Failure Prediction Sampling Around a Failure Centroid
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current statistical yield calculation methods do not scale well for problems with many parameters, making it difficult to effectively detect failures in manufactured objects.
Innovation Solution
A computer-implemented method that classifies samples with uniformly distributed parameter values, determines the centroid of failing samples, generates new samples around the centroid, and uses a nearest neighbor vector space for classification, allowing for efficient failure prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current statistical yield calculation methods are used, then failure detection is possible, but the method does not scale well for problems with many parameters
Solution Approach 1:
The patent segments the high-dimensional parameter space by identifying and focusing on the critical subspace where failures occur. Instead of uniformly sampling all parameters, the method divides the problem into: (1) identifying the failure centroid in the high-dimensional space, and (2) sampling primarily in the directions and regions most likely to produce failures. This segmentation allows accurate failure detection without requiring exhaustive sampling of all parameter dimensions.
Solution Approach 2:
The patent transforms the high-dimensional parameter sampling problem into a different dimensional approach by using a centroid-based method. It identifies the centroid of the failure region in the parameter space and then generates samples based on the distance and orientation from this centroid. This dimensionality transformation allows the method to efficiently handle hundreds of parameters by focusing computational effort on the most relevant regions rather than uniformly distributing samples across all dimensions.
2Measurement precision
If uniform sampling is used across all parameters, then comprehensive coverage is achieved, but the number of samples required becomes computationally infeasible
Solution Approach 1:
The patent performs preliminary action by first identifying the failure centroid and calculating the critical directions in parameter space before generating samples. This preliminary step involves determining which parameter combinations are most likely to cause failures, and then using this information to guide the sampling process. By doing this preparation work upfront, the method avoids the need for exhaustive uniform sampling, significantly reducing computation time while maintaining comprehensive coverage of critical regions.
Solution Approach 2:
The patent changes the sampling parameters from uniform distribution across all dimensions to a non-uniform distribution centered on the failure centroid. It modifies the sampling strategy by using the centroid coordinates and critical direction vectors to generate samples that are concentrated in the most relevant regions of parameter space. This parameter change in the sampling approach allows comprehensive failure detection with far fewer samples than uniform sampling would require.
3Manufacturing precision
If the number of parameters increases to hundreds, then model refinement is possible, but analysis time becomes unreasonable
Solution Approach 1:
The patent segments the analysis process into identifying the failure centroid and then sampling based on critical directions, rather than attempting to analyze all parameter combinations. This segmentation enables the use of refined models with hundreds of parameters by focusing computational resources on the most critical parameter interactions, thereby maintaining analysis speed despite increased model complexity.
Solution Approach 2:
The patent applies dimensionality change by transforming the problem from sampling in hundreds of dimensions uniformly to sampling based on a centroid and critical directions. This approach allows the use of highly refined models with hundreds of parameters while keeping analysis time reasonable, as the method efficiently navigates the high-dimensional space by focusing on the failure-prone regions rather than exhaustively exploring all dimensions.
Data Source
AI summary
A computer-implemented method and computing system are provided for failure prediction of a batch of manufactured objects. The method includes classifying, by a processor sing a simulation, a set of samples with uniformly distributed parameter values, to generate sample classifications for the batch of manufactured objects. The method further includes determining, by the processor, a centroid of failing ones of the samples in the set, based on the sample classifications. The method also includes generating, by the processor, a new set of samples with a distribution around the centroid of the failing ones of the sample in the set. The method additionally includes populating, by the processor, a nearest neighbor vector space using the new set of samples. The method further includes classifying, by the processor, the new set of samples by performing a nearest neighbor search on the nearest neighbor vector space using a distance metric.