Análise e simulação da reamostragem de sinais: uma abordagem baseada na IEC 61850-9-2

• Autor(a) principal: Rosanelli, Adrion Andrei
• Data de Publicação: 2024
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Manancial - Repositório Digital da UFSM
• Texto Completo: http://repositorio.ufsm.br/handle/1/31791
• topic: Reamostragem; Simulação; Processamento digital de sinais; IEC 61850-9-2; Sampled values; Resampling; Simulation; Digital signal processing; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA

Signal resampling is a signal processing technique with several applications, among them, to improve Sampled Values (SV) signals defined by IEC-61850-9-2, which allows the analysis of SV signals in a more accurate and robust way for applications such as the power substations protection. In this sense, this work aims to analyze and compare different resampling schemes, to validate and explore the feasibility of their use. For this, schemes proposed by the literature and using conventional methods implemented in simulation are used. The comparison of these schemes, together with validation of the resampling technique, are performed in simulations with different SV signals and different implementations, considering implementations of the signal directly on the network or the signal as a whole, and evaluating effects such as jitter and random packets loss. To this end, analyzes are carried out in the time and frequency domains (by FFT) and measurements such as RMS and computational time. In these analyses, it was observed that resampling allows to reduce the spectral leakage of SV signals by keeping the relationship between the signal frequency and sampling frequency constant. This also allows a significant reduction in the RMS error, from the order of 10−2A without resampling to the order of 10−7A in some schemes, the most stable in this analysis being the method of “Cubic Spline”. When analyzing the signal effects, it was observed that jitter and random packet loss do not represent significant effects in resampling. Considering the implementation of the SV directly on the network, again the “Cubic Spline” obtained the best performance in the RMS error, and all schemes provided greater stability in the FFT peak and in the RMS, although it results in a phase delay in the fundamental component of the signal. Another analysis developed is computational efficiency, comparing the time of each scheme, in which the best performance was with conventional methods and the worst with “Modified Sinc”. The last simulation carried out was a case study evaluating resampling schemes in SV signals captured by MUs from four manufacturers, in which it was verified that the benefits of resampling are more impactful in manufacturers 1 and 4, being reduced in other cases, and maintaining a phase delay in all situations. With these analyzes it is verified that resampling results in greater precision and robustness of SV signals but adds a processing and phase delay. Among the schemes analyzed, the one that performed best was the “Cubic Spline”, making resampling an attractive technique in some SV applications.