Volumetric Efficiency Calculation??`
Discussion
Can I calculate my VE from my Injector Effective Pulse Width (InjEffPW) and Lambda/AFR?
I have been mulling it over but the result I get seems too high.
This is how I've been trying it. I have simplified the air density and fuel density for ease of maths.
Constants
Air density=1240g/m2 or 1.24g/litre or 0.00124g/cc
Fuel Density= 0.73kg/litre.
Injector flow 590cc/min or Injector mass flow = volume * density- 590*0.73= 430g/min= 430/60000 = 0.007166g/ms
Cylinder capacity 616.7cc
Measured Values
InjEffPW= 9.89 m/s or 0.00989sec
Lambda= 0.90 or 13.08:1
Fuel mass per cycle= InjEffPW * fuel density=
9.89*0.007166=0.0708g
Air mass per cycle= Fuel mass per cycle * AFR =
0.0708*13.08=0.926g
Air Volume per cycle= Air Mass / Air density
0.926/1.24= 0.746lt or 746cc
VE= air consumed/cyl capacity*100
=746/616.7*100 = 120.9%
Seems high for an NA engine, even in the relatively high state of tune I have.
Is my maths wrong?
I'm sure I'm about to get schooled....
I have been mulling it over but the result I get seems too high.
This is how I've been trying it. I have simplified the air density and fuel density for ease of maths.
Constants
Air density=1240g/m2 or 1.24g/litre or 0.00124g/cc
Fuel Density= 0.73kg/litre.
Injector flow 590cc/min or Injector mass flow = volume * density- 590*0.73= 430g/min= 430/60000 = 0.007166g/ms
Cylinder capacity 616.7cc
Measured Values
InjEffPW= 9.89 m/s or 0.00989sec
Lambda= 0.90 or 13.08:1
Fuel mass per cycle= InjEffPW * fuel density=
9.89*0.007166=0.0708g
Air mass per cycle= Fuel mass per cycle * AFR =
0.0708*13.08=0.926g
Air Volume per cycle= Air Mass / Air density
0.926/1.24= 0.746lt or 746cc
VE= air consumed/cyl capacity*100
=746/616.7*100 = 120.9%
Seems high for an NA engine, even in the relatively high state of tune I have.
Is my maths wrong?
I'm sure I'm about to get schooled....
Whilst your methodolgy seems sound, i suspect the accuracy of your constants leaves something to be desired!
Unless you have accurately measured things like injector characteristic (including dead time / low slope), your fuels stoichiometric AFR, and your actual in cylinder conditions at IVC, then there will be a large a range of error in your calculations.
For example,
If we assume your fuel has a stoichiometry at 14.3:1, your injectors have a 1.5ms deadtime, and only actually flow 550cc/min, that returns you a VEman of 94.5%
Unless you have accurately measured things like injector characteristic (including dead time / low slope), your fuels stoichiometric AFR, and your actual in cylinder conditions at IVC, then there will be a large a range of error in your calculations.
For example,
If we assume your fuel has a stoichiometry at 14.3:1, your injectors have a 1.5ms deadtime, and only actually flow 550cc/min, that returns you a VEman of 94.5%
I haven't checked through all your calculations but the rough guide to VE for normally aspirated 4 stroke engines I used to have on my website was to multiply the peak torque per litre by 1.40
An average stock 4v per cylinder road engine producing just over 70 ft lbs / litre would therefore be about 100% VE and a full monty race engine with 93 ft lbs / litre about 130%.
An average stock 4v per cylinder road engine producing just over 70 ft lbs / litre would therefore be about 100% VE and a full monty race engine with 93 ft lbs / litre about 130%.
Max_Torque said:
Whilst your methodolgy seems sound, i suspect the accuracy of your constants leaves something to be desired!
Unless you have accurately measured things like injector characteristic (including dead time / low slope), your fuels stoichiometric AFR, and your actual in cylinder conditions at IVC, then there will be a large a range of error in your calculations.
For example,
If we assume your fuel has a stoichiometry at 14.3:1, your injectors have a 1.5ms deadtime, and only actually flow 550cc/min, that returns you a VEman of 94.5%
I kind of guessed the inaccuracy may have been in things shuch as injector flow rates. AFR variation is not something I even considered.Unless you have accurately measured things like injector characteristic (including dead time / low slope), your fuels stoichiometric AFR, and your actual in cylinder conditions at IVC, then there will be a large a range of error in your calculations.
For example,
If we assume your fuel has a stoichiometry at 14.3:1, your injectors have a 1.5ms deadtime, and only actually flow 550cc/min, that returns you a VEman of 94.5%
The injectors are allegedly rated a 52lb/hr 3bar (Bosch 0280 158 117) but this is the calibration table.
Thanks for the input.
Steve
Pumaracing said:
I haven't checked through all your calculations but the rough guide to VE for normally aspirated 4 stroke engines I used to have on my website was to multiply the peak torque per litre by 1.40
An average stock 4v per cylinder road engine producing just over 70 ft lbs / litre would therefore be about 100% VE and a full monty race engine with 93 ft lbs / litre about 130%.
82lb/ft litreAn average stock 4v per cylinder road engine producing just over 70 ft lbs / litre would therefore be about 100% VE and a full monty race engine with 93 ft lbs / litre about 130%.
82*1.4= 114.8
Not a million miles away from my maths. Certainly within the variables of stoich and fuel flow.
Thanks
Steve
Things worth noting from that injector data:
Min pulse width is 711 uS, the injector injects zero fuel during this time
Battery voltage correction @ 12v is another 1041uS
Low slope (ALOSL) extends to 907uS
As such, with a 9.8ms pulse, the first 1.7ms (1752uS = 711+1041) will do nothing, the next 0.9ms, will start to deliver a small amount of fuel on the low slope, and finally at 2.6mS, the injector will be fully open and be delivering it's rated fuel flow, assuming the delta pressure is as you think it is (inc dynamic fuel rail pressure pulsing etc!)
Min pulse width is 711 uS, the injector injects zero fuel during this time
Battery voltage correction @ 12v is another 1041uS
Low slope (ALOSL) extends to 907uS
As such, with a 9.8ms pulse, the first 1.7ms (1752uS = 711+1041) will do nothing, the next 0.9ms, will start to deliver a small amount of fuel on the low slope, and finally at 2.6mS, the injector will be fully open and be delivering it's rated fuel flow, assuming the delta pressure is as you think it is (inc dynamic fuel rail pressure pulsing etc!)
Min PW won't be zero flow - i.e. not the same thing as injector deadtime. It's usually where the standard deviation/CoV of fuel delivered is > 5% or where linearity deviates by 5% or some similar definition. Hopefully this is a fair bit below your idle fuel requirements.
Edited by AER on Monday 9th September 03:17
Max_Torque said:
Things worth noting from that injector data:
Min pulse width is 711 uS, the injector injects zero fuel during this time
Battery voltage correction @ 12v is another 1041uS
Low slope (ALOSL) extends to 907uS
As such, with a 9.8ms pulse, the first 1.7ms (1752uS = 711+1041) will do nothing, the next 0.9ms, will start to deliver a small amount of fuel on the low slope, and finally at 2.6mS, the injector will be fully open and be delivering it's rated fuel flow, assuming the delta pressure is as you think it is (inc dynamic fuel rail pressure pulsing etc!)
From the Motec data logging;Min pulse width is 711 uS, the injector injects zero fuel during this time
Battery voltage correction @ 12v is another 1041uS
Low slope (ALOSL) extends to 907uS
As such, with a 9.8ms pulse, the first 1.7ms (1752uS = 711+1041) will do nothing, the next 0.9ms, will start to deliver a small amount of fuel on the low slope, and finally at 2.6mS, the injector will be fully open and be delivering it's rated fuel flow, assuming the delta pressure is as you think it is (inc dynamic fuel rail pressure pulsing etc!)
Base PW- what is in the main fuel table (10.3)
Actual PW- Base + dead time + compensations (MAP, IAT, ECT ect applied as a percentage)
Fuel Eff PW is the base value + compensations - dead time. (1.18)
I used Eff PW in my calcs as this negated dead time.
AER said:
Min PW won't be zero flow - i.e. not the same thing as injector deadtime. It's usually where the standard deviation/CoV of fuel delivered is > 5% or where linearity deviates by 5% or some similar definition. Hopefully this is a fair bit below your idle fuel requirements.
That is how I understand the MIN PW, my idle PW is 1.8ms.That calibration table seems pretty complicated, I can't understand why they don't just present it in graph form?
Thanks for the input.
In this case, the fuel actually delivered during the injector non linear region is a tiny proportion of the total fuel injected (As AER says, important at idle however!). It looks like you injPW doesn't include the dead time, so that leaves us with quite a big error still.
As i said above, you should confirm what the stoichiometric ratio is for the fuel you are using (it can vary significantly).
That leaves the basic injector flow as the biggest error potential. Have you actually measured fuel rail pressure at the test condition? If it was set at idle (or engine off) the slope of the spring curve in the fuel rail pressure regulator can result in a large decrease at high power settings when the engine is actually using a large volume of fuel. Depending on the regulator design (some are aftermarket ones are particularly bad in this respect, using short stiff springs, not long soft ones) you could easily be seeing a static 10% reduction in FRPressure at WOT peak power, before you consider dynamic effects. (i assume your ems is sequential? Banked systems where all the injectors fire at once are terrible for fuel rail dynamics....)
As i said above, you should confirm what the stoichiometric ratio is for the fuel you are using (it can vary significantly).
That leaves the basic injector flow as the biggest error potential. Have you actually measured fuel rail pressure at the test condition? If it was set at idle (or engine off) the slope of the spring curve in the fuel rail pressure regulator can result in a large decrease at high power settings when the engine is actually using a large volume of fuel. Depending on the regulator design (some are aftermarket ones are particularly bad in this respect, using short stiff springs, not long soft ones) you could easily be seeing a static 10% reduction in FRPressure at WOT peak power, before you consider dynamic effects. (i assume your ems is sequential? Banked systems where all the injectors fire at once are terrible for fuel rail dynamics....)
The Actusal PW is bas plus compensations plus dead time, i.e what the injector driver delivers.
The Eff PW is the base plus compensations (%)
The only variable I can see is the low slope to high slope transition. The injector flows more on the low slope than the high, but that is only for the first 0.711ms, after that point 0.711 to 9.89 is at the high slope.
The injection is sequential.
The batch fire of the original set up is probably why there was a damper in the fuel line.
I have not measured the fuel pressure, just assuming the FPR is correct OEM (Bosch) 3.8bar.
I think the upshot is in theory it is possible, but there are maybe too many difficult to measure variables to be 100% accurate.
I did not realise that fuel has such a variable stoich. I thought petrol was 14.7 and that is it. How is this tested?
The Eff PW is the base plus compensations (%)
The only variable I can see is the low slope to high slope transition. The injector flows more on the low slope than the high, but that is only for the first 0.711ms, after that point 0.711 to 9.89 is at the high slope.
The injection is sequential.
The batch fire of the original set up is probably why there was a damper in the fuel line.
I have not measured the fuel pressure, just assuming the FPR is correct OEM (Bosch) 3.8bar.
I think the upshot is in theory it is possible, but there are maybe too many difficult to measure variables to be 100% accurate.
I did not realise that fuel has such a variable stoich. I thought petrol was 14.7 and that is it. How is this tested?
stevesingo said:
I did not realise that fuel has such a variable stoich. I thought petrol was 14.7 and that is it. How is this tested?
It's tested (calculated?) by measuring how much C, H and O are in the fuel. "Nominal" test gasoline is 86.6% C by mass and the rest H, but with oxygenated fuels (i.e. alcohol) this is all over the shop these days. Best plan would be to characterise your fuel system more accurately. Pretty easy with a motec that has an injector "test mode" built in! i.e. remove fuel rail, stick injectors into a jamjar on some scales, run x pulse width for y period of time, measure injected quantity.
When i was characterising my returnless system rail pressure control i fab'd up a "return" manifold for my fuel rail and injectors, that sent the fuel either back to the tank, or into a separate catch tank. That way i could measure those sort of things without too much danger of going bang accidentally!
(On a engine test bed we would use a AVL Fuel weigher and a horriba accurate exhaust gas analyser to provide the AFR and fuel mass data.)
When i was characterising my returnless system rail pressure control i fab'd up a "return" manifold for my fuel rail and injectors, that sent the fuel either back to the tank, or into a separate catch tank. That way i could measure those sort of things without too much danger of going bang accidentally!
(On a engine test bed we would use a AVL Fuel weigher and a horriba accurate exhaust gas analyser to provide the AFR and fuel mass data.)
stevesingo said:
Sure, if you don't mind. It is not critical I know, just interested.
Steve
Drop me a PM, I'll let you know what's needed. You will have to be quite serious, there is quite a bit, nothing that can't be found out though.Steve
Edited by Evoluzione on Tuesday 10th September 09:01
Maybe this might be of interest. Can't calculate VE directly, but you can get an idea of what it might be.
http://blackartdynamics.com/Engine/EngineThermodyn...
http://blackartdynamics.com/Engine/EngineThermodyn...
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff