packageTester

Information

Tester

This section provides basic examples demonstrating how to use the SMArtInt library for integrating neural networks into Modelica simulations.

Included Examples

• PI Controller Example
  Demonstrates the use of Recurrent Neural Networks (RNNs), including:
  • A stateless RNN with 250 historical input values
  • A stateful RNN (sRNN) variant
• Pipe Heat Transfer Example
  Replaces a physical pipe model with a Feed-Forward Neural Network. A batched input is used to simulate the spatial discretization.

Both examples are provided for ONNX and TFLite versions of the neural networks and include a corresponding physics-based reference model for comparison.

Contents

Name	Description
ExamplePI
PipeHeatTransferExample