Human Reliability Analysis And Evaluation Method For Manned Submersible Diving And Operation

With the implementation of China’s maritime power strategy,moving towards the deep ocean and the open ocean has become an inevitable demand for expanding national resource reserves and national strategic space.Manned submersibles are important tools for carrying people and equipment into the deep sea for exploration,sampling,and development.China’s manned submersibles have made a series of important breakthroughs and are challenging to a greater depth.It is conducive for the completion of marine exploration tasks to improve equipment performance and enhance the effectiveness of personnel.In the statistics of ship accidents,the accident rate caused by human error is very high.Reducing the errors caused by human factors has an important role in improving the efficiency of personnel operations.The diving and operation tasks of the manned submersible are high-intensity,high-precision,and long-term work under the conditions of complex environments and multiple factors.The harsh deep-sea environment makes oceanauts need to finish a series of important operations in a closed,narrow,high-pressure,dark,unstable temperature and humidity cabin,which puts high demands on the safe operation ability of the oceanauts.In this paper,human reliability analysis(HRA)is performed on the submersible of the manned submersible and the operation process to evaluate the safety of the task and carry out targeted optimization design.The main research contents of the paper are as follows:(1)This paper studied the process of analyzing and evaluating complex tasks using cognitive reliability and error analysis methods(CREAM)and Bayesian network(BN).CREAM pays attention to the context of the situation.It can logically analyze the error factors of the task process and obtain the initial error probability using a complete set of error qualitative and quantitative analysis methods.BN is a graphical mathematical model based on probabilistic reasoning.It can enrich the network by updating the information of some variables as posterior information,and obtain the probability of other nodes after the update.BN can effectively deal with the uncertainty between various causes,reasonably infer the task using statistical theory,and has the ability to identify risk incentives.(2)This paper constructed a human reliability analysis and evaluation method based on CREAM and BN.This method is suitable for analyzing the complex tasks of the severe,complex and volatile environment.We used CREAM to analyze the cognitive behavior of the task,evaluate behavior factors,and obtain the prior probabilities of the task nodes while the statistics are insufficient.We used BN to build the task behavior network,fill the conditional probability of each task node,and update the network to obtain the posterior probabilities.According to the probability ranking,we identified the most likely human error factors and make targeted design improvements.(3)We used the constructed method system to analyze and evaluate human errors and risk factors in the diving and operation of manned submersibles.We found the most error-prone operations and human factors that affect the occurrence of errors,and identified risk factors.(4)We carried out targeted human-machine optimization design based on the identification of risk factors to improve the ergonomics and safety performance of the design objects.