I thought I'd posted this but never saw it show up. Regardless, I've been using a Graham miniscanner to look at specific fuel consumption: The data strongly suggests a fall-off in efficiency around 2,600 rpm and another around 3,200-3,300 rpm. It is part of an on going study: http://hiwaay.net/~bzwilson/prius/pri_power.html Bob Wilson
I have a run from a traffic light to 80 kmh (50mph) and have always used 2700rpm as the target to 8kmh (5mph) below the speed limit and then drop back and slowly go up to the limit. This keeps the AH behind me from getting mad at the "damn" hybrid in front of them. My life time is in the 45mpg US range for 3 years. Could I gain more by dropping back to 2600 I doubt it. Maybe I'm just not that concerened with the last inth of mileage possible. Summer mileage is in the low 50's and winter in the low 40's. But I'll commend you on your graphing using the miniscanner. I've logged some stuff with it and a laptop, mainly BSOC, Voltage and temps on a big hill, 3000' over about 5 1/2 miles. Interesting MG temps and just how far down the hill you get before the BSOC hit's 80% and your into friction braking. Then "B" and watch the BSOC drop. I think I deleted the file as I can't find it on either machine.
Hi Bob, I guess I am confused on your chart. The Red dots are the 91 to 96 % fuel consumperun per ICE watt, right ? Then would not the red dots indicate the worse specific fuel consumption? As there is more fuel consumed per watt on the relative 0 to 100 % scale. But your comment seem to inidcate that where the Red dots are is the best specific fuel consumption. Nice work getting all this data acquistions done!
<div class='quotetop'>QUOTE(donee @ Dec 30 2006, 06:45 AM) [snapback]368565[/snapback]</div> The 5% wide bands are relative to an arbitrary, highest, specific fuel consumption that was on the high end of the larger, efficiency curve. The littlest dots are the highest relative efficiency and the larger ones lower in 5% bands. There are 1,300 samples in the plotted data but by accident, I dropped the upper, 96-100% points. One problem with ODB data is resolving the time each observation occurred. The ECUs, the source of each data, have their own delays and values are captured with separate bus reads. To calculate ICE power, you need values from separate times and even with interpolation, we still have a problem with 'outliers.' But I'm pretty confident the data is accurate and several interesting observations fall out. Unlike a mechanical transmission, the ICE torque is current limited by MG1. This leads to the relatively linear, nearly horizontal torque line from 1,600 to 4,400 rpm. It does slope up a bit in part because at higher speeds, more current can be pushed through MG1. This is one key element that separates the Prius eCVT from a manual transmission. It is all but impossible to 'lug' our Prius engine although that is a technique many of the manual transmission, hypermilers advocate. They can, we can't. The other interesting effect is the fall off of specific fuel consumption in the upper rpm ranges. Investigating what happens at the higher rpm ranges could lead to significant performance improvement in high-speed operation. For example, a smarter cruise control that also has an rpm limit to keep the ICE in an efficient range (or foot control.) Also, it makes sense to look at the exhaust gas flow to see what is going on. Bob Wilson ps. I tried to find an area plot function in excel but there isn't one. The closest I could come up with was a scatter chart and manually change the sizes of the dots to reflect the third variable and ordering the data so the smallest circles are on top. pps. The raw data can be downloaded from a link in the first paragraph: http://hiwaay.net/~bzwilson/prius/pri_power.html
Hi Bob, Thanks for that interesting information. But, you still did not address my question. Your graph indicates that the highest RPM is the most efficient, but again you say that is where the least efficiency is. The Blue Dots (the biggest ones) are where the car uses 71-76 % fuel rate to generate a unit power. 71-76 percent is smaller than 91-96 % (the little red dots). I am still confused. Is there some mistake in the legend ? Should it be 71-76 % Power? Not Fuel ?
<div class='quotetop'>QUOTE(donee @ Dec 30 2006, 04:13 PM) [snapback]368707[/snapback]</div> Hummm, I see the problem. The various sized dots are relative to the peak ICE efficiency. As the ratio of power per unit of fuel falls off, the dots get larger. As would be expected, lower relative efficiency is more common so it is plotted with the largest blue dots. Next come the light blue, smaller dots; the yellow; orange and finally the smallest, red dots. By changing the size to correspond to the order, you can see the pattern. I've changed the legend to show "Fuel Eff. %" as short hand for "Specific Fuel Efficiency %." I may come up with better wording in a day or so. Sorry for the ambiguity. Bob Wilson
<div class='quotetop'>QUOTE(bwilson4web @ Dec 30 2006, 09:33 AM) [snapback]368592[/snapback]</div> How about getting excel to print a polynomial curve fit, and then integrating over ICE rpm?
<div class='quotetop'>QUOTE(tochatihu @ Dec 31 2006, 10:08 AM) [snapback]368926[/snapback]</div> Careful what you ask for . . . The only piece of the puzzle missing is to convert the ICE specific fuel consumption data to values at different speed power requirements. If the ICE efficiency drops off at 70 mph corresponding to the fall off already seen, then we have a basic energy model of Prius efficiency that has good agreement with observations. It becomes a model that leads us towards 'low hanging fruit' of NHW11 performance enhancements. Bob Wilson
I've updated the web referenced charts with HP data points. For fun, take a look at the HP requirement from 65-70-75 mph versus the specific fuel consumption drop off. Funny how that works out, eh? I used to think there was an aerodynamic problem at 65-70 mph and there may still be one. However, the fall-off in ICE specific fuel consumption really looks like a 'smoking gun.' Bob Wilson
