Robust Design of Multi-Energy Systems Accounting for Mixed-Integer Operational Problems
Moritz Wedemeyer · 2026
Read Robust Design of Multi-Energy Systems Accounting for Mixed-Integer Operational Problems by Moritz Wedemeyer (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-26613.txt.
Introduction
Wedemeyer, Mitsos, and Dahmen examine a heuristic for robust multi-energy system design that iteratively adds extreme scenarios to the design problem. They show that piecewise-linear inflow-outflow relations, minimum part-loads, and storage complementarity can each break the heuristic when surplus energy cannot be curtailed. With curtailment allowed, storage complementarity stays safe and convex piecewise-linear relations reformulate linearly, but minimum part-loads can still cause failure. In a case study, raising representative scenarios from four to six or eight yields robust designs.
Catalog record
| CREATOR | Moritz Wedemeyer |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71183 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.26613
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Robust Design of Multi-Energy Systems Accounting for Mixed-Integer Operational Problems Moritz Wedemeyer Abstract a Institute of Climate and Energy Systems, Energy Systems Engineering (ICE-1), Forschungszentrum Jülich GmbH, Jülich 52425, Germany b RWTH Aachen University, Aachen 52062, Germany c RWTH Aachen University, Process Systems Engineering (AVT.SVT), Aachen 52074, Germany d JARA-ENERGY, Jülich 52425, Germany Identifying robust designs for multi-energy systems is computationally challenging. As rigorous approaches are often computationally intractable, heuristics are employed to generate candidate designs. One example is the feasibility time-step heuristic by Teichgraeber et al. [Appl. Energy, 275, 115223, 2020] . We theoretically investigate how 3 common nonconvexities, i.e., piecewise-linear energy inflow-outflow relationships, minimal part-loads, and storage complementarity, affect the robustness of designs identified by this heuristic. We find that, if surplus energy cannot be curtailed, any of these nonconvexities may cause the heuristic to fail. If curtailment is allowed, storage complementarity does not compromise robustness, and convex piecewise-linear inflow-outflow relationships can be reformulated linearly. However, minimal part loads may lead to failure of the heuristic. Furthermore, if the objective function of the operational problem is not jointly convex in the uncertain variables and the operational variables, the heuristic may fail. We consider an illustrative multi-energy system, where minimal part-loads and nonconvex dependence of the objective function on heat-pump efficiency are identified as possible failure modes. We propose a hybrid method that discretizes integer variables and embeds the dual of the lower-level problem into a single-level formulation to verify whether a design identified by the heuristic is robust. The results show that the feasibility time-step heuristic may identify robust designs despite the presence of nonconvexities, but the success is case-dependent. Hence, the heuristic can serve as a first step in identifying robust designs; however, when robustness guarantees are required, rigorous methods are necessary. Keywords : Robust optimization, Multi-energy systems, Robust energy system design, Mixed-integer linear programming 1 Introduction When designing multi-energy systems, it is paramount that these are robust, i.e., they can meet any energy demand that they may face during operation. Robustness is an even stronger concern in isolated energy systems, which are not connected to a superordinate grid. The limited number of users and small geographic extent of isolated systems lead to large relative fluctuations in load and renewable generation, necessitating robustness against uncertainties in both demand and renewable supply. Recently, we transferred a robust design approach from process systems engineering ( 12 ) to energy system design ( 29 ) and demonstrated that if the operational problem is convex, the worst-case scenarios lie at the extreme points of the uncertainty set. Consequently, for convex operational problems, the feasibility time-step heuristic proposed by 25 , based on work by 2 , guarantees feasibility for all scenarios within the convex hull of the historical data. Figure 1 illustrates the feasibility time-step heuristic, which is similar to the algorithm by 4 to solve semi-infinite programs (SIPs). In both approaches, a lower-level problem is iteratively checked for feasibility, and violating scenarios are added to the upper-level problem (design problem). Figure 1 : Illustration of the feasibility time-step heuristic ( 25 ) : A candidate design is obtained using a set of representative scenarios \mathcal{S} . Its operational feasibility is evaluated for all historical scenarios (circles) in the historical data \mathcal{D} . If all scenarios are feasible, the design is deemed robust; otherwise, the set of representative scenarios is augmented with one or more infeasible scenarios from the historical data, and the procedure is repeated. For convex operational problems, the feasibility of all historical scenarios guarantees feasibility for the entire interior of their convex hull (shaded region) ( 29 ) . In this approach, design optimization is performed using a set of representative scenarios \mathcal{S} drawn from the available historical data \mathcal{D} to generate candidate designs, whose operational feasibility is subsequently evaluated for all historical scenarios. If infeasible scenarios are identified, the representative scenario set is augmented by adding one or more infeasible scenarios, and a new design candidate is computed. If all historical scenarios are feasible, the design is deemed robust, and the heuristic terminates. For nonconvex operational problems, such as those involving integer decision variables, there is no guarantee that the design identified by the feasibility time-step heuristic is robust and that scenarios within the convex hull of the historical data are feasible ( 29 ) . Nevertheless, already in the original work by 2 , the heuristic was applied to a mixed-integer linear programming energy system synthesis problem introduced by 27 . The design of multi-energy systems is often formulated as a mixed-integer linear program (MILP) ( 22 ; 18 ; 30 ; 16 ; 20 ) . Integer variables are employed to represent on/off decisions or to approximate nonlinear component efficiency curves via piecewise-linear functions. In our previous work ( 29 ) , we showed that minimal part load operation, if combined with a no-curtailment assumption, can cause the feasibility time-step heuristic ( 25 ) to identify designs for which not all realizations in the convex hull of the historical data are feasible. However, the example was minimal, considered only a single timestep, and relied on the assumption that surplus energy could not be curtailed. In the present work, we investigate the types of nonconvexities that typically arise in energy system design problems and assess whether the feasibility time-step heuristic can reliably identify robust designs in the presence of these nonconvexities. The remainder of this work is structured as follows. Section 2 presents the modeling framework for multi-energy system design and discusses the key nonconvexities arising in the operational problem. Their impact on robustness is then analyzed. Section 3 introduces an illustrative case study that demonstrates the impact of nonconvexities. In addition, a method is proposed to verify whether a design candidate is robust for all scenarios within the convex hull of the historical data, and its application is demonstrated using the case study. Finally, Section 4 discusses the results and summarizes the main findings. 2 Problem Structure Throughout the manuscript, scalar-valued quantities are denoted in regular font, e.g., x , vector-valued quantities are denoted in bold font, e.g., \mathbf{x} , and set-valued quantities are denoted in calligraphic font, e.g., \mathcal{X} . We consider energy system design problems of the form: \displaystyle\underset{\mathbf{x}\in\mathcal{X},\mathbf{z}_{s}\in\mathcal{Z}_{s}(\mathbf{x})\forall s\in\mathcal{S}}{\operatorname{min}} \displaystyle\quad c_{inv}(\mathbf{x})+\sum_{s\in\mathcal{S}}c_{op}(\mathbf{z}_{s}) (PS) s.t. \displaystyle\quad\underset{\mathbf{y}\in\mathcal{Y}}{\operatorname{max}}\underset{\mathbf{z}\in\mathcal{Z}(\mathbf{x},\mathbf{y})}{\operatorname{min}}\quad E_{gap}(\mathbf{x},\mathbf{y},\mathbf{z})\leq 0 with feasible sets \displaystyle\mathcal{X}=\{\mathbf{x}\in\mathbb{R}^{n_{x}}|\ \mathbf{g}_{x}(\mathbf{x})\leq\mathbf{0}\ (Design\ constraints)\} \displaystyle\mathcal{Z}_{s}(\mathbf{x})=\{\mathbf{z}_{s}\in\mathbb{R}^{n_{z}}|\ \mathbf{g}_{en}(\mathbf{x},\mathbf{y}_{s},\mathbf{z}_{s})\leq\mathbf{0}\ (Energy\ system\ model)\} \displaystyle\mathcal{Y}=\{\mathbf{y}\in\mathbb{R}^{n_{y}}|\ \mathbf{g}_{y}(\mathbf{y})\leq\mathbf{0}\ (Uncertainty\ bounds)\} \displaystyle\mathcal{Z}(\mathbf{x},\mathbf{y})=\{\mathbf{z}\in\mathbb{R}^{n_{z}}|\ \mathbf{g}_{en}(\mathbf{x},\mathbf{y},\mathbf{z})\leq\mathbf{0}\ (Energy\ system\ model)\}, where \mathbf{x} represents design decisions such as component capacities, \mathbf{y}_{s} are the fixed uncertainty realizations of the representative scenarios s , and \mathbf{z}_{s} represents the corresponding upper-level operational decisions such as component energy outflow. c_{inv} and c_{op} are the investement and operational costs, respectively. \mathbf{g}_{x}(\mathbf{x}) are the design constraints, e.g., piecewise-linear investment curves, and \mathbf{g}_{y}(\mathbf{y}) are the uncertainty bounds, i.e., constraints describing the convex hull of the historical data. \mathbf{g}_{en}(\mathbf{x},\cdot,\cdot)\leq\mathbf{0} are the model equations that describe the behavior of the energy system. \displaystyle\underset{\mathbf{y}\in\mathcal{Y}}{\operatorname{max}}\underset{\mathbf{z}\in\mathcal{Z}(\mathbf{x},\mathbf{y})}{\operatorname{min}}\quad E_{gap}(\mathbf{x},\mathbf{y},\mathbf{z}) (MLP) An embedded optimization problem, which we refer to as the medial-level problem ( MLP ), ensures the feasibility of the design for all possible uncertainty realizations \mathbf{y} , such as wind-power realizations, using operational recourse decisions \mathbf{z} to counteract the uncertainty realization. We refer to the objective function of problem