Performance-Oriented Drilling Fluids Design System With A Neural Network Approach

摘要

Drilling fluids play a key role in the minimization of well bore problems when drilling oil or gas wells, usually the design of drilling fluids is depended on many experiments with experience. Rule-based and case-based reasoning drilling fluid system was designed with theory of expert system by some researchers. But it is very difficult to get to know and express precious relationship between drilling fluid formulation and its performance. Performance of drilling fluids can be measured with test device when drilling fluid is ready. A performance oriented drilling fluids design system is presented, with supervised artificial neural network algorithm to acquire knowledge by learning from experimental data. The system can be used to design drilling fluid according to specified performance. Experimental results show that drilling fluids designed by the system can satisfy specified performance.