Ali Marzban

Numerical investigation of turbulent flow and heat transfer in flat tube

In the present study, the fluid flow and heat transfer were numerically investigated in a flat tube under the constant heat flux using finite volume method and SIMPLEC algorithm. Also in this study, the range of Reynolds number is 5000–20,000, the range of dimensionless pitch (PN = PD/DL) is 1–2.33, and the range of dimensionless depth (DN = DD/DL) is 0.233–0.433. The use of second-order discretization for solving the governing equations on flow has made acceptable agreement between result between empirical and numerical results. The presence of dimples inside the channel, due to the creation of significant changes in flow physics and temperature field, considerably affects the flow and heat transfer parameters. The results indicate that by increasing Reynolds number, the convection heat transfer (Nusselt number) and the friction factor rise. Also, by decreasing the dimensionless pitch and increasing the dimensionless depth of dimple, Nusselt number and friction factor increase. The changes of Nusselt number are approximately related to the changes of dimensionless pitch and dimensionless depth of dimple, while the changes of friction factor are greatly related to the changes of dimensionless depth than the changes of dimensionless pitch. Based on the figures of average Nusselt number, using dimple in higher Reynolds numbers and constant DN ratio has a positive effect on Nusselt number increase. The main reason is the creation of stronger vortexes and better mixture of flow in higher Reynolds numbers. Hence, in average Nusselt number curves and in each constant DN ratio, the difference between graphs in Reynolds numbers of 2000 and 15,000 is higher than Reynolds numbers of 5000 and 10,000.

Semi-hierarchical analog simulation method for VHDL-AMS

Hierarchical simulation for analog systems, in general, is impossible. Conservative laws in analog systems generate conservative equations that should be satisfied along with the characteristic equations of the entire model. Conservative equations are not written explicitly by designer; but they are extracted from the physical structure of the model. In order to have a correct set of conservative equations, the entire analog model should be flatted and a single network matrix should be produced. In VHDL-AMS because of special port constrains, it is possible to partition the analog models and utilize a semi-hierarchical analog simulation method. In this paper I explain what this method of partitioning is and how it improves the overall simulation speed.