Recently, Phase I of the Full-Process Intelligent Quality Control Center project, jointly developed by IET, Hunan Valin Lianyuan Iron and Steel Co., Ltd. (hereinafter referred to as “Lianyuan Steel”), and Loudi Valin Yunchuang Digital Intelligence Technology Co., Ltd. (hereinafter referred to as “Loudi Yunchuang”), has been successfully accepted.

With the concerted efforts of the three parties, guided by the core philosophy of “platform-based, data-driven, full-line collaboration”, a full-process intelligent quality control platform featuring “one platform, multiple main lines, full empowerment” has been built, establishing a replicable and promotable new paradigm for intelligent quality management in the steel industry.

I. Laying a Data Foundation to Build a Digital Base for Intelligent Quality Control

Centered on a unified industrial internet platform, the project has established a company-level data hub for Lianyuan Steel.

By adapting to various communication protocols including S7OPC, EGD, Modbus, TCP, UDP, and X-COM, the project has completed full-element data collection from plate and coil production lines across Plant 210, Hot Rolling Mill, No.1 Steelmaking & Rolling Mill, Cold Rolling Mill, and Heat Treatment Plant — covering L1, L2, post-steel systems, decision-making systems, surface inspection instruments, flatness meters, and other equipment.Meanwhile, based on the data platform concept, the project has completed data cleaning, integration, governance, and feature extraction. Through spatiotemporal transformation technology, it has realized accurate alignment and reconstruction of full-process data, enabling full-link traceability and connectivity of more than 100 units and 500 nos.ofequipment.Digital products such as digital slab, digital coil, and digital plate have been formed. With high-quality and reusable data service capabilities, the platform provides solid data support for upper-layer applications and breaks down data silos.

II. Multi-line Parallel Development to Create a Closed Loop of Full-Process Intelligent Management & Control

1. Quality Monitoring Main Line: Real-Time Control & Prevention Before Problems Occur

The system integrates AI visual recognition models, multi-modal fusion evaluation models, and deep neural network models to realize online intelligent steel grading for surface quality, composition indicators, performance testing, and process parameters.From slab quality risk rating and hot rolling / cold rolling / heat treatment process compliance judgment to automatic classification of surface defects, the system enables 7×24-hour full-process real-time monitoring, anomaly identification, and automatic interception.It links with the PES system for real-time push and handling of abnormalities, building an intelligent management and control system of “standard+ judgment+handling”.

2. Quality Control Analysis Main Line: Traceability, Diagnosis & Collaborative Optimization

Based on the unique material identifier and full-process genealogy, the system connects the three steps of “analysis, traceability, and collaboration” to achieve data-driven process optimization.The system supports one-click cross-process reverse traceability of multi-dimensional data to quickly and accurately locate the root cause of problems.In addition, using machine learning and multi-modal analysis technologies, the system conducts intelligent diagnosis and root cause analysis for quality defects such as performance issues, cracks, segregation, pitting, and surface inclusions, providing quantitative conclusions for process improvement and continuously promoting quality and process optimization.

3. Collaborative Management Main Line: Online Operation & System Efficiency Improvement

Focusing on the core needs of online business operation for management departments including the Quality Department and Technology Center, core functional modules such as quality operation, quality performance, quality objection, and morning meeting dashboard have been developed.A quality collaborative management mechanism covering all production factors of “man, machine, material, method, and environment” has been established.The system realizes dynamic trend tracking of more than 100 core quality indicators, including qualification rate, scrap & rework rate, defect occurrence rate, and off-schedule rate.A full-process closed-loop management system of “problem discovery – task push – rectification tracking – effect verification” has been constructed, promoting the efficient implementation of quality control and supporting the continuous iteration and optimization of the enterprise quality system.

III. Demonstrating Value & Building a New Intelligent Quality System for Lianyuan Steel

The implementation of this new intelligent quality system has driven Lianyuan Steel’s quality management to transform deeply from experience-driven to data-driven, realizing the intelligent upgrade of consistent quality management and achieving cross-process, full-process, full-specification online unmanned intelligent steel grading.By establishing a full-element management and control mechanism, it realizes pre-control and in-process intervention of quality issues, reduces costs and risks, and sets a model for intelligent manufacturing in the steel industry. It also serves as a typical example of the integration between IET,USTB’s technological R&D and industrial demands.In the future, IET,USTB, Valin Lianyuan Steel, and Loudi Yunchuang will continue to advance the construction of the Intelligent Quality Control Center.By integrating AI large model technology, deepening data intelligent analysis models, expanding intelligent management and control scenarios, upgrading full-process management capabilities, and exploring new paths for intelligent quality control, the three parties will help the steel industry improve quality through intelligence and contribute USTB’s wisdom to the high-quality development of China’s steel industry.