the significance of 12L/100km

[gwatpe]

I am presently collecting data from the OBD2 port using Torque, and have just driven from my home to Melbourne, a distance of approx 800km. The trip was about 8 hours of driving, with a bit of traffic, and mainly flat, but a slight head wind. Have a lot of data to present in a few weeks when I return home.

A significant observation is the predominance of 12L/100km at speeds between 90 and 110kph, that are interspersed with EV only operation, with the PHEV in SAVE mode with the battery operated from 70% full, through 50% and down through 30% and then depleted. The higher speeds reduce the proportion of EV only operation. The %SOC does not seem to have any impact on the performance seen in my PHEV in these constant speed and load lests.

I also have seen a maximum of 15L/100km before the battery helps out at higher power demands, at the 100kph speeds.

I have also tested the time required to recharge the battery from empty. This test still only had the ICE operating at 12L/100km at 100kph, but with no periods where the ICE turned OFF. As the Battery filled up to approx 75%on the gauge, the consumption stayed at around 12L/100km. The consumption did move in an oscillatory way between 12L/100km and down to about 8L/100km up to a point with 15bars. The PHEV is "YES" less efficient as the battery becomes fully charged, but it seems to be about 0.5L/100km worse.

My PHEV has 40000km on the clock now. The speeds recorded were GPS. My PHEV showed 108kph on the dash for actual 100kph. Over the 800km My PHEV returned 8.1L/100km by the car and the PHEV has returned 21%EV. At 100kph my PHEV averaged 24secs EV for every 74secs of ICE operation. This is approx 25% and is close to the 21% returned by the car, that has been impacted on many other factors in 800km. The economy at 8.1L/100km is a result of slower speed operation and some battery usage.

I have not yet established the absolute accuracy of the PHEV L/100km PID. The data I have presented relies on the accuracy of the fuel delivery measurement and the speed, as per the PHEV sensors. Other drivers who are able to record the PHEV OBD data may be able to share their data and compare. I believe the PHEV is operating a particular way and if other drivers record different measurements, then we need to factor this into our discussions.

I will be continuing my recording over the next week or more.

 

[anko]

A significant observation is the predominance of 12L/100km at speeds between 90 and 110kph, that are interspersed with EV only operation, with the PHEV in SAVE mode with the battery operated from 70% full, through 50% and down through 30% and then depleted.

Are you getting 12l/100 km even at 70% SOC? I am pretty sure mine drops to well below 10l/1200 km when SOC is that high. And starts to fluctuate much more as well.

I also have seen a maximum of 15L/100km before the battery helps out at higher power demands, at the 100kph speeds.

I recognise that number, and it makes perfectly sense. Normal state (with a sufficiently depleted battery) is 75% load, which results in the 12 l / 100 km you see. Then 100% load would be 12 * 75% * 100% = 16 l / 100 km. At 100 km/h consumption is probably a bit less than 12 l / 100 km?

I have also tested the time required to recharge the battery from empty. This test still only had the ICE operating at 12L/100km at 100kph, but with no periods where the ICE turned OFF. As the Battery filled up to approx 75% on the gauge, the consumption stayed at around 12L/100km. The consumption did move in an oscillatory way between 12L/100km and down to about 8L/100km up to a point with 15bars. The PHEV is "YES" less efficient as the battery becomes fully charged, but it seems to be about 0.5L/100km worse.

Perhaps you can add Engine Load to your TP dashboard? That would be interesting to see. Were you running A/C? I guess you would be.

At 100kph my PHEV averaged 24secs EV for every 74secs of ICE operation. This is very close to the 21% returned by the car. The economy at 8.1L/100km is a result of slower speed operation and some battery usage.

Were you able to see if this 74 : 24 ratio was impacted by SOC? I would expect it would be, as it is in my tests...

Safe driving.

 

[gwatpe]

Aircon was ON as ambient was well above 30C.

I have not seen any relationship with the economy and %SOC. I have to make a better way of looking at the data first. Excel is just not doing it for me. I will be adding some smart calcs to my data historian to be able to zoom in on a particular test section and have the %EV and average petrol consumption calculated for just that portion of the data. This will allow me to look at bits of the same data log and post some graphs at some different %SOC levels. Will then be able to plot averaged economy vs %SOC and show what my PHEV does at say 90, 100 and 110kph. This will probably not be on the same day, as I am trying to have the PHEV do its thing for more than 5 cycles to avoid battery biasing.

here is an example of some data that needs to be worked on.

at 0 was 70%SOC
at 1 was 50%SOC
at 2 was 30%SOC
at 3 was empty
at 4 was CHARGE mode pressed.

This is my PHEV data, so may be different to what others find.

 

[gwatpe]

I have just finished a revision of my PHEV Torque PRO log file analyzer to allow for zooming within zooming to accurately select a portion of the logged data file. This has allowed me to precisely calculate the average petrol consumption from a constant speed and vehicle loading portion of driving that has periods of EV only and parallel hybrid operation.

I used a section of the data before the 0 and did not have enough data at the section 2.

The analysis started with 15bars and was repeated at 12bars, 8bars, Empty and then recharged back to 15bars. The tests were at 100kph, except the first section with 15bars. The PHEV was operating in SAVE mode, for the first 3 sections, and in NORMAL mode when empty and then in CHARGE mode for the last stage.

Here are the graphs of each section.

first section. almost full battery in SAVE mode
9.16L/100km

second section about 80% full battery in SAVE mode
8.29L/100km

third section about 50% full battery in SAVE mode
8.84L/100km

fourth section empty battery in NORMAL mode
7.81L/100km

fifth section battery close to full again in CHARGE mode
8.47L/100km

One could conclude that operating with a flat battery is slightly more efficient. It seems that any time the PHEV is operated in SAVE mode with some reserve battery capacity, that the PHEV returns lower fuel economy. The PHEV returns the worse economy when operated in SAVE and CHARGE mode with the battery at a high reserve capacity.

As a conclusion, we will need to weigh up a slight improvement in fuel economy against the increased risk of reduced battery life.

One aspect that was evident from the logging is that when the PHEV is operated in CHARGE mode with the battery full, that almost no battery is used, ie no mini cycles, so the PHEV is essentially a petrol car. This is for constant load conditions as tested. Obviously periods of high power will draw down the battery, and this will be replaced accordingly in this mode.

I just need to check the calibration of the PHEV computer so my data can be compared with others.

 

[anko]

Hi gwatpe,

Brilliant stuff. When do you start to take orders?

The way I read this graphs is this:

In the first graph, fuel consumption is all over the place. Every change in driving conditions (wind, slope, etc.) results in a change to the engine load. Just like with a normal car. As a matter of fact, you are driving a normal car at that time, be it a heavy one. At 80% consumption is settling a little bit as, in order to adopt to changing driving conditions, the car can now also play with the charge current rather than just the engine load. But as the battery does not want to accept to much power at this high SOC, this effect is still limited. At <= 50% SOC the battery is willing to accept much more power. As a result, adopting to changing driving conditions can almost totally be done by changing the battery charge current rather than the engine load.

(Side step: I also believe that this is why our CC is so well capable of maintaining a single speed over changing conditions: playing with charge current has a far more direct result than playing with the throttle.)

As a result, average load is higher at times the engine is running. Over here, for normal cars "new driving" (don't know if that is a Dutch thingy) is propagated as a suggested driving style : use a high(er) gear where possible, resulting in higher load at lower RPMs rather than lower load at higher RPMs. What we see here is similar, but instead of lower RPMs we have less REVs. But apart from that, the same thing. Next to that, there are far less fluctuations to the engine load. I believe that too may have a positive effect to the efficiency of the engine.

From memory, I thought I have noticed that, when it comes to how much power the battery will accept, the first step down is at about 50% (on the dash, so 65% real). Below that 50%, measurements should be close together. Above that is where you will find the big differences. From 9.16l/100km in the worst scenario to 7.81l/100km in the best scenario is more than "slightly more efficient" in my book. If I am not mistaken, it is 14 - 15% less.d

May I ask how did you calculate the averages consumption from the TP data? I assume you did take into account that log lines are not necessarily spread out evenly in time? Sure, you have.

BTW: These kind of graphs are not to unfamiliar. See http://www.myoutlanderphev.com/forum/viewtopic.php?p=17113#p17113. But it is nice to seem the results being replicated by somebody else and I like the presentation and usability much better . It got me an awful lot of headache to produce mine :evil:

 

[gwatpe]

I spent about 4 hours to build my analyzer program. Still have to add total km driven and total litres used and average speed to the graphed data in the window. will only take a a few mode minutes hopefully.

To be useful, I don't like having to manipulate data columns. I prefer to just click on a button to turn a graph on or off, and use the mouse to select a portion of a graph. I can turn off all graphs except the interesting one. I can select down to a 2 second accuracy on the zoom window and select a smaller portion if needed. calcs are re adjusted in real time on the new graphed data. selecting a different portion of data is as simple as clicking on the update graph button and then reselecting a different portion to look at with the mouse.

At the moment the program has a fixed data format that I just happened to select from the order I added PID's to log in Torque. If everyone only logged these PID in the same order, then others could look at their data the same way with my app. I will be looking at a customizable app at some point after I return from my trip.

My next testing will include the method of "Trex" using EV and recharge to 50% modes, with the CHARGE button.

Unfortunately my testing will only help me on longer holiday type travels and not so much for where I live.

The average for the data is for the data as displayed on the graph window. Add up all the instant L/100km data points and then divide by the number of points. This is why it is so important to be able to accurately select the start and end of the portion to analyze. If the user is careless the average may be high or low by as much as 10%. You will notice that I have selected data portions with EV and ICE operation to have close to equal spaced portions of EV and ICE and that there are the same number of EV and ICE operations in each graph. The number of cycles was dependent on the quality of the data, to avoid bias. I cannot express how difficult it is to extract data from real world driving with changing road conditions and traffic considerations. Don't try this on a busy motorway.

To keep my data consistent I will check the fuel calibration, only after I have the next phase of data collected.


PS there is a mistake on the fourth graph. shame on me for not checking before posting.

Readers will notice that there is a period of EV operation at the start and end of the graph. This has biased the data to a reduced petrol consumed number. The corrected averaged economy is 8.48L/100km and not 7.81L/100km as shown above. Not as much difference as I had mentioned.

 

[gwatpe]

Have just returned from another interstate drive and had the opportunity to gather some more data. I have confirmed that my PHEV has a factory, internal calibration factor that corrects the raw OBD fuel sensor data, so that the MMCS and dash displays show accurate petrol consumption numbers. The RAW OBD2 petrol L/100km data however needs to be multiplied by 0.88 to provide more accurate data. The Torque PRO has a facility for this to be applied in real time.

My testing has shown the difficulties in obtaining real world data from driving testing. Too short a drive and the battery energy muddies the data, and equally problematic is too long a drive, where terrain and traffic as well as the weather muddy the collected data.

I have tested the interactive CHARGE and EV mode with the battery between empty and half full, against just SAVE mode with the battery at half full capacity. The tests were over 200km each over flat terrain at highway speeds and minimal traffic and towns.

The CHARGE mode had 3 full battery cycles and the SAVE mode had too many to count. The SAVE mode test data shows a trend that the PHEV appeared to try and change the petrol algorithm to feather the petrol consumption. The road condition may have had some contribution, but this would be very difficult to confirm.

Here are some pics

CHARGE mode

and SAVE mode

and some zoomed in sections

early in the SAVE mode test

and late in the SAVE mode test

and expanded zoomed in on the late test data

The PHEV has no hard and fast rules when it comes down to it, and is continually adjusting operating parameters for optimum economies.

I still have to check to see if the battery health has suffered from my testing.

 

[anko]

Thanks for sharing!

The SAVE mode test data shows a trend that the PHEV appeared to try and change the petrol algorithm to feather the petrol consumption. The road condition may have had some contribution, but this would be very difficult to confirm.

Trying to understand what you are saying here. Can you elaborate a bit:
- What is the trend you see?
- What do you mean by "change the petrol algorithm to feather petrol consumption"? And what in these graphs make you think it does so?
- Do you have an explanation as to why speed varies much more in the second half of your save mode test?
- Can you elaborate on your conclusion "The PHEV has no hard and fast rules when it comes down to it, and is continually adjusting operating parameters for optimum economies."?

What I see to me it appears as if in the first half of the Save mode test the car controls speed by playing with the Charge current rather than the throttle, where in the second half, it is playing with the throttle to maintain speed (and is less effective in doing so).

 

[gwatpe]

Thanks for sharing!

The SAVE mode test data shows a trend that the PHEV appeared to try and change the petrol algorithm to feather the petrol consumption. The road condition may have had some contribution, but this would be very difficult to confirm.

Trying to understand what you are saying here. Can you elaborate a bit:
- What is the trend you see?

After about 100km, the PHEV seems to have more difficulty utilizing the battery to maintain the driving needs, and attempts to use the throttle on the ICE instead.

I have put some arrows on a zoomed plot of SAVE mode operation in question.

The yellow arrows show a slight increase in road speed, and then a corresponding reduction in the petrol consumption as the computer adjusts the ICE throttle. The battery is unable to supply and source the small differences in power needs, so the throttle is adjusted instead. The road conditions were essentially the same. [The ambient temperature had risen a few degrees over the duration. The battery had been heavily cycled prior with the CHARGE/EV mode test, so the SAVE mode test may have started with a warmer battery, and this may have cooled slightly during the 4hrs of the SAVE mode test].

The blue arrows are the typical periods of EV only operation associated with SAVE mode operations.

The green arrows show the recording, 2 second period logging interval data fluctuations. The bigger dips in the red L/100km trace are not noise, but recorded changes in petrol consumption.

The horizontal yellowish trace is the 10L/100km level.



- What do you mean by "change the petrol algorithm to feather petrol consumption"? And what in these graphs make you think it does so?

Check the pic above with the changing petrol consumption instead of turning off the petrol motor


- Do you have an explanation as to why speed varies much more in the second half of your save mode test?

The PHEV has lost some of the ability to control speed as the battery in not as supportive

- Can you elaborate on your conclusion "The PHEV has no hard and fast rules when it comes down to it, and is continually adjusting operating parameters for optimum economies."?

I have seen and recorded the way the PHEV operates, change in a single drive. If the rules were more rigid, I should have seen the PHEV operate the ICE ON and OFF in SAVE mode for at least a full tank of petrol over essentially the same road conditions, the same way. FLAT with virtually no traffic and only slow road curves and only a few slows for towns.

What I see to me it appears as if in the first half of the Save mode test the car controls speed by playing with the Charge current rather than the throttle, where in the second half, it is playing with the throttle to maintain speed (and is less effective in doing so).

The PHEV was operating in CC mode. If the PHEV used the same rules, then it would not have changed the operation mode.

My other observation was that CHARGE mode offered no significant benefit with returned petrol economy. The petrol economy seemed to be marginally better towards the end of the SAVE mode test, and this may be due to less processing of electric energy with the battery. We are only talking fractions of a L/100km, so more controlled tests may be required to obtain a more definitive result.

 

[anko]

A short while ago, "feathering the throttle" was presented as a way of (or at least in the context of) optimising fuel economy. Because of this, when you wrote "... the PHEV appeared to try and change the petrol algorithm to feather the petrol consumption", I though you meant to say the car was trying to optimise fuel economy by playing with (feathering?) the throttle. Now I think this is not what you meant at all. Please keep in mind, I am not a native speaker

I also assumed kinda automatically that when you mentioned a trend that you were thinking of a positive trend.

Now I think you were saying "the car is playing with (feathering?) the throttle in stead of playing with (feathering??) the charge current in its attempt to maintain a constant speed and, as I also said yesterday, it is become less effective doing so resulting in variations in speed. Does that make sense?

Big question is: why does the car start playing with the throttle instead of with the charge current? As you suggested, it could be related to "the battery being less supportive". One possible reason for this is clear (IMHO, at least): high SOC. But according to your description this is to the case here, and this is backed up by the high instantaneous fuel consumption.

The fact that you still have periods of EV drive suggest the battery is accepting significant amounts of energy. And again, the fuel consumption level backs this up. If the battery was, for whatever reason, not willing to accept higher charge currents, that would explain the short dips in fuel consumption (when the CC slows down the car a little bit). But it would not explain the peaks in fuel consumption (when the CC slows down the car a little bit). These peaks can have three reasons (again IMHO):

- More additional power is asked than the battery can provide
- The battery has no power to provide at all
- The cars brain is for some reason not using any additional power from the battery

The first seems not very likely, as there are only very small variations in speed.
The second seems not very likely either, as you were running in Save mode.
The third is something we, IIRC, have not observed earlier.

You said the battery might have cooled down after the longer Charge cycle. But still your battery would have been much warmer than the batteries of us poor souls up North. Could it be that the raise in ambient temperature (how warm was it?) has brought the car in a different operating mode?

 

[gwatpe]

I guess what I am seeing is that the PHEV rules of operation are quite dynamic. I believe that for short drives, we may be able to find a particular energy source, like the battery for supplying the PHEV power needs. As the drives become longer, a particular driving style we thought would offer better petrol economy may not be valid. Extrapolation of short drives data to a longer drive is not always going to work.

BTW, the battery gauge showed half full during all of my SAVE mode testing. I don't consider this to be a high charge amount. 14-15 bars maybe.

The ambient temperature was around 25C at the start of the CHARGE mode testing, and ended up at 29C at the end of the SAVE mode testing. There may be some temperature around 28C that the PHEV program changes.

I am presently unable to monitor and log the battery %SOC or battery current with petrol consumption and speed. I have a new OBD2 adapter on order, so will be looking into more extensive logging then.

There was a marginal reduction in petrol consumption as the PHEV used the throttle instead of the battery to adjust power to maintain road speed. The speed control was less effective with the throttle, with +/- 2km/hr variation, compared to occasional +2km/hr variation with the battery. I am not too fussed about the lack of rigid speed control, as I use the CC as a personal over-speed deterrent, and occasional periods of +2kph will be unlikely to get me a ticket.

 

[anko]

I guess what I am seeing is that the PHEV rules of operation are quite dynamic. I believe that for short drives, we may be able to find a particular energy source, like the battery for supplying the PHEV power needs. As the drives become longer, a particular driving style we thought would offer better petrol economy may not be valid. Extrapolation of short drives data to a longer drive is not always going to work.

Hmmm, based on these measurements, I am not convinced yet. There could have been parameters that affect what you see, other than just progression of time.

BTW, the battery gauge showed half full during all of my SAVE mode testing. I don't consider this to be a high charge amount. 14-15 bars maybe.

As I said so, neither do I. On the other hand, I know from measurements that the "supportiveness" of the battery (how willing is it to accept charge currents) goes down one notch somewhere around the 50% mark (not 50% SOC but 50% of the bars). I have seen this change the behaviour of the car from "charge current controlled" to "throttle controlled". I have actually presented pictures of this phenomenon quite a while ago. For a moment, I was thinking that the difference in patterns between your first and second phase could have been explained by your SOC just crossing over that 50% boundary. But I dismissed that possibility, as that should also have resulted in a drop of instantaneous fuel consumption and a reduction of %EV. An crossing that boundary would IMHO also have resulted in an increase of average fuel consumption, rather than a slight drop.

If you are going to do a similar test once more, maybe it is a good idea to keep your SOC slightly away from that boundary (perhaps 30 or 40%) just to eliminate it as a factor.

The ambient temperature was around 25C at the start of the CHARGE mode testing, and ended up at 29C at the end of the SAVE mode testing. There may be some temperature around 28C that the PHEV program changes.

Oh. Only 29 :mrgreen: I thought I heard you or Trex or ... say the 40s were coming up.

There was a marginal reduction in petrol consumption as the PHEV used the throttle instead of the battery to adjust power to maintain road speed.

Please educate me on the English language. When you write "as", do you mean "when" or "because"?

If it was "because", then why would the car not have been programmed to do it this way all the time?

The speed control was less effective with the throttle, with +/- 2km/hr variation, compared to occasional +2km/hr variation with the battery. I am not too fussed about the lack of rigid speed control, as I use the CC as a personal over-speed deterrent, and occasional periods of +2kph will be unlikely to get me a ticket.

Same here. But I prefer to interpret the very constant speed as a sign of good health and of a system that is doing what it was designed to do. Obviously, I have no back up for this preference

 

[gwatpe]

On a longer trip, I will be using whatever gives the best petrol economy. If this happens to be SAVE mode with a full battery, then so be it.

On shorter trips, the battery will still be KING, and used accordingly.

I saw no clear cut benefit on a longer petrol only trip, of heavily cycling the battery with cyclic operation of CHARGE mode and EV mode by pressing the CHARGE button, chasing fractions of a L/100km.

The battery plays a big part in allowing drivers to post the sub 8 L/100km economy figures, for drives at highway speeds of around 100kph.

Maybe the motorways in Europe provide a better driving surface compared to the black top "highways" we have in AUS.

It is going to be difficult to evaluate the PHEV until the required data, until at least %SOC can be recorded, with say petrol consumption and speed to at least confirm the part the battery is playing in the returned petrol economy data.

I will possibly be able to get more data next year on another longer driving trip, hopefully with EvBatMon and a new OBD2 adapter. Will see if others can gather objective data to support a particular driving style in the mean time.