Last month a got a call from FleetCarma asking if was interested in trying one of their products and commenting about it. They told me they saw the information I’ve been posting in this blog and thought I would probably like to use their on-board monitoring application and be able to provide good feedback. When I saw the kind of information I would be able to get with their product, I didn’t think twice before jumping on it.
FleetCarma makes an EV Performance Monitoring Solution that combines a data logger connected to the On-Board Diagnostics (OBD-II) port of the Leaf and a web portal where you can upload the collected data and create performance and monitoring reports. It is primarily meant to be used by fleets but it can also be used by individual drivers like me, interested in monitoring cost and performance. The data logger (called C5 logger) stays connected to the car basically all the time, collecting details about charging, driving, temperature and energy used. Like the SOC-meter, it is capable of decoding the exact state of charge of the Leaf to use in its reports, which gives much better precision than what I’m using right now (the 120V charging time).
The C5 logger collects and saves all telemetry on a 2GB microSD card, which is large enough to store months or possibly years of data. Every time you get a chance – I do it every few days, you unplug the C5 Logger, remove the microSD card and connect it to a computer using a standard microSD card reader. Then log on the web portal and upload all the data files in it. This will cause all reports to be updated automatically. The more frequent you upload the data, the more up-to-date the reports will be.
The main report is a dashboard showing the overall measurements collected and accumulated by the C5 Logger over time. By default, the report covers all the days you used the C5 Logger in the car, but you can also specify start and end dates. The dashboard has several “widgets” that show different measurements, and tabs that show more detailed information. We’ll look at each one of the widgets separately and analyze some of the most interesting data along the way, before digging into the report tabs.
The distance widget shows the total and daily average distances driven throughout the monitoring period. I just noticed that my daily average (66 Km) is almost twice my regular commute to work (34 Km). That means commuting is only half of what we use the Leaf for.
Driving energy shows the “MPG-equivalent” measured during the period, but in metrics (liters per 100 km). I don’t mind MPGe but “km per kWh” would also be a meaningful measurement to show (5.7 km/kWh in our case). This widget also shows the total energy supplied by the battery while driving (283 kWh) and how much energy was lost during the charging process (27 kWh). That basically means 8.7% of the electricity we spend on the Leaf gets lost in the charger. Something for my to-do list: compare these values with the ones measured by my kWh-meter to see how accurate they really are.
The time widget shows a pie chart comparing how much time the Leaf spent running, idling, charging and, well, resting. You can immediately see that almost ¾ of the time the Leaf is basically parked doing nothing. The remaining time is mostly spent charging (21%) and only 6% of the time is actually spent driving. The relatively high charge-to-drive ratio is due to the slow 3.3 kW Level 2 charger of the Leaf, which doesn’t take advantage of the 6.6 kW charging station we have, and also by the fact that about 30% of my charge comes from Level one at work (next widget).
Charging Energy shows how much I charged in total and how much of that charge came from a level 2 charger (71%, mostly at home) and from a level 1 (29%, mostly at work). This is new information for me, something I was never able to collect by any other means. I only had a rough idea of how much charging I was doing at work and how much that would cost to my employer. Now I know exactly! And as I always suspected, it’s not that much: about $2.50 a week on average. It also shows that the Leaf typically starts the day charged at 80% and ends the day with 50% of charge left on average (though this average has a very high variance, not seen here).
The Green House Gas Emissions widget is based on configurable options specified in terms of upstream emissions factors for electricity. I’m using a value of 50 g/kWh based on the mix electricity sources used by Hydro Ottawa during off-peak hours (mostly nuclear and hydro), plus a bit more to reflect some of the charging I do during the day at work. I’m not sure if that’s the most appropriate value to use. If anyone has a better suggestion, please let me know.
The dashboard is a great summary but the most useful reports are behind the other tabs.
Daily Summary gives a very good idea of how my actual driving needs fit within the range the Leaf can provide. The dark blue bar shows how much range I can get with the overnight charge at home (remember I’m charging to 80%). The additional light blue bar shows the extra range I get for charging during the day. The red mark shows how much I actually drove that day. This chart is great! It also shows the average outside temperature measured for each day as well as the energy consumed by the 12 V accessories and an “Eco Score” that rates the efficiency of my driving.
The Daily Summary also includes an histogram showing the state of charge distribution in the beginning and at the end of the day. You can see that the great majority of my days start with an 80% charge, and end all over the place but very rarely below 20%, and never below 15% ! Not much range anxiety for me.
The Daily Utilization chart gives you a great visualization of our driving and charging patterns. I can see the level 2 charging overnight, the commute to work, the charging at work, and all the errands I did in the evening, including some charging at the Science Museum during the EVCO meeting! (Oct 29 evening)
Notice that I configure the on board timer of the Leaf to charge only after midnight, even though the off-peak rates start at 7pm. This is because electricity demand is still fairly high early in the evening, and thermal generation is still going strong at that time. I prefer to charge later at night when demand is very low, and when the electricity is much, much cleaner.
Trip Details show the start time and duration of each trip, as well as distance travelled, state of charge (SOC) at the beginning and at the end, and the energy consumed. Notice the SOC because this is something you’ll never see on the Leaf display (at least until 2013 J). I’m still puzzled with the first to trips in this report (Nov 17). They covered about the same distance (~12 km) and used the same energy (~2 kWh), nevertheless the SOC dropped much more significantly at the second trip, by almost 12 points compared to 9 points. The main difference between the two trips was the duration, where one took almost twice as long. I suspect that might have to do with a visit to the car wash during the second trip. I regenerated the report just for that day and it showed as much “idling” as driving during that day. The mean temperature that day was 0.6C so I’m not sure if I had climate control on but I was certainly using the heated seat and steering wheel. The extra energy definitely showed up in the SOC but not in Electrical Energy Consumed column!
Charge Details show all charging events for both Level 1 and Level 2 charges (and Level 3 as well, if I had ever charged with the DCQC port…). Details also include charging time and duration, charge energy and loss, and the state of charge in the beginning and the end. I noticed a small glitch though, which may be specific for the Leaf. If you have the onboard charging timer configured and you plug in outside the configured charging time, you’ll see a Level 1 event of 0.0 kW, even if you plug in a Level 2 charger. Something similar happen when you unplug a Level 2 charger after charging is complete. This can be seen in the entry #6 above.
Both Trip and Charge Details data can be copy-and-pasted into a spreadsheet for more detailed analysis. In Excel, you have to do a “Paste Special” and select “Paste as Text” in order to parse the tab-separated columns properly. You can also import the Excel file created in IBM Cognos Insight for even deeper analysis.
How does it compare with Carwings so far?
Leaf owners familiar with Carwings will notice that FleetCarma provides much more information and more detailed and useful reports. One thing that particularly called my attention is that FleetCarma also provides more precise and accurate information compared to Carwings. Here’s an example: I ran a quick test by picking one day (Nov 18) and comparing the measurements provided by Carwings and FleetCarm.
|Nr of trips||3||3|
|Distance Traveled (km)||93.93||91.8|
|Electricity Consumption (kWh)||16.94||13.9|
|Average Energy Economy (km/kWh)||5.54||6.6|
The number of trips and the travel time are basically the same, but the 2% discrepancy in the distance travelled is a bit puzzling. I know for a fact that the distance measured by FleetCarma is very precise because measurements of my daily commute are always bang on with the trip odometer. For some reason, Carwings underestimated the distance travelled by more than 2%.
The biggest difference however is in the measurement of electricity consumption. Carwings seems much more conservative while FleetCarma precisely tells the consumption of each trip, which I tend to trust a bit more based on the before and after SOC observed. But the difference is quite alarming: about 3 kWh or 22% lower for Carwings. This makes the energy economy reported by Carwings about 1 km per kWh higher, or almost 20% better than what it actually is!
After using the C5 Logger for about a month, I can say I’m pretty excited with the results it has been able to provide, and they are far superior and more detailed than the ones we get from Carwings. Here are some of the things FleetCarma can do to improve the product even more:
- Streamline the upload process. Maybe a downloadable app for the computer that detects the SD card, uploads all the .BIN files and deletes the LOGS directory (after making a backup on the hard disk). Today these are all manual steps.
- Add Wifi support to the C5 Logger so that it can upload all the files automatically once car is parked at home. No more SD cards, no more computers, and very up-to-date reports online.
- Wifi can also be used to connect to an iPhone/Android phone in the car by enabling Personal Hotspot on the iPhone and have the C5 Logger connect to the Wifi network. This will allow an iPhone app to communicate with the logger and receive telemetry in real time. The app could also communicate via Bluetooth, if the hardware is available, but if Wifi is already there, you can use it instead.
- Add a small display on the C5 Logger to show the current SOC%. Great for 2011 and 2012 Leafs, makes Logger dub as an SOC-meter.
- Add “km/kWh” to Driving Energy widget on dashboard.
- Clarify that “avg start SOC” and “avg end SOC” on the Charging Energy widget refer to beginning and end of the day, not of each trip or charging session.
- Add temperature column to trip details. This will help in monitoring performance vs. temperature more accurately.
- Filter out charging sessions of 0.0 kWh charge. These seem to happen when you plug in the Leaf with the charging timer active but before it is time to charge.
- Add an option to export everything to a .csv file. Include all collected data since the last export. This way we can analyze performance using third-party analytics tools (e.g., IBM Cognos Insight).
- Create a service API that third-party apps can use to report on the monitoring data.
- Add a “favorite” widget right after the login page, so I can jump directly to my car dashboard or to a specific report.
- Auto-select car if “fleet” only has one car (upload page, reports page, etc.)
- Location awareness. Not sure if GPS data is available through the OBD-port. If it is, it would open a great door of posibilites.
In summary, a great product and lots of potential to evolve and become a killer app for EVs!
Update: C5 Logger Spec Sheet says WiFi/Bluetooth support is already in the works (Fall 2012).