Neural Network DEM Contact Model for Agricultural Materials
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional discrete element contact models face challenges in accurately predicting the collective mechanical behavior of agricultural materials due to computational complexity, limited accuracy, and limited constitutive flexibility.
Innovation Solution
A neural network-based discrete element contact model is introduced, which combines Discrete Element Method (DEM) simulations with a feed-forward neural network to predict the mechanical behavior of agricultural materials under various conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional discrete element contact models are used to predict mechanical behavior of agricultural materials, then the models can simulate individual particle movement and interactions, but the computational complexity increases and accuracy is limited
Solution Approach 1:
The patent replaces the traditional mechanical discrete element contact model with a neural network-based predictive system. The neural network learns from training data to predict mechanical behavior directly, substituting the computationally intensive mechanical simulation with a faster machine learning inference process while maintaining or improving accuracy
Solution Approach 2:
The patent performs preliminary training of the neural network using extensive DEM simulation data and experimental measurements before deployment. This preliminary action creates a pre-trained model that can quickly predict mechanical behavior without requiring complex real-time computations, thus resolving the contradiction between accuracy and computational complexity
2Measurement precision
If conventional discrete element contact models are used, then the simulations can be performed with current formulations, but the accuracy of agricultural material models is limited
Solution Approach 1:
The neural network model serves multiple functions: it can predict various mechanical properties (stress-strain relationships, failure characteristics, deformation behavior) across different agricultural materials and loading conditions. This universal approach replaces multiple material-specific constitutive models, enhancing both accuracy and flexibility simultaneously
3Productivity
If DEM simulations are performed using conventional techniques, then the simulations can be conducted, but they are computationally intensive and time-consuming
Solution Approach 1:
The patent substitutes the computationally intensive DEM mechanical simulation with a neural network-based predictive system. The neural network, once trained, can predict mechanical behavior in fraction of the time required for full DEM simulations, dramatically improving productivity and reducing computational time
Solution Approach 2:
The patent creates a simplified copy of the complex DEM simulation behavior through the neural network. Instead of running the full complex simulation each time, the system uses the trained neural network copy to predict outcomes rapidly, maintaining accuracy while reducing computational burden
Data Source
AI summary
A system and method are provided for predicting mechanical behavior of agricultural materials, utilizing neural network-based discrete element method (DEM) models. Test datasets are provided corresponding to observed conditions of respective agricultural materials, for training the neural network which is configured according to predetermined formulations of DEM models identified for a particular industrial process. Variables for contact model parameters are generated during the training as best correlating the test dataset with an observed mechanical behavior of the agricultural materials under the observed conditions. Upon receiving a current input dataset for simulation of the industrial process, the identified DEM model is applied with the generated variables for empirically associating the current input dataset with at least one predicted mechanical behavior, and output signals are generated corresponding to the predicted mechanical behavior within at least a partial simulation of the industrial process.