Industrial production of biogas through co-digestion of waste glycerol and sewage sludge

(1)

Industrial production of biogas through

co-digestion of waste glycerol and sewage sludge

Research upon the effects of addition of crude glycerol to a large scale digestion chamber of a municipal wastewater treatment plant.

Gustav Fröléen froleen@kth.se

Stockholm 2016

Industrial biotechnology School of biotechnology Kungliga Tekniska Högskolan

(2)

I

Abstract

Biogas production poses a good example of how wastes can be a useful resource for society. In this master thesis large scale production of biogas through co-digestion of sewage sludge from Henriksdal wastewater treatment plant with crude glycerol from the biodiesel industry is evaluated with the goal to increase the methane production twofold.

The production of biogas from two full-scale reactors was monitored over time where one reactor was given glycerol and sewage sludge whilst the other reactor received only sewage sludge, thus becoming a reference. 78% of the theoretical methane potential of the added glycerol was obtained at a loading rate double of the control. The consequence of the glycerol addition was an increase in methane production by 74%, a decrease in pH by 0,25 pH-units and in ammonium by 37%. The TS (total solids) increased by 18% while VS (volatile solids) and VFA (volatile fatty acids) concentrations did not change. Finally, bicarbonate alkalinity showed a trend where it dropped by up to 10%, heavy metals concentration did not change and the reduction of VS increased with 38 % at an OLR of glycerol at 1,5 kg VS/m³,day.

It was concluded that with the addition of crude glycerol a near doubling in methane production could be reached, accompanied with improvements in digester condition regarding the reduced ammonium concentration. Through the addition of glycerol there was also a lowering of the pH in the digester. A stability evaluation was performed and showed that the digester remained stable with the addition of glycerol.

(3)

II

Sammanfattning

Biogasproduktion utgör ett bra exempel på hur avfall kan vara en användbar resurs för samhället.

I denna masteruppsats utvärderas storskalig produktion av biogas genom samrötning av avloppsslam från Henriksdal avloppsreningsverk med orenat glycerol från biodiesel-industrin med målet att öka metanproduktionen tvåfaldigt.

Produktionen av biogas från två fullskaliga reaktorer övervakades med tiden där en av

reaktorerna tillfördes glycerol och avloppsslam medan den andra reaktorn enbart beskickades med avloppsslam och därmed blev en referens. 78% av den teoretiska metanpotentialen hos det tillförda glycerolet erhölls vid en belastning som var dubbelt så stor som referensreaktorns.

Konsekvenserna av glyceroltillförseln var en ökning i metanproduktion på 74%, en sänkning i pH med 0,25 pH-enheter och en sänkning i ammoniumkoncentration om 37%. TS

(torrsubstansen) ökade med 18% medan VS (glödförlust) och VFA (organiska syror) var oförändrat. Slutligen visade bikarbonat-alkaliniteten en trend där den sjönk med upp till 10%, tungmetallkoncentrationen förändrades inte och reduktionen av VS ökade med 38% vid en OLR av glycerol på 1,5 kg VS/m³,dygn.

Slutsatsen drogs att med tillförseln av obearbetat glycerol kunde en nära fördubbling av metanproduktionen uppnås tillsammans med förbättringar i rötkammar-förhållanden med avseende på den reducerade ammonium-koncentrationen. Genom tillförseln av glycerol observerades också en sänkning av pH i rötkammaren. En stabilitetsutvärdering genomfördes och visade att rötkammaren bibehöll sin stabilitet med tillförseln av glycerol.

(4)

III

Acknowledgements

There are many people who have greatly contributed to the possibility of carrying out this master thesis. To start with, I would like to thank my three main supervisors Björn Magnusson, Jörgen Ejlertsson, and Martin Johansson who have been central to the progress of my work. A thank you also goes to the project owner Ragnar Stare who have contributed with a lot of valuable

experience within the profession of industrial engineering and project management. I would like to thank Anna Karlsson at Scandinavian Biogas for providing valuable feedback on the thesis report.

As for my orientation in this project at Henriksdal, I would like to thank Andreas Carlsson at SVAB for taking the time and effort of continuously helping me get to know the WWTP and many of the procedures necessary to carry out for the project. Thanks to Ida Andersson and Carina Almé at Scandinavian Biogas R&D in Linköping for teaching me how to perform a number of central analysis methods and helping me with a variety of issues that presented themselves during the master thesis.

I would also like to thank my teacher Gen Larsson, not only for putting up as examiner of this master thesis but also for the many interesting and educational courses she has held. These courses are largely accountable for making me realize my interest in waste water purification and production of biofuels, not to mention that it was she who encouraged me to contact

Scandinavian Biogas in the first place which resulted in this highly appreciated master thesis.

Honestly put, my master thesis and my time at Henriksdal WWTP and the facilities of Scandinavian Biogas in Stockholm and Linköping has without doubt been one of the most exciting, challenging and rewarding times of my life. Never before have I been able to give vent to my curiosity in such an honest way. This thesis has been a pleasure to work with.

(5)

Abbreviations

WWTP: Wastewater treatment plant SBF: Scandinavian Biogas Fuels AB SVAB: Stockholm Vatten AB

CSTR: Continuously stirred tank reactor VFA: Volatile fatty acids

TS: Total solids VS: Volatile solids FS: Fixed solids

OLR: Organic loading rate VSR: Volatile solids reduction

(8)

2

Introduction

All modern societies depend on a handful of basic resources where food, water and energy constitute the foundation. These resources should be utilized in the best possible way, one example is the utilization of wastes for production of biofuels. Digestion of sewage sludge offers a possibility of both reducing the amount of waste sludge to be managed and producing

renewable biofuels. Many of the larger wastewater treatment plants (WWTP) found worldwide are digesting the sludge they receive to produce biogas that can then be sold to external

consumers or be used internally.

The company Scandinavian Biogas Fuels (SBF) wishes to increase the production of biogas from the anaerobic digesters at Henriksdal WWTP and in order to do this, substrate in addition to sludge from the WWTP is needed. Previous lab-scale studies performed by the company have shown that an increase in production of methane by almost 100% is possible if additional organic material in the form of waste glycerol from the biodiesel industry is supplied as a co-substrate . Henriksdal WWTP, owned by Stockholm Vatten AB (SVAB) and located in Nacka, Stockholm, hosts the digesters where the full-scale pilot test for SBFs co-digestion was performed. Data from two full-scale digesters (D5 and D6) were used, both of the model Continuously Stirred Tank Reactor (CSTR) with a total volumetric capacity of 6900 m³each and a liquid volume of 6700 m³. Reactor D6 was used as a reference whilst D5 was used for the co-digestion of sludge and glycerol.

Presently, state of the art technology for digestion of sewage sludge and wastewater covers one- phase systems. Continuous Stirred Tank Reactors (CSTR) and Upflow Anaerobic Sludge Blanket (UASB) reactors are the two main systems employed for wet digestion (total solids <

10%, UASB often utilizes lower TS-percentages), where CSTR is more suited for digestion of waters with high content of suspended solids and the UASB technics is mainly used to treat large volumes of water with low levels of suspended solids. The CSTR is the most common one of these two (Weiland, P. 2010).

The two digesters used in the pilot test have a HRT (hydraulic retention time) of 18 days (Carlsson A., 2015). The two digesters were run under conditions as similar as possible during the pilot test to ensure that a proper reference digester is available. This, in combination with a supply of sludge that varies in quality and quantity due to changes in ingoing concentration of organic material, makes an evaluation in full scale an important step towards implementation.

Goal

The goal of this thesis is to find out if it is possible to increase the production of methane twofold through addition of crude glycerol to a process digesting sewage sludge and evaluate the

performance of the digester during this process.

(9)

3

Strategy

The strategy employed to fulfill the goal is based on the continuous surveillance of the digesters 1) prior to addition of crude glycerol, 2) as the feed of crude glycerol is gradually increased and 3) after the digester has stabilized at the maximum planned glycerol feed. As the digestion involves a complex array of metabolic processes, many parameters can be affected. Firstly, production (and consumption) of VFA is likely to increase with the OLR which in turn can cause a drop in pH (Kolesárová et al., 2011). More easily available substrate could increase the active cell mass, causing an increase in TS and/or VS (Athanasoulia et al., 2014), along with a decrease in ammonium due to assimilation of nitrogen during cell growth. With a raise in VFA, alkalinity may also drop. All of these parameters were surveyed continuously.

In order to evaluate the yield from the conversion of glycerol to methane, the COD in the crude glycerol was followed by chemical analysis during the course of the pilot test and the methane obtained was compared to the theoretical yield of 0,35 Nm³ CH₄/ kg COD. Lastly, heavy metals are present in the crude glycerol (originating from the production of biodiesel) and the

concentration of them in the digested sludge was analyzed before and after introduction of the glycerol.

(10)

4

Hypothesis

The hypothesis is based on previous lab-scale studies carried out by SBF and on outcomes of the pilot test that affects the activities of Henriksdal WWTP.

The first part of the hypothesis is that more than 85% of supplied COD as glycerol will be converted to methane as this is the value obtained in SBFs lab-scale studies. A 85% conversion would at a total loading of 3 kg VS/m³(equal contributions from WWTP sludge and crude glycerol) and day result in a doubling of the methane-production from reactor D5.

The second part of the hypothesis is that there will be changes in VFA, TS and ammonium (see questions below for specifications). The addition of glycerol can affect the digester condition and this can in turn affect on the activities of Henriksdal WWTP. A disturbance in the anaerobic digestion process by the addition of glycerol will likely result in that the digesters concentration of VFA increase with the increase in OLR which can cause destabilization of the digester.

Regarding the activities of Henriksdal WWTP that might be affected by the glycerol addition two parameters are of interest; the digestates concentration of ammonium and its total solids content. The reject water produced from thickening the digestate (see figure 1) is reintroduced to the biological treatment of the WWTP and changes in ammonium concentration in the digestate will thus affect the workload of the WWTP. Changes in the total solids affects the amount of thickened digestate that needs to be disposed, as well as the workload of the centrifuges that thickens the digestate.

In order to evaluate possible negative effects of the glycerol additions on the status of the AD process, the following three questions are posed:

A. Will the concentration of volatile fatty acids (VFA) increase from 150 mg/L to 300 mg/L with the addition of glycerol?

B. Will the total solids (TS, [%, m_TS/m_sample]) increase with more than 10% with the addition of glycerol?

C. Will the concentration of ammonium decrease from 800 mg/L to 500 mg/L with the addition of glycerol?

If the answer to any of these three questions is yes, it is concluded that the hypothesis is false.

(11)

5

Background

Henriksdal WWTP

Henriksdal WWTP treats the wastewater from the municipalities Nacka, Tyresö, Haninge, Huddinge and large parts of central and southern Stockholm. The WWTP thus manages the wastewater from almost 780.000 people and has a volumetric capacity of 250.000 m³ per day (Stockholm Vatten, 2015). The WWT starts with screening of larger solids which is then followed by pre-aeration, pre-sedimentation and biological removal of phosphorus and nitrogen in a activated sludge process. The procedure is finished with a post-sedimentation and

flocculation of the remaining particles that are caught in a sand-filter prior to release of the purified water into the local water Saltviken. The WWTP has seven digesters with a total volumetric capacity of 38.500 m³ available for production of biogas, with D5 and D6 being the largest ones with a liquid volume of 6700 m³ per digester. The digesters of Henriksdal WWTP are fed with primary sludge from the pre-sedimentation and surplus sludge from the active sludge process. An overview of the WWTP can be seen in figure 1.

Figure 1. The structure of Henriksdal WWTP. (modified from Stockholm vatten, 2015).

The digestion in this pilot test was mesophilic at 37 °C and the substrates to be digested, apart from crude glycerol in D5, consisted of a mixture of primary sludge, from the pre-sedimentation of the WWTP, and surplus sludge (also known as biosludge) from the activated sludge process that constitutes the biological purification (see figure 1). The volumetric ratio between the feed of primary sludge and the surplus sludge is approximately 6:1. As supply of sludge and its properties varies with rainfall, snow melting and disposal of organic material from society the amount and composition of the sludge will vary somewhat over time. The sludge from the WWTP accounts for about 1,5 kg VS/m³ day in terms of OLR in D5 and D6.

(12)

6

Biochemistry of digestion

Microbial digestion of organic substrates for the production of biogas is a complex process encompassing complete ecosystems where the waste products of one organism is used as food for other organisms. The process of biogas formation can be divided into four main steps:

hydrolysis, acidogenesis, acetogenesis and methanogenesis (Jarvis & Schnürer, 2009). Figure 2 summarizes the relationship between these steps.

Figure 2. Steps necessary to convert complex organic substances to biogas. The four main steps encompass hydrolysis, acidogenesis, acetogenesis and methanogenesis (modified from Jarvis &

Schnürer, 2009).

Production of biogas is an anaerobic process where, instead of oxygen, organic substances and CO2 are used as electron acceptors. The catabolic pathways for different organic substances will depend on the microbial flora present in the culture, process parameters such as pH, temperature and HRT and the type/structure of the organic material digested (substrate).

(13)

7

The hydrolysis is the initial step where particles and larger polymers are broken down into simpler compounds, making them available for fermentation. Large polymers such as cellulose and proteins takes longer time before available for succeeding digestive steps, as compared to glycerol which can be more directly metabolized by the cells (Yazdani & Gonzalez, 2007). In this sense, digestion of sewage sludge is slower than digestion of glycerol.

After the hydrolysis comes the acidogenesis. Products of this process encompasses mainly organic acids (such as VFA) and alcohols.

The acetogenesis constitutes a critical step in the biogas production. In this step, protons are used as electron acceptors, resulting in the formation of hydrogen gas and in order for this to be thermodynamically favorable, the concentration of the latter product needs to be kept low (Jarvis

& Schnürer, 2009). Thus, consumption of the hydrogen gas by methanogenic organisms is essential in order to allow these oxidative microorganisms to access energy to maintain metabolism and growth. In this step longer VFAs are also metabolized into acetic acid.

The final step of the biogas formation is the methanogenesis. Depending on the substrate, this step can be rate limiting as methanogenesis is a comparatively slow process in the digestion process. Two main pathways exist for methane formation, one using carbon dioxide and hydrogen gas as substrate and one using mainly acetate. Figure 2 shows these two pathways, were they are referred to hydrogenotrophic and acetotrophic methanogenesis.

Glycerol belongs to the mono- and oligomeric category of chemical intermediates and is a natural product formed from the hydrolysis of lipids. External addition of glycerol causes mainly two effects: increased availability of substrate to oxidative microorganisms and decreased

nitrogen/carbon ratio. The easily available substrate introduced in the form of crude glycerol is digested considerably faster than sewage sludge as no hydrolysis is necessary. An increase in the C/N ratio is often beneficial to the digester as excess ammonia can be harmful to the microflora, down to about 200 mg/L (Maes et al., 2013). With the low C/N ratio of sewage sludge, adding carbon rich substrates (such as glycerol) for co-digestion is advantageous for the digester (Heo, N. et al., 2004).

Crucial to the hydraulic retention time of the reactor is the replication time of the methanogenic microorganisms. As these often have a replication time that is measured in weeks, the HRT must be longer than this replication time, otherwise a sufficient concentration of methanogenic

organisms can not be maintained. Likewise the HRT must not be too high for the process to be economically feasible. A too high HRT can also cause excessive amounts of microbes to

accumulate in the digester. Two digestion temperatures are commonly employed; mesophilic and thermophilic temperatures. To some extent, the higher the temperature is the higher the

metabolic activity will be and thus digestion will increase. HRT for digestion in a CSTR usually varies between 10 to 25 days (Jarvis & Schnürer, 2009). Also, a suitable HRT depends on

substrates and access to nutrients. The HRT of the digesters used in this master thesis was on average 18 days.

There are many parameters that may change and affect the anaerobic digestion. VFA, as

mentioned above, is produced during acidogenesis and changes with the OLR of the digester as

(14)

8

well as the biodegradability of the substrates. If too much of the substrate is broken down too quickly the concentration of VFA will rise, a result of organic overloading (too high OLR).

Organic overloading can cause acidification if production of VFA becomes too high for the methanogenic microorganisms to handle in the conversion of VFA to biogas.

pH depends on numerous factors, where production of VFA and consumption of alkalinity are the main ones. Alkalinity buffers the digester so it remains stable but if production of VFA becomes too prevalent the alkalinity can be consumed to the extent where the digester loses stability. Another aspect of potential changes in pH is the concentration of ammonium in the digester. The concentration of ammonium can change with the substrate composition and with changes in the microbial concentration (due to the use of nitrogen for proteins).

TS and VS depends on many parameters, especially the OLR and the biodegradability of the substrates. If the OLR is high the TS can rise as more solids are introduced to the digester. If a substrate with high VS-content is excessively introduced to the digester some of it may pass through the digester undigested (this also depends on the biodegradability of the substrates), hence raising the VS of the digestate.

(15)

9

Materials and methods

Cultivation technology

The pilot phase spanned over a time period of 158 days. On day 70 addition of crude glycerol was initiated at 0,5 kg VS/m³,day. The digesters were allowed to stabilize with the feed until day 124, then the OLR of glycerol was increased with 0,25 kg VS/m³,day every day until the target OLR of 1,5 kg VS/m³,day was reached at day 127. This OLR was then maintained until day 148, giving a total of 21 days at the target OLR. Unless anything else is stated, the data presented in the results section are based on analyses spanning from day 140 to day 148. Presentation of the number of data points used for the different analysis parameters can be seen in table 2.

The feed of the glycerol was initially done batchwise every 6 minutes. The programming for the feed of crude glycerol erred in the beginning of the pilot study, but this was solved by

circumventing the program mathematically. At one point, the sludge valve regulating the feed of sludge to D5 started leaking (day 6 to 13) and the addition of sludge was stopped until the valve had been repaired (day 13 to 18). Due to the batchwise additions of glycerol very large volumes of biogas were produced when the OLR of glycerol was increased. The safety valves did not manage these sharp increases in gas production and the feeding was therefore changed to continuous by exchanging the original open/shut valve for a control valve.

A description of the digester setup and the components of Henriksdal WWTP that is directly connected to the digesters can be seen in figure 3 with table 1 explaining the denotations used.

Figure 3. The, for this study, relevant details of Henriksdal WWTP and the pilot test setup.

Denotations are explained in table 1.

(16)

10 Table 1. Key to denotations in figure 3.

Denotation Meaning

C1 Centrifuge for thickening of surplus activated sludge

C2 Centrifuge for managing digested sludge; separates reject water from thickened digestate.

S1 Sample point for TS and VS measurement on primary sludge.

S2 Sample point for TS and VS measurement on surplus sludge.

S3 Sample point for crude glycerol introduced to D5.

S4 Sample point for analysis of biogas from D5, manual and automatic.

S5 Sample point for digestate from D5.

S6 Sample point for analysis of biogas from D6, manual and automatic.

S7 Sample point for digestate from D6.

(17)

11

Data collected during the pilot test are presented in table 2.

Table 2. Total amount of data points collected for each analysis performed.

Analysis

Number of data points collected during entire

pilot test

after one HRT of maximum OLR

Biogas production and methane concentration

148 9

pH 60 4

VFA 57 4

Ammonium 23 3

TS and VS (for digestate) 39 3

TS and VS (for primary and surplus sludge)

17 1

Alkalinity 24 1

Heavy metals 6 1

Volatile solids reduction 21 2

pH

pH was measured with a pH-meter of the model WTW Inolab pH 730, coupled to a pH-electrode of the model Hamilton Polylite Bridge Lab and a thermometer of the model WTW TFK 325. A two point calibration at pH 4 and 7 was performed once a week, in combination with use of reference solution prior to each measurement in order to ensure reliable results. pH was measured three times a week on digestate from the two digesters.

VFA

VFA was measured with Hach-Lange VFA-tests (product code LCK 365). The digestate was diluted twofold and the analysis was then performed according to the instructions supplied with the analysis-kit. The uncertainty of the analysis method is 1,32 % (Hach Lange, Quality

certificate Technical data for Validation of LCK365). The amount of VFA is given as acetic acid equivalents. Concentration of VFA was measured three times a week on the digestate.

Ammonium

Ammonium was measured with Hach-Lange ammonium tests (product code LCK 302). The digestate was diluted tenfold and the analysis was then performed according to the instructions supplied with the analysis-kit. The uncertainty of the analysis method is 1,49 % (Hach Lange, Quality certificate Technical data for Validation of LCK302). Concentration of ammonium was measured once a week on the digestate.

(18)

12

COD

COD was measured with Hach-Lange COD tests (LCK 014). A sample of crude glycerol was diluted 200-fold in water and the analysis was then performed according to the instructions supplied with the analysis-kit. The uncertainty of the analysis of COD is 0,74 % (Hach Lange, Quality certificate Technical data for Validation of LCK014). COD was measured after every new batch of crude glycerol had been introduced to the glycerol tank, hence the frequency for this analysis rose with the OLR.

TS and VS

TS and VS was measured according to a modification by SBF from the Swedish standard (SS 028112). This was done by weighing in sludge in a weighed, dried crucibles and noting the sample mass. Each filled crucible was then put in an oven at 105 °C for 20 hours. After this the crucible was put to cool in an exsiccator and then weighed again. The remainder of the sample is the TS and it is calculated as the remaining mass of the sample divided by the original sample mass.

After this the crucible was put into a furnace at 550 °C for two hours and thereafter allowed to cool in the exsiccator. The remainder of the sample (the fixed solids) is the VS subtracted from the TS, i.e. the VS has been combusted through the heat of the furnace and no longer remains.

Thus, the VS (given as % of TS) is calculated according to the formula:

𝑉𝑆_{𝑚𝑉𝑆/𝑚𝑇𝑆} =𝑚_𝑇𝑆− 𝑚𝑎𝑓𝑡𝑒𝑟 𝑐𝑜𝑚𝑏𝑢𝑠𝑡𝑖𝑜𝑛

𝑚_𝑇𝑆

The uncertainty of measurement for TS and VS were based on data handed by SBF. These values were 1,0 % for TS and 0,4 % for VS. TS and VS were measured two times a week for digested sludge and once a week for primary sludge and surplus sludge.

Calculations for volatile solids reduction (VSR) can be seen in appendix 1C where two different equations were used, the mass balance equation (MBE) and ash content equation (ACE). This calculation usually requires large amounts of data spanning longer periods of time then the 16 days at the maximum loading rate available from the pilot test. As TS of the primary sludge was considered to be unstable throughout the lapse of the pilot test (see appendix 1C, figure 1) the average feed of primary sludge and surplus sludge was used and VSR was calculated week wise with values for TS and VS of primary sludge, surplus sludge and digestate from the same day of the week.

Though COD was measured for the crude glycerol, its VS was never determined. As the VS of the crude glycerol was needed for calculating the VSR, the analysis sheets supplied by the companies delivering the glycerol was consulted. From this the total mass of 85% organic material was approximated as the total amount of VS for the crude glycerol. Likewise, the FS of the glycerol was approximated to 10%.

(19)

13

Alkalinity

Values for carbonate alkalinity was obtained by sending samples to Eurofins Environment AB Sweden for analysis. According to Eurofins Environment AB Sweden the uncertainty of the measurement is 10 % (Dovberg, M., personal communication). This analysis was done every second week for the digestate.

Heavy metals

Values for heavy metals in the digestate were obtained by sending samples to Eurofins

Environment AB Sweden for analysis once every month. The heavy metals data presented in this master thesis encompassed copper, zinc, mercury, nickel, lead, chromium and silver, which are considered to be especially hazardous. These heavy metals are the most interesting ones for REVAQ-certification, which is an important parameter in the management of the thickened digestate (Svenskt vatten, 2015). The uncertainty of measurement according to Eurofins Environment AB Sweden is 15% for copper, 15% for zinc, 25% for mercury, 15% for nickel, 25% for lead, 15% for chromium, and 20% for silver.

Methane concentration and biogas production

The concentration of methane in the biogas [%], and the production of biogas [Nm³ / h], was measured continuously by stationary analysis equipment. Data from this equipment was recorded every few seconds, where daily average values for methane fraction and gas flow was used in this master thesis. The flow of substrates in and biogas out of the digesters was measured with totalizers which gives one pulse for every m³ of gas passing through. The composition of the biogas (CH4 [%], CO2 [%], O2 [%], H2S [ppm] and other gases [%], mainly N2 and water vapor) was also measured manually with a Geotech Biogas Check on a daily basis.

Calculations on theoretical methane yield from glycerol

The theoretical yield for glycerol converted to methane is given in appendix 1A.

Glycerol feed

D5 was fed with glycerol semi-continuously every 6 minutes. However, as mentioned above, due to difficulties in managing the vastly increased biogas production, continuous feed was initiated at day 120.

(20)

14

Results

The total OLR of each digester and OLR in the form of crude glycerol during the pilot phase can be seen in figure 4. The high peaks could be due to build up and subsequent feeding of primary sludge with higher content of TS, TS that has accumulated during the screening of bulky materials in the pre-treatment of the incoming sludge (see figure 1).

Figure 4. The total OLR of each digester during the pilot phase and OLR in terms of glycerol feed.

All of the results presented were calculated from values obtained after one HRT with maximum OLR if nothing else is stated (day 140-148).

0,00 0,50 1,00 1,50 2,00 2,50 3,00 3,50 4,00 4,50 5,00

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150

O LR [kg V S/m3 , da y]

Day

Organic loading of digesters

Glycerol only D5, total D6, total

(21)

15

Biogas and methane production

Biogas production was 394 ± 41 Nm³/h for D5 and 216 ± 16 Nm³/h for D6 at the end of the experiment. This corresponds to a 82 % increase in biogas production through the introduction of crude glycerol. The development of biogas production during the pilot phase can be seen in figure 5.

Figure 5. Development of biogas production (Nm³/h) during the pilot phase for digesters D5 (amended with glycerol from day 70) and D6 (control).

0,00 0,50 1,00 1,50 2,00 2,50

0,0 50,0 100,0 150,0 200,0 250,0 300,0 350,0 400,0 450,0 500,0

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 Glycerol feed [kg VS / m3, day]

Methane productivity [Nm3/h]

Day

Biogas production

D5 D6 Glycerol feed

(22)

16

Methane production was 239 ± 24 Nm³/h for D5 and 137 ± 10 Nm³/h for D6 at the end of the experiment. This corresponds to a 74% increase in methane production through the introduction of crude glycerol. The development of methane production during the pilot phase can be seen in figure 6.

Figure 6. Development of methane production (Nm³/h) during the pilot phase for digesters D5 (amended with glycerol from day 70) and D6 (control).

0,00 0,50 1,00 1,50 2,00 2,50

0,0 50,0 100,0 150,0 200,0 250,0 300,0

Methane productivity [Nm3/h]

Day

Methane production

D5 D6 Glycerol feed

(23)

17

Yield for conversion of glycerol to methane

At maximum OLR (3 kg VS/m³,day of which 50 % was glycerol), the yield for the conversion of crude glycerol to methane was calculated to be 78 ± 4 % in terms of COD. For calculations on theoretical COD in methane, see appendix 1A. The yield during the course of the pilot test can be seen in figure 7.

Figure 7. Development of yield in terms of CODmethane/CODglycerol during the experiment. No data from the stationary gas measuring equipment was available before day 40.

Due to fluctuations in the stationary equipment measuring the biogas production and composition, a yield that exceeds 100 percent (and a negative yield) is sometimes obtained (figure 7).

-150,0 -100,0 -50,0 0,0 50,0 100,0 150,0 200,0 250,0

40 50 60 70 80 90 100 110 120 130 140 150

Yield [%, COD-methane/COD-glycerol]

Day

Yield (glycerol to methane)

(24)

18

pH

As the OLR increased, pH decreased for D5. After one HRT at maximum OLR, pH was measured to be 6,99 ± 0,04 for D5 and 7,24 ± 0,05 for D6. This corresponds to a difference of 0,25 pH-units. The development of pH can be seen in figure 8.

Figure 8. Development of pH in digestate from digester 5 (D5) and digester 6 (D6) is shown together with the glycerol loading over time.

0,00 0,50 1,00 1,50 2,00 2,50

6,70 6,80 6,90 7,00 7,10 7,20 7,30 7,40

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 131 136 141 146 151 156 Glycerol feed [kg VS / m3, day]

pH-value

Day

pH-values

D5 D6 Glycerol feed

(25)

19

VFA

Throughout the entire experiment, no change in VFA was observed. The concentration of VFA was 143 ± 2 mg/L for D5 and 139 ± 10 mg/L for D6. The development of VFA over time can be seen in figure 9.

Figure 9. Development of concentration of VFA in digestate from digester 5 (D5) and digester 6 (D6) is shown together with the glycerol loading over time.

0,00 0,50 1,00 1,50 2,00 2,50

50 70 90 110 130 150 170 190

[VFA] (mg/L)

Day

VFA-concentration

D5 D6 Glycerol feed

(26)

20

Ammonium

After one HRT at the maximum OLR, the concentration of ammonium was 476 ± 11 mg/L for D5 and 754 ± 5 mg/L for D6. This corresponds to a decrease in ammonium concentration as of 37 %. The development of the ammonium concentration throughout the experiment can be seen in figure 10.

Figure 10. Development of concentration of ammonium in digestate from digester 5 (D5) and digester 6 (D6) is shown together with the glycerol loading over time.

0,00 0,50 1,00 1,50 2,00 2,50

400 450 500 550 600 650 700 750 800 850 900

[NH4+] (mg/L)

Day

NH4+-concentrations

D5 D6 Glycerol feed

(27)

21

TS and VS

As expected, total solids content in D5 increased. The average TS was 2,46 ± 0,01 % for D5 and 2,01 ± 0,02 % for D6 at the end of the experiment. This corresponds to an increase in TS of 18

%. In figure 4 it can be seen that OLR increases at day 55, which could be related to the rise in TS in D5 as of day 80. The TS of D6 also starts to rise at day 89 but it is not clear why this happened, it can however be seen that the TS of D6 shows no liability to change during the pilot test in its entirety and thus the observed rise in TS in D6 at day 89 could be due to the variations caused by the screening of primary sludge in the WWTP. The development of TS can be seen in figure 11.

Figure 11. Development of TS in digestate from digester 5 (D5) and digester 6 (D6) is shown together with the glycerol loading over time.

No change was observed for VS. The average VS was 63 ± 0,7 % (mVS/mTS) for D5 and 62 ± 0,3

% (mVS/mTS) for D6. The development of VS can be seen in figure 12.

0,00 0,50 1,00 1,50 2,00 2,50

1,50 1,70 1,90 2,10 2,30 2,50 2,70

TS (%)

Day

Total solids

D5 D6 Glycerol feed

(28)

22

Figure 12. Development of VS in digestate from digester 5 (D5) and digester 6 (D6) is shown together with the glycerol loading over time.

0,00 0,50 1,00 1,50 2,00 2,50

56,0 57,0 58,0 59,0 60,0 61,0 62,0 63,0 64,0 65,0

VS (%)

Day

Volatile solids

D5 D6 Glycerol feed

(29)

23

Alkalinity

After one HRT of maximum OLR, the alkalinity was 3480 mg/L in D5 and 3850 mg/L in D6.

Thus no change in alkalinity was observed. Since only one alkalinity measurement was

performed after one HRT of maximum OLR, no standard deviation could be calculated. With an uncertainty of measurement of 10%, the observed difference in alkalinity falls short of this uncertainty but there can be seen a trend in figure 13 where the alkalinity of D5 is decreasing.

The development of the alkalinity throughout the experiment can be seen in figure 13.

Figure 13. Development of alkalinity in digestate from digester 5 (D5) and digester 6 (D6) is shown together with the glycerol loading over time.

0,00 0,50 1,00 1,50 2,00 2,50

0 1000 2000 3000 4000 5000 6000

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108 114 120 126 132 138 144 150 156 162 168 Glycerol feed [kg VS / m3, day]

Alkaliniy [mg CaCO3/l]

Day

Alkalinity

D5 D6 Glycerol feed

(30)

24

Heavy metals

The measured heavy metals concentrations during the pilot test can be seen in figure 14 and 15.

Figure 14. Development of concentration of copper and zinc in digestate from digester 5 (D5) and digester 6 (D6) is shown together marks indicating the initiation of the glycerol fed and the maximum OLR of D5.

0 100 200 300 400 500 600 700 800 900 1000

0 20 40 60 80 100 120 140 160

Heavy metal concentration [mg metal / g TS]

Day

Heavy metals (Cu and Zn)

Cu D5 Zn D5 Cu D6 Zn D6 Glycerol feed starts Maximum OLR begins

(31)

25

Figure 15. Development of concentration of mercury, nickel, lead, and silver in digestate from digester 5 (D5) and digester 6 (D6) is shown together marks indicating the initiation of the glycerol feed and the maximum OLR of D5.

After one HRT of maximum OLR the concentration of mercury had increased noticeably and the concentration of nickel and chromium had decreased noticeably in D5 as compared to the

reference period prior to the addition of glycerol to D5. However, the concentration of zinc, lead and silver in the reference digester D6 had increased noticeably during the HRT of maximum OLR in D5. As only one analysis was performed after maximum OLR, no standard deviation can be presented. The concentration of copper before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 16, these values were 410 and 350 mg/kg TS for D5 and 390 and 430 mg/kg TS for D6.

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7

0 5 10 15 20 25 30

0 20 40 60 80 100 120 140 160

Heavy metal concentration [mg metal / g TS] for Hg

Heavy metal concentration [mg metal / g TS] for Ni, Pb, Cr and Ag

Day

Heavy metals (Hg, Ni, Pb, Cr and Ag)

Ni D5 Ni D6 Pb D5 Pb D6

Cr D5 Cr D6 Ag D5 Ag D6

Hg D5 Hg D6 Maximum OLR begins Glycerol feed starts

(32)

26

Figure 16. Concentration of copper in digestate from digester 5 (D5) and digester 6 (D6) before glycerol feed and after maximum OLR.

The concentration of zinc before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 17, these values were 556 and 520 mg/kg TS for D5 and 550 and 660 mg/kg TS for D6.

D5

D6 D5 D6

0 50 100 150 200 250 300 350 400 450 500

Prior to glycrol feed After one HRT of maximum OLR Hea vy m et al conc en tr ation [m g m et al / g TS]

Copper (Cu) concentration in digestate from D5 and D6

D5 D6 D5

D6

0 100 200 300 400 500 600 700

Prior to glycrol feed After one HRT of maximum OLR Hea vy m et al conc en tr ation [m g m et al / g TS]

Zinc (Zn) concentration in digestate from D5 and D6

(33)

27

Figure 17. Concentration of zinc in digestate from digester 5 (D5) and digester 6 (D6) before glycerol feed and after maximum OLR.

The concentration of mercury before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 18, these values were 0,44 and 0,63 mg/kg TS for D5 and 0,43 and 0,45 mg/kg TS for D6.

Figure 18. Concentration of mercury in digestate from digester 5 (D5) and digester 6 (D6)

before glycerol feed and after maximum OLR.

The concentration of nickel before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 19, these values were 22 and 18 mg/kg TS for D5 and 21 mg/kg TS for D6 at both occations.

D5

D6 D6

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7

Prior to glycrol feed After one HRT of maximum OLR Hea vy m et al conc en tr ation [m g m et al / g TS]

Mercury (Hg) concentration in digestate from D5 and D6

(34)

28

Figure 19. Concentration of nickel in digestate from digester 5 (D5) and digester 6 (D6) before glycerol feed and after maximum OLR.

The concentration of lead before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 20, these values were 21 and 18 mg/kg TS and 19 and 24 mg/kg TS for D6.

D5

D6 D6

0 5 10 15 20 25

Prior to glycrol feed After one HRT of maximum OLR Hea vy m et al conc en tr ation [m g m et al / g TS]

Nickel (Ni) concentration in digestate from D5 and D6

D5

D6 D5

D6

0 5 10 15 20 25 30

Prior to glycrol feed After one HRT of maximum OLR Hea vy m et al conc en tr ation [m g m et al / g TS]

Lead (Pb) concentration in digestate from D5 and D6

(35)

29

Figure 20. Concentration of lead in digestate from digester 5 (D5) and digester 6 (D6) before glycerol feed and after maximum OLR.

The concentration of chromium before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 21, these values were 22 and 17 mg/kg TS for D5 and 0,43 and 21 mg/kg TS for D6 at both occasions.

Figure 21. Concentration of chromium in digestate from digester 5 (D5) and digester 6 (D6) before glycerol feed and after maximum OLR.

The concentration of silver before introduction of glycerol and after one HRT of maximum OLR can be seen in figure 22, these values were 3,1 and 2,8 mg/kg TS for D5 and 3,5 and 5,1 mg/kg TS for D6.

D5

D6 D6

0 5 10 15 20 25

Prior to glycrol feed After one HRT of maximum OLR Hea vy me tal conc en tr ation [mg me tal / g TS]

Chromium (Cr) concentration in digestate from D5 and

D6

(36)

30

Figure 22. Concentration of silver in digestate from digester 5 (D5) and digester 6 (D6) before glycerol feed and after maximum OLR.

One of the biggest determinants as to whether the thickened digestate can be used as fertilizer for dispersion on fields is the ratio of heavy metals compared to phosphorus content of the sludge.

The limits of acceptable values, and the values calculated for the digestate before addition of crude glycerol and after one HRT of maximum OLR can be seen in table 3 (Naturvårdsverket, 2013). Data for the content of phosphorus was supplied with the same analysis sheet as the heavy metals. All the calculated values were lower than the acceptable limits.

Table 3. Calculated ratios of heavy metals and phosphorus in digestate after one HRT of maximum OLR for dispersion on fields and corresponding maximum tolerable ratios according to Naturvårdsverket.

Ratio heavy metal/phosphorus [mg metal / kg P]

D5 D6 Acceptable

limit Metal Before glycerol feed After glycerol feed Before glycerol feed After glycerol feed

Cu 12813 11290 12188 11316 21400

Zn 17500 16774 17188 17368 28600

Hg 14 20 13 12 40

Ni 688 581 656 553 1400

Pb 656 581 594 632 1600

Cr 688 548 656 553 2100

Ag 150 132 159 129 180

D5 D5

D6

0 1 2 3 4 5 6

Prior to glycrol feed After one HRT of maximum OLR Hea vy m et al conc en tr ation [m g m et al / g TS]

Silver (Ag) concentration in digestate from D5 and D6

(37)

31

Volatile solids reduction

Reduction of volatile solids increased with the addition of crude glycerol. The two equations used to calculate the VSR gave different results for the VSR of both digesters. The VSR during the beginning, middle and final stage of the pilot test can be seen in table 4. The development of volatile solids reduction can be seen in figure 23.

Table 4. Volatile solids reduction for D5 and D6 using two different equations, the ash concentration equation (ACE) and the mass balance equation (MBE).

VSR D5, ACE [%]

VSR D6, ACE [%]

VSR increase, ACE [%]

VSR D5, MBE [%]

VSR D6, MBE [%]

VSR increase, MBE [%]

Reference period (day 21 - 41)

57 57 -0.61 56 54 3.2

Middle period (OLR

= 0,5, day 77 - 97)

58 49 17 64 50 28

One HRT of maximum OLR (day 140-151)

62 51 21 68 49 38

(38)

32

Figure 23. Volatile solids reduction calculated with the mass balance equation (MBE) and the ash content equation (ACE).

Calculations for VSR can be seen in appendix 1C.

0,00 0,50 1,00 1,50 2,00 2,50

0 10 20 30 40 50 60 70 80 90 100

7 12 17 22 27 32 37 42 47 52 57 62 67 72 77 82 87 92 97 102 107 112 117 122 127 132 137 142 147 Glycerol feed [kg VS/m3, day]

VSR [%]

Day

Volatile solids reduction

D5, MBE D6, MBE D5, ACE D6, ACE Glycerol feed

(39)

33

Discussion

The goal of this master thesis was to see if it is possible to increase methane production twofold through addition of crude glycerol to an anaerobic digestion process treating sewage sludge. This goal was not reached as the increase in methane production was 74%. The hypothesized yield (as a conversion from COD in the form of crude glycerol to COD in the form of methane) of 85 % was not reached as the actual yield was 78%, thus the hypothesis can be discarded. One aspect that needs to be kept in mind is the potential impairment of the digester when it is scaled up. As the mixing of a 7000 m³ digester is not ideal, the concentration of different substrates will fluctuate. Hewitt & Nienow (2007) reports that microbial cells can develop stress responses because of this phenomena, but the response varies depending on the microbes, substrate and digester conditions. Ruffino et al. (2015) mentions that a common yield for a pilot scale digester is 80 % of the corresponding yield of a lab-scale digester. Another important factor to account for is the position of the feed of the crude glycerol. Due to the rapid digestion of glycerol when introduced to D5 under the impellers of the CSTR, biogas production could occur at an early stage of the feed. If this happens directly under the impellers where the feed opening is placed, the resulting gas formation can obstruct the mixing activity of the impellers (Sardeing, R., et al., 2004).

Also, batchwise feed of glycerol (such as the ones performed in SBFs laboratory experiments) allows a higher rate of digestion than continuous feed of glycerol as the latter allows some of the glycerol to escape from the digester because of the volumetric control of D5 and D6 through overflowing.

Due to the crude glycerols content of sulphate, electrons that could have been used for

production of methane were instead incorporated in the synthesis of H2S, which also affects the yield of methane. It is also a concern for the downstream processing of the biogas as H2S is corrosive (Colleran et al., 1995). If a complete reduction of the sulphate in the crude glycerol would occur, this would require 4,7 % of the electrons supplied by the glycerol (see appendix 2B for calculations). This alternative use of electrons can help explain the lower yield of methane.

Sulphide can be toxic to the microbial culture of anaerobic digesters (Parkin et al.,1990) but as the digester was stable with the addition of crude glycerol it can be concluded that the

concentration of sulphide was lower than toxic level and did not affect the biology of the process negatively.

As regards the hypothesis stated in the beginning of the master thesis one can conclude that it is false not only due to the yield being lower than hypothesized but also because the TS increased by more than 10%. A notable decrease in pH was observed but could not be accounted for by the VFA. By means of pH, the concentration of protons increased with 45 nM whereas VFA

contributed with merely 64 uM of protons. The phenomena could be explained in other ways however. The assimilation of ammonium leads to the release of protons as the ammonium is utilized mainly for protein synthesis. If one assumes that every ammonium molecule assimilated produces two protons during protein synthesis, the drop in free ammonium levels in D5 due to the introduction of crude glycerol would produce 30,8 mM of protons (see appendix 1B). This contributes substantially to any acidifying effects in the digester. Furthermore, the drop in carbonate alkalinity equals a decrease in carbonate ions as of 6,1 mM. It is likely that the main

Industrial production of biogas through co-digestion of waste glycerol and sewage sludge