Would Grid Battery storage works, financially, in Singapore?
Disclaimer: All information / data presented here are publicity available data and of my opinion to the best of my understanding. This does not represent the comment of any companies I have worked for, or with.
Grid scale battery operates with different grid economics. To understand the economics of grid scale battery, we first must understand the general purposes of grid-scale battery. In general, the two main propositions are to support the grid with a better power quality, or to harness the wholesale electricity price arbitrage.
In the context of Singapore, the issue of grid power quality is well maintained by Singapore Power. Here, I will focus on the arbitrage of wholesale electricity market. To be more specific, I will try to answer these two questions: (a) does it make commercial sense to install grid-battery storage, and (b) does it make investment sense for consumer as on-site off-peak energy storage.
The Battery
Money comes first. The following table summarizes the properties of Tesla’s Powerwall 2.
Recently, Tesla has adjusted the price of PowerWall to better reflect market values. From the article, it is evident that the cost / kWh used here reflects the extreme lower end of the spectrum.
The Container
To translate grid-battery into physical space, we will use a standard shipping container dimension. Let’s assume that the grid-scale battery will be shipped in a standard-issued shipping container.
Assuming that we can pack the batteries up to 80% of the size of the container. This means that a standard container would ship with 13 Tesla Powerpack 2 batteries with the following specification.
For the ease of calculations, we presume that any financial projections of the batteries can be met technically. This essentially disregards the technical limitation of the batteries and at its connection point. For example, the extreme case of financial calculation may require a 1000 kWh battery to be discharged in 1 minute, which pushes the export power to 1000 x 60 = 60 MW. This is not possible due to physical engineering limitations. For simplicity, we will presume that such energy discharge is possible.
We also assume that the battery is 100% efficient. This means that the battery does not suffer any loss in energy during the charging and discharging process.
Update 2 Nov 2018 — The grid battery used in Singapore’s microgrid system is 200 kWh, at half the volume of the standard container size. So you know I actually oversize the container in an attempt to make it feasible.
The Inverter
Batteries runs on direct current (DC), while electricity grid runs on alternating current (AC). An inverter is needed to convert DC to AC. Let’s assume that Huawei Inverter is used.
We will need approximately 62 inverters to meet the battery output of 2600 kW. This translates to S$ 434,000 in CAPEX. Note that I simplified the engineering assumptions drastically. Life is complicated enough, let’s not make it worse. I got this price off Google but I’ve lost the source. It was from a European B2B e-commerce website.
This means that it would cost S$ 2 million, minimum, to build a container sized portable battery storage system. Pretty pricey.
The Australia Tesla System
To validate the costs, we use the Australia grid connected battery as benchmark for battery system costs. Tesla has commissioned a 100 MW grid-scale battery system in Australia. Using the public source of data.
The equivalent cost for a 2600 kWh of battery system would be S$ 2,496,000. The cost is approximately half a million higher than our estimates as the we did not include engineering costs, consultancy costs, administrative costs, etc. Not to mention that the notoriously high labour cost in Australia. We now know that our estimate is decently close enough.
As the container-sized Tesla battery is a 2,600 kWh battery, the linear scale of annual recurring cost would be S$ 110,760–141,085 per year.
Again, just estimations.
The Fundamental of Trading
The fundamental of trading is a simple economics concept — buy low sell high. The following chart shows the electricity price over the course of one week. While there are fluctuations from time to time due to various reasons, in general it remains relatively stable, with occasional spikes that doesn’t last long.
Recently, the USEP has been more erratic.
The blackout occurred on 18 September 2018 in Singapore sends the USEP price to all time high of S$ 1354.60. USEP remains relatively high for several weeks after the blackout. Although this is an undesirable condition, it will boost the economics of grid-scale battery.
In addition, electricity consumers would still be required to pay for transmission and distribution charges at the point of purchase. For illustration purposes, you can refer to this sample bill to understand how the charges are incurred to consumer.
The 3rd party charges varies at different voltage level.
Grid scale battery is not likely to be connected at 66kV. Therefore, we will use the case of high tension — large transmission charges and low tension charges for simulation.
The Price Arbitrage
The price arbitrage represents the profit for the purchase and sale of electricity. We will start with a simulation that would be unrealistic — purchase wholesale electricity instantaneously at the lowest price and selling it at the highest price over the 24 hour cycle.
For the purpose of confidentiality, let’s just imagine a black box that purchases electricity automatically, and it guarantees that the best electricity deal will be made. Obviously, the algorithm cheated the system. If cheating doesn’t help making the business case, we can throw the business case out of the windows.
Using the actual USEP price from 20 July 2018 to 20 October 2018, the following table shows the transactions that would have taken place. Each row represents a buy-and-sale transaction. Due to the limitation by medium, I’m cropping it piece-wise.
I digress: Despite rising USEP, it is still possible to make daily profit if you are doing it right.
Simplifying the table, the follow chart shows the price arbitrage for every transactions for the past 3 months.
Generally, the energy procurement point occurs during off peak and selling point during peak hours. Some exceptions do happen. The following table shows the monthly gross profits. We can see that the first month is the stable month, which is the expected revenue under normal operating circumstances. The second month is the disaster month (due to the profits generated from blackout). The third mouth is post recovery natural disaster month.
The sum of transaction demonstrates the arbitrage profits in Singapore dollar per MWh. For a 2,600 kWh, or 2.6 MWh system, the profit is a straight forward calculation. The system would have made a profit of S$ 9,408 / MWh x 2.6 MWh = S$ 24,461 over 3 months. The simple forecasted annual profit would be S$ 97,843.
This is just the gross profits off the USEP arbitrage, to fully understand the business model, the transaction costs must be accounted for.
The Economics
There are two scenarios for consideration — grid scale battery arbitrage and on-site energy storage facility. The simulation only considers the trading activity, but neglected the associated transaction costs. The applicable transaction costs include Use of System (UoS) charges and contracted capacity charges. These two scenarios will treat the transaction costs differently and provide different financial outcomes.
Fun fact: if every month and every year is Spooky October 2018, the annual revenue would be S$192,551. This makes the simple payback of 10 years.
(A) Grid Scale Battery
Supposed a company is set up with the purpose of making money by buying energy from the grid when the price is low, and selling it back to the grid when the price is high. That means the company needs to ensure that the price arbitrage is higher than transaction costs, and that it would be sufficient to justify its investment.
The 2600 kWh system would have generated an annual revenue of S$ 97,843. The estimated initial investment cost is S$ 2 million. Hence, the estimated simple payback time is 20 years. This excludes financing cost, annual maintenance costs, operation costs, and many other applicable costs. Without putting in the transaction costs, it is clear that there is simply no business case for any company to operate a grid arbitrage company in Singapore. At least I wouldn’t put my money there.
With this level of annual revenue, it couldn’t even afford Tesla’s annual maintenance cost. Without the need to factor in the transmission and distribution charges, it is a simple conclusion that a grid arbitrage battery company would file for chapter 11 eventually.
Fun fact: if every month and every year is Spooky October 2018, the annual revenue would be S$192,551. This makes the simple payback of 10 years.
(B) Grid Battery as On-Site Energy Source
In this consideration, the grid scale battery is used to purchase energy at lowest price of the day, and consume the energy during peak prices to reduce energy purchase from wholesale electricity. The battery can be either owned by the consumer themselves or third party vendors. The major difference is that the consumer would be paying for lower Use Of System (UoS) charges when purchasing electricity at night and paying lower unit energy costs.
Here we will consider the case of High Tension Large customer, and a 400V low tension customer. We will start by calculating the savings from UoS. The following table summarizes the differences of two connection types:
For the high-tension customer, the on-site battery supply can reduce peak demand. The immediate financial benefit would be the reduction of contracted capacity, which is a monthly payable fee to Power System Operator (PSO). The battery system can supply a continuous power of 60kW, which we can safely assume that a 60kW reduction in contracted capacity.
Putting everything together:
Frankly speaking, I wasn’t expecting the annual savings for HT and LT to be that similar. Nevertheless, we aren’t seeing much difference in terms of savings due to the highly efficient and affordable grid charges.
Conclusion
Based on the current market arbitrage and the prices of batteries, the simple simulation demonstrate the potential returns from the installation of container-sized grid battery. There are two scenarios for consideration — the first involves the sale of electricity back to the grid, and the second scenario for self consumption. The economics and considerations for both scenarios are clearly defined.
From the simulation, it is clear that operating a grid battery arbitrage storage company would not make good investment
From the simulation, it is clear that operating a grid battery arbitrage storage company would not make good investment. Even considering the short term severe wholesale electricity price fluctuation and forecast it to long term, the revenue generated would not be sufficient to pay the installment for the installation of grid battery.
The business case for individual businesses to purchase electricity at night for peak consumption is feeble at best. Even with the benefit of use of system, the current price arbitrage at wholesale electricity price does not make grid battery system financeable. The returns of the grid-battery storage are highly dependent on the nature of business and the engineering consideration. While the initial numbers looks discouraging, I strongly suggest further investigating this possibility before giving it a complete dismissal.
One possible investigation is to understand the level of price arbitrage required to make the business case. This requires additional input, such as the O&M costs, IRR, depreciation rate, etc. The estimation would result in a go-no-go grid arbitrage price, that would determine the economics and commercial viability of the grid project.
Simply put, there are ways to make the business case possible, you just gotta know how.
I’m certain this article would save some companies valuable time in evaluating the business cases in Singapore. You can thank me by buying me a beer =).
If you have any comment / questions, you can reach out to me at hanwen.low@gmail.com. If you want to be notified of the next update, you may drop me an email as well.
