Core Insights - Zhongkong Technology launched TPT 2, a time-series pre-trained transformer model aimed at addressing energy optimization, fault accidents, and product quality fluctuations in the process industry [2][3] - TPT 2 builds on the success of TPT 1, which has been validated in over 110 cases, and is designed to enhance the competitiveness of enterprises in the current economic environment [2][3] - The model is tailored for the specific needs of the oil and chemical industries, providing integrated support from solutions to industrial intelligence [3] Group 1 - TPT 2 focuses on precise perception and dynamic optimization of production processes, contributing quantifiable cost reduction and efficiency improvements in sectors like petrochemicals and chemicals [2][3] - The model has been successfully applied in over a hundred installations across major companies such as China National Petroleum Corporation and China Petrochemical Corporation, demonstrating significant results [3] - TPT 2 achieved an anomaly prediction accuracy of 99.79% and improved ethylene yield by 0.373% at a specific facility [3] Group 2 - The launch event attracted nearly 700 representatives from leading companies, industry organizations, and research investment institutions worldwide [3] - As of the launch date, there were 518 pre-order companies and 112 sets sold, with 3,268 online registered applicants [3] - Industry experts suggest that TPT 2 could lead to the establishment of fully automated factories in the future [4]
全球首个流程工业时序大模型升级面市