In recent years, China's iron and steel industry is facing great changes in the internal and external environment, and users have higher and higher requirements for product quality, especially in the automobile industry, high-end home appliances and other industries, and the personalized demand for products is increasing day by day. At the same time, the industry competition faced by domestic steel enterprises is also intensifying. All iron and steel enterprises attach great importance to quality management. However, the traditional quality management methods have been difficult to meet the requirements of current users for product quality. It has been recognized by more and more iron and steel enterprise managers to make full use of big data to improve the efficiency of quality management and thus enhance the product competitiveness of enterprises.
First, the current quality management problems in iron and steel enterprises
The information systems of most domestic iron and steel enterprises are implemented in stages, production lines and regions, and their functions are not systematic and perfect. Quality design can not fully reflect the individual needs of customers; Production quality performance data are scattered in different information systems, so the quality information between upper and lower processes can not be connected and shared, and the information and quality information in the process of product realization are difficult to trace; There is a lack of effective monitoring on the process quality control of product realization, and it is impossible to track, transfer, trace and improve the verification across processes, and it is impossible to obtain information and take countermeasures in time when the process is abnormal; The process parameters can not be completely judged, which is inefficient, unrepresentative and inaccurate; R&D and process technicians can't get the whole process data completely and quickly, and the system can't provide support for quality design and analysis improvement, which leads to many problems such as low efficiency and unsatisfactory effect. In order to realize the seamless connection with the market and users and enhance the competitiveness of enterprises, iron and steel enterprises must innovate quality management, improve the efficiency and effect of quality management such as quality design, quality judgment and quality improvement, and meet the individual needs of customers. It is necessary to collect and integrate the production quality data scattered in various production lines and systems on a big data platform, and build an integrated quality management information system of "customer demand identification → quality design based on individual demand → process quality monitoring → quality judgment including process parameters → whole process quality analysis and improvement" on this platform to support the whole process quality control and multi-business collaboration.
Second, establish an enterprise-level big data management application platform.
Establish an enterprise-level big data management application platform, namely the factory database. According to the requirements of quality management business, establish quality data collection rules, and collect all the data in the process of product realization, including all the quality information of raw and auxiliary materials procurement, steelmaking, continuous casting, hot rolling, cold rolling, product delivery, sales and user use, on the big data management application platform to centrally and uniformly manage quality data. 1. data acquisition data acquisition can be divided into process real-time data acquisition and product quality data acquisition. According to the set collection requirements, automatically collect data including enterprise information system and field detection instruments. For some common events, states, etc. If it cannot be collected automatically, a corresponding manual data input page is set in each data collection service, and the operator can input the corresponding data according to the actual situation. Data collection is to sort out and summarize the collected production process data and tracking data in a certain format. In steelmaking and continuous casting, the real-time production process data on the production line corresponding to the production heats are recorded with the production heats as the main collection object and the billet number as the collection object. Hot rolling takes the number and length of batches (coils) as tracking units, and accurately collects production process data. The data collection of cold-rolled film-coated pickling, hot-dip galvanizing and color coating production lines is based on accurate material tracking, and the real-time production process data of corresponding strip steel measuring points on the production line are recorded with the steel coil number and strip steel length as tracking units, and the production process data are matched with the steel coil number and strip steel length. 2. Data processing Data processing is a rule customized according to process characteristics and analysis requirements, which makes data become effective information. Establish the relationship between the corresponding data and store them according to the requirements of the storage model. The application platform of big data management can match the process parameters to the corresponding position of slab or coil, so as to realize the collection and matching of quality data of each production line (that is, the time axis is transformed into the position axis).
Third, the application and innovation of quality management under big data
By building an enterprise-level big data management application platform, the product is implemented with process quality information collection, quality design, quality monitoring, online quality judgment, process quality traceability, quality analysis and improvement. Monitor product quality in real time and evaluate the quality level of each production line. Establish a database of related products and metallurgical specifications in the quality management information system as the basis for quality judgment and improvement. Solve the problems of quality control, process optimization and quality analysis improvement through quality management application software and analysis tools.
1, quality design based on big data
Using big data management application platform to establish centralized and unified product specification database and metallurgical specification database to realize standardization and modularization of product quality design. Product specification library module refers to the modular management of product quality design specification database and process design rules. The establishment of product and metallurgical specification library needs to clarify the definition of product essential attributes, product use requirements, special needs of users and other matters, and standardize the management of product process control from steelmaking to final product, such as process path design, production process target control parameter design, product quality control requirements, performance sampling judgment standards and so on. Carry out the concept that product quality is design, design the product quality based on the historical data in the application platform of big data management, and determine the best product design and process design through historical data in the aspects of composition design, process parameter selection and process route determination. The ERP system completes the product design, and the quality information management system completes the process design and the supplementary design of customers' special needs. That is, after checking and supplementing the quality design results released by ERP, a complete product manufacturing process control target, inspection and judgment standard is formed. The results of quality design can automatically form technical documents such as quality plan and control plan according to the prescribed format. The quality management information system supports the rule verification of quality design based on historical data, that is, after the quality design is completed, the rules of quality design are verified with historical data, so as to accurately evaluate the order-taking ability of future products.
2. Use big data to monitor and evaluate process quality.
(1) Based on the real-time big data platform, according to the parameter values in metallurgical specifications, using SPC judgment rules, the important process parameters affecting product quality can be monitored and warned online. Establish a process quality early warning system, provide timely changes and early warning information of important process parameters in the manufacturing process to field operation and quality management positions, and automatically alarm quality abnormal events.
(2) Monitor and analyze key process parameters through SPC rules, and automatically generate control charts and evaluation reports by maintaining judgment standards.
(3) Develop the quality evaluation model of production line, integrate process parameters, product index parameters and production equipment, and automatically evaluate the quality control capability index of production line regularly, so as to facilitate the continuous improvement of product quality.
(4) For some process quality parameters that cannot be directly measured, the software measurement model is used to predict, and the parameters are unified into the monitoring parameters for monitoring.
(5) Establish an expert quality diagnosis system. When there is a quality problem in the production line, use big data to quickly locate the process and key process parameters that cause the problem, and put forward a pre-diagnosis report.
3. Apply big data to realize automatic quality judgment.
Automatic determination of product quality: including hot rolling, cold rolling, casting grading determination of coated products, process product determination and ex-factory inspection determination. When the product production is completed, the quality management information system will automatically judge the product quality according to the pre-maintained quality inspection judgment rules. The data used in the judgment include order information, steel grade information, product physical and chemical inspection results, process quality parameters, process abnormal events, product size, surface quality data, etc.
(1) grading judgment of slab quality: according to the judging rules of slab quality, the process parameters of steelmaking converter, alloy trimming station, LF furnace, RH and continuous casting, and the detection results of slab surface quality, the grading judgment of slab quality is completed, and the final comprehensive quality results will be matched to each slab.
(2) Automatic determination of product surface defects: The automatic determination of steel coil surface quality is based on the accurate identification of surface defects by surface detection system and the maintenance of a set of perfect surface defect detection rules, and finally the automatic determination of defects detected by instruments is realized.
(3) Factory product quality judgment: According to the product quality judgment rules, the system data and pictures of the surface quality inspector of the hot rolling production line are collected to accurately identify various defects and realize automatic judgment. According to the system data and pictures of surface quality inspection instruments of each cold rolling production line, according to the judgment standard and combined with the special needs of customers, automatic judgment can be realized.
4. Traceability of process quality and tracking of surface defects
Based on the application platform of big data management, the whole process quality traceability and analysis of products such as steelmaking, continuous casting, hot rolling, cold rolling and painting can be realized. It can be traced according to various conditions such as material, order, time and steel grade. , to obtain process parameters and quality parameters of multiple processes, and to conduct retrospective analysis of process quality data, process setting data and product quality data, comparative retrospective analysis of process parameters of the same batch of materials, and retrospective analysis of process quality parameters across processes, etc. So as to find out the difference of process and quality parameters in the manufacturing process and locate the cause of the problem.
5. Continuous improvement of quality based on big data.
Apply big data management application platform and statistical analysis tools to establish quality data analysis platform for quality management, quality design and process optimization to provide support for process technicians to continuously improve product quality and new product development. Through the application platform of big data management, the management of customer technical files is realized, including the management of customer basic information, customer raw material procurement information, customer product information, customer quality feedback and customer special needs. You can also make statistical analysis of quality objections based on the quality objection database of customer service system, and trace the whole quality of products with quality objections. An efficient and convenient data analysis tool and KPI report generation tool are established to analyze the quality situation. You can automatically generate statistical reports by shift, day, week, month and year. The innovation of enterprise quality management under big data has realized the refinement and efficiency of quality management and greatly improved the efficiency and level of quality management. The big data management platform collects the whole process quality data from raw and auxiliary materials entering the factory to product delivery to users, so as to realize centralized and unified management and efficient utilization of quality data; The improvement of quality design, quality monitoring, quality judgment and quality analysis based on data and analysis on big data platform is more rigorous, accurate and timely, which is conducive to meeting the individual needs put forward by users and providing basic guarantee for fundamentally realizing the transformation and upgrading of variety structure. However, it should also be pointed out that the accurate matching between real-time data and steel coil is very important, and the accuracy of matching will directly affect the accuracy of defect tracking, and ultimately affect the accuracy of quality traceability and disposal, as well as the effect of product quality analysis and improvement.
;