A Substation Fault Diagnosis Method Based on Substation Alarm Signal Analysis

In the existing substations, when the fault occurs, too many alarm signals will be sent to the substation layer, so that the monitoring personnel cannot find the critical fault information in time. In order to solve this problem, this paper proposes a fault diagnosis method using analyzed alarm signals in substation layer. Firstly, use the string matching algorithm to extract the information of alarm signals and establish signal model and then form the signal information table. Secondly, put the analyzed signals of "accident" and "change-position" level into the device fault causal logic models to determine the failure equipment of the substation. At the same time, if the protection and the circuit breaker incorrect action occur, the inference rules are used to make a preliminary judgment on the cause of the incorrect operation. Finally form a fault brief in substation layer which can assist the monitoring personnel to find faulty equipment and the cause of fault quickly.

I. Introduction

With the completion of intelligent substations and the extensive use of IEDs, the substation collects multiple times of data from various data monitoring systems than before, which increases the difficulty of data processing and filtering while improving data integrity. Especially when a fault occurs, the substation monitoring interface will receive a large number of alarm signals in a short time, so that the monitoring personnel cannot find the key information which reflects the characteristics of the fault in time. If the alarm signals can be filtered and analyzed in the substation layer, substation-level rapid fault diagnosis will be achieved, which avoids sending all the original data to the control center and reduces data traffic significantly. In addition, by making full use of the local rich alarm information of the substation, the uncertainty of diagnosis and alarm can be greatly eliminated, and the direct reason of the fault can be found quickly.