Improved eddy current testing sensitivity on low-conductivity alloys

A To 

This project aims to continue the development of an eddy current testing (ECT) sensor capable of detecting surface breaking defects and fatigue cracks of less than 250 microns in length in safety-critical components. To detect defects at their earliest stage, the sensor needs to be able to find sub-millimetre sized defects. The detection of small defects in titanium and superalloys is difficult because, in general, their low conductivity requires a high-frequency current greater than several MHz, for small skin depths of less than 1 mm. In some of the group’s work, eddy current systems capable of working at frequencies of up to 40 MHz have been made, with good signal-to-noise performance, or the possibility of operating close to coil resonance to improve defect detection has been investigated as part of a wider portfolio of potential approaches to eddy current NDT.

Decreasing skin depth by increasing frequency is difficult, as the capacitance and inductance of wires or cables connecting the eddy current coils to the electronics becomes significant at higher frequencies, increasing the experimental noise in the measurement. While the sensor and supporting electronics has been designed to reduce this, situating the circuitry directly behind the coils, the work in this project has focused on using a combination of techniques to improve defect detection without needing to operate at very high frequencies. A key step in this research is improved analysis of both phase and magnitude data, similar to the concept described in earlier work of the group.

Taking this parametric approach to measure the signal magnitude and phase independently not only improves the sensitivity of the method to defects, but also improves stand-off performance and the ability to measure defects close to the edge of a sample. A number of results are reported, where the authors are trying to better understand the interaction of eddy current and defect through experiment and finite element modelling (FEM).