A Multi-Attribute Knowledge Criticality Framework for Ranking Major Maintenance Activities: A Case Study of Cement Raw Mill Plant

Abstract

Systematic failure analysis enhances the ability of decision makers to implement strategies that are beneficial to systems they manage. However, in industrial maintenance activities such as, Major overhauls, outages, shutdowns and turnarounds (MoOSTs) there is scarcity of knowledge and experience, limiting the effectiveness of such failure analysis. Transformation of knowledgeable actions generated from experts’ tacit based knowledge from performing MoOSTs is encouraged. A key step to achieve such transformation is by prioritizing maintenance efforts by critically assessing relevant maintenance attributes. Criticality analysis of tasks is considered as an effective approach for prioritizing MoOSTs activities. This paper combines a traditional approach for analysing attributes of frequency and consequence factor values ranked by experts using a mathematical relationship to determine critical activities as well as a fuzzy logic system to develop a fuzzy inference system (FIS) for generating fuzzy criticality numbers of MoOSTs activities. In this regard, the traditional method qualitative criticality matrix, and boundary settings by experts provide baseline information for the FIS, to establish If-Then rules and map membership functions of two crisp inputs and output. Practical applicability is demonstrated using a Raw Mill System (RMS) from a cement manufacturing plant. The comparison of results from the two methods shows slight variations in criticality numbers, howbeit a consistent ability to capture critical MoOSTs activities. Moreover, the validity of results obtained by the fuzzy logic system is enhanced and more superior because it can demonstrate sensitivity.