Distribution System State Estimation Based on Enhanced Kernel Ridge Regression and Ensemble Empirical Mode Decomposition

Abstract

In the case of strong non-Gaussian noise in the measurement information of the distribution network, the strong non-Gaussian noise significantly interferes with the filtering accuracy of the state estimation model based on deep learning. To address this issue, this paper proposes an enhanced kernel ridge regression state estimation method based on ensemble empirical mode decomposition. Initially, ensemble empirical mode decomposition is employed to eliminate most of the noise data in the measurement information, ensuring the reliability of the data for subsequent filtering. Subsequently, the enhanced kernel ridge regression state estimation model is constructed to establish the mapping relationship between the measured data and the estimation residuals. By inputting the measured data, both estimation results and estimation residuals can be obtained. Finally, numerical simulations conducted on the standard IEEE-33 node system and a 78-node system in a specific city demonstrate that the proposed method exhibits high accuracy and robustness in the presence of strong non-Gaussian noise interference.