3D Printing Feed Rate Prediction for Overhangs and Thin Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
3D printing faces challenges in accurately determining feed rates for features like overhangs and thin features, which are difficult to calculate due to their anisotropic nature, leading to potential drooping or sagging issues and requiring laborious trial-and-error methods.
Innovation Solution
A method using a trained machine learning system that processes spatial toolpath data, closed loop gain data, and feed rate data to determine revised feed rates for specific toolpath segments, including overhangs and thin features, allowing for automatic determination of optimal feed rates without multiple print trials or visual inspections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If trial-and-error printing methods are used to determine feed rates for anisotropic regions, then feed rate accuracy may be improved through visual assessment, but time consumption and labor increase significantly
Solution Approach 1:
The system performs preliminary analysis of the 3D model to identify anisotropic regions (overhangs, thin features, outer perimeters) before printing begins. Feed rates are pre-calculated based on geometric features and printed layer data, eliminating the need for trial-and-error adjustments during the printing process.
Solution Approach 2:
The system uses feedback from printed layer data to continuously refine feed rate calculations. By analyzing actual printing results and comparing them with expected outcomes, the system adjusts feed rates for subsequent layers to achieve optimal print quality without manual intervention.
2Ease of manufacture
If standard feed rate calculations are applied to anisotropic regions, then printing process simplicity is maintained, but print quality deteriorates due to drooping or sagging
Solution Approach 1:
The system applies different feed rate strategies to different regions of the 3D model based on their geometric characteristics. Anisotropic regions such as overhangs, thin features, and outer perimeters receive customized feed rates, while isotropic regions use standard calculations, thereby maintaining print quality without unnecessarily complicating the overall process.
Solution Approach 2:
The system dynamically changes feed rate parameters based on local geometric features and printing conditions. By adjusting feed rates according to the specific characteristics of each region (overhang angle, feature thickness, position), the system prevents drooping and sagging while maintaining a relatively simple automated process.
3Measurement precision
If multiple print trials are conducted to optimize feed rates for different part types, then feed rate accuracy improves, but productivity decreases due to repeated printing
Solution Approach 1:
The system creates a digital representation (copy) of the 3D model and performs virtual analysis to determine feed rates before actual printing. By working with digital copies and printed layer data rather than repeatedly printing physical prototypes, the system achieves accurate feed rate determination while maintaining high productivity.
Solution Approach 2:
Feed rate optimization is performed in advance through automated analysis of the 3D model geometry and simulated printing conditions. This preliminary determination eliminates the need for multiple physical print trials, thereby achieving both accurate feed rates and high productivity.
Data Source
AI summary
Systems for and methods of providing a feed rate for three-dimensional printing a part are presented. The disclosed techniques include: obtaining computer readable toolpath instructions for the part, where the toolpath instructions specify a nominal feed rate for a toolpath segment and spatial toolpath data of the toolpath segment; providing an input including the spatial toolpath data to a trained machine learning system, where the trained machine learning system has been trained using training data including: training spatial toolpath data, training closed loop gain data, and training feed rate data; obtaining a revised feed rate for the toolpath segment different from the nominal feed rate for the toolpath segment, where the revised feed rate is output from the trained machine learning system; and providing revised computer readable toolpath instructions, where the revised machine learning toolpath instructions include the revised feed rate.