基于实测数据的牵引供电系统并联谐波谐振概率识别方法  

Parallel Harmonic Resonance Probability Identification of Traction Power Supply System Based on Measured Data

作　　者：何逢广 李朝阳[1] 张华赢 艾磊 胡海涛[1] HE Fengguang;LI Zhaoyang;ZHANG Huaying;AI Lei;HU Haitao(School of Electrical Engineering,Southwest Jiaotong University,Chengdu 611756,Sichuan Province,China;New Smart City High-quality Power Supply Joint Laboratory of China Southern Power Grid(Shenzhen Power Supply Co.,Ltd.),Shenzhen 518020,Guangdong Province,China)

机构地区：[1]西南交通大学电气工程学院,四川省成都市611756 [2]南方电网公司新型智慧城市高品质供电联合实验室(深圳供电局有限公司),广东省深圳市518020

出　　处：《电网技术》2024年第5期2084-2094,I0073,I0074,共13页Power System Technology

基　　金：南方电网公司科技项目(090000KK52220014/SZKJXM20220021)。

摘　　要：当前,我国高速铁路电力机车广泛采用脉冲宽度调制的四象限变流器,在其性能改善的同时也带来了频带更宽的高次谐波问题,更易激发牵引供电系统谐波谐振。谐波谐振可能会引起过电压/过电流,严重时引发设备爆炸、保护误动等安全事故。因此,准确识别系统潜在的谐振频率具有重要意义。目前常用的频率扫描分析、谐振模态分析等方法依赖于系统拓扑和模型的精确性,而基于实测数据的传统识别方法的准确性受系统参数波动和背景谐波影响较大。为此,提出了一种基于实测数据的牵引供电系统并联谐振概率识别方法,以电压、电流波形的余弦相似度作为谐振预判条件以减少无效计算,并综合基波/谐波功率因数、系统谐波阻抗幅值等多层次谐振特征进行概率分析,克服了系统参数波动和背景谐波影响。仿真研究、实测数据分析与已有研究方法对比,表明所提方法在负荷波动性强、背景谐波较大时仍然适用,且识别的准确性更高,验证了该方法的有效性和可靠性。Pulse width modulated four-quadrant converter is widely used in the high-speed railway electric locomotives in China at present.While its performance is improved,it also brings higher harmonics with wider frequency bands,which is more likely to stimulate harmonic resonance in the traction power supply system.The overvoltage or overcurrent caused by the harmonic resonance may in turn lead to equipment explosion,protection mis-operation and other safety accidents in severe cases.Therefore,it is of great significance to accurately identify the potential resonant frequency of the system.The commonly used methods,such as the frequency scanning analysis and the resonant mode analysis,depend on the accuracy of the system topology and the models.However,the accuracy of those traditional identification methods based on measured data is greatly affected by system parameter fluctuations and background harmonics.Therefore,this paper proposes a parallel resonance probability identification method for a traction power supply system based on measured data.The cosine similarity of the voltage and current waveforms is used as the resonance prediction condition to reduce invalid calculations.The probability analysis of multi-level resonance characteristics like the fundamental/harmonic power factor and the system harmonic impedance amplitude overcomes the influence of the system parameter fluctuation and background harmonics.Simulation research,measured data analysis and comparison with the existing research methods show that the proposed method is applicable even when there are strong load fluctuation and large background harmonics,and that the recognition accuracy is higher,which verifies the effectiveness and reliability of the proposed method.

关 键 词：牵引供电系统 并联谐波谐振 复独立分量分析法 快速傅里叶变换 谐振识别 

分 类 号：TM721[电气工程—电力系统及自动化]