Surface Scan Projection for Local Variation Detection in Mating Surfaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately determine local variations in complex surfaces of fabricated components that need to mate with further components during assembly, particularly when these surfaces lack a CAD representation or are too complex for simple mappings.
Innovation Solution
A method and system that projects surface scan data onto a coordinate system perpendicular to the local normal vector, using techniques like robust Gaussian regression filtering to identify and project local variations, enabling the generation of shims to smooth the surface for proper mating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional areal filtering methods are used for planar or cylindrical surfaces, then the filtering process is straightforward and well-established, but the method becomes impractical for surfaces without CAD representation or surfaces that are too complex for simple mappings
Solution Approach 1:
The patent divides the complex surface into multiple local patches or regions, each of which can be filtered independently using local coordinate systems. This segmentation allows the filtering method to handle complex global geometries by breaking them down into manageable local segments that can be processed using established areal filtering techniques.
Solution Approach 2:
The patent transitions from global surface filtering to local surface filtering by introducing local coordinate systems at each patch. This dimensional approach allows the method to adapt to complex surfaces by defining filtering planes locally rather than requiring a single global mapping, thereby extending applicability to surfaces without CAD representation.
2Adaptability or versatility
If complex mapping functions are developed to handle arbitrary surfaces, then surface filtering becomes possible, but the process becomes impractical and too complex
Solution Approach 1:
Instead of developing a single complex global mapping function, the patent segments the surface into multiple local patches, each with its own simple local coordinate system. This segmentation replaces the need for one complex mapping with many simple local mappings, reducing overall computational complexity while maintaining versatility.
Solution Approach 2:
The patent applies filtering locally to patches rather than attempting to filter the entire surface globally at once. This partial action approach simplifies the mapping requirement by only needing to define local coordinate systems for each patch rather than a comprehensive global mapping, making the process more practical.
3Manufacturing precision
If local variations in surface are not accurately determined, then assembly precision is compromised, but implementing complex filtering methods increases process complexity
Solution Approach 1:
The patent applies local quality by using areal filtering on local surface patches rather than applying a uniform global filtering method. This allows the filtering to adapt to local surface characteristics, accurately capturing local variations in each region while maintaining a relatively simple overall process through the use of standard areal filtering techniques.
Data Source
AI summary
A method for determining local variations in a surface of a fabricated component includes comparing surface scan data for the fabricated component represented in a first XYZ coordinate system to surface design data for the surface of the fabricated component, determining deviation data values for the surface based on the comparing, wherein the deviation data values are represented as deviation data points, selecting a first data point and neighboring data points from the deviation data points, determining an average normal vector for the first data point based on normal vectors of the first data point and the neighboring data points, defining an XY plane of a second XYZ coordinate system perpendicular to the average normal vector for the first data point, and projecting the first data point and the neighboring data points from the first XYZ coordinate system onto the XY plane of the second XYZ coordinate system.