Data-driven stochastic robust optimization of sustainable utility system

Abstract

Sustainable utility systems reduce reliance on fossil fuels by using renewable energy sources. Multi-scale uncertainties associated with renewable energy and utility systems pose challenges to the modeling and optimization of sustainable utility systems. This study proposes a sustainable retrofit framework for utility systems based on a data-driven stochastic robust optimization approach. Kernel density estimation and fuzzy clustering were employed to capture the uncertainty features in a holistic framework. A life cycle assessment approach was used to calculate the global warming potential (GWP) of the sustainable utility system, and a multi-objective environmental and economic optimization model was developed. A nested decomposition-based algorithm was proposed to solve a large-scale mixed-integer nonlinear programming problem. Finally, a case study of an industrial utility system was conducted to demonstrate the effectiveness of the proposed method. The optimization results show that the proposed method reduces GWP by 9 % after introducing renewable energy and achieves a balance between economic and environmental performance.