Originally Posted by LubricatusObsess
1. Worst case machine elements: valve train, cylinders (HTFS)
2. Major load bearing drivetrain: journal bearings (HTHS)
That's why I thought Gokhan's work was phenomenal. Yes, I know it's an estimate, but I'm unaware of any other work that provides these answers.
His work is very good (I refer to it myself at times) but for the purpose of this post it has to be stated that its "estimate value' is only valid when compared to a laboratory scale in a bulk quantity. That is no real or effective correlation to actual conditions inside a machine so there is no "answer".
That deserves an explanation as to why- here it is.
First- the reason this information isn't commonly available is because of the cost of the testing is prohibitive in many cases so companies don't want to pay for it because it's not a necessary thing to know for their product (or they suspect but don't want to know for a different set of reasons) and those that do keep it a secret. I just happen to have some because I do this on certain equipment for companies in specific applications. I can go general without violating anything.
You listed 3 examples so that's 3 different models so let's just stick to the journal bearing (which I personally do the most of)
Here's the process for me to answer that question for say "JB-1" ( which would have to be repeated for JB-2 is properties differ significantly)
Shear in the oil is directly proportional to the volume, force, temp inside the journal bore ( defined as where the bearing and shaft have contact) and all of that is relative to the FT of the design and subject to whether it is single pass or flood supported (splash or pump)
Then I have to segregate the fluid laminations because there will be a minimum of 3 different ones with their own properties. I will have boundary to bearing, boundary to shaft then the middle section.
Too thin a viscosity I start affecting shear more mechanically because of a fluctuating FT based on fluidic balance relative to asperities and geometry.
Too thick a viscosity then the fluid generates heat and that affects shearing the other way.
Then on an ICE- you have 2 shock pulses to the FT (TDC and BDC when full linear inertia is added)
This is just to start the process.
Now I need to outfit with probes, high speed imaging, accelerometers, acoustics etc. to monitor in real time, build a model with custom tests or head to ansys (which I will do anyway)
Then to validate I have to do a detailed metrology exercise.
Then in your car I have the 2 other regimes (beyond the scope of this post)
All of these forces are acting on an oil in a sump in a car that directly affect shear and rate.
So what we do for "field expedient testing" (less expensive but with a higher risk of error) is get the UT, FLIR, HS camera and accelerometers and make test runs of various viscosities with various additives ( with a detailed flush) to establish baselines.
Measure them with a data logger and compare with sampling (and we pull multiple working zone samples from minimess type valves to actually sample specific working groups to compare with a general sump sample)
Obviously every type of machine cannot have all of these done by design so you do the best you can.
We take all of that then analyze.
That's the difference between a 6 digit and a 7 digit test protocol. Only machines that cost in the millions of dollars or have that much in downtime get this type of analysis and reporting.
Some may come up with a single point test and call it a "miracle revelation" but that's right there with the 1 arm bandit.
If anyone wants a valid answer they have to go the distance- there is no way to accurately correlate a standard test with actual conditions without it
In summary take all of the articles and extrapolations for what they are because none of them directly correlate to what actually has to be known to accurately address the questions on actual shearing inside a given machine ( and other properties of oils as well)
