Question on engine oil passages... 3

[MonicaS]

Is anyone familiar with modeling the effects of changes in oil passage geometry and its effect of pump/filter sizing and adequate delivery to critical surfaces (bearings, cylinder linings, etc.)

 

[ivymike]

yes, I am.

 

[dicer]

Not sure how much the geometry would have on the sizing of pump, there are other variables that are more important to consider.
Besides I know of many engines that have no oil pump.

 

[tbuelna]

MonicaS,

Engine journal bearing oil flows are based primarily on cooling requirements. Assuming all other variables (oil inlet temps, bearing clearances, etc) remain unchanged, then the pressure drop occurring in the lube circuit between the pressure relief valve and the crank main bearing gallery, due to geometry changes, can be of some concern. The cooling oil mass flow at each bearing is partially a function of the deltaP and effective orifice area created at the bearing/journal clearance.

Accurately modeling the entire lube oil circuit, with its flow coefficients and pressure drops for each component and flow passage is not too difficult. It is commonly done with any serious engine design effort, and there are even software applications to perform this analysis.

What you will likely find with such an analysis, is that once you take all of the worst case operating and tolerance conditions into account, your calculated flow requirements needed from a fixed displacement oil pump will seem excessively high. But that's just the nature of the beast with engine lube systems.

As a rule of thumb, engine oil passages should be designed for a maximum flow velocity of 20 ft/sec on the pressure side, and a maximum of 6 ft/sec at the pump inlet to prevent cavitation.

Good luck.
Terry

 

[MonicaS]

Thanks a bunch for your input Terry! I appreciate it.

I had been using CFD to model the individual components, but it would have been way too cumbersome to do an entire oiling system model (complete with passages) using it.

This morning, my company decided to purchase some "system level" CFD modeling software that I can integrate my CFD info into. Hopefully, this will make it a bit easier...

But your points on flow velocities and calculated flow requirements were very valuable and I will keep them in mind as I work on this.

Thanks again!

 

[ivymike]

hmm... what do you mean by "system level" CFD? I'm used to this sort of analysis being done in a 1-D hydraulic circuit simulation, 80% of the time without consideration of heat transfer. bearings are typ. modeled separately and put into the hyd. model as simple restrictions. Lines and geometry changes are considered via classical methods and added to the model as simple restrictions.

"CFD" (not sure what "system level" means) brings to mind 2D or 3D flow distributions, and seems like using a sledgehammer to attack an ant.

 

[tbuelna]

MonicaS,

ivymike is correct. A CFD model is probably way overkill for a lube circuit analysis. The circuit is usually modeled schematically with each passage section, corner, orifice, component, etc. modeled as a simple restriction. These lube circuit diagrams usually resemble electrical schematics, with each flow restriction represented by a "resistor".

Here's a commercial application that might be more suitable:

http://www.lmsgermany.com/lubrication

Good luck.
Terry

 

[MonicaS]

Thanks Terry and Mike,

When I said "system level" I meant a 1D/2D type of model - sorry for the confusion...

But I just watched a webinar earlier this morning on a commercial code that might work for my application. The webinar is over now, but they have a lubrication case study on their website...

http://www.flowmaster.com/pdf/Flowmaster_MAHLE_Powertrain_Case_study.pdf

I'm going to be asking a lot of of questions to these folks over our evaluation period...

Thanks for your notes though!

 

[dicer]

Actually what would be more useful than a computer would be actual time turning the wrenches on various engine types, infield as well as prototypes.
Just an opinion from someone older than 55. I see too much regression in engineering, what I mean is relearning things that were solved in the past, and now forgotten at least it appears that way. Too much diddling on a computer without the proper understanding of the machine being diddled with. I've read countless articles in professional periodicals about this very thing.