Metamodels to enable renewable energy system consideration within thermal and electrical systems optimization

Abstract

Emerging infrastructure systems requiring both thermal and electrical energy will be powered by both renewable and conventional energy sources. To consider competitive energy subsystem designs concurrently within the design of a complete infrastructure system, high speed energy subsystem analysis tools and optimization approaches are required. This work demonstrates the use of metamodels for enabling this concurrent design approach. A renewable energy system model is presented for estimating the performance, cost, greenhouse gas emissions, and water consumption of photovoltaic, wind, concentrated solar power, battery, grid, and steam systems for a given location. High-speed metamodel representations of this model are successfully constructed. A few case studies are presented wherein the optimization time and outcomes are compared between the complete model and the metamodels. Both systems identify cheaper, less emitting, and more water efficient systems by deploying renewable energy sources. The metamodels identify competitive systems exceptionally faster than the complete model, four seconds compared to an hour on average, at no detriment to the cost of the identified system. Therefore, metamodels may be used to enable the preliminary concurrent design of emerging infrastructure systems in a computationally manageable way.

Keywords

• Renewable energy system optimization
• Metamodeling
• Infrastructure design
• Surrogate modeling
• Desalination plant design