Analysis of robust optimization for decentralized microgrid energy management under uncertainty

The present paper provides an extended analysis of a microgrid energy management framework based on Robust Optimization (RO). Uncertainties in wind power generation and energy consumption are described in the form of Prediction Intervals (PIs), estimated by a Non-dominated Sorting Genetic Algorithm (NSGA-II) - trained Neural Network (NN). The framework is tested and exemplified in a microgrid formed by a middle-size train station (TS) with integrated photovoltaic power production system (PV), an urban wind power plant (WPP) and a surrounding residential district (D). The system is described by Agent-Based Modelling (ABM): each stakeholder is modeled as an individual agent, which aims at a specific goal, either of decreasing its expenses from power purchasing or increasing its revenues from power selling. The aim of this paper is to identify which is the uncertainty level associated to the “extreme” conditions upon which robust management decisions perform better than a microgrid management based on expected values. This work shows how the probability of occurrence of some specific uncertain events, e.g., failures of electrical lines and electricity demand and price peaks, highly conditions the reliability and performance indicators of the microgrid under the two optimization approaches: (i) RO based on the PIs of the uncertain parameters and (ii) optimization based on expected values.