The Study Of Drilling Fault Diagnosis Method Base On Neural Network Fuse Technology

Oil is an important strategic resources. Drilling is the basis for the oil industry.The complex, and hidden stratum makes the drilling process prone to many complications and even cause serious accidents.The accidents will affect the well bore quality, drilling speed, spend a lot of manpower, material and money, time, and even threaten the lives of the staff. Therefore, the diagnostic method research of drilling accident is significant.In this paper, common drilling accidents and their signs are researched. And how to get these signs by the drilling instruments or a compound logging tool is researched too. Based on the above, the complex and ambiguous characteristics of drilling accident signs are found out. Then,a drilling accident diagnosis method based on a multiple neural network model,which will be used to fault signs fusion, is identified. And a set of intelligent diagnosis system for drilling accident based on expert systems and neural networks to work together, is developed.Common mechanisms of drilling accidents are researched and the common signs of drilling accident are extract in this thesis. Meanwhile, a drilling accident knowledge base is established. An accident diagnosis method based on multiple neural network fusion different accident signs are studied.At last a drilling accident intelligent diagnosis system is developed by using VB, Access and matlab. The system’s accuracy can achieve94.2%, in a laboratory simulation.