Citation:
Abstract:
Conductors serve as essential components in various electrical and electronic applications (steel, aircraft, and nuclear industries). Therefore, an accurate evaluation of their electrical parameters, in particular their electrical conductivity (σ), remains critical for assessing their performance in industrial processes. Although numerous eddy current based methods exist for conductivity measurement, this study approaches the problem through inverse problem solving. A novel approach integrating Eddy Current Testing (ECT) with Artificial Neural Networks (ANNs) is proposed to determine electrical conductivity from probe impedance measurements. An experimental setup has been developed that includes a custom-designed bobbin coil probe used in conjunction with metal plate samples (targets) and data acquisition and signal processing systems. To validate the introduced approach, conductivity values predicted by the ANN model were rigorously compared with reference measurements obtained using the four-point Direct Current Potential Drop (DCPD) technique. This comparative analysis demonstrates the robustness and measurement fidelity of the proposed approach.