Calculate the flow of cam/head system

Competition engines and 'live' projects only. Good photos to illustrate your post are expected.
Post Reply
Robert
Posts: 86
Joined: August 20th, 2006, 8:17 am
Location: France
Contact:

Calculate the flow of cam/head system

Post by Robert »

Guy you will certainly comment this! At a first glance my reasoning is correct but further explanation is required...

I tried to link 2 important parameters concerning the filling of cylinder:
-The flow of the head function of the valve lift (Q=f (L))
-The lift of the valve function of the crank angle (L=f (CA)).
The idea is to get a curve of the flow function of the crank angle (Q=f (CA)).
If the result is in a table to the Excel format it is quite easy to calculate an integral and the mean of this integral so obtain the mean flow during a cycle of opening valve for a given head.
The goal is to compare for the same head the difference of flow between 2 camshafts and vice versa or mixing both modifications...

So I started from the GC figures found in the 1rst book for the flow of the head and the figures of the Ritmo 130TC camshaft I measured and the ones of the St3 rally camshaft found in the 2nd GC book.
Here below a drawing showing the idea in 3 steps (1) (2) (3)


The method to calculate integral is available on the web (or from your memory...). I used the trapeze method simple to implement in Excel (and giving good results compared to the real integral of the corresponding polynomial function).
The spreadsheet snapshots show the initial tables for Q=f (L) and L= f (CA) and the associated curves.
You can see also the resulting tables for Q=f (CA).
The mean flow is about 34 cfm for the OE cam profile and the mean flow is about the double 68 cfm for the St3 rally cam profile.

Obviously the torque and the power are not doubled too when you change the first camshaft to the second one.
But here I cannot say more because I don't know all the parameters entering in the obtaining of power.

Regards

Robert
PS: sorry I did not put the Excel worksheet as image but I did not find how to do...
Attachments
Steps to get the Flow f(CA)
Steps to get the Flow f(CA)
LiftAngleFlow.JPG (26.15 KiB) Viewed 9701 times
The example with 130TC, ST3 cam and GC head flow
The example with 130TC, ST3 cam and GC head flow
LiftAngleFlow2.JPG (68.29 KiB) Viewed 9701 times
Calcul Integrale Debit GC.xls
Excel File
(31 KiB) Downloaded 410 times
GC_131
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Calculate the flow of cam/head system

Post by Guy Croft »

Thank you for this interesting technical post Robert.

I look forward to studying it when I get time,

G
Michael
Posts: 4
Joined: June 25th, 2007, 8:33 pm
Location: Mellrichstadt, Germany (A)
Contact:

Re: Calculate the flow of cam/head system

Post by Michael »

Since I was thinking/working on exactly the same topic I thought it could help to share my conclusions.

First let me briefly introduce how I got to the point where I was dealing with the same tasks you are thinking about now.

The intention for me was to reveive a comparable value for different configurations of camshafts, cylinder head modifications, valve seats and inlet valves. My method of evaluating the effects of the different mods on head and valves was the use of a "virtual flowbench". I generated CAD geometry models of a Peugeot cylinder head and the respective inlet valves and did 3D flow simulations with Fluent and CFX.To be sure to get comparable results I "adjusted" my simulations with the help of a flowtest diagram for the same head which I found on Guys website.

Since it is just not practical to always look at all the diagrams trying to figure out what the meaning of the results could be and how this or that would work out in the end, I thought of one value to specify how good a certain port and a camshaft profile could work together. By connecting the flow vs. valve lift to the valve lift vs. crank angle values you get a good impression of how big the flow improvement could be in the end. I called my result the port/cam-flow-index (PCFI).

The problem with it is that - in my eyes - it is almost impossible to specify how the flow improvement will affect the horsepower in the end. First thing to mention is that the flowbench results are all generated with one certain value for the negative pressure. In the real engine the negative pressure is not only a function of the piston speed but also of valve position and speed and of the performance of the port. Imagine two ports where one is flowing significantly better than the other one. This means that at the same negative pressure the better port will deliver more air into the combustion chamber. But at the same time the higher amount of air will change the negative pressure at that moment.
This means that the negative pressure graphs for one engine cycle can be totally different when comparing two differently performing ports on the same engine. Same with valves, valve seats, cam profiles and cam timing of course.

All those unknown factors gave me the impression that it is impossible to generate one single value that gives you the opportunity to foresee or calculate the increase in horsepower for a certain setup. Keeping in mind that "all the power lies in the head" it can be a good approach to be at least able to compare the possible effects of different setups. But I'm afraid that all those more or less theoretical values are simply not reliable enough to receive the predictability you (and me, too) are aiming for, since an engine is such a complicated interaction between many parts and too often the changing of one part directly affects the performance of another one.

Finally some pictures of the documentation I did some time ago concerning this topic. Hope you can live with the sometimes german annotations.
This is the CAD-geometry I built for the simulations
This is the CAD-geometry I built for the simulations
Geometry.png (145.25 KiB) Viewed 9601 times
Comparison of measured (by GCRE of course) and simulated flow graph for basic setup
Comparison of measured (by GCRE of course) and simulated flow graph for basic setup
Adjustment.png (32.77 KiB) Viewed 9601 times
Visualization of how to calculate the PCFI value
Visualization of how to calculate the PCFI value
PCFI.png (65.92 KiB) Viewed 9601 times
Example for changes in terms of PCFI for different setups with identical camshaft
Example for changes in terms of PCFI for different setups with identical camshaft
example PCFI changes.png (34.64 KiB) Viewed 9601 times
Finally to give some evidence to my statements concerning negative pressure in the combustion chamber a diagram which shows the gas exchange cycle for three basically identical engines, measured with indicating tool AVL Indicom. You can see e.g. the different valve timing due to different zero positions of the (untriggered) cam phasers by looking at the rising pressure on the left (ex valve closes before piston reaches TDC, in valve opens after piston passed TDC).This work was done by me while I was working for a german engine test/research lab.
Different gas exchange cycles of identical engines
Different gas exchange cycles of identical engines
gas exchange cycles.png (40.26 KiB) Viewed 9601 times
Greetings
Michael
TomLouwrier
Posts: 333
Joined: July 28th, 2010, 3:09 pm
Location: Leiden, The Netherlands
Contact:

Re: Calculate the flow of cam/head system

Post by TomLouwrier »

Interesting stuff!
This ties in with conversations here
viewtopic.php?f=13&t=2743
and here
viewtopic.php?f=4&t=541&start=15
You can approach this in two ways.

The first is to try and model the exact and mathematically predictable behaviour of an engine. You'd be calculating all the flow in all the places and how things interact with each other. Very, very complex and computing power consuming. In my opinion about as useful as trying to predict the climate 20 years in the future. (Try to get next week's weather right, that would impress me.)
Both systems (engine and weather) are fundamentally chaotic so this road is not feasible or even possible.


The second way is to find a set of indicators that, although they do not describe the actual behaviour of all elements involved, give a good indication of what the overall effect is on the system's performance.
Given that we (as in 'humanity', not you and me) have established that testing on a flow bench is a very good way of developing a head, and that 28" H20 will work (Yunnick) but 10" H20 is just as good a yardstick (Croft) you may draw the conclusion that it is, in fact, possible to test against a limited number of well defined conditions and estimate the overall effect on your engine from that.

Just like the oil pressure gauge does not report all actual pressures in all relevant places (cam, crank, turbo shafts etc) but we know if one's OK (2 bar, 6 bar, whatever is 'good'), then the others will be fine too.
In order to read these 'gauges' (you'd be calling them KPI's: Key Performance Indicators) you must test if the reading you see actually tallies with the performance / behaviour you're getting. A lot of testing: finding the right indicators and drawing the right scale behind the needle.

The art is in finding these indicators, proving they work and calibrating their 'scale value'. This means collecting very much data about real and proven engines, doing a lot of statistical analysis in order to see and prove correlations and thus setting up the indicators you need.
When that works you must set up rules that describe the behaviour you expect when you change certain things (cam timing, port size and shape, valve lift, etc etc).
This part is the 'expert system'; it must be filled by humans with a lot of knowledge and experience before it can do anything. It has to learn from us.
Do not underestimate the human mind, it is a very complex and powerful thing. We don't know half of how it works or what it can do. When you ask Guy advice on cam timing and jet setting for a certain setup, have you ever thought about what happens inside his head? Something buzzes and clicks, then the right answer comes out. Amazing.

Next step would be confronting your expert rule base with real world, empirical data. Prove your rules (formulas) work and calibrate the values, or disprove a rule and drop it. Classic scientific approach.

This, by the way, is how my company functions, be it in quite another field of expertise (financial analysis of complex and large scale IT environments; matching technology, capacity, costs and usage). Yes, I'm dreaming of having both time and opportunity to sett up one of our systems for an analysis like the one described here. People like messrs Croft and McMullen would have the ideal brains to pick for teaching the 'expert system'. Collecting field data for the statistics is not difficult at all, just a lot of work.

There are of course some very sophisticated engine simulation packages around. I'm not sure how they work, but expect a hybrid approach: partly fundamental / mathematical and partly empirical / statistical.


regards
Tom
GC_29
WhizzMan
Posts: 459
Joined: August 13th, 2010, 8:05 pm
Location: Amsterdam, the Netherlands

Re: Calculate the flow of cam/head system

Post by WhizzMan »

One of the reasons I find David Vizard interesting to read, is that he is one of the persons that spends a lot of his time trying to find those KPIs and come up with models that will help calculate performance and improve on them. He can afford to "spend all day" experimenting with just lobe separation angles, or valve seat width or angles or such. He's proven his previous theories wrong more than once, but that doesn't mean he has nothing interesting to say or isn't building engines that win races. Not all of his stuff translates well to other configurations than he's used for testing, but a lot of it does.
Book #348
TomLouwrier
Posts: 333
Joined: July 28th, 2010, 3:09 pm
Location: Leiden, The Netherlands
Contact:

Re: Calculate the flow of cam/head system

Post by TomLouwrier »

Yup.
Vizard, Croft, Yunnick, Helgersen. Men who work with more than 'just' engineering and working knowledge of engines: they practice proper scientific methods.
I'm with you on this one, Whizz.

Tom
GC_29
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Calculate the flow of cam/head system

Post by Guy Croft »

I am finding it hard to make time to reply to Robert's thought-provoking post though I must say it looks great fun.

This in Michael's post caught my attention:

"All those unknown factors gave me the impression that it is impossible to generate one single value that gives you the opportunity to foresee or calculate the increase in horsepower for a certain setup"

And he is quite right. The intake system and valve event taken as a whole are merely a 'facilitator' for pressure waves to do their work. That's where the real power comes from. Get the shape and volume of the tract right and time the valve event 'right' and the pressure wvaes can ram a decent charge into the cylinder and the engine works really well. In saying 'all the power is in the head' - well of course it is - all the crankcase assy does in actuality is transfer the energy to the drivetrain and generate enormous losses in doing so. But get the inlet layout (in the broad sense) 'wrong' and it diminishes the effect of the pressure waves. Get the valve event wrong or relative inlet-exhaust lengths and the engine will never 'light up'. I mention these things to underscore what Michael (in particular) has said - you can't predict power direct from flowtest. You can certainly archive and correlate dyno and associated flowtest results and make quite good comparative assements but that is another matter altogether.


G
Robert
Posts: 86
Joined: August 20th, 2006, 8:17 am
Location: France
Contact:

Re: Calculate the flow of cam/head system

Post by Robert »

Hello Michael,

I see that you had the same concern than me.
One remark: I did like you the extrat of the polynomial equation up to 6th order like you shown on your picture. Then I calculated the integral with this polynom and I was surprised that I did not find the same result than the trapeze method. Finally I used this equation to re-draw the curve using the same absice points and the curve is not the same ! Explaining the bad results. I tried with a degree 4 and there the curve was right and also the integral the same as the trapeze method. So may be Excel is not so accurate some time (but software bugs are common...). All that to say I'm using only the trapeze method working well and directly using table without passing through a curve modelization given by Excel.

All,
More globally, I still wonder why the increase of power is not so high when the mean flow on a cycle is doubled (let say 20% of power increase: 150 to 180 HP with this kind of camshaft difference).
May be my results are wrong (To check the calculation is possible with my file).
There are other effects to take into account (for sure). I think about some but not sure I'm the best person to list them so if you can try to give some limiting factors there it will be great to continue the debate.
See you soon on GCRE !!

Regards

Robert
GC_131
Robert
Posts: 86
Joined: August 20th, 2006, 8:17 am
Location: France
Contact:

Re: Calculate the flow of cam/head system

Post by Robert »

Hello,

I think I made a mistake when I measured the lift of the 130TC cam.
In effect if I consider the lift integral of 1557 Degree x mm for a 132 cam in this topic :
viewtopic.php?f=6&t=1809&hilit=lift+integral#p9530
it seems that the 896 deg x mm I found is not the good number. Specially if I compare to the 1842 deg x mm for the GC ST3.

So is it possible to get the right lift function the crank angle for a 130TC cam ?

Thank you !

In attachement a new version of the Excel file with the calculation of the lift integral and the drawing of the total cycle instead of the half as I did the first time.

Robert
Attachments
Calcul Integrale Debit GC 2.xls
Lift integral, flow integral calculation
(59 KiB) Downloaded 416 times
GC_131
Robert
Posts: 86
Joined: August 20th, 2006, 8:17 am
Location: France
Contact:

Re: Calculate the flow of cam/head system

Post by Robert »

Good evening,

I confirm that the values of the 130TC cam I wrote in the excel file are wrong. I apologize for that. I don't know where these numbers come from. I made measurements years ago and I retrieved them but they are not enough precise and trustable to be communicated.
Nevertheless the integral method is right.

So I’m looking for the right lift values for the 130TC camshaft in order to correct my mistake.

Guy, do you have this camshaft values in your folders ?

Merci !

Robert
GC_131
Robert
Posts: 86
Joined: August 20th, 2006, 8:17 am
Location: France
Contact:

Re: Calculate the flow of cam/head system

Post by Robert »

Hello it's me again...

On the forum I found some weeks ago a cam profile of the 130TC (I cannot find it any more but it was a drawing not a table of values).
So I did a new version of the file.
Because the figures are not sure at 100%, let's say it is a described method more than a true result. To me it is enough to understand "what happens" when you use different cams in the same head (the reverse is possible too).

I think it is also a way to predict approximately the increase of power due to the change of an element (cam or head) because you can see the flow increase.

Of course it is not a professional who is writing... and I have no pretention at all.

Any comment is welcome!

Robert
PS:slight correction in the attached file
Attachments
Flow Integral GC 4.1.xls
(95.5 KiB) Downloaded 427 times
GC_131
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Calculate the flow of cam/head system

Post by Guy Croft »

ALL - PLEASE ASSURE ME THESE ARE NOT FULL GC CAM MAPS ON SHOW?

G
Post Reply

Who is online

Users browsing this forum: No registered users and 34 guests