Series Hybrid Mode

[gwatpe]

I have been recently recording how my PHEV has operated when subjected to higher power demands and a depleted battery in NORMAL mode.

On the last test, a section of freeway up a reasonably steep grade for about 8km, My PHEV switched from Parallel Hybrid operation to Series Hybrid. I have been reasonably critical of the engine screaming, with the associated noise that detracts from the typical peaceful drive offered by the PHEV, but there is a more important aspect with petrol economy.

The diagnostics I measured, were petrol consumption, throttle position, ICE load, Battery Power, Generator Power, and Electric Motor Power, as well as a few more that are not important in this test. It is quite difficult to present data, and I have yet to add functionality to my windows PHEV data historian to calculate kWh, so I can present so many kWh for a certain amount of petrol, so for simplification of the numbers, I have extrapolated the power to be equivalent to an hour of operations and distance to be 100km.


The road grade was very constant for the whole 8km, with only slow curves, typical of most modern freeways, with a normal hot mix type road surface, and the PHEV was operated with CC at 90kph.


The first section was in Parallel hybrid mode, with petrol consumption at about 16L/100km, and battery power was about 14kW. The Petrol consumption was equivalent to about 50kW. The ICE operated at 100% load at about 2330rpm, with a very constant throttle position at 63%.

The PHEV then converted to Series hybrid operation, with petrol consumption at 35L/100km and Generator output at about 74kW, and 4100rpm, with approx 12kW returned to the battery.

The power needed to propel the PHEV was equivalent to about 64kW. Do not have a pure electric number from battery alone as it was greater than the battery could deliver by itself.

In Parallel hybrid mode, the 16L/100km was equivalent to 50kW directly to the wheels through the clutch, plus the 14kW from the battery comes out at 64kW.

In Series hybrid mode, the 35L/100km produced 74kW, minus the 12kW returned to the battery, equals 62kW, so similar amounts of energy to propel the car.

There is a 26kW difference in power, due to the battery sinking and sourcing the power. This means that approx 19L/100km was used for the 26kW.

In Parallel hybrid mode, the PHEV only needed 16L/100km for an effective 50kW, but in series hybrid, returned 19L/100km for effectively 26kW.

The critics may now see why I choose to keep a reserve in the battery, to maintain parallel hybrid operations. Allowing the PHEV to slip back to series hybrid mode could totally remove any savings in petrol by running the battery down first.

Here is a pic of the data

The white trace is the throttle position that the PHEV computer set, to maintain the CC at 90kph. Notice how much the throttle is varied. The PHEV is not doing a very good job, and this could have contributed to the high petrol consumption. The driver had no input here, as the PHEV was doing all the controlling in CC. The high noise of the PHEV ICE screaming masked the behavior.

Now that I am aware of this operation style of the PHEV, I will endeavor to gather more data to show some more detail in another way, maybe with my right foot controlling the gas pedal.

 

[maby]

Interesting - and good to see some objective data that seems to support the conjecture that the car "operates better" with a moderate amount of charge in the battery! Half way round the M25 yesterday, some idiot suddenly pulled in in front of me leaving me no space to manoeuver - I was very pleased to have plenty of charge in the battery to be able to pull out into the fast and accelerate away hard.

 

[anko]

Interesting - and good to see some objective data that seems to support the conjecture that the car "operates better" with a moderate amount of charge in the battery!

The "conjecture that the car "operates better"", has always been about drivability and performance. This graph says nothing about this. At best, Gwatpe himself only comments on the drivability aspect in a subjective manner. The conjecture was never about fuel consumption (what this thread is all about). Although I have suggested / claimed / am trying to prove that maintaining SOC in a "just in time" fashion could actually improve fuel economy. But that being said, the graph only looks at more or less efficient ways of spending SOC, not on the cost of maintaining it. And, like you said yourself, we should not be making conclusions about longer trips based only upon info about short sections of the trip.

I was very pleased to have plenty of charge in the battery to be able to pull out into the fast and accelerate away hard.

You say you appreciate objective data. Well, I have provided several charts that, based upon actual measurements, show how power output is not affected by low SOC, at least not as long as you are at 22% SOC or above, so outside the Turtle zone. Gwatpe has tried to get into the turtle zone to in order to measure what the impact on power output would be, but he didn’t even manage to get into the Turtle zone. And then you counter with a claim that you were only able to quickly drive away from a situation thanks to a high SOC, without any form of back up. Doesn’t seem you appreciate objective data that much after all .

From a content perspective, there is a small issue, IMHO. In his post, Gwatpe estimates, based upon petrol consumption, that the engine produced 50 kW of output in parallel mode. How he gets to the estimation, he doesn’t say, but he confirms the number by what he sees happening in serial node: (In serial mode: 74 kW from generator - 12 kW for recharging means 62 kW is needed for driving, where in parallel mode: 50 kW from engine + 14 kW from battery means 64 kW is needed for driving). The problem I see is that, although the generator may take 74 kW of physical input, the electrical output is limited to 60 kW. This means that in serial mode the amount of power needed for driving was approx. 60 - 12 = 48 kW. Consequently, the amount of power produced by the engine in parallel mode would have been no more than roughly 34 kW. The real engine output will be somewhere in between (looking at the engine torque curve, max output at 90 km/h is definitely a decent bit less than 45 kW). At this time, we do not know exactly where it is / was, but for the time being it does invalidate the general conclusion Gwatpe presents, regarding differences in efficiency in serial vs. parallel mode.

Of course I too think that parallel mode will be more efficient than serial mode. But whether the positive effect of parallel mode at low speeds outweighs the negative effect of driving around with a high SOC has yet to be proven, if you ask me.

 

[maby]

...

Of course I too think that parallel mode will be more efficient than serial mode. But whether the positive effect of parallel mode at low speeds outweighs the negative effect of driving around with a high SOC has yet to be proven, if you ask me.

My comments are quite clearly subjective - gwatpe's measurements seem to be the first published that show evidence of some measurable differences in the way the car operates with lower levels of charge that are not low enough to trigger the turtle.

Whether or not the positive effects of parallel mode outweigh the negative effect of driving around with a high SOC depends entirely on your criteria. You are totally focussed on fuel economy and emissions and make a case for running on a low SOC on the basis of those criteria. I really don't care very much about either fuel economy or emissions - I'm much more interested in performance/driveability/battery service life - and, subjectively, keeping the SOC up seems to have benefits from these criteria. The received wisdom is that the life expectancy of a lithium battery is a function of the number of charge-discharge cycles it goes through - in the case of my car, this is about three per week - so I'm hoping for a long service life.

While I can see some theoretical basis for your belief that running on a low SOC will improve fuel economy, several people have run tests now and nobody has been able to demonstrate any difference large enough to stand out over background noise. Until they do, I'll stick to the policy of running EV on my short weekday trips and Save on the long runs. One or two mains charges gets me through the week, so I'm hoping that my battery will stay healthy longer than average.

 

[Tipper]

Having a car with a 5-year warranty I'm not that worried about long term battery life as we plan to change the car well within that 5 years, probably at around 4 years old. The warranty was a real deciding factor in our purchase decision particularly with respect to battery life/replacement cost.

In addition, selling a hybrid car with no warranty I expect to be difficult, other than any Mitsubishi trade in incentives of course, and probably very uneconomical!

I'm also very keen to see the ASX PHEV next year as a slightly smaller car would suit our driving profile better.

 

[gwatpe]

The problem I see is that, although the generator may take 74 kW of physical input, the electrical output is limited to 60 kW.

I have checked again my original log file, and have added the total Electric motor power to my plot program. It turns out that the PHEV also recorded about 48kW to propel the PHEV up the hill. I have to admit that that the total power was an overestimate. My apology. The 48kW at 90kph, leaves the ICE supplying 34kW equivalent to the wheels directly in parallel hybrid plus the 14kW from the battery.

The car has returned data that has been recorded and then plotted. The PHEV returned a Generator power of 74kW. This is electrical power, calculated from presumably generator current and battery volts. There is no direct way of getting physical generator input power, unless by some internal fudge factor, that is used to provide the OBD generator power number. I cannot change this power number, without introducing a reverse fudge factor. If this data is wrong, then this puts in question all of the data that we retrieve from the OBD port.

If we have to introduce factors to correct calibrations, then I am asking for an adventurous PHEV driver to provide some.

This is a Torque dash screen showing some dials I now have to look at instead of the boring MMCS screen. credit goes to another esteemed forum member for some serious hard work on his proxy server.

Any manipulation of the data, either external factors, or even time averaging can be seen as corrupting the data, and we may as well give up providing data, as it can always be seen as subjective and not to be believed.

PS edit: My previous drives, a Datsun 180B, I had for about 8 years, when it rusted out. A Suzuki LJ50, for about 9years, A Toyota Corolla, 1prior owner, was 18 years when traded. A Honda Prelude 4WS was 20 when retired, and then a Nissan Maxima was 17 when traded for the PHEV. Never had a car from new, and am hoping to have my PHEV for well past 10 more years. Never had a car with warranty before, as all prior were over 1 when I purchased them.

 

[anko]

I have to admit that that the total power was an overestimate. My apology.

Please, NO!!! Do not apologise for making mistakes. If mistakes are no longer allowed, everything will come to a stop here.

The car has returned data that has been recorded and then plotted. The PHEV returned a Generator power of 74kW. This is electrical power, calculated from presumably generator current and battery volts. There is no direct way of getting physical generator input power, unless by some internal fudge factor, that is used to provide the OBD generator power number. I cannot change this power number, without introducing a reverse fudge factor. If this data is wrong, then this puts in question all of the data that we retrieve from the OBD port.

I was totally under the impression you were using my extended PID list, to which I have added Generator Power myself as a calculated PID:

GENP,Power,213C02,((A*256+B)/10-1000)*((C*256+D)-20000)/9548.8,-60,+60,kW,

In plain English: Power = Torque * RPM / some factor which translates RPM to radials per second.

So, that power number does not come from the OBD port, but is calculated based on some other numbers that do come from the OBD port. I have started a thread a while ago, seeking help on how to use Current 1 and Current 2 and Voltage to calculate E-motor power consumption (or generator power production), but this has lead to nothing. This is why I have stayed with the mechanical side of things.

Unless you have found an OBD request / response that presents real electrical output of the generator (which you would hopefully share with us), I would expect your reading to be the mechanical input to the generator. Does that make it useless? I don't think so. Although it is more difficult to make solid calculations it still allows us to understand a lot about what is going on in the car.

 

[anko]

gwatpe's measurements seem to be the first published that show evidence of some measurable differences in the way the car operates with lower levels of charge that are not low enough to trigger the turtle.

Seriously?
http://www.myoutlanderphev.com/forum/viewtopic.php?p=21683#p21683(by all means, a topic title "Impact of SOC on performance / drivability" in which you participated)
http://www.myoutlanderphev.com/forum/viewtopic.php?p=21482#p21482
Have you totally missed these? Or are my contributions on this subject simply ignored by you because they might prove you wrong? That is too bad, as I put an awful lot of effort in producing them. In an as objective as possible manner.

Whether or not the positive effects of parallel mode outweigh the negative effect of driving around with a high SOC depends entirely on your criteria. You are totally focussed on fuel economy and emissions and make a case for running on a low SOC on the basis of those criteria. I really don't care very much about either fuel economy or emissions - I'm much more interested in performance/driveability/battery service life - and, subjectively, keeping the SOC up seems to have benefits from these criteria. The received wisdom is that the life expectancy of a lithium battery is a function of the number of charge-discharge cycles it goes through - in the case of my car, this is about three per week - so I'm hoping for a long service life.

While I can see some theoretical basis for your belief that running on a low SOC will improve fuel economy, several people have run tests now and nobody has been able to demonstrate any difference large enough to stand out over background noise. Until they do, I'll stick to the policy of running EV on my short weekday trips and Save on the long runs. One or two mains charges gets me through the week, so I'm hoping that my battery will stay healthy longer than average.

Depends on criteria, you say? The subject of this topic was clearly "the effect of serial mode on petrol consumption". Nothing more, nothing less. If you do not care about fuel consumption that much, then why do you get involved in this discussion, if I may ask?

 

[Neverfuel]

gwatpe's measurements seem to be the first published that show evidence of some measurable differences in the way the car operates with lower levels of charge that are not low enough to trigger the turtle.

Seriously?
http://www.myoutlanderphev.com/forum/viewtopic.php?p=21683#p21683(by all means, a topic title "Impact of SOC on performance / drivability" in which you participated)
http://www.myoutlanderphev.com/forum/viewtopic.php?p=21482#p21482
Have you totally missed these? Or are my contributions on this subject simply ignored by you because they might prove you wrong? That is too bad, as I put an awful lot of effort in producing them. In an as objective as possible manner.

Whether or not the positive effects of parallel mode outweigh the negative effect of driving around with a high SOC depends entirely on your criteria. You are totally focussed on fuel economy and emissions and make a case for running on a low SOC on the basis of those criteria. I really don't care very much about either fuel economy or emissions - I'm much more interested in performance/driveability/battery service life - and, subjectively, keeping the SOC up seems to have benefits from these criteria. The received wisdom is that the life expectancy of a lithium battery is a function of the number of charge-discharge cycles it goes through - in the case of my car, this is about three per week - so I'm hoping for a long service life.

While I can see some theoretical basis for your belief that running on a low SOC will improve fuel economy, several people have run tests now and nobody has been able to demonstrate any difference large enough to stand out over background noise. Until they do, I'll stick to the policy of running EV on my short weekday trips and Save on the long runs. One or two mains charges gets me through the week, so I'm hoping that my battery will stay healthy longer than average.

Depends on criteria, you say? The subject of this topic was clearly "the effect of serial mode on petrol consumption". Nothing more, nothing less. If you do not care about fuel consumption that much, then why do you get involved in this discussion, if I may ask?

Keep it clean please gentlemen!

 

[gwatpe]

Apart from an assumption I had made re some PID, that are calculated and not PHEV data, the fact that series hybrid mode gave economy of about 35L/100km compared to Parallel hybrid mode at 16L/100km with a small change in how the battery was used at the same vehicle loading, highlights a good reason to keep a %SOC reserve.


It may be possible to work out some 10min data for comparisons. I say 10mins, as this is about how long driving with a full battery, and at 50kW load, a driver would expect to travel. Possibly about 15km at 90kph, up a steep grade. We could calculate an approx electricity cost, for me about $1.80 from offpeak grid power @ $0.20/kWh.

Traveling in Parallel hybrid mode for 10min at 90kph with about 50kW load, 14kW from the battery and 16L/100km @ $1.20/L gives $2.90 for petrol, plus about $0.30 for electricity.

Travelling in Series hybrid mode for a total of 10mins and for the same conditions has 12kW going back to the battery and 35L/100km, so $6.30 for petrol minus $0.25 for electricity.

The driver can choose how to use the electricity. Series hybrid costs approx double per km compared to driving in parallel hybrid mode, in conditions I outlined above. These are more extreme conditions, and certainly low speed, low power demands, series hybrid offers economy savings. We have to be adaptable, and apart from benefits of cleaning out the fuel systems, series hybrid operations under high power demands, is not economical with petrol consumption.

Every time I drive to the city, If I used NORMAL mode[flatten the battery first mode], the high power series hybrid operation played a part. Preserving the battery, for the drive home, makes complete sense now and my mods, to help me maintain the battery levels, seem vindicated. Better economy and a more pleasurable drive overall.

 

[jaapv]

Keep it clean please gentlemen!

Don't worry, this forum is moderated. If you notice something that does not pass muster you can report it and a moderator will take action if needed. The little red exclamation mark at the lower right hand side.

 

[anko]

We know $1,00 of electricity will take you further than $1.00 of petrol. Otherwise, we would probably not be driving hybrid vehicles :mrgreen: But anyway, because of this it is rather dangerous to simply add cost for electricity and cost for petrol, without looking at what that money will bring you in terms of mileage. So, lets try to bring it back to cost per distance traveled.

The difference to the (dis)charging rate of the battery between serial and parallel drive is 26 kW (+12 vs -14 kW). This is (by approximation) equivalent to 26 / 34 = 76% of what is needed for propelling the car (I would not call that "a small change in how the battery was used", BTW). So, the $6.30 worth of petrol in serial mode gives you about 3/4 more miles than the $2.90 in parallel mode. And then we are looking at 1.76 * $2.90 = $5,10 versus $6.30 to travel the same distance.

Of course, looking at this isolated section of a trip, parallel mode must be king when it comes to fuel consumption, because only part of the power needed for driving needs to be converted twice (mech to E and E to mech), where in serial mode all power needed for driving needs to be converted twice. But IMHO the overall difference is not nearly as big as you suggest.

An approach of looking at the ratio of money spent on that isolated section of your trip in serial mode versus the same in parallel mode results in ($6.30 - $0.25) / ($2.90 + $0.25) = 1.92.
An approach of looking at the ratio of cost per distance travelled in serial mode versus the same in parallel mode results in 6.30 / 5.10 = 1.24.

So, cost per distance travelled goes up by 'only' 24% when dropping out of serial mode. And to me that kind a makes sense, because if the difference was as big as suggested earlier, then it would probably be counter productive to have a hysteresis cycle for speeds > 65 km/h at all.