标题：A NOVEL METHOD FOR REAL-TIME QUANTITATIVE EVALUATION OF DRILLING RISK BASED ON BP NEURAL NETWORK AND MONTE CARLO SIMULATION USING IN OIL& GAS DRILLING ENGINEERING
摘要：Petroleum drilling is a high-input,high-risk and high-tech concealed underground project.There are a large number of complex,random and uncertain factors,which can easily lead to a variety of risks such as complexity and accidents,which seriously affects the economic benefits and environmental protection of the system.Drilling site monitoring parameters provide a wealth of construction information and are important data for timely diagnosis of drilling risks.The essence of drilling risk assessment is to study the internal laws between complex downhole accidents and geological parameters and monitoring parameters,and describe them with a functional relationship.Comprehensively considering the influencing factors of the drilling process and the monitoring parameter selection principles such as "high sensitivity,good stability,strong correspondence,small calculation amount and easy direct acquisition",9 logging parameters such as mud outlet flow rate,hook load,pump pressure,turntable speed,turntable torque,weight on bit,mud volume,mud density,and drilling speed were established as identification variables.Five kinds of drilling conditions,such as normal,lost circulation,overflow,stuck pipe and drilling tool accident,were determined.Finally,a real-time drilling risk assessment method based on neural network and Monte Carlo simulation was established.This method can analyze the downhole working conditions by using on-site engineering parameters and logging parameters,identify the drilling risk types,and consider the uncertainty of monitoring parameters and identification models,and calculate the risk probability of the corresponding downhole complex accidents.The example analysis shows that the results of the drilling risk monitoring and evaluation calculated by the theoretical method are basically consistent with the actual engineering,and can meet the requirements of drilling risk monitoring and evaluation.It can provide reliable technical support for safe and efficient drilling and well site environmental protection.
关键词：Drilling engineering;environmental protection;risk assessment;real-time monitoring;neural network;Monte Carlo simulation
