Rolling Mill Gap Monitoring for Cylinder Wear Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing monitoring systems for rolling mill stands struggle to accurately and efficiently track wear and shape variations across multiple mill stands, leading to unpredictable product quality and reduced production efficiency due to uncontrolled wear cascade effects.
Innovation Solution
Implementing a neural network system that utilizes input data from various sources to predict and adjust cylinder gaps and speeds, integrating physical models and sensor data for precise wear monitoring and timely intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If magnetic sensors are used to monitor workpiece section, then the monitoring system is compact and relatively cheap, but wear to channels leads to shape variations that are not detected and reliability is reduced
Solution Approach 1:
The monitoring system is segmented into multiple independent measurement points along the rolling mill line. Each magnetic sensor monitors a specific location, and the combined data from all segments provides comprehensive coverage of workpiece dimensional changes while maintaining the advantages of individual magnetic sensors.
Solution Approach 2:
The system implements continuous feedback by constantly monitoring workpiece section at multiple points and using this information to detect channel wear patterns. The feedback loop enables real-time identification of wear-induced shape variations that single-point sensors would miss, thereby improving reliability without abandoning magnetic sensor technology.
2Measurement precision
If optical sensors are used to monitor workpiece section, then detection precision is improved, but the system requires significant space and increases the dimensions of the rolling mill
Solution Approach 1:
The system merges multiple magnetic sensors at different positions into a unified monitoring architecture. By combining measurements from several compact magnetic sensors along the rolling line, the system achieves comprehensive monitoring capability that would otherwise require a single large optical sensor, thus maintaining high measurement precision while minimizing space requirements.
3Productivity
If the number of mill stands working in series is increased to improve productivity, then output is improved, but the problem of wear and its cascade effects is aggravated
Solution Approach 1:
The monitoring system provides continuous feedback on workpiece section and dimensional changes at each mill stand. This feedback enables real-time detection of wear patterns and their cascade effects across the series of mill stands, allowing operators to maintain control over wear propagation while operating at high productivity levels.
Solution Approach 2:
The system performs preliminary detection of wear trends by continuously monitoring workpiece dimensions before significant wear cascade effects occur. This early warning capability allows for preventive maintenance scheduling and gap adjustments before wear propagates through the entire mill line, maintaining reliability in high-productivity configurations.
4Manufacturing precision
If adjustments to cylinder gap are made frequently to maintain workpiece features, then manufacturing precision is improved, but production continuity is disrupted
Solution Approach 1:
The system implements continuous feedback monitoring of workpiece section and channel wear, enabling precise detection of when adjustments are actually needed. This continuous information flow allows for optimized adjustment timing that maintains manufacturing precision while minimizing disruptions to production continuity, as adjustments are made only when wear reaches critical thresholds rather than on fixed schedules.
Data Source
Figure 1
AI summary
Method for monitoring wear to cylinders of the cages of a rolling mill, in particular for bars or rods, comprising the following steps: reading by a neural network (5) of a plurality of data (1, 2, 3) relating to the initial conditions (3) of one or more rolling cylinders, in particular one or more pairs of cylinders each belonging to a rolling cage, to the settings (2, 3) and to the running of the process (1); generation by the neural network (5) of signals (6) relating to the state of wear of the cylinders.