Gross error processing in state estimation: Comparing the residual and the error tests

Although being the most used method for detection and identification of gross errors in measurements in the state estimation process, the largest normalized residual test fails in determined situations. Aware of this fact, the authors of this paper aim to clarify the reasons for these failures and to present an alternative method for processing gross errors, based on the geometrical interpretation of the errors and the innovation index of the measurements. It will be shown in the simulations that the largest normalized residual test often fails if the gross error is present in a measurement with low innovation index, however this drawback does not occur when using the geometrical methodology. Both methods are applied in the gross error processing in different error scenarios, single and multiple, in order to make a comparison between their performances. The simulations were performed in the IEEE 14-bus test system.