مدلسازي فرايند تزريق بخار در يكي از مخازن كربناته شكافدار نفت سنگين

چكيده انگليسي :

Modeling of Steam Injection in a Fractured Heavy Oil Carbonate Reservoirs Shirin Mohammadi Date of Submission February 13 2012 S mohammadi@ce iut ac ir Department of Chemical Engineering Isfahan University of Technology Isfahan 84156 83111 IranDegree M Sc Language FarsiSupervisor Mohammad Reza Ehsani mrehsani@cc iut ac irAbstractIncrease of oil consumption and price of the crude oil in the world and declining oil supplies in recent years hascaused an increasing attention to heavy oil and extra heavy oil production methods As far as substantial parts ofsuch reservoirs are fractured and fractured reservoirs are not easily exploited by using conventional technologiesand requires more energy time and high cost so more recognition of oil recovery methods from fracturedreservoirs and finding details of each methods especially the correct understanding of effective oil productionmechanisms are effective in selection of suitable and economical methods So in this study three dimensionalmodeling of steam injection in a fractured heavy oil carbonate reservoir was performed First steam injectionprocess was modeled in a non fractured reservoir then the model accuracy was assessed This model showedgood results compared with experimental results model Willman and Shutler This model considered bothcapillary and gravity effects in the energy equation and a variable porosity The results showed that steaminjection in a non fractured reservoir increases the oil recovery to 70 In the next step steam injection processwas modeled in a fractured reservoir The amount of oil recovery in this reservoir was estimated about 30 Thesensitivity analysis for identification and evaluation of steam injection process in both non fractured andfractured reservoirs was performed The results indicated that some parameters such as injection pressure steamquality pre heating permeability and location of fracture affect on the oil recovery Oil recovery from carbonatefractured reservoirs is less compared with non fractured reservoirs and breakthrough occurs sooner Steaminjection pre heating and cyclic steam injection into Kuh e mond reservoir has been studied too The resultsshowed that pre heating and cyclic steam injection increase oil recovery Finally our modeling results werecompared with the results of the CMG simulator The amount of oil recovery obtained from our modeling has anerror of 1 67 compared with the CMG result in a horizontal fractured reservoir In the next part steamdistillation mechanism during steam injection was studied and the yield of steam distillation as an effective oilproduction mechanism was predicted by the artificial neural network model Multi Layer Perceptron was chosenas our neural network model Different one hidden layer and two hidden layer networks built by our neuralnetwork model and examined by input parameters We used Supervised training algorithm so the desiredresponses steam distillation yield entered as outputs of models After building the main structure of neuralnetworks step of training and testing begun finally in cross validation step the simulation results compared withexperimental results In this comparison the numbers of neurons in hidden layers changed for example four typesof One hidden layer and two hidden layer neural networks by different neurons number 3 4 5 6 built and wereinvestigated The results showed one hidden layer neural network had less average relative error ARE compared with some previous models The average relative error of this model was obtained about 2 KeywordsEnhanced oil recovery steam injection modeling fractured reservoir neural network multi Layer perceptron steam distillation mechanism

محمدي، شيرين

مدلسازي فرايند تزريق بخار در يكي از مخازن كربناته شكافدار نفت سنگين

ازدياد برداشت , مدلسازي , شبكه عصبي , مكانيسم تقطير بخار