Corrosion detection in industrial pipes using guided acoustics and radial basis function neural network

Detection of corrosion in industrial pipes and tubes is a challenging task. This paper presents an efficient and a low-cost approach for corrosion detection and classification. The proposed approach combines the guided acoustic waves and Radial Basis Function Neural Network (RBF-NN) classifier. Mean, RMS, Variance, Kurtosis and Skewness are used as distinct features for the RBFNN classifier to classify corrosion and no corrosion state. Network is trained to reduce the Mean Squared Error (MSE) up to 0.0074. The proposed technique shows promising results to detect and classify corrosion with an accuracy of 99.45%. The specificity and sensitivity of the classifier are found to be 99.54% s and 99.6% respectively.