engBaader, Florian JosephBardow, AndréMitsos, AlexanderSimultaneous real-time scheduling of multi-energy systems and dynamic production processesinfo:eu-repo/classification/ddc/620demand responseintegration of scheduling and controlmixed-integer dynamic optimizationmixed-integer linear programmingsimultaneous schedulingVolatile renewable electricity production causes a temporal mismatch of supply and demand, which can be reduced by consumers using optimal production scheduling to shift their demand in time. However, if production processes with nonlinear dynamics must be scheduled simultaneously with multi-energy systems (MESs) introducing discrete on/off decisions, the resulting optimization problems are typically not solvable in real-time today. This thesis reformulates simultaneous dynamic scheduling (SDS) problems of MESs and nonlinear production processes to mixed-integer linear programs (MILPs), which can be solved in real-time. These MILP reformulations rely on tailored scheduling models consisting of three piece-wise affine (PWA) parts: (1) process output models, (2) data-driven process energy demand models, and (3) MILP energy system models. For part 1, we present two alternatives: First, we use a scale-bridging model (SBM), which is easy-to-apply but requires heuristic tuning. We use two cooled continuous stirred tank reactor (CSTR) case studies to show that our approach can capture the major part of the nonlinear potential in real time, and a heated distillation column case study to show that our approach can reduce the number of states substantially. Second, as a rigorous alternative to the heuristically tuned SBM, we derive dynamic ramping constraints (DRCs), which are first restricted to flat processes with only one scheduling relevant variable. These DRCs consider linear dynamics of high order with PWA constraints. We use that, for flat processes, a nonlinear model can be transferred to a linear model by coordinate transformation. Again, we use a CSTR case study to show that our approach can capture the major part of the nonlinear potential. For non-flat processes, we develop heuristic DRCs, based on simulation experiments. For the heated distillation column, these heuristic DRCs perform similarly to SBMs regarding both optimization runtime and operational costs. Lastly, we show that DRCs can also consider multiple scheduling-relevant variables by applying DRCs to an electrolyzer with slow temperature dynamics. We outperform a quasi-steady-state scheduling through optimized temperature dynamics. This thesis thus offers reformulations that strike reasonable compromises between optimization runtime and solution quality for SDS of production processes and MESs. While SBMs can be applied with less effort and knowledge, DRCs offer more dynamic flexibility for flat processes.Aachen : RWTH Aachen University 1 Online-Ressource : Illustrationen, Diagramme (2023). doi:10.18154/RWTH-2023-07390 = Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2023info:eu-repo/semantics/doctoralThesisinfo:eu-repo/semantics/publishedVersionRWTH Aachen University2023info:eu-repo/semantics/openAccessDEhttps://publications.rwth-aachen.de/record/962430https://publications.rwth-aachen.de/search?p=id:%22RWTH-2023-07390%22StudentsStudent Financial Aid ProvidersTeachersResearchersinfo:eu-repo/semantics/altIdentifier/doi/10.18154/RWTH-2023-07390