DEVELOPMENT OF AN ARTIFICIAL NEURAL NETWORK FOR HYDRAULICALLY FRACTURED HORIZONTAL WELLS IN TIGHT GAS SANDS

Open Access
Author:
Kulga, Ihsan Burak
Graduate Program:
Petroleum and Mineral Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
April 19, 2010
Committee Members:
  • Turgay Ertekin, Thesis Advisor
Keywords:
  • ARTIFICIAL NEURAL NETWORK
  • HYDRAULIC FRACTURING
  • HORIZONTAL WELLS
  • TIGHT GAS SAND RESERVOIRS
Abstract:
Increasing demand on fossil fuels and the decline in their production promote producing hydrocarbon from unconventional sources. Natural gas existing in tight reservoirs has a huge potential for the oil industry with the improvement of drilling and production technology. Hydraulically fractured horizontal wells are a proven technology to produce oil and gas from hese tight gas reservoirs. Hydraulic fractures increase the surface area in contact with production zones which means an increased well productivity. They also decrease well drawdown and create highly conductive paths to the wellbore. Arti cial neural networks (ANN) are widely used technology in science and engineering and can be applicable anywhere that there are problems of forecast, classi cation or control. There are many studies utilizing ANN involving topics on hydrocarbon recovery. The present study aims to develop an ANN tool for hydraulically fractured horizontal wells in tight gas sands. The ANN model predicts monthly production data with an error of less than 10%. The principal advantage of proposed ANN model is that it can accommodate several production regimes for any given data sets in the range of reservoir parameters for this ANN tool.