Vehicle-to-Grid (V2G) technology enables Electric Vehicles (EVs) to discharge power from their batteries into electricity grids. Since its conception in 1997, V2G has been motivated by three sustainability goals, as a way of (1) providing the energy storage capacity required to “facilitate large-scale integration of intermittent-renewable energy resources” (Kempton & Letendre, 1997), (2) reducing material consumption by “using electric vehicles when they are parked and underutilized” (Kempton et al., 2008), and (3) by ”foster[ing] EV adoption ... add[ing] a revenue stream for EV owners” (Chen et al., 2020).
There are many avenues by which V2G could contribute to these goals. These include supporting power systems by storing renewable energy generation (Hemmati & Mehrjerdi, 2020; Pillai & Bak-Jensen, 2011; Tarroja et al., 2016) and by providing frequency regulation services to balance power supply and demand on a second-by-second basis (Black et al., 2018; DeForest et al., 2018; Han et al., 2010; Sortomme & El-Sharkawi, 2012). Value streams for EV owners include revenue from energy market arbitrage (Kern et al., 2022) and electricity bill management (Erdinc et al., 2015; Ortega-Vazquez, 2014; Thompson & Perez, 2020). These applications have all been demonstrated in modelling, laboratory testing and field trials, as well as being the subject of social research (Lucas-Healey et al., 2024; Lucas-Healey, Sturmberg, et al., 2022; Sovacool et al., 2018; Sovacool & Hirsh, 2009).
Another potential application of V2G EVs is as the reserve capacity needed as a backup in case of unforeseen power system contingencies, such as storms downing power lines or large generators tripping offline. This mechanism is illustrated in Figure 1. It has been envisioned since at least 2005 (Kempton & Tomić, 2005) and fits well with the general observation that V2G is better suited to creating value through power applications rather than energy storage applications (Thompson & Perez, 2020) In Australia the provision of this service is remunerated on the basis of potential power capacity made available for this service entirely irrespective of how much (or little) the power reserves are called upon. While this application has been demonstrated in laboratory tests (M. M. Haque et al., 2021) it has, to the best of our knowledge, not been demonstrated in a real-world grid contingency.
This paper reports on the response of 16 V2G EVs to a contingency that occurred in the Australian national grid on the 13th of February 2024 which caused 90,000 customers in Melbourne to lose electricity supply (Preliminary Report-Trip of Moorabool-Sydenham 500 KV No. 1 and No. 2 Lines on 13 February 2024, 2024). The EVs are part of a fleet of 51 Nissan LEAF vehicles that deployed as part of the Realising Electric Vehicle-to-Grid Services (REVS) project, together with 51 bidirectional chargers (Lucas-Healey, Jones, et al., 2022) and charge control systems designed to provide frequency support services (M. M. Haque et al., 2021).
At the time of the contingency the 16 EVs were plugged in to chargers at six properties across Canberra, about 500km from the location of the downed transmission lines that caused the contingency. Figure 2 presents the power system frequency and power imports recorded on high-speed power meters at the six properties (25ms time resolution, frequency data is averaged across the properties while power is summed). It shows that the contingency underfrequency trigger of 49.85Hz was crossed at 13:09:43, at which point the power imports of the properties quickly began to decline from an average of 799.9kW in the minute before the event to 699.0kW at 13:09:57 and 662.9kW 13:11:18. Such a decrease in power demand is precisely what is needed to correct the power supply-demand imbalance evidenced by an underfrequency (see Figure 1).
This decrease in import power exceeds the combined discharging capacity of the 16 chargers, which is limited to 80.0kW due to a 5.0kW export limit being imposed on each charger to meet the phase balancing requirements specified by the distribution network operator (Lucas-Healey, Jones, et al., 2022). Figure 3 shows the charging powers of the 16 EVs, as recorded by the chargers (at a lower temporal resolution of 30 seconds). This reveals that an additional 27.1kW of the observed decrease in property import power can be attributed to the termination of charging by four V2G vehicles as their first step in responding to the contingency. This data furthermore confirms that all 16 vehicles discharged at their (constrained) maximum of 5kW within 60 seconds of the frequency trigger (power ramping being limited by the implementation of Australian Standard 4777.2:2020 (M. Haque et al., 2022)) and remained at this power level for ten-minutes, as specified by the Australian National Electricity Market rules for contingency frequency services (Australian Energy Market Operator, 2017). The total impact on the V2G response was therefore 107.1kW, which is consistent with Figure 2 – accounting for the variability in the properties’ other loads.
Taken together, these results demonstrate the success of V2G in providing a high quantity of high-speed, high-quality frequency response services to the power system, and doing so from assets (EVs and chargers) that were otherwise idle (or making the power supply-demand balance worse by charging).
However, the data also reveal shortcomings of the current V2G implementation (and the market driven approach to frequency control that they are designed to). This can be seen in the chargers’ behaviour immediately after the ten minutes of required discharging, when nine chargers began to charge at their maximum power of 6.3kW. This led to a power swing of 129.1kW in two minutes from 13:19 (from –80.04kW to 49.07kW), of which 99.3kW occurs in the minute from 13:20 (from –49.6kW to 49.07kW). These results would be worse were it not for a software bug that led two chargers to continue to discharge (for two and ten minutes).
While this behaviour is facilitated by the specifications of the frequency services market, the addition of load to the power system – particularly so rapidly – is antithetical to the recovery of the power system from a contingency. In this case, the EVs added load before the power system frequency had recovered to 50Hz (at 1:23 (WattClarity, n.d.)). This highlights an opportunity for improving power system security through better integration of EV charging, especially as delaying the charging of these EVs for a while longer would almost certainly have had no impact on their fleet use and would have avoided the carbon emissions of gas peaker plant that tend to be called upon during such contingencies.
This point is further emphasised by considering that there were 23 non-V2G enabled EVs plugged in at the 6 properties during the event, which were charging at an average of 30.8kW. Stopping this charging, together with the charging of the four V2G EVs that were charging at the outset of the contingency, would have created a contribution of 57.9kW to power system frequency restoration without the use of V2G. While this is less than the 80.0kW contribution from V2G discharging, it illustrates the potential to make very substantial contributions to power system security from well managed unidirectional (non-V2G) EV charging.
Another way to illustrate this is that the blackout imposed on 90,000 customers is equivalent to stopping 6,000 EV’s charging at 5kW. For society, one of these is a far preferable option. Thinking further ahead, once the vehicle fleet is fully electrified, it would only require 2.25% of the 6.2 million vehicles in the state of NSW and Australian Capital Territory (which is a part of the NSW power system) to stop charging at 5kW to provide the 700MW of flexible load required to (on average) cover all frequency contingency services.
In conclusion, our results demonstrate – for the first time in a real-world power system – that V2G can effectively respond to frequency contingencies to meaningfully contribute to power system security (in other words, to ‘keeping the lights on’). While the actions of the 16 V2G EVs during this one event reveal avenues for improvement in this particular V2G implementation and frequency control market, the most significant issue raised is the opportunity to for non-V2G EVs to contribute to a more secure and sustainable power system.