Study Of Drilling Accident Pre-warning System Based On Multi-source Information

This paper takes drilling process as research object, and it proposes a drilling accident pre-warning method based on multi-source information. First of all, by combining domestic and international status quo, starting from multi-source information source and drilling accident mechanism, we sum up all kinds of accidents related information source. Then study the multi-source information fusion technology, such as neural networks, D-S evidence reasoning, support vector machine, fuzzy reasoning, etc. Select the appropriate warning method according to the characteristics of each type of accident, for example, choose fuzzy reasoning method to diagnose the failure of shock absorber, SVM technology to diagnose the failure of drillpipe, and demonstrate the reliability of these diagnostic methods. Finally, develop a set of drilling early warning system based on multi-source information, establish the before drilling database, drilling database, knowledge database, drilling accident database. Studying the drilling accident system based on multi-source information can be beneficial to improve the on-site staffs level of technology and forcast the engineering abnormal condition, protect drilling safety, reduce drilling cost, and improve the overall effectiveness of exploration and development.