An Approach to Steady‐State Temperature Prediction of Open‐Type Disconnector Contacts

ABSTRACT

Open‐type disconnectors ensure safe power grid operation, yet predicting their steady‐state contact temperature rise remains limited. This study systematically investigated the steady‐state temperature prediction of disconnector contacts through experimental platform development, data processing and modelling analysis. First, a dedicated experimental platform was constructed to measure the influence of six key factors: current, contact resistance, ambient temperature, humidity, wind speed and irradiance, on contact temperature. Subsequently, a preprocessing framework combining the isolation forest (IF) and synthetic minority oversampling technique (SMOTE) was introduced, and a support vector regression (SVR) prediction model optimised by an improved whale optimisation algorithm (IWOA) was developed—with these factors as inputs and steady‐state contact temperature as the output. Experimental results under lab‐scale controlled conditions showed that the IWOA‐SVR model achieved average performance metrics of MSE = 0.519, MAE = 0.401, MAPE = 1.11% and R ²  = 0.982. Compared with the best‐performing benchmark model, the proposed approach exhibited clear improvements, particularly in high‐temperature scenarios where the maximum relative error was only 2.55%, significantly lower than those of other models. These findings demonstrate the effectiveness of the proposed method within the experimental scope and provide valuable foundational insights for the overheating risk assessment of disconnector contacts in substations.