Learning Forum

[mkrsmanovic]

I'm modeling a large thermal energy storage based on the actual shop drawings of the tank. Unfortunately I cannot share the actual propriatery model, but the approach I'm taking right now is:

- Solver: Pressure-Based, Absolute Velocity Formulation, Transient Time, Gravity on.

- Models: Energy on, Viscosity is Realizable k-ε, Standard Wall Functions, Full Buoyancy Effects.

- Fluid: water (single phase), Boussinesq density.

- Boundary conditions: mass-flow-inlet for inlet; and I tried outflow, mass-flow-outlet and pressure-outlet for outlet.

- Operating conditions: Gravity on, Variable-Density Parameters also on.

- Methods: SIMPLE scheme, Second Order Upwind for Energy, but I left First Order Upwind for k and ε.

Anything I didn't mention is left at the default values.

 

The inlet - hot temperature charging - is at high level, and the outlet - low temperature discharging - is at low level. I set the density, temperature and thermal expansion coefficient under Fluid and Operating Conditions to match the low temperature state. I set the time step so that my Courant number is below 1, and I set the number of steps to capture one full volume change based on the actual charge/discharge mass flow rate.

 

The thing with my simulations is, I'm seeing significant temperature diffusion from top, hot layer, and towards the very bottom of the tank. I can observe stratification visually and thermocline appears to be fairly tight just by looking at the contours. But the temperature at the bottom of the tank, near the outlet, rises sharply early on and reaches 1/2 or 2/3 of the inlet-outlet temperature difference range. The whole time thermocline appears to be located in the upper 1/4 of the tank height, and doesn't drop towards the low level.

 

I don't expect this to be correct based on the temperature sensors reading from similar tanks, nor did I expect to see it after reading different papers on this type of simulations. I would expect the temperature at the outlet to not change at all during the most of the duration of the sim and raise only towards the end of the sim.

 

I initially set the outlet to outflow, but after reading the notes on fully developed flow I tried other outlet boundary conditions. Still, the results are fairly similar, which makes me think there is something else to fix in my model.  Any ideas on what I can modify in my approach?

[Rob]

Typically academic work would be published under the terms of the licence, so please ensure you're within any terms agreed with the University. 

Given the water density isn't changing, try using a velocity inlet as it's generally more stable. Pressure outlet is correct, and 0 Pa is fine as a value as you're running single phase: check the outlet backflow temperature.  If you're seeing excessive diffusivity, check the fluid you have in the cell zone(s) as you won't be the first (nor the last) to accidentally set up water correctly and then run the model with air....  What temperature did you initialise the tank to?