Tubular Reactor Surrogate Model Application

Application ID: 120381


This alternative version of the Tubular Reactor app demonstrates how computational speed can be significantly increased by using a surrogate model instead of a full finite element model.

A surrogate model is a simplified, computationally efficient approximation of a more complex and resource-intensive model. By enabling faster evaluations, the surrogate model enhances interactivity.

This app demonstrates the following:

  • Adjusting input parameters via sliders, with near-instantaneous updates to the solution retrieved from the surrogate model
  • Comparing the surrogate model solution with the full finite element solution
  • Efficient geometry sampling for data generation to be used for surrogate model training.

In this case, the surrogate model is a Deep Neural Network (DNN). The surrogate model has 5 input parameters in total: 3 for the activation energy, thermal conductivity, and heat of reaction, and 2 for the spatial coordinates.

This application does not require any add-on products.

This application example illustrates applications of this type that would nominally be built using the following products: