratio suggest that the digestion system was still below its opti-
ing rates (expressed as VS) from 0.957 g L1 d1 to 3.18 g L1 d1) mal organic loading rate.
(Fig. 2c). The biogas production rate increased from 0.4 L L1 d1 The concentration of ammonium ions during continuous
up to 1.38 L L1 d1 at the end of the experiment. These values cor- fermentation remained between 1.5 g L1 and 2.5 g L1 (Fig. 2e).
respond to a production efciency (referring to organic matter mea- These values are in the upper range of concentrations optimal for
sured as VS) of 0.42 L g1. The methane concentration at the anaerobic digestion and a further increase in ammonium ion con-
beginning of the experiment was 57% and slightly decreased with centration could cause inhibition problems (Nielsen and
the increase of the organic loading rate to 50% in the last part of Angelidaki, 2008).
experiment (Fig. 2b). The presented results show that the continuous fermentation
During the whole experiment the pH value oscillated between process had good stability. We did not encounter any problems
7.52 and 7.78 (Fig. 2d). These values are only slightly above the with acidication of the reactor nor the accumulation of
ammonium.
The results of modelling and fermentation showed a good cor-
Total biogas production

relation with methane concentration. The difference in the total

1000 (A)
biogas production between the experimental results and the com-
800
puter simulation was only 7.8%. The difference between the biogas
600
[L]

production obtained from the experiment and from the simulation

400 Experiment
is presented in Fig. 2A. The predicted gas production was slightly
200 Simulation
overestimated in the rst part of experiment (days from 1st to
0
0 5 10 15 20 25 30 35 40 25th) and underestimated in the second part of the experiment
Time [d] (days from 25th to 35th). A good correlation between the simula-
tion results and the experimental data for continuous fermentation
conrms the accuracy of the kinetic parameters optimised for
Methane concentration

70%
65%
(B) batch experiments. Further increases in accuracy might require
60% optimisation of other model parameters, for example ammonia
55% inhibition constant acetate uptake. These parameters may
50% signicantly inuence the simulation when the concentration of
45% inhibitory agent is high.
40%
35% 4. Conclusions
0 5 10 15 20 25 30 35 40
Time [d]
Our results show that the anaerobic digestion of rapeseed oil-
cake with cattle manure is one of the possible ways to utilize mate-
Organic loading rate

3,5 3,5%
(C)
VS mass fractio in

3 rial which cannot be used as fodder. Batch fermentation

3,0%
[g L-1 d-1]

2,5 experiments and composition analysis show that the anaerobic

sludge

2 2,5%
digestion of rapeseed oilcake is faster than the digestion of other
1,5 2,0%
1 OLR substrates often used in agricultural biogas plants. The fermenta-
1,5%
0,5 VS tion process showed good stability indicated by low VFA concen-
0 1,0%
0 5 10 15 20 25 30 35 40
tration, pH values in the optimal range and good biogas
production efciency. Thus the hydraulic retention time in systems
Time [d]
treating this substrate might be shorter than 20 days.
8 (D) 0,6 The method of calculation of kinetic constants using values
7,9
7,8 0,5 optimised on the basis of batch experiments proved to be suitable
FOS/TAC

7,7 0,4 in the modelling of continuous fermentation. The model of a con-

7,6
pH

7,5 0,3 tinuous fermentation process based on the optimised hydrolysis

7,4
7,3 0,2
7,2
pH constants and substrate composition analysis showed a 7.8% devi-
0,1
7,1
7
FOS/TAC
0
ation from experimental results. Such accuracy should be sufcient
0 5 10 15 20 25 30 35 40 for the use of this optimisation method in the modelling of full
Time [d] scale processes, which operate in conditions close to optimal for
anaerobic digestion.
3000 (E)
2500 Acknowledgments
NH4+ [mg/l]

2000
1500
This work is a result of the training nanced as a part of the pro-
1000
500
ject Development of the potential and educational offer of the
0 University of Wrocaw the chance to enhance the competitive-
0 5 10 15 20 25 30 35 40 ness of the University (No. POKL.04.01.01-00-054/10). The
Time [d] research was also nancially supported from the key project
POIG.01.01.02-00-016/08 The model agro-energetic complexes
Fig. 2. Parameters describing reactor operation during continuous fermentation.
(A) Total biogas production; (B) volume fraction of methane; (C) organic loading
as an example of distracted cogeneration based on local and
rate (g L1 d1) and mass fraction of volatile solids in the digested sludge (%); (D) pH renewable energy sources and by Wroclaw Centre of Biotechnol-
and FOS/TAC ratio; (E) concentration of ammonium ions in reactor (mg L1). ogy, programme. The Leading National Research Centre.
110 S.J. Jabonski et al. / Waste Management 35 (2015) 105110

References Monlau, F., Latrille, E., Da Costa, A.C., Steyer, J.-P., Carrre, H., 2013. Enhancement of
methane production from sunower oil cakes by dilute acid pretreatment. Appl.
Energy 102, 13.
ANKOM Technology, 2010. ANKOM RFS Gas Production Base Kit. http://
Naumann, C., Bassler, R., 1993. Die Chemische Untersuchung von Futtermitteln.
www.ankom.com/.
VDLUFA-Verlag, Darmstadt.
Batstone, D.J., Keller, J., Angelidaki, I., Kalyuzhnyi, S.V., Pavlostathis, S.G., Rozzi, A.,
Nelder, J.A., Mead, R., 1965. A simplex method for function minimization. Comput. J.
Sanders, W.T.M., Siegrist, H., Vavilin, V.A., 2002. Anaerobic Digestion Model No.
7, 13.
1. Scientic and Technical Report 9. International Water Association (IWA),
Nielsen, H.B., Angelidaki, I., 2008. Strategies for optimizing recovery of the biogas
London.
process following ammonia inhibition. Bioresource Technol. 99, 8001.
Biernacki, P., Steinigeweg, S., Borchert, A., Uhlenhut, F., 2013. Application of
Ramachandran, S., Singh, S.K., Larroche, C., Soccol, C.R., Pandey, A., 2007. Oil cakes
Anaerobic Digestion Model No. 1 for describing anaerobic digestion of grass,
and their biotechnological applications a review. Bioresource Technol. 98, 9.
maize, green weed silage, and industrial glycerine. Bioresource Technol. 127,
S
en, N., Kar, Y., 2011. Pyrolysis of black cumin seed cake in a xed-bed reactor.
94.
Biomass Bioenergy 35, 304.
Bohdziewicz, J., Piotrowski, K., Cebula, J., 2012. Kinetyka chemiczna fermentacji
Staubmann, R., Foidl, G., Foidl, N., Gbitz, G., Lafferty, R., Valencia Arbizu, V., Steiner,
metanowej makuchu rzepakowego. In: Cenian, A. et al. (Eds.), Ekoenergetyka
W., 1997. Biogas production from Jatropha curcas press-cake. Appl. Biochem.
Biogaz. Wydawnictwo Gdanskiej Szkoy Wyzszej, _ Gdansk, pp. 2427.
Biotechnol. 6365, 67.
Brzska, F., Hanczakowska, E., Koreleski, J., Strzetelski, J., Swiatkiewicz, S., 2010.
Ucar, S., Ozkan, A.R., 2008. Characterization of products from the pyrolysis of
_
Pasze rzepakowe w zywieniu zwierzat. Polskie Stowarzyszenie Producentw
rapeseed oil cake. Bioresource Technol. 99, 6.
Oleju, Warszawa, pp. 910.
van Soest, P.J., Wine, R.H., 1967. Use of detergents in the analysis of brous feeds IV.
Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1998. 2540 Solids, Standard Methods for
Determination of plant cell-wall constituents. J. Assoc. Ofcial Anal. Chem. 50,
the Examination of Water and Wastewater, 20th ed. APHA American Public
9.
Health Association, Washington, DC.
Vavilin, V.A., Fernandez, B., Palatsi, J., Flotats, X., 2008. Hydrolysis kinetics in
Coyatte, C., Schenk, H., 2011. Main Agricultural Products, Agriculture and Fishery
anaerobic degradation of particulate organic material: an overview. Waste
Statistics Main results 20092010. Publications Ofce of the European Union,
Manage. 28, 51.
Luxemburg, p. 50.
Verein Deutscher Ingenieure, 2006. Fermentation of organic material.
Deublein, D., Steinhauser, A., 2008. Biogas from Waste and Renewable Resources.
Characterization of the Substrate, Sampling, Collection of Material Data,
VCH Verlag GmbH & Co. KGaA, Weinheim.
fermentation tests, Verein Deutscher Ingenieure (VDI), Dsseldorf.
Fernndez-Cegr, Victoria., ngeles De la Rubia, M., Raposo, F., Borja, R., 2012. Effect
Wett, B., Eladawy, A., Ogurek, M., 2006. Description of nitrogen incorporation and
of hydrothermal pretreatment of sunower oil cake on biomethane potential
release in ADM1. Water Sci. Technol. 54, 76.
focusing on bre composition. Bioresource Technol. 123, 29.
Wett, B., Schoen, M., Phothilangka, P., Wackerle, F., Insam, H., 2007. Model-based
Garcia-Heras, J.I., 2003. Reactor sizing, process kinetics, and modelling of anaerobic
design of an agricultural biogas plant: application of Anaerobic Digestion Model
digestion of complex wastes. In: Mata-Alvarez, J. (Ed.), Biomethanization of the
No. 1 for an improved four chamber scheme. Water Sci. Technol. 55, 8.
Organic Fraction of Municipal Solid Wastes. IWA Publishing, London.
Wichern, M., Lbken, M., Horn, H., Schlattmann, M., Gronauer, A., 2008.
Koch, K., Lbken, M., Gehring, T., Wichern, M., Horn, H., 2010. Biogas from grass
Investigations and mathematical simulation on decentralized anaerobic
silage measurements and modeling with ADM1. Bioresource Technol. 101, 65.
treatment of agricultural substrate from livestock farming. Water Sci.
Leubner, G., 2000. The seed biology place.
Technol. 58, 72.
Lbken, M., Wichern, M., Schlattmann, M., Gronauer, A., Horn, H., 2007. Modelling
Wirtschaftlichkeitsrechner Biogas, 2012. Kuratorium fr Technik und Bauwesen in
the energy balance of an anaerobic digester fed with cattle manure and
der Landwirtschaft. https://www.ktbl.de/.
renewable energy crops. Water Res. 41, 96.