|
924Board.org Discussion Forum of 924.org
|
View previous topic :: View next topic |
Author |
Message |
Cedric
Joined: 27 Aug 2004 Posts: 2614 Location: Sweden
|
Posted: Fri Dec 16, 2016 11:46 pm Post subject: Engine performance simulations 924 Turbo |
|
|
To continue on the performance topic i will show some data from a GT-power simulation i have been working on from time to time, wirlds first 1d simulation of a 924 engine as far as I know. My car is a reference and then i tried to build up the engine model. Lots of things are approximated geometry wise since i havent had my 931 close by. Also the cam profiles are approximated and port flow curve, having those would have been really heplful. Maybe I can flow the head here at work some day, but not in the near future though.
Specs:
Turbine, newer K26 scaled to get similar performance as engine test.
Comp, 2664 and 2670 digitalized from old maps
Ignition and combustion
Ignition curve is from 931 s2 manual, combustion characteristics is approximated, i have no predictive combustion/knock model, to bad.
Pressure losses are semi simulated, semi guestimated.
Charge cooler is similar performance to cgt, ca 45degC out of the IC.
This is run at a load sweep similar to the one from my dyno session, at 300rpm/s
Here is an example plot of engine power. I had to rev the 2660 and even the 2664 to the moon to boost 1 bar at 6500rpm, definietly not safe and recommended. CarreraRSR, can your 2660 comp really keep the 1bar at 6500rpm?
and compressor efficiency, here you can see why porsche went for the 2670 on the GTS and 951, the other compressors are crashing through, and need dangerous revs to pump that air. Compressor outlet temperatures are really really high with the smaller comps at 1bar 6500rpms.
If you could help me with geometry data for the engine parts the simulation model will only get better and better. I do this for a living as a gas exchange/turbo system expert, but on heavy duty diesel engines, but then you have hundreds of parameters from test cell that you can use for calibration, and cad models do disctretize. So a completely different world..
I can rebuild this to do whatever I want, for eaxmple change to twin scroll manifold, optimizing cam timing, twin turbo setup, large or small intercoolers, tuned intakes, change bore and stroke, comrpession, add driveline to simulate accelerations etc etc Only limitation is time
More interesting info will come further on... _________________ 1980 924 Turbo
www.instagram.com/garagecedric/ |
|
Back to top |
|
|
Carrera RSR
Joined: 08 Jan 2010 Posts: 2309 Location: Somerset, UK
|
Posted: Sat Dec 17, 2016 12:52 am Post subject: |
|
|
Hi Cedric
I rarely rev mine over 5800rpm for several reasons
*Previous and current dynos always showed torque and power dropping off from 5600-5800rpm, so why spank it with ever decreasing return. Could be caused by 2660 ?
*The current cam has timed with advance to induce earlier boost/power with the inevitable loss at the top end. Its a Kent 270 fast road not stock.
*I have usually turned up boost for low rpm improvements which can be lost at the top end of rpm. If I turn down the boost below 1bar then I can rev higher more easily, but I have lost the kick in the butt from 2500-4500rpm
Yes, I can hold up to 1.4bar all the way to 5800rpm or so. 1.0bar was proven to hold all the way up to 6400rpm on the dyno.
*Post IC temps are usually 6-8deg above ambient air temps on long pulls. No more than 6 deg on shorter boosts.
*EGTs are circa 850deg on long boost at circa 1.2bar
I do need to get the car properly dyno'd now all the bugs are ironed out to find the optimum tune. All I know is 1.0bar and 225bhp crank is lively. 1.2bar is more livelier again on the butt dyno. So the 2660 cannot be wimping out completely at higher boost! Sometimes boost up to 1.4bar, but thats where I believe via the butt dyno that there's potentially too much turbo stress occurring and no power gain. But the boost noise is addictive!!
Hope that helps
Steve _________________ 1980 931 - forged pistons, Piper cam, K27/26 3257 6.10 hybrid turbo, 951 FMIC, custom intake, Mittelmotor dizzy & cam pulley, H&S exhaust, GAZ Gold, Fuch'ed, Quaife
Now www.924board.org/viewtopic.php?t=34690
Then www.924board.org/viewtopic.php?t=31252 |
|
Back to top |
|
|
gegge
Joined: 27 Jul 2007 Posts: 1124 Location: Sweden
|
Posted: Sun Dec 18, 2016 2:53 am Post subject: |
|
|
The standard 931 head flow 143 cfm@9mm and 28” of H20 and drops back to 140 cfm with higher lift due to increased turbulence. It is almost a straight line from zero lift to 9 mm which pivot and levels out. The 924 head benefits from a larger radius at higher lift. A simply DIY in order to remove and smooth out edges raise the flow to more than 160cfm@10mm before it settles. This show just how important the area prior to the valve seat is. This is a 10-15% increase in flow AND output.
Stock cam is a fairly hot for a turbo application with 223 degrees@0,05” which is the most used measure of duration. Porsche specify it 228 degrees at 1mm instead. The 12mm lift is accomplished by taking advantage of the large diameter of the followers.
When I am trying to simulate the 924 engine in less advanced programs, I usually start with a natural aspired version. It is important to get the curvature of the torque through the rpm range right before you add boost. Since the head is choking above 5000 rpm, there is really not much gaining revving more. With the benefit of additional boost something happens above 5000 rpm. The head breathe a lot easier with support from the turbo. But still, at 5500-6000 even with a decent IC the song is over unless you port the head, swap the turbo for something bigger or go wild is the cam territory.
Do you have any numbers for the backpressure? Preferably with the logshaped infamous 931 exhaust manifold. If so, cam evaluation can be made with interesting outcomes. Guessing overlap and tight LSA isn´t going to be a problem...
With reference to the SAE paper in a previous topic: The 931 wastegate was actuated by either the boostpressure or the backpressure and it opens at 0,7 bar dumping the exhaustgases directly into the exhaustpipe. Therefore, the way I see it, the backpressure is lower than boost until 5200 rpm with the #6 turbine. Are you able to replicate that number? (The same situation can be seen in the early 911 diagram where P3 is significantly lower than P2 until 4300 rpm. Guessing that was the K26 related 3LDZ)
Would you mind running a sim with the #8 hotside? Maybe with a K27 3070? Please… _________________ Carl Fredrik Torkildsen
924 turbo -81 Carrera GT RESTOMOD
924 turbo -80 Dolomite De Luxe
924 -85 DP kit, BBS RS, M030 and tuned engine
924s -86 Black on black turbo with Fuchs |
|
Back to top |
|
|
morghen
Joined: 21 Jan 2005 Posts: 8884 Location: Romania
|
|
Back to top |
|
|
Carrera RSR
Joined: 08 Jan 2010 Posts: 2309 Location: Somerset, UK
|
Posted: Sun Dec 18, 2016 7:34 pm Post subject: |
|
|
Good point on the head flow Carl. When I first rebuilt the motor I had a ported head. Unfortunately this head had big issues, possibly cracked internally. So I swapped to a basic refurbed head with no porting. Would be nice one day to switch to a Mittelmotor CNC head and upgrade the turbo compressor. But at the moment it ain't broken so am reluctant to pull apart until something does go pop. _________________ 1980 931 - forged pistons, Piper cam, K27/26 3257 6.10 hybrid turbo, 951 FMIC, custom intake, Mittelmotor dizzy & cam pulley, H&S exhaust, GAZ Gold, Fuch'ed, Quaife
Now www.924board.org/viewtopic.php?t=34690
Then www.924board.org/viewtopic.php?t=31252 |
|
Back to top |
|
|
Carrera RSR
Joined: 08 Jan 2010 Posts: 2309 Location: Somerset, UK
|
Posted: Sun Dec 18, 2016 7:58 pm Post subject: |
|
|
Just read your post again Cedric. Your model shows a scaled 2660/#4. No idea if it makes a difference but mine is 2660/#6. But agree, I wish I had not listened to advice and rebuilt it with 2670 or similar compressor.
A question on intake temps. How much difference does the temps make on potential power? For my car to run at 45degC the ambient air would need to be 35degC. Appreciate the CGT tmic is suffering more with heat soak than my 951 ic. Just curious.
On another note. We dynod a stock but very highly maintained S2 on the same dyno and on same day as mine. The dyno also ran several 924's 2.0 that same day. They all mostly ran 10-20% less than the Porsche figures of 125bhp on the 'corrected' engine output. The stock S2 turbo ran 150bhp. Mine at 0.7bar ran 175bhp so some of the mods and ic must have added something. At 1.0bar it ran at 225bhp. So a 50% increase over the stock S2 set up. So a 0.3 bar increase is giving a 50bhp uplift. _________________ 1980 931 - forged pistons, Piper cam, K27/26 3257 6.10 hybrid turbo, 951 FMIC, custom intake, Mittelmotor dizzy & cam pulley, H&S exhaust, GAZ Gold, Fuch'ed, Quaife
Now www.924board.org/viewtopic.php?t=34690
Then www.924board.org/viewtopic.php?t=31252 |
|
Back to top |
|
|
Cedric
Joined: 27 Aug 2004 Posts: 2614 Location: Sweden
|
Posted: Mon Dec 19, 2016 4:27 pm Post subject: |
|
|
Carrera RSR wrote: | Just read your post again Cedric. Your model shows a scaled 2660/#4. No idea if it makes a difference but mine is 2660/#6. But agree, I wish I had not listened to advice and rebuilt it with 2670 or similar compressor.
A question on intake temps. How much difference does the temps make on potential power? For my car to run at 45degC the ambient air would need to be 35degC. Appreciate the CGT tmic is suffering more with heat soak than my 951 ic. Just curious.
On another note. We dynod a stock but very highly maintained S2 on the same dyno and on same day as mine. The dyno also ran several 924's 2.0 that same day. They all mostly ran 10-20% less than the Porsche figures of 125bhp on the 'corrected' engine output. The stock S2 turbo ran 150bhp. Mine at 0.7bar ran 175bhp so some of the mods and ic must have added something. At 1.0bar it ran at 225bhp. So a 50% increase over the stock S2 set up. So a 0.3 bar increase is giving a 50bhp uplift. |
Running a #6 instead would likely be
The 2660 might also be a completely different wheel design, just like the 2670 (not only a trim change compared to 2664), and that could make a difference, in my simulation the 2660 is only a scaled 2664 map.
But my quick analysis of that turbo is that i think Porsche choose it to show of how responsive and fuel efficient a turbo engine can be in a sports car. High end performance and tunability was not prioritized. Thats why they took such a big steg up in compressor size for the GTS, where they knew the customers will race it and tune it.
High boost at lower rpms is probably fine for the 2660 since you run in the sweet spot of the comp map, its just at higher flows/rpms it gets really tough. I cant see how that would save your engine, it will only be more prone to detonation since you must run with more closed WG=more backpressure=more burned gases in the combustion. 1.4bars must be hilarious
But if you would change to a 2664 billet, 2667 or 2670 its a pretty small job to do.
Regarding the charge cooler, I will ad such a simulation later, where i run different levels of cooling performance _________________ 1980 924 Turbo
www.instagram.com/garagecedric/ |
|
Back to top |
|
|
Cedric
Joined: 27 Aug 2004 Posts: 2614 Location: Sweden
|
Posted: Mon Dec 19, 2016 5:28 pm Post subject: |
|
|
It takes 30-40minutes to digitize a map, so im not sure i will have time. Really odd compressor, only used by a few liebherr genset/earth movers etc. Being a K27 3257 is obviously a diesel spec compressor, big wheel small trim, optimized for high pressure ratios rather than flow. The risk with going to bigger wheels are that you might put your turbine at worse efficiency instead when you run it at a different RPM, but not a deal breaker on a petrol engine.
Where did you get the map from, is it really that exact compressor? _________________ 1980 924 Turbo
www.instagram.com/garagecedric/ |
|
Back to top |
|
|
Cedric
Joined: 27 Aug 2004 Posts: 2614 Location: Sweden
|
Posted: Mon Dec 19, 2016 6:04 pm Post subject: |
|
|
gegge wrote: | The standard 931 head flow 143 cfm@9mm and 28” of H20 and drops back to 140 cfm with higher lift due to increased turbulence. It is almost a straight line from zero lift to 9 mm which pivot and levels out. The 924 head benefits from a larger radius at higher lift. A simply DIY in order to remove and smooth out edges raise the flow to more than 160cfm@10mm before it settles. This show just how important the area prior to the valve seat is. This is a 10-15% increase in flow AND output.
Stock cam is a fairly hot for a turbo application with 223 degrees@0,05” which is the most used measure of duration. Porsche specify it 228 degrees at 1mm instead. The 12mm lift is accomplished by taking advantage of the large diameter of the followers.
When I am trying to simulate the 924 engine in less advanced programs, I usually start with a natural aspired version. It is important to get the curvature of the torque through the rpm range right before you add boost. Since the head is choking above 5000 rpm, there is really not much gaining revving more. With the benefit of additional boost something happens above 5000 rpm. The head breathe a lot easier with support from the turbo. But still, at 5500-6000 even with a decent IC the song is over unless you port the head, swap the turbo for something bigger or go wild is the cam territory.
Do you have any numbers for the backpressure? Preferably with the logshaped infamous 931 exhaust manifold. If so, cam evaluation can be made with interesting outcomes. Guessing overlap and tight LSA isn´t going to be a problem...
With reference to the SAE paper in a previous topic: The 931 wastegate was actuated by either the boostpressure or the backpressure and it opens at 0,7 bar dumping the exhaustgases directly into the exhaustpipe. Therefore, the way I see it, the backpressure is lower than boost until 5200 rpm with the #6 turbine. Are you able to replicate that number? (The same situation can be seen in the early 911 diagram where P3 is significantly lower than P2 until 4300 rpm. Guessing that was the K26 related 3LDZ)
Would you mind running a sim with the #8 hotside? Maybe with a K27 3070? Please… |
I could maybe scale my flow curves, but a complete curve would be better. But it need to be converted to a more scientific setup, but that could be handled. I dont have huge confidence though in numbers from different tuning guys, best would be if someone close by had a completely std head and borrowed it to me, then I could do it at work. Then we could also maybe have a look at the tumble and swirl levels, which is one of the most important part if you want an knock resistant high boost engine..
Same with the cams, I scaled around some "default" lift curves, but a complete measurment on valve lift vs cam angle would be what I need.
I will rebuild my model so I can run i staticaly in at every 100rpm or so, then I could get much better measurments. The log manifold isnt a bad idea, its the way to go for a responsive engine. But they kind of screwed it up by adding a loooon pipe down to the waste gate which adds significant ammounts of volume.
What is interesting of course is the amounts of residual gas left in the cylinder, and that is also something you have to consider when optimizing cams. I could get some plots where we can look at backflow and how cylinders are interacting during the gas exhange phase. what I really need is to check my remote desktop setup so I can run this from home during the vacation _________________ 1980 924 Turbo
www.instagram.com/garagecedric/ |
|
Back to top |
|
|
WEASEL149
Joined: 19 Aug 2005 Posts: 595 Location: UK, Sheffield
|
Posted: Mon Dec 19, 2016 6:41 pm Post subject: |
|
|
Very interesting thread Cedric, thanks for sharing.
I'd be interested to know your thoughts on how a twinscroll manifold and turbine housing interact with the 931 engine.
I have a few different Holset HX35's; some with 12cm twinscroll and one with a rare 10cm twinscroll. _________________ 1979 UK 932 |
|
Back to top |
|
|
Cedric
Joined: 27 Aug 2004 Posts: 2614 Location: Sweden
|
Posted: Mon Dec 19, 2016 6:51 pm Post subject: |
|
|
WEASEL149 wrote: | Very interesting thread Cedric, thanks for sharing.
I'd be interested to know your thoughts on how a twinscroll manifold and turbine housing interact with the 931 engine.
I have a few different Holset HX35's; some with 12cm twinscroll and one with a rare 10cm twinscroll. |
Thats pretty easy to check.. Do you have some more detailed specification on that HX35? I happen to have access to hundreds of holset maps at work.
Would be interesting to see response, pumping work and combustion residuals with twinscroll. Probably much better..
But prio1 is to work more on the model itself. _________________ 1980 924 Turbo
www.instagram.com/garagecedric/ |
|
Back to top |
|
|
gegge
Joined: 27 Jul 2007 Posts: 1124 Location: Sweden
|
Posted: Mon Dec 19, 2016 10:59 pm Post subject: |
|
|
Cédric wrote: | Same with the cams, I scaled around some "default" lift curves, but a complete measurment on valve lift vs cam angle would be what I need.
I will rebuild my model so I can run i staticaly in at every 100rpm or so, then I could get much better measurments. The log manifold isnt a bad idea, its the way to go for a responsive engine. But they kind of screwed it up by adding a loooon pipe down to the waste gate which adds significant ammounts of volume.
What is interesting of course is the amounts of residual gas left in the cylinder, and that is also something you have to consider when optimizing cams. I could get some plots where we can look at backflow and how cylinders are interacting during the gas exhange phase. what I really need is to check my remote desktop setup so I can run this from home during the vacation |
I have the cam data you need, lift per degree
Agree that dynamic backpressure is a concern running overlap when using a logmanifold. The exhaustvalve of cyl 1 is opening prior to cyl2 closing, shooting the cumbustion back in. Same thing with cyl 3 which opens with a massive pressurewave just before cyl 1 close. The exhaustflow of cyl 2,3 and 4 are more or less directed the same way. The WG appendix is just weird. Look at old Saab turbo logmanifolds, that is good design!
Instead of using wide lobe angle separation for turbo angines, it would be of interest investigating the possibilty using a Miller setup. The Miller engine have longer duration on inlet compared to exhaust. That is one way to eliminate backflow from one cyl to another. Exhaust flow is very different from inletflow. Exhaust need good lowlift flow at early stage only with high pressuredelta over the valveseat and inlet at max pistonspeed i.e mid to highlift flow with almost no pressuredelta. _________________ Carl Fredrik Torkildsen
924 turbo -81 Carrera GT RESTOMOD
924 turbo -80 Dolomite De Luxe
924 -85 DP kit, BBS RS, M030 and tuned engine
924s -86 Black on black turbo with Fuchs |
|
Back to top |
|
|
Cedric
Joined: 27 Aug 2004 Posts: 2614 Location: Sweden
|
Posted: Tue Dec 20, 2016 4:40 pm Post subject: |
|
|
Last night i did some recalibration on my model, and some fine tuning on cam lift, head flows, combustion etc. Seems pretty spot on at the moment, but i still have to check some things around the combustion and volumetric eff, which might be of.
Here is my future setup and Steves setup both running 1bar boost, like i said before i dont have a 2660 map so I scaled it from the 2664. But it makes a hughe difference if i scale it from 2670 or 2664, since those two have a very different design. Also change the IC so outlet tempmax is around 35-36 degrees. I end up with 244hp and 330Nm with the 2670@1bar, seems like I en up pretty good if you look at GTS numbers.
Red and pink curves are reference simulation and my hub dyno data(x10%)
The 2660 overspeeds at the top end, not a good place to be, even though its possible to get the boost.
I ran about 672 different cams during the night, well se when i have the time to look at that.. _________________ 1980 924 Turbo
www.instagram.com/garagecedric/ |
|
Back to top |
|
|
morghen
Joined: 21 Jan 2005 Posts: 8884 Location: Romania
|
Posted: Tue Dec 20, 2016 7:20 pm Post subject: |
|
|
Cedric wrote: |
It takes 30-40minutes to digitize a map, so im not sure i will have time. Really odd compressor, only used by a few liebherr genset/earth movers etc. Being a K27 3257 is obviously a diesel spec compressor, big wheel small trim, optimized for high pressure ratios rather than flow. The risk with going to bigger wheels are that you might put your turbine at worse efficiency instead when you run it at a different RPM, but not a deal breaker on a petrol engine.
Where did you get the map from, is it really that exact compressor? |
It is an odd one indeed. Low flow fits with the 924 engine
Just like the 924 engine block that was designed to be a turbo diesel earth mover now mine has the "proper" turbocharger attached to it
I used the #6 turbine with that compressor and it spools up almost the same as the factory one.
The map is the only K27 map I found that is not a 911 K27. Butt dyno seems to match the behavior on the graph with real life experience though.
Surely not the map of that particular compressor.
It was a best effort thing from my side to graph that RPM line over that compressor map as my knowledge and understanding of such a complex system is limited but I did the effort knowing that its going to be an approximation at best.
I guess you need the actual map of the compressor to do an accurate simulation. I think I found the trim specs and the wheel part number and specs for it but its as far as I could go. _________________ https://www.the924.com |
|
Back to top |
|
|
Carrera RSR
Joined: 08 Jan 2010 Posts: 2309 Location: Somerset, UK
|
Posted: Tue Dec 20, 2016 8:10 pm Post subject: |
|
|
My last dyno sheet shows 2660#6 @ 1.0bar FMIC = 328Nm @ 3400rpm and 225bhp @ 5500rpm corrected engine, so does relect your models Cedric
One positive of my turbo/cam set up is early spool and a good kick in the butt early in the rev range. But agree it doesn't like going beyond 5500rpm. For road use I find this fine where you often cannot spank it fully through the gears at high rpm for longer periods where you would on track. But a turbo rebuild with a more efficient compressor needs to be considered for more bhp and more top end rpm. Will be interesting to see what compressor is the best all rounder. _________________ 1980 931 - forged pistons, Piper cam, K27/26 3257 6.10 hybrid turbo, 951 FMIC, custom intake, Mittelmotor dizzy & cam pulley, H&S exhaust, GAZ Gold, Fuch'ed, Quaife
Now www.924board.org/viewtopic.php?t=34690
Then www.924board.org/viewtopic.php?t=31252 |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|