A Time-series-based approach for robust design of multi-energy systems with energy storage

Abstract

This work proposes a mixed-integer linear program approach to consider the uncertainty of input data in the optimal design of distributed multi-energy systems involving both conventional and renewable-based conversion technologies, as well as storage units. The design procedure determines the minimum-cost combination of technology selection, size and operation. Traditionally, distributed multi-energy systems are designed using deterministic optimization methods, implying that the input data are known when the system optimization is performed. However, such input data are commonly affected by significant uncertainty, making the deterministic solution possibly suboptimal or even unfeasible. Recently, both robust and stochastic optimization have been applied to the optimal design of multi-energy systems. Nevertheless, when including energy storage in the analysis, the traditional techniques are complicated by the short- and long-term evolution of the input data of the underlying optimization problem, as well as their complex interactions. Moreover, the analysis of the uncertainties characterizing such input data for the optimal design of multi-energy systems, as well as the evaluation of their impact on the system design, have been investigated in little details. The approach proposed in this work is based on the analysis of the historical time-series representing the input data of the mixed-integer linear program for different years. First, the most important input data in terms of optimality and robustness of the system design are identified. Moreover, the most relevant features of the corresponding time-series are determined and assessed. Then, this information is used to build a custom set of input data which translates into a system design able to guarantee both security of supply and cost optimality.