Experimental Study and Signal Analysis of Acoustic Emission from Typical Damages of Strain Clamps for Transmission Lines

Strain clamps (dead-end clamps) are critical components of transmission lines, whose operation status affects the safety and stability of the entire power system. It is essential to realize the detection of the crimping quality of strain clamps and achieve effective identification of damages in strain clamps. In this study, a test rig for strain clamp damage monitoring based on acoustic emission (AE) was designed and built to acquire AE signals generated by strain clamps which are respectively under-crimping, over-crimping and with cracks under different experimental conditions. It is found that the power spectrum of the AE signals generated by strain clamps with different crimping states is mainly concentrated in the range of 40 kHz to 100 kHz. However, additional power spectrum peaks are observed near 640 kHz for strain clamps which are with cracks and over-crimping. Additionally, peak value, ring count, energy and RMS can be identified as appropriate characteristic parameters reflecting the damage states of strain clamps through parametric analysis. Combining the backpropagation (BP) neural network algorithm, this study established an identification model for typical strain clamp damages. Through testing AE signal samples from strain clamps which are under-crimping, over-crimping and with cracks, the model achieves identification accuracies of 100%, 70%, and 80%, respectively, with an overall accuracy of 83.33%, thereby verifying the effectiveness of the proposed model.