A benchmark experiment to assess the performance of electrical resistivity tomography data inversion algorithms in the framework of nondestructive evaluation

Electrical Resistivity Tomography (ERT) is an effective, non-invasive technique widely used in geophysics and civil engineering to identify the resistivity field of the ground or of the concrete. Over the last decades, the field of ERT has seen significant advancements, particularly in the development of measurement equipment and inversion algorithms. Despite these advancements, there remains a gap in experimental benchmarking, which is essential for validating the results and ensuring the reliability of ERT techniques in practical use cases. This manuscript presents a comprehensive procedure for setting up physical models suitable for ERT experiments at the lab scale. These physical models, consisting of gelatin and water blocks in a tank, enable testing of multielectrode sensors, of DC-resistivity measurement equipment, and, most importantly, of inversion algorithms. This work is particularly useful to the Non-Destructive Testing (NDT) and geophysical research communities, as it ensures that the procedure is fully reproducible. Additionally, the data generated in this study are openly accessible. A comparative analysis of two popular inversion algorithms is performed on the generated data to illustrate the relevance of the proposed benchmarks and to assess the performance of commonly used inversion algorithms in the ERT field.