Steam Crack PORTABLE By Nicolo
Cast aluminum track shoes reinforced with metal matrix composite (MMC) inserts at heavy loading areas such as center splines and sprocket windows are light in weight, and can resist high temperature and wear. Various defects such as disbonds at the insert-substrate interface, cracks and porosity in the MMC layer, etc. can be introduced during the manufacturing process and/or in service. This paper presents a portable ultrasonic system to automatically inspect tank track shoes for disbond. Ultrasonic pulse/echo inspection has shown good reliability for disbond detection. A prototype sensor array fixture has been designed and fabricated to prove the feasibility. Good agreements between the sensor fixture results and ultrasonic C-scan images were obtained.
Application of a self-compensating technique for ultrasonic inspection of airplane structures makes it possible not only to detect cracks in the different layers of joints but also to obtain information on crack sizes. A prototype computerized ultrasonic system, which utilizes the self-compensating method, has been developed for non-destructive inspection of multilayered airplane structures with in-between sealants, such as bolted joints in tail connections. Industrial applications of the system would require deployment of commercially available portable modules for data acquisition and processing. A portable ultrasonic flaw detector EPOCH II manual scanners and HandiScan, and SQL and FCS software modules form themore PC-based TestPro system have been selected for initial tests. A pair of contact angle-beam transducers were used to generate shear waves in the material. Both hardware and software components of the system have been modified for the application in conjunction with the self-compensating technique. The system has bene tested on two calibration specimens with artificial flaws of different sizes in internal layers of multilayered structures. Ultrasonic signals transmitted through and reflected from the artificial flaws have bene discriminated and characterized using multiple time domain amplitude gates. Then the ratios of the reflection and transmission coefficients, R/T, were calculated for several positions of the transducers. Inspection of measured R/T curves shows it is difficult to visually associate curve shapes with corresponding flaw sizes and orientation. Hence for online classification of these curve shapes, application of an adaptive signal classifier was considered. Several different types and configurations of the classifiers, including a neural network, have been tested. Test results showed that improved performance of the classifier can be achieved by combination of a back-propagation neural network with a signal pre-processing module. less