Energy balance in compressor

jmenendez · February 20, 2021, 09:39

Hello everyone

I need to separate all energy types from the balance, but I don't get it with CFD-POST tools. I am using a model with:

Turbo-compressor (micro-scale), about 600W of power.
Steady case.
Real gas.
No conjugate heat transfer, only fluid domains (total energy). Energy equation with viscous work term activated)
Turbulent with SST k-Omega, Mach<0.5 (average).
The thermal boundary layer is modeled with $y^+\sim 1$
No gravity (volumetric).
Inlet conditions: T=20º, 1 to 4 bar variable, outlet: mass flow.

Convergence is very good: very small imbalances, RMS convergence below 1e-5, variables do not vary with interactions, etc. I use the classic formula: $\Delta h_{0,12}=q_{12}-w_{12}$ where is the total enthalpia,

heat and

work. 1 inlet, 2 outlet. In CFD-Post I use:

Wall Heat Transfer function for q
work w, torque_z (rotor surface without shroud) dot speed.
$\Delta h_{0,12}=h_{02}-h_{01}$ total enthalpy, with Total Enthalpy in Std Frame function.

All expressed in watts using mass flow.

and the result is neither good nor consistent. I am not getting a correct value (approximately zero), so how can I separate all the energy contributions and thus check where it comes from (work, heat, dissipation, etc.)?

Thank you very much for your help.

Opaque · February 20, 2021, 11:32

Let us go back to basics, and list exactly which terms you think you need to complete the imbalance calculation verification:

1 - Inlet: advection of XXX, ...?
2 - Outlet: advection of XXX, .. ?
3 - Walls: conduction heat flux, ... ?
4 - Volumetric: ?

Once you have that list, we can help with how to compute each one of them.

Keep in mind which heat transfer model you are using to determine exactly which fluxes/sources are needed.

jmenendez · February 20, 2021, 12:18

Quote:

Originally Posted by Opaque

Let us go back to basics, and list exactly which terms you think you need to complete the imbalance calculation verification:

1 - Inlet: advection of XXX, ...?
2 - Outlet: advection of XXX, .. ?
3 - Walls: conduction heat flux, ... ?
4 - Volumetric: ?

Once you have that list, we can help with how to compute each one of them.

Keep in mind which heat transfer model you are using to determine exactly which fluxes/sources are needed.

Thank you so much for your help. I have edited the first message adding information.

Opaque · February 22, 2021, 08:43

great!! Thank you for editing the initial post.

Could you please post the imbalance summary at the end of the output file? i.e. H-Energy balance summary.

The terms you have computed in CFD-Post should match some of the values reported in those rows. Let us see which ones are correct, and which ones we need to figure out.

jmenendez · February 22, 2021, 17:16

Quote:

Originally Posted by Opaque

great!! Thank you for editing the initial post.

Could you please post the imbalance summary at the end of the output file? i.e. H-Energy balance summary.

The terms you have computed in CFD-Post should match some of the values reported in those rows. Let us see which ones are correct, and which ones we need to figure out.

Hello

Here is the information of the imbalaces in the three domains: inlet, impeller (R1) and diffuser (S1):

Code:

 +--------------------------------------------------------------------+
 |                           H-Energy-Inlet                           |
 +--------------------------------------------------------------------+
 Boundary         : Inlet Hub                               1.4171E-01
 Boundary         : Inlet Inlet                             2.1989E+02
 Bnd Src/Visc Work: Inlet Inlet                            -1.8412E-05
 Boundary         : Inlet Shroud                            2.6594E-01
 Bnd Src/Visc Work: Inlet to S1 Side 1                      7.8412E-06
 Domain Src (Neg) : Inlet                                  -1.8258E-01
 Domain Src (Pos) : Inlet                                   1.8258E-01
 Domain Interface : Inlet to R1 (Side 1)                   -2.2029E+02
                                                           -----------
 Domain Imbalance :                                        -4.5853E-04

 +--------------------------------------------------------------------+
 |                            H-Energy-R1                             |
 +--------------------------------------------------------------------+
 Bnd Src/Visc Work: Inlet to S1 Side 2 1                    3.5795E-03
 Boundary         : R1 Blade                                1.1637E+00
 Boundary         : R1 Hub                                  5.7607E-01
 Boundary         : R1 Shroud                               1.1157E-01
 Bnd Src/Visc Work: R1 Shroud                               8.3487E-01
 Bnd Src/Visc Work: R1 to R1 Internal 2 Side 1              1.3052E-04
 Bnd Src/Visc Work: R1 to R1 Internal 2 Side 2              1.2352E-03
 Bnd Src/Visc Work: R1 to R1 Internal Side 1                1.0359E-04
 Bnd Src/Visc Work: R1 to R1 Internal Side 2                1.3081E-03
 Bnd Src/Visc Work: S1 to R1 Side 1 1                       1.7367E-01
 Domain Src (Neg) : R1                                     -2.4993E+01
 Domain Src (Pos) : R1                                      2.4993E+01
 Domain Interface : Inlet to R1 (Side 2)                    2.2030E+02
 Domain Interface : R1 to R1 Internal 2 (Side 1)           -3.0471E+01
 Domain Interface : R1 to R1 Internal 2 (Side 2)            3.0471E+01
 Domain Interface : R1 to R1 Internal (Side 1)             -4.3890E+01
 Domain Interface : R1 to R1 Internal (Side 2)              4.3890E+01
 Domain Interface : R1 to S1 (Side 1)                      -2.2317E+02
                                                           -----------
 Domain Imbalance :                                        -2.7862E-03

 +--------------------------------------------------------------------+
 |                            H-Energy-S1                             |
 +--------------------------------------------------------------------+
 Boundary         : S1 Hub                                 -5.9928E-01
 Boundary         : S1 Outlet                              -2.5993E+02
 Bnd Src/Visc Work: S1 Outlet                              -1.4654E-04
 Boundary         : S1 Shroud                              -7.2501E-01
 Bnd Src/Visc Work: S1 to R1 Side 1                         6.6675E-02
 Domain Src (Neg) : S1                                     -3.5929E+01
 Domain Src (Pos) : S1                                      3.5929E+01
 Domain Interface : R1 to S1 (Side 2)                       2.6110E+02
                                                           -----------
 Domain Imbalance :                                        -8.4797E-02

If I do an energy balance $\Delta H= Q-W$

Total heat through the walls, I use Wall Heat Flux function
Work (energy introduced by the external motor, I do not know this data), I use torque_z
Total enthalpy outlet-inlet, I use Total Enthalpy in Std Frame between outlet-inlet

The geometric domain is a passage: an inlet, an outlet, the side surfaces are periodic, and the hub, shroud, blade, are with constant temperature contour condition.

In CFD-Post:

1. Heat: -7.475W, with areaInt(Wall Heat Flux)@all wall surfaces.

2. Work: 39.35W, with torque_z()@Rotor omega .

3. Total enthalpy: 39.89 W (edited, is -32.2W, not -39.89W), with massFlowInt(Total Enthalpy)@Outlet +massFlowInt(Total Enthalpy)@Inlet (outlet and inlet of machine, not diffuser).

Opaque · February 22, 2021, 18:19

If you take the contributions from the summary, and add them, you should get the correct terms.

Flow @S1 Outlet - Flow @Inlet Inlet = 40.04 ~ 39.89

Now you have to add up all the walls or split the summation into components, say for areaInt(Heat Flux)@Inlet Walls + areaInt(Heat Flux)@R1 Walls + areaInt(Heat Flux)@S1 Walls will allow understanding which row is not producing the correct value.

NOTE: You should use Heat Flux and not Wall Heat Flux for the best matching. It is documented somewhere.

and then, add the values for Boundary Src/Visc Work similarly and see what values you obtain.

There should be nothing else to balance. The term you call "Work" computed from torque()... is a workaround way to balance the equations. The energy equation does not have any "torque term", it has viscous work.

Add them up, and let us see what it leads to.

jmenendez · February 22, 2021, 19:32

Quote:

Originally Posted by Opaque

If you take the contributions from the summary, and add them, you should get the correct terms.

Flow @S1 Outlet - Flow @Inlet Inlet = 40.04 ~ 39.89

Now you have to add up all the walls or split the summation into components, say for areaInt(Heat Flux)@Inlet Walls + areaInt(Heat Flux)@R1 Walls + areaInt(Heat Flux)@S1 Walls will allow understanding which row is not producing the correct value.

NOTE: You should use Heat Flux and not Wall Heat Flux for the best matching. It is documented somewhere.

and then, add the values for Boundary Src/Visc Work similarly and see what values you obtain.

There should be nothing else to balance. The term you call "Work" computed from torque()... is a workaround way to balance the equations. The energy equation does not have any "torque term", it has viscous work.

Add them up, and let us see what it leads to.

Sorry, there was an error in the previous post, I just edited it. when I do in CFD-POS massFlowInt(Total Enthalpy)@(Inlet+S1 Outlet) I get -32.2W, not -39W, sorry. And, if I sum wall heat: Houtlet + Hinlet +Qwalls, the result is 39.89W.

I'm trying to follow the method you've suggested by comparing the energy of the solver with the one I get from CFD-Post. But there's one thing I don't quite understand:

1. The energy that appears in the solver output, is total H integrated into the surface, is it not?,

2. but I'm interested in separating the kind of energy, how can I do this? I'm trying to balance energy, and I use: Total enthalpy outlet + Total enthalpy inlet = total heat + work, $\Delta H= Q-W$

Total enthalpy I get from massFlowInt(Total Enthalpy)
Total heat from areaInt(Heat Flux), as you suggest. So how could I get the total (net) of the heat (all the heat involved, viscous dissipation, by surfaces, etc.). If I integrate areaInt(Heat Flux) into the entire control volume, would that be enough?
how could I post it to CFD-Post, what function is there for this? I just thought I'd use the torque.

When I do massFlowInt(Total Enthalpy)@(INLET+OUTLET) I get -252.424W, I dont understand this value.

Thank you.

Opaque · February 23, 2021, 08:46

The report in the output file lists the different terms discretized in the equations. For the Total Energy equation (steady state)

div (advection +/- diffusion +/- viscous work) - Sources = 0

I use +/- just to cover for sign conventions of what the "word" implies.

From the above, we need to only balance the surface terms, i.e. div( ), with the sources in the volume. Since there is no heat generation in your setup, the surface terms MUST balance to 0.

The report you posted is not fully converged (notice the domain imbalances increasing from inlet to outlet); however, good enough to understand them.

Keep in mind the flows already have the ANSYS CFX convention sign in them; therefore, careful when adding/substracting them. ANSYS CFX uses the convention
+ -> Energy in
- -> Energy out

Your convention for

Delta H = Heat - Work

uses the

Heat in --> +
Work in --> -

The equivalent equation for ANSYS CFX would be

Delta H = Heat + Work

February 20, 2021, 09:39	Energy balance in compressor	#1
jmenendez Member Join Date: Feb 2019 Posts: 37 Rep Power: 7	Hello everyone I need to separate all energy types from the balance, but I don't get it with CFD-POST tools. I am using a model with: Turbo-compressor (micro-scale), about 600W of power. Steady case. Real gas. No conjugate heat transfer, only fluid domains (total energy). Energy equation with viscous work term activated) Turbulent with SST k-Omega, Mach<0.5 (average). The thermal boundary layer is modeled with $y^+\sim 1$ No gravity (volumetric). Inlet conditions: T=20º, 1 to 4 bar variable, outlet: mass flow. Convergence is very good: very small imbalances, RMS convergence below 1e-5, variables do not vary with interactions, etc. I use the classic formula: $\Delta h_{0,12}=q_{12}-w_{12}$ where is the total enthalpia, heat and work. 1 inlet, 2 outlet. In CFD-Post I use: Wall Heat Transfer function for q work w, torque_z (rotor surface without shroud) dot speed. $\Delta h_{0,12}=h_{02}-h_{01}$ total enthalpy, with Total Enthalpy in Std Frame function. All expressed in watts using mass flow. and the result is neither good nor consistent. I am not getting a correct value (approximately zero), so how can I separate all the energy contributions and thus check where it comes from (work, heat, dissipation, etc.)? Thank you very much for your help. Last edited by jmenendez; February 20, 2021 at 14:19.

February 20, 2021, 11:32		#2
Opaque Senior Member Join Date: Jun 2009 Posts: 1,804 Rep Power: 32	Let us go back to basics, and list exactly which terms you think you need to complete the imbalance calculation verification: 1 - Inlet: advection of XXX, ...? 2 - Outlet: advection of XXX, .. ? 3 - Walls: conduction heat flux, ... ? 4 - Volumetric: ? Once you have that list, we can help with how to compute each one of them. Keep in mind which heat transfer model you are using to determine exactly which fluxes/sources are needed. aero_head likes this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

February 22, 2021, 08:43		#4
Opaque Senior Member Join Date: Jun 2009 Posts: 1,804 Rep Power: 32	great!! Thank you for editing the initial post. Could you please post the imbalance summary at the end of the output file? i.e. H-Energy balance summary. The terms you have computed in CFD-Post should match some of the values reported in those rows. Let us see which ones are correct, and which ones we need to figure out. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

February 22, 2021, 18:19		#6
Opaque Senior Member Join Date: Jun 2009 Posts: 1,804 Rep Power: 32	If you take the contributions from the summary, and add them, you should get the correct terms. Flow @S1 Outlet - Flow @Inlet Inlet = 40.04 ~ 39.89 Now you have to add up all the walls or split the summation into components, say for areaInt(Heat Flux)@Inlet Walls + areaInt(Heat Flux)@R1 Walls + areaInt(Heat Flux)@S1 Walls will allow understanding which row is not producing the correct value. NOTE: You should use Heat Flux and not Wall Heat Flux for the best matching. It is documented somewhere. and then, add the values for Boundary Src/Visc Work similarly and see what values you obtain. There should be nothing else to balance. The term you call "Work" computed from torque()... is a workaround way to balance the equations. The energy equation does not have any "torque term", it has viscous work. Add them up, and let us see what it leads to. jmenendez and aero_head like this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

February 23, 2021, 08:46		#8
Opaque Senior Member Join Date: Jun 2009 Posts: 1,804 Rep Power: 32	The report in the output file lists the different terms discretized in the equations. For the Total Energy equation (steady state) div (advection +/- diffusion +/- viscous work) - Sources = 0 I use +/- just to cover for sign conventions of what the "word" implies. From the above, we need to only balance the surface terms, i.e. div( ), with the sources in the volume. Since there is no heat generation in your setup, the surface terms MUST balance to 0. The report you posted is not fully converged (notice the domain imbalances increasing from inlet to outlet); however, good enough to understand them. Keep in mind the flows already have the ANSYS CFX convention sign in them; therefore, careful when adding/substracting them. ANSYS CFX uses the convention + -> Energy in - -> Energy out Your convention for Delta H = Heat - Work uses the Heat in --> + Work in --> - The equivalent equation for ANSYS CFX would be Delta H = Heat + Work aero_head likes this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
2D air heater with chtMultiRegionFoam: Converged solution violates energy balance?	Reineke	OpenFOAM Running, Solving & CFD	0	February 3, 2020 09:57
Energy balance in conjugative/ porous simulation; Domain Energy is greater than Input	Vishnu_bharathi	CFX	2	June 27, 2018 10:08
Natural convection, problems with total energy balance	MdoNascimento	STAR-CCM+	0	March 13, 2014 12:15
volumetric reaction and energy balance?	m.beh	FLUENT	1	October 17, 2011 17:21
ATTENTION! Reliability problems in CFX 5.7	Joseph	CFX	14	April 20, 2010 15:45