Early warning of the abnormal response of the generator set electrical control based on SIP concept

Abstract

An early warning method based on Session Initiation Protocol (SIP) concept for generator set electrical control abnormal response is proposed. The first is to clean the electrical data of the generator set by SIP. Using Supervisory Control and Data Acquisition (SCADA) and Generalized Linear Models (GLM) algorithm, a linear model is constructed to analyze the electrical control of generating units. Using the Autoregressive Integrated Moving Average Model (ARIMA), an abnormal response early warning model for electrical control of generating units is established. BP neural network is used to train the abnormal response data of the generator set electrical control. According to the current response data, the model prediction is realized, and the early warning of the abnormal response of the generator set electrical control is effectively realized. The simulation results show that the proposed method can effectively reduce the early-warning error, false alarm rate, and early-warning delay of generator electrical control abnormal response.