AI News

SUPCON Builds AOP with TPT Large Model and UCS, Replicating Industrial AI Across 29 Sub-sectors

Securities TimesUpdated Nov 16 18:48

AI summary

SUPCON has integrated its self-developed time-series large model TPT with the Universal Control System (UCS) to build the Autonomous Operating Plant (AOP) system, which has been replicated across industries and plant units to 29 sub-sectors. Wanhua Chemical's Ningbo Chlor-Alkali Company built digital employees based on TPT, reducing total plant alarms by 50% with warning accuracy exceeding 95%; Chambroad Petrochemical achieved advance warnings of process anomalies by 5 to 30 minutes and reduced alarm counts by 90%. SUPCON also completed industrial chain cooperation agreements with Huawei, Alibaba Cloud & Qwen Office, and others.

Body · AI translation

Recently, the "Defining the Next Generation of Industry: AI Empowering New Industrialization In-Depth Tour (Process Industry) and Industrial AI Achievement Release Conference" was held in Ningbo. The event was guided by the Ningbo Municipal Bureau of Economy and Information Technology and jointly hosted by Supcon Technology (688777), the National Artificial Intelligence Application Pilot Base (Petrochemical and Chemical Direction in the Manufacturing Sector), Zhenhai Refining & Chemical, and Wanhua Chemical.

At the conference, Supcon Technology stated that the focus of competition in industrial AI is shifting from single-point technology validation to system-level replicable capabilities. According to the "2026 China Industrial AI Industry Development Research Report," China's industrial AI market is expected to grow from 116.3 billion yuan in 2026 to 303.4 billion yuan in 2030, and technology penetration has entered an accelerated channel.

It is understood that the core of competition in process industry AI lies in whether cross-industry migration can be achieved. Single-point projects can be delivered through customized development, but customization means that every new project must pay implementation costs again; only by consolidating scattered AI capabilities into a reusable, composable general foundation can large-scale deployment be economically feasible.

Supcon Technology's approach is to deeply integrate its self-developed time-series large model TPT with the universal control system UCS to build the autonomous operating plant system AOP, creating a complete closed loop of "perception-analysis-decision-execution." The conference disclosed that the TPT industrial AI foundation platform and the AOP system have been replicated across 29 sub-sectors through two paths: horizontal replication and vertical expansion.

At the conference, several representative process industry enterprises shared practical results after industrial AI entered core production links, covering different industries and different application levels. Wanhua Chemical Ningbo Chlor-Alkali Company built multiple types of digital employees based on the TPT large model; total equipment alarms dropped by 50%, early-warning accuracy exceeded 95%, and processes such as wastewater treatment, precise feeding, and electrolysis load allocation achieved autonomous operation, creating more than 10 million yuan in value annually. Shandong Chambroad Petrochemical's refining units have achieved early warnings for process anomalies 5 to 30 minutes in advance, reduced alarm counts by 90%, and generated more than 5 million yuan in comprehensive annual economic benefits per unit; the company will subsequently expand full-link closed-loop optimization to more units and scenarios. Inner Mongolia China Coal Yuanxing built a "coal quality prediction-risk warning-production optimization" decision chain based on the TPT large model; the gasifier unit fluctuation rate decreased by 35.98%, operating frequency dropped by an average of 85.62%, effective gas output increased by 0.55%, and the comprehensive economic benefit per furnace increased by 8.16 million yuan.

It is understood that the equipment characteristics, working condition complexity, and data foundations of industries such as chlor-alkali, coal chemical, and petrochemical are all different. Supcon Technology stated that results across multiple industries and multiple units mean its replication capability has been validated in cross-scenario migration.

It is worth noting that at the conference site, Supcon Technology completed industrial chain cooperation signings with Huawei, the National Artificial Intelligence Application Pilot Base (Petrochemical and Chemical Direction in the Manufacturing Sector), and Alibaba Cloud & Qianwen Office, and simultaneously held a signing ceremony for the Industrial AI Data Alliance, adding important new strength to the data alliance. At the same time, Supcon Technology's Industrial AI Data Center released the Industrial AI Data Alliance co-built dataset and the Industrial AI Data Sharing Community. (Wen Run)

Source

Securities Times