Document Type : Review Paper
Authors
- Jassim Mohammed J. Al Said Naji 1
- Mohammed Abdullah Ahmed 2
- ِAlyaa Mahmood Ali 1
- Hijran Mageed Hammad 3
1 University of Technology-Iraq
2 University of Technology-Iraq.
3 Ministry of Oil, Mid. Land Oil Company
Abstract
Sand production is a worldwide issue for all oil and gas wells are producing from sandstone reservoirs. Along past century many of authors and companies worked on putting suitable solutions for this situation that causing many problems as decreasing of fluid production reach to well shutdown. In present paper many published manuscripts that dealt with employment artificial intelligent approaches in predicting the onset of sand production are reviewed. The reviewed artificial intelligent approached are developed to perform one main target: the sand onset production prediction. This main target is detected by employment different ways such as artificial neural networks, generalized regression neural network, feedforward neural network, genetic algorithm, particle swarm optimization and support vector machine. Many influencing parameters on sand production initiation are used as inputs for these models likewise: total vertical depth, transit time, gas and water flowrates, formation cohesive strength, bottom-hole shut-in and flowing pressures, drawdown, critical drawdown pressure, effective overburden vertical stress, interval length, perforation density, and sand free production duration in years. The models results were in many terms including predicted critical drawdown pressure for sanding onset, or for making sand production probability in term of numbers (minus one, zero or positive one). The main conclusion from this review, the accurate artificial intelligent approach for sand onset production need to accurate and large data set as well as the results accuracy proportional to progress and development in artificial intelligent tools.
Keywords
', './data/ijogr/coversheet/cover_en.jpg'))