Abstract
Eddy current techniques are favoured for measuring the properties of conductive coatings on magnetic substrates because the skin effect ensures that only the coating or region at or just below the surface of a magnetic substrate is inspected. The data acquired from eddy current sensors, however, is affected by a large number of variables, which include sample conductivity, permeability, geometry, and temperature, as well as sensor lift-off. The multivariable properties of sample coatings add an even greater level of complexity. This research project is therefore motivated by the need for a measuring instrument, which can intelligently adapt to the large number of variables that affect eddy current measurements on steel. Sensor model optimisation against experimentally generated data is undertaken, leading to the development of accurate and fast inversion models based on artificial neural networks. Neural network architecture, operation and training are discussed, which includes an algorithm for neural network normalisation and calibration. System tests are fmally undertaken on a wide range of plated samples. This research demonstrates that an intelligent measuring system incorporating a ferrite-cored sensor can provide high accuracy while operating over a wide frequency range.