4.4     An example simulation

It is possible to demonstrate the use of the QW-HFM module using the same tasker file (*.ta3 file), which has been described in the previous section. No modification to this file is needed, only a modification in the project is needed to allow the heat flow. We check Allow Heat flow option in QW-Editor and invoke File-Export and Run ( button in Simulation tab). In QW-Simulator perform File-Open on heat_auto.ta3. Each Modify_Media_Parameters command (also Modify_Media_Parameters_Multiple and Modify_Media_Parameters_Multiple_T command) will automatically make a call to the hfm.exe application pointed in Set Tools.

Fig. 4.4-1 An example media definition file containing the electromagnetic and thermal properties of “food”.

In order to simulate the process of microwave heating inside of the example oven heat_auto.pro it is necessary to include thermal media properties in food.pmo. The file is been presented below. Although, a more accurate result could be obtained with thermal properties given as function of the temperature, one can expect that the presented approach is also reasonably accurate since the heating time is short (the time-step of the QW-BHM module is only 10s).

In order to compare various results of the analysis, the heat_auto.pro model has been run three times. The first time the QW-HFM module was not invoked which resulted in the final temperature field distribution shown already in Fig. 3.1-4. The distribution has been repeated in this section so that the reader can refer to them easier. The same model has also been simulated after coupling the QW‑HFM application to the QW-BHM module. Because two computational modes are available in the QW-HFM module the simulation has been done twice: using the external FDTD algorithm implementation and using the coupling to the Fluent package. The latter simulation has been performed at the Warsaw University of Technology, within the Eureka E!2602 project.

The initialisation files used for these simulations have been presented below – only one option has been changed.

Attention:

The QW-HFM module working in the Fluent mode utilises specialised routines that facilitate communication with the Fluent package. These routines – available in FluentDumpField2.c – should be saved into the folder containing the project which is to be simulated. A copy of the file is available in ..BHM\Heat\ folder in example directory and also on the installation DVD in \Utility\BHM\Fluent\ folder.

Fig. 4.4-2 The comparison of the temperature distribution in the lossy sample of medium “food”: a) the heat transfer not taken into account; b) the heat transfer effect modelled with the FDTD algorithm implemented in the hfm.exe application; c) the heat transfer effect modelled with the Fluent package.

The results of the three simulations have been compared in Fig. 4.4-2. As expected, taking into account the heat transfer effect slowed down the rise of the temperature inside of the sample. Also, the diffused temperature field is much smoother than the distribution calculated without using the QW-HFM module. In this case the results obtained with the QW-HFM application running in the FDTD mode are exactly the same as the field obtained with the external CFD package.

The simulation described above has been performed without taking into account the phase-change effect. In order to check how much the effect will affect the final temperature distribution the simulation has been repeated with the phase-change model active both in the FDTD mode and the Fluent mode. The corresponding initialisation files were modified only by adding the following command in the [General] section:

UseNonlinearModel = true

The comparison of the simulation results obtained with the FDTD mode and the Fluent mode has been shown in Fig. 4.4-3. It may come as a surprise that the temperature distributions hardy differ from the results obtained previously. This is due to the fact that the phase-change effect will most clearly show its influence in scenarios when the heating time (the BHM time-step length) is extremely long. In typical problems involving standard food products and time-steps equal to a few seconds the phase-change model will not change the final results to a large degree.

Fig. 4.4-3 The comparison of the temperature distribution in the lossy sample of medium “food”: a) the heat transfer not taken into account; b) the heat transfer effect modelled with the FDTD algorithm with phase-change model; c) the heat transfer effect modelled with the Fluent package and its phase-change model.

Fig. 4.4-4 The comparison of the temperature distribution in the lossy sample of medium “food”: a) the heat transfer not taken into account; b) the heat transfer effect modelled with the FDTD algorithm without phase-change model; c) the heat transfer effect modelled with the FDTD algorithm with phase-change model.

As already said, the influence of the role of the phase-change model can only be clearly shown for heating times long enough to let the sample reach the phase-change temperature. This is why the simulations described above have been repeated but, this time, the heating lasted cir. 8 minutes. One BHM iteration took 10s, exactly like in the previous example, and a complete simulation cycle lasted 49 iterations. The resulting temperature distributions within the sample have been presented in Fig. 4.4-4. Because results obtained with FDTD internal algorithm and Fluent solver are very close to each other, the figure shows only a comparison between temperature field obtained without employing the QW-HFM module, temperature field calculated with FDTD solver without the phase-change model, and the temperature field obtained with FDTD solver using the phase-change model.