Dynamic State Transition Model-based Prognosis Approach for Degradation Prediction of Aerial Bundled Cables

Aerial Bundled Cables (ABCs) comprise multiple insulated phase conductors. As different cables are bundled together using a main insulation, the moisture penetrates within the insulation, resulting in degradation initiation when subjected to different loading conditions. Sophisticated degradation prediction techniques based on the current health state of the insulation are required to enable timely maintenance action, leading to failure prevention. This insulation degradation with respect. Time is observed by acquiring thermal images from the installed cables, and the corresponding thermal degradation parameters (TDP) are determined at each measuring instant. A particle filter (PF) based Bayesian framework is used in the proposed study to predict the TDP for future instants. In PF, state transition models play an important role in future predictions. The selection of the state transition model is critical; as different models represent different degradation rates during ABC’s life cycle. Therefore, our paper uses a Weibull distribution-based dynamic state transition model within the proposed framework. Weibull distribution's shape parameter is estimated using the Maximum Likelihood Estimation (MLE) method using historical data at each prediction step. A comparison of prediction accuracy between a prognosis framework based on single (constant) prediction density and dynamic prediction density is also presented in the proposed work. However, our proposed model indicates that the TDP can be predicted with an average accuracy of 0.013, 0.0281 & 0.042 oC/A when using one, two & three-step predictions, respectively.