Study On Risk Analysis And Accident Prediction Of Oil And Gas Exploration Industry

Oil & gas exploration industry is a key part of the oil industry.Due to the complex risk factors existing among exploration operations,the oil & gas exploration accidents may cause significant injuries,fatalities and economic losses.However,there is little systematical risk analysis method and accident prediction model available related to the underlying factors contributing to the oil & gas accidents and correlations under which these occur.In order to reduce the risk level and prevent accident of oil & gas exploration,this study proposes a risk assessment and accident prediction method.This study depends on the Bureau of Geophysical Prospecting INC.,China National Petroleum Corporation.The specific research contents are as follows:1.Literature review of the risk assessment and accident prediction methods.Based on the literature review of the accident causation theory,typical risk assessment and accident prediction method,the basic definition,operation characteristics and basic process of risk analysis of oil and gas exploration operations can be identified.The risk assessment of oil and gas exploration operations is analyzed from the perspective of Hall three-dimensional structure.These researches provide a theoretical basis for the risk assessment and accident prediction methods for oil and gas exploration.2.Risk identification of oil & gas exploration.Considering the characteristics of oil and gas exploration operations,the Hierarchical Holographic Modeling(HHM)method is used to analyze the risk of oil and gas exploration from multiple perspectives and dimensions.Based on the industrial accident data among 1970-2017,the construction method of the accident database is proposed,and the historical data is analyzed statistically.Based on the improved grey correlation analysis method,the key contributing factors are identified.3.Dynamic risk assessment of oil and gas exploration based on Bayesian network.The risk of oil and gas exploration can change dynamically.This study proposes a dynamic risk assessment method of oil and gas exploration based on Bayesian Network(BN).The backward analysis and sensitivity analysis method is used to identify the key causes of the accidents,and calculate the time-variant occurrence probability.This method can reduce the level of risk and control the high-risk factors.4.Accident prediction model based on wavelet neural network.The oil and gas exploration accident data are random.Combined with the wavelet analysis theory and the BP neural network,this study constructs a wavelet neural network(WNN)accident prediction model for oil and gas exploration.The training data and test data are used respectively to validate the WNN and BP model.The results are comparatively analyzed.This model will be helpful to predict likely accidents under different working conditions.The results are primarily useful for accident prediction of oil and gas exploration in field conditions,in addition for relative precaution making.