Thermal energy storage simulation with turbulence and bouyancy

HighPlains · July 15, 2024, 15:35

I'm modeling a large thermal energy storage - water only, no phase change - based on the actual shop drawings of the tank. I'm comparing my simulation results with the actual data from flow meters and temperature sensors.

Tank has a hot water inlet at the high level, and discharges low temperature water at low level. The temperature data shows historically stable cold water column with a constant temperature profile up to 80-85% of water column height. The top 5-6% is the warm zone where water temperature is fairly close to the inlet water temperature. The remaining thermocline zone is well defined.

However, in all my simulations I'm getting results showing significant temperature diffusion from the top. The thermocline is "smeared" throughout the height and the difference between the maximum and minimum water temperature is significantly lower than in reality. The maximum temperature in the tank is significantly lower than what's the actual case. The low temperature zone at the bottom warms up and the minimum temperature increases from its initial value.

The details of my model are below, together with what I tried changing in it. What would you try differently to make this accurate?

- Solver: Pressure-Based, Absolute Velocity Formulation, Transient Time, Gravity on (0,0,-g).
- Models: Energy on, Viscosity is Realizable k-ε, Standard Wall Functions, Full Buoyancy Effects.
- Fluid: water. I tried: 1) Boussinesq density and 2) density as expression.
- Boundary conditions: I tried 3) mass-flow-inlet and 4) velocity inlet; and I tried 5) outflow, 6) mass-flow-outlet and 7) pressure-outlet for outlet.
- Operating conditions: Gravity on, Variable-Density Parameters off.
- Methods: SIMPLE scheme, Second Order Upwind for Momentum, TEK, TDR, and Energy.
- Time Advancement 8) Fixed and 9) Adaptive.
Anything I didn't mention is left at the default values.

My hand calcs and Adaptive method both gave me time-step size on the order of magnitude of 1e-4s. I did a couple of those simulations but since they take forever I steadily increased the time-step until I reached 0.5s. No major change in results.

The inlet and outlet are curved surfaces, so I'm not sure if I should change anything in momentum direction settings for inlet and outlet.

July 15, 2024, 15:35	Thermal energy storage simulation with turbulence and bouyancy	#1
HighPlains New Member Join Date: Jul 2024 Posts: 1 Rep Power: 0	I'm modeling a large thermal energy storage - water only, no phase change - based on the actual shop drawings of the tank. I'm comparing my simulation results with the actual data from flow meters and temperature sensors. Tank has a hot water inlet at the high level, and discharges low temperature water at low level. The temperature data shows historically stable cold water column with a constant temperature profile up to 80-85% of water column height. The top 5-6% is the warm zone where water temperature is fairly close to the inlet water temperature. The remaining thermocline zone is well defined. However, in all my simulations I'm getting results showing significant temperature diffusion from the top. The thermocline is "smeared" throughout the height and the difference between the maximum and minimum water temperature is significantly lower than in reality. The maximum temperature in the tank is significantly lower than what's the actual case. The low temperature zone at the bottom warms up and the minimum temperature increases from its initial value. The details of my model are below, together with what I tried changing in it. What would you try differently to make this accurate? - Solver: Pressure-Based, Absolute Velocity Formulation, Transient Time, Gravity on (0,0,-g). - Models: Energy on, Viscosity is Realizable k-ε, Standard Wall Functions, Full Buoyancy Effects. - Fluid: water. I tried: 1) Boussinesq density and 2) density as expression. - Boundary conditions: I tried 3) mass-flow-inlet and 4) velocity inlet; and I tried 5) outflow, 6) mass-flow-outlet and 7) pressure-outlet for outlet. - Operating conditions: Gravity on, Variable-Density Parameters off. - Methods: SIMPLE scheme, Second Order Upwind for Momentum, TEK, TDR, and Energy. - Time Advancement 8) Fixed and 9) Adaptive. Anything I didn't mention is left at the default values. My hand calcs and Adaptive method both gave me time-step size on the order of magnitude of 1e-4s. I did a couple of those simulations but since they take forever I steadily increased the time-step until I reached 0.5s. No major change in results. The inlet and outlet are curved surfaces, so I'm not sure if I should change anything in momentum direction settings for inlet and outlet.