When analyzing fault waveforms occurred in the railway power system, different RMS computation results are derived depending on analysis software. This is because RMS computation method is applied differently for each software. In order to compare and analyze the calculated values in each software, this paper describes the three methods most commonly used in RMS computation. And it was compared how the calculation results were derived when each computation method was used through the actual field fault waveforms. The conventional RMS computation results are well-affected by signals with exponentially decaying DC offset. RMS computation based on DFT detects RMS value without harmonic components using only fundamental component, therefore it is more robust than other methods in case of signals with exponentially decaying DC offset. RMS computation based on the orthogonality of real and imaginary part of DFT results can rapidly reach steady-state but it fluctuates when signals have harmonic components. Through this paper, power engineers will be able to understand and analyze each of the RMS computational characteristics.

中文翻译：

---

铁路电力系统故障数据分析的均方根计算性能比较

在分析铁路电力系统中发生的故障波形时，根据分析软件得出不同的RMS计算结果。这是因为 RMS 计算方法适用于每个软件。为了比较和分析各个软件中的计算值，本文介绍了 RMS 计算中最常用的三种方法。并通过实际的现场故障波形，比较了使用每种计算方法时的计算结果是如何得出的。传统的 RMS 计算结果受 DC 偏移呈指数衰减的信号影响很大。基于 DFT 的 RMS 计算仅使用基波分量检测没有谐波分量的 RMS 值，因此在具有指数衰减 DC 偏移的信号的情况下，它比其他方法更稳健。基于 DFT 结果实部和虚部正交性的 RMS 计算可以快速达到稳态，但当信号具有谐波分量时，它会波动。通过本文，电力工程师将能够理解和分析每个 RMS 计算特性。