Digital Chemical Engineering最新文献

In the Industry 4.0 revolution, industries are advancing their operations by leveraging Artificial Intelligence (AI). AI-based systems enhance industries by automating repetitive tasks and improving overall efficiency. However, from a safety perspective, operating a system using AI without human interaction raises concerns regarding its reliability. Recent developments have made it imperative to establish a collaborative system between humans and AI, known as Intelligent Augmentation (IA). Industry 5.0 focuses on developing IA-based systems that facilitate collaboration between humans and AI. However, potential conflicts between humans and AI in controlling process plant operations pose a significant challenge in IA systems. Human-AI conflict in IA-based system operation can arise due to differences in observation, interpretation, and control action. Observation conflict may arise when humans and AI disagree with the observed data or information. Interpretation conflicts may occur due to differences in decision-making based on observed data, influenced by the learning ability of human intelligence (HI) and AI. Control action conflicts may arise when AI-driven control action differs from the human operator action. Conflicts between humans and AI may introduce additional risks to the IA-based system operation. Therefore, it is crucial to understand the concept of human-AI conflict and perform a detailed risk analysis before implementing a collaborative system. This paper aims to investigate the following: 1. Human and AI operations in process systems and the possible conflicts during the collaboration. 2. Formulate the concept of observation, interpretation, and action conflict in an IA-based system. 3. Provide a case study to identify the potential risk of human-AI conflict.

Imaging and image-based process analytical technologies (PAT) have revolutionized the design, development, and operation of crystallization processes, providing greater process understanding through the characterization of particle size, shape and crystallization mechanisms in real-time. The performance of corresponding PAT models, including machine learning/artificial intelligence (ML/AI)-based approaches, is highly reliant on the data quality used for training or validation. However, acquiring high quality data is often time consuming and a major roadblock in developing image analysis models for crystallization processes.

To address the lack of diverse, high-quality, and publicly available particle image datasets, this paper presents an initiative to create an open-access crystallization-related image database: OpenCrystalData (OCD, at www.kaggle.com/opencrystaldata/datasets). The datasets consist of images from different crystallization systems with different particle sizes and shapes captured under various conditions. The initial release consists of four different datasets, addressing the estimation of particle size distribution using in-situ images for different categories of particles and detection of anomalous particles for process monitoring purposes. Images are collected using various instruments, followed by case-specific processing steps, such as ground-truth labeling and particle size characterization using offline microscopy. Datasets are released on the online collaborative platform Kaggle, along with specific guidelines for each dataset. These datasets are aimed to serve as a resource for researchers to enable learning, experimentation, development, and evaluation and comparison of different analytical approaches and algorithms. Another goal of this initiative is to encourage researchers to contribute new datasets focusing on various systems and problem statements. Ultimately, OpenCrystalData is intended to facilitate and inspire new developments in imaging-based PAT for crystallization processes, encouraging a shift from time-consuming offline analysis towards comprehensive real-time process insights that drive product quality.

Polymer grade scheduling, maritime surveillance, e-food delivery, e-commerce, and military tactics necessitate multiple agents (e.g., extruders, speed boats, salesmen) capable of visiting (or completing) dynamically changing locations (or tasks) in minimum time and distance. This study proposes a novel methodology based on clustering and local heuristic-based evolutionary algorithms to address the dynamic traveling salesman problem (TSP) and the dynamic multi-salesman problem with multiple objectives. The proposed algorithm is evaluated on 11 benchmark TSP problems and large-scale problems with up to 10,000 instances. The results show the superior performance of the proposed methodology called the dynamic two-stage evolutionary algorithm as compared to the dynamic hybrid local search evolutionary algorithm. Furthermore, the algorithm's applicability is illustrated through various scenarios involving up to four salesmen and three objectives with dynamically changing locations. To demonstrate real-world relevance, a maritime surveillance problem employing a helideck monitoring system is solved, wherein the objective is to minimize the patrolling route while visiting faulty vessels that threaten marine vessels. This study provides a general framework of TSP which finds application in several sectors, including planning and scheduling in chemical and manufacturing industries, the defense sector, and the e-commerce sector. Finally, the results showcase the effectiveness of the proposed methodology in solving the dynamic multiobjective, and multiple salesmen problem, which represents a more generalized version of the TSP.

The use of immersive technologies and digital twins can enhance training and learning outcomes in various domains. These technologies can reduce the cost and risk of training and improve the retention and transfer of knowledge by providing feedback in real-time. In this paper, a novel virtual reality (VR) based Bioreactor simulation is developed that covers the set-up and operation of the process. It allows the trainee operator to experience infrequent events, and reports on the effectiveness of their response. An embedded complex simulation of the bioreaction effectively replicates the impact of operator decisions to mimic the real-world experience. The need to train and assess the skills acquired aligns with the requirements of manufacturing in a validated environment, where proof of operator capability is a prerequisite. It has been deployed at UK’s National Horizons Center(NHC) to train the trainees in biosciences.

An issue that can limit the long-term value of information published in peer-reviewed engineering publications is the inability of readers to readily access data contained within a publication. This paper discusses experiences in changing the expectations for data sharing by authors in a large, disciplinary engineering journal, the AIChE Journal, in ways that seek to balance the burdens on authors and the benefits to readers.

Autoencoder-based reduced-order machine learning models have been developed for modeling and predictive control of nonlinear chemical processes with high dimensionality such as discretization of reaction–diffusion processes. However, in the presence of data noise, autoencoders may over-fit the training data and subsequently learn an inaccurate low-dimensional representation of the process variables. This leads to an inaccurate prediction model when the models are integrated with model predictive control (MPC). To address this issue, this work develops a novel machine-learning-based reduced-order modeling method by integrating SpectralDense layers into autoencoders and incorporating them with recurrent neural networks. We demonstrate that the new architecture of autoencoders using SpectralDense layers is more robust against over-fitting than conventional autoencoders in the presence of data noise, which improves the prediction accuracy in MPC. A diffusion–reaction process simulation example is used to demonstrate that the robust autoencoders outperform those using conventional layers for reduced-order modeling in predictive control.

Cold start is a critical operating scenario for the proton exchange membrane fuel cell (PEMFC), particularly in the field of transportation. Under sub-freezing temperatures, the water inside the cell will freeze and obstruct gas flow paths as well as cover catalyst reaction sites, resulting in a failed startup. This study proposes an optimization method for the -30°C cold start of PEMFC based on a data-driven surrogate model to improve cold start performance and reduce irreversible damage to the cell. A validated PEMFC cold start mechanism model is utilized as the basis for developing an extreme learning machine (ELM) based data-driven surrogate model, which is trained using data collected from the mechanism model and has higher computational efficiency compared with the original model. In addition, the NSGA-II multi-objective optimization algorithm is employed to optimize the current loading strategies and operating parameters using the surrogate model as fitness function. The objectives are to enhance the minimum voltage and reduce startup duration time. Moreover, experimental validation confirms the effectiveness of the proposed method. The test results demonstrate that a cold start from -30°C is achieved within 97 s, with the minimum voltage reaching 0.44 V. Notably, there is a reduction in startup time by 26 s and an increase in the minimum voltage by 0.06 V compared to the base case. This study establishes a foundation for researchers to adjust operating settings during cold start based on diverse applications and requirements.

Cryogenic processes with mixed refrigerants are prevalent in energy-intensive chemical industries, enhancing energy efficiency while reducing costs and unit size. However, the curse of dimensionality and process design constraints pose significant hurdles for effective screening and optimization. To tackle this, we developed a neural network model for natural gas liquefaction prediction. Trained on an extensive Aspen HYSYS database, our ML model accurately simulates LNG processes, with an impressive $R^2$ test value of 99.63, operating almost ten million times faster than HYSYS. It effectively addresses vital process design constraints, including liquid slugging and temperature cross, crucial for optimization. By integrating the ML model with genetic and Nelder–Mead algorithms, we achieve an 8.9% reduction in total exergy, outperforming Aspen HYSYS within the same time frame. Our study underscores ML’s significance in modeling energy-intensive chemical processes, providing insights into the exergy profile and enabling feature importance analysis.