On the Technology of Fuzzy Neural Network Identifying Complex Reservoir

Reservoir identification plays an important role in the oil and gas exploration, especially in the complex reservoir analysis. The complex geological environment gives rise to the complexity of the reservoir, which is taken shape as a result of multiple geological factors, thus making the underground oil and gas reservoirs possess a strong nonlinearity and fuzzy features. This paper is based on the fuzzy neural network theoretically, and takes advantage of the fuzzy neural network technology to identify complex reservoir. In accordance with different complex reservoir and their features, this technology makes fully use of well-logging, well-drilling, core stone and test to choose different parameter for different complex reservoir as input parameter. In the light of different input parameter and the number of middle layer neurons, this skill uses genetic algorithm to optimize initial parameter, and takes BP algorithm to improve convergence speed. This technology applies to carbonate reservoirs in Tarim, carbonate reservoirs in Long Gang, Sichuan, Karamay volcanic reservoir and water-sensitive reservoir, its coincidence rate reaches 83.62%, achieving fruitful applied outcomes.