Fuzzy logic applied to registration of alarms and events in substations with IEC 61850

When fault occur in electric power systems, protection and automation schemes take all necessary actions to properly isolate the problem, assuming a safe or emergency operating state. On these occasions, human operators should review all messages, events and alarms from SCADA to understand the underlying situation. Only after this analysis, based on previous experience and knowledge, the operators may order the system restoration to a normal operating state. The data to be understood are usually conflicting, confusing and numerous, since they were sequentially triggered by many devices, resulting in an avalanche of information. Due to these characteristics a full human-based analysis is difficult, slow and prone to failure. This paper addresses a Fuzzy logic based diagnosis system, which can quickly interpret the avalanche of data resulting from an undergoing fault situation, to properly access, assist and ease the human-based operation of the power system. The system used a novel scheme, with fully compliant IEC 61850 communication scheme, solely based on GOOSE and MMS messages, to acquire data from all substation equipment, as long as a modern programmable controller, to execute all artificial intelligence routines needed.

I. Introduction

The demand for more energy, without increasing the capacity to generate and transmit power, causes the electrical systems to work at extreme limits, reducing reliability, quality, and affecting the power supply continuity to costumers. In this context, to achieve greater energy availability, one of the few solutions is to promote the integration and automation of the whole electrical system, from its generation, transmission, distribution, up to the customers, in an intelligent and standardized way, allowing better resource management, reduced losses, and lower costs.