Thread Content
Sinochem News Recently, the Artificial Intelligence Technology Achievements Exchange Meeting of the 2026 National Petroleum and Chemical Industry Science and Technology Innovation Conference was held in Dalian. Participants said that artificial intelligence is reshaping the R&D and production chain of the petrochemical industry and bringing tangible benefits to the development of the industry. In the research and development process, artificial intelligence is subverting the traditional "trial and error" research and development model. Luo Yi, chief expert at PetroChina Petrochemical Research Institute, said that his team applied a high-throughput computing platform in the design of olefin polymerization catalysts. The generative AI model can automatically propose candidate catalyst structures based on target performance, which is expected to significantly shorten the research and development cycle. Ou Qi, deputy director of the Basic Research Department of Sinopec Petrochemical Research Institute Co., Ltd., said that the molecular representation pre-training model built by his team based on self-supervised learning, relying on 60 million molecular conformation data, can predict physical properties such as dielectric constant, viscosity, thermal conductivity, etc. with high accuracy. The single-system prediction is less than 1 second, which greatly reduces research and development costs. It has been successfully used in the research and development and experimental verification of environmentally friendly immersed coolants. In the production process, artificial intelligence is creating real benefits. Hu Yuxiang, Marketing Solutions Director of SUPCON Technology Co., Ltd., introduced TPT, a large time series model dedicated to the process industry, which uses an AI large model as the core to build an intelligent closed loop of "identification-evaluation-decision-execution". In a certain 10-million-ton oil refining project, after conducting heat exchange network simulation and optimization transformation through the TPT large model, the annual fuel gas saving was 15 million yuan, and the annual benefit was approximately 40 million yuan. Lu Weifeng, deputy director of China Petroleum Exploration and Development Research Institute, said that after the large pumping well production optimization model developed by the company was applied to more than 3,800 wells in Xinjiang oil fields, the accuracy of working condition diagnosis exceeded 93%, and fault troubleshooting was reduced from days to minutes.