Applications of xROM™

Thanks to ANSYS® dynaROM™, we prove, in three steps, that our xROM™ is able to learn physics from measured data.

Step 1 – dynaROM™ is able to learn the solution of a 3D/4D simulation solver:

Step 2: Even when the available data is limited to a few observations, dynaROM™ is still able to learn the output of the 3D/4D solver restricted to these observations:

dynaROM has been validated on more than 100 test cases.

Step 3  – This validates the application of xROM™ to measured data:

The model created by xROM™ from measured data is potentially the restriction of a 3D/4D complex model to the measured quantities.

xROM™ is even better than a 3D/4D complex model, since the latter has to be calibrated.

 


Step 1: Solid/fluid test case

We consider a coupled solid/fluid case. A flow at given velocity and temperature profile is injected in a cavity. A heated solid bar is fixed in the middle of this cavity. The flow temperature is changing with time. The geometry and mesh are given in the next figure.

The data of a simulation corresponding to the “tri” profile (in red in the figure below) was used by dynaROM™ to generate a dynamic reduced order model. Then the model can predict the flow evolution for three other temperature profiles: square, sqtri, sinus, defined in the following figure.

The results of these simulations are compared to the reference ones in the following videos. Only the temperature part of the field is presented.

  • “Sinus” profile

 

  • Sqtri profile

  • Square profile

 

Step 2: Piquage test case

The geometry of the test case is given below.

The dynamical reduced order model is applied to two configurations:

  • First excitation
Temperature excitation at inlet
Temperature at fluid outlet Temperature at solid outlet

 

  • Second excitation
Temperature excitation at inlet Temperature at fluid outlet Temperature at solid outlet

 

Step3: Mechanical test case
The response of mmechanical structural nonlinearities is close to the chaos. The goal of the test case is to build a neural network model from the measured displacement of a nonlinear structure. In the figures below we compare the best result obtained by the state of the art and the results obtained by xROM.

 

5) Thermal building
This work is carried out in collaboration with ACTIS – EMM (European Multifoil Manufacturer). It breaks with the state of the art in deriving thermal performances from in situ testing thanks to a breakthrough technology. This work has lead to the participation in the European CEN/TC89/WG13 standardization committee and in the international ISO/TC163/SC1/WG16-ISO9869-1 standardization committee.On this example, the difference between internal and external temperatures is simulated from the power consumption and some meteorological information.One can observe a good agreement between the measured data and simulated temperature difference.This is the basis of the Adagos patent (WO2014147148A2) for thermal building assessement.