A Forecasting-aided State Estimator Based on Enhanced Autoformer-UPF for Active Distribution Networks

In response to the issue of incomplete measurement configurations in active distribution networks, which leads to limited accuracy in traditional state estimation and subsequently results in decreased estimation precision, this paper introduces a Forecasting-aided state estimator for active distribution networks based on an enhanced Autoformer-UPF framework. The proposed Forecasting-aided State Estimator employs a predictive model based on Autoformer, and utilizes the Whale Optimization Algorithm (WOA) for hyperparameter optimization to enhance estimation accuracy. Leveraging its multi-head self-attention mechanism, this model significantly improves prediction accuracy for long time series data, consequently enhancing the precision of state estimation. Simulation results on the IEEE 39-bus standard system demonstrate the effectiveness and robustness of the proposed method under both normal operating conditions and scenarios with presence of bad data.