| As the environment of offshore oil & gas platform is very crucial and there are a large number of flammable and explosive substances in the platform which is distributed with intensive operation equipment,the risk of fire explosion accident is high.It is found b y research that human error is an important cause of the accident.In order to reduce the risk of fire explosion accidents,the paper,supported by Natural Science Foundation of Shandong Province and the Fundamental Research Funds for the Central Universit ies,takes offshore oil drilling production platform work as the research object and studies human reliability analysis in fire and explosion accident of offshore oil drilling production platform.The details are as follows:(1)Based on the existing theory for human reliability analysis,as well as combined with the definition and classification of human error in offshore,the types of human error in offshore are analyzed from the perspectives of individuals and organizations.Combining the two classification criterion of Performance Influencing Factors(PIF)defined by the Safety and Health Executive(HSE)and Cognitive Reliability and Error Analysis Method(CREAM),the behavior form factor classification scheme of PIF for offshore oil drilling production platform work is proposed purposefully.(2)For the most frequently occurring fire explosion accident caused by oil and gas leakage and blowout,the paper analyzes the root cause of accident and predict possible accident consequences using the Bowtie model.According to the oil and gas leakage fire and blowout fire emergency process,the paper identifies the possibility of human error step by step and draw the human error table of emergency operations systematically.(3)According to the 173 major fire and explosion accidents reports occurring during 1984-2015,28 accident investigation reports were selected as the samples to extract human reliability data and make preliminary statistical analysis of human-equation errors.In order to overcome limitation in the numbers of accident investigation reports,three methods were applied comprehensively to expand and collect personnel data for offshore drilling platform operations,including accident investigation reports,statistical surveys and expert judgments.On the basis of marine drilling operations formation,the five-level questionnaires were designed with 26 criteria.The paper interpreted PIF relevance based on theoretical model of information,decision,and action in crew context(IDAC)in team situation and built relevance model between PIF and Personnel Behavior in marine drilling.Structural equation model was built with SPSS and AMOS.And the above model has been simulated and corrected enough times to verify the relevance between PIF and personnel behavior.(4)Amended CREAM is proposed to break through the limited use in nuclear industry.The new relation model between PIF score and control modes based on the BP neural network was supposed to cover the disadvantage of artificial specification and raise the accuracy.According to the designed calculation steps,human error probability of daily drilling and production operation is calculated quantitatively.Finally,an example of the application of this method was given.The results show that the quantitative analysis method is practical.(5)Using the behavior observation BBS tool,the field data of the unsafe behavior of the welding process during the period from April 2015 to March 2016 are collected by observing the behavior of the workers.This provides a data base for the prediction of human error frequency.Then,the gray GM(1,1)model,the gray Markov GM(1,1)model and the Markov forecasting model were used to predict the unsafe behavior frequency during the welding operation,and the prediction results are compared and analyzed. |
PDF Full Text Request