An ANN based forecast for IED network management using the IEC61850 standard

Abstract

This paper presents the development of a forecast approach to manage IED networks using the IEC 61850 Standard. This is achieved by monitoring a special network occupancy parameter, which properly measures the traffic and behavior of Generic Object Oriented Substation Event (GOOSE) messages in the underlying network bandwidth. A small-scale station bus network was setup in a laboratory to test the proposed strategy. The tests revealed that IED functions may be deteriorated when the network parameter surpasses 5–15% of the available bandwidth, depending on the IED manufacturer and model. A piece of software was created to continuously calculate these parameters in the network, as well as a specially designed Artificial Neural Network, responsible for forecasting the bandwidth occupation. The software monitors and reports any possible changes in the performance of GOOSE messages, which makes it possible to use adaptive strategies in order to ensure the proper operation and performance of the protection and automation schemes.