Accurate detection of zero-value insulators in transmission line porcelain insulator strings remains a challenging problem. The infrared imaging and harmonic electric field methods, which rely heavily on environmental conditions and multidimensional image processing. In this study, a full-scale coupled-field model was developed in COMSOL, to closely approximate the real operating environment of ±800 kV DC systems to detect zero-value. On this basis, a differential spatial electric field criterion was proposed, where the field intensity differences between adjacent insulator segments were extracted as features. These features were then cascaded with a probabilistic neural network (PNN) to enhance detection accuracy. The proposed electric field–based method requires only one-dimensional data for classification, thereby simplifying implementation and improving robustness for practical applications. The results show that the electric field curve at the zero-value insulator exhibits a pronounced “dip”, particularly under wet pollution conditions, while the electric field of adjacent normal insulators increases. Using this approach, the position of zero-value insulators in the string was identified with an accuracy of 96.3 %. This study adopts a simulation-based verification approach and has not yet completed field measurements and cross-calibration. In the future, scaled experiments and UAV-based line application tests will be conducted for validation.
扫码关注我们
求助内容:
应助结果提醒方式:
