LCOS explained with some examples

TL;DR

Levelized Cost of Storage (LCOS) is a great metric for determining per kWh cost of storage ($/kWh). LCOS has a simple and flexible formula that can be adjusted for varying degrees of complexity of the system allowing us to compare different technologies and conditions directly. LCOS can also shed light on how seemingly good deals in the residential sector are economically poor decisions or demonstrate why round-trip efficiency matters a lot when it comes to economic viability of storage.

Levelized Cost of Storage (LCOS) is a great first order guide

Since we are interested in heuristic guides for navigating the energy transition, we should be aware of some of the tools and metrics at our disposal. For storage, we have the metric Levelized Cost of Storage (LCOS), which gives us an apples-to-apples number ($/kWh) to compare the costs of different storage technologies. LCOS is like the Levelized Cost of Energy (LCOE) that many are familiar with. Roughly speaking, it is the net present value (NPV) of all costs divided by the NPV of all energy delivered. In the spirit of heuristics, the LCOS is a good first order guide to choosing the most cost-effective technology for an application.

More formally, the LCOS is given by:

\[ LCOS = \frac{Sum \medspace of \medspace NPV \medspace of \medspace all \medspace costs}{Sum \medspace of \medspace NPV \medspace of \medspace all \medspace energy \medspace delivered} = \frac{\sum\limits_{i}^{lifetime}{C_i} }{\sum\limits_{i}^{lifetime}{E_i}} \]

Where \(C_i\) and \(E_i\) are the cost and energy delivered in year \(i\), respectively. The costs may include upfront capex costs, any operating costs, and replacement costs. Perhaps the most confusing aspect is the NPV, which accounts for the fact that money is worth more today than in the future (because you could invest it). This is done using a 'discount rate' which reflects how much you value having money sooner.

The same holds for energy in the denominator, if you can sell the energy today, then you can invest that dollar earned today instead of selling the energy in a year and having a dollar then. Energy today is worth more than energy in a year.

Since the costs and energy delivered are fundamental concepts, one can include complexity like O&M costs, degradation, cycling frequency, losses, and so on into the equation to fine-tune the LCOS to a specific use case. The LCOS, however, only gives you an idea of cost, not about value or economic viability (since this requires a revenue component). And while the LCOS equation itself is quite simple, the inputs may require complexity to ascertain, for example the true cycling frequency of a battery subject to variable charging and discharging schedules.

Let’s use two examples to better understand how LCOS can be a useful tool for choosing a storage technology (or choosing to pass on storage).

Example 1: How expensive is my home battery really? Why they’re not all they’re chalked up to be

Battery storage is often an upsell for solar installers looking to make a few more dollars on their customers. “You can increase your self-consumption and reduce your energy costs” and “Batteries pay for themselves since you have free energy” are some of the promises that installers pitch. Are these promises true?

Let’s say you are in an area where solar makes sense, odds are your electricity costs are above $0.20/kWh, let’s use $0.25/kWh. In the case that you have neither net metering nor an export option for your energy, then we can say your solar energy is “free” insofar as you have no opportunity cost of using it (what it really means is that your system is oversized^[1]). This means that your storage must cost less than $0.25/kWh for the system to make sense for you, since otherwise you’d be better off without the battery and purchasing from the grid at $0.25/kWh.

Assumptions:

• $600/kWh installed cost
• 15-year lifetime
• $0 operating costs
• 2% degradation a year
• 95% roundtrip efficiency
• 365 cycles per year (shifting excess solar every day)
• 8% discount rate

\[ LCOS = \frac{\text{\textdollar} 600}{\sum_{i=0}^{14}{\lbrack 365 \times 0.95 \times (1-0.02)^{i} \rbrack \times (1-0.08)^i}} = \text{\textdollar}0.22/kWh \]

In this case you would be slightly better off with the battery since $0.22 is less than $0.25. However, in a more realistic scenario your battery won’t be cycled (or fully cycled) every day since you’ll have low solar days or low demand days. Let’s say that on average you charge/discharge your battery 200 times a year.

\[ LCOS = \frac{\text{\textdollar} 600}{\sum_{i=0}^{14}{\lbrack 200 \times 0.95 \times (1-0.02)^{i} \rbrack \times (1-0.08)^i}} = \text{\textdollar}0.39/kWh \]

In this case getting the battery is not worth it. Without doing a full analysis we could quickly gauge whether a battery would or wouldn’t make sense for us – for residential it almost never does without other revenue streams.

Example 2: Roundtrip efficiency matters when you have non-zero electricity prices

Many envision a future in which zero or even subzero electricity prices during high renewable generation periods are frequent. While exciting, it’s not clear that this will carry on indefinitely and certainly not on the scale that we would need for “free” energy (see this post). In the case of “free” energy there is an economic case to be made for storage technologies that have low capex costs even with low round-trip efficiencies (RTE). Let’s look at a cost comparison for a more realistic scenario between an LFP battery (high capex, high RTE) and iron air (low capex, low RTE).

In this case we will add the cost of energy to the LCOS extending the model to make it a LCOES (Energy + Storage) model. (This is a great example of how flexible LCOS can be.) Now we can consider the impact of RTE on costs, since having to buy more electricity to return the same kWh should be considered in the overall operating cost.

In our example we will assume:

System:

• $0.05/kWh electricity
• No degradation
• 8% discount factor
• 365 cycles per year
• 20-year lifetime
• O&M costs are 2% of capex costs

Battery A (LFP):

• 95% roundtrip efficiency
• $200/kWh installed price

Battery B (Iron Air):

• 50% roundtrip efficiency
• $20/kWh installed price

Equations

\[ LCOS_A = \frac{\text{\textdollar} 200 + \sum\limits_{i=0}^{19}{\frac{\text{\textdollar}0.05}{0.95} \times (1-0.08)^{i}} }{\sum_\limits{i=0}^{19}{365 \times (1-0.08)^{i}} } = \text{\textdollar}0.096/kWh \]

\[ LCOS_B = \frac{\text{\textdollar} 20 + \sum\limits_{i=0}^{19}{\frac{\text{\textdollar}0.05}{0.50} \times (1-0.08)^{i}} }{\sum\limits_{i=0}^{19}{365 \times (1-0.08)^{i}} } = \text{\textdollar}0.106/kWh \]

In this relatively frequent cycling example, the two batteries come out about even though the iron air battery was priced ridiculously cheap. This shows that efficiency losses can offset any advantages in capital costs and we should be careful in ignoring this.

This is also a good time to mention that LCOS can be very misleading if you don’t consider all the important factors. In the example above, if you don’t consider the RTE and just compare the LCOS based on the capex factors, iron air would look incredibly attractive with an LCOS roughly 10X lower than LFP. This is why one should be cognizant of the parameters included in the LCOS calculation.

As another aside, we modeled a high frequency cycling scenario (daily cycling). Iron-air batteries are being considered for longer term storage in which cycling frequency goes down and therefore the operating costs decrease and upfront capital costs become more dominant. Iron-air aren’t being considered for daily cycling applications, the above was an exercise to show how LCOS can deliver surprising results for otherwise seemingly simple problems.

LCOS is a robust and simple tool for quickly identifying storage opportunities

While LCOS and LCOE are similar metrics, I believe LCOS to be slightly more robust. The biggest shortcoming of LCOS for intermittent renewables is that it doesn’t consider the value of the electricity generated (higher value during peak vs mid-day). There are metrics like Levelized Avoided Cost of Electricity (LACE) that attempt to add this component but then become much more complex. LCOS doesn’t suffer from this shortcoming.

The examples above demonstrate that LCOS can be a great first order guide to whether storage will make sense or not. In cases where the LCOS is much too high given a target value, one doesn’t need to consider further analysis. This can also help guide high level investment into specific technologies. In an upcoming post, we will explore how technologies can be categorized by their attractiveness for different storage durations.