Is filter media surface area size proportional....

[PandaBear]

Is filter media surface area size proportional to the amount of surface area oil pass through under load?

The reason I ask is that it seems like the oil will only pass through the "inner fold" edge of the media from the outside to the inside.

In that case, an oil filter will start to load up from the inner fold first, then gradually load up to the top. Think of them as a lot of Vs and the inner edge of the fold being the bottom tip of the Vs. The surface area of the media isn't important, but rather the number of folds and/or the inner tube's diameter is the limiting factor.

The outer fold/edge portion of the media under pressure will collapse and thus no oil will pass through it. The only portion the oil will pass through is the inner fold/edge, until they are full and the nearest surface will start passing oil.

Since your application is limited to a particular thread size, the size of the filter or its media surface area doesn't matter.

 

[Gary Allan]

Quote:
The reason I ask is that it seems like the oil will only pass through the "inner fold" edge of the media from the outside to the inside.


Well, let's think about this a bit. What stops oil from passing through every square inch of the media? The exterior is subject to the same PSID over its entire span. Now one might think that momentum is involved due to the general direction of travel when the oil does its cascading toward the interior ..and you may be correct ..but it's a relatively slow process if viewed in non-transitional states.

Unlike a fuel filter, which has such extremely low velocity that it's a still well, your oil filter keeps most of its particles in suspension. Seeing as you're taking two 90° turns (one before and one after the media) I reason that you're going to have some laminar shearing effect on flow. That is, the oil is going to manage the shortest pathway to the outlet. It's (initially) going to flow fastest through the top edge of the media instead of pushing through the longer path for a 100% uniform distribution across the entire media. Now 100% of the media may pass oil ..but the flow rate will decrease as you travel to the closed end of the can.

People still have the image of the oil filter being a high velocity environment. Now the individual streams of oil through each pore may be screaming, but if you stretch out 125-165^" of media and compare it to the .50^"+/- outlet, you can see that the stuff is typically going from a .5" pipe to a 3.0X3.0 (or more on both dimensions) intermediate pipe. Just ignore that the filter media is there for a second and you can see that the stuff is way slowed down. Some pores will get more flow initially ..and will clog sooner/earlier in the filter's life cycle. As they do, the flow will slow at those points and increase elsewhere.

 

[dnewton3]

Interesting concepts. I'm not sure we can "prove" any one particular point of view.

I'll agree with Gary that it's likely that the end of the filter closer to the mount sees a higher fluid flow rate. As you progress towards the opposite end, your velocity will drop off to some degree. It won't be zero, but it would likely be some amount less.

Here's where it gets tricky. Presuming that a slower flow rate provides better filtration (Gary's often touted this point), then the further away from the base mount you go, the better filtration you'd see.

Yet, filtration must also be a slave (at least in some portion, but perhaps not proportional?) to the amount of contamination presented per unit of time. In other words, if more oil is flowing past the base end of the filter media, does it allow more opportunities for contamination to be caught?

In simple terms, a slower rate will provide a "better" opportunity for catching particles; a faster rate will provide "more" opportunities for catching particles. In the end, does it equal out?

Let's apply some numbers here. These are ficticious numbers, but I chose them so that the math was easy. Presume you have an overall contamination of 50 particles per gallon of fluid; the engine's environment produces this given, and we have to presume some kind of a constant. I realize that it fluctuates in reality, but still it always produces the same rate of contamination that goes into the filter. Also, I know that they're not all the same size in reality, but just presume that regardless of size, the media will catch what it can, given it's physical limitations of pore size. Here's the model: if you can catch 60% of the particles at 5 gpm, is it the same as catching 30% of the particles at 10 gpm?

Here, we see 10 gal/min multiplied by 50 particles/gal gives us a "contamination rate" of 500 particles per minute. Multiply that times it's effective filtration percent (.30) and you get a capture rate of 150 particle per minute.

Now consider the 5 gal/min flow, multiplied by the same 50 particles/gal contamination, provides 250 particles / min. With a greater filtration rate (at .60), we still see the same effective number of particle caught in the realive time frame; 150 particles per minute.

So, I realize that the flow rates and effective filtration of the media are MUCH closer together. Perhaps there's only a 5% or less difference in flow rate from one end of the filter to another; who really knows? We could narrow the numerical gaps of my example, but the mathematical relationship would not change. For any given known contamination (xxx particles per volume of fluid presented), the total capture rate per unit of time would likely be very similar if not nearly identical.

The TOTAL flow rate through the filter is always equal (what goes in, must come out), but each linear inch of media, from base to top, sees some difference in flow relative to another. However, because different flow rates across the media in the one same filter likely catch nearly the same amount of the particulate, I would presume that the media would be evenly loaded.

Now, my example speaks to the linear loading from base to closed end. The question the OP posed was more of loading from one end of the "V" to the other. If you imagine the "Vees" as alternating "head to toe", you get a picture of reduced flow right at the narrow closed end of each "V", both inner and outer. So the flow would be better towards the middle of the pleat. Here again, my model of flow and capture ratio and unit of time all play into account.

It's been my experience, when taking filters apart and doing non-clinical analysis (eyeballing and using a magnifying glass) that the overall loading of filter media is reasonably even. You'll occasionally see clumps of goo from some sludge event that break away and gets snagged at the media, but overall, the media loading seems very consistent in texture, form, coloration, perceived density, etc. If you pull the media out lengthwise, you tend to see "lines" of darker concentration where the pleat ends are, but I've noticed that this is not true "loading" at the crisp fold, but rather the effect of squeezing out the oil at the fold. It's like wringing out a towel; as you compress the material at the fold of the media pleat, it produces a coloration change due to the fact that you're squeezing out a bit more oil at that fold by UNfolding it.

Gary has always professed that better filtration comes from slower flow, for any given media material, and he has given many fine examples. I believe that overall filtration is also subject to the "rate of opportunities". It's a question of percentage of success.

In the day to day application of vehicle filtration, it's likely a moot point. We're debating finite analysis here, and today's engines, oils, and filters are so good that it just doesn't matter much, if at all.

 

[Gary Allan]

Quote:
In the day to day application of vehicle filtration, it's likely a moot point. We're debating finite analysis here, and today's engines, oils, and filters are so good that it just doesn't matter much, if at all.


It does fall along the lines of whether you need a 10kt nuke or a 350mt nuke for a postage stamp size target ...but it is fun to ponder the finer points of such things.

 

[PandaBear]

Thanks for the insight. I understand that as long as the filter doesn't collapse, the filtration load on each part of the media should be fairly even.

What I am more interested in is whether the media "collapse" near the top and the flow stop, and only the inner folds, due to the proximity to the axis/center, do not collapse and remains the only available filter media surface.

You guys are right, since the color of the used filter medias are fairly even across the surface, the flow should be even across the media.

 

[dnewton3]

Well, let's be careful here.

I agree with Gary's insight that flow will vary in the filter, depending upon design and linear sectional view. To say that "the flow should be even across the media" is not true, and I think Gary and I agree on that.

My position is that the particulate loading of the media will be reasonably similar across the media, because the slower flow in some areas will filter with "better" efficiency. However, the faster flow in other areas will provide more opportunities, so even though the section "misses" more due to a faster flow, it also "catches" more because it sees more chances.

Gary's position is that (given media of the same type and construction, which is the only fair way to judge this debate) a slower flow results in better filtration. He's shown good examples of this.

I have no proof of my theory. I can model it easily mathematically. But to prove it, we'd have to be in the micro-lubricant fluid flow world; we'd have to shrink ourselves down (like in the movie "The Incredible Journey") and travel along the engine's life-blood lube stream.

Some several weeks back, Gary and I debated this to the Nth degree; each side with merit. Being an engineer, I have to be able to "model the therom" mathematically to make it work in my head. It comes down to the concept of whether or not you believe that the flow/filtration relationship is proportional or not. If proportional, then a larger filter would not filter "better", because everything would be linear; Gary has shown some evidence that this isn't the case. However, to be NON-proportional there are two alternatives. One would be uneven loading of filter media, because the slow flow at any near still-well would approach a much greater efficiency, thereby blinding off at a much quicker rate. Or, the other side of NON-proportional elemental analysis would be true, and the filteration would get worse with slower flow, which clearly isn't the case.

For me, I believe it to be near proportional, but I cannot show why a larger filter will, in reality, filter better when it's flow rate is slowed. Gary's theorm cannot be resolved mathematically, as it results in a huge filter going into bypass almost immediately, or never catching anything at all, at the infinite ends of possibility. Cleary, we're both right and both wrong. I would suspect that the truth is a combination of several elements. My therom of (near) proportional flow results in even loading, and does explain why we see filter media with reasonably diversified particulate distribution upon examination, even with clinical particle count analysis.

Gary is correct; it's fun to ponder the size nuke required for such infinitesimal tasks.

 

[Steve S]

Originally Posted By: Gary Allan
Quote:
In the day to day application of vehicle filtration, it's likely a moot point. We're debating finite analysis here, and today's engines, oils, and filters are so good that it just doesn't matter much, if at all.


It does fall along the lines of whether you need a 10kt nuke or a 350mt nuke for a postage stamp size target ...but it is fun to ponder the finer points of such things.
No need to ponder bigger is better.

 

[Gary Allan]

Quote:
have no proof of my theory.


Nor do I. It's more a "reasoned" thing.

I think that we've got a few constants and a few variables that are melded enough to confound a "one rule". Much/many of my belief(s) aren't in conflict with yours ..but can be somewhat static in their foundations. There are always exceptions and qualifications to any of our self constructed "rules".

As you infer, the "truth" may be more elusive on our level. We can even offer reasoned pseudo/para facts ..that when used in qualification, appear to support one notion or belief. You're also viewing a continuum in some "start:end" scenario where you've got an element that has some form of "half life". There goes those constants and variables thingie.

For example, if you merely had a flow manifold and stretched out your media across one rectangular plane, the flow would divide across the entire span of the media at the same flow rate. Assuming that the velocity wasn't so low that particle suspension would remain intact (that is, particles would not sink to the lower levels of the media), AND that pore distribution was fairly uniform, then the media should saturate at in an even manner. Each square inch would pass the same amount of fluid and trap the same amount of removable particles. Filter efficiency, assuming you're not talking in too large proportions, should be somewhat variable due to velocity.



But, I still think the sensible assumption is that flow will more conform to laminar instead of turbulent models ..at least in most reasonable flow rate levels.

I would think that, again under my laminar flow model, that smaller pores would rendered inert in the upper strata of the media due to higher opportunity for exposure to flow, while the lower strata would eliminate smaller particles due to lower velocity. That is, the actual loading rate may be uniform from a material weighted view, but flow alterations due to pore blockage, could surely vary.


..but ..media "collapse" is just about unheard of. The media is built to a given tolerance of PSID. The bypass valve setting is less than that tolerance. Otherwise, it would be a waste of a component. The bypass valve is to prevent oil starvation by limiting the transitional resistance that a filter can present to flow. The bypass valve setting is there to protect the media from being breached.

 

[Gary Allan]

Originally Posted By: Steve S
Originally Posted By: Gary Allan
Quote:
In the day to day application of vehicle filtration, it's likely a moot point. We're debating finite analysis here, and today's engines, oils, and filters are so good that it just doesn't matter much, if at all.


It does fall along the lines of whether you need a 10kt nuke or a 350mt nuke for a postage stamp size target ...but it is fun to ponder the finer points of such things.
No need to ponder bigger is better.


In most views of scale, yes.

Dave and I debated this in the post that he referenced. There's a diminishing rate of return in going bigger.

Let's say that there's a pore variance that spans from 10um to 50um on a given piece of media that is passing a constant 5gpm. Here's where we come into those "variables and constants" morphology that confounds the view. We don't know how to perceive the "standard" media relationship between size and flow design. That is, "just what is typical or standard"

That means that we're also poking in the dark about the effects of "bigger" in terms of efficiency improvements with reduced velocity. So, we may see a substantial improvement, using identical media, if we just use more square inches of the stuff ..BUT..let's triple or quadruple the media ..and assuming that all particles stay in suspension, you've just provided enough 50um holes to pass everything in the oil.

If you've ever seen some type of "size sorter" (eggs, nuts, anything) and mis-engineered it so that the items were randomly hitting the sorting gutters/holes, you might never get the smaller items into the smaller item bins.