Processing-Condition Search with Fixed-Parameter Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional dimensionality reduction methods are ineffective in searching for optimal processing conditions when some parameters are fixed and cannot be changed, as they alter unchangeable parameters, making it difficult to achieve desired processing results in industrial processing machines.
Innovation Solution
A processing-condition search device that classifies parameters into variable and fixed categories, applies dimensionality reduction separately to each, and uses a learning model to search for optimal variable parameter values while retaining fixed parameters, allowing for efficient retrieval of optimal processing conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional dimensionality reduction is applied to all parameters, then the search problem becomes easier to handle, but unchangeable parameters are also altered which prevents finding optimal conditions for fixed parameters
Solution Approach 1:
The patent divides parameters into two distinct groups: variable parameters (which can be changed) and fixed parameters (which cannot be changed). This segmentation allows the system to apply dimensionality reduction only to variable parameters while preserving fixed parameters unchanged, thus resolving the contradiction between simplifying the search problem and maintaining the optimality of processing conditions.
Solution Approach 2:
The patent extracts fixed parameters from the dimensionality reduction process entirely. By separating fixed parameters out of the feature generation pipeline and using them directly without transformation, the system avoids altering unchangeable parameters while still reducing the dimensionality of variable parameters to simplify the search problem.
2Measurement precision
If the number of parameters is increased to include material characteristics and environmental factors, then the processing result prediction becomes more accurate, but the number of combination patterns becomes enormous requiring more trial-and-error
Solution Approach 1:
The patent applies dimensionality reduction to transform high-dimensional variable parameters into lower-dimensional features while preserving the essential information. This dimensionality change reduces the number of combination patterns from enormous to manageable levels, significantly decreasing the time required for trial-and-error while maintaining prediction accuracy through the use of fixed parameters.
Solution Approach 2:
The patent changes the state of parameters by categorizing them into variable and fixed groups, and applying different processing approaches to each. Variable parameters undergo dimensionality reduction to reduce search space, while fixed parameters are preserved as-is. This parameter transformation enables efficient searching across material characteristics and environmental factors without requiring exhaustive trial-and-error.
Data Source
AI summary
A process-condition search device includes: a parameter classifying unit that classifies a plurality of parameters into a plurality of variable parameters and one or more fixed parameters; a first dimensionality reducing unit that generates, from the variable parameters, first features whose dimension is equal to or smaller than a first dimension; a second dimensionality reducing unit that generates, from the one or more fixed parameters, a second feature whose dimension is equal to or smaller than a second dimension; a machine learning unit that generates a learning model by learning the relationship between the first features, the second features, and a plurality of evaluation values; a third dimensionality processing unit that generates a third feature whose dimension is equal to or smaller than the second dimension from one or more target fixed parameters, which are the one or more fixed parameters; an optimal-processing-condition search unit that uses the third feature and the learning model to search for an optimal value of features of the target variable parameters; and a dimension restoring unit that specifies a retrieved processing condition from the optimal value and the one or more target fixed parameters.