The battery still holds 70 % of its capacity. Is it still delivering for your project?

The answer usually begins with a degradation curve and a capacity warranty. A percentage, a term and, apparently, a certainty: after ten years, the battery will retain at least 70 % of its capacity.

But to know what that figure means in a project you have to keep reading. How is that percentage verified? What use does the warranty allow? And, above all, will the remaining capacity be enough to provide the service that justified the investment?

These are different questions, and confusing them can leave a significant risk out of the analysis.

What a capacity warranty measures

In the warranty conditions of a manufacturer I will call A, for its 30 and 50 kW commercial units, residual capacity is verified at 25 °C, with a tolerance of one degree, by constant charge and discharge at 0.2C, on balanced modules and with measurement on the DC side. A second manufacturer, B, sets similar conditions for one of its lines: individual module, 25 °C with a tolerance of two degrees, and charge and discharge at 0.2C.

That precision is necessary. To compare one measurement with another, a reproducible procedure is needed. The problem appears when we take that result as a description of what will happen at the plant.

A battery doing peak shaving may deliver a lot of power for short periods. One intended for backup may spend months waiting for an outage. Another replacing diesel generation may work with deep discharges and variable temperatures. Each will have a different operating history, even though all three are bought with a warranty highlighting the same percentage of residual capacity.

That is why, before incorporating the manufacturer's curve into the project model, you have to understand what it represents: how capacity is measured, under which conditions it is guaranteed and how far those conditions correspond to the intended use.

That review can change the assessment of applications that, on paper, seem a perfect fit for a battery.

Two conditions worth clarifying before buying

In the three public warranty documents I was able to review in full, I found two examples especially relevant for Latin America. In one of manufacturer B's lines, the performance warranty is conditional on use of approximately one cycle per day and expressly excludes under-use. It also requires an indoor, climate-controlled installation between 15 and 35 °C. For a battery intended mainly for backup, or for peak shaving a few hours a month, that condition deserves clarification before buying.

In the case of manufacturer A, the conditions are void if the system operates predominantly or permanently in islanded operation, with no connection to the grid or to a backup system. If the project seeks to replace a diesel generator, that clause requires reviewing the planned configuration and confirming its coverage in writing.

These examples do not allow judging an entire manufacturer, let alone the market. They do show why the warranty must be reviewed together with the operating strategy. A commercially attractive application may not be contemplated in the conditions of the chosen product.

Years, cycles and cumulative energy: coverage ends with the first

The same reading is needed with the duration of coverage. The term in years usually shares the stage with another limit: the cycles or the cumulative energy the battery may process. The warranty ends when the first is reached.

Manufacturer A, for example, sets twelve years or 6,000 full cycles, with 70 % residual capacity. Manufacturer C guarantees 70 % of the system's energy for ten years or until the cumulative energy limit stated in its table is reached, whichever comes first.

For a 15.36 kWh model from the latter manufacturer, the table lists 49.05 MWh. Dividing that energy by the nominal capacity gives approximately 3,190 equivalent cycles, while the commercial brochure advertises 6,000 cycles. That calculation does not by itself describe the whole life of the battery, but it does leave an important distinction: the cycle figure in the catalogue and the energy limit covered by the warranty are not interchangeable.

If the economic model assumes intensive use for ten years, you have to check that coverage does not run out earlier.

What the owner receives when making a claim

Even when a loss of capacity gives the right to claim, another question remains: what does the owner receive and which costs remain theirs?

In the documents reviewed, the manufacturer may choose among different remedies, such as replacing components with new or used ones, supplying additional capacity or compensating the difference financially according to the market value at the time. Recovering capacity, moreover, is not necessarily equivalent to recovering the economic result lost while the system performed below expectations.

Manufacturer A expressly excludes lost savings and lost profits, as well as reimbursement of labour and transport. Manufacturer C limits its total liability to the purchase price. The customer must also demonstrate the loss of capacity through the required procedure; if there is a discrepancy requiring an independent laboratory, they may have to advance the cost of the test.

What the contract measures, by market

The warranty is, therefore, one part of the project's protection. To understand what risk remains with the owner, it has to be compared with the obligations and revenues of the asset.

And there the analysis changes according to the market.

In the Brazilian case included in this review, the capacity contract requires a minimum round-trip efficiency of 85 % at the metering point for fifteen years, and ties the fixed revenue (receita fixa) to availability. In Spain, the scheme described remunerates firm power, with availability obligations during stress hours and periodic tests. In Colombia, for storage awarded as a grid asset, availability considers both power and available usable energy, and efficiency must be maintained over the operating life.

The Puerto Rico case makes the difference between retaining a percentage of capacity and fulfilling a service obligation especially visible. For a 100 MW and 400 MWh system, the regulator required removing the clauses for progressive reduction of available energy and maintaining the 400 MWh for twenty years, replacing or expanding components when necessary.

In a project with that obligation, knowing the expected degradation is only the beginning. Then you have to decide how the committed energy will be maintained, when intervention will be needed and who will pay for it.

When there is no contract: savings and the architecture of the equipment

In commercial and industrial (C&I) installations, that same discussion usually appears in a less visible form. When the value of the system depends on the user's savings and not on a multi-year availability contract, the loss of performance shows up on the bill. There may be no contractual penalty, but there are savings lower than expected. And those savings, as the exclusions above show, may fall outside the warranty.

Resolving this also requires knowing the architecture of the equipment. In a large-scale system, the addition of capacity over the life of the project can be planned. In many commercial units, however, the product is a closed cabinet with a fixed number of packs: expanding capacity means installing other cabinets in parallel. Replacing a pack is a different maintenance operation, with its own requirements for compatibility and state-of-charge synchronisation.

There are manufacturers that advertise more flexible designs, able to combine new and old packs or to expand capacity independently. If the project depends on that possibility, it should be documented as a feature of the contracted equipment.

Buying more capacity from the outset does not automatically solve the problem either. Capacity that sits idle ages by calendar and has a cost from day one. Whether oversizing is advisable depends on the service: maintaining several hours of autonomy poses a different need from covering brief power peaks.

As a reference, the NREL commercial storage sheet of 2022 includes in its operation and maintenance costs two capacity additions, equivalent to 20 % of the cells, in years 10 and 20. It is an assumption of that model, but it illustrates something that should appear in any assessment: maintaining performance over time requires a strategy and a budget.

The questions before signing

That is why, before signing, I would try to have these six questions answered:

1. How is the guaranteed capacity measured, and how does that measurement relate to the service my installation needs?
2. With the intended use, which limit will I reach first: the years of coverage, the cycles or the cumulative energy?
3. Does the warranty expressly allow my operating mode, including periods of low use or islanded operation?
4. If capacity falls below the guaranteed level, what remedy can the manufacturer offer and which costs remain mine?
5. How will that loss of capacity be verified, who will keep the records and who will pay for a possible dispute?
6. What power, usable energy and efficiency does the project need to maintain to meet its commitments or produce the expected savings?

The answer to the opening question is not only in the manufacturer's curve. It is in the relationship between that curve, the warranty, the planned operation and the performance that sustains the value of the project.

A battery can retain 70 % of its capacity and comply with its warranty. What we need to know before buying it is whether, under those conditions, it will still deliver for us.

Limits of this note

This review is based on three public warranty documents that I was able to consult in full. Their conditions do not represent the whole market and may differ from those negotiated for a specific project. Manufacturer names are omitted because the purpose is to examine the method of evaluation. The NREL reference corresponds to 2022; the analysis of the Spanish case does not include an estimate of the firmness coefficient.

Douglas Rosales Baptista, founder and technical director of Kyklios. Civil engineer with a master's degree in solar energy. On the owner's side, he has led in Portugal the hybridisation with batteries of 624 MWh on an operating solar portfolio, and in Spain the development of 226 MWh of storage to hybridise an existing portfolio.

Sources