Digital Chemical Engineering最新文献

The term hybrid modeling refers to the combination of parametric models (typically derived from knowledge about the system) and nonparametric models (typically deduced from data). Despite more than 20 years of research, over 150 scientific publications (Agharafeie et al., 2023), and some recent industrial applications on this topic, the capabilities of hybrid models often seem underrated, misunderstood, and disregarded by other disciplines as “simply combining some models” or maybe it has gone unnoticed at all. In fact, hybrid modeling could become an enabling technology in various areas of research and industry, such as systems and synthetic biology, personalized medicine, material design, or the process industries. Thus, a systematic investigation of the hybrid model properties is warranted to scoop the full potential of machine learning, reduce experimental effort, and increase the domain in which models can predict reliably.

Due to the enormous potential of modelling, graph-based approaches have been used for various applications in the process industries. In this study, we propose a fault detection framework through graphs by utilising its attributes in the form of node embeddings. Shallow embedding methods are deployed to generate node embedding vectors. Shallow embedding methods are broadly classified into matrix factorisation and skip-gram-based methods. Node2vec and Deepwalk fall under skip-gram models, while GraphRep and HOPE constitute the Matrix factorisation methods. Node embedding values generated from these methods are then fed to the variational auto-encoder, which ranks the nodes in reconstruction loss value. The node embedding reconstruction loss values exceeding a particular threshold are considered outliers. The proposed work has been validated on NPCIL power-flux data and the benchmark Tennessee Eastman data. The results indicate that skip-gram models, especially Node2vec-VAE, outperformed the matrix factorisation methods for both the above-mentioned datasets.

The potential for greater algorithmic efficiency for some problems on quantum computers compared to classical computers is appealing in many fields including, for example, the process systems engineering field. While quantum algorithms have been studied for a variety of applications related to optimization, molecular modeling, and machine learning, there remain many applications in process systems engineering, including process control, where it is not clear how quantum computing algorithms would be beneficial. One idea for attempting to understand when a quantum algorithm might provide benefits for control is to start with algorithms that would be expected to benefit “similar” problems (e.g., optimization problems) and to see if controllers can be implemented within those algorithmic frameworks. Therefore, in this work, we study the use of a quantum computing algorithm related to Grover’s algorithm, which is an amplitude amplification strategy that can search an unordered list with improved efficiency compared to a classical algorithm for the task. It has been extended to perform a search for optimal paths over a graph. Given its potential utility for search and optimization, this is an example of an algorithm where we might wonder if it could be adjusted or used to provide speed-ups for large control problems if the controller could function within this algorithmic framework. This work provides the first steps toward attempting to address this question by investigating how optimization-based control problems would fit into this framework. A process described by $\dot{x}=x+u$ is considered as a test case. The modified Grover’s algorithm requires the optimization problem to be mapped into quantum gates. We discuss ideas for attempting to represent an optimization-based controller known as model predictive control (MPC) in the modified Grover’s algorithm framework. We test how various parameters of the control and quantum algorithm designs, including fundamental parameters in MPC such as the number of sampling periods and length of the sampling periods, impact the success of using the quantum algorithm for the MPC. We provide analyses regarding why the results are what they are to give perspective on how quantum computing algorithms work and intersect with engineering problems.

This research focuses on encrypted distributed control architectures, aimed at enhancing the operational safety, cybersecurity and computational efficiency of large-scale nonlinear systems, where only partial state measurements are available. In this setup, a distributed model predictive controller (DMPC) is utilized to partition the process into multiple subsystems, each controlled by a distinct Lyapunov-based MPC (LMPC). To consider the interactions among different subsystems, each controller receives and shares with the other controllers control inputs computed for its particular subsystem. As full state feedback is not available, we integrate an extended Luenberger observer with each LMPC, initializing the LMPC model with complete state estimate information provided by the observer. Furthermore, to enhance cybersecurity, wireless signals received and transmitted by the controllers are encrypted. Guidelines are established to implement this proposed control structure in any large-scale nonlinear chemical process network. Simulation results, conducted on a specific nonlinear chemical process network, demonstrate the effective closed-loop performance of the encrypted DMPC with state estimation, utilizing partial state feedback with sensor noise. This is followed by a comprehensive comparison of the closed-loop performance, control input computational time, and suitability of encrypted centralized, decentralized, and distributed MPC frameworks.

Due to the complexity of biochemical systems, the development of traditional phenomenological models is limited if the underlying mechanics are not entirely known. As an alternative, hybrid model frameworks, consisting of data-driven models complemented with first principles models like conservation law, are starting to be used for complex systems. In this work, a comparison of the modeling capabilities between a data-driven model and a hybrid model was developed. The enzymatic production of Galactooligosaccharides (GOS) with the effect of metallic ions was considered as case study. Compared with the experimental results, predictions from data-driven model achieve an $R^2$ of 0.9188 in the best training fold, and the hybrid model an $R^2$ of 0.9696 in the best training fold. Illogical predictions were avoided by including non-phenomenological first-principles constraints into the hybrid model. Finally, an optimization analysis was carried out to find the highest GOS productivity using the hybrid model, optimization results present a deviation of 5.99 % compared to the highest productivity found from experimental data.

Traditionally, tuning of PID controllers is based on linear approximation of the dynamics between the manipulated input and the controlled output. The tuning is performed one loop at a time and interaction effects between the multiple single-input-single-output (SISO) feedback control loops is ignored. It is also well-known that if the plant operates over a wide operating range, the dynamic behaviour changes thereby rendering the performance of an initially tuned PID controller unacceptable. The design of PID controllers, in general, is based on linear models that are obtained by linearizing a nonlinear system around a steady state operating point. For example, in peak seeking control, the sign of the process gain changes around the peak value, thereby invalidating the linear model obtained at the other side of the peak. Similarly, at other operating points, the multivariable plant may exhibit new dynamic features such as inverse response. This work proposes to use deep reinforcement learning (DRL) strategies to simultaneously tune multiple SISO PID controllers using a single DRL agent while enforcing interval constraints on the tuning parameter values. This ensures that interaction effects between the loops are directly factored in the tuning. Interval constraints also ensure safety of the plant during training by ensuring that the tuning parameter values are bounded in a stable region. Moreover, a trained agent when deployed, provides operating condition based PID parameters on the fly ensuring nonlinear compensation in the PID design. The methodology is demonstrated on a quadruple tank benchmark system via simulations by simultaneously tuning two PI level controllers. The same methodology is then adopted to tune PI controllers for the operating condition under which the plant exhibits a right half plane multivariable direction zero. Comparisons with PI controllers tuned with standard methods suggest that the proposed method is a viable approach, particularly when simulators are available for the plant dynamics.

This paper addresses the liquefied natural gas (LNG) sales planning problem over a pipeline network with a focus on uncertain demands. Generically, the total profit is maximized by seeking optimal transportation and inventory decisions, and robust optimization (RO) has been a viable decision-making strategy to this end, which is however known to suffer from over-conservatism. To circumvent this, a new goal-oriented data-driven RO approach is proposed. First, we adopt data-driven polytopic uncertainty sets based on kernel learning, which yields a compact high-density region from data and assures tractability of RO problems. Based on this, a new goal-oriented RO formulation is put forward to satisfy to the greatest extent the target profit while tolerating slight constraint violations. In contrast to traditional min–max RO scheme, the proposed scheme not only ensures a flexible trade-off but also yields parameters with clear interpretation. The resulting optimization problem turns out to be equivalent to a mixed-integer linear program that can be effectively handled using off-the-shelf solvers. We illustrate the merit of the proposed method in satisfying a prescribed goal with optimized robustness by means of a case study.

This paper aims to outline an approach for case-based chemistry and chemical engineering education for sustainability. Education for Sustainability is assumed to offer a holistic approach to equip students with the knowledge, skills, values, and attitudes needed to contribute to a more sustainable society in their future careers. While Case-Based Education traditionally focuses on disciplinary learning in simulated settings, it can also effectively teach essential sustainability-related skills like integrated problem-solving, critical thinking, and systems thinking. The approach we propose is “case hacking”, which should be understood as utilizing existing business cases while incorporating supplementary resources to align the assignment with intended learning objectives. This expansion of the cases involves, among other things, introducing additional questions and assignments, perspectives from stakeholders previously unexplored in the original case, and the integration of recent research articles from relevant fields. We advocate for the use of case hacking when educators want to harness the educational benefits of Case-Based Education while emphasizing the complexity of sustainability-related challenges faced by industrial companies today. As an illustrative example, we demonstrate the process of hacking a case related to Green Chemistry in the pharmaceutical industry, highlighting specific challenges for chemistry and chemical engineering education. We hope this example will inspire educators in these disciplinary contexts to engage with the case hacking approach as they navigate the complex terrain of sustainability.

Water is a natural and essential resource for humans, animals, and plants to persist. However, only ⁓2.5 % of the freshwater resources are available, while the remaining ⁓97.5 % is saline water, which is unsuitable for humanity. According to the WHO, water scarcity will worsen by 2050. As a result, numerous researchers, scientists, and engineers are working in this field to improve water resources with advanced treatment technologies. Aside from the multiple water resources, desalination is critical in converting saline water to fresh water. In line with a recent update from the International Desalination Association (IDA, Reuse Handbook 2022–23), approximately ⁓22,757 desalination plants are operating worldwide, providing ⁓107.95 million cubic meters of freshwater per day (m³/day). Furthermore, in this digital age, artificial intelligence (AI) techniques, such as gray wolf optimization (GWO), sine cosine algorithm (SCA), artificial neural networks (ANN), multi-verse optimizer (MVO), fuzzy logic systems (FLS), moth flame optimizer (MFO), particle swarm optimization (PSO), artificial hummingbird algorithm (AHA) and genetic algorithms (GA), are playing a vital role and capable of deep analysis of real-time desalination plant for saving time, energy, human efforts, and money. This study focuses on the critical review and various aspects of current-age PSO-ANN techniques for desalination plants. In this regard, recent datasets of the Web of Science (WoS), provided by Clarivate Analytics, state that about >54,856 records (1965–2023) of desalination and around > 180 records (2008–2023) of PSO-ANN techniques are available globally. These records involve research articles, reviews, proceedings, letters, books, chapters, and editorial materials. Finally, this review article is specific and analyzes the various perspectives of PSO-ANN techniques in the water desalination process, promoting plant engineers and researchers to improve plant performance with minimum effort and time.

This paper explores the integration of large language models (LLMs), such as ChatGPT, in chemical engineering education, departing from conventional practices that may not be universally accepted. While there is ongoing debate surrounding the acceptance of LLMs, driven by concerns over computational instability and potential inconsistencies, their inevitability in shaping our communication and interaction with technology cannot be ignored. As educators, we are positioned to play a vital role in guiding students toward the responsible, effective, and synergetic use of LLMs. Focusing specifically on distillation column design in undergraduate mass-transfer courses, this study demonstrates how ChatGPT can be utilized as an auxiliary tool to create interactive learning environments and simulate real-world engineering thinking processes. It emphasizes the need for students to develop critical thinking skills and a thorough understanding of LLM principles, taking responsibility for their use and creations. While ChatGPT should not be solely relied upon, its integration with fundamental principles of chemical engineering is crucial. The effectiveness and limitations of ChatGPT are exemplified through two case studies, showcasing the importance of manual calculations and established simulation software as primary tools for guiding and validating engineering results and analyses. This paper also addresses the pedagogical implications of integrating LLMs into mass transfer courses, encompassing curriculum integration, facilitation, guidance, and ethical considerations. Recommendations are provided for incorporating LLMs effectively into the curriculum. Overall, this study contributes to the advancement of chemical engineering education by examining the benefits and limitations of LLMs as educational aids in the design process.