基于SVM的直流电能质量扰动分类算法

Classification algorithm of DC power quality disturbance based on SVM

近年来分布式能源发电的规模逐渐扩大,更具优势的直流供电技术也随之飞速发展,未来电网的主要模式将大概率趋向交流电网和直流电网混联的电网结构。其中交流电能质量问题已逐渐成熟,而相应的直流部分仍有待研究。对直流电能质量扰动波形进行有效区分与准确辨识,是研究直流电能质量必不可少的环节。直流电能质量扰动波形分类算法研究对直流配电网的实际工程推广应用具有重大意义。文中通过搭建交直流仿真系统,模拟各类直流电能质量问题并监测异常波形,对不同扰动源造成电能质量扰动问题波形进行了特征分析。通过训练二元树结构的支持向量机辨识扰动波形的类型。仿真算例验证了该方法的可行性与高准确率。

In recent years,with the rapid development of distributed energy generation,the DC power supply gradually shows the advantages of technology,and the main mode of power grid will tend to be the hybrid structure of AC/DC power grid in the future.The problem of AC power quality has been gradually mature,while the corresponding DC part is still to be studied.It is an essential step to classify and accurately identify the disturbance waveform of DC energy quality.The research on classification algorithm of DC power quality disturbance waveform is of great significance to the practical engineering application of DC distribution network.In this paper,the AC/DC simulation system is built to simulate various DC power quality problems and monitor abnormal waveforms.The characteristics of the waveforms of power quality disturbance caused by different disturbance sources are analyzed.The type of disturbance waveform is identified by training the support vector machine of binary tree structure.The simulation example verifies the feasibility and high accuracy of the proposed method.