NISCO deploys swarm intelligence steel model cutting annual costs 15 million dollars

Nanjing Iron & Steel Co., Ltd. (NISCO) has joined forces with Huawei and other partners to deploy the world's first steel model utilizing swarm intelligence, named YuanYe, across various applications. This deployment has enabled NISCO to shift its core operations from experience-based decision-making to data-driven intelligence collaboration.

NISCO's intelligent operations center, featuring 5-dimensional digital twins, provides integrated production and operations management across spaces, boundaries, organizations, and equipment. Leveraging industry-leading products such as Ascend, Kunpeng, Huawei Cloud Stack, and Pangu models, NISCO has built a strong digital and intelligent foundation, applying AI to every stage of production.

The YuanYe model, which combines multiple AI models with steel industry metallurgical mechanism models, creates a dual-drive, multi-model cluster architecture with a 'right brain' for probabilistic computing and a 'left brain' for precise decision-making.

By integrating sensing data from millions of operational technology (OT) data points across 26 business domains, NISCO has achieved real-time digital twin mapping within its intelligent operations center. This enables the entire steelmaking process to become visible, controllable, and intelligently assessable. Consequently, all of NISCO's key equipment is now connected and digitally controlled, resulting in a 50% reduction in product development cycles and over CNY100 million (approximately US$15 million) in annual cost savings.

When YuanYe is applied to the ironmaking stage, it makes traditionally invisible blast furnaces transparent by utilizing internal sensor data and metallurgical mechanism analysis for real-time dynamic simulations. This shift from reactive remediation to proactive pre-control increases prime-grade molten iron yield by over 10% while cutting unit energy consumption significantly.

In the rolling stage, YuanYe's rolling-force prediction model achieves millisecond-level response, improving thickness tolerance to ±0.2 mm and reducing thickness deviation to 1.9%, thus meeting a world-leading standard. This enhancement leads to a 98% hit rate and a steady annual output of over 2.5 million tons of steel plates.

In quality inspection, the YuanYe system, which combines robots and AI models, doubles metallographic inspection efficiency. Inspection accuracy has reached over 95%, and inspection time per sample has decreased from 10 minutes to 3-5 minutes, increasing overall inspection efficiency by 80%. In energy and environmental control, the YuanYe architecture powers the S-Link Intelligent Agent for Gas and Flue Gas Sulfur Control.

This agent integrates 602 process parameters, multi-modal image, and knowledge capabilities to issue sulfur-risk warnings an hour in advance with 95% accuracy, leading to a nearly 50% reduction in natural gas co-firing consumption and over CNY600,000 (approximately US$90,000) in annual economic benefits.