arXiv subjects

Pascal Schäfer

Publications and source records attributed to Pascal Schäfer.

4 recordsLinked to original sources

arXiv2026

Context is all you need: Towards autonomous model-based process design using agentic AI in flowsheet simulations

Agentic AI systems integrating large language models (LLMs) with reasoning and tooluse capabilities are transforming various domains - in particular, software development. In contrast, their application in chemical process flowsheet modelling remains largely unexplored. In this work, we present an agentic AI framework that delivers assistance in an industrial flowsheet simulation environment. To this end, we show the capabilities of GitHub Copilot (GitHub, Inc., 2026), when using state-of-the-art LLMs, such as Claude Opus 4.6 (Anthropic, PBC, 2026), to generate valid syntax for our in-house process modelling tool Chemasim using the technical documentation and a few commented examples as context. Based on this, we develop a multi-agent system that decomposes process development tasks with one agent solving the abstract problem using engineering knowledge and another agent implementing the solution as Chemasim code. We demonstrate the effectiveness of our framework for typical flowsheet modelling examples, including (i) a reaction/separation process, (ii) a pressure-swing distillation, and (iii) a heteroazeotropic distillation including entrainer selection. Along these lines, we discuss current limitations of the framework and outline future research directions to further enhance its capabilities.

cs.AI↗

arXiv2025

On Computing and Pricing of Adjustable Robust Chemical Process Designs

Model-based process simulation can be used to derive designs and operating conditions of chemical processes that optimally balance multiple objectives, such as quality, costs, or environmental impacts. This work focuses on identifying designs that hedge against uncertainties in model parameters to ensure feasibility, taking the possibility to adjust operating conditions into account. An adaptive scheme is proposed to pinpoint the relevant scenarios in a discretized uncertainty space; these scenarios are then fed into a multi-objective adjustable robust optimization framework reducing the computational burden compared to the consideration of all potential scenarios. Furthermore, we propose a method to quantify the cost or price of robustness, i.e., the compromise which has to be made in comparison to the nominal design case in order to hedge against uncertainty. The conceptual findings are illustrated with an industrially relevant case study.

math.OC↗

arXiv2025

Pareto-optimal treatment of uncertainties in model-based process design and operation

Model-based process design and operation involves here-and-now and wait-and-see decisions. Here-and-now decisions include design variables like the size of heat exchangers or the height of distillation columns, whereas wait-and-see decisions are directed towards operational variables like reflux and split ratios. In this contribution, we describe how to deal with these different types of decisions in a multicriteria framework, offering an adjustability for the wait-and-see variables while at the same time respecting optimality guarantees on process KPIs.

math.OC↗

arXiv2020

Adaptive Learning of Hybrid Models for Nonlinear Model Predictive Control of Distillation Columns

Nonlinear model predictive control (NMPC) requires accurate and computationally efficient plant models. Our previous work has shown that the classical compartmentalization model reduction approach for distillation columns can be enhanced by replacing parts of the system of equations by artificial neural networks (ANNs) trained on offline solved solutions to improve computational performance. In real-life applications, the absence of a high-fidelity model for data generation can, however, prevent the deployment of this approach. Therefore, we propose a method that utilizes solely plant measurement data, starting from a small initial data set and then continuously adapting to newly measured data. The efficacy of the proposed approach is examined in silico for a distillation column from literature. To this end, we first adjust our reduced hybrid mechanistic/data-driven modeling approach that originally builds on compartmentalization to a stage-aggregation procedure, tailoring it for the application within the adaptive learning framework. Second, we apply an adaptive learning algorithm that trains the ANNs replacing the stationary stage-to-stage calculations on newly available data. We apply the adaptive learning of the hybrid model within a regulatory NMPC framework and conduct closed-loop simulations. We demonstrate that by using the proposed method, the control performance can be steadily improved over time compared to a non-adaptive approach while being real-time applicable. Moreover, we show that the performance when using either a model trained on excessive amounts of offline generated data or the original high-fidelity model can be approached in the limit.

math.OC↗