Intelligent Supervision Systems for Improving the Industrial Production Performance in Oil Wells

An Intelligent Supervision Scheme for the Industrial Production is presented in this work. Such scheme is tested for gas lift (GL) oil wells. The proposal is based on the possible production assessment, the process variables (specifically, the bottom-well pressures), and the operational scenarios detection for the process (in the case of study, as an oil producing well), with the objective of optimizing the producing performance of the well. The proposal combines intelligent techniques (Genetic Algorithms, Fuzzy Classification, Neo-Fuzzy systems) and Energy Mass Balance. The scheme in this specific study allows establishing the oil or gas flow that a well can produce, taking into account the completion geometry and the reservoir potential, as well as the financial criteria related to the well’s performance curves and the commercialization cost of the oil and gas. The possibility of estimating bottom-well variables gives it a great operational significance to the presented approach; due to installation costs and bottom-well technology maintenance are very high, turning out to be unprofitable to produce the well. Key-words: nodal analysis, supervision system, gas lift wells, automation, fuzzy logic, neo-fuzzy systems, evolutionary computation.