A framework for the optimal integration of solar assisted district heating in different urban sized communities: A robust machine learning approach incorporating global sensitivity analysis

Abstract

A promising pathway towards sustainable transaction to clean energy production lies in the adoption of solar assisted district heating systems (SDHS). However, SDHS technical barriers during their design and operation phases, combined with their economic limitation, promote a high variation in quantifying SDHS benefits over their lifetime. This study proposes a complete multi-objective optimization framework using a robust machine learning approach to inherent sustainability principles in the design of SDHS. Moreover, the framework investigates the uncertainty in the context of SDHS design, in which the Global Sensitivity Analysis (GSA) is combined with the heuristics optimization approach. The framework application is illustrated through a case study for the optimal integration of SHDS at different urban community sizes (10, 25, 50, and 100 buildings) located in Madrid. The results reveal a substantial improvement in economic and environmental benefits for deploying SDHS, especially with including the seasonal storage tank (SST) construction properties in the optimization problem, and it can achieve a payback period up to 13.7 years. In addition, the solar fraction of the optimized SDHS never falls below 82.1% for the investigated community sizes with an efficiency above 69.5% for the SST. Finally, the GSA indicates the SST investment cost and its relevant construction materials, are primarily responsible for the variability in the optimal system feasibility. The proposed framework can provide a good starting point to solve the enormous computational expenses drawbacks associated with the heuristics optimization approach. Furthermore, it can function as a decision support tool to fulfill the European Union energy targets regarding clean energy production.