Probabilistic properties of measurement for thickness of pipe with metal loss

Kaida, Takuyo*; Ishizaki, Yoichi*; Okajima, Satoshi 

Probabilistic approach for Fitness-For-Service (FFS) assessment of pressure equipment in petroleum and petrochemical plant helps to quantify the safety margin and rationalize the assessment. As for local metal loss assessment, several studies have been made on reliability analysis of the damaged equipment by use of limit state function, denoted by deterministic FFS model in published standard. Stochastic properties of random variables are needed in the reliability analysis. Remaining thickness of the pressure equipment is one of the important variables in limit state function for metal loss assessment. The thickness readings on a prescribed grid, which characterize the remaining thickness and size of the region of metal loss, might be varied by each inspector. There has been, however, no studies on stochastic properties of thickness readings for the local metal loss assessment. In this study, the stochastic properties of thickness readings were investigated by analyzing of the results of Round-Robin test. The cause of variation of thickness readings was clarified for the purpose of either protection of failure by pressure or that of leakage. The results of investigation indicated that standard deviation of thickness readings in the region of metal loss was more than that of thickness away from metal loss. Finally, the reliability of pressure equipment with local metal loss was analyzed using stochastic properties.