Multi-objective and machine learning strategies for addressing the Water–Energy–Waste nexus in the design of energy systems

Abstract

This paper presents a multi-objective strategy coupled with fuzzy C-means to address synergies and conflicts around the water–energy–waste nexus. The proposal deals with an optimal design and operation scheme in a multi-objective framework where the objective functions are linked to the nexus. The objective functions are normalized to obtain subsets of functions that are used to assess the performance of the nexus and compute trade-off optimal solutions. The soft clustering algorithm is used to determine levels of synergy among Pareto optimal solutions. The proposed strategy has the potential to address problems with many-objective functions, in which the Pareto front has representation limitations, allowing for the identification of conflicts. The soft clustering algorithm indicates different levels of synergy among the Pareto optimal solutions based on the level of membership. The focus of the analysis is on distributed generation systems. As a demonstration, the coupling of a combined heat and power unit with a biodigester and a thermal storage system, interconnected to the local grid. The objective functions, used as descriptors of the nexus, are the use of waste, energy efficiency and water consumption, which are used with the economic performance of the system to configure a problem of many-objective functions.