The BSFC chart thread (post 'em if you got 'em)

[Occasionally6]

Originally Posted by bikenfool

Care to elaborate? That link doesn't say anything about how to get BFSC data. I think measuring torque or hp accurately outside of a dyno would be tough.

Really easy, at least with a manual transmission. Measure the acceleration with an accelerometer and calculate the force required to achieve that acceleration using the mass of the car. Work back from the tires with the gear ratios to find torque at the flywheel. With the engine speed known you can then calculate the power being used.

With a torque convertor between the engine and transmission it's a bit harder to obtain engine power but it will still be possible to find the power from the transmission back.

Originally Posted by P-hack

tough, a little bit, for any hope of accuracy you would have to do some coastdown tests and have you and your vehicle weighed with a consistent fuel load (and tire psi) and warmup to determine CDA and CRR. Plus you want consistent atmospheric conditions.

HP to the wheels would be a function of drag at speed and change in velocity.

Fuel consumption would be monitored by an injector (or other accurate source).

Speed is the VSS signal (or rpm if using different gears).

Map is indicitive of load.

So you log fuel rate, speed, map. Accelerate at different map values using your foot and determine the hp required (and fuel consumed)between steps as you accelerate. Build the bsfc chart row by row by keeping map consistent during each run (As practical).

Standard testing, run in both directions and average, repeat. Flat test area with no wind is best.

It's so easy, you could gather a lot of data, analogous to logging fuel trims when engine tuning. Define cells of load vs rpm and fill them in as you drive. Most errors will average out.

Yes, you would want to ignore the data collected while the engine was at lower than normal operating temperature.

Because of the effect of road speed on aero and rolling drag maybe produce separate maps for each gear ratio. Enough coast down tests should get a reasonable measurement of these losses anyway.

[ever_green]

direct injection bsfc

[bikenfool]

Originally Posted by Occasionally6

Really easy, at least with a manual transmission. Measure the acceleration with an accelerometer and calculate the force required to achieve that acceleration using the mass of the car. Work back from the tires with the gear ratios to find torque at the flywheel. With the engine speed known you can then calculate the power being used.

With a torque convertor between the engine and transmission it's a bit harder to obtain engine power but it will still be possible to find the power from the transmission back.

.

You're neglecting the external loads, aero, rolling, road grade.

[Occasionally6]

Originally Posted by bikenfool

You're neglecting the external loads, aero, rolling, road grade.

Nope, they just get rolled in with the internal engine losses. We don't really care about the flywheel BSFC in a vehicle, just the most efficient engine load and speed.

If you do want to quantify those things measurement while coasting out of gear will do it.

[mwilliamshs]

Ford 300/4.9L Inline 6 cylinder, carburated, as used in commercial applications for generators, compressors, pumps, etc. Very similar to its automotive specs for HP and TQ and BSFC are considered accurate for automotive applications by those in the know.

http://www.industrialengines.ca/prod...D/4_9LSPEC.PDF

This engine is in my 1989 Ford E-150 and was used in Ford trucks and vans from 1966-1996.

[joemac]

Does anyone have recommendations on the best way to generate BSFC maps from engine dynamometer data for posting here? I have rpm, torque/BMEP, and BSFC over multiple points but generating and plotting iso-BSFC lines through the data isn't straight forward.

[P-hack]

I wouldn't mind having a look at the data and see if I can script up something. Can you zip it up and post it?

[joemac]

When scripting this, how are the iso-BSFC lines fitted through the data?

[P-hack]

connect the like bsfc dots, probably some approximating depending on the resolution of the data. i.e. group them in bands of 10 g/kwh, then find the center of that band, don't really know without some data to experiment with...

I would think that just coloring the bsfc points the same within each band with alternating contrasting colors across bands would probably be good enough for starters.

[bikenfool]

Matlab and the open source clone, Octave, have contour funtions that might be useful. You might need to do a curve fit to the data and generate more points to get a good plot.