随着船舶电力系统中接入越来越多的电力电子设备，谐波污染现象变得愈发严重。当前的谐波检测还无法实现较小的谐波频率、相位、幅值检测误差，提出小波包和神经网络相结合的船舶电力系统谐波检测方法。设计由微处理器、信号采集板、上位机构成的船舶电力系统信号采样装置，实施电力系统电流信号与电压信号的采样。通过小波包算法提取采集信号高频部分的有效值。基于神经网络思想设计Elman神经网络谐波检测器，实现船舶电力系统谐波检测。其中在输出层中通过主成分分析方法实施神经网络的输出优化，并实施Sigmoid激励函数的改进，以降低检测误差。测试结果表明，该方法平均谐波频率检测误差、平均谐波相位检测误差、平均谐波幅值检测误差的区间均值分别为0.16 Hz，0.20°，0.12 V，整体误差很低，同时克服了基波分离延时问题。
As more and more power electronic devices are connected to the ship power system, the harmonic pollution is becoming more and more serious. The current harmonic detection cannot achieve smaller harmonic frequency, phase and amplitude detection errors, so a harmonic detection method of ship power system combining wavelet packet and neural network is designed. A signal sampling device for marine power system composed of microprocessor, signal acquisition board and upper computer is designed to sample current signal and voltage signal of power system. The effective value of the high frequency part of the collected signal is extracted through the wavelet packet algorithm. Based on the idea of neural network, Elman neural network harmonic detector is designed to realize the harmonic detection of ship power system. In the output layer, the principal component analysis method is used to optimize the output of the neural network, and the sigmoid excitation function is improved to reduce the detection error. The test results show that the interval mean values of average harmonic frequency detection error, average harmonic phase detection error and average harmonic amplitude detection error are 0.16 Hz, 0.20 ° and 0.12 V, respectively. The overall error is very low, and the problem of fundamental wave separation delay is overcome.
2022,44(19): 110-113 收稿日期：2022-05-30
DOI：10.3404/j.issn.1672-7649.2022.19.021
分类号：U664.8
作者简介：王贺(1989-)，男，硕士，工程师，研究方向为电力系统接地与保护等
参考文献：
[1] 周祎隆, 傅晓红, 夏骏. 22000 TEU集装箱船电力系统谐波抑制方案[J]. 船海工程, 2020, 49(3): 5–9
[2] 赵禹灿, 姜旭, 孙福寿, 等. 基于谐波特性的光伏孤岛与低电压穿越同步检测方法[J]. 电力系统保护与控制, 2022, 50(3): 41–50.
[3] 陈倩, 王维庆, 王海云, 等. 基于混合卷积窗四谱线插值和改进全相位的谐波检测组合优化算法[J]. 电力系统及其自动化学报, 2020, 32(8): 1–6.
[4] 李金, 张喜铭, 时伯年, 等. 一种基于瞬时无功功率理论的改进谐波检测算法[J]. 电气传动, 2020, 50(3): 92–96+106.
[5] 王恒利, 高山, 季圣贤. 一种三相二重化串联型逆变器谐波分析[J]. 海军工程大学学报, 2022, 34(4): 7–12+19.
[6] 黄星宇, 罗萍萍, 龚锦霞, 等. 新能源并网系统次同步谐波相量检测方法[J]. 电力系统保护与控制, 2020, 48(13): 38–44.
[7] 田书娅, 贾清泉, 林丽娟, 等. 面向谐波和电压综合治理的电压检测型APF与SVG协同配置[J]. 电力系统自动化, 2021, 45(24): 149–157.
[8] 李振华, 胡廷和, 杜亚伟, 等. 一种优化窗函数及其在电网谐波检测中的应用分析[J]. 高压电器, 2020, 56(10): 239–246+252.
