In order to monitor the loose state of transmission tower bolts, this study discusses the application of voice print recognition technology in this field. By collecting the sound signals under different loosening states of the transmission tower bolts and using Mel cepstrum coefficient (MFCC) feature extraction method, the sound signals are converted into feature vectors for analysis, and the loosening state of the bolts is effectively recognized. The experimental results show that the characteristic value obtained in the bolt loosening test is 4.3766, which is higher than the MFCC characteristic critical value 4.0, indicating that the bolt loosening problem exists. At the same time, in the bolt fastening test, the result obtained is 3.8364, which is lower than the MFCC characteristic critical value 4.0, indicating that the bolt is not loosened. It can be seen that whether the bolt loosening test or the bolt tightening test, the method based on MFCC feature vector has a considerable recognition effect, and can judge the bolt state through the eigenvalue method. This study provides a new non-contact technique for real-time monitoring of bolt loosening state of transmission tower, and provides important guarantee for the safe operation of transmission lines. This technology has a broad development prospect in practical application, and provides an important reference and guidance for further research and practice in related fields.
Read full abstract