Implementation of Artificial Neural Networks to Determine Enhanced Oil Recovery Method
- 作者：Anas Puji SantosoEdo Pratama
- 会议时间：2013-09-16
- 关键词：Enhanced Oil Recovery ; Artificial Neural Networks ; Backpropagation
- 作者单位：Petroleum Engineering Department, UPN "Veteran" Yogyakarta Jl.SWK 104 (Lingkar Utara) Condongcatur, Yogyakarta 55283, Indonesia
- 母体文献：2013油气田监测与管理联合大会论文集
- 会议名称：2013油气田监测与管理联合大会
- 会议地点：西安
- 主办单位：中国石油学会
- 语种：chi
摘要
It is hard to make a decision on early stage field development planning with limited reservoir data.To planning Enhanced Oil Recovery (EOR) needs long time complicated research and very expensive.In ordinary, we need reservoir simulation tool which used to predict reservoir behavior so that can be the hint for field development.Reservoir simulation need complete reservoir data and spend much time to get the good result.Usually, field development planning strategy made in the early stage of field exploitation.It is related with how to arrange the Plan of Development (POD) in a field includes the budget to develop that field.A tool is needed to help in making decision with limited reservoir information.By using the tool, it is possible to choose EOR method which will be applied to the particular field with limited reservoir data.We can applicate Artificial Neural Networks (ANN) theory to make a tool which can use to choose EOR method which compatible based on applicability level.Screening criteria that is used based on reservoir data that applied EOR in different condition.The data that needed are key reservoir properties parameter, such as: oil gravity, oil viscosity, oil composition, oil saturation, formation type, net thickness, average permeability, depth, and temperature.Backpropagation learning method use to learned the learning data.ANN model capable to processing different data, eliminate arbitrary approach to making decisions, and capable to fast investigate and high accuracy.In this research resulting a software which is named PiXel.