Prediction of Static Voltage Stability in a Power System with Large-scale Electric Vehicles

Abstract

This paper takes a classic 3-machine 11-bus system integrated with scaled and disorderly electric vehicles (EVs) as an example, and uses the power system analysis toolbox PSAT to analyze the static voltage stability of the system which appears limit-induced bifurcation (LIB) phenomenon before the saddle-node bifurcation (SNB). Large-scale disorderly charging of EVs will make the load parameter λLIB at the LIB point become smaller, and discharging of large-scale EVs can increase the λLIB. Local compensation capacitor groups dispersedly at the low voltage buses and various types of photovoltaic power plants installed on load region can, in some extent, increase the λLIB, and reduce the adverse effects caused by EVs charging. The support vector machine (SVM) algorithm is used to predict the LIB point in the 3-machine 11-bus system with scaled and disorderly EVs. After training, the SVM optimal model can fleetly and accurately predict the LIB point and is convenient for computing the system’s load margin index according to real-time charging power of EVs. The prediction method based on SVM is beneficial to improve the system static voltage stability, and suitable for on-line application.