How does dyno sw calc AFR from air flow & fuel flow measurements?

[BradH]

How does dyno software calculate the air-fuel ratio from the air flow reading of the turbine on the carb and the fuel flow measurements?

Inputs?
- Air flow CFM
- Fuel flow Lb/Hour
- ???

Formula?

Thanks - Brad

[Stan Weiss]

cfm * 60 = cfh

cfh * air density = air flow lbs hr

Stan

[MadBill]

It bears reminding that AFR calculated by measured mass/volume is not the same as O2 sensor readouts based on Lambda values multiplied by the known or assumed stoichiometric ratio for the fuel in question. In addition to the stoich variable, the O2 sensors can only assess the combusted mixture; i.e. if there's some percentage of raw fuel in the exhaust due to poor atomization, over-scavenging, marginal ignition, etc. the equivalent unreacted oxygen mass will skew the readings in the lean direction relative to the meter number. If the meter and sensor data matches, great. If not, well-calibrated meter readings should presumably get the nod.

Perhaps David Redzus or ?? can expound on whether such an AFR differential can be a useful combustion metric...

[BradH]

It bears reminding that AFR calculated by measured mass/volume is not the same as O2 sensor readouts based on Lambda values multiplied by the known or assumed stoichiometric ratio for the fuel in question...

Understood.

I'm still trying to figure out how the following two AFRs were calc'd based upon the info provided. The air temp was 69* F, the barometric pressure was 29.68" Hg, the Vapor Pressure determined using a sling psychrometer was .35" Hg. Whatever I'm doing... it's wrong.

Fuel = 263.6 #/hr
Air = 807
AFR = 14.01

Fuel = 241.8 #/hr
Air = 689
AFR = 13.05

[Rick360]

Is the air reading an RPM from the air turbine? or does it have some known units?

Rick

[BradH]

It's a SF 901 dyno; the CFM #s are from a SuperFlow 4150 air turbine; not sure what version of WinDyn software the dyno uses.

[Stan Weiss]

It bears reminding that AFR calculated by measured mass/volume is not the same as O2 sensor readouts based on Lambda values multiplied by the known or assumed stoichiometric ratio for the fuel in question...

Understood.

I'm still trying to figure out how the following two AFRs were calc'd based upon the info provided. The air temp was 69* F, the barometric pressure was 29.68" Hg, the Vapor Pressure determined using a sling psychrometer was .35" Hg. Whatever I'm doing... it's wrong.

Fuel = 263.6 #/hr
Air = 807
AFR = 14.01

Fuel = 241.8 #/hr
Air = 689
AFR = 13.05

They both look like they just about work out using 0.0763 for air density.

► (807 * 60 * .0763) / 263.6 =14.01535
► (689 * 60 * .0763) / 241.8 = 13.04484

Stan

PS Are you able to get the SFD file?

[BradH]

The online calculator I was using to come up w/ an air density # generated something waaaaaayyy bigger than 0.0763. "Highly likely" a case of GIGO, or I was using an inappropriate calculator for this.

If have printouts of SFD files, but not necessarily every page of possible output for each pull. Not sure what info you'd think I'd want to look at...

[Stan Weiss]

Brad,
I had the confused symbol after the line with 0.0763 because based on the number you posted the air density should be lower than that more like low 0.074x.

I would have like to seen if the SFD file had this information

Stan

ab-brad-air-density.gif

[user-23911]

You haven't factored in fuel density.

Fuel volume is measured.
You then have to input temp and density to get fuel mass.

[David Redszus]

It bears reminding that AFR calculated by measured mass/volume is not the same as O2 sensor readouts based on Lambda values multiplied by the known or assumed stoichiometric ratio for the fuel in question...

Understood.

I'm still trying to figure out how the following two AFRs were calc'd based upon the info provided. The air temp was 69* F, the barometric pressure was 29.68" Hg, the Vapor Pressure determined using a sling psychrometer was .35" Hg. Whatever I'm doing... it's wrong.

Fuel = 263.6 #/hr
Air = 807
AFR = 14.01

Fuel = 241.8 #/hr
Air = 689
AFR = 13.05

They both look like they just about work out using 0.0763 for air density.

► (807 * 60 * .0763) / 263.6 =14.01535
► (689 * 60 * .0763) / 241.8 = 13.04484

Stan

PS Are you able to get the SFD file?

They both look like they are correct if you use 0.07384 lbs/ft^3 as a conversion for CFM to air mass.

► (807 * 60 * .07384) / 263.6 = 13.56
► (689 * 60 * .07384) / 241.8 = 12.62

The above is based on using the SAE standard for water vapor content (0.90%). The higher the H2O, the lower the air density.
Fuel density is not an issue if fuel use is given in mass units. If volume units are used, fuel must be converted to mass.

[user-23911]

It bears reminding that AFR calculated by measured mass/volume is not the same as O2 sensor readouts based on Lambda values multiplied by the known or assumed stoichiometric ratio for the fuel in question. In addition to the stoich variable, the O2 sensors can only assess the combusted mixture; i.e. if there's some percentage of raw fuel in the exhaust due to poor atomization, over-scavenging, marginal ignition, etc. the equivalent unreacted oxygen mass will skew the readings in the lean direction relative to the meter number

The WB ssensor works in 2 different ways depending on whether the mixture is leaner or richer than stoich.

It should be perfectly accurate at lambda 1, fairly accurate at the lean end of the scale but inaccurate at the rich end.

At the lean end it measures the difference in concentration between atmospheric o2 and exhaust o2.
At the rich end of the scale , it measures the concentration of combustibles in the exhaust. It doesn't and can't measure the o2 concentration when richer than lambda 1.
There isn't unburnt fuel as such but H2, CO and HC.
At lambda 1 there's close to zero o2 and close to zero combustibles.

What you need to know is that WB sensors are nowhere near as accurate as most people try to make out. In particular at the rich end due to no calibration at the rich end.
At a lambda of 1 the dyno AFR should match the WB AFR perfectly.That's the only time it will unless you've done a custom calibration of the WB from the dyno AFR.


Our engine dyno has orifice plate and electronics to measure air flow.
To get air mass (g/sec) you need temp, pressure, humidity.

For the fuel, it measures volume. You need to know the temp and specific gravity to get mass in g/sec from the volume flow.

Then you need to also know the calorific value of the fuel.
Depending on what part of the world you're in, it can vary with summer or winter blends from the pump.

From that you'll get the actual AFR for a lambda of 1.

WB sensors are programmed to indicate an AFR of 14.7? for lambda 1 even though different fuel will vary a bit in the real number.

[David Redszus]

They both look like they are correct if you use 0.07384 lbs/ft^3 as a conversion for CFM to air mass.

► (807 * 60 * .07384) / 263.6 = 13.56
► (689 * 60 * .07384) / 241.8 = 12.62
The above is based on using the SAE standard for water vapor content (0.90%). The higher the H2O, the lower the air density.
Fuel density is not an issue if fuel use is given in mass units. If volume units are used, fuel must be converted to mass.

The numbers above may look right but they are probably wrong.

The reason is that CFM can never be used to determine air/fuel ratio in a running engine.

Suppose a 5.0L V8 running at 6000 rpm. The displacement is five liters and the volume of air ingested is also five liters.
At 6000 rpm, the volume of air would be 529 CFM. The engine does not care about air volume; it only pays attention to air MASS.
Air is compressible (and stretchable as well), therefore the density of the ingested air will vary from cylinder to cylinder, cycle to cycle, and at every crank angle. For standard air density and a volumetric efficiency of 1.0, our engine would ingest 38.77 lbs/min.

If volumetric efficiency were only .90, or 1.10, it would ingest 34.89 lbs/min and 42.64 lbs/min respectively.
But the volume of air would remain unchanged.

While it is easy to assume that air density can be determined by a weather station using baro pressure, temperature and relative humidity, it is the local air conditions actually entering the combustion chamber that really matter. Inlet air can be heated underhood or cooled by fuel evaporation; local pressures will change density, and the air mass flow rate will change with each crank angle.

The engine only responds to the MASS of fuel available. But fuel systems meter fuel by volume and not mass. The volume of fuel required will be determined by the fuel stoich value, specific gravity (density), and required enrichment.

Not all fuel sent to the engine will fully evaporate, some will remain as carbon deposits, some will be blown out the exhaust.
No matter how much fuel we deliver, the engine can only burn a maximum determined by available air mass and fuel stoichiometry. We can burn less fuel but cannot burn more fuel.

The net result is that stable air/fuel ratios are only obtained in steady state, pre-mixed flames like a Bunsen burner, gas stove or welding torch.

A fast responding lambda sensor will indicate what it sees which may not be what we may think. Combustion misfires will indicate lean, oxygen content in the exhaust will vary with crank angle. The presence of other gases will cause a shift in readings of the sensor. Gas temperatures will also affect the accuracy of readings.

If Lambda readings are sampled fast enough, the data traces resemble unintelligble hash; mostly un-readable. Smoothing or slowing the sample rate provides a cleaner trace but at a loss of accuracy.

Real engines do not produce clean, smooth, single value air/fuel mixture data traces. All readings must be represented by a range of values.

Lab techniques employing the use of high speed 5 gas analyzers can be used to determine the true mixture ratio, and GC analysis can reveal emission composition components. But these tools are out of reach for most racers. The best bet currently available is to use wide band lambda sensors (one per each cylinder), sample at 500-1000 Hz, and log into a data logger.