基于核函数的PCA-KSICA储能系统典型工况在线分析方法

PCA-KSICA Online Analysis Method of Energy Storage System Typical Operating Conditions Based on Kernel Function

考虑储能系统运行过程复杂的内在机制以及实时监控的要求,针对常规动态过程在线监控方法在特征提取和变量识别中的难点问题,结合PCA策略与KSICA方法,对储能系统测量数据进行PCA白化预处理,抵消各通道测量数据间的二阶相关性,建立核函数,提升快速性,有效解决ICA分离算法由于数据量大、迭代次数多引起收敛速度慢的问题。基于核函数的PCA-KSICA储能系统运行工况在线分析方法,不仅可实现复值域独立成分分析,完成时间空间非平稳状况下源信号的在线估计,且可在仅高斯变量情形下保持鲁棒性能,避免产生发散行为,将该方法用于储能系统工况的非线性多元特征提取,分析性能指标,在线运行结果验证了所提出方法的有效性和可行性。

This paper proposes a PCA-KSICA online analysis method of energy storage system typical operating conditions based on kernel function. By combining the PCA with the KSICA, the two order correlation between the original data is eliminated by PCA whitening process, and by building kernel function, the slow convergence speed of the ICA algorithm is effectively solved. KSICA is a theoretical analysis of a certain criterion for complex-valued independent component analysis(ICA) with a focus on blind speech extraction(BSE) of a spatio–temporally nonstationary speech source. The proposed method is shown to share the important fixed-point feature with the Fast ICA method, although an improvement with the proposed method is that it does not exhibit the divergent behavior for a mixture of Gaussian-only sources that the Fast ICA method tends to do, and it shows robust performance in online implementations. It is used in the feature extraction of the energy storage system, the results verify the effectiveness and feasibility of the proposed method.