Research On The Intelligent Warning System For Drilling Accidents Based On SVM And PSO

Oil drilling is a complex underground engineering influenced by many vague, random anddynamic uncertain factors. Accidents might occur at any time in the process of drilling, whichseriously threaten the safety of drilling; therefore, the need of real-time monitoring of the stateof drilling and accurately warning drilling accidents will be of great urgency for the drilling.The current methods of predicting the drilling accidents include artificial judgment, expertsystem and neural network, etc. however, their prediction accuracy and adaption of thecomplex drilling conditions are deficient. Consequently, this paper proposes the research onthe intelligent warning system for drilling accidents based on PSO-SVM, which will be ofgreat significance for improving the drilling quality and guaranteeing the safety of drilling.Firstly, this paper analyzes the present situation of drilling accidents warning, introducesthe related basic theory of drilling accidents and analyzes the principle of support vectormachine (SVM) and particle swarm optimization (PSO) algorithm in detail. Secondly, themodel of the intelligent warning system for drilling accidents based on PSO-SVM has beenstudied and its concrete implementation steps of the core algorithm were given. Throughanalysis of system requirements, this paper have detailed designed the system’s architecture,each function module and its database structure. Finally, the system has been implemented byC#language and verified using the actual instances of accidents.This paper focuses on the construction process of the model of intelligent warning systemfor drilling accidents based on PSO-SVM, this model uses PSO algorithm to optimize SVMparameters to solve the problem that the SVM parameters are not easy to choose, which willimprove greatly the accuracy of drilling accidents prediction, provide drilling technologypersonnel with decision support in prevention and control of drilling accidents.