Prediction of the Ferrite-Core Probe Performance Using a Neural Network Approach

Abstract

Physical properties and coating thickness of the materials are measured using the eddy current principles. Most publications on the eddy current method are directed toward air-core coils operated at high frequency. But low-frequency ferrite-core probes provide the advantage of enhanced signal-to-noise ratio and increased resolution for electronic detection. A ferrite-core probe design has several parameters, which includes the selection of ferrite material and the design of the probe structure and core. In this article a total of five variables, which are most influential on probe performance, were considered. A three-layer back-propagation neural network model is used to correlate these variables to probe performance using the data generated based on experimental observations.

Keywords:

• Eddy current
• Ferrite-core probe
• Neural network