In the field of biochemical process design, the accurate modeling of microbial growth is essential for the development and optimization of biological reactors used in the production of high-value compounds. Achieving this objective requires a detailed understanding of how environmental factors-such as pH and nutrient availability-influence microbial dynamics across the four distinct growth phases: lag, exponential, stationary, and death. Traditionally, reactor design relies heavily on the Monod model, which provides a simplified representation of microbial growth, focusing primarily on the exponential phase under constant operating conditions (1). However, this model presents substantial limitations when applied to dynamic environments where key parameters vary over time. To overcome these constraints, the present study proposes a data-driven modeling approach using a multilayer perceptron (MLP) artificial neural network for the prediction of microbial growth trajectories under varying... [more]

A survey was conducted of 103 students and lecturers who had recently participated in chemical engineering design courses concerning their opinions on the use of Generative Artificial Intelligence (Gen-AI) in their capstone design education. Participants were at universities in Europe, the Middle East, North America, and South America, from at least eight different language groups. The survey found little difference in responses between students and lecturers, except for uptake, in which students reported higher rates of familiarity and adoption of Gen-AI tools than instructors. Both groups were net-positive generally on the use of Gen-AI in the classroom, reporting relatively high confidence in the ability to assess results, the general positive benefits of using Gen-AI in their chemical process design education, and the likelihood of using them in the future. However, participants reported that their trust in the results of Gen-AI tools was relatively low.

As Industry 5.0 (I.D. 5.0) reshapes the engineering education landscape, Higher Education Institutes (HEIs) have evolved to integrate Generative Artificial Intelligence (GenAI) via strategic curriculum revamps to meet Education 5.0 (E.D. 5.0) competencies. EN.540.202 (Introduction to Chemical & Biological Process Analysis) is the first core course at Johns Hopkins University and was revamped in Fall 2025 to create more rigorous course content and the conscious creation of new weekly graded problem sets, which did not rely on prior course content/textbook-based solved examples. Problem sets were fed as Effective Prompt Engineering (EPE) inspired prompts to ChatGPT, and AI-elicited responses were compared. AI was able to perform fundamental calculations, offer detailed explanations, unit conversions/checks, proactive information (outside the problem scope), and graphical information. Key challenges and pitfalls observed were terminology misinterpretation, lack of visual representation, d... [more]

With the ongoing advances in generative artificial intelligence (GenAI), the initial skepticism surrounding its tools is gradually diminishing. In fact, tools such as ChatGPT, Copilot and similar, are often used in everyday tasks, both in our personal lives and in educational contexts. Educators may use them for content creation, grading exams, or automating repetitive tasks. Students resort to them to better understand a topic, get feedback on an assignment and brainstorm ideas. Research has shown that, if used correctly, these tools can spur and support both teaching and learning. However, these continuous advancements and the increasing number of available tools also require more research to benchmark all these models and, if possible, provide quantifiable indications of which tool is better to use for which specific subtopic. As such, we created FermBench, a dataset of fermentation knowledge, which can be used to benchmark various large language models (LLMs). The models selected f... [more]

Recurrent neural nets (RNNs) are now commonly used for the surrogate modeling of process systems, leading to better control and faster real-time optimization. However, when trained with small training data sets, most experiments show that RNNs exhibit poor generalization abilities outside the range of the training data space. Nonetheless, recent advances in deep learning research have shown that certain characteristics of the loss landscape of trained models, such as the flatness around the local minimum, tend to relate to generalization ability. This paper investigates this phenomenon for the case of RNN surrogates of the well-known Three Tanks case study, which is representative of many continuous processes. We trained a total of 200 LSTMs (long short-term memory networks) differing in initialization, architecture, and training dynamics on the same data of 500 samples. The number of model parameters ranges from 238 to 11, 353. We estimated the loss curvature of each trained model usi... [more]

In this work, we propose a methodology for monitoring the performance of model predictive controllers (MPCs). A sequence of binary classification machine learning models, organized in cascade, called Cascade Machine Learning Models (CMLM), is evaluated to give a diagnosis of the control conditions. The proposed methodology was assessed using two case studies: a benchmark problem (the van de Vusse reactor under nonlinear MPC, NMPC) and a simulated industrial debutanizer column under commercial MPC. The ML models evaluated were the Random Forest and the Multilayer Perceptron. The results show that the proposed approach outperforms both a single multiclass model and traditional MPC performance monitoring methodologies, while remaining adaptable and scalable to larger applications.

This research presents a novel framework for the surrogate-based dynamic optimization of control schemes within chemical separation and purification processes such as the biorefinery downstream processing. The current study investigated the downstream of an enzymatic bioreactor responsible for the synthesis of 5-hydroxymethylfurfural value-added derivatives, focusing on the critical balance between operational costs and productivity. Two high-fidelity long short-term memory neural network-based surrogate models were developed to predict energy consumption and economic gain, both achieving a coefficient of determination (R2) exceeding 0.97. These models were subsequently integrated into a multi-objective optimization architecture to address an operating efficiency testing scenario characterized by stepwise inflow parameter changes. By exploring the resulting Pareto front, an optimal set of operational (control) settings was identified and validated. The results demonstrate that while en... [more]

In recent years, research for systematic process design approaches has gained traction, especially with the rise in popularity of generative machine learning models and reinforcement learning. However, works from the literature will often focus on proof-of-concept studies, limited to a specific process synthesis problem. Despite showing promising results, it is not clear how easily these methodologies could be transposed to new applications, and whether they would be successful. In this context, this work evaluates the possibility of using a Natural Language Processing model, which has already proven itself for thermodynamic cycle generation, for another different case: seawater desalination. The processes generated by this model will initially be those using reverse osmosis processes aimed at desalinating a seawater solution containing 25000 ppm of NaCl. Results show that the model has been successful in designing structural reverse osmosis desalination processes without defining asse... [more]

Microstructure segmentation is essential for quantitative materials analysis; however, supervised deep learning demands substantial annotation, whereas general-purpose foundation models such as the Segment Anything Model (SAM) offer limited domain-specific semantic control. This paper presents MatStudio, a human-in-the-loop framework for microstructure segmentation that is proposed and implemented end to end in this work. MatStudio couples an interactive workflow for batchwise micrograph annotation and model adaptation with a dual-head convolutional architecture and SAM-guided boundary refinement. The loop combines sparse supervision with SAM-assisted labeling, task-specific training, and iterative batch-level correction, typically converging within two to three cycles.. The network comprises a shared encoder initialized from a pretrained backbone and two decoders: a UNet-style segmentation head that jointly predicts class labels and pixelwise uncertainty, and a prototype branch that m... [more]

The increasing demand for energy, water, and chemical products signals the need for more sustainable and efficient process design methodologies. Traditional methods for conceptual process design constrains the exploration of novel and intensified process alternatives, as they rely on prior knowledge in defining the design space. Previous studies employing bottom-up approaches, such as phenomena building blocks (PBBs), suggest that the synthesis of complex bottom-up flowsheets remains computationally challenging and is thus limited to the synthesis of individual units of operation. This work proposes a bottom-up, data-driven framework for process synthesis and intensification based on phenomena building blocks (PBBs), in which process flowsheets are constructed from their underlying physical and chemical phenomena rather than conventional units of operation. The proposed framework introduces a phenomena-based text representation and data collection module. Furthermore, a sequence traini... [more]

Real-time monitoring of metabolite concentrations is critical for optimising bioprocess performance. While Raman spectroscopy offers a non-invasive solution, translating spectra into metabolite concentration estimates requires robust machine learning models. Foundation models such as TabPFN demonstrate exceptional predictive performance but suffer from high inference complexity when trained on large calibration datasets, hindering their use in real-time laboratory settings. This study proposes a batch Active Learning (AL) strategy to efficiently calibrate TabPFN using a minimal subset of data. We employ a weighted K-means clustering strategy that balances model uncertainty and dataset diversity to select the most informative calibration samples. We evaluated this method on a dataset of nearly 7, 000 Raman spectra covering eight substances. Our AL strategy achieved a mean R² score greater than 0.95 with approximately 1, 000 samples, significantly outperforming random sampling. Notably,... [more]

Fault diagnosis is essential for ensuring safety and efficiency in chemical process industries. Conventional diagnostic systems often generate raw numerical outputs that require extensive human interpretation, increasing the operator's workload and slowing decision-making during abnormal events. To overcome these limitations, this work introduces a model context protocol (MCP)-integrated fault diagnosis framework, where a Large Language Model (LLM) functions as the MCP client, coordinating multiple diagnostic tools through a unified protocol. Within the proposed framework, the LLM interacts with specialized diagnostic tools, including a convolutional neural network-based fault diagnosis model and an ensemble-based variant for uncertainty-aware analysis. The LLM synthesizes the outputs of these tools and generates operator-oriented natural-language reports that summarize diagnostic results and explicitly communicate uncertainty, thereby supporting more transparent and efficient decision... [more]

Machine learning (ML) has attracted growing interest in process systems engineering for its potential in process design, synthesis, and optimization. By learning complex patterns from data, ML methods complement traditional first-principles modelling and heuristic approaches, particularly for conceptual process design and the exploration of alternatives. Although current text-based representations capture unit-level connectivity, they lack a holistic view of process intent, equipment hierarchy, and contextual information to guide learning and inference. Consequently, models trained on such linear token sequences tend to reproduce syntactic structure rather than underlying process reasoning, thus limiting interpretability and user control. In this work, we introduce a contextual framework for representing process flowsheet information in ML models that embeds process engineering logic directly into the model inputs. The approach combines a structured, text-based representation of proces... [more]

Batch manufacturing is central to fine chemicals, pharmaceuticals, and bioprocessing. Its operation evolves across phases and recipes, which yields high-dimensional trajectories and strong batch-to-batch variability. Meanwhile, key quality-indicative variables are often measured offline and cannot be used as online model inputs. This work presents an integrated deep learning framework that unifies soft sensing and process monitoring in a single module using only process variables as inputs. A multi-scale Temporal Convolutional Network with multiple kernel sizes extracts complementary dynamic features from sliding windows. These features are concatenated and pooled into a compact representation that feeds two task branches. A variational autoencoder branch reconstructs the input window and provides fault monitoring signals via reconstruction deviation while regularizing the latent space through KL divergence. In parallel, a prediction branch estimates the quality-indicative variable dir... [more]

The accurate prediction of pure component properties has consistently been a critical issue in fields such as chemical engineering, biomedicine, and environmental science. In recent years, end-to-end deep learning methods have shown significant improvement over traditional machine learning approaches. This is due to their ability to automatically learn task-relevant representations from raw molecular data. In addition to accurate property prediction, researchers have increasingly focused on how specific fragment structures influence molecular properties. However, existing fragmentation methods based on predefined rules and group libraries struggle to capture novel molecular structures, which hampers the development of new materials and drugs. To address these challenges, this work proposes a hierarchical molecular fragmentation method. This method can automatically segment molecules into multiple fragments containing key functional groups. Then a three-branch graph attention network wa... [more]

Laboratory safety is procedure-dependent: required personal protective equipment (PPE) and permissible actions vary across experiments and across experimental steps, yet most vision-based monitoring remains appearance-driven and often produces generic warnings without reliable procedural context. We propose a multimodal framework for step-aware safety monitoring in laboratory videos. The framework first localizes procedural context through clip-level step prediction and protocol alignment to identify the experiment and current step. Given this context, it retrieves step-specific safety constraints, extracts evidence of step-relevant equipment and interactions using an equipment database, and prompts a video-capable vision-language model (VLM) to generate structured (JSON) monitoring reports supported by retrieved constraints and visual evidence. Experiments on protocol-annotated molecular biology lab videos show that our approach improves the mean score from 0.4352 to 0.6430 and reduce... [more]

Near-infrared (NIR) sorting systems in waste sorting plants operate under multiple settings, creating distinct input-output relationships that challenge predictive modeling. Conventional neural networks, such as multilayer perceptron (MLP), often suffer from catastrophic forgetting under continual training, limiting reliability across settings. This study evaluates Kolmogorov-Arnold Networks (KAN) for continual regression modeling of multi-setting NIR systems. KAN assign nonlinear transformations to network edges using localized spline grids, enabling structural isolation between input regions. We introduce controlled input-space manipulations (shifting successive settings to adjacent or non-overlapping grid regions) and compare KAN performance with MLPs of comparable parameter count. We also examine single-input versus multi-input configurations to assess dimensionality effects. Results show that KANs with sufficient input-space separation maintain previously learned knowledge with pe... [more]

The rapid integration of renewable energy sources is increasing the volatility and non-stationarity of modern power systems, posing significant challenges for data-driven forecasting models. In particular, concept drift and uncertainty in exogenous inputs such as weather forecasts can severely degrade predictive performance over time. This work proposes a lightweight two-layer forecasting framework that decouples prediction from adaptation. A traditional offline regression model is augmented by an online meta-learner that continuously generates adaptive meta-features, enabling the system to respond to structural changes and noisy inputs without repeated retraining. The framework is evaluated on two real-world case studies. First, concept drift is addressed in nuclear power production forecasting, where abrupt and gradual capacity changes are inferred through an online meta-learner. Second, feature uncertainty is mitigated in day-ahead solar production forecasting by correcting noisy we... [more]

The unprecedented expansion of Artificial Intelligence is adding increasing electricity demand to Europe's power system. While incumbent plans pursue a net-zero future by 2050, they fail to consider the implications of large-scale AI-based data centres. In this study, a spatially explicit optimisation model is developed to assess how hyperscale data centres may reshape energy infrastructure investment, and emissions trajectories, across different AI demand growth scenarios. The results indicate that, after 2030, AI capacity deployment increasingly shifts toward regions with the ability to expand nuclear and gas-based generation, as firm and flexible power sources are essential for supporting the deployment of high-capacity AI data centres. By 2050, AI-driven electricity demand under high growth scenarios may reach up to 450 TWh, corresponding to 7% of total Europe's demand, with installed AI capacity reaching approximately 85 GW. This additional load leads to an increase of nearly 25 M... [more]

Supplementary material for: Generative AI for the optimal design of seawater desalination processes (ESCAPE 36, Sheffield, June 2026)

Supplementary material for the article "Utilizing Machine Learning for Phenomena-based Synthesis of Intensified Process Flowsheets", submitted to The 36th European Symposium on Computer Aided Process Engineering (ESCAPE 36). The document includes information about the heurstic and samplic logic rules used in generating the initial dataset, and the grid search results for hyperparamter optimization.

This is supplementary material for the paper "Generative AI in Process Design Instruction: A Survey of Students and Faculty" in Systems and Control Transactions. The supplementary material contains reference information for the paper. Specifically, it contains the survey questions used in the study, the raw data results of that survey, and a ChatGPT transcript of a session in which ChatGPT was used to synthesize a flowsheet of an ammonia synthesis process and perform an analysis of the conceptual design.

Job Safety Analysis (JSA) is critical for proactively identifying workplace hazards, assessing their potential consequences, and implementing effective control measures. However, traditional JSA methods can be inefficient and prone to errors, particularly in complex industrial environments. This paper introduces AutoJSA, a knowledge-enhanced framework that leverages large language models (LLMs) to automate and optimize the JSA process. We collected 73 high-quality JSA reports from a chemical engineering company and divided the JSA workflow into three key tasks: hazard identification, consequence identification, and control measure generation. Two approaches - fine-tuning and retrieval-augmented generation (RAG) - were employed on a base LLM (GLM-4-9B-Chat) to adapt it for these domain-specific tasks. Experimental results demonstrate that both fine-tuning and RAG significantly improve task performance relative to the unmodified model, with fine-tuning generally providing larger gains. W... [more]

Maintenance is critical for industrial plants to ensure operational reliability and worker safety. In process industries, fouling, the accumulation of solid residues in equipment, poses a significant challenge, causing inefficiencies and productivity losses. Effective modeling of fouling evolution over time is essential for maintenance planning to prevent equipment from operating under suboptimal conditions. Traditional approaches to fouling prediction include equation-based models, which offer high precision but may struggle with continuously changing process bound-aries, and machine learning techniques, which are more adaptable but less effective at capturing rapidly evolving trends driven by complex underlying physics. This study introduces an innova-tive hybrid machine learning approach for predictive maintenance, combining the strengths of both methods. Pressure differential is modeled using an equation-based approach that links pressure data with fouling thickness, while the foul... [more]

Mechanistic population balance modeling (PBM) has advanced the design of pharmaceutical crystallization processes, enabling the production of active pharmaceutical ingredient (API) crystals with desired critical quality attributes (CQAs), such as purity and crystal size distribution. However, PBM development can sometimes be resource-intensive, requiring extensive design of experiments (DoE) and high-quality process data, making it impractical under fast-paced industrial development timelines. This study proposes a machine learning (ML)-based workflow for developing ‘fit-for-purpose’ digital twins of crystallization processes, leveraging industrially available DoE data to link operating conditions with CQAs. Validated on industrial data for a commercial API with complex crystallization challenges, the workflow efficiently identifies optimal operating conditions, demonstrating the potential of data-driven digital twins to accelerate the development of pharmaceutical processes.