Rolling Force Signal Monitoring for Roll Eccentricity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing roll state monitoring systems suffer from accuracy deterioration due to noise and abnormal values, leading to incorrect identification of roll eccentricity and state determination, particularly when relying on single rolling force detection values.
Innovation Solution
A roll state monitor device that detects rolling force variations at multiple rotation positions, accumulates values, and applies correction coefficients to identify roll eccentricity, using a roll gap equivalent value and plastic coefficient for accurate determination, and employs statistical methods for state determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If single rolling force detection value is used to calculate identification value, then calculation process is simple, but identification accuracy of roll eccentricity amount deteriorates due to noise and abnormal values
Solution Approach 1:
The patent divides the rolling force detection data into multiple segments corresponding to different rotation positions of the roll. Instead of using a single detection value, the system collects rolling force values at multiple discrete rotation positions (e.g., 0°, 90°, 180°, 270°), processes each segment separately, and then integrates them to calculate the roll eccentricity amount. This segmentation approach filters out random noise and abnormal values while preserving the actual eccentricity signal.
Solution Approach 2:
The patent employs periodic sampling of rolling force at regular rotation position intervals throughout the roll's rotation cycle. By systematically collecting data at periodic intervals (e.g., every 90 degrees of rotation), the system captures the periodic nature of roll eccentricity while averaging out random fluctuations and abnormal transient values, thereby improving identification accuracy.
2Measurement precision
If multiple accumulated values are acquired by accumulating for each rotation position, then identification accuracy of roll eccentricity amount is improved, but calculation complexity increases
Solution Approach 1:
The patent transforms the raw rolling force detection values into accumulated values by applying accumulation operations across multiple rotation positions. This parameter transformation converts the problem from directly analyzing noisy individual measurements to analyzing smoothed accumulated data, where the signal-to-noise ratio is significantly improved. The accumulation process effectively integrates the periodic eccentricity signal while canceling out random noise.
3Measurement precision
If rolling force variation value is used for each rotation position, then determination accuracy of roll state is improved, but susceptibility to noise and abnormal values increases
Solution Approach 1:
The patent extracts the rolling force variation value by separating it from the absolute rolling force value. Specifically, the system calculates the variation component (e.g., deviation from mean value or difference from reference position) and uses only this extracted variation signal for roll state determination. This extraction process removes the influence of abnormal values and noise present in the absolute force measurements, leaving only the meaningful eccentricity-related variation signal.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A roll state monitor device comprising: rolling force detecting means (6) configured to detect a rolling force signal of a monitored roll selected from an upper roll set (3, 4) and a lower roll set when a rolled material (1) is rolled between the upper roll set (3, 4) and the lower roll set, the upper roll set (3, 4) having at least one roll and the lower roll set having at least one roll; signal extracting means (210) extracting from the rolling force signal a rolling force high frequency signal having a frequency equal to or larger than a predetermined frequency which is set in advance; and roll state determining means (212) configured to determine a state of the monitored roll based on a test result of a statistical test method for a plurality of rolling force values included in the rolling force high frequency signal.