基于奇异值分解信号提取的电力电子变压器能耗模型在线辨识方法

Online Identification Method of Power Electronic Transformer Energy Consumption Model Based on Singular Value Decomposition Signal Extraction

电力电子变压器端口能耗模型的非线性和多模态特性使得损耗建模和解算非常困难,并且实测端口的数据量受到信号干扰也会导致建模不够精确。为此,提出了一种基于奇异值分解的信号特征提取方法,可以将隐含在噪声中的特征信号提取出来,并对提取后的数据进行能耗模型的在线参数识别。首先分析了电力电子变压器的能耗模型,接着对实测的含干扰数据进行奇异值分解信号提取,通过试验验证了所提出的模型的有效性。对能耗模型的参数识别,可以为电力电子变压器运行效率优化提供参考。

The nonlinear and multi-modal characteristics of the power consumption model of the power electronic transformer port make it very difficult to model and solve the loss,and the amount of data measured at the port is disturbed by the signal,which will lead to inaccurate modeling.To this end,a signal feature extraction method based on singular value decomposition was proposed,which can extract the feature signal hidden in the noise,and perform online parameter identification of the energy consumption model on the extracted data.First analyze the energy consumption model of the power electronic transformer,then extract the signal from the singular value decomposition of the measured data with interference,and verify the validity of the proposed model through experiments.The parameter identification of the energy consumption model can provide a reference for the optimization of the operation efficiency of power electronic transformers.