3D Print Thermal Deformation Prediction With Recurrent Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing finite element simulations for temperature and deformation calculations in additive manufacturing are computationally intensive and unsuitable for real-time, printing-parallel monitoring due to long calculation times, leading to increased computing time and inability to analyze deviations in geometry during the printing process.
Innovation Solution
A computer-aided method using a trained recurrent neural network to determine temperature distribution and deformation data, which is less computationally intensive and allows for real-time calculations, enabling real-time monitoring and adjustments during the additive manufacturing process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If finite element simulations are used for temperature and deformation calculations, then calculation accuracy is improved, but calculation time increases significantly
Solution Approach 1:
The patent pre-calculates temperature and deformation data using finite element simulations for various printing parameters before actual printing. These pre-calculated data are stored in a database, allowing the system to retrieve results instantly during printing without performing real-time simulations, thus resolving the contradiction between accuracy and calculation time
Solution Approach 2:
The patent creates a digital twin or virtual model of the printing process that replicates the physical printing behavior. By copying the essential thermal and mechanical characteristics into a simplified computational model, the system can predict temperature and deformation outcomes rapidly without requiring full finite element simulations during actual printing
2Manufacturing precision
If real-time monitoring and parameter adjustments are implemented during printing, then manufacturing precision is improved, but computational complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where sensors monitor actual printing parameters (temperature, position, speed) in real-time, compare them with planned parameters, and automatically adjust printing conditions to compensate for deviations. This closed-loop control improves geometric accuracy without requiring complex real-time simulations by using pre-calculated correction strategies
Solution Approach 2:
The patent introduces an intermediate lookup table or database that stores pre-calculated temperature and deformation data for various printing scenarios. This intermediary structure allows the control system to quickly retrieve appropriate parameters without performing complex calculations, reducing computational complexity while maintaining precision
3Manufacturing precision
If geometry deviations are analyzed during printing, then manufacturing precision is improved, but calculation time increases
Solution Approach 1:
The patent pre-calculates expected geometry deviations for various printing scenarios using finite element simulations and stores these predictions in advance. During actual printing, the system only needs to compare real-time sensor data with pre-calculated deviation patterns, enabling rapid detection without performing time-consuming simulations during the printing process
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Computer-implemented method for determining a temperature distribution and deformation data of an additively manufactured component, comprising the following method steps: - reading (S1) input data (INP), wherein the input data comprise o a power value of a print head of a 3D printer at a specified time, o a speed value of the print head at the specified time, o a next position of the print head after a specified time step, and o temperature distribution and deformation data of the component as well as power values, speed values, and positions of the print head at previous times, - providing (S2) a trained recurrent neural network (NN) that is trained to determine a temperature distribution and deformation data of the component depending on the input data (INP),- Determining (S3) the temperature distribution (T1) and the deformation data (V1) of the component at the specified time as a function of the input data (INP) by means of the trained recurrent neural network (NN), and - Outputting (S4) the determined temperature distribution and the determined deformation data (T1, V1) of the component.