基于人工神经网络的稻谷干燥过程控制策略(IJITCS-V6-N11-4)

I.J.InformationTechnologyandComputerScience,2014,11,28-35PublishedOnlineOctober2014inMECS()DOI:10.5815/ijitcs.2014.11.04Copyright©2014MECSI.J.InformationTechnologyandComputerScience,2014,11,28-35ArtificialNeuralNetworkBasedControlStrategiesforPaddyDryingProcessShekharF.LilhareResearchScholar,GHRCE,Nagpur,440016,[e-mail](correspondingauthor)Dr.N.G.BawanePrincipal,S.B.JainInstituteofTechnology,Management&Research,Nagpur,441501,[e-mail]—Paddydryingprocessdependsuponambientconditions,paddyquality,temperatureandmassofhotdryingair.Existingtechniquesofpaddydryingprocessarehighlynonlinear.Inthispaper,aneuralnetworkbasedautomatedcontrollerforpaddydryingisdesigned.Thedesignedcontrollermanagesthesteamtemperatureandblowermotorspeedtoachieveconstantpaddydryingtime.ALayerrecurrentneuralnetworkisadoptedforthecontroller.Atmosphericconditionssuchastemperatureandhumidityalongwiththesizeofthepaddyareusedasinputtothenetwork.Experimentalresultsshowthatthedevelopedcontrollercanbeusedtocontrolthepaddydryingprocess.Implementationofdevelopedcontrollerwillhelpincontrollingthedryingtimeatalmostconstantvaluewhichwilldefinitelyimprovethequalityofrice.IndexTerms—Paddy,DryingTime,NeuralController,Simulation,ControlStrategiesI.INTRODUCTIONRiceisthesecondlargestproducedcerealcropintheworldandroutinelyeatenbyalmosthalfofitspopulation.Nearlyhalfofthetotalpaddyproductionisprocessedinparboilinganddryingplants.Wholegrainofriceisnaturalchoiceofthepeople.Parboilinganddryingprocessisappliedtohardenthericekernelinordertoenhanceheadriceyieldinmilling[1].Thewellknownreasonforpaddyparboilinganddryingistogethighriceyield,highernutritionalvalueandresistancetospoilagebyinsectsandmould[2,3].Paddydryingisahighlyenergyconsumingprocess.Paddydryersaredividedintotwobroadcategories,unheatedairdryerandheatedairdryer.Naturalunheatedairdryersarenotusedincommercialpaddydrying.Heatedairdryerssuchasre-circulatorybatchdryer,baffeldryer,rotarydryerandLouisianastateuniversity(LSU)dryerarecommerciallyusedforparboiledpaddydrying.LSUdryerismostpreferredbypaddyprocessorsincentralpartofIndiabecauseofweatherconditionsandavailabilityofskilledoperators.Ambienttemperatureandrelativehumidityhavegreateffectondryingprocess[4,5].Knowledgeofambienttemperatureandrelativehumidityisneededtosupplyrequiredmass&temperatureofdryingairforefficientdryingperformance.Thevariabletemperaturedryingusingfuzzycontrollerwithfeaturesofnon-linear,timevariabilityandothervariablesduringpaddydryingprocess,improvetheproductionefficiencyofdryerandreducedryingcostofhighmoisturecontentpaddy[6].Theworkin[7,8]proposeanadvancedcontrolstrategyusedtodesigntemperature,relativehumidityandairvelocitynonlineartrackingcontrollers,inordertoovercomeitshighlynon-lineardynamicsoverwholedryingoperatingconditions.Inonlinequantitativefeedbacktheory(QFT)basedself-tuningcontroller[9]foraconveyorbelttypegraindryerplantamathematicalmodelofthegraindryerplantwasobtainedusingsystemidentificationandreal-timeinput/outputdata.QFTtechniquepotentiallyhasexcellentsolutiontowardsthisproblemduetoitswellknowncapabilitytoachieverobustperformanceregardlessparametervariationsuchaslongdelaytime,presenceofdisturbancesandplantuncertainty.Anautomaticmultiparameterprocessingsystemimprovethetiresomeprocedureforgraindryinginadeepfixed-beddryer[10].Dryingexperiments[11]conductedtoassesstheeffectsofinletdryingairtemperature,feedmassflowrate,onenergyefficiency,entropygeneration,improvementpotentialrate,andsustainabilityindex.Studyconductedbyresearchers[12,13,14]showthatthesurfacedryingprocesscanbecontrolledbythreevariablestemperature,humidityrate,anddirectionofairflowratefromexternalsource.Valuesofthesevariablesshouldbeoptimizedtoobtainmaximumefficiencyanddryingperformance.Theoreticalmodelpredictsthatpaddydryingtimeisbasedoninitialgrainmoistureandrelativehumidity[15].Thepaddydryingat35°Ctemperaturewithtypicaldryingconditions,takelessthan5hoursfortemperingandhighertemperatureshortenedthetemperingtime[16].Therelativehumiditydoaffectstemperingtime.Speedofdryingdependsontemperatureandhumidityofdryingair[17].Dryingrequirementsofagriculturalandfoodproductsvaryfromproducttoproduct.Eachproducthasitsspecifieddryingconditionforbestquality[4].Inheatedairdryingoperation,dryingairtemperatureandrelativehumiditycanbecontrolledbyblowermotorspeedandsteamtemperature.Experimentsconductedonroughricedryingatdifferentairvelocity&temperatureArtificialNeuralNetworkBasedControlStrategiesforPaddyDryingProcess29Copyright©2014MECSI.J.InformationTechnologyandComputerScience,2014,11,28-35showthatthereislesseffectofairvelocitywhenriceisdriedatlowertemperature,buthavesignificanteffectofairvelocitywhendriedathighertemperature[18].Itisalsoseenthattheeffectoftemperatureismoreimportantthanairvelocityeffect,howeverifbothairparametersareconsideredtogether,thedryingeffectishigher.Theworkin[19]presentstheinfluenceofdryingtemperatureonricetastequality.Dimensionsareimportantfactorofpaddytobestudiedforquantitativeanalysiso