4 shows once again the comparison between the two approaches. Furthermore, the temperature difference between the membranes and the bottom surface of the micro-cavity is 10 K less than the one existing between the membranes and the top surface.įor what concerns the heat flux of the side walls of the microbolometers, Fig. Since the gap between the bottom surfaces of the microbolometers and the bottom surface of the micro-cavity is quite narrow, the number of incident particles used for the computation of the heat fluxes are probably not enough if to consider just 5000 time steps. A possible explanation for this comes from the statistical nature of the DSMC approach. A good agreement between the solutions coming from the two different methods is observed both for the hotter and for the colder microbolometers, that is partially lost if to consider the heat fluxes obtained for the bottom surfaces for low values of Δ T: the committed relative error for the heat flux evaluated at the bottom surface of the hot membrane at Δ T = 15 K is 5.5%. The considerable difference in temperature between the membranes and the bottom surface of the micro-cavity T W, B and between the two membranes Δ T is aimed at reducing the characteristic statistical noise of the DSMC approach. Here, the parameter Δ T have been swept from 5 K up to 35 K, considering a fixed temperature for the second microbolometer of 343.15 K. The necessity of having a 2D model for evaluating the heat flux at the lateral sides of the microbolometers derives from their subtlety: the thermal field spreading in the thin gap between two adjacent membranes is strongly affected by the ones spreading towards the top and the bottom of the unit cell, leading to spatially varying heat fluxes.Ī first comparison between the heat fluxes obtained from the DSMC and the approximate diffusive approaches is shown in Fig. The DSMC study has been carried out by considering 250000 simulated particles, tracked for 5000 time steps (observed to be enough for obtaining a steady state solution).įor what concerns the approximate diffusive analysis, conducted by using (7) for the thermal conductivity of argon, two 1D models for the characteristic dimensions L top and L bottom and one 2D model for the characteristic dimension L side have been implemented, and a β = 1.5 has been used. Depending on the value of the Knudsen parameter, three different approaches may be used to characterize the phenomenon under investigation: for \(\). Where λ is the mean free path of the molecule and L the characteristic dimension of the system.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |