Yet another WAVE discussion

[Bazanaius]

Hey guys, I know there has been a lot of threads on here recently about WAVE and so I'm loathe to start more, but I havn't been able to find anything about my question on them..

We had WAVE up and running for the first time last week and were pleasantly surprised with how quickly we got to grips with it. We had an approximate engine model within 3 hours, and we're now obviously working hard to refine it and to then start using it as a development tool. My question is really more a wondering, regarding what people use for the miriad of values that teams like us (without a dyno or much test equipment) just can't measure? I'm talking about liner temps at various rpm, burn durations, friction coefficients and heat transfer coefficients for all kinds of things! etc. etc.

We are working on putting together a flow bench asap to give us as much data about our various parts as we can, but I was just wondering what people's exeperiences of these variables were, and how they dealt with them or what their thoughts were on the required accuracy of some of them?

cheers,

Baz

[Bazanaius]

Hey guys, I know there has been a lot of threads on here recently about WAVE and so I'm loathe to start more, but I havn't been able to find anything about my question on them..

We had WAVE up and running for the first time last week and were pleasantly surprised with how quickly we got to grips with it. We had an approximate engine model within 3 hours, and we're now obviously working hard to refine it and to then start using it as a development tool. My question is really more a wondering, regarding what people use for the miriad of values that teams like us (without a dyno or much test equipment) just can't measure? I'm talking about liner temps at various rpm, burn durations, friction coefficients and heat transfer coefficients for all kinds of things! etc. etc.

We are working on putting together a flow bench asap to give us as much data about our various parts as we can, but I was just wondering what people's exeperiences of these variables were, and how they dealt with them or what their thoughts were on the required accuracy of some of them?

cheers,

Baz

[Charlie]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Bazanaius:
I was just wondering what people's exeperiences of these variables were, and how they dealt with them or what their thoughts were on the required accuracy of some of them?
</div></BLOCKQUOTE>

You have the tools to answer this question yourself. Think about it. http://fsae.com/groupee_common/emoticons/icon_smile.gif

[Professor Gas Can]

Bazanaius - I would tend to focus on more important things before you worried about such things as liner temps and HTC's.

Mass flow rate is the purpose of 1D simulation and is what predicts the highly coveted HP/TQ graphs you see on everyone's design boards. There are many ways to get MFR but the easiest off the top of my head is to use a MFR sensor. Last time I checked you don't need a dyno for it. It's not pretty, but I betcha it'll dial it in pretty close (within 10 g/s). Nay sayers will say the waves from the engine will screw it up because of the reverse flow effects from the manifold, but mount it upstream of the throttle and I don't foresee too many problems. If anyone gives you trouble, tell them to give you money so you can analyze the chemical composition of the exhaust gasses (combined with A/F ratio) so you can back calculate the MFR.

After that, concentrate on the things that affect mass flow rate into the cylinder - valves mainly. Get CD-CA's dialed in, the correct profiles...etc. Good start with the flowbench but keep in mind that manifold effects are not steady state. Steady state is a good starting point, but you'll catch flak from fellow FSAEers and design judges that transient data is needed...CFD is your friend in this case, of which Boundary conditions can be predicted using ur 1D model.
Then consider the nightmare that is combustion, fuel and spark timing to get your cylinder pressures. (Good luck here without DAQ or DYNO)

Overall, the message is don't be relying on the model to give you anything of significance if you can't get it correlated. You might be 25 HP off and not even know it. We at GM can predict hp within 2% accuracy, so that's the top. I'd say a 7-10% agreement is realistic for FSAE and 30%+ for an uncorrelated model... Also, do not get overly excited and place any confidence in any intake or exhaust you design using the model if it's not correlated. One of the biggest pet-pieves of mine is seeing people's design boards claiming they've "optimized" parts of their car (intake and exhaust) but don't have a correlated engine model. Of course having a dyno solves all of this ambiguity.

Oh right, tutorial/sample models is a great way to start assumptions on HTC's and head temps. It's up to you to be an engineer and make the right assumptions after that. I apologize, this was a bit of a rambling post (good stuff though, hopefully).

[Bazanaius]

Hi guys, thanks for your responses.

Charlie - I agree, about estimating the required accuracy of the variables. I think I worded that a bit wrong.

Prof Gas Can, good points that I think will definitely start us off on the right track. I agree that correlating the model is v. important. We are looking to run this model in tandem with some track testing on last year's car (with modified parts being tested) and some dyno work to gather as much info as we can and correlate our model. Unfortunately the dyno isn't ours, or I'd have been on it already!

We'll certainly investigate MFR further - I guess this is a further way to correlate the model to your actual results?

This is our first year with this software (and the number of people on the team who are keen enough/know what they're doing) and so I'd be surprised if we get beyond the steady state stuff this year, but certainly dialling in our valve attributes and behaviour is something we can work on. It really is a case of taking what we can get with what we have available and then building on it as we can - it would be awesome to carry this development on with the guys next year.

Thanks again for your responses - if anyone else has any thoughts please post away!

B

[Bazanaius]

Hi Guys,

we're slowly getting there - meeting up tomorrow eve to work on it more. Something I wonder if anyone can help with - does anyone have the intake/exhaust valve profiles that they could send me?

I've had a look on the internet and come up with a couple but one I'm not sure about and the other is just an image so not terribly accurate.

any ideas would be great.

cheers!

B

[Bazanaius]

gents,

another question on WAVE simulations I'm afraid.

on setting up fuel injectors - I notice the setting for fuel/air ratio to be set in the injector properties. I wonder if it is simply possible to specify a quantity of fuel to inject? I would be interested to know the effect of a non-symmetrical plenum and our ECU supplies a quantity of fuel from the map, rather than supplying enough for a fuel/air ratio desired. Being able to set the amount of fuel injected (for each case) would be good, so I was wondering if anyone knew whether this is possible?

cheers in advance,

Baz

[Charlie]

I don't have access to WAVE at the moment, but I believe there is a setting that lets you specify an injector size and a pulsewidth instead of an A/F ratio.

However I don't feel this level of detail is necessary. In the end you just need to find/decide the A/F ratio that works for you and just use it.

WAVE is not going to give you a very accurate result of fuel droplet propagation through a non-symmetrical program... it's a 1D simulation. Use it for volumes, runner lengths, cam timing, etc.