Fuzzy inference system application for oil-water flow patterns identification

Prediction of oil-water two-phase flow pattern provides an effective solution for reducing oil production costs. In this research, the fuzzy inference system (FIS) is utilized to predict fluid patterns and establish a new adaptable prediction model. This paper takes No. 10 industrial white tap water as research objects simulate fluids, analyzes changes pipeline angle, total convective cut. A data set containing 60 samples was used create model, Mamdani model established using MATLAB software. The results show that compared with BP neural network algorithm, forth in present has higher accuracy reliability, can achieve real-time monitoring effectively reduce errors, especially case decision-making. addition, demonstrated entire logging process non-vertical wells, use greatly save costs while ensuring safe operation equipment.