Abstract:

To reduce human factor errors in traditional expert scoring methods, based on analysis of maintenance risk characteristics, a risk assessment model based on the correlation between unsafe information and hazards was proposed. The average value of the number of unsafe events and the number of non-conformities (violations and other deviations from requirements that have not yet constituted unsafe events) correlated with each hazard within a certain period of time was taken as the risk possibility. If multiple hazards jointly lead to an unsafe event or non-conformity, the number of associations for each hazard is added once. The proportion of occurrence times of each consequence in the total occurrence times of all consequences was taken as the proportion of the severity of such consequence in the total severity of consequences that may be caused by the hazard. The total consequence severity that the hazard may cause was determined by the sum of the product of each consequence's severity and its corresponding proportion. The model was used to analyze a large maintenance organization and statistical analysis was made on nearly twenty thousand unsafe incidents and non-conformance records in the past three years. The results indicate that the evaluation results are consistent with the distribution of hazards found during the inspection of the local regulatory authority in the same period, and the risk assessment model based on the correlation between unsafe information and hazards can objectively quantify the risk level.

Key words: unsafe information, hazard, maintenance organization, risk assessment model, possibility of risk, severity of consequence