Based on an analysis of the role of industrial control and optimization technologies in the Industrial Revolution, as well as the current situation and existing problems of operational decision-making (ODM) for industrial process, this paper introduces the concept of intelligent ODM in industrial process, shapes its future directions, and highlights key technical challenges. By the tight conjoining of and coordination between industrial artificial intelligence (AI) with industrial control and optimization technologies, as well as the Industrial Internet with industrial computer management and control systems, an intelligent operational optimization decision-making methodology is proposed for complex industrial process. The intelligent ODM methodology and its successful application demonstrate that the tight conjoining of and coordination between next-generation information technologies with industrial control and optimization technologies will promote the development of industrial intelligent ODM. Finally, main research directions and ideas are outlined for realizing intelligent ODM in industrial process.
With the emergence of general foundational models, such as Chat Generative Pre-trained Transformer (ChatGPT), researchers have shown considerable interest in the potential applications of foundation models in the process industry. This paper provides a comprehensive overview of the challenges and opportunities presented by the use of foundation models in the process industry, including the frameworks, core applications, and future prospects. First, this paper proposes a framework for foundation models for the process industry. Second, it summarizes the key capabilities of industrial foundation models and their practical applications. Finally, it highlights future research directions and identifies unresolved open issues related to the use of foundation models in the process industry.
To achieve sustainable development goals and the requirements of a circular economy, a new class of intelligent computer-aided methods and tools is needed. Artificial intelligence (AI) techniques have been gaining much attention due to their ability to provide options to tackle the challenges we are currently facing. However, the successful application of AI to solve problems of current interest requires AI to be integrated with associated process systems engineering methods and tools that are already available or being developed. In this perspective paper, we highlight the use of a collection of process systems engineering methods and tools augmented by AI techniques to solve problems related to process manufacturing, with a focus on chemical product design, process synthesis and design, process control, and process safety and hazards.
Recent advances in artificial intelligence (AI) have led to the development of sophisticated algorithms that significantly improve image analysis capabilities. This combination of AI and microscopic imaging is transforming the way we interpret and analyze imaging data, simplifying complex tasks and enabling innovative experimental methods previously thought impossible. In smart manufacturing, these improvements are especially impactful, increasing precision and efficiency in production processes. This review examines the convergence of AI with particle image analysis, an area we refer to as “particle vision analysis (PVA)." We offer a detailed overview of how this technology integrates into and impacts various fields within the physical sciences and materials sectors, where it plays a crucial role in both innovation and operational improvements. We explore four key areas of advancement—namely, particle classification, detection, segmentation, and object tracking—along with a look into the emerging field of augmented microscopy. This paper also underscores the vital role of the existing datasets and implementations that support these applications, which provide essential insights and resources that drive continuous research and development in this fast-evolving field. Our thorough analysis aims to outline the transformative potential of AI-driven PVA in improving precision in future manufacturing at the microscopic scale and thereby preparing the ground for significant technological progress and broad industrial applications in nanomanufacturing, biomanufacturing, and pharmaceutical manufacturing. This exploration not only highlights the advantages of integrating AI into conventional manufacturing processes but also anticipates the rise of next-generation smart manufacturing, which is set to revolutionize industry standards and operational practices.
As industrial production progresses toward digitalization, massive amounts of data have been collected, transmitted, and stored, with characteristics of large-scale, high-dimensional, heterogeneous, and spatiotemporal dynamics. The high complexity of industrial big data poses challenges for the practical decision-making of domain experts, leading to ever-increasing needs for integrating computational intelligence with human perception into traditional data analysis. Industrial big data visualization integrates theoretical methods and practical technologies from multiple disciplines, including data mining, information visualization, computer graphics, and human-computer interaction, providing a highly effective manner for understanding and exploring the complex industrial processes. This review summarizes the state-of-the-art approaches, characterizes them with six visualization methods, and categorizes them based on analytical tasks and applications. Furthermore, key research challenges and potential future directions are identified.
Metal-organic frameworks (MOFs) hold great potential for gas separation and storage, and graph neural networks have proven to be a powerful tool for exploring material structure-property relationships and discovering new materials. Unlike traditional molecular graphs, crystal graphs require consideration of periodic invariance and modes. In addition, MOF structures such as covalent bonds, functional groups, and global structures impact adsorption performance in different ways. However, redundant atomic interactions can disrupt training accuracy, potentially leading to overfitting. In this paper, we propose a multi-scale crystal graph for describing periodic crystal structures, modeling interatomic interactions at different scales while preserving periodicity invariance. We also propose a multi-head attention crystal graph network in multi-scale graphs (MHACGN-MS), which learns structural characteristics by focusing on interatomic interactions at different scales, thereby reducing interference from redundant interactions. Using MOF adsorption for gases as an example, we demonstrate that MHACGN-MS outperforms traditional graph neural networks in predicting multi-component gas adsorption. We also visualize attention scores to validate effective learning and demonstrate the model’s interpretability.
With the intelligent transformation of process manufacturing, accurate and comprehensive perception information is fundamental for application of artificial intelligence methods. In zinc smelting, the fluidized bed roaster is a key piece of large-scale equipment and plays a critical role in the manufacturing industry; its internal temperature field directly determines the quality of zinc calcine and other related products. However, due to its vast spatial dimensions, the limited observation methods, and the complex multiphase, multifield coupled reaction atmosphere inside it, accurately and timely perceiving its temperature field remains a significant challenge. To address these challenges, a spatial-temporal reduced-order model (STROM) is proposed, which can realize fast and accurate temperature field perception based on sparse observation data. Specifically, to address the difficulty in matching the initial physical field with the sparse observation data, an initial field construction based on data assimilation (IFC-DA) method is proposed to ensure that the initial conditions of the model can be matched with the actual operation state, which provides a basis for constructing a high-precision computational fluid dynamics (CFD) model. Then, to address the high simulation cost of high-precision CFD models under full working conditions, a high uniformity (HU)-orthogonal test design (OTD) method with the centered L2 deviation is innovatively proposed to ensure high information coverage of the temperature field dataset under typical working conditions in terms of multiple factors and levels of the component, feed, and blast parameters. Finally, to address the difficulty in real-time and accurate temperature field prediction, considering the spatial correlation between the observed temperature and the temperature field, as well as the dynamic correlation of the observed temperature in the time dimension, a spatial-temporal predictive model (STPM) is established, which realizes rapid prediction of the temperature field through sparse observation data. To verify the accuracy and validity of the proposed method, CFD model validation and reduced-order model prediction experiments are designed, and the results show that the proposed method can realize high-precision and fast prediction of the roaster temperature field under different working conditions through sparse observation data. Compared with the CFD model, the prediction root-mean-square error (RMSE) of STROM is less than 0.038, and the computational efficiency is improved by 3.4184 × 104 times. In particular, STROM also has a good prediction ability for unmodeled conditions, with a prediction RMSE of less than 0.1089.
In the pharmaceutical industry, model-based prediction is a crucial stage in process development that allows pharmaceutical companies to simulate different scenarios toward improving process efficiency, reducing costs, and enhancing product quality. Nevertheless, ensuring the quality of formulated pharmaceutical products through the management of raw material variations has always been a challenging task. In this work, data-driven chance-constrained recurrent neural networks (CCRNNs) are developed to address the issue arising from raw material uncertainty. Our goal is to explore how, by proactively incorporating uncertainty into the model training process, more accurate predictions and enhanced robustness can be realized. The proposed approach is tested on a fluid bed dryer (FBD) from a continuous pharmaceutical manufacturing pilot plant. The results demonstrate that CCRNN models offer more robust and accurate predictions for the critical quality attribute (CQA)—in this case, moisture content—when material variations occur, compared with conventional recurrent neural network-based models.
The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants capable of evading both convalescent and vaccine-triggered antibody responses has underscored the pivotal role of T-cell immunity in antiviral defense. Here, we develop the ConFormer network for epitope prediction, which couples convolutional neural network (CNN) local features with Transformer global representations to enhance binding prediction performance, and employ the deep learning algorithm and bioinformatics workflows to identify conserved T-cell epitopes within the SARS-CoV-2 proteome. Five epitopes are identified as potential inducers of T-cell immune responses. Notably, the multi-valent vaccine composed of these five peptides significantly activates cluster of differentiation (CD)8+ and CD4+ T cells both in vitro and in vivo. The serum of mice immunized with this vaccine is able to neutralize the five major SARS-CoV-2 variants of concern. This study provides a candidate peptide vaccine with the potential to trigger antiviral T-cell responses, thereby offering the prospect of immune protection against SARS-CoV-2 variants.
With growing concerns over environmental issues, ethylene manufacturing is shifting from a sole focus on economic benefits to an additional consideration of environmental impacts. The operation of the thermal cracking furnace in ethylene manufacturing determines not only the profitability of an ethylene plant but also the carbon emissions it releases. While multi-objective optimization of the thermal cracking furnace to balance profit with environmental impact is an effective solution to achieve green ethylene manufacturing, it carries a high computational demand due to the complex dynamic processes involved. In this work, artificial intelligence (AI) is applied to develop a novel hybrid model based on physically consistent machine learning (PCML). This hybrid model not only reduces the computational demand but also retains the interpretability and scalability of the model. With this hybrid model, the computational demand of the multi-objective dynamic optimization is reduced to 77 s. The optimization results show that dynamically adjusting the operating variables with coke formation can effectively improve profit and reduce CO2 emissions. In addition, the results from this study indicate that sacrificing 28.97 % of the annual profit can significantly reduce the annual CO2 emissions by 42.89 %. The key findings of this study highlight the great potential for green ethylene manufacturing based on AI through modeling and optimization approaches. This study will be important for industrial practitioners and policy-makers.
Industrial decarbonization is critical for achieving net-zero goals. The carbon dioxide electrochemical reduction reaction (CO2RR) is a promising approach for converting CO2 into high-value chemicals, offering the potential for decarbonizing industrial processes toward a sustainable, carbon-neutral future. However, developing CO2RR catalysts with high selectivity and activity remains a challenge due to the complexity of finding such catalysts and the inefficiency of traditional computational or experimental approaches. Here, we present a methodology integrating density functional theory (DFT) calculations, deep learning models, and an active learning strategy to rapidly screen high-performance catalysts. The proposed methodology is then demonstrated on graphene-based single-atom catalysts for selective CO2 electroreduction to methanol. First, we conduct systematic binding energy calculations for 3045 single-atom catalysts to identify thermodynamically stable catalysts as the design space. We then use a graph neural network, fine-tuned with a specialized adsorption energy database, to predict the relative activity and selectivity of the candidate catalysts. An autonomous active learning framework is used to facilitate the exploration of designs. After six learning cycles and 2180 adsorption calculations across 15 intermediates, we develop a surrogate model that identifies four novel catalysts on the Pareto front of activity and selectivity. Our work demonstrates the effectiveness of leveraging a domain foundation model with an active learning framework and holds potential to significantly accelerate the discovery of high-performance CO2RR catalysts.
Microelectromechanical system (MEMS) high-temperature pressure sensors are widely used in aerospace, petrochemical industries, automotive electronics, and other fields owing to their advantages of miniaturization, lightweight design, simplified signal processing, and high accuracy. In recent years, advances in semiconductor material growth technology and intelligent equipment operation have significantly increased interest in high-temperature pressure sensors based on the third-generation semiconductor silicon carbide (SiC). This review examines the material properties of SiC single crystals and discusses several technologies influencing the performance of SiC pressure sensors, including the piezoresistive effect, ohmic contact, etching processes, and packaging methodologies. Additionally, it explores future research directions in the field. The review highlights the importance of increasing operating temperatures and advancing sensor integration as critical trends for future SiC high-temperature pressure sensor research and applications.
The passive containment heat removal system (PCS) is one of the key passive safety systems of China’s third-generation advanced pressurized water reactor—Hua-long Pressurized Reactor (HPR1000), used to prevent overpressure of large concrete containment under severe accident scenarios. This paper provides an overview of the development of the HPR1000 passive containment heat removal system, including its operating principles and configuration, internal heat exchanger design, feasibility tests, engineering-scale PCS verification tests, comprehensive tests on PCS-containment coupling characteristics, among other key supporting studies. These extensive studies demonstrated that the PCS of HPR1000, which is designed based on flashing-driven open natural circulation and efficient condensation heat transfer theory, can work effectively and ensure the integrity of the containment under various accident scenarios. The system has been applied to Fuqing No. 5 and No. 6 nuclear power units and Zhangzhou No. 1 and No. 2 units of China’s first million-kilowatt third-generation nuclear power HPR1000. It is also applied to K-2/K-3 units of Karachi Nuclear Power Plant in Pakistan.
Electric tractors (ETs) with mounted implements form operating units. There are significant differences in parameters such as shape, firmness, and moisture content of the soil in contact with the tractor and implements when working in complex terrains such as field stubble, waterlogged silt, and loose/firm terrain. These differentiated dynamics prevent cooperation between ETs and operating implements under independent control, resulting in poor quality operations and low energy efficiency. We propose a control mechanism for ETs and implements to achieve full life cycle management of collaborative control tasks, instantaneous intertask interaction, and a multitask synchronization mechanism. To address the internal redundant communication problems caused by traditional distributed microcontrol units, we break through the underlying technology of unit data processing and interaction and develop an integrated high-performance controller structure with high processing capacity and high- and low-speed communication interfaces. On the basis of hierarchical stepwise control theory, a hierarchical real-time operating system is designed. This system realizes a preemptive kernel response of computational tasks and competitive-collaborative synchronization among tasks; overcomes the low-latency response of collaborative control tasks, instantaneous information interaction, and multitask synchronization problems; and provides system-level support for deep collaborative operation control of units. To demonstrate and validate the proposed collaborative control mechanism, a plowing collaborative operation management strategy is designed and deployed. The experimental results show that the communication delay of collaborative tasks is as low as 83 μs, the solution time of complex collaborative equations is as low as 46 ms, the mechanical efficiency of the ET is increased by 9.07%, the efficiency of the drive motor is increased by 9.72%, the stability of the operation speed is increased by 106.25%, and the stability of the plowing depth reaches 94.98%. Our work meets the hardware and software requirements for realizing complex collaborative control of ET units and improves the operational quality and operational energy efficiency in real vehicle demonstrations.
Angiogenesis is essential for supporting tumor progression and metastasis. However, the potential role of the epitranscriptome in regulating angiogenesis remains unclear. Here, we identify the RNA N6-methyladenosine (m6A) reader insulin-like growth factor 2 (IGF2) messenger RNA (mRNA)-binding protein 2 (IGF2BP2) as the top enriched m6A regulator in hypervascular colorectal cancer (CRC), with its expression correlating with poor prognosis. Knockdown of IGF2BP2 in CRC cells suppressed their ability to promote pro-angiogenic phenotypes in endothelial cells in vitro, as well as vascular abnormalization, tumor progression, and metastasis in vivo. Supporting these findings, intestine-specific Igf2bp2 knock-in mice exhibited accelerated azoxymethane (AOM) plus dextran sulfate sodium (DSS)-induced CRC through enhanced angiogenesis and vascular abnormalities, whereas intestine-specific Igf2bp2 knockout inhibited tumor growth by normalizing tumor vasculature. Mechanistically, IGF2BP2 binds to m6A-modified cell migration inducing and hyaluronan binding protein (CEMIP) mRNA and enhanced its stability, leading to increased secretion of CEMIP. Secreted CEMIP interacts with membrane glucose-regulated protein 78 (GRP78) on endothelial cells, activating pro-angiogenic signaling. Importantly, targeting IGF2BP2 through genetic ablation, lipid nanoparticle (LNP)-encapsulated small interfering IGF2BP2, or the chemical inhibitor (CWI1-2) synergized with anti-angiogenic drugs to suppress tumor growth in multiple CRC models. Together, these findings suggest that targeting IGF2BP2 is a promising strategy to enhance the efficacy of anti-angiogenic therapy in CRC.
Extraction unit operation is the first step in traditional Chinese medicine (TCM) product manufacturing, and it is crucial in determining the quality of the produced medicine. However, due to a lack of effective multimodal monitoring and adjustment strategies, achieving high quality and efficiency remains a challenge. In this work, we proposed an artificial intelligence (AI)-based robot platform for the multi-objective optimization of the extraction process. First, a perception intelligence method for multimodal process monitoring was established to track active ingredient transfer and production changes during the extraction process. Second, a digital twin model was developed to reconstruct the field information, which interacted with real-time monitoring data. Furthermore, the model performed real-time inference to predict future production process states by using the reconstructing information. Finally, according to the predicted process states, the autonomous decision-making robot implemented multi-objective optimization, ensuring efficient process adjustments for global optimization. Experimental and industrial results demonstrated that the platform could effectively infer component transfer dynamics, monitor temperature variations, and identify boiling states, ensuring product quality while reducing energy consumption. This pharmaceutical robot could promote the integration of AI and pharmaceutical engineering, thereby accelerating the iterative development and improvement of China’s pharmaceutical industry.
As the world’s largest digital economy, China has a significant demand for data centers, which are energy-intensive. With an annual growth rate of 28% in installed capacity, these centers are primarily located in the developed eastern region, where land and energy resources are limited. This localization poses a major challenge to the industry’s net-zero goal. To address this, China has launched a bold initiative to relocate data centers to the western region, leveraging natural cooling, clean energy, and cost-effective resources. By 2030, this move is expected to reduce emissions from the data center sector by 16%-20%, generating direct economic benefits of approximately 53 billion USD. The success of this initiative can serve as a model for other countries to develop their internet infrastructure.
With the anticipated growth in air traffic complexity in the coming years, future civil aviation transportation system (CATS) is transforming into a complex cyber-physical-social system, surpassing all previous experiences in the history of civil aviation safety management. Therefore, a new safety concept based on a system-of-systems (SoS) perspective is proposed for the next-generation aviation. This article begins by elucidating the complexity of existing aviation risks and emphasizing the necessity for an updated safety concept. It then presents the challenges of current safety management and potential solutions from the new SoS perspective. To address future risks, the concept of SoS safety is introduced with the inspiration of the human immune system in terms of capability, logic, and architecture, which can serve as a guiding framework and methodology for safety engineering in complex large-scale CATS. This concept indicates the transition from “process and outcome-oriented" to “capability-oriented" intelligent safety management. Our research highlights the development directions and potential technological areas that need to be addressed at different stages of SoS safety. The integration of SoS design and operation through rapid iterations enabled by artificial intelligence (AI) will ultimately achieve endogenous SoS safety.
Placement optimization is a crucial phase in chip design, involving the strategic arrangement of cells within a limited region to enhance space utilization and reduce wirelength. Chip design enterprises need to optimize the placement according to design rules to meet customer demands. While mixed-cell-height circuits are widely used in modern chip design, few studies have simultaneously considered the non-overlapping cells, rails alignment, and minimum implantation area constraints in the placement optimization problems. Hence, this study involves preprocessing the non-linear parts and developing a mixed-integer linear programming model to reduce the cost of legalizing chip placements for businesses. Furthermore, this study designs and implements an exact algorithm based on Benders decomposition, utilizing dual theory to obtain an optimal cut and iteratively solve for the coordinates of cells. Numerical experiments across various scales validate the performance of the algorithm. Through a detailed analysis of the shape of the chip region division, the proportion of different types of cells, the total number of cells and bins, and their impact on the placement, we derive some potentially useful design insights that can benefit chip design enterprises.