Loading [Contrib]/a11y/accessibility-menu.js

Welcome
What's New
Discover new features and enhancements.
Get Started
New to Twin Activate? Learn the basics here.
Tutorials
Start using Twin Activate with our interactive tutorials.
User Guide
Learn about the features and functionality available in Twin Activate.
Reference Guides
Explore the Extended Definitions, OML Guides, Block Library, API Guide and Glossary.
- Extended Definitions for Advanced Users
  PDF file with in-depth information on key topics in the User's Guide.
- Scripting Guide for OpenMatrix Language
  OpenMatrix is a mathematical scripting language.
- Reference Guide for OpenMatrix Language Functions
  The Reference Guide contains documentation for all functions supported in the OpenMatrix language.
- Block Reference Guide
  Describes all of the blocks in the installed Twin Activate library.
- Function Guide for Twin Activate
  Provides information about developing and simulating models through the Twin Activate Application Program Interface.
- romAI
  - Introduction
    romAI is an application that is comprised of libraries from the Altair Composeand Altair Twin Activate software products.
  - Design Fundamentals
    romAI serves two purposes.
  - Theoretical Framework
    Some system theory terminology must be understood.
  - ROM Framework
  - Workflow
    The typical workflow has four steps.
  - Installation
    The romAI tool is enabled through Extension Manager.
  - GUI
    The first steps of the workflow are accomplished through the GUI.
  - Pre-Processor Tab
    The Pre-Processor tab contains the optional pre-processing step.
  - Builder Tab
    The Builder tab facilitates the training step.
  - Post-Processor Tab
    The Post-Processor tab lets you visualize and evaluate the trained ROM.
  - ROM Generation Tips
  - Batch Training
    The training procedure can be carried over without using the GUI.
  - Batch Execution
    The simulation and evaluation of a romAI model can be also done via batch in the Compose OML editor without the need to deploy the trained romAI model into a Twin Activate block, using the romAI_execute function.
  - romAI Block
    After you install the romAI library in Twin Activate, a new palette appears in the Palette Browser.
  - Examples
    - Battery Cell Thermal Behavior
    - Nonlinear Spring Mass Damper System
    - Bucket into Pile of Bulk Material
    - Auto Exploration
    - Repeat Training
      During the training of a model in romAI, the Neurons are initialized randomly therefore two models can predict different values even though they have been trained with same Hyper Parameters.
    - Finite-State Machine Mass Spring Damper System
    - 2DoF Pendulum
      An example of a mechanical system of a pendulum in a cart.
  - Frequently Asked Questions
- Glossary
  Key terms associated with the software.
Frequently Asked Questions
You've got questions? We've got answers!

Auto Exploration

Consider the surface data shown below.

The files used in the example can be found at the following location:

<Twin Activate Installation> \hwx\databases\activate\extensions\romAI\_bin\romAIDirector\demos\romAI\auto_exploration

The auto_explore.oml file can be used to plot the surface data and write it out as .CSV file.

The task is to find the best combination of the number of hidden layers and number of neurons to fit the given surface data. The data is inside the surf_data.csv file and will be used to fit multiple static models in one training. Follow the steps below to run an auto exploration.

Load the surf.data.csv file in the Builder tab.
Select ‘x’ and ’y’ as input and ‘z’ as output.
Select the Auto Exploration option in the Settings tab and set the auto exploration attributes as shown below:
- Model Type: non linear
- (Hidden) Activation Function: tanh and relu
- Hidden Layers:
  - Min: 2
  - Max: 3
  - Incr: 1
- Neurons X (Hidden) Layer:
  - Min: 50
  - Max: 150
  - Incr: 50
- Epochs: 1000
- Test Split Ratio: 0.1
Start the training procedure.
After the training is complete, the Post-Processor opens and all the models generated from the training get stored in the auto_1 folder. The best model gets automatically selected in the post-processor. The best performing model is indicated by a value of 1 in the Test Loss Index column.
Select the Predict on Dataset option in the Hyper Planes section to see predictions and target data in the same plot.

We can see that the best performing model is surf_demo_relu_100_100_100 out of all the combinations given for exploration. In this way, you can do a grid search on Hyper parameters for best combinations.