مكان يابي نقاط بهينه حفاري در ميادين نفت خيز جنوب به روشهاي هوشمند

چكيده انگليسي :

Optimizing Well Placement Procedurein South Oil Reservoirs Using Artificial Intelligence Approaches Maryam Jannesari maryam5394@gmail com 2012 02 13 Department of mining engineering Isfahan University of Technology Isfahan 84156 83111 IranDegree M Sc Language Farsi1 Nader FathianPour fathian@cc iut ac ir2 AhmadReza Mokhtari ar mokhtari@cc iut ac irAbstract The optimization process is considered as maximizing or minimizing a predefined objective functionusing a structured algorithm under predefined limitations In optimizing well placement procedure thisfunction is defined in exploration and exploitation stages separately and needs to include several effectivefactors such as geological petrophysical and economical parameters simultaneously Due to large quantity ofinvolved parameters and the uncertainty associated with some of them application of intelligent optimizationmethods such as evolutionary algorithms is inevitable Optimizing the placement of new wells in an oil fieldis essential if productivity is to be maximized The computational demands for this problem are substantial as many function evolutions are required and each entails a full reservoir simulation though surrogatemodels can be used in some cases It is therefore essential that underlying optimization algorithm beefficient and robust Used approach in this thesis is Particle Swarm Optimization PSO which is one of thementioned algorithms This algorithm is a relatively new approach for global optimization The algorithmattempts to mimic social interactions exhibited in animal groups e g flocks of birds in flying Like GA PSOconsists of population of solutions here referred to as particles rather than individuals In our problem theparticles are blocks of oil field In this thesis the exploratory objective function was defined as the multiplication of the 3D reservoirPorosity Estimation variance and Permeability PEPr in one of the south oil field reservoirs The ParticleSwarm Optimization PSO algorithm was then applied on the defined objective function throughout thedefined search space which was specified by the extent of 3D Kriging estimations The optimum welllocations given by PSO algorithm for the first three priorities were cross validated through analyzing theirPEPr function values The results show that obtained optimum value for objective function 71 2955 is inmaximum range of objective function values in the studied area The high facility of PSO and its ability tofind extermum targets of objective function in well placement problem approved in this chapter are anotherquality of proposed method In Exploitation and producing stage of Oil reservoirs the Net Present Value NPV was defined as theappropriate objective function for the available data from other south oil field reservoirs The Particle SwarmOptimization PSO algorithm was then applied on the predefined objective function throughout the searchspace encompassing all possible locations as the potential oil wells The optimum well placements given byPSO algorithm for six priorities were cross validated through analyzing their NPV function values Theresults show that the obtained mean value for objective function for all six proposed well locations is431 850 which is placed among maximum of the NPV values in the range of 280 8 to 438 41 Finally thesensitivity of the proposed well locations as a function of the production rate was assessed and the resultswere found to be very consistent and stable if the increment or decrement of overall production is distributedevenly among all production wells Key words Well placement Optimization evolutionary algorithms Particle SwarmOptimization Net Present Value

جان نثاري، مريم

مكان يابي نقاط بهينه حفاري در ميادين نفت خيز جنوب به روشهاي هوشمند

الگوريتم هاي تكاملي , الگوريتم بهينه سازي ازدحام ذرات , ارزش خالص فعلي