FACTOROPTIMIZATIONFORANAEROBICBIOGASGENERATIONFROM PALMOILMILLEFFLUENTUSINGARTIFICIALNEURALNETWORK

ABSTRACT

 

 

 

A preliminary study of anaerobic digestion of Palm Oil Mill Effluent (POME) was undertaken using a 1L gasometric assembly. Biogas generation at varying pH values (6.0 to 9.0) and temperature range (30^oC to 60^oC) were investigated in 14days. The results revealed optimal biogas yields at pH of 7.0 and temperature of 50^oC. An Artificial Neural Network (ANN) model was applied in other to optimize the yield in anaerobic digestion of POME using experimental data from an integrated Bioreactor. The integrated bioreactor system comprised of a standard up flow anaerobic sludge bed (UASB) (Module 1), a centrally pump- stirred twin vessel bioreactor (Module II) and a standard continuous stirred tank (CSTR) (Module III). Operational data (Module I, II and III), were collected and employed in the analysis using Neural Network toolbox embedded in MATLAB R2015a. The study considered the effect of digester operational parameters, such as temperature (T), pH, C:N ratio, HRT (days), % C, % N, % P, % S and % COD reduction on the cumulative methane yield. A Two-layer feed forward neural network was trained with Levenberg-Marquardt Back Propagation training algorithm (trainlm) to simulate the digester operations. This was capable of predicting the yields of the methane production of the anaerobic digestion process as factors were varied simultaneously. The performance of the ANN model was validated and demonstrated the effectiveness of the model to predict the methane production accurately. The model with neural network structure 9-5-1 produced the best performance with the generated models showing a high correlation coefficient (R²-value). The linear regression between the network outputs and the corresponding targets showed a good correlation indicating that the outputs matched the targets closely based on the obtained ANN regression R² values which ranged between 0.98262 to 0.99719. Application of the functional relationship of the methane process generated a set of model equations, which revealed that the methane yield can be predicted by a family of power series functions of the type:

y = 2.7 x 10^-1x⁴ – 94x³ + 1.2 x 10⁴x² – 7 x 10⁵x + 1.5 x 10⁷ (R² = 0.9976), where y is the

methane yield and R² expresses agreement between observed values of methane yield and theoretical constructs. These equations are novel and have high predictive and analytical capabilities that can be used to predict methane generation as dependent on factors such as % COD reduction, pH and HRT. Finally, the proposed model has an acceptable generalization and can be used for further predictions for data sets with similar experimental conditions.

Keywords: Anaerobic digestion, artificial neural network, integrated bioreactor, POME, Optimization and model.

 

 

TABLEOFCONTENTS

 

Page

Titlepage                                                                                                                                                          i

Certification                                                                                                                                                    ii

Dedication                                                                                                                                                        iii

Acknowledgements                                                                                                                                     iv

Abstract                                                                                                                                                            vi

TableofContents                                                                                                                                         vii

ListofTables                                                                                                                                                  xi

ListofFigures                                                                                                                                                xiii

ListofAbbreviations                                                                                                                                  xxi

 

CHAPTERONE:                       INTRODUCTION                                                                       1

1.1         Background to Study                                                                                                                   1

1.1.1     Alternative Energy Sources                                                                                                      2

1.1.2     Biomass Energy                                                                                                                             3

1.1.3    Biogas Generations as a Waste Treatment Technology                                              4

1.1.4     Waste Generation in Palm Oil Mills                                                                                     5

1.1.4.1 Gaseous Emission                                                                                                                         6

1.1.4.2 Environmental Regulations of Effluent Discharge                                                       6

1.1.4.3 Energy Sources                                                                                                                               7

1.1.4.4 Palm Oil Mill Effluent Treatment System                                                                        8

1.1.4.5 Optimization Methods                                                                                                                 8

1.1.5.1 Simplex Methods                                                                                                                           9

1.1.5.2 Optimization Methods In Anaerobic Digestion Processes                                       10

1.1.6      Optimization Using Artificial Neural Network                                                             12

1.2         Statement of Problem                                                                                                                  13

1.3 Research Aim and Objectives          15 1.4 Hypothesis of The Study             15

1.5         Significance of Study                                                                                                                  16

1.6         Scope of The Study                                                                                                                      17

 

CHAPTERTWO:LITERATUREREVIEW 18 2.1 Anaerobic Digestion of POME 18

2.1.1     POME Treatment                                                                                                                          19

2.2        Anaerobic Digestion Compared with Alternative Methods                                     20

2.2.1     Conventional Treatment Systems                                                                                          21

2.2.2     Anaerobic Filtration                                                                                                                     22

2.2.3     Fluidized Bed Reactor                                                                                                                24

2.2.4    Up-Flow Anaerobic Sludge Blanket (UASB) Reactor                                                26

2.2.5    Up-Flow Anaerobic Sludge Fixed-Film (UASFF) Reactor                                      28

2.2.6     Expanded Granular Sludge Bed (EGSB)                                                                           30

2.2.7    Continuous Stirred Tank Reactor (CSTR)                                                                        31

 

VI

2.2.8    Anaerobic Contact Digestion                                                                                                   32

2.3        Comparison of Various Anaerobic Treatment Methods of POME                        33

2.4        Factors Affecting Anaerobic Digester Performance                                                    36

2.4.1     Effect of pH                                                                                                                                      37

2.4.2     Effect of Mixing                                                                                                                            38

2.4.3     Effect of Temperature                                                                                                                 39

2.4.4     Effect of Organic Loading Rates                                                                                           41

2.4.5     Effect of Hydraulic Retention Time                                                                                     41

2.4.6     Effect of Carbon: Nitrogen Ratio                                                                                           42

2.4.7     Effect of Nutrients for Bacteria                                                                                              42

2.4.8    Effect of Chemical and Physical Pre-Treatment                                                            43

2.5         Mechanism of Anaerobic Digestion                                                                                     45

2.5.1     Hydrolysis                                                                                                                                         45

2.5.2     Acidogenesis                                                                                                                                   45

2.5.3     Acetogenesis                                                                                                                                   46

2.5.4     Methanogenesis                                                                                                                             46

2.5.5    Biochemical Mechanism of Anaerobic Digestion                                                         47

2.5.6     Comparative Reactor Performance                                                                                      48

2.5.7       Identification of Responsive Microorganisms                52 2.5.8               POME Stabilization in terms of COD removal               53

2.6         Kinetic Studies                                                                                                                                54

2.6.1     Theoritical Mathematical Model                                                                                           56

2.6.2     Advanced Model for Biogas Yield                                                                                       58

2.6.3     Simple Model for Biogas Yield                                                                                             59

2.6.4     Summary of Bio-Chemical Kinetics Models                                                                   60

2.7        Empirical Optimization Conditions                                                                                      63

2.7.1    Reactor Optimization by application of Simplex Models                                          63

2.8         Computer Assisted Modelling                                                                                                 64

2.8.1     Data Simulation                                                                                                                             64

2.8.2     Predictive Analytics                                                                                                                     64

2.8.3     Multilayer Perceptron                                                                                                                 65

2.8.4 Artificial Neural Networks            65 2.8.5 Uses of Artificial Neural Nets              67 2.8.6 Types of Artificial Neural Networks 69

2.8.7    Algorithms and Logic Systems Associated with ANN                                               70

2.8.8     MATLAB                                                                                                                                          70

2.8.9     Curve Fitting Tool Box                                                                                                               71

2.9         Functions Used to Process Data in ANN                                                                           71

2.9.1     Step Function                                                                                                                                   72

2.9.2     Radial Basic Function                                                                                                                 73

2.9.3    Linear Combination and Continuous Log-Sigmoid Function                                 73

2.9.4 Learning Paradigms          75 2.9.5 Gradient Descent         76 2.9.6 Backpropagation 76

2.9.7     Boltzmann Learning                                                                                                                    78

2.9.8     Hebbian Learning                                                                                                                         79

 

 

 

 

 

2.10     Application of ANN in Environmental Waste Management System

Modelling                                                                                                                                          79

2.10.1 ANN for Modeling of COD Removal from Waste Aqueous Antibiotic

Solutions                                                                                                                                            80

2.10.2 Modelling with ANN for Dosage Control and Color Removal from

Waste Textile Water                                                                                                                     82

2.10.3 Application of ANN for Prediction of Methyl Tert-Butyl Ether (MTBE)

Degradation                                                                                                                                      83

2.10.4 Application of ANN for Predicting Nitrogen Oxides Removal Efficiency 85

2.10.5 Application of ANN in Air Quality Prediction                                                                85

2.10.6 Simulation of Process Control using nonlinear Systems Control                           87

2.11      Application of ANN in Experimental Data Analysis of Biogas Production 88

2.11.1 Application of ANN in Biogas Production from Sago Wastewater using

Anaerobic Tapered Fluidized Bed Reactor                                                                       88

2.11.2 Application of ANN in Biogas Production from fresh Organic Waste                 88

2.11.3 Application of ANN in Biogas Production                                                                        89

2.11.4 Application of ANN in Biogas Production from Field-Scale Landfill

Bioreactors                                                                                                                                       90

2.11.5 Application of ANN in Biogas Production from Anaerobic Co-Digestion

of Leachate with Pineapple Peel                                                                                            91

 

 

CHAPTERTHREE:                 METHODOLOGY                                                                       94

3.1         Research Materials                                                                                                                       94

3.1.1     Laboratory Equipment Used                                                                                                    94

3.1.2     Laboratory Reagents Used                                                                                                        94

3.2         Reactor Rig                                                                                                                                      95

3.2.1    Definition of Bioreactor Modules                                                                                          98

3.2.2     Biogas Collector                                                                                                                            98

3.3        Selection of Substrate and its Volume                                                                                 99

3.4         Laboratory Analysis                                                                                                                     99

3.4.1 Characterization of the Feed                                                                                                       99

3.4.2     pH Determination                                                                                                                         100

3.4.3 Nitrogen/Crude Protein Determination 100 3.4.4 Determination of Total Organic Carbon (TOC) 101 3.4.5 Determination of Volatile Solids 102

3.4.6     Determination of Biochemical Oxygen Demand (BOD)          103 3.4.7                 Determination of Chemical Oxygen Demand (COD)                104

3.4.8     Determination of Alkalinity                                                                                                     106

3.4.9     Determination of Density                                                                                                          106

3.5 Preliminary Studies (Gasometric Assembly) 106 3.5.1 Experimental Variables (Preliminary Studies) 106 3.5.2 Preparation of Feed Slurry for the Variable Determination 108 3.5.3 Analysis of Results of Preliminary Studies 108

3.5.4    Gas Chromatography (GC) Analysis of Product Mixture                                         109

3.6        Digestion of POME Using Integrated Reactors                                                              109

3.6.1    Experimental Initialization of the Integrated Bioreactor                                            109

 

 

3.6.2   Maintenance of Anaerobic Conditions                                                                                110

3.7        Modeling and Optimization of Biogas Production Using Artificial Neural

Network                                                                                                                                              110

3.7.1    MATLAB Neural Network Toolbox                                                                                    110

3.7.2    Neural Network Fitting Tool Box                                                                                         111

3.7.3       Neural Network Architecture 112 3.7.4            Neural Network Size   113

3.7.5     Neural Network Training                                                                                                          114

3.7.6    Model Validation                                                                                                                           117

3.7.7    Artificial Neural Network Model for Biogas Yields Obtained                                117

 

CHAPTERFOUR:RESULTSANDDISCUSSION                                                               119

4.1         Results on Preliminary Studies                                                                                               119

4.1.1    Data on Variable Factor Determination                                                                              122

4.1.2     ANOVA based on Preliminary Analysis                                                                           123

4.1.3 Methane Yield from Integrated Bioreactors                                                                        126

4.2         Artificial Neural Network Model                                                                                          130

4.2.1    Graphical Representations of the ANN Training Results                                          131

4.2.3    Interpretation of Figures and Graph from ANN                                                             155

4.2.4     Data Mining and Normalization                                                                                            156

4.2.5    Mathematical Modelling to Circumvent Kinetics of Biodegradation                  156

4.3        MATLAB Code for Neural Networks Model                                                                  158

4.4        Functional Relationship Between Percentage COD Reduction and the

Biogas Yields                                                                                                                                  160

4.5        Functional Relationship Between pH and the Biogas Yields                                   166

4.6       Functional Relationship Between Time (HRT) and the Biogas Yields               174

4.7        Chromatographic Analysis of the Biogas Produced                                                     181

4.8        Feed Stock/ Waste (Spent Slurry) Characterization                                                     181

4.9        Economic Analysis of Biomethane Processing                                                              182

 

CHAPTERFIVE:   CONCLUSIONANDRECOMMENDATION                            185

5.1         Conclusion                                                                                                                                        185

5.2         Contribution to Knowledge                                                                                                      186

5.3         Recommendations                                                                                                                         187

 

REFERENCE                                                                                                                                                188

 

 

CHAPTER ONEINTRODUCTION

 

                                     1.1       BACKGROUNDTOSTUDY

 

Theglobeisfacedwiththetwincrisesofsafeenergygenerationandclimatechange.Global

 

warming as a result of CO₂increaseintheatmosphereduetotheuseoffossilfuelscauses

 

greenhouseeffect(Courtney&Dorman,2003;Ejike&Anyile,2011;Ramadhas,Jayaraj,&

 

Muraleedharan,2004).Hence,theneedtoexploreandexploitnewenergysourcesthatare

 

renewableandecologicallysustainableispertinent(Dow&Downing,2006;Mills,2009).

 

Furthermore,ithasbeendemonstratedthatenergysupplyandutilizationcanbeoptimized

 

byincorporatingdifferentfactorsthatenhancebestmanagementpracticesintheIndustry

 

(Ejike&Anyile,2011).Biogasproductionhasemergedasoneofthepromisingalternatives

 

tofossilfuelsources,andthedevelopmentofthistechnologywillhelphumanityto

 

overcomethecurrentenergycrisisandalsoprovideacleansourceofenergytocombatthe

 

phenomenonofglobalwarming(Levin,Pitt,&Love,2004;Nath&Das,2011).

 

Anaerobicdigestion(AD)isaversatilebiochemicalprocess,whichiscapableofconverting

 

mosttypesofbiodegradableorganicmatterunderanaerobicconditionintoamixtureof

 

gasesconsistingofmethaneandcarbondioxide(Blonskaja&Vaalu,2006).ADprocessis

 

controlledbyseveralprocessparametersandthedeterminationoftheoptimumvaluesof

 

theseparametersarecrucialforbioprocessdevelopmentandscale-up(Wang&Wan,

 

2009b).ADprocessesaredescribedtobenon-linear,complexandunsteady;thus,itis

 

challengingtodevelopaprecisephysical-basedformulatocharacterizetheirphysical

 

performance.Similarly,thedevelopmentofaccurateADprocessmodelscontinuestobaffle

 

expertsduetothenon-linearnatureofthebiochemicalnetworkinteractionsthatoccur

 

duringbioprocesses(Franco-Lara,Link,&Weuster-Botz,2006).Modellingand

 

optimizationofbioprocessescontributestoincreasedunderstandingoftheprocessinputsfor

1

 

 

 

optimumyieldandproductionrate.Variousmodellingalgorithmshavebeenappliedin

 

biogasproductionprocesses,andresultshaveshownthatmodellingandoptimizationcan

 

enhancebiogasyields(Sewsynker-Sukai,Faloye,&Kana,2017).

 

However,theimprovementofbiogasproductionrequiresarobustprocessmodelthat

 

accuratelyrelatestheeffectofinputvariablestotheprocessoutput.Artificialneural

 

networks(ANNs)haveemergedasoneofthelatesttoolsforoptimizationandmodelling

 

complex,non-linearprocesses.ANNsareappliedtopredictvariousprocessesthatareuseful

 

forvirtualexperimentationsandcanpotentiallyenhancebioprocessresearchand

 

development(Sewsynker-Sukaietal.,2017).

 

                                1.1.1   AlternativeEnergySources

 

Duetotoday’senergydemandandthepossibilityofdepletionofnon-renewableenergy

 

resources,theneedtoexploreandexploitnewenergysourcesthatarerenewable,secureand

 

ecologicallysustainableispertinent.Alternativeenergysourcesevolvinginthelasttwo

 

decades are gaining attention all over the world because they do not cause increase in CO₂

 

compositionoftheatmosphere.Theyarealsoknowntoenhanceenergyconservationand

 

providenewjobopportunities(Ejike&Anyile,2011;Mshandete,Björnsson,Kivaisi,

 

Rubindamayugi,&Mattiasson,2006).Inordertoencouragethediversificationofenergy

 

generation,variousprotocolshavebeendevelopedtogiveassistancetoenergygeneration

 

methodswhichdonotdirectlydependonpetroleumandnaturalgasreserves(Wayman&

 

Parekh,1990).Variousalternativeenergysourcesinclude;solarenergy,geothermalenergy,

 

nuclearenergy,hydro-power,windenergyandbiomassenergy.

 

 

 

Thedifferentalternativeenergysourceshavepotentialnottoincreasetheconcentrationof

 

carbondioxide (CO₂) in the atmosphere. However due to the contribution of CO₂byfossil

 

fuels,theyarenolongerregardedasthebestoptioninmodernenergysupply(Ejike&

 

Anyile,2011;Ejike,Chinedu&Egbujor,2012).ThepredictionmadebyDowandDowning

 

 

(2006) has led to several protocols on the management of CO₂emissionandother

 

greenhousegases.Themanagementsystemhasledtowhatisreferredtoas“CarbonCredit”.

 

Acarboncreditisagenerictermforanytradablecertificateorpermitrepresentingtheright

 

toemitonetoneofcarbondioxideormassofanothergreenhousegaswithacarbondioxide

 

equivalent.Carboncreditandcarbonmarketarecomponentofnationalandinternational

 

attemptstomitigatethegrowthinconcentrationofgreenhousegases(ClimateChange

 

Glossary,2008).Biomassenergybeinglessenvironmentallydemandingisadaptableto

 

operationindevelopingcountries.

 

1.1.2BiomassEnergy

 

Thisinvolvesplantremainsbeingconvertedintosmallermoleculeswhichcouldbeburntto

 

generateenergy.Twoprocessesarewidelyknown.First,thecelluloseofplantcellwallcan

 

be hydrolyzed to glucose which in turn is fermented into ethanol and CO₂(Singh,2001).

 

Secondly,anaerobicdigestionprocessofplantandotherorganicmatterbyaconcertiaof

 

microorganismstogivebiogas(Ghosh,Samanta,&Ray,2011).Theuseofbiogasfrom

 

biomass as an energy source is regarded as CO₂ neutral, because CO₂releasedduring

 

combustion of the biogas is the CO₂thattheplanthasassimilatedduringphotosynthesisto

 

createorganicbiomass(Jewel,1974).Thisisincontrasttoburningfossilfuelswhich

 

resultedfromplantthatlivedover20millionyearsbefore.

 

Biogasgenerationisthereforeoneofthemostdesirableoptionsamongthevariousother

 

alternativesourcesofrenewableenergycurrentlyavailable,asitrequireslesscapital

 

investmentperunitproductioncost(Ghoshetal.,2011).Ithasbeendemonstratedearlier

 

thatbiogasproductionfromcropresiduesiseconomicallyfeasibleonafarmscalelevel50–

 

500KW (Wayman & Parekh, 1990). Fermenting 8000 m³ofslurryformedfromorganic

 

fractionsofsolidwasteandanimalwastewithadepressionof10%generallygives

 

electricityat16Ev/KWh(Hoffman,2001).

 

 

 

 

Introductionoftheuseofrenewableenergyisusuallyapoliticallydrivendecision.The

 

reasonisthatrenewableenergyismoreexpensivebutitsusebylargepopulationsdecreases

 

thecost.Europeancommunityhasagreedonatarget,forconversionofenergybased

 

sourcesandtheirstandardwillbe20%renewableenergybytheyear2020(European

 

EnergyCommission2010).

 

                                1.1.3BiogasGenerationsAsAWasteTreatmentTechnology

 

Biogasaregeneratedfromfoodindustrywastesuchasdairywaste,chocolate,sugarcane,

 

beefhoof,potatoes,oilseed(e.g.olive,palmfruit,coconutfruit),brewerywaste,vegetable,

 

suchastomatoes,mangoandguavajuices.Othersubstratesforbiogasgenerationinclude

 

municipalorganicwaste(MOW),sewage,manurepoultry,piggeryandcattlefarms

 

(Baharuddimetal.,2010;Demirel,&Schere,2008;Fang,Angelidaki,&Boe,2010).

 

Biogastechnologyhasbeenshowntobeasustainablealternativeforwastetreatmentin

 

municipalandindustrialwastedisposalsystem.Indeed,thetownofLinkopinginSweden

 

obtained90%oftheirrequiredenergyneedfromprocessinganimalmanureparticularlypig

 

manureimportedfromtheNetherland.Ninety-eightpercentorganicwastesfromhousehold

 

inthecityarerecycledthroughprocessingtobiogasandusedforelectricitygeneration

 

(Philipsson,2001).InLinkopingCity,70%ofthecityvehiclesrunonbiogasfrom

 

municipalorganicandanimalwastes.

 

LinkopingCitysupportstheslightlyhighercostofbiogasgeneratedenergythroughcarbon

 

tradeexchangewhichisastockexchangeforcarboncredit.HenceLinkopinghashighly

 

validatedandsophisticatedprocesses,whichallowstheirassessmentasacleanenergy

 

developmentenvironmentintheinter-governmentalpanelonclimatechange(IPCC,2012)

 

(KyotoWorkDataPoliciesInfrastructures,2007).Thisisinlinewiththerequirementofthe

 

KyotoProtocolonthecapsandquotaforgreenhousegas(KyotoProtocolTargets,2010).

 

 

 

 

1.1.4WasteGenerationInPalmOilMills

 

The main produce of a palm oil mill is the crude palm oil (CPO). Other products are liquid

 

effluent and solid wastes (empty fruit bunches, palm kernel, fibre and shell), which may

 

haveasignificantimpactontheenvironmentiftheyarenotdealtwithproperly.The

 

productionofpalmoilresultsinthegenerationoflargequantitiesofwastewatercommonly

 

referredtoaspalmoilmilleffluent(POME).Typically,1tonneofcrudepalmoil

 

productionrequires5-7.5tonnesofwater;over50%ofwhichendsupasPOME(Ma,

 

1999).ThePOMEcomprisesacombinationofwastewaterfromthreemainsourcesviz.

 

clarification(60%),sterilization(36%)andhydrocyclone(4%)units(Ma,1999).

 

 

 

POMEcontainsvarioussuspendedcomponentsincludingcellwalls,organelles,shortfibres,

 

aspectrumofcarbohydratesrangingfromhemicellulosestosimplesugars,arangeof

 

nitrogenouscompoundsfromfreshfruitbunch.Fromenvironmentalperspective,fresh

 

POMEisahotandacidicbrownishcolloidalsuspension,characterizedbyhighamountsof

 

totalsolids(40,500mg/l),oilandgrease(4000mg/l),COD(50,000mg/l)andBOD(25,000

 

mg/l)(Ejike,Chinedu&Egbujor,2012;Singh,Ghumen,&Grewal1999;Ma,1999).The

 

chemicalcharacteristicofatypicalPOMEissummarizedinTable1.1.

 

Table 1.1 Typical characteristics of POME (Ma, 1999).

 

Parameter pH	Average	Metal	*Average
	4.7	Phosphorous	180
Oil and Grease	4000	Potassium	2270
Biochemical Oxygen	25000	Magnesium	615
Demand (BOD5) Chemical Oxygen Demand (COD)	50000	Calcium	439
Total Solids	40500	Boron	7.6
Suspended Solids	18000	Iron	46.5
Total Volatile Solids	34000	Manganese	2.0
Ammonical Nitrogen	35	Copper	0.89
Total Nitrogen	750	Zinc	2.3
* All in mg/l except pH.

*

 

 

 

                                1.1.4.1  GaseousEmission

 

Palm oil mills are generally self-sufficient in terms of energy requirements due to the

 

availability of adequate quantities of fiber and shell materials that are used as solid fuel in

 

the boiler. There are two sources of air pollution in the mills, the boiler and incinerator that

 

arecausedbyincompletecombustionofthesolidwastematerials;wastefiber,shell

 

materials and empty fruit bunch (Thani, Hussin, Wan Ramlah, & Sanusi.,1999).Themain

 

practice of treating POME is by using ponding and/or open digesting tank systems (Ma,

 

1999). The emission of greenhouse gases (CH₄ and CO₂)fromtheseopensystemstothe

 

atmospherehasbeenrecentlyreportedasasourceofairpollutionfromthepalmoilmills

 

(Yacob, Hassan, Shirai, Wakisaka, & Subash,2005).

 

                                1.1.4.2    EnvironmentalRegulationsofEffluentDischarge

 

Theenvironmentalcontrolinpalmoilindustrywasgrantedalicensedapproachthatwould

 

permitintimatecontrolofindividualfactories.Thisapproachalsoprovidesamechanismfor

 

permittingvariableeffluentstandardstobeappliedbasedonthedemandsofprevailing

 

environmentalcircumstances.Theenvironmentalqualityregulationsforthecrudepalmoil

 

industrywerethefirstsetofregulationspromulgatedundertheEnvironmentalQualityAct

 

(EQA), 1977, for control of industrial pollution sources (Thani et al.,1999).Theeffluent

 

dischargestandardsapplicabletocrudepalmoilmillsarepresentedintheTable1.2.

 

Table 1.2 Effluent discharge standards for crude palm oil mills (EQA 1974, 2005).

Parameter	unit	Parameter limits	Remarks
BiochemicalOxygenDemand	mg/l	100
ChemicalOxygenDemand	mg/l	*
TotalSolids	mg/l	*
SuspendedSolids	mg/l	400
OilandGrease	mg/l	50
AmmoniacalNitrogen	mg/l	150	Valueoffilteredsample
TotalNitrogen	mg/l	200	Valueoffilteredsample
pH	–	5-9
Temperature	°C	45

Note:*Nodischargestandardafter1984.

 

                                1.1.4.3  EnergySources

Duetotheincreaseindemandforenergy,costsavingandtheprotectionoftheenvironment,

 

otherenergysourcesthatprimarilycompriseofoil,naturalgas,hydropower,coal,solar

 

powerandbiomassarenowbeingexploited.AspresentedinTable1.3,naturalgas,

 

hydropower,andbiomassenergyresourcesinMalaysiaaregenerallyabundant.

 

 

 

Table 1.3Energy resource potential in Malaysia (ASEAN, 2003).

Energy resources	Amount	Unit
Oilreserve	5.0	Billionbarrels
Gasestimatereserve	2402	Billioncubicmeters
Coalprovenreserve	–	Milliontonnes
Hydropowertechnically	72	Twh/y
FeasibleBiomass	665	MW
Geothermalpotential	–	MW
Windenergypotential	–	MW

 

 

InMalaysia,themostextensivestudyontheuseofbiomasshasbeenonpalmoilwastes,

 

whichcanbeutilizedtomeettheenergyrequirementofthepalmoilmillsandtheelectricity

 

needsoftheworkers.Thetotalenergypotentialofthebiomassisestimatedtobeabout5%

 

ofMalaysianelectricitydemand.Therefore,renewableenergyhasbeenidentifiedby

 

Malaysian government as the 5^thfuelunder‘TheNewFive-FuelDiversificationStrategy’

 

(EuropeanEnergyCommission,2002).Therearedifferenttypesofbiomassgeneratedbya

 

palmoilmill.Fromthefourbiomasssources,threeofthem(emptyfruitbunch,fiberand

 

shell)canbedirectlyburnedasfuelwhilePOMEmustfirstbeanaerobicallyconvertedto

 

methane.Therefore,itisessentialforahighrateanaerobicbioreactortobeappliedasitcan

 

servedual-functioni.e.wastewatertreatmentandenergygeneration(organicconversionto

 

methane).

 

 

 

 

 

 

 

                                1.1.4.4PalmOilMillEffluentTreatmentSystem

 

Palm oil industries are facing tremendous challenges of meeting the increasingly stringent

 

environmental regulations. Over the past decades, several cost-effective treatment

 

technologies comprising of anaerobic, aerobic and facultative processes have been

 

developed for the treatment of POME (Ma,1999).Morethan85%ofpalmoilmillsuse

 

solely ponding systems due to their low costs (Thani et al.,1999).Ithasbeenreportedthat

 

onlyafewmillsareequippedwithbiogasrecoverysystems(Yeoh,2004).Longhydraulic

 

retentiontimes(HRT),lowtreatmentefficiency,highsludgeproduction,extensivelandarea

 

requirement and emission of large amount of greenhouse gases (CO₂ and CH₄)aresomeof

 

thedrawbacksofconventionalPOMEtreatmentmethods.However,theoptimizationand

 

applicationofefficient,stableandeconomichighrateanaerobictreatmentsystemsarebeing

 

investigated (Yacob et al.,2005).

 

                               1.1.5    OptimizationMethods

 

Optimizationwithmodellingandsimulationhasbeenanintegralpartofmoderndesign

 

practice.Themainaimofanydesignproblemistofindasetofgood,feasibleandideally

 

bestsolutionstoagivenproblem,whichmeasuresuptodesigncriteriasuchastheminimum

 

costs,highperformance,sustainability,recyclabilityandenergyefficiency.Thestringent

 

requirementsofminimizingenvironmentalimpactandcarbonfootprintrequireaparadigm

 

shiftinscientificthinkinganddesignpractice.Despitethesignificantprogressmadeinthe

 

lastfewdecades,therearestillchallengesrelatedtononlinearity,scaleoftheproblem,time

 

constraintandthecomplexityofthesystem(Yang,Koziel,&Leifsson,2013).Mathematical

 

modelinghasmostmodelsbasedonpartialdifferentialequationsthatarenotanalytically

 

solvable,hence,approximatemethodsandnumericalmethodsarethealternative.For

 

complex tasks in which the actual system behaviors have no explicit or closed expression

 

(equations), approximate learning-based models such as neural networks are very useful and

 

effective in practice (Yang et al., 2013). An algorithm development is one of the main activities in optimization. An algorithm is an iterative procedure for solving a class of

 

problemsandmustbefinite(optimalsolutionsarefoundinafinitenumberofiteration).

 

 

 

Inmathematics,optimizationisthedisciplineconcernedwithfindinginputsofafunction

 

thatminimizeormaximizeitsvalue,whichmaybesubjectedtoconstraints.Inrecenttimes,

 

researchershavesolutionstoreal-lifeproblemsasaresultofnewtechnological

 

developmentsinalgorithmsandcomputerhardware(Ba˜nos,Manzano-Agugliaro,Montoya,

 

Gil,Alcayde,&Gomez,2011;Yangetal.,2013;Yu,Wensel,Ma,&Chen,2013).

 

Optimizationdoesnotessentiallymeanfindingtheoptimumsolutiontoaproblem,sinceit

 

maybeunworkableduetothecharacteristicsoftheproblem,butinrecentdecadesmany

 

authorshaveproposedapproximatemethods,includingheuristicapproachesandartificial

 

neuralnetworks(ANN),tosolvetheseproblemsinsteadofusingtraditionaloptimization

 

methods,suchaslinear-programming(LP),SteepestAscent,Simplexmethod,quadratic

 

programming(QP),etc.(Yuetal.,2013).However,tilldatemostcomputational

 

optimizationmethodshavefocusedonsolvingbothsingle-objectiveandmulti-objective

 

problems.Thisapproachhasseveraldrawbacksasitonlyreturnsasinglesolutionasaresult

 

ofthesearchprocess,whichbecomesanimportantlimitationinthedecision-making

 

process,wherethedecisionmakermustselectonesolutionfromseveralalternatives(Ba˜nos

 

etal.,2011).

 

1.1.5.1SimplexMethods

 

TheSimplexmethodsarewell-knownalgorithmusedforlinearprogramming.Itpresentsan

 

organizedstrategyforevaluatingafeasibleregionsverticesandthishelpstofigureout

 

optimalvalueoftheobjectivefunctioninalinearprogram.Itisclearthatonefactoris

 

crucial to the method: which variable should replace which. Simplex method is very

 

efficient in practice. However, its worst-case complexity is exponential as demonstrated

 

with carefully constructed examples (Carreira-Perpinan, 2017). There is also a complexity of polynomial functions for both average and worst cases in linear programming. Linear

 

Programming Models can be used easily to solve two variables, but more than three to five

 

variables involves the simplex method as it is an iterative method which by repeated use

 

gives the solution to any n variable. Simplex Methods deals with standard form

 

(standardization)andthereisneedtosolveafewsetofequations.InSimplexMethod,the

 

primaryalgebraictaskofiterationistotransform,usingGauss-Jordanelimination,andthe

 

constraintequationsfromagivenconfigurationtoanewconfigurationthatcorrespondsto

 

thenextbasicfeasiblesolution.Suchtransformationsarerepeatedmanytimesinthecourse

 

ofthesolutionofaproblem(Carreira-Perpinan,2017).

 

1.1.5.2OptimizationMethodsInAnaerobicDigestionProcesses

 

AnaerobicDigestion(AD)processesaredescribedtobenon-linear,complexandunsteady;

 

thus,itischallengingtodevelopaprecisephysical-basedformulatocharacterizetheir

 

physicalperformance.ThedevelopmentofaccurateADprocessmodelscontinuestobaffle

 

expertsduetothenon-linearnatureofthebiochemicalnetworkinteractionsthatoccur

 

duringbioprocesses(Franco-Laraetal.,2006).Modellingandoptimizationofbioprocesses

 

contributestoincreasedunderstandingoftheprocessinputsforoptimumyieldand

 

productionrate.Variousmodellingalgorithmshavebeenappliedinbioprocesses,and

 

resultshaveshownthatmodellingandoptimizationcanenhancebiogasyields(Sewsynker-

 

Sukaietal.,2017).

 

Mathematicalandstatistical-basedmodelsprovidevitalinformationfortheunderstanding,

 

analysisandpredictionofbioprocessesandtheyarerequiredfortheoptimizationofkey

 

parametersinordertoimprovetheprocessoutput(Escamilla-Alvarado,Rios-Leal,&

 

Ponce-Noyola,2012;Yuetal.,2013).Thesebioprocessmodelscanprovideinsightonthe

 

individual as well as the interactive effect of the various input parameters on the target

 

output. Nevertheless, the non-linearities associated with microbial fermentations have

 

limited the use of these bioprocess models. Non-linear systems as opposed to linear systems are not standardized, which leads to deviations between the results obtained. The

 

implementation of bioprocess models that are able to efficiently encapsulate these non-

 

linearities is of paramount importance for optimization and scale-up of the bioprocess

 

(Ahmadian-Moghadam,Elegado,&Nayve,2013).

 

Traditionally,modellingandoptimizationofbioprocesseshavebeencarriedoutusingthe

 

one-variable-ata-timeapproach(OVAT),factorialDesignofExperiment(DOE)and

 

responsesurfacemethodologyRSM(Nath&Das,2011;Whiteman&GueguimKana,

 

2013).Theseapproacheshavebeenextensivelyusedandtheirconceptsaswellas

 

limitationsarewellknown.Forexample,OVATdoesnotconsidertheinteractiveeffectof

 

parametersontheprocessand,therefore,theoptimumsetpointsmaybecompletelyignored

 

(GueguimKanaetal.,2012;Wang&Lu,2005).Moreover,itisunfeasibleforthesearchto

 

accomplishanappropriateoptimuminarestrictedamountofexperimentalset-ups(Lotfy,

 

Ghanem,&El-Helow,2007).ThefactorialDOEhasbeenshowntobeunappealing,sinceit

 

istime-consuming,resourcedemandingandlabourintensivewhenthenumbersofinput

 

factorsareincreased(Wang&Wan,2009a).Ontheotherhand,theRSMdisregardsthe

 

‘lessimportant’parameterswithalimitedunderstandingoftheirpossibleinteractiveeffects

 

onthebioprocessoutput(GueguimKanaetal.,2012;Desaietal.,2008).

 

Artificialintelligencetoolshaveemergedasapromisingmethodformodellingand

 

optimizationofbioprocesses.Someoftheseincludeartificialneuralnetwork(ANN),

 

geneticalgorithm(GA),fuzzylogic,antalgorithmandparticleswarmoptimization,allof

 

whichareconsideredsuitableinthedesignofbioprocessesforresearchanddevelopment

 

(Sewsynker-Sukaietal.,2017).Inthelastdecade,ANNhasbeenappliedinmultivariate

 

non-linearbioprocessresearchanddevelopment.Theyareefficientforthedevelopmentof

 

bioprocessmodelsdevoidofpreviousinformationwithregardtothekineticsandmetabolic

 

fluxesthatoccurwithinthecellsandcellsurroundings(GueguimKanaetal.,2012).ANN

 

models simulate the linkage that exists in biological neurons with extraordinary capability for learning, analysis, association and adaptation. Some recent studies have shown that the

 

analysis using ANN yielded better results, in terms of R² values, than the linear regression

 

method. Furthermore in the area of predictive accuracy test, the ANN has a higher accuracy

 

than multiple regression analysis. Other comparative studies show that ANN has higher

 

correlationcoefficientthantheresponsesurfacemethodology(RSM),henceabettermodel

 

forpredictionoftheexperimentaldata(Sewsynker-Sukaietal.,2017).

 

                                1.1.6   OptimizationUsingArtificialNeuralNetwork

 

ArtificialNeuralNetworksareabstractcomputationalmodels,roughlybasedonthe

 

organizationalstructureofthehumanbrain.TheattractivenessofANNscomesfromtheir

 

remarkableinformationprocessingcharacteristicsthatismainlypertinenttononlinearity,

 

highparallelism,faultandnoisetolerance,aswellastheirlearningandgeneralization

 

abilities,(Strukov,Snider,Stewart,&Williams,2008).ANNconsistsofaninputlayer,one

 

ormorehiddenlayersandanoutputlayer.Theneuronsofthehiddenlayerassistthe

 

networkinestablishingthecomplexassociationsthatsubsistbetweentheinputandoutput

 

parameters.Asshownschematicallyinfigure1.1,eachcircularnoderepresentsanartificial

 

neuronandanarrowrepresentsaconnectionfromtheoutputofoneneurontotheinputof

 

another(Graves,&Schmidhuber,2009b).

 

 

Figure1.1AschematicdiagramofanArtificialNeuralNetwork

 

 

An ANN is able to learn process input and output without underlying assumption about the

 

distribution of data. They are powerful in data processing and analysis and can be used for

 

optimization of processes, which are highly complex and nonlinear. ANN can be trained to

 

solve certain problems using a learning method and sample data. In this way, identically

 

constructedANNcanbeusedtoperformdifferenttasksdependingonthetrainingreceived.

 

Withpropertraining,ANNiscapableofgeneralization,andhastheabilitytopredict

 

analysis(Graves,&Schmidhuber,2009b;Kurzweil,2012).OptimizationsusingANNsare

 

concernedwiththeminimizationofaparticularcostfunctionwithrespecttocertain

 

constraints.ANNisshowntobecapableofhighlyefficientoptimizationduetoitsabilityto

 

capturenonlinearbehaviour,andthusprovidesamodelthatlinkstheprocessinputstothe

 

correspondingoutputparameters.(Ciresan,Meier,Masci,&Schmidhuber,2012b).

 

 

 

                                1.2      STATEMENTOFPROBLEM

 

Biogasreactorsarecomplexinstallations.AlmostallbiogasreactorsinNigeriahavefailed.

 

HumanwastebiogasplantatZariaprisonwasabandonedbecauseofthereactorfailure.The

 

reactorusedwasIMHOFFtank,whichisanimprovedversionofseptictankandisreported

 

tohaveexcessivesedimentationintheinletchannels.Italsohasexcessivefoulingofsurface

 

andfloatingofsludge.Acowdungbasedbiogasplantatthefodderfarmofthenational

 

animalproductionresearchinstitute,Zaria(NAPRI)isreportedtohaveexperiencedlower

 

gasyield.Thereactorusedistheupflowanaerobicsludgeblanket(UASB).TheUASBhave

 

inherentproblemwhenoperatedunderhydrodynamicconditionsandthereispossibilityof

 

ioninhibitionfromtheun-ionisedvolatilefattyacids(UVFA’s)whichmayleadto

 

disintegrationofthesubstrategranuleleadingtothedeathofmethanogensunderhydro

 

dynamic condition resulting to incomplete mixing between the granule and block waste. The

 

utilization of POME from palm oil mills for biogas production can prevent the harmful

 

environmental impacts associated with the discharge into the ecosystem. Several studies on POME treatment have been carried out using various high rate anaerobic reactors but these

 

reactor designs have at some point experienced their own fate of failures. Such failures are

 

due to complex biochemical processes which are often poorly understood and difficult to

 

control and run. However, as an important technology, the optimization of anaerobic

 

digestionprocessescanassistitsoperationandprocesscontrolaswellasmaximizing

 

methaneproduction.

 

 

 

Traditionally,theoptimizationofexperimentalprocessesisachievedbyrepetitionsof

 

experimentswhilevaryingonefactorafteranother.Modernmethodsforoptimizationapply

 

thesimplexproceduresandoftengivegoodresults.Theuseofsimulationasanoptimization

 

toolismoreadvantageous,asvariousmodellingformatsapplyanumberofexpressions

 

whichrequiresolutionstopolynomialequations.Thesemethodsareallstochasticbecause

 

theresultsaredeterminedabinitio.However,theuseofANNforsimulationandmodelling

 

arenon-stochastic;thisisbecausetheprogramusedallowstheusertopreselectthe

 

distributionwhichcanbemodifiedbythesystemdynamicstoreachatarget.Thistargetis

 

whatthesimulatorregardsastheoptimumdistribution.Inaddition,mathematicalmodelling

 

ofanaerobicdigestioncanbeverycomplexandbecauseofthecomplexitiesinvolved,the

 

digestionprocessisoftenmodelledasablack-box.ANNmodelscanbeusedtoovercome

 

someoftheproblemsofcomprehendingtheanaerobicdigestionprocesses.Theapproachof

 

ANNisparticularlyeffectiveformodellingprocesseswheretheunderlyingrelationships

 

betweeninputandoutputdataareunknown,complexandnonlinear.Otheroptimization

 

modellingsystemslikeregressionmodelling,linearprogramming,andgenericalgorithm,

 

requiresequations(mathematicalmodelling)butANNgivesadatadrivenmodelwhichis

 

much preferred because of its high degree of accuracy that is the closeness of the simulated

 

orpredictedoutputtothetargetwhichistheexperimentaloutput.

 

 

 

In view of these considerations, this research’s concern, which has been the challenge of

 

many researchers and other relevant stakeholders is, how can we model an experimental data

 

in order to remove their complexity, contain their non-linearity and unsteadiness in the

 

anaerobicbiogasgenerationfrompalmoilmilleffluentforanoptimizedbiogasyield?

 

 

 

                                1.3      RESEARCHAIMANDOBJECTIVES

 

Thisresearchaimsatdesigningamodelfortheoptimizationoffactorsaffectinganaerobic

 

biogasgenerationfrompalmoilmilleffluent(POME)usingartificialneuralnetwork.

 

Thespecificobjectivesforwhichtheresearchisbeingcarriedoutare:

 

1. Toexaminerelationshipbetweenparameterssuchas%C,%N,CODreduction,

 

hydraulicretentiontime(HRT),pH,%S,%P,C/NRatio,temperature,organic

 

loadingrate(OLR)intermsofreactorefficiency.

 

2. TodevelopamodelbasedonArtificialNeuralNetwork(ANN)foroptimizationof

 

biogasinanintegratedbioreactor.

 

3. ComparisonofthepredictiveabilityusingANN,ofsimulatedmethaneyieldswith

 

experimentalmethaneyieldsoftheintegratedbioreactor.

 

4. Todeterminethefunctionalrelationshipbetween%CODreduction,HRTandpHon

 

methaneyieldrespectively.

 

 

 

1.4      HYPOTHESIS OF THE STUDY.

 

Thehypothesisofthestudyisthetheoreticalbasiswhichhasmotivatedanddriventhis

 

research.ANNisoneoftheworld’sfastgrowingmodernmethodsusedintheanalysisof

 

data.Itisnon-stochasticandassuchcanbeusedincomplexcomputationalmodeltopredict

 

various experimental processes which are hard to solve by various conventional computer

 

based and programming techniques. The use of ANN as an optimization tool for biogas

 

generation is relatively new in Nigeria. The theory underlying this study is to apply ANN in modelling, simulation, prediction and interpreting of experimental data in order to obtain

 

useful information and enhance decision making. This work was designed to optimize the

 

methaneyieldofanintegratedbioreactorusingNeuralNetworkToolboxofMATLAB

 

R2015a and to determine the functional relationship between the % COD reduction and

 

biogasyield.

 

Secondly,themulti-layeredfeedforwardarchitectureusedtodesigntheANNmodelis

 

decidedbytestingdifferentnumbersofhiddenneuronsinthehiddenlayer.Levenberg-

 

MarquardtBackPropagationtrainingalgorithm,trainstheneuralnetworksgeneratinga

 

modelinformofMATLABcodeforprediction.Optimizationofatargetclosetotheoutput

 

validatestheperformanceoftheneuralnetworkmodel(Graves,&Schmidhuber,2009b;

 

Hinton,Osindero,&Teh,2006;Strukovetal.,2008;Ciresanet.al2012b).

 

Thirdly,anoptimizedbiogasgenerationsystemsupportstheefforttoassessthecarbontrade

 

exchangeasoperatedintheinter-governmentalpanelonclimatechange(IPCC,2012)since

 

NigeriaisasignatorytotheKyotoprotocols.Thereforeestablishingasystemofconversion

 

ofpalmoilmilleffluenttobiogasenergyeitherforheatingorelectricitygenerationwill

 

enablethecityorindustriesinvolvedtoearncarboncreditwhichinturncanactuallyoffset

 

anyextracostinvolvedinchangingfromfossilfuelenergygenerationtoaclean

 

developmentmechanism(Kyotoworkdatapoliciesinfrastructures,2007).

 

Fourthly,energyresourcesbasesuchaspalmoilmilleffluent,municipalorganicwaste,

 

domesticsewage,essentiallywillre-educatethesectionofthepopulationtoadapttocleaner

 

environmentandalsogeneratesustainablejobs.

 

                                1.5       SIGNIFICANCEOFSTUDY

 

Thestudypresentsanopportunitythathelpstofostereffectiveandefficientenvironmental

 

wastemanagementofPOME(Ejike,Chinedu&Egbujor,2012).Thereisneedtofinda

 

bettermethodforoptimizingchemicalexperimentsefficientlythroughtheoreticalbasis.The

 

core of a control/optimization system is the model describing the process. Despite several attempts to model the processes that occur in a digester, up to now classical mathematical

 

modeling was only possible when severe simplifications of the process representation were

 

performed (Sulaiman, Nikbakht, Tabatabaei, Khatamifar, & Hassan, 2010 ; Holubar et al.,

 

2003). The main reason for this situation is that the mechanisms ruling this processes are not

 

sufficientlywellunderstoodtoformulatereliablenon-linearmathematicalmodels.Asan

 

alternative,artificialneuralnetworkshaveadistinctiveadvantageoversomeothernon-

 

linearestimationmethodsusedforbio-processes,becausetheydonotrequireanyprior

 

knowledgeaboutthestructureoftherelationshipsthatexistbetweenimportantvariables

 

(Holubar et al., 2003).

 

Optimizationinvolvesvaryingmorethantwofactorssimultaneouslytherebyminimizingthe

 

errorontheyieldwhileidentifyingvalidrelationshipsamongthevariables.Consideringthe

 

hugeinvestmentcostrequiredforthesetupofbiogasplant,coupledwiththecomplexityof

 

workandtimespanningintomonths/yearsneededforestablishment,thedeploymentof

 

ANNhasbecomeaneffectiveandefficienttooltopromotethepredictionofproportionof

 

allneededparameters(factors)foroptimalyieldortarget.

 

 

 

1.6SCOPEOFTHESTUDY

 

Thestudyinvolvesthefollowingareasofactivity:

 

1. PreliminaryStudiesonbiogasgenerationfrom

 

2. DevelopingamodelbasedonArtificialNeuralNetwork(ANN)foroptimizationof

 

biogasinanintegratedbioreactor.

 

3. Optimizingtheyieldofbiogasinanintegratedbioreactor,byapplying

 

4. Determinefunctionalrelationshipbetweenthe%CODreduction,pHandHRTon

 

Biogasyieldrespectively.

 

These were performed by the application of ANN system embedded in MATLAB

 

software R2015amodel.

 

 

 

FACTOROPTIMIZATIONFORANAEROBICBIOGASGENERATIONFROM PALMOILMILLEFFLUENTUSINGARTIFICIALNEURALNETWORK

 

 

Sharing is caring!