基于集合经验模态分解的增强核岭回归配电系统状态估计

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

针对配电网量测信息存在强非高斯噪声时会大幅干扰基于深度学习的状态估计模型滤波精度的问题,提出了一种基于集合经验模态分解的增强核岭回归状态估计方法。首先,使用集合经验模态分解筛除量测信息中的多数噪声数据,保障了后续滤波对数据可靠性的要求。然后,通过构建增强核岭回归状态估计模型,建立了量测信息与估计残差之间的映射关系,输入量测信息后可以得到估计结果与估计残差。最后,在标准IEEE 33节点与某市78节点系统上进行数值仿真,结果证明了该方法在强非高斯噪声干扰下具有较高的精确性和鲁棒性。

Addressing the significant interference in the filtering accuracy of deep learning-based state estimation models caused by strong non-Gaussian noise in distribution network measurement information,an enhanced kernel ridge regression state estimation method based on ensemble empirical mode decomposition is proposed.Firstly,the ensemble empirical mode decomposition is utilized to filter out most of the noise data in the measurement information,ensuring the reliability of data for subsequent filtering.Subsequently,an enhanced kernel ridge regression state estimation model is constructed to establish a mapping relationship between measurement information and estimation residuals.By inputting the measurement information,the estimation results and estimation residuals can be obtained.Finally,numerical simulations are conducted on the standard IEEE 33-node system and a city-level 78-node system,demonstrating that the proposed method exhibits high accuracy and robustness under the interference of strong non-Gaussian noise.