LIMITAÇÕES DO ESTIMADOR DE ESTADO DE MÍNIMOS QUADRADOS PONDERADOS: UMA ANÁLISE BASEADA EM ESTUDOS DE CASO
LIMITAÇÕES DO ESTIMADOR DE ESTADO DE MÍNIMOS QUADRADOS PONDERADOS: UMA ANÁLISE BASEADA EM ESTUDOS DE CASO
-
DOI: https://doi.org/10.22533/at.ed.1281125170313
-
Palavras-chave: Estimação de Estado, Erros grosseiros, Mínimos Quadrados Ponderados, Observabilidade, Sistemas Elétricos de Potência.
-
Keywords: State estimation, Bad data, Weighted Least Squares, Observability, Electric Power systems.
-
Abstract: State Estimation is an essential tool for real-time monitoring and operation of Electric Power Systems, providing consistent estimates of state variables from redundant measurements. The Weighted Least Squares method has been consolidated as the classical approach due to its mathematical robustness and statistical foundation; however, its performance may be compromised in the presence of gross errors. This article presents an analysis of the limitations of Weighted Least Squares based State Estimation through three case studies applied to a 5-bus test system. In the first scenario, it is shown that when all measurements become critical, the normalized residual test incorrectly identifies the suspicious measurement, and the removal of this measurement leads to loss of observability. In the second scenario, the Chi-square test fails to detect the presence of a gross error. In the third scenario, although multiple errors are present, both the Chi-square and normalized residual tests correctly identify the faulty measurement; however, removing this measurement compromises observability, making the estimation process infeasible. The results demonstrate that, under different adverse configurations, the classical state estimator exhibits structural limitations that compromise the reliability of the estimation process, highlighting the importance of alternative methods for gross error processing in Electric Power Systems
- Hugo Andrés Ruiz Flórez
- Gloria Patrícia López Sepúlveda
- Victor Brasil Santos Nogueira