Roller Mill Wear Prediction Model for Flat Rolled Goods
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current rolling mill technologies face challenges in predicting and managing roll wear accurately, leading to frequent and often premature roll changes, which reduce productivity and risk non-compliance with quality specifications due to the inability to measure wear in real-time during the rolling process.
Innovation Solution
An operating method for a rolling mill that uses a wear model to predict future roll wear by analyzing expected future variables and actual wear data, providing critical information to operators for timely decision-making on roll changes and adjusting the wear model based on actual wear data for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If roll change intervals are extended to reduce downtime and improve productivity, then productivity increases, but the risk of excessive wear and quality specification non-compliance increases
Solution Approach 1:
The wear model predicts future roll wear before it actually occurs, allowing operators to plan roll changes in advance based on predicted wear levels rather than reacting to actual wear measurements. This preliminary prediction enables optimization of roll change intervals to maximize productivity while maintaining quality standards.
Solution Approach 2:
The system continuously monitors actual rolling parameters and feeds this information back to the wear model, which adjusts its predictions accordingly. This feedback mechanism ensures that wear predictions remain accurate even as rolling conditions change, enabling reliable decision-making about roll change timing.
2Reliability
If roll changes are performed frequently in a conservative manner to ensure quality, then quality specification compliance is maintained, but productivity decreases due to increased downtime
Solution Approach 1:
By predicting future wear levels before they occur, the system allows operators to extend roll change intervals confidently, performing changes only when predicted wear reaches critical thresholds. This eliminates the need for conservative early changes while maintaining quality compliance.
Solution Approach 2:
The system changes the parameter from actual wear measurement (which requires roll removal) to predicted future wear (which can be calculated during operation). This parameter transformation enables continuous monitoring without stopping production, thereby maintaining productivity while ensuring quality.
3Measurement precision
If direct measurement of roll wear is performed by removing the roll, then measurement accuracy is achieved, but production downtime increases
Solution Approach 1:
The wear model acts as an intermediary that translates easily measurable rolling parameters (rolling force, speed, material properties) into accurate wear predictions. This intermediary approach provides measurement accuracy without requiring direct physical measurement that would halt production.
Solution Approach 2:
The system replaces the mechanical measurement process (which requires removing and physically measuring the roll) with a computational model that calculates wear based on operational parameters. This substitution eliminates the need for production stoppage while maintaining measurement accuracy.
4Measurement precision
If the wear model is adapted during commissioning but not corrected during operation, then initial model accuracy is achieved, but prediction accuracy degrades over time due to unaccounted wear variations
Solution Approach 1:
The system continuously feeds actual rolling parameters and wear measurements back to the wear model during operation, allowing the model to adapt to changing conditions and maintain accuracy over extended periods. This continuous feedback prevents model degradation that would occur without operational corrections.
Solution Approach 2:
The wear model transitions from a static configuration (adapted only during commissioning) to a dynamic system that continuously updates its parameters based on operational data. This dynamic adaptation ensures the model remains accurate throughout the roll's service life, not just during initial calibration.
Data Source
Figure 1
Figure 2
Figure 3~6
AI summary
The invention relates to a roller mill for rolling flat rolled goods (1), comprising at least one roll stand (2) having rollers (3). The flat rolled goods (1) are rolled in the roll stand (2). During rolling of the flat rolled goods (1) in the roll stand (2), actual parameters (I) of the flat rolled goods (1) and/or the roll stand (2) are captured. Based on an initial state of at least one roller (3) of the roll stand (2), an expected current actual wear (V) of the at least one roller (3) is derived by means of a wear model (8), using the actual parameters (I) captured during the previous roller run of the particular roller (3). Based on the expected current actual wear (V), using the wear model (8) and expected actual future parameters for flat rolled goods (1) to be rolled in the future, an expected future actual wear (V') of the at least one roller (3) is derived for at least one position (P) of the future roller run. The position (P) of the future roller run, the expected future actual wear (V'), and/or information derived from at least one of said values are made available to an operator (9) of the roller mill.