This is a very common type of interface, used any time a fluid exchange heat with its container wall.

Covered in detail in the GunnsFluidHeatExchanger (heat convection) and GunnsFluidTank (heat conduction) link GunnsDraw help pages.

In this interface, the fluid aspect always owns the temperature of the fluid, and the thermal aspect always owns the temperature of the wall. The “wall” can be a tank shell, or a pipe, or the blades of a fan — basically any solid surface of the device that the fluid comes in contact with. The fluid aspect calculates the amount of heat transfer (flux or Q) between the fluid and the wall, based on the difference in temperature and some kind of heat transfer coefficient. The sign convention is that positive Q goes from the fluid to the wall, and negative Q goes from the wall to the fluid. The heat can go in either direction, always from hot to cold. This Q value is sent to the thermal aspect across the interface. That Q is absorbed in the thermal aspect of the device, and some of the Q can be spread out among the surrounding network. This changes the device temperature, which is sent back to the fluid aspect, and the cycle repeats.

Most fluid links are only designed to interface with one thermal aspect. This includes valves, tanks, pumps & fans. In this case there is one set of T & Q interface, and T is assumed to represent the average temperature of the entire solid device. Other links, such as the heat exchangers, can interface with more than one thermal aspect. The fluid flow path though the HX is divided into “segments”, each with its own thermal aspect. This can be used to get more accuracy in the integrated model.

This page has details on the exact terms to interface for various fluid link types.

We typically use this interface for these kinds of devices:

• Coldplates and fluid-to-fluid heat exchangers. This is essential for modeling any ATCS or ECLSS THC system.
• some pumps/fans, if their thermal aspect is needed. In space vehicles, pumps & fans are typically cooled by the fluid that they pump. In the real-world ISS, one time they left a ventilation fan running but blocked its flow path, so the air stagnated at the fan and couldn’t carry its waste heat away, and the fan overheated.
• some tanks. For most tanks we don’t care about the fluid & wall interaction. In some gas tanks, particularly those that undergo large & rapid pressurizations, the gas gets very hot or cold so it’s important to model that heat being equalized with the tank’s surroundings through the wall.
• some valves & orifices. This is rare — we usually don’t bother with thermal aspects for valves.

Here are some examples of what this interface looks like in action. First we show what it looks like when a hot wall is cooled down to the fluid’s inlet temperature over time. In this case there’s no other external heat fluxes on the thermal aspect so it eventually reaches steady-state at the fluid inlet temperature with zero Q:

Now we start adding 80 W of external Q to the thermal aspect. This starts heating it up again until it reaches steady-state with that 80 W of Q passing into the fluid and carried away:

This is a closed-loop interface with lag, so it is possible to misconfigure it so it becomes unstable. Instability will appear as chatter & noise in the Q and T terms. It is usually cause by the thermal aspect’s thermal capacity being too small relative to the fluid link’s heat transfer coefficient. The easiest way to remove the instability is to increase the thermal capacity in the thermal aspect.

Integration with a thermal aspect is optional. You can leave the fluid aspect not integrated, and configure it to exchange heat flux with a constant-temperature boundary condition. This initial wall temperature can be initialized from the GunnsDraw drawing. Also, you can disable all heat flux at any time by zeroing the link’s heat transfer coefficient.

Previous Page / Next Page