Control of TCSC and SVC using Least Square Support Vector Regression (LS-SVR) to improve voltage stability

Abstract

This paper proposes the application of Least Square Support Vector Regression (LS-SVR) for controlling Flexible AC Transmission Systems (FACTS) in order to meet voltage stability requirement. Transient voltage stability is a very fast phenomenon. Therefore, the proposed approach is aimed to provide a quick response to prevent voltage collapse. Generally, time response consists of two parts. Firstly, control center receives signals from the field and then process those signals to determine the appropriate setting of FACTS devices according to load level and location of fault. Secondly, FACTS devices react based on the signals sent by control center to prevent voltage collapse. The total response time should be shorter than the time to voltage collapse. Two kinds of FACTS devices, Thyristor Controlled Series Capacitor (TCSC) and Static VAR Compensator (SVC), are used to represent series and shunt type devices, respectively. To prove the effectiveness of the proposed approach, IEEE 14 buses is used as test system. In addition, comparison study between application of LS-SVR and Extreme Learning Machine (ELM) is also presented.