Rolling Mill Control Using Forward Modeling for Coil Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Rolling mill production systems face inefficiencies due to the inability to predict and ensure that coil-shaped end products meet customer specifications regarding geometry and weight, leading to potential scrap and reduced efficiency, as changes in product dimensions and weight during processing are not adequately accounted for until the final stages.
Innovation Solution
A method for controlling rolling mill production systems that involves modeling the processing of slabs through multiple production units, using physical data to predict and check conformity with quality criteria, allowing for early interruption of production if specifications are not met, and adjusting processing parameters through positive list identifiers to accommodate deviations, thereby minimizing waste and optimizing production.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the product is processed through all production units without intermediate checking, then the production flow remains continuous and simple, but the end product may not meet customer specifications leading to scrap and wasted processing steps
Solution Approach 1:
The patent applies preliminary action by performing a forward model calculation before the product actually passes through all production units. The control device calculates expected final product properties based on current physical data and planned processing steps, allowing early detection of potential specification violations. This enables intervention before unnecessary processing occurs, resolving the contradiction by ensuring manufacturing precision through prediction without requiring complex real-time monitoring of every processing stage.
Solution Approach 2:
The patent implements feedback by continuously comparing the modeled expected final product properties against customer specifications. When a deviation is detected, the system provides feedback to interrupt further processing or adjust parameters. This feedback mechanism ensures that only products meeting specifications proceed through all production units, eliminating scrap while maintaining a relatively simple control structure based on periodic checking rather than continuous complex monitoring.
2Reliability
If physical data is measured at every production unit, then product quality can be continuously monitored, but measurement time and processing complexity increase
Solution Approach 1:
The patent applies copying by creating a virtual model (digital twin) of the product's processing path through forward model calculations. Instead of physically measuring the product at every production unit, the system copies the essential physical data taken at fewer points and uses this copied information to simulate and predict the final product properties. This approach ensures reliability through accurate modeling while significantly reducing measurement time and processing complexity compared to actual physical measurements at every stage.
Solution Approach 2:
The system performs preliminary modeling calculations using physical data measured at selected points before the product completes its journey through all production units. This preliminary action allows the system to predict final properties and determine quality assurance outcomes without waiting for actual completion and measurement of all processing steps, thereby reducing loss of time while maintaining reliability.
3Loss of substance
If the production process is interrupted early when specifications are not met, then scrap is minimized, but additional control checks and modeling increase system complexity
Solution Approach 1:
The patent replaces mechanical/physical trial-and-error processing with a computational modeling system. Instead of actually processing material through all production units and then discovering defects, the system uses forward model calculations to substitute for the physical processing in prediction. This computational approach minimizes scrap by identifying non-conforming products before full processing occurs, while the modeling system complexity is offset by eliminating the need for extensive post-processing inspection and rework infrastructure.
4Manufacturing precision
If forward model calculation is performed for every product, then specification compliance is ensured, but computational load and processing time increase
Solution Approach 1:
The patent applies partial action by performing forward model calculations selectively rather than for every single product. The control device measures physical data at selected points and performs modeling only when necessary to determine specification compliance. This approach ensures manufacturing precision for critical decisions while reducing computational load and energy consumption by avoiding redundant calculations for obviously compliant products or when measurement data is insufficient to warrant full modeling.
Data Source
AI summary
A method of controlling a rolling mill production system for production of a coil-shaped end product from a slab, the production including processing the slab by sequentially arranged production units, the processing by the production units resulting in a respective strip-shaped product having physical data, the method including modeling, under consideration of the physical data, the processing of a testing product by a plurality of production units arranged downstream from a given production unit while taking into account the physical data. If the modelling shows that, under consideration of the physical data, one of the products resulting from processing by the downstream production units does not meet a predetermined quality criterion, the intended manufacture of the product is interrupted and a signal relating to the interrupting is outputted.