Uncertain Scenario Based MicroGrid Optimization via Hybrid Levy Particle Swarm Variable Neighborhood Search Optimization (HL_PS_VNSO)

Dharmesh Dabhi,Kartik Pandya

doi:10.1109/access.2020.2999935

Abstract

Within the MicroGrid environment, the Energy Resource Management (ERM) problem becomes highly complex due to the uncertainty related to the Renewable Generation (RG) such as Photovoltaic power generation (PV), Electric Vehicle (EV) trip with Grid to Vehicle (G2V) or Vehicle to Grid (V2G), Energy Market price and load demand with Demand Response (DR) programs. Each of these issues should be tackled while optimizing revenues and reducing the running costs of Virtual Power Player (VPP) that collects each of these types of energy resources from the MicroGrid. This article presents a new hybrid optimization algorithm called “Hybrid Levy Particle Swarm Variable Neighborhood Search Optimization” (HL_PS_VNSO) to solve the ERM problem. Its key aspect is the hybridization of the Particle Swarm Optimization (PSO) and the Variable Neighborhood Search Optimization (VNS) algorithm with the enhanced step length using Levy Flight. The effectiveness of the proposed approach is measured by a 25-bus MicroGrid with 500 uncertain scenarios of the aforementioned uncertainty. The results of HL_PS_VNSO are compared with the most advanced optimization algorithms. The findings show that HL_PS_VNSO's results are superior for the Average Ranking Index (A.R.I) and Ranking Index (R.I). For effective comparative analysis of algorithms, the traditional statistical method called One-way ANOVA Tukey Analysis is used. The results from this analysis also prove the superiority of HL_PS_VNSO over the most advanced optimization algorithms.

Highlights

In the MicroGrid context, the growing emergence of uncertain Distributed Energy Resources (DERs) such as solar PV, Demand Response (DR) programs, Electric Vehicles with G2V or V2G feature, Energy Storage Systems (ESSs) and the uncertain market price of energy are challenging the functioning of distribution networks
The GECAD group in collaboration with Delft University organized the competition [35] on Energy Resource Management (ERM) problem in a microgrid with 100 uncertain scenarios at IEEE-WCCI/CEC conference, 2018
To verify the efficacy and robustness of algorithms, for all competing algorithms we find the 20 final solutions for the ERM problem

Summary

INTRODUCTION

In the MicroGrid context, the growing emergence of uncertain Distributed Energy Resources (DERs) such as solar PV, Demand Response (DR) programs, Electric Vehicles with G2V or V2G feature, Energy Storage Systems (ESSs) and the uncertain market price of energy are challenging the functioning of distribution networks. The GECAD group in collaboration with Delft University organized the competition [35] on ERM problem in a microgrid with 100 uncertain scenarios at IEEE-WCCI/CEC conference, 2018 In this competition, Variable Neighborhood Search Optimization- Differential Evolutionary Particle Swarm Optimization (VNS-DEEPSO), Enhanced Velocity Differential Evolutionary Particle Swarm Optimization (EVDEPSO) and PSO with Global Best Perturbation (PSO-GBP)/Chaotic Evolutionary Particle Swarm Optimization Algorithm (CEPSO) got the first, second and third rank respectively. 2. The HL_PS_VNSO algorithm has the near-optimal solution for a model developed by GECAD group for solving the day ahead energy resource scheduling problems in a MicroGrid with high uncertainty owing to Solar PV generation, load forecasting, EVs trip and Electricity Market price. For more details about the constraints please refer [39]

MODELING OF UNCERTAINTY

TEST CASE AND RESULT ANALYSIS

CONCLUSION