• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Electric Car - Tesla Type S

ponderingturtle said:
In previous discussions of that movie it has appeared to be less than honnest.
Click to expand...
Agreed. It had more to do with a bad business decision and general short sightedness than an obvious political agenda.

Toyota bought the tech and created the Prius.
 
technoextreme said:
Im making snide comments because unlike you I've actually read more than a press release. Hell, I've actually run fuel cells before.
Click to expand...

So, you think you are the only one here who knows anything about them. I'll make sure I tell that to the guys who I know who work with them every day.

Unfortunately for you, there is real world proof that they work in cold climates. Are they at perfect efficiency? No, but neither is an ICE.
 
paximperium said:
Well it technically isn't a separate tank, the residual gas below the main piping and occasionally a smaller separate bulge below the main gas is usually called the reserve tank and it is more often filled with sediment so it isn't a good idea to use it all the time.
Click to expand...
If there isn't a separate switchable pickup for the fuel pump, it makes no difference. In any case, I think it's better to use the reserve every so often to clear any sediment so you *can* use if if you really need it sometime. Fuel systems have filters anyway.

Sorry if I gave the impression that it was a separate tank.
Click to expand...
Just the way *I* read it. ;)

I believe the floater is still the most common system in use but some newer cars have a computer than monitors it and reads it more accurately to the shape of the tank.
Click to expand...
GIGO. If the computer uses a float to get its reading it's just displaying spurious accuracy.

As an aside, the fuel computer in my 2003 Audi is rubbish compared to the one in the 1993 Saab I had before.
 
TrueSceptic said:
Thanks, but is it actually a problem in practice?
Click to expand...

The torque ripple in SR motors is a problem which has to be designed around by using clever control schemes. E.g.,

http://ieeexplore.ieee.org/Xplore/l.../iel5/41/20022/00925594.pdf&authDecision=-203

Summary:This paper addresses torque-ripple reduction in current-fed switched reluctance motors (SRMs). Ripple-free torque production in SRMs requires an accurate model that is often too complex for practical implementation. The algorithm proposed here combines the use of a simplified model with adaptation. Explicitly, it includes dynamic estimation of low harmonics of the combined unknown load torque and the ripple in the produced torque (due to model simplification), and adds appropriate terms to the commanded current to cancel these harmonics. Several simulations are presented first, suggesting that our method is effective for constant-speed reference commands, even when a very simple model is used in control design. Experimental results are included next to demonstrate that the algorithm performs well in reducing the torque ripple. Finally, limitations of the adaptive algorithm are explored and quantified
Click to expand...
 
A hydraulic accumulator doesn't use electricity. In it's simplest form, it is just a tank containing air or some other gas and hydraulic oil. When braking, a pump pushes hydraulic oil into the tank and pressurizes the gas. Power can than be drawn from it by using the gas pressure to push the oil out of the tank to drive a hydraulic motor.
 
CORed said:
A hydraulic accumulator doesn't use electricity. In it's simplest form, it is just a tank containing air or some other gas and hydraulic oil. When braking, a pump pushes hydraulic oil into the tank and pressurizes the gas. Power can than be drawn from it by using the gas pressure to push the oil out of the tank to drive a hydraulic motor.
Click to expand...
Isn't that the new tech that people are developing?
I read that it was more efficient than hybrid systems.
 
A hydraulic accumulator doesn't use electricity. In it's simplest form, it is just a tank containing air or some other gas and hydraulic oil. When braking, a pump pushes hydraulic oil into the tank and pressurizes the gas. Power can than be drawn from it by using the gas pressure to push the oil out of the tank to drive a hydraulic motor.
Click to expand...

When oil hits $200 watch for that being applied in trucks....

already is at UPS

hydraulic_hybrid_3_hi.jpg

http://blog.wired.com/cars/2008/10/ups-hydraulic-h.html
 
69dodge said:
I don't follow. Why can't floats be accurate?
Click to expand...
They can, but given that fuel gauges used to be, and often still are, crude analogue dials that show the amount to the nearest 1/4 tank, and don't even tell you how many gallons or litres you've got, what would have been the point?

But anyway, my complaint is about the accuracy a digital display implies, when the data source might have an error of +/- 5%.
 
LOL, haven't been to this forum in quite a while. And browsing the threads I remember why. Seems the AGW debate on this site is the same ol', same ol'. Not much point in engaging in that, but I'm happy to add some (hopefully useful) info on metrology, which I know a little about ;)

TrueSceptic said:
They can, but given that fuel gauges used to be, and often still are, crude analogue dials that show the amount to the nearest 1/4 tank, and don't even tell you how many gallons or litres you've got, what would have been the point?

But anyway, my complaint is about the accuracy a digital display implies, when the data source might have an error of +/- 5%.
Click to expand...

Seems to me people shun analogue systems inappropriately these days. No reason why an analogue system can't be very accurate. The dials are just moving-coil ammeters which are pretty sensitive, linear and accurate. The gauge is just a potentiometer - quite capable of accurate measurement well below 5%.

There are two main problems measuring fuel in a tank. One (already mentioned) is linearity due to the shape of the tank, which is important but for reasons perhaps not obvious to the lay person. The second is measurement errors due to tank slosh, angular orientation, and friction.

Friction could be an important factor because the force on the float is related to the difference between the float location and the level - the level has to move sufficiently for the buoyancy (or weight) of the float to overcome the stiction in the potentiometer. Although this isn't much of an issue, because tank slosh kind of does this automatically.

Slosh and angular orientation introduce most of the error, causing deviation from the correct value. This needs to be corrected for. The simplest method (used by older cars) is just to slug the output with a big filter. This produces an averaging effect, smoothing out these errors. The problem is, angular orientation can create a large error that holds for a long period of time - for example, driving up a long hill, or even just a heavily cambered road. This averages out (eventually...), but how long do you make the filter time constant? If you make it really long (say, five hours), you will remove much of the variability and get an accurate estimate of the mean, but if you're using fuel at a rate of 10 litres per hour, you could be up to 25 litres in error because you haven't accounted for the rate of use...

That's where modern systems come into play. The trip computer of a modern, fuel injected car knows with quite a high degree of accuracy the consumption rate (which is how it displays MPG). Remember, to pass a modern MoT, the car needs to track fuelling accurately to meet emissions requirements. If you know the consumption rate, you can form a state vector describing the tank and then use an alpha-beta filter to get a long-term measure of the tank state. The alpha-beta filter blends the rate of use of the fuel with the absolute measure in the tank, effectively averaging over a very long period of time but accounting for usage rate. This gives a very accurate indication of the level.

If your engineers are smart enough, they may try to go as far as implementing a Kalman filter, which could allow a more complex filter model, include additional measures etc. etc. (ETA: the alpha-beta filter can be shown to be a "special case" of the Kalman filter, the latter being more general)

Two notes about using an alpha-beta filter or Kalman filter. Firstly, your inputs must be linearised. Hence the need for linearisation correction mentioned above. Secondly, you need to recognise when the tank level has been changed outside of the model (i.e. you just filled up), and the filter state must be re-initialised. I noticed this with an old Renault 19 I had with a trip computer - I once put in just a couple of litres in (can't remember the reason, must have been a good one!) and the digital fuel gauge didn't notice. The software must have a threshold for change that triggers the filter to re-initialise, and I didn't put in enough for the threshold to pass, introducing a two litre "error" in the filter state. The filter will, of course, wash out this error over time, but it matches my view of how these things are probably done.

Note 1: I'm not an automotive engineer, but I know enough about metrology to explain how I would go about it. Of course, there are lots of different systems and different people would do it different ways; this would be my approach.

Note 2: While mentioning modern digital systems implementing alpha-beta filters, these filters can be implemented in analogue components as well. The only reason they didn't would be the cost / benefit - since you already have the processor, in the digital world the filter implementation comes for "free" in the software.
 
Last edited:
BenBurch said:
You can just log fuel injector applications and know to a high degree of accuracy how much fuel has been supplied to the engine.
Click to expand...

One still needs a mechanism to measure the tank level when refueled. The combination of fuel consumption rate - level state estimates as Coolsceptic outlined are effective. However, estimates are subject to discontinuities such as when re-fueled. Was the car on an incline at the gas station? DId the fuel change characteristics from ETOH additives? It would be interesting to see what model automotive engineers actually use to accomodate these factors.
 
TrueSceptic said:
They can, but given that fuel gauges used to be, and often still are, crude analogue dials that show the amount to the nearest 1/4 tank, and don't even tell you how many gallons or litres you've got, what would have been the point?

But anyway, my complaint is about the accuracy a digital display implies, when the data source might have an error of +/- 5%.
Click to expand...

I had a 1963 VW bug, which had a mechanical linkage (just a rotating wire in a sleeve), running from the float in the gas tank (in the front)to the guage. I think it was about as accurate (not very) as the guage in my current car.
 
69dodge said:
I don't follow. Why can't floats be accurate?
Click to expand...

They can be. All they require is a perfectly level tank. So you could calibrate them perfectly, as long as you had a perfectly level tank. Theoretically if you stick it in the middle, you have a good estimator. This doesn't work well. We had one car where you could gain a quarter of a tank by going downhill (and go down to empty heading uphill).

You could do an okay estimation with 4 floats, one in each corner, but there'd still be significant technical issues that plague all floats (bob, sticking, general lousy accuracy).

Best manner would be mounting the tank on a suspension and measuring the weight, though you could probably do many more exotic techniques.
 
Last edited:
You must log in or register to reply here.

Back
Top Bottom