Research On Probability Of Human Error In Civil Aviation Maintenance Based On Improved SLIM

The civil aviation maintenance is conducted to ensure the safety of aircraft operations.Nowadays, the hardware has been more and more reliable with the increasingly development of civil aviation technical, which leading the portion of accidents caused by human factors growing rapidly. So in order to assure operation efficiently and safety, it is necessary to analyze the human factors of aviation maintenance qualitatively and quantitatively.In the qualitative analysis, the HFACS-ME method is used to analyze the factors influencing human errors, because the traditional human factor analysis model could not classify the human factors of aviation maintenance accurately in detail. According to the HFACS-ME indicator system, the indicator factors are modified through expert interviews to meet with the practical situation of aircraft maintenance. Based on the factors, the aviation maintenance human error indicator system is established. In the quantitative analysis, the Success Likelihood Index Method(SLIM)is chosen from several human reliability analysis methods to forecast the human error probability, due to the characters of multi-tasking units and multi-workers it fits in aviation maintenance. However, as first generation method SLIM is flawed, for reason of the performance shaping factors(PSFs)with its weight have to be given through expert estimation, which has its own drawback of the uncertainty during the determination of PSFs. The Cognitive Reliability and Error Analysis Method(CREAM)has representative and concluded common performance conditions(CPCs)which could be used to replace the PSFs of SLIM, to eliminate the experts’ subjectivity. And the triangular fuzzy number is used to represent the scores made by experts, then the weights of PSFs is determined by fuzzy judgment matrixes. Due to the differences of knowledge, experience,ability and preference among experts, the weights of experts should be considered. And the Bayesian formula is used to modify the experts’ weights, making the group opinions more accurate and objective. At the last part, five representative human errors are chosen for probabilistic forecasting using developed SLIM. And the probability of each human error is concluded, which is meaningful in precaution and identification human errors in aviation maintenance.