gpsman1 — In that case, I accept the results of your careful test, and congratulate you on producing data that convinces me that the 2005 FEH's fuel efficiency can indeed benefit from driver intervention, to maximize the amount of pure-EV-mode operation that occurs. This tells me that there's a lot of scope for improvement in Ford's control software. I wonder whether the 2008 FEH is significantly better in this regard? It would be interesting to have an owner of the new FEH repeat your experiment in his/her vehicle.

In the TCH, gliding in 'N' removes the small amount of regenerative braking that normally occurs when coasting with your foot off the accelerator in 'D'. Neither MG1 nor MG2 operates while in 'N'. What you have done sounds to me a bit more like "pulse-and-glide" driving, which we know can significantly improve FE compared with normal "in-gear" driving — use the ICE to get up to speed; shut off the ICE, put the vehicle in 'N', and glide down to a lower speed; then repeat. This eliminates all the ICE losses during the glide phase. In your case, the ICE was already "off" while gliding in EV-mode, and I guess I could have achieved the same effect as putting the vehicle in 'N' by carefully feathering the accelerator while leaving it in 'D'. This being the case, what you did is equivalent to optimizing EV-mode while driving in 'D'.

Now, if only some TCH owners would rise to the occasion and either repeat my test or do a good one of their own!

Stan

 

Ford has stated in public documents, that part of their control strategy was to make the car "feel" as normal as possible, and to make the hybrid part of it to be as transparent to the driver as possible.

Ford has stated that part of their programing was to not have too many switches between EV and ICE, switches that may become an annoyance to the driver... wheather good for FE or not. The Ford car, by itself, will not go between EV and ICE in intervals shorter than 30 second of each, for example. A driver can, through driving methods, coax the car into EV with intervals of less than 30 seconds of ICE, for one example.

 

OK Stan, I'm going to make one last try to convince you that your initial hypothesis was indeed fallacy and not fact.

Look at your post #29, part 1.) You assume (correctly) that the BSFC @6.47kW is 250g/kWh.
Unfortunately, you also assume in part 3.) that the BSFC @9.84kW is also 250g/kWh. IT IS NOT! The BSFC @ 9.84 kW drops to 245, per the published BSFC contours for the Prius. A trivial error? NO! The additional 3.37kW is thus obtained at an incremental effciency of 34.5%! The break-even energy storage-retrieval point is thus 32.4/34.5=93.8%. Yes, you can't get there with the Prius HV pack, but start dropping your calculation speeds and you will see this change dramatically. In fact, just dropping your speed to 35MPH for the Prius results in about break-even FE by going EV. Perhaps it was just a coincidence, but the 6.47kW power level is almost exactly at the knee of the efficiency curve for the Prius. Once below this, the efficiency decreases with decreasing power become substantial (can't run the ICE less than 1000 RPM). Consider going from 5.00kW @ BSFC=266 to 8.37kW @ BSFC=246. Your incremental efficiency is then 37.5%, so the break even point is 30.5/37.5=81.3%. Doable with EV? P&G? Hills? All three are methods of energy storage-retrieval, and at lower speeds they will probably all give you better results than pure ICE cruising. What about forcing the ICE to produce more power through MG1/MG2 loading without storing? ALWAYS worse for MPG. More power from the ICE always means more gas, even if the incremental effciency approaches 100%.

For the FEH, with its less flat efficiency curves and larger HV battery and motor/generators pushing EV gives fantastic results. I would assume that for the Camry it would be less than the FEH, but perhaps better than the Prius.

 

Some more observations/corrections. On slide 53 of Miller's paper he mislocates the 6.47kW operating point for the ICE. It should be below the 8kW constant power curve, not above it. The best BSFC is thus no better than 260g/kWh. And for 9.84kW, the BSFC is certainly no more than 240g/kWh. So, per Stan's post #29 the steady state ICE cruising MPG is 66.6MP, not 69.3MPG. And the MPG while charging is found from: (9.84kW*240g/kWh)/740 g/L/64.4km/h=4.96 L/100km or about 46.7MPG. So, for EV-charge, repeat the resulting MPG is given by .8 miles of EV followed by 1.5 miles of charging @ 46.7 MPG for a total MPG of 71.6 MPG! So, you can steady-state ICE cruise @ 66.6MPG, or push EV to get 71.6MPG, a 7.5% advantage. So, this very palatable fact gives us a very convienent truth - EV rocks!

 

DesertDog — I apologize for the long delay in responding to your recent posts #123 and 124. I've been away overseas for part of this time, and have been rather busy since I got back. So I haven't had the opportunity to give them the needed careful consideration until now.

Yes, I agree with you that I was wrong not to have taken into account the increased thermodynamic efficiency of the ICE during the recharge phase (3) in my post #29. The higher loading of the engine due to the NiMH battery recharging would indeed have increased its efficiency significantly, as you point out. And, yes, some of Miller's "dots" are indeed misplaced in his efficiency diagrams, and mean that the ICE BSFC numbers that I used need to be revised. By the way, I never claimed that it was impossible to beat Toyota's ECUs in normal driving, only that it seems to me to be unlikely that you can generally do so without resorting to extreme measures, like pulse-and-glide. We have the benefit of knowing what's coming up on the road ahead (which the car doesn't know), but the car has the benefit of having full knowledge of the ICE's and MGs' parameters, and so can optimize its moment-by-moment strategy better than you can.

However, I must disagree with your use of an ICE power of only 9.84 kW during recharging. That number is the one I calculated for the ADDITIONAL (not total) ICE power required during recharging. The total ICE power needed during this phase is thus 6.47 + 9.84 = 16.31 kW. At this power, I read an even better BSFC figure of ~220 g/kWh from Miller's diagrams. Using your example speed of 40 miles per hour, what we need to compare is:

• fuel usage for steady-state driving at 40 miles per hour for a given distance (or time) at an ICE power of 6.47 kW and BSFC of 260 g/kWh;
• pure EV driving at 40 miles per hour for half this distance (or time), followed by fuel usage for ICE-powered steady-state driving at 40 miles per hour for the other half of the distance (or time), during which the NiMH battery is being recharged, at an ICE power of 16.31 kW and BSFC of 220 g/kWh.

I do not understand why you use 0.8 miles of pure EV followed by 1.5 miles of recharging. I assumed that the battery recharging would be done at the SAME rate as the discharging. This would imply the same distance (or time) for both phases (2) and (3) — each being half of the total. This being so, I compute (using 1 hour for the total time for the sake of simplicity, since it doesn't affect the answer):

• 1 h (40 miles) @ 6.47 kW => 6.47 kWh => 1682 g of fuel => 2.273 L of fuel => 3.53 L/100 km => 66.6 mpg [as also computed by you] for steady ICE-powered driving;
• 0.5 h (20 miles) of EV @ 5.85 kW => 2.925 kWh from the battery during the pure-EV phase (no fuel used), followed by 0.5 h (20 miles) of ICE power @ 16.31 kW => 8.155 kWh => 1794 g of fuel => 2.424 L of fuel => 7.53 L/100 km => 31.2 mpg during the recharge phase, giving 3.77 L/100 km => 62.5 mpg for the overall trip of 1 h (40 miles).

But, this is a REDUCTION in FE, not the increase that you claim! Am I missing something? Or was your calculation indeed incorrect? Granted, it's not a great reduction in FE, and the way the numbers fall will depend on the assumptions regarding mechanical-to-electrical-to-chemical efficiencies (and vice versa) in the MGs, inverter/converter electronics, and NiMH battery. Although these calculations are for a Prius, it's interesting to note that the TCH has significantly increased efficiencies, relative to the Prius, in almost all MG and electronics parameters (see the attached very interesting ORNL report — but they don’t measure the NiMH battery). This would seem to confirm my comment in post #58 that the results of my careful test (post #48) show lower losses in my TCH than I had initially guesstimated. I am not aware of any published BSFC data for the TCH's 2.4-L ICE. It's no doubt essentially similar to the Prius' 1.5-L ICE, but without hard data one cannot calculate any numbers for the TCH.

Stan

 

I forgot to attach the promised Oak Ridge National Labs (ORNL) file to the preceding post, and now I find that the system won't let me attach it thereto. It won't even let me attach it hereto either! It's probably the size of the file, even though it's well under the stated 10 Mbyte limit for PDF attachments. (I've had this problem frequently on this site.) So, the best I can do is point you to the URL where it can be downloaded:
http://www.osti.gov/bridge/purl.cove...922327-5ktfCi/

Stan

 

Hi Stan,

Only thing wrong with your calculations is that a 16.3kW constant power curve overlaid on the BSFC contours clearly intersects the 210 BSFC contour, not the 220, implying less than 210 is possible. But, using the 210 number gives your revised calculations a near wash for EV vs. steady state of 65.5 MPG vs. 66.6 MPG. Again, I think the 260 is probably too low for 6.47kW, but I'll accept it so as not to waffle at this point. But, how did you arrive at 9.84kW additional ICE load for normal charging? This seems high as it is about 2X what the FEH does for its slightly larger capacity pack. But at any rate, these calculations were for 40 miles per hour, which is obviously too high for EV, even on a level road. And, given the TCH's larger (and thus less flat efficiency) ICE and ostensibly more efficient MGs (from your reference), one would reasonably believe that it would tilt toward EV being slightly better than steady state cruising. Lower speeds should be much better for EV vis-a-vis steady state ICE cruise.

I would do the calculations for 35 miles per hour and 30 MPH, but I need to know whether to use the same 9.84kW charging load or some other number? Everything else can be found from Miller's paper, although the BSFC number below 8 kW are going to be really tough to extrapolate with enough accuracy.

Most of my EV is done @ 30-35 MPH. It is not as good as P&G, but it is definitely better than ICE cruising. I still believe that this is also true for the TCH.

EDIT:After looking at this more, I really have a hard time accepting the 9.84kW number. Accordng to your calculations, this gives an electrical gen-store-retrieve-use of 41.7%, much lower than I have ever seen posted. Most numbers are in the mid 50's. The numbers in Miller's paper would yield high 50's. Aso, this much power would require that MG2 provide most of the charging, which I don't think happens under normal circumstances. MG1 couldn't even absorb that much power at that RPM (~1150) if it wanted to.

Stan, please explain in detail how you got that number and what you think the MG1/2 split for that power level would be.

 

DesertDog — Your observations hit on an important point. In my calculations of the increased power required from the ICE during the battery-recharge phase, I was forced to make a somewhat incorrect simplifying assumption — namely, that the ICE was operating at the same efficiency during the 40-miles per hour recharge phase as it is during normal ICE-powered driving at 40 miles per hour. This will be wrong, since with the higher load at the same road speed the ICE's efficiency ought to be somewhat higher. I don't believe that it is easy (or even possible) to compute the "correct" figure from the information provided in Miller's paper. If you have any ideas about resolving this matter, I would be interested to hear them.

In the absence of a "correct" ICE efficiency figure for the recharge phase, I used the same figure as that given by Miller for normal ICE-powered driving at 40 miles per hour: 0.69466 (= Pem/Pe on Slide #52). I also assumed the same mechanical-to-electrical loss factor for the battery recharge phase as for the EV-powered (discharge) phase: 0.91*0.94 = 0.8554 (same Slide #). The overall discharge-recharge mechanical-electrical-mechanical loss factor is thus 0.8554^2 = 0.7317, and when combined with the assumed ICE power efficiency factor of 0.69466, yields an overall loss factor of 0.50829. Dividing the required MG2 power of 5 kW by this factor gives the needed additional ICE power of 9.84 kW during recharge. (My corresponding calculation was parsed somewhat differently in Post #29.) This gives the total ICE power number of 6.47+9.84 = 16.31 kW that I used in my calculation.

I agree with you that I indeed misread the labelling of Miller's BSFC curves, and that the number I should have used for the 40-miles per hour recharge phase at 16.31 kW is 210 g/kWh. Recalculating my numbers using this figure leads to the following revised final bullet in Post #125:

• 0.5 h (20 miles) of EV @ 5.85 kW => 2.925 kWh from the battery during the pure-EV phase (no fuel used), followed by 0.5 h (20 miles) of ICE power @ 16.31 kW => 8.155 kWh => 1712 g of fuel => 2.314 L of fuel => 7.19 L/100 km => 32.7 mpg during the recharge phase, giving 3.59 L/100 km => 65.5 mpg for the overall trip of 1 h (40 miles).

This is in agreement with your calculation, and is effectively the same as the 66.6 mpg number for normal ICE-powered driving. I don't believe that we have the data necessary to resolve this matter. The TCH's electrics being even more efficient than the Prius's would render the results a wash for the TCH as well. This is indeed what I found in my experiment (Post #48).

You could try doing calculations for a lower EV speed (Miller gives numbers for 30 miles per hour in Slide #50) and the same 40-miles per hour recharge speed as I used, but will I think be stymied by the same lack of data for the recharge phase as I found. I'd be most interested to learn what your calculations reveal, but yes, I would expect better results. I chose to use as nearly as possible the same speeds (~40 miles per hour) for both the pure-ICE, pure-EV, and recharge modes, so as to make the comparison as fair as possible (the fewer variables being altered the better). You still need to convince me that "EV rocks!" but it may not suck as much as I first thought!

Stan

 

DesertDog — A few additional clarifying comments:

• I think that my application of the ICE's power efficiency factor of 0.69466 to the electro-mechanical loss factor of 0.7314 may be misleading you into thinking that the latter is inordinately low, but I think that Miller's 73% number for this is probably reasonable. Even with the ICE's factor included, the overall factor is 0.50829 (i.e., ~51%), not the 41.7% you cite. I'm not sure where you get your number.
• The additional mechanical power that needs to be provided to MG1 by the ICE is 6.84 kW (in order to get the needed 5.85 kW = 6.84 kW x 0.8554 to the NiMH battery). This is well within MG1's capability, I believe. To deliver this additional mechanical power, the ICE needs to produce an additional 9.84 kW (= 6.84 kW/0.69466), which is well within its capability too.
• We don't actually know how the ICE's and MG1's rpm's are being apportioned (at the road speed of 40 mph) when the ICE has the additional load of recharging the battery. Without this knowledge I think we're stymied.

Another possibly relevant typo in Miller's paper is in Slide #50, where Pb should be -3.3 kW (and not -3.9 kW).

Stan

 

Stan, with my different, more primitive approach to instumentation as previously described (electrical triple tach), I can add a few observations to this discussion. Relative to battery charging, (other than regen) it is not possible to isolate the function to one MG. The two work in tandem to control ICE revolutions as a primary function, and any charging activity is a seconday activity that may involve either one, or both, depending on the drive train dynamics.

When a sustained battery charge (as opposed to a high speed maintainence charge that I will discuss later) starts, the rotation and speed of MG1 is changed by the ECU such that the ICE to axle drive ratio is very slightly adjusted to a lower ratio. At the same time, I propose that some generated current is taken from either MG2 or MG1 or both as required to maintain the ICE to axle drive ratio set by the ECU. I can see this on my triple tach as a change in MG1 and ICE revs. Although I cannot see a rev change in MG2 (or the drive line), the fact that the ratio between MG1 and the ICE has changed implies a torque change to the drive line, and thus to MG2. (Unfortunately, I cannot record the triple tach yet, so I cannot show you any meaningful data.) A dead giveaway to the fact that this condition can exist is the fact that the MG1 can start the ICE at a stoplight when MG2 is stationary and then control its RPMs. This is only possible if there are holding forces on MG2 (brakes) that prevent (by exerting a opposing torque) its rotation. In both of these situations the torque exerted by MG1 to govern the ICE is the control here, any power that MG1 can produce from that effort is a by product. The only difference in the two cases is that when in motion and MG2 is spinning the ECU has the additional potential to either extract additional energy from MG2, or put energy into it as Assist.

What I do not know is whether the ECU "thinks" in terms of RPM, Torque or Current in the control program. One interesting thing to watch is what the triple tach does when the battery is at a topped out SOC. The ECU "looks" for ways to bleed off some stored energy. Just as it manages charging for a low or decreased SOC, it manages "discharging" for a battery with an SOC above some target value.

In the early days, we speculated that the car had an inclinometer or other way to sense a change in road gradient. It does, but it is done by sensing a change in torque, RPM or a departure from a steady state current situation (don't know which) when the drive train changes speed. That is why the battery will intermittently show an assist state in high speed cruise. The ECU is constantly switching from an under SOC target state to an above SOC target state and back again in an effort to maintain a target SOC.

The other advantage of this SOC "toggle" is that it results in a more constant RPM on the ICE with smaller rev changes because within certain limits the assist is taken as an alternative to a ratio change when more power is needed by the road conditions, and the charge is taken in the opposite situation where the road conditions permit some of the energy being produced to be taked as generated current. But in both of these situations, the ICE seems to be attempting to maintain the ICE revs at a constant.

Consider this little scenario and quiz -- I suspect not many people on this forum know the answer, although you and a few others might. What happens on a long downgrade when the SOC tops out? We know that the TCH does not just give up its simulated drive train drag. Since it can no longer regen because there is nowhere for the current to go, how does it maintain the electrical load necessary to maintain the artificial drag?