Risk Assessment Model Of Small-batch Customization Process Failure Mode And Its Sensitivity Analysis

Process Failure Mode and Effects Analysis is a process reliability analysis and quality control technology,which has been widely used in process management under industrial mass production mode.However,under the small-batch customization production mode,the risk assessment model is still insufficient.For example,the evaluation value is a vector rather than a scalar,which is not convenient for intuitive understanding and application.The expert confidence level information is not fully exploited,and the The degree of discrimination of the model is reduced;the sensitivity analysis of the model parameters is not performed,so there is no guidance to develop evaluation criteria for evaluation parameters that are of great value to engineering practice.Aiming at the shortcomings of the existing research,the paper first constructs a risk factor evaluation model based on generalized trapezoidal fuzzy numbers.The risk assessment model introduces the generalized trapezoidal fuzzy number and fully exploits the confidence information of the expert evaluation opinions.It not only has a good discrimination degree,but also establishes a sorting scalar threshold and a conversion method between the risk factor threshold and the traditional threshold.Secondly,based on the coupling relationship of input evaluation parameters and the characteristics of each evaluation parameter,the paper establishes a distribution model of 12 input evaluation parameters including three factors: Severity,Occurrence and Detection.Among them,the lower bound and the upper bound of the true value of the risk factor obey the truncated skew normal distribution,and the deviation tendency parameter gives the distribution law as a discrete distribution.The confidence parameter obeys the beta distribution.Based on the influence of the distribution parameters on the probability density function in the distribution model,the range of values of each distribution parameter is determined.Finally,in order to apply the proposed risk assessment model to guide the development of scientific evaluation parameter selection criteria,the paper performs a global sensitivity analysis based on Monte Carlo method for 12 random evaluation parameters.According to the change of the sensitivity index,the input evaluation parameters which have significant influence on the output ranking index under different conditions are determined,and the criteria for the input evaluation parameters are established based on this.The research results show that the risk assessment model proposed in this paper has better discrimination than the existing model.The proposed cut-off threshold can not only filter the failure modes with higher risk order than the threshold,but also can filter the risk factor;the sensitivity analysis results not only verify the accuracy of the input random evaluation parameter distribution model,but also the established value criteria can help the evaluator to make a more scientific estimation of the risk level of the failure mode.The research results not only enrich the risk assessment theory of the existing small batch custom production process failure mode,but also has good practical application value for engineering practice.