In-cylinder pressure measurement

[CrazyDave]

I have built a GT-Power model of the f4i, with all data completed except in cylinder pressure measurement data.

I was considering purchasing either a Kistler or cheap alternative probe.

I have spoken to several universities and they expressed that they have taken this measurement in the past.

Is anyone willing to share their measurements with me? I am considering on distributing my model freely in a paper and on a forum such as this.

What is everyone's experience with this type of modeling?

David Ostberg
Lawrence Tech FSAE

[CrazyDave]

I have built a GT-Power model of the f4i, with all data completed except in cylinder pressure measurement data.

I was considering purchasing either a Kistler or cheap alternative probe.

I have spoken to several universities and they expressed that they have taken this measurement in the past.

Is anyone willing to share their measurements with me? I am considering on distributing my model freely in a paper and on a forum such as this.

What is everyone's experience with this type of modeling?

David Ostberg
Lawrence Tech FSAE

[drivetrainUW-Platt]

why would you wana know in cylinder pressures anyways?

[Garlic]

Read the first sentence of his post.

[Nima]

I know that the manifold pressure is near to the 100kpa... in the same reason the in-cylinder pressure may be near 100kpa with the pressure head loss that depends on your intake piping...
finding out the in-cylinder pressure get the very complicated computing ... so it seems wisely to consider it 100kpa...

if I'm not right, correct me...

[Greg 08]

You are fairly accurate with that nima, however some ltu people (dr. claw?) were posting about their turbo earlier (around design report date?) so that wouldn't work. Also, I think he wants instantaneous data as his engine goes through compression, not at bdc. Probably some high dollar testing, boy I wish our engine team could do that........

Greg
MTU FSAE

[Kamil S]

i don't know too much about having actual measurements, but I have a Matlab model that would generate your in cylinder dimensionless pressure as a function of crank angle.

Otherwise, as you mentioned, you need to get one of these:

http://www.public.asu.edu/~ksalloum/myspace/engine%20007.JPG

[CrazyDave]

I have found an alternative cylinder pressure measurement; it has the sensor with associated circuitry for $800. It has a 12V source, 0V reference and an output. Ill find out the company and post it if anyone is interested. The thing is its time and money, and if someone out there is willing I would rather exhange data then spend money

The reason I want the measurement, is we mess around with different pistons adjust 15 other variables and don't necessarily know where improvements and interactions are in the system.

I want to know what peak pressures are, see how pre-ignition occurs in this situation, and lastly get my model to account for the pressure profile correctly. Right now I am using default multipliers, which I don't know if they accurately describe what is occurring in the cylinder. I figure it's not a bad way to engineer the engine/transmission. Btw, I am not very concerned with the higher output torque, because the life of the gears is diminished, but I do not believe to the point of concern. My focus is on the engine.



Nima, if you want a log P-V diagram of what I have you are welome to it. I will post it when i get a chance.

David Ostberg
LTU FSAE

[Jersey Tom]

Not very concerned about higher output torque? That's the #1 thing I would be going for.

[CrazyDave]

To each his own on the gear life problem, btw, when i said not concerned I would still limit to 60ft-lb

I've learned some interesting things on manifolding with this model I would love to see what kind of cylinder pressures it equates to. For instance on a turbocharged model certain intake manifold geometry I have a Voleff of 1.64 (ref atm.) on one cylinder and 1.12 on another cylinder....

LogPV (http://community.webshots.com/photo/548106081/2991107390074791088lAIQUr)

compressor (http://community.webshots.com/photo/548106081/2069426380074791088JDRAij)

[CornellGixxer]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by CrazyDave:
To each his own on the gear life problem, btw, when i said not concerned I would still limit to 60ft-lb
</div></BLOCKQUOTE>

Im surprised you are running into or anticipating gear failures due to increased torque production. Unless you've got some uber stiff clutch springs or huge weight transfer I would expect clutch slip or spinning tires before any gears give out. I know ive seen one or two roasted clutches when the stock springs where left in but other than that traction was the limiting factor.

[CrazyDave]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by CornellGixxer:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by CrazyDave:
To each his own on the gear life problem, btw, when i said not concerned I would still limit to 60ft-lb
</div></BLOCKQUOTE>

Im surprised you are running into or anticipating gear failures due to increased torque production. Unless you've got some uber stiff clutch springs or huge weight transfer I would expect clutch slip or spinning tires before any gears give out. I know ive seen one or two roasted clutches when the stock springs where left in but other than that traction was the limiting factor. </div></BLOCKQUOTE>

my commment was in reference to it being the #1 problem. I recently read a paper that described a in-line two cylinder engine that was turbocharged, they ran into a problem with torque. I was just stating i dont think its a concern, and that the tyre is the limiting factor so it appears we agree.

Power is not produced in an area where high torques are required, I just thought 60ft-lbs is a good maximum for indicated torque. What is your stance on this issue? I love your electronic acutated wastegate, the flexabilty of that is marvelous. What is the actuator that you are using at Cornell? We are using a motec boost controler, but are running into problems with near the redline our stock wastegate being still too stiff.

[Mike Claffey]

I have never used GT power but I'm curious to know how well a quickly made model correlates to your dyno figures, and how difficult it is to calibrate your model. (rough figures in hours anyone?)

Were also at the stage of requiring a kistler p-transducer for furthur accuracy in our engine sims, and the figures we have been quoted are fairly pricey - no way thats in the budget this year. So in answer to your question CrazyDave were looking for cylinder pressure data to possibly tune our combustion model - and we are usually happy if what we see on the dyno comes within 5 or so % of the dyno data, with correct trends in the torque curve.

Mike C
UWA FSAE

[repeatoffender]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Mike Claffey:
I have never used GT power but I'm curious to know how well a quickly made model correlates to your dyno figures, and how difficult it is to calibrate your model. (rough figures in hours anyone?)

Were also at the stage of requiring a kistler p-transducer for furthur accuracy in our engine sims, and the figures we have been quoted are fairly pricey - no way thats in the budget this year. So in answer to your question CrazyDave were looking for cylinder pressure data to possibly tune our combustion model - and we are usually happy if what we see on the dyno comes within 5 or so % of the dyno data, with correct trends in the torque curve.

Mike C
UWA FSAE </div></BLOCKQUOTE>

GT Power is a fancier version of desktop dynos or virtual fourstroke.

Its still a one dimensional simulation program that mimics higher order simulations via correction factors and correlated empirical data in the form of tables. Such as your compressor map, which i assume you had to individually measure the points off the map. The only advantage i can see over other programs is that it already contains fuel - air cycle information.

To setup a full model only takes about a day, but to tweak it to give reasonable results will take a long time and a dedicated dyno testing program to supplement.

Dont get me wrong the program has serious potential but to get any data that actually means or correlates to the physical engine is a difficult and time consuming process.

Im also interested in this 5% error combustion model mike. That seems pretty accurate to me, is it steadystate or transient? Also to what depth does this model account for? i.e. ignition, burn, multi-zonal, fuel-air energy, residual gas recirculation, turbulence models etc?

Theres some good references in SAE papers on burn and multi zone models if you want the links i could dig them up for you. But a google search should provide what your looking for.

[Mike Claffey]

repeatoffender, when I say "we are happy if it comes within 5%" I mean when it does we are really happy, when it doesn't, well its just another day in the lab. I also made that comment refering to a more general power or torque curve generated from an engine sim. I wouldn't think our combustion model is anywhere near close to being 5% near reality. Thanks for the info on gt power btw.

I personally start to disbelieve most cfd results when they are getting below 5-10% of reality.. because I know how much fudging goes on (hard to measure or estimated parameters etc) to make them.

[CrazyDave]

Are you (UWA) using the F4i in the future? What software have you been using for simulation? GT-Power is $100 for a licence to us, but i dont know about the international issue.

Would you be intested in a collaboration? I am currently working getting the model to co-simulate with STAR-CD due to the complex intake manifold geometry we use.

The model i made isnt a quick one day model. I suppose its possible with doing lookup tables as you suggest, but the model i made is certainly more complicated.

We have done cam and lift profiles and the flowbench data from a standard and extrude honed head. The (baseline model is complete and matches well to the f4 bike data i found, but our dynos havent been very operational for awile... so the answer to correlation is i dont know.


The model was used in the turbo selection and decision process to establish saftey limits and expecations.

[CrazyDave]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Mike Claffey:
I have never used GT power but I'm curious to know how well a quickly made model correlates to your dyno figures, and how difficult it is to calibrate your model. (rough figures in hours anyone?)

Were also at the stage of requiring a kistler p-transducer for furthur accuracy in our engine sims, and the figures we have been quoted are fairly pricey - no way thats in the budget this year. So in answer to your question CrazyDave were looking for cylinder pressure data to possibly tune our combustion model - and we are usually happy if what we see on the dyno comes within 5 or so % of the dyno data, with correct trends in the torque curve.

Mike C
UWA FSAE </div></BLOCKQUOTE>

so your saying that the torque is w/in 5%, how close is the pressure profile is what I am wondering?



How many hours to calibrate the model to the profile? that is very quick infact GT-power has a profile matching program they include in excel format to get a wiebe fit. (less than an hour) you can also convert the pressure profile to a heat release rate profile and use that directly.

How long it takes to build the model is how long you want to spend onit.. haha... i would say i have 100+ hours into the model(s).. you can get carried away with playing on DOE`s and optimization studies really quickly

[Mike Claffey]

CrazyDave, unfortunatly I dont have time right now to write a proper answer, but software wise simulation is done using WAVE/Vectis, and I've been coupling WAVE to Star-CD (restrictor/exhaust). Our intake guy will most probably be using WAVE/Vectis for his 1D/3D stuff. I thought I might have been alone with the Star work, so I'll try and get in contact with you through email.

We are using the F4i this year, but I can't comment on what is happening in the future. It sound's like our model was created in much the same way as you have created yours.