Forged Component Quench Prediction Using Geometry-Based ML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Forging processes, particularly quenching, are complex and time-consuming, often leading to unsuitable components due to distortions like cracks and warping, and require extensive modeling for each geometry variation, which is inefficient and resource-intensive.
Innovation Solution
A computing device uses a trained machine learning model to predict the impact of quenching processes on forged components by associating heat transfer coefficients with component geometries, allowing for rapid prototyping and resource-efficient prediction of mechanical properties without full computational modeling for each geometry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full computational fluid dynamics modeling is performed for each geometry variation, then prediction accuracy is improved, but computational time and resource consumption increase significantly
Solution Approach 1:
The system performs full CFD modeling in advance for a comprehensive set of reference geometries and stores the results in a database. When a new geometry needs prediction, the system retrieves pre-computed results from the database rather than performing new expensive simulations, thus achieving fast predictions without sacrificing accuracy for common geometry types.
Solution Approach 2:
The system creates a database of pre-computed CFD results for reference geometries and uses these copied results to predict properties of new geometries through similarity matching, avoiding the need to perform expensive full simulations for each new geometry case.
2Measurement precision
If separate modeling is performed for each geometry variation, then prediction accuracy for specific geometries is improved, but overall productivity decreases due to repetitive modeling work
Solution Approach 1:
The system builds a universal database that stores CFD results for multiple different geometries. This single database serves multiple purposes: it provides accurate predictions for any queried geometry, enables rapid prototyping by quickly retrieving results for variations, and supports supplier selection processes, thereby achieving multi-functionality that boosts overall productivity.
Solution Approach 2:
By pre-computing and storing CFD results for various geometries in advance, the system eliminates the need to perform repetitive modeling work for each geometry variation. Users can quickly query the database for predictions on different geometries without restarting the full modeling process, significantly improving productivity.
3Manufacturing precision
If detailed quenching process modeling is performed for each component, then manufacturing precision is improved, but device complexity and operational difficulty increase
Solution Approach 1:
The system extracts and stores key quenching parameters and geometric features from the complex CFD simulations in a structured database format. When predicting for a new component, only the essential extracted features are needed for similarity matching and prediction, simplifying the input requirements and reducing the complexity of operating the system while maintaining high manufacturing precision.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The approach reduces computational resources and time required for modeling, enabling rapid identification of better geometries and efficient supplier selection by predicting quench distortions and mechanical properties with high accuracy.
Implementation Method 1
The computing device may extract the features of the geometry of the target forged component, providing the extracted features of the target geometry for the target forged component as an input to the trained machine learning model. The trained machine learning model may, upon receiving the extracted features of the target geometry for the target forged component, identify one or more of the training forged components having a similar geometry
Implementation Method 2
CFD relies on partial differential equations (PDEs) to solve a mathematical model of how the fluid dynamics may facilitate transfer of heat from the hot forged component to the quenching medium
Implementation Method 3
Quenching may, for components forged from metals such as iron and steel, increase the hardness of the forged component relative to forged components that do not undergo quenching
Implementation Method 4
CFD relies on partial differential equations (PDEs) to solve a mathematical model of how the fluid dynamics may facilitate transfer of heat from the hot forged component to the quenching medium
Implementation Method 5
A mechanical finite element method (FEM) model may utilize the output of the CFD model to produce data representative of location-specific quench impacts on the hot forged component
Data Source
AI summary
In general, various aspects of the techniques enable predictive modeling for forged components. A computing device comprising a memory and a processor may be configured to perform the techniques. The memory may store a trained machine learning model that associates training features extracted from data representative of a plurality of training forged components to a plurality of training model results. The memory may also store data representative of a target forged component. The processor may perform a geometrical analysis with respect to the data representative of the target forged component to extract target features, and apply the trained machine learning model to the target features to obtain predicted model results for the target forged component. The processor may also output the predicted model results.