This dataset contains simulated SAFE, FEA, and experimental data used to generate most of the figures featured within the manuscript.
The SAFE data was generated by solving the equation of dynamic equilibrium using the commercial software, COMSOL. These models predicted the existence of acoustic waves in the resistance seam weld waveguide. All files are provided with no solutions saved to minimise storage requirements. The SAFE data was analysed using MATLAB through the information provided in the manuscript and was used to generate Figure 5, 6, & 7. Additional models are provided replicating the results observed in: https://doi.org/10.1016/j.ymssp.2019.02.026.
The FEA data was generated via timestep models solved in the commercial software, POGO, from Imperial College London. This confirmed the findings of the SAFE model above and verified that one type of wave could be excited and readily detect defects. Like before, the data generated was analysed using MATLAB, PogoPro, and ParaVIEW. The data generated by these models was used to create Figure 8, 9, 10, 11, 12, & 13. Again, additional models are provided replicating the results observed in: https://doi.org/10.1016/j.ymssp.2019.02.026.
The experimental data was acquired using the equipment described in the manuscript. These experiments showed the findings from the simulated efforts existed in reality, and explored the nuances of exciting such a wave out with a simulated noiseless environment. Specialist equipment such as a Polytec MSA-100-3D laser Doppler vibrometer was used but could be easily swapped for an equivalent. A wavemaker from Macrodesigns LTD was used to excite the transducer. This allows for sinusoidal excitation of the transducer which is uncommon. Again, this can be replaced for an equivalent. Details of the transducer, geometry and materials of the experiments are all given in the manuscript. The data generated was again analysed using MATLAB, and was used to create Figures 17, 18, 20, 21, 23 & 25.