Inspection of copper canisters for spent nuclear fuel by means of ultrasonic array system. Modelling, defect detection and grain noise estimation

摘要

The work presented in the report has been split into three overlapping tasks which have the following objectives: (1) development of beam-forming tools, and verification of modeling tools; (2) investigation of detection and resolution limits; (3) evaluation of attenuation, estimation and suppression of grain noise. For beam-forming tools, a method of designing steered and/or focused beams in immersed solids is presented based on geometrical acoustics. Presently, the beam designs are only related to delays but not to apodization. These focused, steered beams are intended to be used for sizing defects and inspecting the regions close to canisters outer walls. The modeling tool developed previously for simulating elastic fields radiated by planar arrays into immersed solids has been verified by comparing with the results obtained from PASS, a software developed by Dr. Didier Cassereau, France. The results from our modeling tool are in excellent agreement with those from PASS. Since the array coming with the ALLIN ultrasonic array system is not planar, but cylindrically curved in elevation, and it works not in transmission mode, but in pulse echo mode, the above modeling tool for the planar arrays cannot be applied directly. Therefore, the modeling tool has been upgraded for the ALLIN array. The theory underlying this modeling tool is the extended angular spectrum approach (ASA) which was developed based on the conventional ASA that only applies to planar sources. Experimental verification of the modeling tool has shown that the results from the tool agree very well with the measurements. To quantify the fields from the ALLIN array and to facilitate the comparison of simulated results with the measured ones, the ALLIN array system has been calibrated based on the existing functionality, and an analytical model has been proposed for simulating measured acoustic echo pulses. To investigate the detection and resolution limits, we have carried out a series of experiments. These experiments have demonstrated that use of focused, steered beams is a very effective solution to the inspection of the zone close to the outer walls of copper canisters, and they have also indicated the most suitable beam angle for this inspection. For evaluation of attenuation, the log-spectral difference method and the spectral shift method have been employed. Measurements were made on copper specimens of different grades. The results have shown that the spectral shift method gives a stable estimation of attenuation when the echoes from front and back surfaces of a specimen are used. Therefore, the spectral shift method has been chosen for the attenuation evaluation. For estimation of grain noise, two statistical models, i.e., the independent scattering model (ISM) and the K-distribution model (KDM), are used. The ISM has been applied to estimate grain noise in three copper specimens with different grades. The results have shown that the model gives good prediction under the approximation which is expected to be valid for the early time portion of a signal when the main beam has not been significantly attenuated. They have also demonstrated that the figure of merit (FOM) obtained from the ISM can be a good parameter used for depicting grain noise severity. The KDM has been further exploited and applied to evaluate grain noise from welds in copper canisters, and also appl