Application of Artificial Neural Network on Prediction of Drilling Rate in Deep Well

摘要

Artificial Neural Network has high nonlinearity and fault-tolerant quality. It can simulate the brain of human being to solve the complex problems of drilling. In this paper, a model of Back-Propagation Neural Network is constructed and corresponding software is programmed to analyze the main factors that influence the drilling speed in deep well and to predict the drilling speed. The on-site application of the software in two wells tested shows that the average penetration rate of the third spuding in is 20.56% higher than the well drilled in the year 1997 and the average rate of round trip is 21.24% higher. The degree of accuracy predicted of the drilling rate is above 90%.