Small-Scale Biogas Upgrading System Modeling Tool Development

(1)

and Management

Small-Scale Biogas Upgrading System Modeling Tool Development

David Saldarriaga

Master of Science Thesis

KTH School of Industrial Engineering and Management Department of Energy Technology

EGI-2013-MJ232X Unit of Heat and Power

SE-100 44 Stockholm 2018

(2)

(3)

Master of Science Thesis TRITA-ITM-EX 2018:625

Small-Scale Biogas Upgrading System Modeling Tool Development

David Saldarriaga Approved

2018-September-26

Examiner

Anders Malmquist

Supervisor

Anders Malmquist

Commissioner Contact person

Abstract

The potential of biogas to decarbonize society depends partially on the success of small-scale systems.

Two specific locations where biogas units can be implemented are considered in this study: small farms and small isolated populations. The access to energy sources, either traditional or renewable for these is often restricted and typically costly. Clean raw biogas can be the energy source to satisfy power generation, cooking, and heating needs. Upgrading widens the options to transport fuel and energy storage helping positively in the unlinking between production and demand.

Water upgrading has three essential and exclusive benefits that make it a highly feasible solution for these isolated locations or small agricultural units. It is in general available in these places; it has a very low environmental impact if leakage or malfunctioning of the system, and has no toxicity per se.

To aid the development of the biogas industry focused on small-scale systems a fast, easy to use, low cost, customizable tool is needed to help the design process of the high-pressure water upgrading units. The present study covers the development of such a tool. In the present report, the basis of the model to solve the mass balance of the system and to calculate the dimensions of the scrubber are described. The scrubber model is an implementation of the NTU-HTU model proposed by Billet and Schultes in two major publications (Billet, 1995) and (Billet & Schultes, 1999). The strategies used to solve the set of closed loop equations, and iterations are presented in a block diagram fashion. The tool was developed in visual basic for applications using Excel as the hosting application.

The results of the tool are compared against those obtained from the same model ran in Aspen Plus.

To perform such a comparison, 540 cases were used. The cases are the result of running three nominal raw biogas flows, using three different packing materials, varying the raw biogas and water flows, varying pressure, temperature, and height of the scrubber, and varying the pressure of the flash tank.

Three sensitivity analyses are performed to check the influence of some variables in the model. One

is designed to check the influence of the exponent choice for dimensionless numbers in the calculation

of the volumetric mass transfer coefficient as an example of the various points where the Billet model

(4)

is adjusted to follow the behavior of packed columns. Another analysis consists of comparing the results of height using different packing materials to see how the six packing constants affect the results of the calculations. The third analysis is performed to check the influence of the methane absorption model where three different approaches were used.

The results show that the tool behaves coherently, and a validation step can be implemented via real experimental tests comparison. The tool has several points where adjustments can be made, like the mentioned exponents for the dimensionless numbers, or the constants used in the interfacial area calculation and the correction of the 𝐶

_𝑆

and 𝐶

_𝐹𝑙

packing constants when inversion point has been reached.

The implementation of high-pressure water scrubbing together with a flash tank can achieve slip values as low as 0.25% or even lower. The lower the slip, the higher the energy needed to upgrade is.

Thus, there is a trade-off between slip and internal energy demand.

(5)

Sammanfattning

Potentialen hos biogas för att minska samhällets kolberoende är delvis beroende avframgången för småskaliga system. I denna studie betraktas två specifika platser där biogasenheter kan användas:

små lantbruk och små isolerade byar. Tillgången till energikällor, såväl traditionella som förnybara, är ofta begränsad och är vanligtvis dyr för dessa platser. Renad biogas kan bli enenergikällan för att tillgodose kraftproduktion, matlagning och uppvärmning. Biogasuppgradering breddar användningsalternativen till användning som drivmedel och till energilagring, vilket bidrar på ett positivt sätt till koppla isär energiproduktion och energiförbrukning.

En vattenskrubber har tre viktiga och unika fördelar som gör den en väldig användbar lösning för dessa isolerade platser eller för små jordbruksenheter. Vanligtvis är vatten tillgängligt på dessa platser, det har en synnerligen låg miljöpåverkan om det blir läckor eller funktionsfel i systemet och det är icke-giftigt.

För att stödja utvecklingen av biogasindustrin med inriktning på småskaliga system behövs ett snabbt, lättanvänt, billigt och anpassningsbart verktyg för att genomföra design en högtrycks vattenskrubber.

Den här studien omfattar utvecklingen av ett sådant verktyg. I föreliggande rapport beskrivs grunden för modellen för att lösa systemets massbalans och att beräkna och dimensionera skrubbern.

Skrubber-modellen innehåller en implementering av NTU-HTU-modellen som Billet och Schultes föreslagit i två stora publikationer (Billet, 1995) och (Billet & Schultes, 1999). De strategier som används för att lösa uppsättningen av slutna loop-ekvationer och iterationer presenteras i ett blockdiagram. Verktyget har utvecklats i VBA (Visual Basic for Applications) och använder Excel som programvarumiljö.

Resultaten som erhålls med det utvecklade verktyget jämförs med de som erhålls från samma modell i Aspen Plus. För att utföra en sådan jämförelse användes 540 testfall. Fallen är resultatet av modelleringen av tre nominella biogasflöden, där tre olika typer av fyllmaterial/fyllkroppar används.

För varje testfall varieras rå biogas, vattenflöde, tryck, temperatur och höjd av vattenskrubbern, samt trycket i flash-tanken.

Tre känslighetsanalyser utförs för att kontrollera påverkan av vissa variabler i modellen. Den första är utförmad för att kontrollera exponeringsvalets inverkan på dimensionslösa tal vid beräkningen av den volymetriska massöverföringskoefficienten, som ett exempel av flera punkter där Billets modell är justerad för att följa vattenskrubbers beteendet. Den andra analysen består av att jämfora skrubberns beräknade höjd när olika fyllmaterial används för att evaluera hur påverkar de sex fyllmaterial konstanterna resultaten av beräkningar. Och den tredje analysen utförs för att kontrollera påverkan av metans absorptionsmodell när tre olika metoder användes.

Resultaten visar att verktyget uppträds på ett koherent sätt och ett valideringssteg kan baserat på detta genomföras med verkliga experimentella jämförelser. Verktyget har flera delar där justeringar kan göras, som t.ex. exponenterna för dimensionslösa tal och konstanterna som används i interface- beräkningarna, samt korrigeringen av 𝐶

_𝑆

och 𝐶

_𝐹𝑙

konstanter när punkten av inversion har uppnåtts.

Implementeringen av högtrycks vattenskrubbers tillsammans med en flash-tank kan medföra en slipp

(förklaras senare i rapporten) så låg som 0,25% eller ännu lägre. Ju lägre slipp är desto högre är

energin som behövs för att uppgradera biogasen. Det finns således en avvägning mellan slipp och

intern energiförbrukning.

(6)

(7)

Abbreviations and Symbols

Abbreviations

AD Anaerobic digester, anaerobic digestion CBG Compressed biogas used as vehicle fuel CHP Combined Heat and Power

CH4 Methane

CM Combined CH4 absorption model

CNG Compressed natural gas used as vehicle fuel CO2 Carbon dioxide

DEA Diethanolamine

EQ Equilibrium CH4 absorption model FE Flash equilibrium CH4 absorption model GHG Greenhouse gases

HTU-NTU Height of the transfer unit – Number of transfer units H2S Hydrogen sulfide

LBG Liquefied biogas LHV Lower Heating Value LNG Liquefied natural gas

LPSA Low-pressure swing adsorption

MEA Monoethanolamine

NH3 Ammonia

Nm³ Normal cubic meters. Normal refers to the volume of gas measured at 1.01325 bar and 273.15 K. (1 atm and 0°C)

N2 Nitrogen

O2 Oxygen

PSA Pressure swing adsorption. Also: Pressure swing absorption

RB Raw Biogas

UK United Kingdom

VBA Visual Basic for Applications (Microsoft) VOC Volatile organic compound

Variable Names

𝐴 Cross section area of the absorption column. Also: variable used in the Peng-Robinson equation of state

𝑎 Total packed surface area per packed volume. Also: variable used in the Peng-Robinson equation of state. Also: variable used in the flash tank solving equations

𝐵 Variable used in the Peng-Robinson equation of state

(11)

𝑏 Variable used in the Peng-Robinson equation of state. Also: variable used in the flash tank solving equations

𝑎_ℎ Hydraulic area

𝑐 Molarity concentration. Also: variable used in the flash tank solving equations 𝐶𝐹 Inter-stage compression factor

𝐶_𝐹𝑙 Packing constant used to calculate the flow resistance 𝜉𝐹𝑙

𝐶_𝐿 Packing constant used to calculate the volumetric mass transfer coefficient 𝑐_𝑝 Specific heat at constant pressure

𝑐_𝑣 Specific heat at constant volume

𝐶_𝑆 Packing constant used to calculate the flow resistance 𝜉_𝑆 𝐷_𝑥 Diffusivity coefficient for phase 𝑥.

𝑑 Diameter of the absorption column

𝑑_𝑥 Diameter of the absorption column for the loading or flooding point 𝑑ℎ Hydraulic diameter

𝑓𝑦 CH4 absorption flow factor used for the internal flow of the scrubber

𝑓_𝑘 Wall factor

𝐹𝑟 Froude number

𝐻 Total height of the absorption column

𝑖 or 𝑗 In an iteration an “i” case. Also: the “i” component in a summation ℎ𝑙 Liquid holdup of the absorption column

𝐾 Temperature-dependent Henry’s constant

𝑘_𝑖𝑗 Binary interaction parameter used in the Peng-Robinson equation of state 𝐿 Liquid. Also: Aqueous flow

𝑀 Molecular weight

𝑁 Molar flow correspondent to mass transfer between phases

𝑛 Number of compression stages. Also: Total components of a mixture

𝑚 Mass

𝑚_𝑦𝑥 Slope of the equilibrium curve

𝑃 Power

𝑝 Pressure

𝑅 Universal gas constant

𝑅𝐵 Raw biogas flow

𝑅𝑒 Reynolds number

𝑇 Temperature

𝑆𝑐 Schmidt number

𝑠𝑓 Slip as a fraction 𝑠𝑙𝑖𝑝 Slip as molar flow

(12)

𝑢 Velocity or load 𝑢𝐿

̅̅̅ Local liquid velocity 𝑢_𝑉

̅̅̅̅ Average effective vapor velocity

𝑉 Vapor flows

𝑣̅ Specific volume

𝑋 Aqueous. Also: Liquid molar load 𝑥 Vapor mole fraction

𝑌 Vapor molar load

𝑦 Vapor mole fraction Sub-indices

𝑎 Species “a”. Also: Absorption column

𝑎𝑖𝑟 Air

𝑎𝑑 Additional temperature

𝑎𝑡𝑚 Atmospheric. Also: Ambient conditions

𝑏 Bulk

𝑏𝑔 Biogas

𝐵 Base. Also: Blower

𝑐 Critical value

𝐶 Compressor

𝐶𝑂 Cooler

𝑑𝑒𝑣 Deviation

𝑒 Equilibrium

𝑓 Flash tank

𝐹𝑙 Flooding point

𝑖 Interphase

𝑖 or 𝑗 In an iteration an “i” case. Also: the “i” component in a summation 𝐿 Liquid phase or related to the wet surface of the packing

𝑘 Related to the wall factor 𝑓_𝑘

𝑚 Mixture

𝑚𝑎𝑠 Mass

𝑚𝑒𝑐 Mechanical

𝑚𝑜𝑙 Mol. Also: Molar

𝑁𝑣𝑜𝑙 Normalized volume. Normal refers to the volume of gas measured at 1.01325 bar and 273.15 K.

(1 atm and 0°C)

𝑃 Pump

𝑜 Top of the column

𝑜𝑝 Operative. Also: Operational

(13)

𝑜𝑇 Overtemperature

𝑂𝑉 Overall

𝑜𝑣 Overpressure. Also: Overtemperature

𝑟 Reduced value

𝑆 Loading point. Also: Stripper 𝑆₀ Thickness of the liquid film

𝑇 Temperature-dependent

𝑢 Bottom of the column

𝑉 Vapor phase

𝑣̅ Specific volume used in the Peng-Robinson equation of state

𝑣𝑜𝑙 Volume

𝑥 Loading point or Flooding point. Also: Liquid or Vapor. Also: “x” Compressor stage 2 Refers to CO2 as species

4 Refers to CH4 as species Greek letters

𝛼 Variable used in the calculation of the Peng-Robinson equation of state 𝛽 Mass-transfer coefficient

𝛽 ∙ 𝑎_𝑝ℎ Volumetric mass transfer coefficient

Δ Delta or change

𝜀 Void fraction

𝜂 Dynamic viscosity. Also: Efficiency 𝜉 Resistance factor

𝜌 Density

𝜎 Surface tension

𝜏 Time

𝜑 Variable used in the Peng-Robinson equation of state

𝜔 Acentric value

(14)

List of Figures

Figure 1 Number of industrial AD plants commissioned each year in the world from 1981 to 2016. In total, there is one plant that lacks an upgrading system and another one that performs only

partial upgrading. (IEA Bioenergy, 2017) ... 18

Figure 2 Total Number of upgrading plants per country until 2016 (IEA Bioenergy, 2017) ... 19

Figure 3 Added capacity year by year from 1981 to 2016. The red curve depicts the total cumulative capacity in Nm³/h. ... 20

Figure 4 Plant type and size in Nm³/h distribution until 2016 (IEA Bioenergy, 2017). ... 21

Figure 5 Left: Total capacity technology distribution mix in 2016. Right: Sizes 0 – 100 Nm³/h range technology capacity mix (IEA Bioenergy, 2017). ... 21

Figure 6 Biogas paths. The amount of biogas used in each path is dependent on several factors. The flow into upgrading is, thus, variable. ... 23

Figure 7 Left: Bubble plate tower. Right: Picture of a tray with bubble caps for a distillation column. Figure adapted and picture obtained from (Sölken, 2018). ... 28

Figure 8 Water scrubber process. Adapted from (Wellinger, et al., 2013) and (Bauer, et al., 2013). ... 29

Figure 9 Different industrial packing materials. 1 Cylindrical ring. 2 Berl-Saddle. 3 Novalox® Saddle. 4 Pall® ring (metallic). 5 Pall® ring (plastic). 6 Interpack®. 7 VSP®. 8 Igel®. 9 Novalox® - M. 10 VFF.Power-Pack®. Images obtained from (Vereinigte Füllkörper-Fabriken GmbH & Co. KG, 2016) ... 30

Figure 10 Structured packing. Images obtained from (Raschig Gmbh, 2018). ... 30

Figure 11 PSA plant configuration. Four columns continuous process. Image adapted from (Wellinger, et al., 2013). ... 31

Figure 12 Different membrane upgrading configurations. Image adapted from (Hoyer, et al., 2016). ... 32

Figure 13 Gas - Liquid membrane technique. ... 33

Figure 14 Biosling process scheme. Figure obtained from (Biosling AB, 2012) ... 34

Figure 15 AD process description (Sofer & Zaborsky, 1981). ... 35

Figure 16 Scrubbing process. Left: Packed column. Center: Equilibrium and operating curves given in molar load units. Right: Equilibrium and operating curves given in molar fraction units. Figures adapted from (Billet, 1995). ... 41

Figure 17 Validation of constants fitted to laboratory data. In light blue the results of this study in black original data. The original plot was extracted from (Carroll, et al., 1991). ... 44

Figure 18 Left: Liquid holdup and phase distribution for a low liquid load. Center: High liquid load. Right: Schematic representation of packing and phase interaction. The right figure was adapted from (Billet, 1995) ... 47

Figure 19 Loading and flooding points. Figure adapted from (Billet, 1995) ... 48

Figure 20 Section of the scrubbing column. The flow of CH4 𝑽 and H2O 𝑳 constant. ... 53

(15)

Figure 21 Penetration model - Higbie Model. The time t of residence of the liquid packet is the time needed to flow through a bubble of diameter d. The concentration n the liquid is cL and in

the vapor phase cV. ... 54

Figure 22 Example of the height calculation while varying the composition of the upgraded biogas. Notice the non-linearity of the function. ... 60

Figure 23 High-pressure water upgrading cycle with a flash tank and stripper. ... 65

Figure 24. Simplified process omitting the stripper. ... 66

Figure 25 Qualitative description of the CH4 mass transfer behavior as a function of the height of the column. Source Author of the present study. ... 68

Figure 26 Left: Equilibrium concentration. Right: Flash equilibrium CH4 concentration. ... 68

Figure 27 High-pressure water upgrading cycle with auxiliary equipment highlighted... 72

Figure 28 User Interface of the modeling tool developed. The interface has additional fields not shown in the figure. The additional fields are where the packing material is selected, and the results of the scrubber are displayed like the velocities of the vapor and liquid phases, the slope of the equilibrium curve, the liquid holdup, the loading and flooding point loads among others. ... 77

Figure 29 Modeling tool general block flow ... 78

Figure 30 Loop block calculation strategy. ... 78

Figure 31 Calculation strategy to solve for the height of the scrubber... 80

Figure 32 Slicing of the column. Strategy for applying HTU-NTU method. Qualitative description of the density of the vapor phase behavior in the column as a function of the height. ... 81

Figure 33 Block layout of the Aspen model. (Aspen Technology, Inc, 2017) ... 84

Figure 34 General comparison results of the slip calculations. Left: Whole range. Right: Detail between 3% and 9% of slip ... 86

Figure 35 Slip deviation graphical representation. Left: All the data plotted. Right: Zoom to data up to 1% slip. ... 86

Figure 36 Histograms for Slip deviation comparison. For the scrubber model, 7 points were omitted in the histogram since they are outliers that distort the histogram shape in the central focal area. ... 87

Figure 37 Concentration calculation deviation graphical representation. All the data plotted. ... 88

Figure 38 Histograms for concentration calculation deviations comparison... 89

Figure 39 Results comparison for the calculated height of the column. ... 89

Figure 40 Histograms for the height deviations. ... 90

Figure 41 Average and Standard deviations for the calculations considering only the scrubber -No flash tank- ... 92

Figure 42 Average and Standard deviations for the calculations considering scrubber and flash tank. ... 92

Figure 43 Sensitivity analysis. Calculated slip solving for yo4 on the Scrubber model. FE, CM and EQ absorption models. ... 93

(16)

Figure 44 Sensitivity analysis. Calculated slip solving for yo4 on the Cycle model (Scrubber + Flash). FE, CM and EQ absorption models. Left: All the data plotted. Right: Slip Detail from 0.025 to 0.5%. ... 94 Figure 45 Height comparison between the results obtained using tool with the original Billet exponents

of 3/4 and 1/2 and those obtained with the tool using 1/2 exponent. ... 95 Figure 46. Height comparison between the results obtained using tool with Pall ring and those obtained

with the tool using Ralu Flow, Ralu Ring, Hiflow ring, and NOR PAC ring, packing materials. ... 96 Figure 47 Packing materials compared. Left to Right: Pall Ring, Hiflow ring, NOR PAC, Ralu Flow, Ralu

ring, Hackette. Source (Raschig Gmbh, 2018) except for Hiflow Ring which was obtained from (RVT Process Equipment GmbH, s.f.) ... 96 Figure 48 Valtra Diesel - Biogas tractor. (Valtra, 2018) ... 104

(17)

List of Tables

Table 1 Typical raw biogas characterization (Wellinger, et al., 2013) ... 20

Table 2 Comparison based on operating parameters and quality performance between upgrading techniques. Sources (blue small font in the table): 1 (Kadam & Panwar, 2017), 2 (Hoyer, et al., 2016), 3 (Bauer, et al., 2013), 4 (Ong, et al., 2014), 5 (Allegue & Hinge, 2011), 6 (Biosling AB, 2012). ... 36

Table 3 Comparison based on contaminant tolerance between upgrading techniques. Sources (blue small font in the table): 1 (Kadam & Panwar, 2017), 2 (Hoyer, et al., 2016), 3 (Bauer, et al., 2013), 4 (Ong, et al., 2014), 5 (Allegue & Hinge, 2011), 6 (Biosling AB, 2012). ... 37

Table 4 Constants reported by (Sander, 2015) used in Equations 17 and 18. ... 44

Table 5 CO2 Henry's constants for Equations 17 and 18. Source author of the present study. ... 45

Table 6 Values of efficiencies, overpressure, additional temperature and others used in the power calculations. Source ... 75

Table 7 Structure of the comparison runs. The values of flow and "from" "to" are approximations of the values used. RB: Raw biogas flow, Lo Liquid flow, pf: pressure of the flash tank, ps: pressure of the scrubber, Ts: temperature of the scrubber, Hs: height of the scrubber. ... 82

Table 8 Slip comparison results ... 87

Table 9 Compositions concentrations results ... 88

Table 10 Height calculation deviations results ... 90

Table 11 Sensitivity analysis results summary. The method the tool uses is CM —blue text—. ... 92

(18)

1 Proposal

1.1 Background

Replacement of fossil fuels as an energy source is imperative to mitigate the climate change impact. Biogas obtained from Anaerobic Digestion (AD) of organic matter is a promising option. The use of AD dates back to ancient times. Countries like China and India have thousands of artisanal digesters. However, the raw biogas has low heat value due to the presence of a high carbon dioxide (CO2) concentration. Therefore, upgrading is often necessary depending on the final use of the biogas. For example, if it is injected into the natural gas grid or used as Compressed Biogas Gas for vehicles (CBG), the methane content must be around 97%.

The industrial use of upgraded biogas did not experience a significant development until 2006 - 2007 mainly in Germany, Sweden, and the United Kingdom as it can be seen in Figure 1 and Figure 2 (IEA Bioenergy, 2017).

Figure 1 Number of industrial AD plants commissioned each year in the world from 1981 to 2016. In total, there is one plant that lacks an upgrading system and another one that performs only partial upgrading. (IEA Bioenergy, 2017) Following (Wellinger, et al., 2013) or (Sibiya, et al., 2017), the sources for biogas production can be classified as:

• Agricultural Sources that can be further classified as wastes and energy crops.

o Animal manure and slurries o Plant residues

o Energy crops

• Industrial waste and byproducts

• Municipal organic waste

o Sorted solid organic waste o Sewage

• Aquatic Biomass

According to (Wellinger, et al., 2013) the most important feedstock for the AD is the agricultural sources.

Nevertheless, according to (Sibiya, et al., 2017), the use of these biomass sources has reported process

(19)

instabilities and low efficiencies. The author also points out that co-digestion has been recommended as a remedy to these issues.

There are two pathways to produce biogas, biochemical and thermochemical. The biochemical pathway consists of producing biogas using microorganisms to process the biomass. The most widely used nowadays is the AD. The thermochemical path consists of producing biogas using physical mechanisms to process the biomass. Generally, two main techniques are used, pyrolysis and gasifier. While pyrolysis produces a mixture of solid, liquid and gases, gasification processes will mainly produce only gases. The final products of these thermochemical processes depend on the feedstock, process design, and ultimate purposes of the products (Luque & Lin, 2016).

As a fuel, biogas has the potential to serve as a dispatchable energy source. This capability helps to balance energy systems where intermittency of the source is high, like in Solar and Wind power. It can play an essential role as ancillary services for power networks. For transport fuel, it holds a high potential with proven technology – Compressed Natural Gas (CNG) – that has been successfully implemented in the past two decades (Budzianowski & Brodacka, 2017).

Since one of the most used technologies to produce Biogas is AD, the use in situ offers the possibility for agricultural dependent population to gain access to clean sustainable energy as a source for electricity, heating, transport fuel, and cooking. The impact may boost welfare as well as considerably reduce the environmental impact by reducing deforestation, replacement of fossil fuels, and reduction of greenhouse gases (GHG) (IRENA, 2016).

Depending on the final use of the biogas, it is convenient to clean and upgrade it. Biogas can be used directly – with minor cleaning – as an energy source for a Combined Heat and Power (CHP) unit. Nevertheless, the demands of electricity and heat are seldom constant and equal to the biogas production. Having additional uses or storage of upgraded biogas will improve the energy efficiency and flexibility of the system.

The typical characterization of biogas composition is presented in Table 1. Cleaning is the process of removing the impurities and pollutants present in the raw biogas. Some of the most common ones are hydrogen sulfide (H2S) and Ammonia (NH3). Upgrading is the process of increasing the CH4 concentration by reducing the CO2 present in the mixture (Luque & Lin, 2016).

Hydrogen sulfide possesses a serious corrosive potential to equipment, and it can also produce SOX gases when combusted. In order to eliminate these effects, the raw biogas must be cleaned. There are several common ways of treating this issue, but one of the most effective is trough water scrubbing.

Figure 2 Total Number of upgrading plants per country until 2016 (IEA Bioenergy, 2017)

(20)

The presence of CO2 lowers the energy content of the biogas mixture, and it consumes energy to transport, compress, heat or cool. Depending on the final use, it is more efficient to increase the CH4 concentration, so the lost energy associated with its handling is reduced. Further, if the biogas is intended to be used as a transportation fuel or injected to a natural gas network, it should comply with regulations that determine the concentration levels of CH4 as well as other components, like nitrogen, oxygen, and water.

Table 1 Typical raw biogas characterization (Wellinger, et al., 2013)

Compound Unit Range

Methane (CH4) mol-% 35 – 75

Carbon dioxide (CO2) mol-% 15 – 50

Nitrogen (N2) mol-% 0 – 5

Oxygen (O2) mol-% 0 – 1

Hydrogen sulfide (H2S) ppm 10 – 15.800

Ammonia (NH3) ppm 0 – 150

Total chlorine (Cl) mg/m³ 0 – 100

Total fluorine (F) mg/m³ 0 – 100

Siloxanes mg/m³ 0 – 0.2

Ethylbenzene ppm < 0.34

Copper µg/m³ < 20

Even though the AD is spread along the globe, the upgrading is still not implemented massively. Only industrial operations have incorporated the process in recently. As depicted in Figure 1 and Figure 2 the number of upgrading units have grown only since 2006, mainly in Germany, Sweden, and United Kingdom (IEA Bioenergy, 2017).

The upgraded-biogas added capacity and total cumulative is shown in Figure 3. The average size of an installed plant from 2007 to 2016 is 930 Nm³/h. Until 2016, the total worldwide installed capacity is only 527000 Nm³/h (IEA Bioenergy, 2017). This capacity production is only a small fraction of the natural gas consumption per hour for the same year of approximately 404 million Nm³/h (BP, 2017).

The plant size is depicted in Figure 4. The average size is 1 022 Nm³/h. The most common technology type is the water scrubber, followed by membrane, chemical scrubber, Pressure Swing Adsorption (PSA), organic physical scrubber, low-pressure swing adsorption (LPSA), amine scrubber, and cryogenic separation. The capacity mix for 2016 is presented in Figure 5. The water scrubbing techniques double its closest competitor.

Figure 3 Added capacity year by year from 1981 to 2016. The red curve depicts the total cumulative capacity in Nm³/h.

(21)

There is no consensus about the classification of the biogas plant regarding its size. For this study, a plant producing 100 Nm³/h or less is considered “small-scale.” This classification follows the one suggested by (VALORGAS, 2011). Nevertheless, other authors as (Wellinger, et al., 2013) consider small-scale systems only up to 100 kW (approximately 25 Nm³/h of raw biogas).

Figure 4 Plant type and size in Nm³/h distribution until 2016 (IEA Bioenergy, 2017).

Regarding the small-scale plants with upgrading, only 42 out of 548 are smaller than 100 Nm³/h. For these small systems, the most used technology is membrane, followed by PSA, water scrubber, other not identified, and organic scrubber. Concerning the number of units of the 19, eight are membrane, five are PSA, four are water scrubbers, one is an organic scrubber, and one is not identified.

Figure 5 Left: Total capacity technology distribution mix in 2016. Right: Sizes 0 – 100 Nm³/h range technology capacity mix (IEA Bioenergy, 2017).

According to (Lems & Dirkse, s.f.), for small-scale plants to be economically feasible, the best approach is to use the biogas locally or as compressed biogas as vehicle fuel (CBG). They found a limit between 20 and 25 Nm³/h to achieve a payback time of the upgrading plant of five years. Nevertheless, if the system has a CHP unit, the payback can be reduced to three to four years. It is important to notice that their study relies on membrane separation technique.

(22)

An innovative example of small-scale upgrading system is the one produced by Biosling. It consists of a water scrubber process that develops in a rotating coil filled with biogas and water. According to (Biosling AB, 2012) the biogas can be upgraded to 95% CH4 content using only the rotating coil. The biogas rated flow for the smallest unit is 0 to 36 Nm³/h. See further down for a more detailed explanation of the system.

There is an increasing focus on small-scale biogas operations, for example, the “Record Biomap” which is a research project that intends to “bridge the gap between research and industry and to accelerate innovation in small and medium scale biomethane production, thus shorten the time to market of the most promising technologies.” Some of the processes that are supported by the project are: Algal-bacterial upgrading (University of Valladolid), Blue BONSAI membrane technology (Apex), Ash filter (RISE), in-situ methane enrichment (RISE), vacuum pressure swing adsorption (NeoZeo), vacuum swing adsorption Biogame (AzzeroCO2 SRL), MR methane reactor (ED Biogas International AB), Pressureless cryogenic conversion of biogas into dry ice and liquid biomethane (Hochschule Landshut), G-PUR Membrane technology (CEBB CentraleSupélec), Biological methanation (Electrochea GmbH), and Biogas to bio-LNG system (Cryo Pur).

(Rogstrand, 2018).

Water, as the main consumable of water-upgrading systems, has an important characteristic that makes it a potential feasible technology to be implemented in small scale. In general, water is available where biomass as feedstock is located for AD applications, like farms, or agricultural locations. The supply of chemicals, or special consumables like membranes or spare parts for moving objects, increases considerably the cost of operation for remote locations, which is often the case of agricultural sites. Technical assistance costs also build up the farther the plant is located from central cities. The use of water as the absorber of CO2 reduces these expenses.

Another benefit of using water as the absorber is the low environmental impact. Handling dangerous materials carries the risk of leaks, accidents, and pollutions due to malfunctioning of the equipment. On the other hand, the final disposal of used materials increases the operational costs, especially if the system is an isolated location.

Using water as an absorber can be done in two ways. The simplest way is to pass it through the scrubbing column just once and dispose it of. The other way is to regenerate the water after the scrubbing process in a stripper column. It is evident that the first method demands a large fresh water supply, like a river or lake.

The impact of this method is that the CO2, H2S, and other substances will go into the wastewater, and if no treatment is performed then these compounds will pollute the water. The latter method releases the CO2

and H2S from the water in the stripper column. Nevertheless, after some cycles the water must be replaced because of contaminants build up.

As it can be seen, water scrubbing imposes an environmental impact, but it is case dependent, and it should be evaluated as a system. There are some alternatives to treat the wastewater of an upgrading plant, but the methods are dependent on the biogas composition, especially the pollutants present.

In general, all upgrading techniques use approximately the same amount of energy (Budzianowski, et al., 2017). It is important to notice that an upgrading plant needs to be flexible due to the fluctuations in the available raw biogas flow. This is because, whatever needs to be upgraded is dependent on the balance between the CHP unit, the AD production rate, the upgraded fuel demand, the cost of electricity and substitute fuel (e.g. petrol or diesel), among other variables that influence the final amount of biogas that goes through one path or another. Figure 6 illustrates this concept.

(23)

Figure 6 Biogas paths. The amount of biogas used in each path is dependent on several factors. The flow into upgrading is, thus, variable.

Commonly, isolated communities have unsatisfied energy needs. Biogas can be the solution, but the current development of upgrading systems have been centered around big scale systems with high quality and strict regulations for the final biogas composition (Bauer, et al., 2013). Those studies, pilot plants and current operation plants are not intended for such isolated locations. Therefore, another piece of the energy shift towards renewables and increased sustainability of energy systems is to find a solution for these needs.

Biogas holds a promising potential where simple technologies, low environmental impacts, ease of use, and low cost can be combined to supply affordable and clean energy for all.

1.2 Problem Formulation

International Micro Biogas AB (IMBAB) is a startup company that is investigating small-scale biogas systems.

Their focus is to provide self-standing units capable of providing an essential share of the energy demand of small farms and agricultural locations. The units would consist of an anaerobic digester, an upgrading unit, and a CHP plant. IMBAB has preselected high-pressure water scrubbing technique as the upgrading technology.

One of IMBAB’s goals is to develop “in-house” knowledge about their processes. Thus, to be able to predict and model their biogas systems. Modeling the upgrading unit will provide valuable information about the internal energy consumption; the upgraded quality of the biogas; the water flow needs; the sizes of accessory equipment like compressors, blowers, water pumps, and coolers, and the columns physical dimensions, I.E., height and diameter.

Their request is to develop an Excel-based model that can provide all this information in a user-friendly way.

The user should be able to produce results in an agile manner without the need of investing hours to set up the system.

(24)

1.3 Objective

The focus of this thesis is to develop a modeling tool that describes the performance of a small-scale water- based scrubber system for biogas upgrading. Small-scale refers to a production rate of raw biogas between 10 and 100 Nm³/h. The results of the model must provide:

• The upgraded flow and quality

• The height and diameter of the upgrading column

• The methane loss of the system (see slip definition in page 27, section 2)

• The power demand for ancillary equipment, I.E., compressors, water pumps, air blower and coolers

• The energy balance of the system

1.4 Methodology

The methodology proposed is based on numerical, empirical and analytical models, with a comparison of the model results against the results of similar models developed in a commercial software.

The project is based on the following hypotheses:

• The performance of an upgrading biogas system can be modeled by knowing the characteristics and properties of the packing elements of the scrubbing column.

• The influence of the methane loss is profoundly influenced by the presence and operating parameters of a flash tank.

The development of the model requires solving two different problems. One is solving the mass balance and calculation of operation parameters of the system conformed by a scrubber, flash tank, and a stripper. The second one is to calculate the height and diameter of the scrubber.

For the first problem —Solve the mass balance—, the challenge lays in the recirculation flows. The proposed methodology to solve the system is to implement an iterative process where the recirculation flows are updated in each step using the results of the previous step. This process is carried out until there is no significant change in the flows. This convergence is reached if the difference between two consecutive steps is less than a stopping value, typically set to 10^-10.

The advantage of this iterative process is that it resembles the operation of the startup of the system, each step resembles a transient period, and the behavior of the convergence process can be followed and analyzed. If no solution is found, for example, if the flows tend to zero or infinity, or there is a residual value that does not decrease with additional iterations, there is the possibility to see the evolution of the process, and infer the possible problems and propose solutions to overcome the difficulties.

One of the disadvantages of the process is that it might require large amounts of processing time if the number of iterations rises. Nevertheless, since the equations used in the mass balance are all solved analytically, each iteration process is expected to be significantly short in time. Another disadvantage is that a non-convergence case takes a long time to be pointed out since the iterations rise and the residue does not change, or the flows would tend to zero or infinity. This inconvenience could be avoided if some checks are performed before starting the iteration process. For example, one could check that the flow of water is enough to carry out the maximum separation, meaning that the slip would be very close to zero. If this check is passed, it means that there is a solution to the separation process.

For the second problem —calculate the height and diameter of the scrubber— the proposed methodology is to implement a variation of the one proposed by Billet and Schultes in two core publications (Billet, 1995)

(25)

and (Billet & Schultes, 1999). The variation consists in splitting the column into slices and applying the Billet and Schultes methodology to solve for the height and diameter of the slice.

The Billet and Schultes methodology is an application of the HTU-NTU method using Higbie’s penetration model as the governing mechanism for the mass transfer. In this method, a critical parameter to calculate is the liquid holdup of the column. This value is calculated by iteration, and it varies from zero to one since it is a fraction of the void volume. In this case, there is an asymptotic behavior of the function. The proposed methodology for this iteration is solving by sweeping the function from 0 to 1 by 0.1 increments of the liquid holdup. When the function changes sign, it means a solution between the two points has been found. Then a new sweeping process would be carried out between those points where the solution was found but with 0.01 increments. The increments will reduce until a stopping value has been reached, typically between 10^-

10 and 10^-16. For example, let us assume that the first sweeping process shows that the solution is found between 0.2 and 0.3. Then a second sweeping process is carried starting with 0.2, followed by 0.21, 0.22 and so on. Let us assume that the solution is found between 0.25 and 0.26; then a third sweeping process is carried out using 0.251, 0.252, 0.253… The process finishes when a sufficiently small interval of a solution is achieved.

This iteration method has a potential problem of jumping over the asymptote. If such is the case, a solution is to start with a smaller increment sweeping process, for example using 0.01. The processing time is increased but is more likely to achieved convergence.

Another important aspect of the Billet and Schultes methodology is that it is a combination of a theoretical framework developed using fluid and hydrodynamics, and empirical correlations result from extensive experimentations. The correlations are introduced to correct or adjust the theoretical model to the presence of the packing material. The correlations used are the ones reported by the authors in the referred publications.

1.5 Implementation

The project is divided into three general stages. First, a literature study and a theoretical framework will be developed. The second stage comprises the model development and implementation in Visual Basic for Applications (VBA). A comparison will be made against the same model carried in the commercial, specialized software like ASPEN Plus. The third and last stage consists of performing a sensitivity analysis on some dominant parameters and evaluate their impact on the performance of the tool.

1.6 Problems and Limitations

The main limitation identified is the scarcity of available data on small-scale biogas upgrading systems. In literature, it is usually considered small-scale those systems with throughput less than 100 Nm³/h. The aim of this study is centered at sizes between 10 and 100 m³/h.

Biogas is a mixture of several gases, but the main components are methane (CH4) and carbon dioxide (CO2).

The primary contaminant substance is hydrogen sulfide (H2S). Other contaminants are commonly present like ammonia (NH3) or volatile organic compounds (VOC). For this study, only CH4 and CO2 will be considered as the core upgrading need. All other substances present will be disregarded.

The results and sources are theory-based, and no laboratory tests, experimental data or real application will be developed or studied.

(26)

The physical details of the equipment are not covered nor analyzed in the study. The influence or limitations due to material compatibility, physical strength or resistance, controlling processes and other limitations are neglected in this study.

The process to be studied is based on physical absorption. All chemical reactions dynamics and related details are not part of the scope of the project and are thus disregarded.

Due to the low speed of biogas generation processes, all phenomena to be evaluated and analyzed will be considered only a steady-state process. The transient phenomena between the different stages of the digesting process are neglected and disregarded.

All models are maximum second-order mathematical models. The influence of higher order terms is neglected in the current study.

The water to be used in the process will be considered as drinking quality water and the effects of different substances present in the water, like minerals or dissolved gases, will be neglected.

1.7 Expected Outcomes

A VBA model implemented in Excel as hosting application of a water-based, small-scale biogas system, where different parameter can be changed like raw gas flow, raw gas composition (CH4, CO2), process pressure and temperature of the different streams, and pressure losses where relevant.

A written report detailing the mathematical framework, the model structure, results, sensitivity analysis, discussion, conclusions and possible future studies.

(27)

2 Upgrading Techniques

This section will discuss the most used techniques for upgrading. A comparison of benefits and drawbacks will be presented for each. The following descriptions are mainly based in (Cheremisinoff, 2000)

An important concept used in biogas upgrading is the “slip.” It is defined as the amount of methane that is removed —and lost— from the original biogas flow. Slip can be expressed as the total amount lost in Nm³/h, mol/h or kg/h, or as a fraction of the total incoming methane flow.

2.1 Physicochemical Upgrading Systems

2.1.1 Absorption

Absorption is the flow of molecules of a substance into another substance. Absorption can be either physical or chemical. Physical absorption depends on the solubility of a substance into the other. Chemical absorption includes chemical reactions between the substances.

The principle is based on the absorption difference between the gases. For biogas, using water as the absorber, CO2 is around 25 times more soluble than CH4 at 25°C. For other absorbers, commonly referred as amine, like polyethylene glycol, monoethanolamine (MEA), diethanolamine (DEA) the difference is even more significant, one to two magnitude orders. This higher solubility reduces the absorber flow, operating pressure, column diameter and equipment installation costs. Another benefit is the lower CH4 slip, typically 0.04 to 0.1% - one order of magnitude lower than water scrubbing-. Nevertheless, the energy requirements to regenerate, the contamination of the absorber due to the presence of H2S or NH3 are drawbacks of these absorbers. Due to complexity to regenerate, contaminant buildup monitoring, amine breakdown, and foaming this upgrading technique is not used in small-scale systems (Deublein & Steinhauser, 2011) (Ong, et al., 2014).

Since the substances must get into direct and close contact with each other, it is necessary to achieve the highest transfer surface between them. Usually, and this is the case in the present study, one of the substances is a gas and the other is a liquid. To make them get in contact, there is some typically used equipment found in the industry like plate columns and packed towers among others.

The absorber can be regenerated and then reused in a continuous process. The regeneration process is often called stripping —the contrary to absorption—. The stripping process is the same physical process that happens in absorption but with opposite direction. Diffusion of molecules in the liquid-gas interphase have two driving forces. One is the concentration difference in the liquid-liquid side, and the other one is the partial pressure difference in the gas-gas side. See section 3 further where the process is detailed.

In the following subsections, the equipment that is commonly used for mass separation absorption is described. There are other types of equipment like Venturi scrubbers and cooler absorbers that are not covered in this study since their application is related to specific chemical processes.

2.1.1.1 Agitated Vessels

The gas and the liquid are introduced in a closed vessel, and mechanical agitation will provide the intimate contact between the substances. This method is useful when a slow reaction occurs within the substances.

One drawback is that when a continuous flow is needed, a series of vessels will have to be used. Another disadvantage is that it presents more pronounced pressure drops than other continuous flow equipment.

(28)

2.1.1.2 Bubble-plate Towers

Plate towers and packed columns are, typically, counterflow equipment that operates continuously. They are also used for distillation purposes. The liquid falls in steps through trays, while the gas flows up and passes through cups that form bubbles in the liquid contained in the tray. Figure 7 presents the basic principle. This method is useful when a low liquid rate is needed, when a slow reaction is needed, when there is a difficult absorption task, or if there is a risk of clogging due to solids deposition. The latter, since it is easier to clean via utility holes than packed towers-. Pressure losses are slightly higher than in packed towers (will be described below.)

Figure 7 Left: Bubble plate tower. Right: Picture of a tray with bubble caps for a distillation column. Figure adapted and picture obtained from (Sölken, 2018).

2.1.1.3 Spray Towers

Spray towers are empty vessels where at the top a sprayer is found. The sprayer will atomize the liquid to form small droplets which are free to fall. In their trajectory, they meet a gas flow going upward. The effectiveness of the method depends on the application. There are some disadvantages in its application.

The total surface tends to reduce as the droplets fall due to the coalescence effect, which takes place when two or more droplets merge to form a bigger one. Another drawback is that the liquid in contact with the gas is only a small layer of the surface of the droplet, the center of the droplet will remain at initial concentration, thus reducing the effectiveness of the process. For difficult absorption, a series of towers can be implemented. (Cheremisinoff, 2000)

2.1.1.4 Wet scrubbers – Packed Towers

Wet scrubbers -packed towers- consist of a column filled with what is called packing material. The liquid flows down contacting the packing material. The sum of all the wetted surfaces accounts for a large total absorbing surface. The gas can flow co-current, cross-current or countercurrent, being the later the most common. The liquid flow is controlled so that the column does not flood, and all the packing material gets wet. If the flow is too low, some part of the packing will remain dry reducing the total absorption area. If the flow is too high, the column will fill up with liquid and become flooded.

In some cases, the liquid load is several times higher than the vapor flow, and the liquid becomes the continuous phase where the gas bubbles upward. This operation state is called phase inversion, and one of the characteristics is that it presents a relatively high-pressure drop of the gas. (Billet, 1995)

(29)

The general process is graphically explained in Figure 8. The liquid is recirculated between the scrubber and the stripper. Depending on the biogas contamination the liquid is replaced more or less frequently due to chemical or thermal breakdown or degradation, and contamination buildup. This process can use different washing liquids like water or amines. A common practice is to use a flash recovery tank, where part of the CH4 dissolved in the liquid is recovered by a drastic pressure drop.

There are several versions of the process, depending on the source and final use of the biogas, some of the options are (Wellinger, et al., 2013), (Bauer, et al., 2013), and (Allegue & Hinge, 2011):

• One or two stages of compression with cooling and hydrate separation upstream from the scrubber.

• H2S removal treatment upstream from scrubbing —external from digester or in situ—.

• With or without a flash unit.

• With or without regeneration of water.

• With or without waste gases treatment.

• With or without water filtration.

Figure 8 Water scrubber process. Adapted from (Wellinger, et al., 2013) and (Bauer, et al., 2013).

Depending on the scrubbing substance and the final use of the biogas a dryer is needed downstream or upstream of the scrubbing column. For example, if water or amine are used, the dryer is installed downstream since the gas will exit saturated. The most used scrubbing substances, water and amine solution, will separate CO2 and H2S at the same time, but H2S will be regenerated in the stripping process and must be treated in the waste gas flow. Therefore, a common practice is to clean the biogas from H2S upstream of the scrubber.

There is a wide range of commercial packing materials available. In Figure 9, some examples of packing material are shown.

(30)

Figure 9 Different industrial packing materials. 1 Cylindrical ring. 2 Berl-Saddle. 3 Novalox® Saddle. 4 Pall® ring (metallic). 5 Pall® ring (plastic). 6 Interpack®. 7 VSP®. 8 Igel®. 9 Novalox® - M. 10 VFF.Power-Pack®. Images obtained from (Vereinigte Füllkörper-Fabriken GmbH & Co. KG, 2016)

Another option for packing material is the so-called structured or stacked packing. These consist of organized and specially arranged packing units. In general, they have lower pressure drops and higher liquid flows but can result in higher building costs and difficulty to perform cleaning maintenance routines. Two examples of structured packing are shown in Figure 10.

Figure 10 Structured packing. Images obtained from (Raschig Gmbh, 2018).

2.1.2 Adsorption

Adsorption is the physical process of adhesion of molecules of a substance on the solid surface of other substance. Contrary to absorption, the adsorbed substance does not penetrate the other substance; it stays adhered to the surface. Adsorption can be physical called physisorption, or chemical called chemisorption.

(Cheremisinoff, 2000)

The physical principle behind physisorption is driven by intermolecular interaction, electrostatic, van der Waals forces, or by the physical structure of the pores in the adsorber. It is the interaction between these different properties of the adsorber and the adsorbate that will determine the degree of adsorption.

For physical adsorbers, regenerations is also a common practice. The process is driven by one or a mix of pressure, temperature of concentration swing.

Chemisorption is based on the phenomenon that chemical bonds will develop on the surface with none or little, diffusion between the adsorbate and the adsorber. Typically, these reactions are exothermic and with greater heat generation than the heat rejection taken place when physisorption occurs.

(31)

The most used adsorber is active carbon. It is considered as a physical adsorbent. Commercially, it is found in natural grain —granular—, pellets or powder. The granular type is mostly used for gas-solid adsorption processes. The size of the pore and the properties of the adsorbate define how effective the adsorption process is.

The gas-adsorption process is designed in different ways, but the most common is the bed type. A flow of gas is forced to go through a granular active carbon bed. The adsorption process will decrease as the concentration of the adsorbate increases in the adsorber.

The adsorption phenomenon is widely used in industry and especially in the chemical sector. Regeneration can be done in multiple ways depending on the adsorbate and if chemisorption or physisorption is in place.

For this study, only the pressure swing adsorption process will be described as is widely used in the biogas upgrading process.

2.1.2.1 Pressure Swing Adsorption (PSA)

In the biogas upgrading process, this technique is usually carried out using active carbon as the adsorber material, and regeneration occurs with a pressure drop to free the adsorber of the CO2. The process is graphically described in Figure 11.

Figure 11 PSA plant configuration. Four columns continuous process. Image adapted from (Wellinger, et al., 2013).

As it can be seen, is a batch process, where adsorption takes place in one vessel, while regeneration takes over in the other vessel. When the adsorber is close to saturation, the vessels are switched. Regeneration is achieved by lowering the pressure —pressure swing— in the saturated vessel. The typical setup for continuous operation is a four-column configuration, known as the Skarstrom cycle (Hoyer, et al., 2016). In this setup, one of the vessels is always depressurizing, so there is a continuous flow of upgraded biogas.

Typical adsorbent materials are active carbon, zeolites, silica gels, and carbon molecular sieves. According to (Hoyer, et al., 2016) there is active research in metal-organic new adsorbents and combining different adsorbers into a single column to clean and upgrade at the same time.

One of the typical drawbacks of this system is that the raw biogas must be cleaned from H2S and dried before entering the columns. H2S will react with the adsorber while CO2 will be physically adsorbed. The chemical binding will limit the lifespan of the adsorber as an increasing loss of active material will be available through the cycles. To avoid this problem, an active carbon filter is placed upstream the columns and H2S, as well as other contaminants, are removed and thus protect the adsorber in the column (Hoyer, et al., 2016).

(32)

2.1.3 Membrane Permeation

Another technique to separate gases, liquids, or mixes of them is done using the selective permeability properties of one or a set of membranes. For biogas upgrading, hollow polymeric fibers are used. One crucial factor is the selectivity of the membrane to allow the flow of CO2 and avoid CH4 slip (Hoyer, et al., 2016).

Typically, the membranes used are sensitive to liquids, oils, and impurities. Some conditioning should be carried out before the upgrading. Low concentrations of H2S will be separated with CO2, in addition, they do not harm the membrane markedly. Nevertheless, if condensates of water are present, they will react to form acids that do harm the membrane and will significantly lower the lifespan of the system. Other contaminants, like NH3 when dissolved in water from condensation, or volatile organic compounds (VOC) can harm the membrane irreversibly. Commonly a set of filters or a scrubber is used before passing the gases through the membrane (Hoyer, et al., 2016).

Commonly, operating pressures are around 10 – 20 bar. Compared to other upgrading processes, the pressure is high. This elevated pressure can be useful for subsequent compression for storage or compressed biogas as vehicle fuel (CBG). Nevertheless, for some other applications, the pressure must be reduced meaning that the energy stored as pressure will be lost or has to be recovered using additional equiptment (Ong, et al., 2014).

Depending on the application, the process can consist of one, two or three stages of membranes as shown in Figure 12. Adding stages will further reduce the recycled gas and the CH4 slip, but it will increment the energy intensity of the process (Hoyer, et al., 2016).

Figure 12 Different membrane upgrading configurations. Image adapted from (Hoyer, et al., 2016).

Organic membrane materials are preferred over inorganic due to higher surface-volume ratio, superior structural integrity, lower production cost, advanced stage of development, stable at high pressures (Ong, et al., 2014). The hollow fibers are bundled into small units to allow ease of maintenance and replacement. This set up of several units working in parallel give this upgrading system the ability to perform well at partial loads, by closing some of the membrane units (Wellinger, et al., 2013).

Small-Scale Biogas Upgrading System Modeling Tool Development