Separating CD and MD Variations in Sheet Scanning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately separating cross-machine direction (CD) and machine direction (MD) variations from scan measurements in sheet-making processes, as these measurements are often mixed due to diagonal traversing paths of sensors, leading to contaminated or distorted CD variation components and incomplete detection of fast MD variations.
Innovation Solution
The system identifies CD and MD variations by comparing power spectra of scan measurements taken at different scanning speeds, using dominant spectral components at the same spatial frequencies for CD variations and temporal frequencies for MD variations, and employs inverse transformation to separate these components effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If scan measurements are taken using diagonal traversing paths, then the sheet property variations are captured across the width of the sheet, but the MD and CD variations become mixed and difficult to separate
Solution Approach 1:
The patent segments the mixed scan measurement signal into separate MD and CD variation components by taking multiple measurements at different scanning speeds. Each measurement captures a different mixture of MD and CD variations, and by segmenting and analyzing these measurements through power spectrum comparison, the system isolates the pure CD variation component from the mixed signal.
Solution Approach 2:
The patent changes the scanning speed parameter to obtain multiple measurements with different temporal frequency characteristics. By comparing power spectra across different scanning speeds, the system identifies spectral components that remain constant (pure CD variations) versus those that change (MD variations), thereby separating the two types of variations through parameter variation.
2Productivity
If a single scanning speed is used, then the measurement process is simple and fast, but fast MD variations cannot be fully detected and CD variations are contaminated
Solution Approach 1:
The patent performs preliminary measurements at multiple scanning speeds before analyzing the CD variations. By collecting measurements at different speeds in advance, the system prepares the data needed to identify and eliminate contaminated spectral components, ensuring accurate CD variation detection without requiring complex real-time processing during the measurement phase.
Solution Approach 2:
The patent creates multiple copies of the measurement process at different scanning speeds. Each copy provides a different perspective on the same sheet variations, and by comparing these copies through power spectrum analysis, the system identifies the true CD variation signal that appears consistently across different scanning conditions.
3Measurement precision
If multiple scanning speeds are used to separate CD and MD variations, then the separation accuracy is improved, but the measurement time and system complexity increase
Solution Approach 1:
The patent makes the scanning system multi-functional by using the same scanner to perform both rapid measurements at multiple speeds for separation purposes and accurate measurements for final analysis. The system universally applies the power spectrum comparison method across different scanning speeds, allowing one device to achieve both speed and accuracy functions that would otherwise require separate systems.
Data Source
AI summary
CD variations and/or MD variations in scan measurements are determined from spectral components of power spectra of scan measurements taken using two or more scanning speeds. Dominant spectral components having the same spatial frequencies identify CD variations and dominant spectral components having the same temporal frequencies identify MD variations. Dominant spectral components are extracted from a noisy power spectrum (PS) by sorting all spectral components into an ordered PS. A first polynomial representing background noise of the ordered PS is used to set a first threshold. Spectral components of the ordered PS that exceed the first threshold are removed to form a noise PS. A second polynomial representing the noise PS is used to set a second threshold. Spectral components of the PS that exceed the second threshold are identified as dominant spectral components of the PS.