Bayesian Texture Assessment for Coating Formulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for formulating color matches in coatings are inefficient due to their reliance on brute force approaches, neural networks that are inflexible and require significant maintenance, and often exclude necessary pigments, leading to suboptimal results.
Innovation Solution
A Bayesian belief system that partitions processing steps into smaller, multidirectional pieces, utilizing a feed-forward design to identify texture qualities and physical properties of coatings, allowing for flexible pigment identification and formulation, and incorporating a tolerance module to ensure accurate matches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If brute force combinatorial approaches are used to validate all available pigments, then complete pigment selection is achieved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent segments the pigment selection process into distinct phases: initial broad screening using spectral data, intermediate filtering based on texture and coarseness properties, and final validation. This multi-stage approach divides the computationally intensive brute force method into manageable segments that can be processed efficiently at each stage, reducing overall processing time while maintaining complete pigment selection.
Solution Approach 2:
The system performs preliminary actions by pre-characterizing pigments with texture properties, coarseness metrics, and spectral signatures before the actual formulation process. This pre-processing creates a structured database that enables rapid filtering and comparison during formulation, eliminating the need to analyze all pigment combinations from scratch and significantly reducing computation time.
2Measurement precision
If neural networks are used to select color matches, then both linear and non-linear relationships are addressed, but system flexibility and adaptability decrease
Solution Approach 1:
The patent implements a dynamic, multi-stage formulation system where the approach can adapt at each stage. Unlike rigid neural networks, the system can switch between different analysis methods, adjust filtering criteria, and incorporate new pigment data dynamically. The formulation engine allows users to modify parameters, add constraints, and re-run analyses with updated requirements, providing the flexibility and adaptability that fixed-structure neural networks lack.
3Productivity
If predefined subsets of pigments are selected for use, then processing speed increases, but necessary pigments may be excluded
Solution Approach 1:
The system uses parameter changes strategically at different stages. Initial filtering uses spectral parameters to quickly eliminate incompatible pigments, maintaining speed. Then texture parameters and coarseness metrics are applied to further refine the candidate set. Finally, the system adjusts parameters to ensure all potentially necessary pigments are considered for validation, balancing speed with completeness and accuracy.
4Ease of operation
If user interaction is minimized in the formulation process, then ease of operation improves, but ability to correct errors and adjust parameters decreases
Solution Approach 1:
The formulation system incorporates feedback mechanisms that allow users to review results, identify errors, and request corrections. The system provides feedback on formulation quality, highlights potential issues, and enables iterative refinement. Users can adjust parameters, add constraints, or modify target specifications based on feedback, maintaining ease of operation while preserving the ability to correct errors and improve results.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
A computer implemented method. The method includes identifying, using a processor, a texture in a target coating, wherein identifying comprises applying a Bayesian process, and assigning, using the processor, a texture value adapted for use by one of a search engine and a formulation engine.