Does dimpling intake ports work? (No)

[James363]

Hi all,
I have just finished writing my final year dissertation after taking a break from work to finish my university degree. My employer suggested that I should investigate the effects of adding dimples to the internal surface of an intake port in order to determine if it is beneficial to intake port performance. The port tested was a Jaguar Lightweight E-Type intake port from a wide angle Jaguar head, which had been previously optimised using CFD to improve flow through the port. After nearly 300GB of simulations which have been running for months, I have not been able to improve the flow through the port using single dimples on their own, or covering the whole port in dimples, or even dimpling just the valve guide or valve stem itself. Putting dimples into an intake port does have some beneficial characteristics, but these are outweighed by the increase in skin friction.
Probably not of any interest to most of you, but at least next time you see those awesome pictures of CNC dimpled intake ports on websites, you can rest easy knowing that they don't help performance, especially if the whole port is dimpled!
If anyone would like to ask any questions, feel free, I'll try to answer them, or if you'd like to read the dissertation (I wouldn't, it's long) do let me know.

 

[STEVE.M]

I did discuss the possibility of a 4 valve XK head with Harry Weslake over a cuppa in Rye.
He said it would lead to too many stress fractures.
Just like the one`s going on with that Yamaha head..

 

[James363]

Yeah I can imagine, plus it'd be a bit redundant, since no race series that jags of the time regularly compete in would allow it. You'd need to completely redo all the casting as well, so wouldn't be a cheap job...

 

[R3k1355]

Isn't the general rule of thumb that shape is everything?
Finish being neither rough nor smooth.

 

[bozothenutter]

does it hurt performance
what are the beneficial characteristic you're hinting at?

 

[James363]

The simulation calculates the mass flow rate of the pot at different levels of valve lift as well as vorticity and swirl. When you dimple the port, you slow the air down wherever you put the dimples. The dimples do help prevent the flow separating as it passes the valve and valve guide, but the increase in skin friction due to the dimples means the mass flow rate is still lower. The idea of putting dimples is like putting them on a golf ball, it reduces the drag as the air passes over the surface

 

[bashracing]

Did port size make any difference,
Say Making the port 2mm oversize to compensate for the slower air velocity?

 

[bozothenutter]

The simulation calculates the mass flow rate of the pot at different levels of valve lift as well as vorticity and swirl. When you dimple the port, you slow the air down wherever you put the dimples. The dimples do help prevent the flow separating as it passes the valve and valve guide, but the increase in skin friction due to the dimples means the mass flow rate is still lower. The idea of putting dimples is like putting them on a golf ball, it reduces the drag as the air passes over the surface

Ah...the golfball bit...but golfballs rotate aswell. (https://www.scientificamerican.com/article/how-do-dimples-in-golf-ba/)
I remember golfballdimples were being put on anything and everything in the 90's...

 

[bozothenutter]

Did port size make any difference,
Say Making the port 2mm oversize to compensate for the slower air velocity?

wouldn't that make velocity even lower?

 

[bozothenutter]

Oh, and now do some sims on F4R heads.....:smilingimp:

 

[bashracing]

Velocity would be lower but cfm would be higher, I wondered if the larger port would compensate for the lower velocity air where the port was dimpled around the valve stem area and so improving overall flow

 

[James363]

To be fair, the rotation on golf balls isnt the reason for the dimples, a dimpled ball Vs a non dimpled ball static in a windtunnel show the dimples still reduce drag. If anyone has a spare head with good intake ports they'd be willing to send me, I'd be more than happy to 3D scan the port and work on the design of the port as well?

 

[James363]

I only experimented keeping the geometry the same, although it wouldn't be too difficult to increase the port size and find out. That said the increase in surface friction is a loss in energy, which will still be present in the larger port. Any loss in energy contributes to the volumetric efficiency being below 1 (for a NA engine) so the idea is to reduce any losses in energy throughout the intake system.

 

[scullyg40]

So what type of surface finish would you suggest? The opposite of this is a polished finish, but there are some contradictions about this as well, saying air stays attached, and thus reducing air flow.

 

[James363]

Yeah, honestly at this point I'm not sure. Dave Crissel, from Allstsge cylinder heads (very highly respected in the world of golfs and superbikes) ports by hand and suggests a slightly rough surface finish, like a finely sanded surface. I would be inclined to agree, having not experimented too much with this, although I suspect there may be gains to be had using a slightly rough surface finish.

 

[497adam]

Didn’t the likes of David Vizard test things like this to death in the 70s


Sent from my iPhone using Tapatalk

 

[James363]

When I did my literature review I couldn't find any published papers of anyone testing this before, although I'll bet a good few top names have given it a try over the years! The benefit of CFD is you can run 50 simulations without having to buy and port 50 cylinder heads though.

 

[R3k1355]

When I did my literature review I couldn't find any published papers of anyone testing this before, although I'll bet a good few top names have given it a try over the years! The benefit of CFD is you can run 50 simulations without having to buy and port 50 cylinder heads though.

and with a ton of data to back you up it's a much more scientific approach, you're writing this all up as a paper?

 

[James363]

Yeah, well I've finished the paper if you'd like me to pm it to you. Obviously only a small number of Sims were noteworthy enough to make it to the paper though.

 

[497adam]

When I did my literature review I couldn't find any published papers of anyone testing this before, although I'll bet a good few top names have given it a try over the years! The benefit of CFD is you can run 50 simulations without having to buy and port 50 cylinder heads though.

Sorry mate wasn’t to knock your paper. Just meant lots of ideas were tested on the flow bench and then on the dyno. I was more getting at rough verses smooth ports rather than dimples.


Sent from my iPhone using Tapatalk

 

[STEVE.M]

Yeah, well I've finished the paper if you'd like me to pm it to you. Obviously only a small number of Sims were noteworthy enough to make it to the paper though.

Did you read this before you started?
http://ljs.academicdirect.org/A14/112_123.pdf

 

[James363]

Did you read this before you started?
http://ljs.academicdirect.org/A14/112_123.pdf

I didn't actually, I was aware of the requirements for good port geometry, and the port was designed to optimise mass flow rate. But that looks really interesting, I'll be sure to read it in the next few weeks because I would like to really closely examine the affects of surface roughness. Thanks for sending it!

 

[Ezzat96]

Hi all,
I have just finished writing my final year dissertation after taking a break from work to finish my university degree. My employer suggested that I should investigate the effects of adding dimples to the internal surface of an intake port in order to determine if it is beneficial to intake port performance. The port tested was a Jaguar Lightweight E-Type intake port from a wide angle Jaguar head, which had been previously optimised using CFD to improve flow through the port. After nearly 300GB of simulations which have been running for months, I have not been able to improve the flow through the port using single dimples on their own, or covering the whole port in dimples, or even dimpling just the valve guide or valve stem itself. Putting dimples into an intake port does have some beneficial characteristics, but these are outweighed by the increase in skin friction.
Probably not of any interest to most of you, but at least next time you see those awesome pictures of CNC dimpled intake ports on websites, you can rest easy knowing that they don't help performance, especially if the whole port is dimpled!
If anyone would like to ask any questions, feel free, I'll try to answer them, or if you'd like to read the dissertation (I wouldn't, it's long) do let me know.
View attachment 1362487

Hi all,
I have just finished writing my final year dissertation after taking a break from work to finish my university degree. My employer suggested that I should investigate the effects of adding dimples to the internal surface of an intake port in order to determine if it is beneficial to intake port performance. The port tested was a Jaguar Lightweight E-Type intake port from a wide angle Jaguar head, which had been previously optimised using CFD to improve flow through the port. After nearly 300GB of simulations which have been running for months, I have not been able to improve the flow through the port using single dimples on their own, or covering the whole port in dimples, or even dimpling just the valve guide or valve stem itself. Putting dimples into an intake port does have some beneficial characteristics, but these are outweighed by the increase in skin friction.
Probably not of any interest to most of you, but at least next time you see those awesome pictures of CNC dimpled intake ports on websites, you can rest easy knowing that they don't help performance, especially if the whole port is dimpled!
If anyone would like to ask any questions, feel free, I'll try to answer them, or if you'd like to read the dissertation (I wouldn't, it's long) do let me know.
View attachment 1362487

Hello james, my name is Ezzat and im currently planning on doing same paper like yours (dimpled vs normal polish on perfomance) and i would love to read your dissertation. Im very impressed on how did you manage to finish all of those! I mean how long and how much effort does it take? Can i contact you through mail? mine is muhammadezzatluqman@gmail.com . Your help is a big contribution dude.

 

[James363]

Hi Ezzat,
I'll send it through when I get to my PC. It took about 8 months of running my home built computer round the clock to get all the results, on 8 cores at 5GHz AMD FX8350. I'm not sure I'd have done it for fun, but for the dissertation it was worth it ha ha!
I'll be in touch mate

 

[Ezzat96]

Hi James, thanks for the reply.
Wow now my impression gets higher! You also built a powerful rig to just to do the paper? That passion! So the results worth it. Actually before i found this forum and your thread, i'm also planning on dimpling my yamaha head too (but really pricey tho). Now i'm thinking twice Haha!
Yeah big thanks dude we'll sure be.

 

[James363]

Well, I already had the rig, I do a bit of sim racing, so when I built it I beefed it up because I knew I was going to have to run these sims. I wouldn't spend any money on it myself, just get a real nice geometry shape and call it a day there. The dimpling all over the port is certainly a non-starter, way too much skin friction. In places you might see a benefit on some heads, but I didn't. I've just emailed you the paper btw.

 

[R3k1355]

Papers been accepted and stuff now? Where did you get it in?

 

[Slamz]

Pretty interesting this, did you look into the effects of dimpled ports on the mixture of fuel and air? The whole point of not having a mirror polished intake port is to do with fuel atomisation, so it would be interesting to see if the dimples have much effect on this.

 

[R3k1355]

The whole point of not having a mirror polished intake port is to do with fuel atomisation

So this would mean the rules now change for the latest generation of direct injection engines?

 

[Slamz]

So this would mean the rules now change for the latest generation of direct injection engines?

Yep, a big positive to direct injection, is you can perfect fuel atomisation within the cylinder itself, meaning a more efficient burn, and better economy and/or power.

I forget direct injection is a thing half the time, as I hardly ever mess with cars with it.

 

[James363]

The paper was just my final year project for my engineering degree, it hasn't been published or anything, although it was featured in my universities annual research feature journal thing.
Anyway, there seems to be a tremendous amount of smoke and mirrors surrounding the truth with intake ports, a lot of it fuelled by things people quote as fact which actually are not, and stem from some 12 year old on pistonheads quoting theory as fact ha ha.

Anyway, the gist of it is that a rougher surface texture helps energise the boundary layer of the fluid. This very thin turbulent region can be used to create a low pressure which helps "suck" the flow to the contour of the surface in question. As an incredibly rough rule of thumb, the rougher the surface the greater the effect, however the greater the skin friction penalty incurred. Also, the effect of "sucking" the flow to a contour becomes reduced at high reynolds numbers (read: flow speeds). Dimpling is an incredibly rough surface, so works very well with low flow speeds. My paper essentially concludes that the flow in most parts of the intake port are too fast to take advantage of the dimples. When the surface is too rough, it can actually trip the flow off of the surface earlier too, which is what happens around the short side curve on the bottom of the intake port if you put dimples.

Due to the effect only causing turbulence in the outer sort of 1mm layer of fluid in the port, mixing doesn't really come into it at all, which I was quite surprised by, given what I'd read on the internet before starting the project.

I found that the best surface finish was between 0.8Ra and 1.2Ra, so a smooth machined surface but not polished, although the difference is honestly marginal. The only results from my sims which weren't marginal at all, was dimpling the whole intake port, which cased the flow rate to drop by over 10%, which is why I wrote this post in the first place to warn people of the potential harm it causes if you apply a blanket surface texture to the whole port! The only limited success I had was by putting a couple of dimples in the region surrounding the valve guide, which caused the flow to reattach and hold its speed a lot better as it goes round the valve guide and stem. However, I wasn't able to achieve a significant benefit to mass flow rate through the port, unfortunately, only marginal gains.

 

[Ezzat96]

Already received your paper in the email dude. I just a have look at it. Believe me, thats some pretty heavy stuff there;D.Thanks James.
Maybe I'll do the dyno run the engine to see numbers. Sure thing will update here.

 

[bozothenutter]

Awesome that you are getting back to us with results, thanks mate!

 

[bozothenutter]

saw a guy advertising headwork with dimples on FB......was sooooo tempted to point him this way.....

 

[James363]

saw a guy advertising headwork with dimples on FB......was sooooo tempted to point him this way.....

Hahaha definitely should have!

 

[White16valver]

Kudos for doing an awesome piece of research that will actually be useful in the real world!

 

[bloke]

did you do wet flow testing?

 

[James363]

did you do wet flow testing?

No, the petrol/air interface in CFD would have put my simulation run times up from about 8 hours to about 250, and I'd have needed an extra 100GB of RAM or so. In a carburetted engine like the one I studied, it would have been good to do, but since it's largely irrelevant now days, with port injection going out of date, it isn't really all that necessary. Also, at 250m/s, the fuel doesn't really run down the side of the port as some people seem to imagine, more blast through the port without interacting specifically with the port walls.
There is an SAE technical paper somewhere I found that compares a simulation run with a simulated intake charge and one run without. The gist of it was that the two were interchangeable for flow rate optimisation, so long as the Reynolds number variance within the port was kept below a certain margin.

 

[DaveL485]

Going back a few years but i'm pretty sure I read some research on this and it was determined that there was no benefit in a port, and that dimples only work on a golf ball because they create a thin turbulent layer of air that clings to the ball's surface allowing the smoothly flowing air to follow the surface a little further around the back side of the ball, thereby decreasing the size of the wake.

The port doesnt have a back side, or wake, so it won't benefit. Unless maybe you shape your valve like a golf ball

 

[DaveL485]

If you fancy a good nerd session the posts from "PumaRacing" on this thread are very interesting. Bit of reading for your next lunchbreak/skive/poop

3/5/7 angle valve jobs? - Page 1 - Engines & Drivetrain - PistonHeads UK