Equipment criticality assessment — Global methodology

This article presents the overall methodology for assessing the criticality of the consequences of equipment failures in a complex industrial system.

This assessment is of strategic importance for managers, providing them with an instant view of their company's performance, and enabling them to implement improvement actions if necessary. The methodology recommended in this article is based on seven main steps.

The first step emphasizes the need to define the precise objectives of a criticality study, which must be validated and endorsed by the company's senior management. These criticality assessments may concern economic, strategic, financial, regulatory, technical or organizational aspects. They can be used to establish performance indicators and/or draw up dashboards.

The second stage proposes the organization of a working group bringing together all the players required to carry out the criticality study.

Based on the objectives defined in the first stage, the third stage corresponds to the precise definition of the system, its equipment and its limits.

The fourth step describes the main methods of functional analysis. These methods are essential for creating functional and/or material trees for complex industrial processes. The aim of these methods is to help you understand your reasoning by visualizing the path between the causes and various effects of failures, and to define the appropriate criterion level. The FAST, SADT ^® , IDEF0 and Apte ^® methods will be briefly described.

The fifth stage proposes a selection of the most appropriate metrics and performance indicators for the objectives sought. Specific descriptions will be given of "lagging indicators" and "leading indicators". Among the vast array of "key" metrics and performance indicators, severity and frequency of occurrence rating scales, risk priority ranking (RPN) and the Farmer curve will be presented. Reference will also be made to the main international standards defining organizational, economic and technical criticality indicators.

The sixth step provides a typology of the reliability data needed to construct criticality metrics. As these data are by nature random variables, only statistical estimates are possible. Point or interval estimation methods are described, and the principles of frequentist and Bayesian approaches to estimating probability distributions are recalled.

The seventh step establishes an inventory of international reliability data banks (MIL-HDBK-217F, OREDA, NSWC, Telcordia Issue 3, RIAC 217Plus, IEC TR 62380, FIDES, CHINA GJB/z 299B,...