Here is an idea I've been toying with... let me know what you all think and such.
A real time and/or trip vehicle efficiency meter/logger.
Instead of indirect fuel usage rates like MPG ... how about more direct vehicle efficiency numbers.
No small project.
Instead of doing the much more complicated method of looking at the real time efficiency of each part.
Here is an idea for what might be a simpler method of reasonably estimating net vehicle real time efficiency.
We compare the Energy In vs Energy Out.
To determine Energy Out:
Four parts:
Aerodynamics:
We have the OEM ( Cd * A ) figure to estimate the figure / shape ... to determine the real time aerodynamic power out ... we need the air speed , and the air density ... both of which we can measure real time with a variety of different methods , and sensors on the market.
Rolling Resistance:
I think this will take the most work ... but is still doable ... the Tire's Cr will have to be better quantified under different conditions ( temperature , weight, speed , etc ) ... and input what the vehicles weight is ... we could use a general manual input ... or if we got real fancy we could calculate it the real time 'effective weight' by measuring the distance to the ground near each wheel combined with a better quantified understanding of the spring rates of the suspension system ... but a manual input would be far easier.
Gravity:
For this one we need a slope / tilt sensor , kind of like an artificial horizon on air planes ... combined with the vehicle weight.
Acelleration:
Is the vehicle speeding up or slowing down?
Combine those four and we should be able to reasonably estimate the real time net vehicle energy output.
To get the energy input is much easier ... we need a flow rate ... gallons per minute, etc ... and watt meter on the IMA battery to add in the electrical energy ... we can use the known average gasoline energy content ( kwh ) per gallon , and we can even adjust it when we put in E10 , or E15 , or E85 ... etc.
energy in ... energy out ... we have all our pieces.
- - - - - -
Now why would we want to do this? ...
Well it gives a better idea of how the total system is performing ... ~100 MPG might be excellent ... or if cruising down a hill where no gasoline is needed ... it will be a lower efficiency... when going up a steep hill ... seeing ~50 MPG might not seem very good ... but if you know how efficient you are doing at going up the hill ... you can tell weather or not it actually is really good ... or or not.
This also helps to slightly remove some of the obscuring things like getting lower MPG with E10 or E15 or E85 vs straight gasoline ... his MPG might have gone down ... but with a meter like this he could know if his efficiency changed.
It also would allow people to better see if their technique is getting better or worse without depending on MPG alone , which varies significantly based on hot or cold air temperatures ... varying driving route conditions .... etc ... If one's technique improves and you are now driving more efficiently ... that is a good feedback to have ... but with just MPG , if that improvement happened as the weather got colder ... the MPG might go down ... thus giving incorrect feedback , suggesting to the driver what he did was not an improvement.
- - - - - - -
Not something to be done over night ... but I figured I would see what others thought about it this idea I have been toying around with recently.
Very interesting I like the idea if you can make it work.
Battery data can be obtained from BCM
Fuel data can be obtained from ECM (Probably) We really need to to decode the signals from the ECM to the display cluster to get access to the mpg data. Needs a fair bit of work and hacking into ecm to display harness.
We perhaps could use ODBII data not sure what is available at that port (Theres a list somewhere) but we could get it. If live fuel injection pulse data is available then with injector specs and rpm we could perhaps work out fuel being consumed on a second by second basis.
Your energy out devices will be interesting.
We really need a black box solution with all these fancy sensors in it where practicable and a simple display or output to a laptop with one single connection to the car. Probably the odbII port for power and ECM data.
There are spare pins at the odbii port so for instance BCM data could be fed through this to our device.
Very interesting I like the idea if you can make it work.
Thanks ... but even if I can get it all to work reasonably well ... and I am liking the idea more and more as I think about it ... It will take a while... rapid design and development speed are not among my personal strengths.
Quote:
Originally Posted by retepsnikrep
Battery data can be obtained from BCM
Fuel data can be obtained from ECM (Probably) We really need to to decode the signals from the ECM to the display cluster to get access to the mpg data. Needs a fair bit of work and hacking into ecm to display harness.
We perhaps could use ODBII data not sure what is available at that port (Theres a list somewhere) but we could get it. If live fuel injection pulse data is available then with injector specs and rpm we could perhaps work out fuel being consumed on a second by second basis.
Your energy out devices will be interesting.
We really need a black box solution with all these fancy sensors in it where practicable and a simple display or output to a laptop with one single connection to the car. Probably the odbII port for power and ECM data.
There are spare pins at the odbii port so for instance BCM data could be fed through this to our device.
I had originally thought of flow rate sensor on the fuel line ... but you are 100% correct ... between the instant MPH and MPG displays we already have the Gallons per Hour information.
I know both of those signals are on the OBDII line along with outside air temp ... scan gauge can display the MPH and the MPG ... and that would be a nice easy way to plug in.
Which takes care of the Gasoline input ... and the real time watt rate of the battery can be had from the plug on the BCM.
So in 2 plugs we could have all the input energy we would need to know... and would be fairly easy to install for that half of it.
- - - - -
The output energy would require some additions to the OEM sensors.
Real time relative wind speed could be had from small RC airplane sensors that already do this... air density I know there are small chips that do this as well ... I've seen small hand held electronic weather stations that have it built in ... I just haven't sourced such a chip yet... so with two small sensor chips , combined with the Cd and frontal area we already know ... we should be able to get a reasonable real time aerodynamic energy use.
Accelerating , or decelerating we should be able to do from the MPH data we already have from the OBDII port.
I will have to look into the artificial horizon and self balancing devices type instrumentation a bit more ... the gravity tilt is definitely doable from those types of devices ... but I do not yet know enough about them... but I have seen ~$50 toy RC motorcycles that self balance ... so ideally that might also be something someone makes on a chip somewhere.
Quantifying the Cr of the tires will take equipment I don't have ... So I'll have to see what can be found for that ... ideally they would just give us the data ... I strongly suspect the manufacture already has it anyway... if they won't give it up, that will take a bit of doing... the varying temperature part especially.
The last piece would be weight ... we could use a manual input ... which would be much simpler ... but I am still looking a bit more into the real time 'relative' weight idea ... measuring the suspension deflection at each of the four wheels under different amounts of weight I suspect should be doable reasonably well even without any real fancy devices ... then a distance sensor to detect that deflection real time as we drive would tell us the real time 'relative' weight ... this might be nice also because once it is set up it should account for the change of weight from refueling ~10 gallons ... or a passenger ... or hitting a bump etc.
Even if we get the 'real time' calculations of all these things down to display updates once every second or so ... I think that might be fine ... a simple numeric digital display , maybe with a data out for those who want to log it would be all that would be needed... of course faster updates would be nice ... but I think about once a second might be enough to be reasonably functional.
- - - - -
hmmm ... I am continuing to like the idea ... will have to see what I might be able to turn up.
I might be able to help with the effect of Crr at the stock Insight weight, and at different temperatures, and with the tires at 55 psi or so throughout the year.
I have been performing coast down testing on the same road, at the same time each morning on the way to work for about one year now.
Of the sixty or runs, about half have been thrown out due to variances that are not easily explained or the run was made at too slow of a speed (before I knew how to perform them), or there was too much wind at the time, but the remainder show some fairly concrete estimates regarding Crr versus ambient temperature.
Of course the Cd of the car is also estimated, but that is may not be relevant as my these runs were recorded with 75% of the underside smoothed out with fiberglass panels, which is much smoother than the stock Insight.
It will probably be several more weeks before I am ready to post results.
By the way, the math for this stuff can get very interesting as I use a typical hand-held GPS, and the speed readings vary enough to make analysis of the true road speed quite "challenging". There are two custom Visual Basic programs, one used to "average" GPS data, and a second program to find the "optimum" ratio between Cd and Crr for the measured dV/dT data set during the coast-down.
What I really need is a microcontroller that senses tire RPM's and converts that into elapsed time, then stores it on an SD card, which can then be down-loaded to the computer for analysis.
One of these days, I will figure out how to program the microcontroller and make such a system!!
What I really need is a microcontroller that senses tire RPM's and converts that into elapsed time, then stores it on an SD card, which can then be down-loaded to the computer for analysis.
One of these days, I will figure out how to program the microcontroller and make such a system!!
Jim.
I think that's a fairly easy project I might be able to help with, it would need a stick on magnet and sensor on one of the wheels. I'm don't understand how that benefits you though?
We already have a pulses per mile VSS output on the car and we could time the gaps between pulses and store those results? Speedometer.
My understanding is lacking on the subtleties of all this though. Bit dim.
I might be able to help with the effect of Crr at the stock Insight weight, and at different temperatures, and with the tires at 55 psi or so throughout the year.
I have been performing coast down testing on the same road, at the same time each morning on the way to work for about one year now.
I appreciate it ... any help is welcome.
While 'coast down' testing is useful in some applications ... How would you decouple the 4 separate forces/effects? ... #1> Gravity effect of slope ... #2> Aerodynamic drag ... #3> Rolling resistance...#4>acceleration &/or Deceleration ???
From the way I have seen most people go coasting tests ... there is rarely any effort at all to decouple these 4 separate forces... So I am interested in what you might suggest.
You might also want to try to see if this can be made to work for an EV. Where Battery Data comes from the data port on an Emeter. Accel/decel can be from the speedo signal. No fuel data needed, obviously. All the rest of the data comes from the sensors discussed above. In the setup program you would slect different parameters depending on gas or EV.
EVer's are forEVer looking for a data reader that will give them information about how the battery pack is performing and how much energy is left under 'these' conditions. So far there are really only a couple 'readers' that people are sort of content with, the Emeter, EVisionTM and Mark Brueggmann's meter Simple EV Charge Control.
This might also satisfy the discussion of how to convert gas mpg and elec watthours/mile to a common energy usage.
__________________
Rush
#1 - 2000 Silver #4965, not working now, 175k Miles, 61 LMPG - will probably turn into all Electric
#2 - 2000 Silver #4095, 185k Miles, 62 LMPG, OBDIIC&C by Peter, GCIM1 by Mike on bat pack www.TucsonEV.com
Tucson AZ
You might also want to try to see if this can be made to work for an EV.
A 100% pure EV , could be a little easier actually.
The hardest part of the project is the other stuff ... vehicle energy / power output part.
The same kind of 4 part energy use output is the same... wouldn't matter what the prime mover is ... could be a straight ICE, a rocket , whatever.
A Pure EV would only have the electric energy use input ... a hybrid like our Insights has that electric part plus a gasoline input part as well ... and the gasoline input part would have to be varied as the driver uses 100% gasoline ... E10 ... E15 ... E85 , etc.
Quote:
Originally Posted by Rush
In the setup program you would slect different parameters depending on gas or EV.
Any variation in major components would change the accuracy of the system.
If you change to a different tire ... you would need to update the Cr profile for the different tire ... or rolling resistance calculations will be less accurate.
If you change anything about aerodynamics ... you will need to update the Cd profile for the difference ... weather it be a different car , changes to the Gen 1 Insight's Cd Profile , etc...
If you change your suspension system ... the active 'relative' weight system would need to be updated with the profile of that suspension system.
etc ... etc...
Quote:
Originally Posted by Rush
So far there are really only a couple 'readers' that people are sort of content with,
I am not really intending on designing a BMS ... that is a worth while device and all that ... but as you wrote ... there already are many types of them ... I don't know of any vehicle efficiency meters ... so I am more interested in that direction.
Quote:
Originally Posted by Rush
This might also satisfy the discussion of how to convert gas mpg and elec watthours/mile to a common energy usage.
The problem with MPGe ... or converting between wh/mile and MPG ... etc ... is that there are different ways to do it ... people would have to agree to a standard method.
Each method is easily understood by itself ... but there are disagreements about which way to do it ... what to include or not include ... energy content is different than pollution is different than $ Cost .... etc...
Perhaps this project could be as simple as adding a little programming to the MPGuino? That device already has instantaneous fuel consumption and vehicle speed passing through it, and an arduino and a display. It wouldn't be able to account for hills, wind, or assist, but it would still be immensely useful to know where the sweet spot lies, and how large the lean burn advantage is.
Here's an effort at something very similar. I think Kubark42 went after too much at once and came up with nothing. For one thing, parameters governing how much energy the vehicle uses (m, CRR, CdA, etc) vary far too much and too unpredictably to make real-time determinations of their values without inputing them manually beforehand. I would prefer a less expensive system with one or zero accelerometers, no 3D GPS, and an acceptance of the fact that hills will screw up my data.
While 'coast down' testing is useful in some applications ... How would you decouple the 4 separate forces/effects? ... #1> Gravity effect of slope ... #2> Aerodynamic drag ... #3> Rolling resistance...#4>acceleration &/or Deceleration ???
From the way I have seen most people go coasting tests ... there is rarely any effort at all to decouple these 4 separate forces... So I am interested in what you might suggest.
IamIan,
Forces #1, #2 are resolved by post processing the coast down data, via a big spreadsheet.
If one coasts from lets say 60 mph down to 5 or so, and model the ideal coast down, then compare the ideal with actual, you can then determine what mix between forces #1 and #2 it took to create the measured coast down data.
It's an iterative process involving making guesses on the car Cd and Crr, crunching some numbers and then seeing how well the theoretical matches reality. After several iterations you get to a point where the model has the best match with actual testing. A custom VB program solves for these two variables in the blink of an eye.
The largest amount of time has been spent trying to average the high uncertainty from the GPS data. With the GPS, even an uncertainty of 0.5 mph causes a large variance in the dV data (deceleration) that is unacceptably high.
To do this more accurately would involve a Hall-effect sensor mounted to the tire rim and simply count tire rotations during the coast down and then place the output on an SD card for later retrieval.
The AutoGuide.com network consists of the largest network of enthusiast-owned enthusiast-operated automotive communities.
AutoGuide.com provides the latest car reviews, auto show coverage, new car prices, and automotive news. The AutoGuide network operates more than 100 automotive forums where our users consult peers for shopping information and advice, and share opinions as a community.
Efficiency Meter
