基于时序建模的光纤电流互感器随机噪声卡尔曼滤波方法

发布时间：2018-04-29 07:24

  本文选题：随机噪声 + 测量精确度　； 参考：《电机与控制学报》2017年04期

【摘要】：针对光纤电流互感器(FOCT)随机噪声特性及其对继电保护、电能计量等间隔层设备的影响,建立FOCT随机误差的时序模型,并采用滤波方法有效提高了FOCT测量精确度。首先,预处理和统计检验FOCT原始数据,获取数据随机特征;根据赤池信息准则(AIC)准则选择时间序列模型的阶次,求出模型系数建立FOCT随机误差的ARMA(2,1)模型,并检验其适用性;采用卡尔曼滤波方法对FOCT输出数据进行滤波处理。总方差分析结果表明:建立的FOCT时序模型经卡尔曼滤波后,随机噪声幅值明显减小,方差值降低了两个数量级,各项随机噪声的误差系数均下降一个数量级,采用的时序建模和卡尔曼滤波方法能有效减小FOCT的随机噪声,提高电流信息的测量精确度。
[Abstract]:Aiming at the random noise characteristics of optical fiber current transformer (FOCTT) and its influence on relay protection, electric energy metering and other spacer equipment, the time series model of FOCT random error is established, and the accuracy of FOCT measurement is improved effectively by using filtering method. Firstly, we preprocess and statistically test the original FOCT data to obtain the random characteristics of the data, select the order of the time series model according to the red pool information criterion and calculate the model coefficients to establish the ARMA-2Q1) model of the FOCT random error, and test its applicability. Kalman filtering method is used to filter the FOCT output data. The results of total variance analysis show that the amplitude of random noise is obviously reduced, the square difference is reduced by two orders of magnitude, and the error coefficient of each random noise is reduced by one order of magnitude after Kalman filter. The time series modeling and Kalman filtering method can effectively reduce the random noise of FOCT and improve the measurement accuracy of current information.
【作者单位】： 云南电网有限责任公司电力科研究院;中国南方电网公司电能计量重点实验室;东南大学仪器科学与工程学院;
【基金】：南方电网科技项目(YNKJ0000124) 云南电网科技项目(HLZB20150738)
【分类号】：TM452.94
，

本文编号：1818977