PROBABILITY OF DETECTION MODEL FOR GAS TRANSMISSION PIPELINE INSPECTION

In nondestructive testing, uncertainties due to variabilities in inspection parameters can affect the accept/reject decision of a flaw. The concept of probability of detection (POD) offers a measure for quantifying the capabilities of nondestructive evaluation (NDE) systems in the presence of various sources of uncertainty. This paper presents a model based POD evaluation technique for magnetic flux leakage (MFL) inspection of natural gas transmission pipelines. The technique consists of two parts: finite element (FE) prediction and a parametric functional model. The procedures and results of evaluating PODs with respective to experimental parameters such as sensor liftoff, magnetization level, and sensor spacing are described in this paper.