Tilting-pad thermo-hidro-dynamic bearing models #690
Conversation
verg1lio
changed the title
|
Hello and good afternoon! This is the pull request created for the tilting pad bearings model we are implementing. Since we have a working code on Matlab, it is a matter of translating and converting it to object-oriented programming style. This work has been started recently, and should take some time until it nears its completion, but if possible, we would like to request a sample of the results you obtain from your commercial code, so that we can standardize our code outputs. Changes in the output generation will undoubtedly incur in many structural modifications inside the code, so the sooner we have a defined standard to follow the better. We have a sample of this that has been provided for the cylindrical bearing model, and a similar one for the tilting-pad bearing should be enough to set the standard for this code too. Thanks in advance, and have a great evening! (@bela-bs , @marcusfilipesr , @hiagopinacio , @raphaeltimbo , @rodrigomoliveira1 ) |
|
Hello everyone. I am joining you in this work. Any questions about hydrodynamic bearing will be available to help. |
@elyqg is our expert on the bearing models, and both me (@verg1lio) and @marcusfilipesr will be assisting on the coding and debug. |
|
Hello, @verg1lio, good morning. It's a very nice initiative. Besides, I think @JuliaMota and @davidjgm can help you out, since they work mainly with the |
hiagopinacio
force-pushed
the
dev_tilting
branch
from
0ab3f65 to
965a980
Compare
|
Hello there! Just a quick update here... I have been translating the iterative loops to build the variable fields inside the bearings, and since they are all one inside the other and very much recursive in nature, they will be done all at once. So, on Trello (https://trello.com/c/zm7tJ29x/11-bearing-tilting-pad-thermohydrodynamic) my Tilting pad bearing card will be updated in a burst when the loops are done. After that the process should be more streamlined, but until then i'll focus here to get it right. Hope you guys are having a productive week! |
Co-authored-by: Marcus Filipe Sousa Reis <[email protected]>
Co-authored-by: Marcus Filipe Sousa Reis <[email protected]>
|
Hi there!
|
|
Dear all (@raphaeltimbo , @marcusfilipesr , @hiagopinacio , @elyqg , @bela-bs ), good morning! This THD tilting pad bearing model is accurately matching our results from other methodologies and previous benchmarks found in the specific literature. I believe it is ready for merging in the main ROSS code. Is there something else to make on it or we can proceed with the merging? Thanks, and have a great day! P.S. I am working locally on the THD thrust tilting pad bearing model, and should finish its rough translation soon. Once that is done it will be committed to its branch for debugging and corrections to its code. |
I have also provided the relevant information for a simulation on MaxBRG for comparison, but never got the results back. I therefore utilized literature benchmarks and experimental tests performed in our lab (LMEst) for this model validation. The code IS VALIDATED both experimentally and numerically, as of today, matching our benchmark results to its numerical precision. |
|
Dear @raphaeltimbo, I hope to find you in better health! We (@verg1lio and @elyqg) will take this week to add the K and C coefficients determination to this methodology, adapting the modular code made for the THD cylindrical bearings model. I will get back to you once its done, to remove the WIP from here and do the final conference and merge. Cheers! |
|
Hi @verg1lio ! Thank you for the message. |
verg1lio
changed the title
|
Dear @raphaeltimbo , good morning! I believe that this function is now finished and ready for merging. @elyqg and I did the inclusion of the K and C coefficients determination procedure to it, and it is working as intended. If there is any modification or correction still to be done, please let us know. I am currently doing the hard translation of the Thrust THD bearing model from Matlab to Python, and have about a third of it done as of now. In the next few weeks I will upload the translated code to our Thrust bearing branch and start the debugging. Have a great day! |
There was a problem hiding this comment.
Hi @verg1lio !
I have added some comments to the tilting_pad.py file.
Also, notice that the CI is failing because the code has not been formated with black.
Additionally, we will need to have a class that will inherit from BearingElement and implements the .K() and .C() methods so that the user can add a tilting pad bearing to their model.
See here how this was done to the fluidflow bearing:
https://github.com/ross-rotordynamics/ross/blob/09ace08c6aed999bdf9ded887690637f92dad0e5/ross/bearing_seal_element.py#L917
This part we could do in a separate PR in the future.
ross/fluid_flow/tilting_pad.py
Outdated
| from decimal import Decimal | ||
|
|
||
|
|
||
| class Tilting: |
There was a problem hiding this comment.
Please add documentation to the Class explaining how the user should provide each argument to instantiate an object.
Also, I think a name such as class TiltingPadBearing would be more descriptive.
There was a problem hiding this comment.
Alright, i changed the name and am adding the documentation atm.
There was a problem hiding this comment.
Done, the example details will be added in the next PR, but all the variables are described in detail and with relevant units indicated.
ross/fluid_flow/tilting_pad.py
Outdated
| ## Pressure border conditions ==================================================== | ||
| # for i in range(0, self.nZ - 1): # Loop in Z | ||
| # for j in range(0, self.ntheta - 1): # Loop in THETA | ||
| # if P[i, j] < 0: | ||
| # P[i, j] = 0 |
There was a problem hiding this comment.
Remove code that is commented out.
ross/fluid_flow/tilting_pad.py
Outdated
| matcoef_test = np.loadtxt( | ||
| "/home/verg1lio/Desktop/ross/ross/fluid_flow/matcoef.txt" | ||
| ) | ||
| b_test = np.loadtxt("/home/verg1lio/Desktop/ross/ross/fluid_flow/b.txt") |
There was a problem hiding this comment.
These paths won't work in other systems.
Do we need to have the matcoef and b files in the repository?
There was a problem hiding this comment.
We do not need them, these were validation files, and did their job just fine. I'm keeping the testing code there commented out, just in case we need to do the same task once merged into the main ROSS code.
ross/fluid_flow/tilting_pad.py
Outdated
|
|
||
| def run(self): | ||
| Fhx, Fhy = self._forces() | ||
| print(f"Fhx = {Fhx}\nFhy = {Fhy}\n") |
There was a problem hiding this comment.
I would recommend removing all the print from the code.
These can be very useful for debugging, but at the final code the user might not be interested in these print and it may clutter there shell/ide when running the code.
There was a problem hiding this comment.
Alright, I am commenting them in case debugging is needed in the next PR.
ross/fluid_flow/tilting_pad.py
Outdated
| psi_pad = np.array([x[0], x[1], x[2], x[3], x[4], x[5]]) | ||
| npad = 6 | ||
|
|
||
| # Radial clearance | ||
| Cr = 250e-6 | ||
|
|
||
| # Oil tank temperature | ||
| Tcuba = 40 | ||
|
|
||
| alpha = 0 # * (2 * np.pi * 5) * alpha | ||
|
|
||
| # Geometric parameters for the bearing -------------------------------------------- | ||
| # Journal radius | ||
| R = 0.5 * 930e-3 | ||
|
|
||
| # Pad radius | ||
| Rs = 0.5 * 934e-3 # [m] | ||
|
|
||
| # Pad thickness | ||
| esp = 67e-3 # [m] | ||
|
|
||
| # Pad arc | ||
| betha_s = 25 # [degree] | ||
|
|
||
| # Pivot position (arc pivot/arc pad) | ||
| rp_pad = 0.6 | ||
|
|
||
| # Bength of bearing | ||
| L = 197e-3 # [m] | ||
|
|
||
| # Angular position of the pivots [degrees] | ||
| sigma = np.array([0, 60, 120, 180, 240, 300]) | ||
|
|
||
| # Bearing loading | ||
| fR = 90.6e3 # [N] | ||
|
|
||
| # Rotor speed | ||
| wa = 300 # [rpm] | ||
|
|
||
| # Reference temperature | ||
| T_ref = Tcuba # [Celsius] | ||
|
|
||
| # Thermal properties for the oil ---------------------------------------------------- | ||
| # Thermal conductivity | ||
| kt = 0.07031 * np.exp(484.1 / (Tcuba + 273.15 + 474)) # [J/s.m.C] | ||
|
|
||
| # Specific heat | ||
| Cp = (16.5 * np.exp(-2442 / (Tcuba + 273.15 + 829.1))) * 1e3 # [J/kgC] | ||
|
|
||
| # Specific mass | ||
| rho = 0.04514 * np.exp(9103 / (Tcuba + 273.15 + 2766)) * 1e3 # [kg/m**2] | ||
|
|
||
| # Reference viscosity | ||
| # mi_ref=0.0752 | ||
| mi_ref = 5.506e-09 * np.exp(5012 / (Tcuba + 273.15 + 0.1248)) # [N.s/m**2] | ||
|
|
||
| # Bearing Position --------------------------------------------------------- | ||
|
|
||
| # Discretized Mesh ------------------------------------------------------ | ||
|
|
||
| # Number of volumes in theta direction | ||
| ntheta = 48 | ||
|
|
||
| # Number of volumes in x direction | ||
| nX = ntheta | ||
|
|
||
| # Number of volumes in z direction | ||
| nZ = 48 | ||
|
|
||
| # Number of volumes in neta direction | ||
| nN = 30 | ||
|
|
||
| mancal = Tilting( | ||
| R, | ||
| Rs, | ||
| esp, | ||
| betha_s, | ||
| rp_pad, | ||
| L, | ||
| sigma, | ||
| fR, | ||
| wa, | ||
| kt, | ||
| Cp, | ||
| rho, | ||
| mi_ref, | ||
| E, | ||
| phi, | ||
| psi_pad, | ||
| npad, | ||
| Cr, | ||
| alpha, | ||
| Tcuba, | ||
| ntheta, | ||
| nX, | ||
| nZ, | ||
| nN, | ||
| ) | ||
| mancal.run() | ||
| # mancal.coefficients() |
There was a problem hiding this comment.
Add a test file and move this part of the code to there.
There was a problem hiding this comment.
Alright, this is the next task in line. It should be done in a bit.
There was a problem hiding this comment.
Done! I will run a few tests with it and leave some default values for conference of code validity.
|
Hi @verg1lio ! |
Good morning @raphaeltimbo , and thanks for the input! I will get to work on this right away, and work out the details for it to be successfully merged to the main. Have a great week, |
|
Hi there!
|
|
I'll be closing this PR since it has not any updates for a while. |
Creation of the hydrodynamic tilting-pad bearing model for their properties determination inside ROSS.
This is a WIP, currently being ported to Python from a functional and validated Matlab code. The code is in Running through all relevant formulation. The current development target is the post-processing phase, translating the bearing fields into practical K and C coefficients.