Prediction and experimental validation of weld dimensions in thin plates using superimposed laser sources technique

Gas Metal Arc Welding (GMAW) is one of the primary techniques used to join thin structures together. The quality of the weld plays an important role in structure integrity and product safety. Weld dimensions such as penetration depth, leg length, throat thickness, and reinforcement height are key to the quality of welds. Therefore, it is crucial to accurately measure them. Previous research has shown that non-destructive evaluation using laser generated bulk waves and electromagnetic acoustic transducer (EMAT) reception is an efficient and effective way to monitor weld quality in thick structures. Laser generated Lamb waves have the potential to be used to monitor weld quality in thin structures. However, due to the fact that laser generated Lamb waves in thin structures are broadband and dispersive, the complexity of ultrasonic signals is greatly increased.;The objective of this research is to develop a method to measure important weld dimensions in thin plates by using laser generated ultrasound. This research comprises three aspects: First, to develop a technique that can generate narrowband Lamb waves in thin plates. Secondly, to develop a signal processing procedure to extract useful information from the ultrasonic signals to evaluate weld dimensions. Thirdly, to develop prediction models to predict weld dimensions by using the reflection coefficients of narrowband Lamb waves.;The technique named superimposed laser sources (SLS) technique is developed to generate narrowband Lamb waves in thin plates. By using the superimposed laser sources, one has the flexibility to generate desired wavelengths of Lamb waves. The advantage of generating narrowband Lamb waves with fixed wavelengths is that the dominant frequency content and traveling speeds of different wave modes can be determined from the dispersion curves.;The signal processing procedure developed in this research is used to reduce the complexity of the signals of Lamb waves in thin structures. It includes wavenumber-frequency (k-o) domain filtering and synthetic phase tuning (SPT). The k-o domain filtering technique helps to filter out the unwanted wave components traveling at the direction that are irrelevant to our analysis and the SPT technique is used to amplify and isolate a particular Lamb wave mode. The signal processing procedure facilitates the calculation of reflection coefficients of Lamb waves that result from the presence of weld joints.;Reflection coefficients that result from the welds can be calculated for A0 and S0 Lamb wave modes for ten discrete wavelengths of interest. Two methods, the direct method and the indirect method, are used to develop models that use reflection coefficients as predictors to measure these weld dimensions. The assumptions made in these two methods are intrinsically different. In the direct method, weld dimensions are assumed to be functions of the reflection coefficients. But in the indirect method, it is assumed that the reflection coefficients are functions of the weld dimensions. Different approaches are taken to identify significant predictors that are used in the prediction models. Both models are shown to effectively predict weld dimensions in thin plates and they are complementary to each other. Furthermore, from the model developed by the indirect method, the response of each reflection coefficient to the change of weld dimensions can be shown. The results provide us a way to investigate the interaction between Lamb waves and geometry of welds. The advantages and disadvantages of these two methods are discussed, and the detailed discussion about the sources of errors is presented.;The weld dimensions measurement techniques and procedures developed in this research have resulted in a new nondestructive and noncontact method for measuring important weld dimensions in thin plates. The techniques and procedures have great potential. They can be applied to other types of thin structures such as curved thin plates. They can also be applied to evaluate welds made by other types of welding processes such as friction stir welding. They will help to improve the quality and efficiency of the welding process on thin structures and reduce costs, material waste and human injury.