The 24th International Workshop on Electromagnetic Nondestructive Evaluation (ENDE2019)

215 / 2019-04-06 19:49:42

PREDICTION OF THE HARDNESS OF X12M USING BARKHAUSEN NOISE AND CHEBYSHEV POLYNOMIALS REGRESSION METHODS

Non-destructive Evaluation,Hardness Prediction,Chebyshev polynomials regression,Barkhausen Noise

Plenary track > QNDE and AI for ENDE

Barkhausen noise (BN) is an efficient non-destructive evaluation signal that could be used to predict the properties of materials such as hardness, residual stress and carbon content. Current BN signal analysis methods failed to describe the highly variated BN signal and achieved the limited regression accuracy. In this paper, two multi-variable regression tools was firstly employed to predict the property of material. Combined with Chebyshev polynomials, our regression tools were designed on the basis of cascaded regression and mutual-information-based feature selection. As implied by the experimental results on predicting the hardness of Cr12MoV material (i.e. X12m), our proposed method outperforms other comparative methods including neural network and partial linear square regression method.