Four months in, SA Tesla battery is showing mixed results in energy arbitrage
April 5, 2018
Tesla’s big battery in South Australia was officially switched on in late November 2017. At 100 MW/129 MWh, it is the largest lithium ion battery in the world and signals the beginning of a new era in how we manage the electricity system. In this article, we take a look at how the battery has operated in its first four months.
Before we delve any further, it is important to note that the Tesla battery is really two systems:
- 70 MW/39 MWh is contracted to the South Australian state government for the purpose of providing grid stability services;
- The remaining 30 MW/90 MWh can be used by the Hornsdale operator to trade in and arbitrage the energy market.
The economic viability of large scale batteries is largely dependent on how much revenue they are able to extract from the wholesale energy market and the various ancillary service markets. In this article, we will focus on the economic performance of the battery in the energy market only.
We use a software called NemSight by Creative Analytics (part of the Energy One group) to analyse the operation and bidding behaviour of the Tesla battery.
Tesla battery is being heavily utilised
The Tesla battery is certainly not sitting idle in the market. Figure 1 shows the operation of the Tesla battery on a five minute basis from 1 Dec 2017 to 31 Mar 2018. As can be seen, the battery is being utilised very frequently. In fact, in 63% of the dispatch intervals, the battery was either being charged or discharged. Only 37% of the time was the battery not being used in any way. Furthermore, nearly 40% of these zero utilisation periods occurred in December while the battery was still being tested.
From Dec 2017 to Mar 2018, the Tesla battery consumed an average of 116 MWh per day for charging. In contrast, it delivered an average of 94 MWh per day back into the grid. From this we can work out that the average efficiency of the battery has been 82%. As expected, the efficiency of the battery is lower under real world conditions than the spec sheet efficiency of 88%, which is calculated at 25°C.
Tesla battery made $1.4 million in the energy market, but is losing money 47% of the time
Figure 2 shows our estimate of the value that the Tesla battery received from selling electricity into the energy market versus the cost of buying electricity to charge the battery. We estimate the total net revenue from the energy market to be just under $1.4 million. The overwhelming majority of this came in January 2018 when the energy market experienced the highest volatility. In contrast, the Tesla battery barely made any money in December and March.
When arbitraging the spot market, the aim of the game is the same as trading stocks: buy low, sell high. A crucial difference is that you not only have to sell at a price which is higher than your buy price, but you also have to cover the cost of the extra energy that is needed to charge the battery (because energy efficiency is less than 100%). Therefore, the days with the most volatile pricing offer the biggest opportunities for arbitrage. Figure 3 shows that the Tesla battery made 95% of its net revenue in just five (very volatile) days.
If we exclude these five days, the average net revenue for the battery is a measly $530 per day. In fact, on 57 days (47%) the Tesla battery actually lost money in the energy market. The total losses over the 57 days add up to about $135,000. As mentioned earlier, this analysis looks at the revenue from the energy market only. It does not include any of the revenue received from providing ancillary services. Nonetheless, our numbers do suggest that the operators of the battery will need to be careful to avoid needlessly cycling the battery for little financial gain. This is an important consideration because the lifetime of a battery is strongly related to how many times it is cycled.
Furthermore, “buy low, sell high” is not as easy as it sounds. The energy market is incredibly complex. In our experience, a successful bidding strategy needs to be underpinned by advanced predictive analytics and co-optimisation. Failure to do so can result in the asset significantly undershooting revenue expectations. Some food for thought as the race to build big batteries begins.