One additional piece of information are the numbers of Prius on Ebay with high mileage in excess of 150,000 miles in operating condition. What seems to be missing are "Prius with dead battery" postings on Ebay. They do have salvage vehicles but these are from accidents. Of course some may still be running using batteries from salvage vehicles.

Battery lifetime is an important consideration but we have no open source data that tells us what that value is. The quote that started this thread does not include the lab report that would let us know the test protocol.

This begs the question of how to design an experiment to replicate or at least provide some open metrics for expected battery life. Fortunately, I have a spare battery pack and four, cell assemblies from another vehicle. I've also got the technology needed although the cold cell may be a challenge. A proper life-cycle test needs to include not only the loads but temperature cycles. Ideally, it would include samples across different temperature profiles.

As for posting practices, my posts bring facts and data to the table and treat the rest as noise. Some of the best posts I've read have been from others who "did the experiment" or brought new facts and data to the table beyond these 13 words:

"Hybrid vehicle battery expected life is 150,000 miles based on laboratory bench testing."

One of my speculations is that it may be possible to substitute an entirely different battery solution for our current, NiMH batteries. Right now, I'm thinking a relatively small, ultra capacitor system could provide the surge capability needed allowing a smaller, lighter battery assembly to work instead. The key is understanding the energy flows.

GOOD LUCK!
Bob Wilson

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

I tend to somewhat agree with Bob's position (if I understand it correctly that is). It is almost unheard of for folks who have been trained as engineers to regard any marketing and product brochures as as little more than noise. I have to admit that I am guilty of this too, and as a result I am often very dismissive of information that non-technical people write. Sorry, been burned a few times already.

Besides, it is common to find at the end of a brochure or promotional materials disclaimers regarding accuracy of the statements and assertions contained in the material. Check the brochure in question and you are likely to find such a disclaimer.

Engineers don't produce such disclaimers and neither do technical papers and empirical data... and that, is what we should really look for as authoritative information.
But, as it has often been said, when lacking the technical information the best path to the truth is still the experimental path and raw field data. In my opinion, all other competing options (especially brochures and promotional content) are best regarded as "noise" and hardly authoritative.


Cheers;

MSantos

bwilson4web — I don't understand the green curve in your post #18. You label it as dW/dt. Do you mean watts/second? Is it the time derivative of one of the other curves? What physical significance does this have?

Stan

The problem is the absolute kinetic energy of the vehicle, 0.5*m*(v**2), is at any one time, a very large number, much larger than the engine and battery watts. However, the change in the kinetic energy does a much better job of showing how engine and battery energy change the kinetic energy. The only thing missing was altitude to show the change in potential energy.

For example, if I repeat the experiment on a flat street using just the battery (aka., the engine is off), the change in kinetic energy should be almost equal in size but opposite in sign to the battery energy expended to reach a given speed. Of course there are drag effects that need to be accounted for but that is the basic physics. You might notice that my kinetic energy (KE) line should actually use the opposite sign but missing is the potential energy (PE) changes due to the 7 m. high overpasses.

Energy accounting for a system often allows us to find subtle effects not previously detected. But it does take attention to detail when running the experiment. I'd be happy to go into more detail or try different ways of explaining what is going on. Feel free to ask or send a PM.

Bob Wilson

ps. This is the chart:

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

Bob, I get it, or at least I did until I read your box "engine braking". The ICE, as a load seems low. Are you referring to ICE rotation in fuel cut, or the ECU simulated braking load, or the "B" ECU simulated load?

It is hard to tell without PE or at least a sum(PE,KE) plot. ICE RPM overlay would also make it easier to read becasue one could more easily determie ICE state.

It is the ignorant among us that will eventually destroy us all.

That is exactly what is going on. On the backside of the overpass, gravity gets the car going faster than the cruise control set speed. In an ordinary car, no problem, it just closes the throttle, the speed creeps up and then drag eventually brings the car back to speed. But to our 'anal-retentive', hybrid cruise control, that isn't good enough. Not only does it shut the throttle and turn off the injectors, it is quiet happy to use some of the MG2 regenerative energy to reverse the torque on MG1 . . . to spin the engine!

Often when I travel that route over a number of overpasses, traffic permitting, my protocol is:

1. approach overpass a little on the fast side (like the rest of the traffic)
2. shift to "N" just as the ICE starts to pull and coast up and depending upon traffic, coast down the reverse slope
3. when speed approaches the flat cruise value, shift to "D" and resume cruise control

Now we're not talking some earth shattering MPG improvement but rather it keeps my engineering efficiency side of the brain from not dwelling on the limitations of the overly constrictive cruise control.

It might work better if I share some of the old data. But recently I've standardized my Graham scanner monitoring and have a data reformatter that makes it much easier to analyze. I'm also working on integrating GPS data with the Graham data (see below.) Within the limits of GPS, this will give location, altitude, heading and a secondary velocity vector (the primary comes from MG2 rpm that is fixed to the tire rotation.)

My thinking is combined Graham and GPS data may give enough information to handle correcting some of the GPS errors and a chance at measuring total energy state. There will be problems with 'outliers' due to the coarse sampling nature of the Graham scanner but I think I've found ways to at least detect these problem data points. I've also found that seven or more satellites significantly improves the GPS data quality.

Give me a week or so to work out the details and I'll be able to replicate this graph over the same route with a lot more detail. Best of all, Google takes longitude and latitude coordinates so folks can easily see the route. I haven't tested Topozone but it too may support global coordinates so we can look at a topographical map data.

In the past, I've felt bad about posting charts and graphs with incomplete data until I realized waiting for perfection could take a long, long time. I felt it was better to have 'something' and some insights than to be blind until we could get perfect data. I've just tried to make sure the limitations are clear and unambiguous.

Bob Wilson

Prius Monitoring with Graham Scanner and GPS

Using a Graham Davies Mini-scanner, I have settled on polling and recording the following data items:

1. MG2 rpm (also provides MG2 torque Nm)
2. MG1 rpm (also provides MG1 torque Nm)
3. throttle air flow grams/sec.
4. MG2 torque Nm (also provides MG2 rpm)
5. MG1 torque Nm (also provides MG1 rpm)
6. battery current amps

MG2 has a fixed ratio with the vehicle velocity and with the mass, kinetic energy. MG1 and MG2 rpm gives ICE rpm. MG1 torque has a fixed ratio to ICE torque. With rpm and torque, we can calculate the power input or output of the ICE, MG1 and MG2. Using a fixed or a simple current-to-voltage assumption, we can calculate battery power input or output. Finally, the throttle air flow gives the fuel flow rate and actual consumption is indicated by ICE torque. So in total I have:

• vehicle kinetic energy
• ICE power
• MG1 power
• MG2 power
• battery power
• gasoline consumption

With the GPS data, I get a rough longitude, latitude, heading, velocity and altitude. However, GPS data is somewhat imprecise. But it may be possible to use the vehicle velocity to calculate the correction needed on the GPS data. This may lead to sufficient precision to give potential energy. Certainly, the position information should be enough to show routes and with a GIS map, a true route. If that GIS map includes topographical data, then the altitude and potential energy will be known with very useful tolerances.

If not, I'll have to get an electronic barometer and mount it so it is as immune to air speed changes as possible. This may be the ultimate, potential energy map. I've also considered getting a computer readable, magnetic compass to do 'dead reconning' navigation . . .

A little over the top perhaps?

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

bwilson4web — But, assuming that the PE stays constant (e.g., constant altitude) for simplicity, then:

change in KE = work done = integral of W with respect to t

and not the derivative of W with respect to t. So now you have me really puzzled!

Stan

At every sample point, I have the vehicle velocity, which changes, and the mass with myself is relatively fixed. So using this formula, I calculate the KE:

KE = 0.5 * M * (V**2)

M ~= 1,360 kg

50 miles per hour ~= 22.352 m/sec
51 miles per hour ~= 22.799 m/sec
52 miles per hour ~= 23.246 m/sec

KE@50 miles per hour = 339,736.1 Joules (353460.2-339736.1=13724.1)
KE@51 miles per hour = 353,460.2 Joules (367456.0-353460.2=13995.8)
KE@52 miles per hour = 367,456.0 Joules (367456.0-339736.1=27719.9)

Assume each sample interval is 1 second, which is pretty close to the actual values:

T1@50mph = 339,736.1 (J), dW/dt = ?
T2@50mph = 339,736.1 (J), dW/dt = 0
T3@51mph = 353,460.2 (J), dW/dt = 13,724.1 (dW)
T4@51mph = 353,460.2 (J), dW/dt = 0 (W)
T5@52mph = 367,456.0 (J), dW/dt = 13,995.8 (dW)
T6@52mph = 367,456.0 (J), dW/dt = 0 (W)
T7@50mph = 339,736.1 (J), dW/dt = -27,719.9 (dW)
T8@50mph = 339,736.1 (J), dW/dt = 0 (W)

Kinetic energy is normally in Joules, which is Watt-seconds. However, in this case, we want to find the change in Joules per sample interval, which by happy accident is about 1 second each. This becomes the energy added or subtracted to effect the speed changes.

I'm already calculating the engine and battery in watts and it was simple enough to show the KE as watts. A happy accident, the one second sampling interval applies regardless of engine, battery or vehicle kinetic energy. But if I do everything in Joules and adjusting for the precise interval, we still get similar ratios. They are all using the same time base.

For those confused:

Joules = watt seconds (watts over a one second time interval)

Normally we'd integrate the watts over time with the interval being 0-1 seconds. However, our sampling interval, by happy accident, is about 1.1 seconds (if I remember correctly.)

BTW, I had good results last night analyzing the GPS NEMA data. I actually plotted my route in excel and plotted altitude change versus trip time. The next step is to translate longitude and latitude changes into meters and calculate the velocity. Then I'll normalize the Graham and GPS time and get a chance to look a the total energy state of my Prius over a trip from Huntsville AL to Knoxville TN.

Bob Wilson

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

Ah, I overlooked the Cruise Control impact downhill. It is not only fuel cut drag in coast, but additionally induced drag courtesy of MG1 as demanded by the CC.

For the NEXGEN hybrid, Toyota needs to look into more smarts for the CC. i.e. if the charge state is OK, why send a charge to the battery on a gradual decline. Better to convert the small PE available to KE to reduce demanded power. For steeper grades, go ahead and charge because if you don't then you will need to use the power anyway to drive MG1 to slow down as demanded by the CC. [vicious circle!] Also maybe a acceptable range for cruise that can be tightned or loosened by a control. That would permit the CC to accept slight slowing in small inclines and slight speed increases on short declines without the madness of ICE [STOP:GO:STOP:GO] that we see now. It should be a simple enough design problem.

It is the ignorant among us that will eventually destroy us all.

bwilson4web — If I now understand you correctly, you are saying that what you denote by 'W' is not power (say in watts) but energy (say in watt.seconds = joules). [Yes, I now recall that this has come up before, and you corrected your data to energy units by multiplying/dividing by the time interval.] But then what you denote by the derivative dW/dt is actually power and not the time derivative of power! Is the 'W' you use to denote the 'Battery' and 'ICE' traces on your graph then also energy, or are they actually power? 'W' is the standard symbol for power, not energy. This is most confusing.

If 'W' is power, then 'dW/dt' cannot be KE, as implied by your graph label: 'KE dW/dt'. If 'W' is energy, then 'dW/dt' is power, and again it cannot equal KE. Perhaps I could understand your graphs if you'd just tell me the true units for each of the three traces: Battery W; ICE W; KE dW/dt.

Since you labelled the variables 'W', I assumed that the vertical axis was labelled in power in watts. It now appears that I was wrong in this assumption. You say: "I'm already calculating the engine and battery in watts and it was simple enough to show the KE as watts." But KE cannot be measured in watts! I think your 'W' and 'dW/dt' must be energy changes in joules, and that none of them is power in watts. All the graphs use the same vertical scale. So then, do all the graphs represent just changes in energy, i.e., power expended over a ~1.1 second interval: power x time = energy? If that is the case, the vertical scale is then approximately in joules (watt.seconds), and not in watts for any of the traces. [Note that the units of a "change in energy" are still energy units.]

Stan