Additive Manufacturing Condition Search Device for Optimizing Process Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In powder bed fusion additive manufacturing, determining optimal additive manufacturing conditions is challenging due to numerous factors such as heat source output, scanning speed, and stacking thickness, requiring extensive time and resources to construct a process window and often resulting in inefficient condition search processes.
Innovation Solution
An additive manufacturing condition search device that includes a defect database, two machine learning units, and a specification unit to output and optimize additive manufacturing conditions, monitor and evaluate modeling processes, and determine defect information, thereby improving the efficiency of searching for optimal additive manufacturing conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If comprehensive additive manufacturing conditions are set to cover wide control factors and ranges, then the completeness of condition search is improved, but the number of evaluations increases and convergence time increases
Solution Approach 1:
The system performs preliminary actions by pre-defining hierarchical condition categories (filling conditions, surface conditions, support conditions) and their sub-parameters before the actual search begins. This preliminary structuring allows the machine learning model to systematically explore the condition space without redundant evaluations, reducing convergence time while maintaining comprehensive coverage.
Solution Approach 2:
The condition search space is segmented into hierarchical categories (filling, surface, support) with multiple levels of sub-parameters. This segmentation allows the system to independently optimize each category using machine learning, preventing the exponential increase in evaluations that would occur if all parameters were searched simultaneously as a single large space.
2Measurement precision
If many evaluation samples are manufactured to determine optimal conditions, then the accuracy of condition optimization is improved, but the cost and time required increase enormously
Solution Approach 1:
Instead of physically manufacturing and evaluating numerous samples to determine optimal conditions, the system creates a virtual copy through machine learning modeling. The ML model learns the relationship between additive manufacturing conditions and outcomes from a limited set of actual evaluations, then predicts optimal conditions without requiring proportional physical verification for each predicted parameter combination.
Solution Approach 2:
The mechanical process of physically manufacturing and evaluating numerous samples is replaced by a computational machine learning system. The ML algorithm processes condition parameters and predicts manufacturing outcomes, substituting the mechanical evaluation process with a computational model that achieves higher accuracy with fewer physical experiments.
3Adaptability or versatility
If many samples with poor manufacturing results are included in evaluations, then the comprehensiveness of condition testing is improved, but the adverse effects on good conditions increase
Solution Approach 1:
The system incorporates feedback mechanisms where the machine learning model continuously learns from evaluation results and adjusts its predictions. When poor manufacturing results are encountered, the model uses this feedback to refine its understanding of condition boundaries, thereby preventing the propagation of harmful effects to subsequent evaluations of good conditions through iterative improvement.
Solution Approach 2:
The system converts the harmful effect of poor manufacturing results into beneficial information by using these failures as training data for the machine learning model. The poor results provide valuable boundary information that helps the model better define the optimal condition space, turning what would be detrimental into a useful learning signal for improving overall condition identification accuracy.
4Ease of operation
If manual work is used to allocate conditions according to control factors, then the flexibility in condition selection is improved, but the time required for allocation increases
Solution Approach 1:
The machine learning system performs self-service by automatically allocating and selecting optimal additive manufacturing conditions based on the control factors and desired outcomes. The system independently processes the condition space, applies learned patterns, and generates allocations without requiring manual intervention, thereby maintaining flexibility while dramatically reducing the time previously required for human expertise in condition selection.
Data Source
AI summary
An additive manufacturing condition search device includes a defect database that accumulates a material, shape information, an additive manufacturing condition, monitoring information during modeling, and defect information in association with each other, a first machine learning unit that outputs an additive manufacturing condition corresponding to material information and device information, and outputs a new additive manufacturing condition from a combination of a plurality of the additive manufacturing conditions and the defect information, a specification unit that causes an additive manufacturing apparatus to perform modeling by the additive manufacturing condition, acquires the monitoring information during modeling, and acquires the shape information and the defect information by inspection of a modeled object, a second machine learning unit in which a model trained by using the defect database as train data estimates defect information of the modeled object from the monitoring information and stores the defect information in the defect database, and a determination unit that determines whether or not the defect information of the modeled object has achieved an evaluation target value.