Application of statistical criteria for improving the efficiency of risk assessment methods

Abstract

Aim. The aim of this paper is to improve the efficiency of the failure mode and effects analysis (FMEA) method through the verification of expert judgment correctness by means of statistical methods. Therefore, the paper deals with the matters of improving the quality of products and services in various enterprises through risk-oriented approaches. Methods. For the purpose of improving the efficiency of the failure mode and effects analysis (FMEA) method, it is suggested to increase the number of experts, while making the expert evaluation an independent process, i.e. by separating the experts from each other. The resulting expert judgment is proposed to be considered as a random value. The correctness of the expert judgment is suggested to be evaluated by means of statistical criteria methods, e.g. Grubbs’ criterion methods. The proposed evaluation methods are not limited to the Grubbs’ criterion methods. This criterion can be replaced by the Cochran’s criterion or Shewhart charts. Each of the suggested methods enables more efficient estimates with lower risks in the process of service provision or product manufacture. The paper proposes statistical methods with the example of the Grubbs’ criterion. All indicators of the integral estimation of the failure mode and effects analysis are submitted to statistical verification. Results. Such data verification as part of an independent expert evaluation enables a higher reliability of expert judgment and significantly reduces the number of risks at the enterprise. Such risks may include bribing or collusion of experts involved in the performance of the failure mode and effects analysis. Independent expert judgments after expert evaluation are verified by means of statistical methods. Sharp spikes in independent expert opinions will justify repeated expert evaluation, while complete agreement of evaluations will eliminate doubts regarding the quality of the performed assessment. The use of statistical methods for the evaluation of every indicator of the integral FMEA estimation will allow increasing its reliability. A combination of those approaches enables an independent estimation as part of various projects evaluation, including the evaluation of industrial products or provided service using failure mode and effects analysis, elimination of the human factor in the estimation procedure, significant reduction of risks. Conclusions. Failure mode and effects analysis (FMEA) was performed. Additionally, the method was improved by means of independent expert assessment. The consistency of the results of such evaluation is verified by means of statistical methods. The performance of such verification of independent expert opinion is demonstrated through the Grubbs’ criterion. Expert opinion can also be verified by means of Cochran’s criterion, Shewhart charts. The proposed approach is a combination of the failure mode and effects analysis (FMEA) method and statistical methods with the example of Grubbs’ criterion.