Multiple design options for sizing off-grid microgrids: A novel single-objective approach to support multi-criteria decision making

Abstract

When designing a microgrid, developers usually regard economic metrics, and occasionally consider reliability and environmental aspects. However, sociopolitical, supply chain, and geographical facets, among others, are often never included in project-design because they are really difficult to model, especially, in the context of developing countries. Traditional planning methodologies offer an optimal solution, disregarding solutions with similar profitability but different size of components, even when these second-best solutions can better fit the non-considered intangible developer needs. In this paper, we define the concept of Multiple Design Options (MDO) for a single-objective optimization. We propose a novel methodology (MDO-PSO) for sizing stand-alone hybrid energy systems that, by using Particle Swarm Optimization, identifies the optimal solution and post-processes the search history to select second-best options of interest. While searching for the traditional optimum, the proposed iterative algorithm stores all tried configurations. Moreover, a Pareto-like frontier, denoted as MDO-Pareto, is proposed to highlight the tradeoff between Net Present Cost (NPC) and CAPEX. The proposed Pareto-like frontier is also compared to a standard multi-objective optimization to illustrate how MDO-PSO successfully captures multiple goals. The numerical case study for a PV-battery-diesel-tank system in Uganda confirms that MDOs differing up to 32% in capacity can achieve NPC values within 2%–5% optimality with both load following and predictive MILP rolling-horizon operating strategies, thus suggesting important implications for business decision making.