The Prediction Of Cementing Quality Based On Fuzzy Neural Network

The main purpose of cementing is to seal the annulus space, prevent the flow of formation fluid, and achieve the separation of different formations. The factors that affect cementing quality are numerous and miscellaneous, especially in the deep wells, ultra deep wells, and HPHT wells. In order to gain high quality cementing quality and separation of different formations, a full-scale, systematic, and clear understanding of the factors that affect cementing quality is needed. By the analysis and study of various factors, a method fuzzifying the influence of the factors was proposed, and a prediction system was established accordingly. In the system, a network, based on the BP network, using the fuzzified parameters as input data, was achieved by the adjustment and calculation of network weights. The application interface is completed through modular design, so it not only embodies the advantages of fuzzy neutral network but also ensure the stability and improvement.