ULTRASONIC IMAGING SYSTEM FOR FLAW CHARACTERIZATION

Abstract

Among all the nondestructive evaluation (NDE) techniques, the ultrasonics is the most widely used technique. This is capable to detect and characterize any type of defects in test specimens and determine their size, shape and orientation. The in-flow of digital computers has helped in making the ultrasonic inspection process fully automatic and operator independent. The development of present ultrasonic imaging system is an effort towards the detection of flaws in test materials and thereafter to image them to know their shapes. The system is fully automatic and user-friendly. The hardware circuitry, in present ultrasonic imaging system, is based on conventional pulse echo method employing a single transducer of 2.5 MHz nominal frequency and acting as transmitter and receiver, both. The signal transfer to the personal computer is accomplished by using the analog to digital conversion facility of a digital oscilloscope (Nicolet 320). A motorized mechanical system is developed to effect the probe movement for scanning. The detection of flaws (target echoes) is achieved by processing the recorded signal by either of the two signal processing methods : cross correlation and linear area estimation. In cross correlation, the relationship between transmit and received signal at individual sample points is computed. Here, first of all, the FFT values for the transmit and received signal are computed separately and from these values their cross power spectrum is evaluated. Finally, the inverse FFT of cross power spectrum values yields the cross correlated values. The pinpointed localization of target echoes is done by using step averaging and thresholding. A new technique, named as Linear Area Estimation (LAE) technique , is suggested here. The technique consists of three steps : band pass filtering, LAE operation and step averaging & thresholding. The band pass filtering is performed to remove any signal (noise) falling out of the echo frequency range. The filtered signal Is LAE operated. This enhances the signal to noise ratio to a considerably high level. Through the step averaging, the target echo envelopes are formed. Thereafter, pinpointed location of the target echoes are extracted by applying the threshold cut-off procedure and equal energy distribution criterion . In LAE method, the processing is done only for the received signal and thus a large amount of computations required for the transmit signal is eliminated. In addition, this technique is much faster and requires less memory as compared to the cross correlation technique. All the three types of images viz. A, B and C-scan, are formed at CRT monitor. To form the A-scan image the radio frequency (rf) waveform is recorded and processed by either of the two signal processing methods and the target echoes are extracted with their amplitude and location. At CRT monitor the recorded signal waveform and the processed signal waveform both are displayed simultaneously. Additionally, the echo information with its amplitude and location is also displayed. Thus, a complete on-line information is provided at monitor in each A-scan image. The time Interval between the transmit echo and the first received target echo is determined and is accordingly displayed in its different forms of images. A study has been done to determine the empirically optimized values of the different parameters required for the two signal processing methods. This has been made possible by generating the transmit and received signal by a simulation algorithm. At later stage, the algorithm for signal generation is exploited to simulate the process of ultrasonic scanning by generating the received signal from the software simulated test object. The simulated test object ( rectangular block ) can have any dimension with a single flaw of either rectangular, cubic, cylindrical or spherical shape having placed any where in the object. For this specified test object an A-scan image for individual scan points can be generated. The B or C-scan images have been formed by collecting A-scan information of individual scan points over a scan line or surface, respectively. The evaluation of the system was done with using the ultrasonic signals received from different types of test pieces e.g. varied lengths of brass bar and rectangular blocks of different materials such as aluminium, steel and brass. The extracted information was found to be satisfactory. The scanning process simulation part of the system is very useful in generating a data base for flaws of different shapes and sizes which can be further utilized for identifying the flaws.