DEVELOPMENT OF GENETIC ALGORITHM WITH OBJECTIVE FUNCTIONS BASED ON RESERVOIR INPUT DATA FOR DETERMINATION OF MULTI-VERTICAL WELL IN MULTI-LAYER RESERVOIR ON AN COMMINGLED AND SELECTED

• 主要作者: HARYADI (NIM : 22210015), FEBI
• 格式: Theses
• 語言: Indonesia
• 在線閱讀: https://digilib.itb.ac.id/gdl/view/20099
• 機構: Institut Teknologi Bandung

The process of determining the well location for both new and mature field is one important factor that determines the amount of oil recovery. The process has been conducted by the conventional trial and error method that is time consuming and requires a lot of work, especially for large oil or gas field. Hence, a new robust method is proposed by modify Genetic Algorithm (GA) to solve this optimization problem. <br /> <br /> <br /> The Genetic Algorithm will be applied to optimize well location candidates by evaluating a proposed fitness function (objective function) to maximize the production of wells penetrating multilayer reservoir. By employing basic reservoir properties of each layer obtained from a reservoir simulation model, i.e., permeability, porosity, oil saturation, pressure of reservoir, and grid thickness as the GA's objective function parameters, the optimum coordinates of new wells are determined. <br /> <br /> <br /> A model of oil field has been applied to validate the proposed method. The result of GA are the locations of wells in two-dimensional coordinates (x,y). The wells are vertical wells which is perforated in commingled scenario. There are two perforation scenarios, i.e. perforation in all productive layers and perforation in the layer that meet certain constraints only. Furthemore, by using a reservoir simulator, the results of GA will be compared with the result of conventional method. By using the tested field model, GA gives better results than conventional method. GA gave recovery factor 3% greater than conventional method.