Smart grid resiliency improvement using a multi-objective optimization approach

Abstract

The enhancement in temperature will increase energy consumption and reduce the power grid’s efficiency. Due to global warming, improving grid resilience under extreme heat is very important. However, there is no metric to measure grid resilience based on temperature. In this research, a novel probabilistic metric for measuring network resilience is introduced in which the probability of temperature occurrence is calculated according to historical temperature data. This metric also depends on the energy remaining in the battery, which is calculated using a neural network. In addition, one of the main steps for power grid expansion is power demand forecasting. Classical models can not accurately estimate the future power demand; therefore, a novel model is introduced based on a twin support vector machine and quantile regression to forecast power demand. A multi-objective optimization problem is presented to minimize investment and operational costs and improve power grid resilience. The performance of the proposed method is investigated using numerical simulations. Results indicate both reliability and resilience are enhanced by optimal sizing and siting of PV panels and batteries. In addition, it is observed that global warming can reduce the resilience of the power grid and increase total investment and operational costs.