Study on a real time model predictive based power system SIPS

A Model Predictive based system integrity protection scheme (SIPS) using a Thyristor Controlled Series Compensator (TCSC) to improve power system transient stability will be described in the paper. Supervised learning (SL) is utilized to predict the power system dynamics by assuming each control action has been taken. Furthermore, Monte-Carlo approaches are used to analyze the generators' separation pattern and determine the location to apply the TCSC control. The SIPS is operated in real time and based on PMU measurements. The detection and control processes for unstable conditions are performed every discrete time interval, so the control solution including the amount and locations of TCSC control is refreshed successively. The proposed SIPS has been tested and verified in IEEE 39-bus system. Simulation results show it can effectively stabilize the system after a large disturbance even when the system conditions are complex. Moreover, the testing experience in the multimachine system has enhanced the operational confidence in the use of this type of SIPS.