Nonparametric Statistical Approach For Failure Prioritisation in Reliability Centered Maintenance at Petrochemical Laboratory

Abstract

The petrochemical company, a major producer of PTA in Indonesia, has an important laboratory to ensure product quality, with capillary electrophoresis as one of the vital instruments for determining quality. The maintenance strategy currently implemented, namely preventive and corrective maintenance, is not optimal in preventing sudden downtime. The Reliability Centered Maintenance (RCM) method is proposed as a more systematic approach. In the RCM process, Failure Mode and Effect Analysis (FMEA) is used to identify failure risks; however, the conventional FMEA method has limitations in determining the Risk Priority Number (RPN), which can result in identical values. This study uses the Reliability Centered Maintenance (RCM) method, which includes qualitative and quantitative analyses. One of the qualitative analyses is determining FMEA, with the output being the RPN value. The qualitative analysis of the RCM method resulted in identifying critical instruments and critical instrument capability limits, creating a critical instrument system block diagram, and identifying FMEA, which produced 45 critical instrument failure modes (FM). The quantitative study proposed a nonparametric statistical approach, namely the Mann-Whitney and Kruskal-Wallis tests, to optimize failure priority ranking. The Mann-Whitney test results for FM with two identical RPN values (U_value>37;p>0.05) showed insignificant results and through expert consideration was able to distinguish priorities among 16 FM with identical RPN, while the Kruskal-Wallis test results for FM with more than two identical RPN values (H_value<5.99;p>0.05) showed insignificant results and through expert consideration was able to distinguish priorities in 13 FM with three or more identical RPN values.