Defect detection and signal processing for cast stainless steel

Abstract Reliable ultrasonic inspection of cast stainless steel components in the primary piping loop of pressurized water reactors (PWRs) has not been satisfactorily demonstrated to the US utilities. The inspection for flaws is difficult because strong reflections from the grain boundaries could mask the flaw signal response; in addition, the ultrasonic beam can be redirected by the ‘wave guide’ effect of the coarse grain structure. This paper discusses signal processing approaches that may be relevant to inspection of such materials. Specifically, spatial averaging—a proven technique for enhancing signals that exhibit spatial coherency—can be useful in enhancing signal-to-noise ratio (SNR) for flaws the length of which are larger than the average grain size. Several specimens of PWR main coolant piping, each a cast steel elbow welded to forged steel containing fatigue flaws, were examined with a field-usable ultrasonic imaging system and with a personal computer (PC) based data acquisition and imaging system, both developed under EPRI funding. In addition, inspection data were examined from a component in a PWR plant during a scheduled outage in 1987. The images produced before and after signal processing showed improvements in flaw detectability.