Adaptive ultrasonic inspection of specimens with varying profile using phased array techniques

Phased array techniques, providing a dynamic control of the ultrasonic beam , have been widely used in medical imaging, and subsequently in non-destruc tive testing. Such techniques allow to steer and focus the ultrasonic beam radiated by a transducer split into an array of individually addressed elem ents, according to specific delay and amplitude laws, supplied by a multi-c hannels acquisition system((1)). In the nuclear industry, those beamforming abilities have been commonly use d to inspect specimens with varying angles and focal depths, or using secto r scanning and tomography techniques((2,3)). The French Atomic Energy commi ssion (CEA) has developed the FAUST, (an acronym for Focusing Adaptive ULtr aSonic Tomography) system, supported by the French Institute for Nuclear an d Safety Protection (IPSN), to fulfil major requirements of safety authorit ies concerning improvement of defect characterisation and adaptability to v arious configurations (complex geometry and state of surface). Such a syste m has been used to perform depth focusing using theoretical((4)) and experi mentally extracted delay laws((5)) in order to optimise lateral and axial r esolution for inspections carried out on planar specimen containing calibra ted defects. Specific developments devoted to cladded specimen inspection h ave also been carried out((6)). This paper deals with specific adaptive modes dedicated to more complicated specimen. Two specific configurations: a drastic varying profile simulatin g a misaligned specimen, and the inspection of toric specimen (a cast elbow ), scanned from the inside to inspect both curved walls of different curvat ures, have been investigated using different phased array techniques. Unlik e sector scanning schemes using beam steering with a fixed contact array, t he technique described hereby aims at preserving the ultrasonic beam charac teristics while scanning a specimen with varying surface profile, for the m isaligned specimen, whereas the inspection of the toric specimen only requi res two different dealy laws for each inspection.