Human Reliability Analysis Of Ship Collision Avoidance Based On CREAM

In order to ensure the safety of navigation and keepmarine economic corridor open,the researchers have been devoted to the study of ship collision avoidance.Researches have been shown that 90% of the ship collision accidents are directly or indirectly related to human factors.Therefore,it is very important to study the human reliability of ship collision avoidance.Considering that the study of human factors starts late in maritime field and lacks specialized theories and methods,the related research is carried out with the help of theories and methods of human reliability analysis in nuclear power and aviation fields.Firstly,on the basis of Swain&Guttmann human error model,behavior-forming factors are decided using the way of human factor classification in human factors guide of IMO.According to the CREAM theory,the information processing is divided into four stages: observation,interpretation,planning and execution.And then human error model of ship collision avoidance is built to reveal the mechanism of human error in the process of collision avoidance and identify human error and its influencing factors.Secondly,according to the identification result of human error caused by ship collision avoidance,the correlation between human error and ship collision accident is analyzed using the method of grey correlation analysis.The results show that the four types of human errors are closely correlated to ship collisions,such as "improper observation," improper ship operation,"decision making error" and "non compliance rules".And the main influencing factors of those four types of human errors are identified by combining expert judgment and membership function,fuzzy subset and fuzzy statistical test.In order to find out the root cause of human error,the root cause analysis of human error of ship collision avoidance is propoesd based on CREAM.Thirdly,in order to predict the reliability of ship collision avoidance,the improved common method and expanding method of CREAM are presented respectively.The correctness and feasibility of the two methods are verified in the application.When improving the common CREAM,it not only improves the CPC factor influence rule,but also takes into account the different influence of 9 kinds of CPC factors on the probability of human error.The viewpoint of using DEMATEL method to obtain the weight of CPC factor is put forward.In order to obtain more accurate prediction results,the CPC factor level evaluation of expanding CREAM is also improved using fuzzy theory and membership function.A new method to obtain the weight factors of CPC for 4 kinds of cognitive functions is obtained.Finally,the influencing factors of human error is classified into individual factors,ships and equipment,external environment,organization and management.The correlation between those factors and ship collision accidents is analyzed.According to the results,the measures to improve the human reliability of ship collision avoidance are put forward.