Evaluation of Pressure Relief Valve Maintenance Intervals Using a Weibull Analysis with Bayesian Updating

Abstract: Spring operated (pressure) relief valve (SORV) maintenance intervals for the Department of Energy's Savannah River Site (SRS) were evaluated using a Risk-Based Inspection Methodology approach published by the American Petroleum Institute (API RP 581). The API approach is characterized as a Weibull analysis with Bayesian updating and then modified using SRS SORV proof testing experience. In addition, the applicability of the API approach was evaluated using Monte Carlo Simulation. The SORVs were classified according to their service severity, and their failure rates were estimated using Weibull parameters provided in the API recommended practice. Initial Weibull parameters were updated using SRS's historical proof test records as submitted to the Center for Chemical Process Safety Process Equipment Reliability Database (CCPS-PERD). The statistical results provide for an evaluation of SRS's maintenance intervals based on actual proof test experience. Current practices are reviewed, and recommendations for extending or reducing inspection intervals are discussed. The Weibull analysis with Bayesian was used to estimate the SORV's probability of failing on demand (PFD) and also the remaining service life in the non-fail condition (ratio of proof test to set pressure < 1.5) if left in place past the scheduled maintenance time. Maintenance intervals were compared to the limits of 95% reliability t_0.95. The maintenance interval could be extended if t_0.95 is greater than the scheduled maintenance interval.