Testing Vehicle to grid (V2G) possibilities in the INVADE project, with surprising results
Next to the large scale Smart Charging of electric vehicles in the public domain ElaadNL also tested charging and discharging of a V2G-car as part of the INVADE Horizon 2020 project. What are the surprising results of the tests, and the possibilities for balancing the grid with vehicle to grid?
The V2G-tests were done with a Nissan Leaf on a DC bidirectional charging station, using Chademo. - Foto: Nynke Arends
Getting started… selecting a car and charging station
The purpose of the tests was to research the new possibilities within the Open Chargepoint Protocol (OCPP 1.6) when including the reverse energy-flow in DC-charging. To do the tests a car with V2G possibilities and a charging station with V2G were off course necessary. Obtaining these proved to be a challenge, since both cars and charging stations are not easily available yet.
After an inventory process on V2G chargers a 10kW 3 phase DC V2G charger from the Belgian charge point manufacturer eNovates was installed at the testing area of ElaadNL in Arnhem. Of the three cars capable of reverse energy-flow charging on DC (two BEV’s and one PHEV) the Nissan Leaf was selected and used for the pilot.
Developing testing procedures for V2G
Since V2G is a new technology for which standard test procedures did not exist prior to the work in INVADE, we developed a new test procedure for testing V2G cars and chargers, for both AC and DC V2G, for testing future cars and chargers. This procedure is now the standard in the Netherlands. The tests for INVADE as well as future tests are carried out in the Dutch testing facilities at the Elaad Test Laboratory premises, which was partly funded by INVADE budget and which was officially opened by Secretary of State Van Veldhoven op April 18th, 2018. See: https://www.elaad.nl/news/secretary-of-state-van-veldhoven-opens-test-lab-for-research-on-charging-of-electric-cars/
Figure 1: Test runs of a session performing charging and discharging alternately.
Interesting things we found about V2G in practice
During the first tests we noticed a few interesting things regarding V2G, car and charging station:
-The discharge speed depends on the current state of charge (SOC) of the car, and decreases when the state of charge is lower, meaning that when the batteries are less charged the speed of discharging is lower. The difference is not yet specified in detail but is significant. We measured several hundreds of Watts of difference.
-The Nissan Leaf allows discharging until 10% state of charge, that is until the batteries are 90% empty.
-When the state of charge reaches the lower or upper limit (10% or 100%) the charging/discharging sessions stops and cannot be restarted without driver interference (in practical terms: driving back and forth). This is default behavior of the CHAdeMO protocol and has to be taken into account when programming the charging commands. As limits are car specific, the upper and lower boundaries of sessions will have to be based on the car with the most narrow range.
-Sessions lasting over 24 hours are automatically stopped, which is normal behavior for a regular charging station but not desirable for a V2G charger. This has been reported to the charge point manufacturer to be fixed.
-The car does not have any indicators to show charging or discharging. There is also no indicator for charging/discharging speed. The only indicator present, shows the amount of time needed to finalize charging, in terms of full hours.
-The blind power usage of the charger (when connected but not charging) is high, at between 600 VAr and 700 VAr (reactive). This is a poor result which has to be improved, as it makes it costly to run V2G-charging sessions at idle mode and even at powers below 3*8A.
V2G demonstration at the ElaadNL site in Arnhem during an INVADE meeting - Foto: Nynke Arends
Experiment results: shared car vs lease car
We wanted to find out which type of car would be best suited for balancing the power grid, an electric car that is used as a shared car or one that is a leased car. We did this at the charging laboratory at the Elaad office parking lot (40 regular EV chargers, 1 V2G charger with a Nissan Leaf and 1 stationary battery). The building consumes energy mostly on weekdays between 4:30 and 20:30. Only baseload usage in the weekends.
The Nissan Leaf has known arrival- and departure times, registered in Outlook. We wrote a program that translates the scheduled appointments of the car to set point state of charge for the car and battery. In this program, the battery followed the opposite regime as the actual car, mimicking V2G cars.
When the Nissan Leaf was used for small trips during the day (departure and return during an office workday), the car was to be 100% charged upon departure and returns with unknown state of charge. Charge/discharge transactions were to be supporting the predicted consumption of the building during the period the car was available. Charging stays under the maximum capacity profile of the building, similar to how charging works in the Large Scale public pilot we did in INVADE.
The stationary battery was used as if it was a number of V2G cars operated as leased vehicles, virtually arriving between 8 and 9 am with a random battery capacity of around 30-50% state of charge and leaving between 17:00 and 18:00 pm with 100% fully charged batteries. It was our hypothesis that the V2G Shared car had more added value for supporting the grid than the leased V2G cars (simulated by the battery).
Another use case: V2G car vs stationary battery
We did another use case to test if a stationary battery has more added value for supporting the grid than a leased V2G car. The Nissan Leaf in this use case is present at the V2G-charger following a work day pattern, meaning it arrives in the morning and only leaves again at the end of the working day. A prerequisite is that the car needs to have 100% state of charge at departure. Arrival state of charge is variable. The charging and discharging of the stationary battery supports the predicted load on the parking premises. The stationary battery is operated as a battery: It is 100% full at the start of the working day, is gradually discharged to reduce load on the parking premises and is charged again at night. Our second hypothesis was that a stationary battery has more added value for supporting the grid than a leased V2G car.
Figure 2: On May 16th, the use case 1 scenario could be carried out in the morning, but there was a scheduled drive (use case 2) at 15:30 so the car responded by charging immediately after the discharge support, to supply the car with sufficient state of charge.
Conclusions: we were wrong!
So what conclusions can we draw from this used cases? The good news is: we are able to include V2G in the charging commands of a group of EV chargers and the discharge option can be of assistance to the prevent too much strain on the local grid. However, this is mostly the case when the car follows a typical leased car pattern, i.e. it arrives in the morning and leaves at the end of the afternoon. The shared car behavior results in loss of days in which V2G can be executed. In the tests the shared car suffered from the unpredictability of the drive reservations. Even though we had supplied all possible drivers with information about the procedure at hand and the possible consequences for their reservation, we experienced multiple late-notice reservations. We also experienced reservations in which the car was returned too late to be of assistance to balancing the local grid. In conclusion, our hypotheses was false: a leased car is the preferred V2G car in our tests, even though it is not 100% available at the start of the working day.
With respect to the performance of the stationary battery compared to the V2G car, we succeeded in operating the two simultaneously. The leased V2G car had sufficient battery in the morning to assist the charging, but the stationary battery was still needed. Whether a stationary battery has more added value for supporting the grid than a leased V2G car strongly depends on the amount of leased V2G cars present. Only one car will not be of significant impact, but even after driving to work or home, the electric cars have enough remaining state of charge to be of assistance. Therefore, if a company or neighborhood has a high predictability of available V2G cars, the cars can help to reduce the morning and evening peak, even if they just arrive from a drive at the time they are needed.
INVADE is a European subsidized project from the Horizon 2020 Research and Innovation program. The abbreviation INVADE is derived from the objective for the integration of electric vehicles and bateries in the distribution network for accessing distributed and centralized energy storage.
The INVADE project started on January 1, 2017 and runs until the end of 2019. In the project, 12 partners work together at 5 test locations (the Netherlands, Norway, Germany, Spain and Bulgaria). Extensive information about the INVADE project can be found at: https://h2020invade.eu/