Thread Content
“The purpose of integrating digital and physical elements is to reshape the value of industries. It is very important to figure out how artificial intelligence (AI) can help us analyze, optimize, and achieve better integration between industry and finance. This requires a concept based on the entire life cycle of petrochemical plants, starting with digital plant design, making full use of plant operation data, and integrating information technologies such as AI with expert knowledge systems in order to achieve it. ”On July 15, at the \"Discussing the Intelligent World – Huawei Smart Chemicals Summit Forum,\" jointly hosted by the Chinese Chemical Society and Huawei, Sun Lili, an academician of the Chinese Academy of Engineering and chief scientist at Sinopec Group Corporation, offered the following insights on the integration of digital technologies and traditional practices in the chemical industry. During the conference, several industry experts noted that AI is bringing about innovations in the R&D, production, and operation paradigms of the chemical industry, accelerating the overall value restructuring of the chemical industry chain. Innovating research paradigms significantly shortens the industry’s innovation cycle. “The chemical industry is currently at a historical crossroads of industrial transformation.” ”Hu Jie, deputy secretary-general of the Chinese Chemical Society, noted that according to statistics, in 2024 the number of enterprises in China’s petrochemical industry that met certain size criteria exceeded 32,000, with their annual revenue reaching 16 trillion yuan. However, the total profit of the entire industry was only 789.71 billion yuan, a decrease of 8.8% compared to the previous year. The questions surrounding development behind these figures call for the industry to embark on a profound industrial transformation, and the core driving force behind this transformation is undoubtedly the digitization and automation of chemical technology. In recent years, major chemical companies both domestically and internationally have been leveraging AI to drive innovation in the chemical industry, thereby promoting its development toward higher quality and greater value. For example, at the World Congress on Chemical Engineering, CNPC presented its newly developed Kunlun Rubber large model, which integrates experimental processes, simulation techniques, and a vast amount of literature data. This model enables a rubber design workflow that can be self-iterating and optimized; it also allows for precise molecular structure design and prediction of formulation properties based on the requirements regarding rubber components and specific performance characteristics. Such a tool can help researchers improve their R&D efficiency and reduce the costs associated with trial and error. Communication companies such as Huawei also leverage their core technological advantages in areas like 5G, cloud computing, and big data to assist research institutions and enterprises in developing more large-scale scientific computing models, thereby accelerating the process from the development to the actual implementation of new chemical products. “Today, AI for science – namely artificial intelligence-driven innovation in science, especially in the area of molecular structure prediction for chemical research – has reached a high level of development. The next step is to capture tacit knowledge and enhance the learning and generalization capabilities of large AI models, in order to overcome the ‘black box problem’ associated with such models. ” **Feng Enbo, deputy director of the Innovation Center for Intelligent Control in Process Manufacturing, believes that the next generation of industrial AI will be like \"super PhDs\" – entities possessing both the capabilities of \"humanities graduates\" (in terms of generative language skills) and those of \"science graduates\" (in terms of ability to derive mechanisms). Such AI systems can achieve self-evolution and optimization of models through reinforcement learning, and they can make the outcomes of large-model applications more accurate by uncovering and accumulating tacit knowledge. Reengineering production models to enhance factory flexibility: Numerous digitalization cases in the chemical industry presented at the World Congress on Chemical Engineering show that, in factories, digital technologies such as AI are profoundly reshaping the production models of chemical plants, enabling a shift from a traditional experience-based approach to one based on precise optimization through data-driven methods and intelligent algorithms. Li Bojun, the APC+RTO project manager at Yuntianhua Group, shared at the meeting the gasification RTO optimization project carried out in collaboration with Huawei. He said that chemical production processes involve many nonlinear elements, making it quite difficult to develop mechanistic models; whereas AI-based RTOs reduce reliance on such mechanisms and enable the identification of hidden patterns through large amounts of data. Yuntianhua leverages Huawei’s predictive large-model capabilities to optimize coal blending, generating economic benefits of over 10 million yuan; it also reduces carbon dioxide emissions by more than 20,000 tons per year. “The application of AI algorithms has given traditional chemical companies new vitality. We have made the transition from relying on manual experience in the past to relying on digital systems today, and the focus of production management has shifted from ‘managing people’ to ‘managing data’. ”Li Bojun said, “Next, we will apply the experience gained from RTO optimization to other gasification units within the group, and continue to expand this approach to areas such as synthetic urea and phosphate fertilizer production, thereby continuously improving the company’s digital capabilities.” ” Empowering operation and maintenance management to achieve preventive maintenance of equipment. Currently, the digitalization of the chemical industry has reached an advanced stage; many large enterprises have already completed the initial stages of digitalization and are now facing more complex challenges such as predictive equipment management, as well as issues related to safety and environmental protection. Li Junpeng, President of Huawei’s Chemicals and Building Materials Division, said that chemical companies generally face three major challenges in the area of equipment operation and maintenance management: a lack of the ability to detect abnormal trends early, a lack of tools for analyzing complex failures, and a lack of assistance in formulating maintenance strategies. Currently, Huawei has collaborated with industry partners to develop end-to-end solutions based on its predictive AI models for device condition monitoring, fault diagnosis, lifespan prediction, and maintenance management, enabling the prediction of equipment failures two weeks to two months in advance. In addition, Huawei has also undertaken various initiatives with industry clients in areas such as improving process optimization and developing new products. Based on the preliminary results, AI can indeed help industries enhance safety, efficiency, and environmental sustainability building on their existing foundations. Gao Jinji, an academician of the Chinese Academy of Engineering, proposed a more cutting-edge concept of \"artificial self-healing\". He pointed out that the intelligence development taking place in the industry today is an imitation of the human mind and conscious thinking, with the aim of making machines smarter. The potential of AI goes far beyond serving as an intelligent brain for processing and solving data; it can also evolve in the direction of autonomous self-healing. By imitating the self-healing mechanisms of human unconscious thought, it can conduct rapid, accurate, and intelligent diagnosis based on early warnings, identify the causes and conditions behind faults, and then carry out real-time, precise self-healing adjustments to suppress and eliminate those faults during operation. Gao Jinji believes that the self-healing technology for \"autonomous health\" in such equipment will become an inevitable trend in the future development of AI, representing a new path for the evolution from intelligent process manufacturing to intelligent, safe, and unmanned process manufacturing, thereby achieving true safety and automation. “According to Accenture’s projections, by 2030 AI technology will generate over $400 billion in value for the global chemical industry. This industrial transformation driven by digital intelligence will not only boost operational efficiency by 25% to 40%, but will also systematically reshape every stage, from laboratory research and development to factory production, marking the full onset of an era of a ’silicon-based revolution’ in the chemical industry. ”Hu Jie said.