In the following pages the data of OpenFOAM and Sparta best simulations respectively, as chosen in Chapter 6, are presented. The data chosen are the ones which could be interesting for a future study on the thermal shield, that is the Mach number, pressure, overall temperature and secondary temperatures. They are shown from Figure 6.4 to 6.12.
As it is possible to see there is a dierence in the secondary temperatures between the two software. In fact it is important to know that ve kinds of dierent temperatures exist and the analysis of these can be performed using dierent models. The meaning of the ve temperatures and why it is important to take into consideration dierent kinds of temperatures instead of only the overall one, will be treated in the following Paragraph 6.4. For the moment it is important to know that dierent kinds of temperature exist and they are dierent according to what physical model is used or, also, to what names are given to these temperatures.
The comparison between the two software can be done only with the same data plotted that is the Mach number, pressure and temperatures. Considering the unreal conditions of this experiment there are no referent experimental data. The comparison between the two software can be done
Figure 6.4: Mach Number Evaluated from OpenFOAM Best Simulation
Figure 6.6: Overall Temperature Evaluated from OpenFOAM Best Simulation
Figure 6.8: Translation Temperature Evaluated from OpenFOAM Best Simulation
Figure 6.10: Pressure Evaluated from Sparta Best Simulation
Figure 6.12: Vibration Temperature Evaluated from Sparta Best Simulation
done in the next Paragraph 6.3. Figure from 6.13 to 6.16 show the comparison between the two software and in Table 6.1 the same comparison is shown considering the absolute values. The values taken into consideration in Table 6.1 are the same as the sensitivity analysis plus the maximum pressure and maximum internal temperature (rotation temperature for Sparta). It is possible to see immediately that the results in their absolute values do not match, even if the shape of the curves are almost the same. Therefore it is possible to say there is no dierence in the physical meaning parameters evaluated, but probably the dierence is in the way of evaluation of each parameter. The way of evaluation does not mean the formula used, because that is reported in the documentation and it is the same for both software, but in the way of moving and colliding particles, applying boundary conditions,... so in how is structured the DSMC solver is structured. Actually, as can be seen, the overall temperature curve of the two software is quite dierent, so probably it is not the same parameter which is being evaluated. Anyway this subject will be examined more in depth later.
To establish what are the correct results, some accurate reference data are necessary, for com- parison, as is reported in the introduction of this paragraph. However, analyzing the graphs it is possible to make some important observations. For the moment the comparison between the two overall temperatures is not considered, instead, the pressure parameter will be discussed in parallel. From the rest of the data it is possible to see that the part of the OpenFOAM curves characterized by a negative gradient starts farther from the surface with respect to the Sparta one. This dierence is about 0.4 m. This means that the border of the shock layer is farther
Figure 6.13: OpenFOAM and Sparta Mach Number Comparison
Figure 6.15: OpenFOAM and Sparta Overall Temperature Comparison
Figure 6.16: OpenFOAM Internal Temperature and Sparta Vibration Temperature Compari- son
OpenFOAM Sparta Position(x) Ma=1 7.840e-2 3.550e-2
(m)
max Overall T 6.613e4 7.329e4 (K)
Shock Layer T 4.366e-2 2.000e-2 (m)
max p 2.439e3 1.484e3 (Pa)
Shock Layer p 5.012e-2 4.000e-2 (m)
max Int T/Rot T 6.561e4 4.587e4 (K)
Table 6.1: OpenFOAM and Sparta Relevant Data Comparison using Absolute Values
from the surface according to OpenFOAM data with respect to Sparta ones. The Mach number graph can be considered proof of this data where the position of Mach number equal to one is 0.07804 m for OpenFOAM and 0.0355 m for Sparta, so the dierence is about 0.4 m.
The part of the curve with a positive gradient for all the dierent temperatures has almost the same inclination and position. For the pressure, the analysis is equivalent except for the fact that the gradient is always negative and there is a dierence in the starting initial value. In fact, the initial gradient of the pressure slope is the same for both the curves, but it becomes higher than -1 in a position closer to the capsule surface in the Sparta case with respect to the OpenFOAM one, with a dierence of about 0,4 m. This means that the width of the shock is larger in OpenFOAM with respect to Sparta. Considering the point H3 is in the transition regime, at the limit of the continuous one, it seems that Sparta data are more accurate. However it is impossible to arm this because there are no reference data.
Another remarkable observation is that OpenFOAM absolute values of the pressure and internal temperature are higher then the Sparta ones. This observation is useful to analyze the Sparta overall temperature. In fact, from the comparison of the two overall temperatures it is clear that the shape of the two curves are dierent. From the OpenFOAM documentation, but also analyzing the shape of the OpenFOAM internal temperature and comparing it with some other temperature slopes deriving from literature, it is clear that what OpenFOAM refers to as internal temperature is a vibration / electron / electronic temperature in a Two-temperature model. The documentation about Sparta is not so clear about the temperature model used. However if an
Figure 6.17: OpenFOAM Translation Temperature and Sparta Overall Temperature Compar- ison
temperature (Figure 6.17) it is evident the slopes here are comparable. In addition the part of the OpenFOAM curve with a negative gradient starts farther from the surface with respect to the Sparta one by 0.4 m, like the dierence noticed in the rest of the graphs between OpenFOAM and Sparta simulations. Furthermore the characteristic of the pressure and internal temperature, That is that the OpenFOAM resulting data have higher absolute values with respect to Sparta ones, is evident here too. Instead if a comparison is made of both overall temperatures, besides having dierent shapes, the Sparta absolute value is higher than the OpenFOAM one, while the dierence in the position of the part of the slope with a negative gradient, noticed also in the other graphs analyzed, is respected.
Therefore probably the overall temperature in Sparta is the vibration / electron / electronic temperature in the Two-temperature model. Anyway this topic will be discussed in further detail and deepened in the next Paragraph 6.4.
Analyzing the graphs of each software it is possible to note the OpenFOAM plots have more data points with respect to the Sparta ones, so the curves are more accurate and smoother. In fact in Sparta it is possible to dump only a datum for each cell, but , as is presented in the sensitivity analysis, it is possible to increase the number of data points decreasing the cell size. Therefore in principle it is possible to have as many data points as required. The setting of the data points depends directly on the cell size, so if a certain number of data is required, probably it is impossible to obtain the most ecient parameter setting or the best results, as discussed in the sensitivity study. For this reason it is possible to say that OpenFOAM can oer a better
quality of results between the two DSMC software analyzed.
Besides, OpenFOAM has a very exible way of sampling data from the simulation which allows for as many data points as required and along whatever curve desired inside the domain without making any variations to the simulation parameters. In fact sampling is an operation done after the end of the simulation. For this reason it is possible to say that OpenFOAM has the most ecient and correct sampling data system too.
In the next Paragraph 6.3 an attempt will be made to verify which results are correct, Open- FOAM or Sparta.