Intelligent assessment of rock structure degradation based on entropy features of piezoelectric signals and their algorithm-optimized fusion

The process of assessing structural degradation is crucial, as it ensures structural integrity and prevents potential hazards, while also aiding in the development of maintenance strategies. This article introduces an entropy-based methodology for extracting structural degradation features from piezoelectric signals to evaluate the degradation of rock structures. Initially, the Fourier transform decomposes the piezoelectric signal into various harmonic components, followed by the reconstruction of frequency-band component signals through the superposition of harmonic signals within a predetermined frequency band. The optimal frequency bandwidth and the frequency position of the reconstructed components are determined using a predetermination dataset based on the correlation coefficient. An intelligent multifeature fusion algorithm, based on a genetic algorithm, is designed to assign weights to each feature and introduce a bias to correct fusion errors. Experimental studies on granite demonstrate over 90% accuracy in assessing the rock health index using piezoelectric signals to identify degradation states.