基于递归型定频采样变数据窗离散傅氏算法的大型旋转电机启机保护

Starting protection of a large rotary electric machine based on a recursive fixed frequency sampling variable data window discrete Fourier transform algorithm

介绍了大型旋转电机在变频工况下的启机保护。针对现有启机保护存在抗干扰能力不足和计算量偏大的问题,提出一种基于递归型定频采样变数据窗离散傅氏算法。该算法在保证计算精度的前提下,大幅减少计算量。首先根据系统频率,确定数据窗长,然后采用递归型离散傅氏算法计算模拟量幅值。当经过一个固定时间或频率变化后,插入一次非递归全波傅氏校验,消除累计误差。在此基础上,以变频启动工况下的调相机启机保护为例,将新算法与常规峰值算法相结合,对启机保护进行改进。在保护装置上对改进后的调相机启机保护进行了实验验证。结果表明在不同系统频率下保护均能够准确计算故障中的基波分量,且递归型变数据窗傅氏算法与峰值算法相结合能够确保启机保护的可靠性。

The starting protection of large-scale rotating electrical machines under frequency conversion conditions is introduced.There are problems of insufficient anti-interference ability and large amount of calculation in the existing starting protection.Thus a recursive fixed frequency sampling variable data window discrete Fourier transform(FFSVDW-DFT)algorithm is proposed.The algorithm greatly reduces the amount of calculation but with calculation accuracy.First,the precise one-cycle data window length is calculated according to the system frequency,and then the recursive discrete Fourier transform algorithm is applied to calculate the magnitude of the phasor.When a fixed time elapses or the frequency changes,a non-recursive full-wave Fourier check is inserted to eliminate accumulated errors.Then,taking synchronous condenser starting protection under frequency conversion conditions as an example,the proposed recursive FFSVDW-DFT algorithm and the conventional peak algorithm are combined to improve starting protection.The improved protection is experimentally verified on the protection device.Experimental results show that the protection can accurately calculate the fundamental wave component in the fault at different system frequencies.In addition,the combination of the recursive FFSVDW-DFT and the peak algorithms can ensure the reliability of starting protection.