(33b) A Statistical Approach for Using Test and Inspection Results to Adjust Maintenance Strategies

William M. Bradshaw, ABSG Consulting Inc., 10301 Technology Drive,
Knoxville, TN, 37932 (865) 671-5829, WMBradshaw@ABSConsulting.com

Jan Wagner, ABSG Consulting Inc., 10301 Technology Drive, and
Oklahoma State University (emeritus), (405) 372-8963, jan.wagner@okstate.edu

Keywords: Mechanical integrity; inspection, testing, and preventive maintenance (ITPM)

Abstract

Standards, such as American Petroleum Institute (API) standards, provide widely accepted methods and intervals for inspecting fixed equipment such as pressure vessels, piping, and storage tanks. Test methods and test intervals for safety instrumented systems are typically established during the detailed design phase. However, there is little or no guidance regarding inspection, testing, and preventive maintenance (ITPM) activities and intervals for a broad range of equipment that serves a process safety function.

Over the years, most plants have identified ITPM activities and established intervals based on factors such as (1) turnaround intervals, (2) auditors’ opinions, (3) historical practices, (4) risk assessments, (5) manufacturers’ recommendations, and so forth. Most facilities have not used ITPM results and in-service failure data to determine whether (1) ITPM intervals can be increased while meeting reliability requirements and (2) test and in-service failures of a safety system indicate inadequate performance or are simply the result of statistical noise.

In this paper, we present a statistical method for evaluating the likelihood of an in-service failure given (1) a failure rate distribution, (2) the facility’s maximum allowable failure rate (e.g., a risk reduction factor used in a layer of protection analysis), and (3) the user’s willingness to get the wrong answer. In addition, we provide a statistically valid method to use ITPM and in-service failure rate data to determine the likelihood that reliability requirements are being met.