Combining Multi-Objective Optimization, Principal Component Analysis and Multiple Criteria Decision Making for ecodesign of photovoltaic grid-connected systems

Abstract : Photovoltaic grid-connected systems (PVGCS) promise to be a major contributor of the future global energy system. Even if no GreenHouse Gases (GHG) are emitted during their operation phase, emissions are generated by the use of fossil fuel-based energy during the manufacture, building and recycling of the components. An integrated ecodesign framework that simultaneously manages technical, economic and environmental criteria for the design and sizing of PVGCS (cradle-to-gate approach) is presented in this work. A Multi-Objective Optimization problem embedded in an external multi-objective Genetic Algorithm (NGSA II) optimization loop generates a set of Pareto solutions representing the optimal trade-off between the objectives considered. Then a decision-making tool (M-TOPSIS) selects the solution providing the best compromise. The Life Cycle Assessment (LCA) method was selected to assess the environmental impact. Five commercial PV technologies were evaluated to generate alternatives of PVGCS configurations through a set of 18 objectives (two technical and one economic criteria as well as the 15 midpoint categories of the IMPACT 2002+ method). After a statistical analysis of the first results, the Principal Component Analysis (PCA) method was applied to remove redundant objectives, thus leading to only four contradictory objectives. The results highlight the advantage of the use of thin-film PV modules over crystalline-Si based PV modules.