Cycle Life Needs a Finish Line: Reading Battery Aging Claims

By Daoyu Guan

Two battery brochures promise thousands of cycles. One counts shallow cycling to a defined capacity threshold. The other does not disclose its cycling window or the remaining capacity at the final measurement. The larger number is easy to remember, but the two headlines do not yet describe comparable evidence.

An export buyer needs to know what was cycled, under which conditions, how deterioration was measured, and what ended the test. An English product page should answer those questions before turning a laboratory result or simulation into a service-life promise. A cycle count without its conditions is an incomplete purchasing specification.

Separate aging from the counting convention

Battery aging can be examined through more than one mechanism. The System Advisor Model's official battery-life documentation distinguishes calendar degradation, which occurs over time, from degradation associated with cycling. Its modeling options use conditions such as depth of discharge, temperature, and state of charge. These are descriptions of simulation models, not certification criteria for a particular battery.

Those distinctions expose a weakness in the familiar calculation “cycles divided by cycles per day equals years.” The division can describe a hypothetical operating schedule. It does not, by itself, predict the battery's actual life while time, storage conditions, operating limits, and the required performance also change.

Define what the supplier counts as one cycle. A documented full-equivalent-cycle convention can help compare accumulated throughput, but it should not be substituted for the actual sequence used in an aging test. Two operating histories with the same normalized throughput need not be the same experimental protocol.

State the finish line before comparing the race

“End of life” needs an operational definition. A procurement program might choose a remaining-capacity threshold, a power-delivery requirement, or another application-specific acceptance condition. A battery can cross one threshold while still being useful for a different application. The endpoint should therefore name the performance requirement rather than imply that all batteries become universally unusable at the same percentage.

For the fictional examples in this article, an engineering team chooses an 80% remaining-capacity threshold. That number is an invented project requirement here, not a universal rule, a cited standard, or a recommendation for every chemistry. A real project needs its own justified requirement and measurement procedure.

Ask whether the advertised count means the test reached that threshold, stopped while still above it, or estimated the crossing with a model. A completed observation and a projected endpoint should have different labels. If the final report only shows that testing stopped at a stated cycle count, do not silently turn it into a measured failure point.

Preserve the conditions that make the number meaningful

Use a condition table beside a cycle-life statement. The table below is a proposed evidence request, not a prescribed battery standard.

Do not force every supplier's report into a single number before reading these fields. First establish whether the intended comparison is technically defensible. Where protocols differ, present the difference explicitly or commission the necessary evaluation; a clean chart cannot repair incomparable inputs.

A fictional capacity calculation

Suppose a fictional test record reports an initial measured capacity of 25.0 Ah and a later capacity of 23.2 Ah. Both were measured under the same reference procedure. Relative to that measured baseline, the later result is 23.2 divided by 25.0, or 92.8%. The difference is 1.8 Ah, equivalent to a 7.2% decrease from the baseline.

Now suppose the product's nominal label is 24.0 Ah. Relative to that different denominator, 23.2 Ah is approximately 96.7%. Neither arithmetic result is mysterious. They answer different questions. A report that changes the denominator without saying so can make the same observation look like a different aging outcome.

For this fictional project's 80% threshold, the capacity level relative to the 25.0 Ah measured baseline would be 20.0 Ah. The 23.2 Ah observation remains above that threshold. It does not tell us which future cycle will reach 20.0 Ah. Estimating that crossing requires further evidence and an explicitly identified method.

These numbers are a teaching example. They are not measurements of a commercial cell, a Semvi product capability, a battery warranty, or a qualified engineering prediction.

Do not flatten a changing test into a single curve

A team may alter temperature, update the charging procedure, or replace a failed measurement channel during a long program. Some changes can be properly controlled and documented. They should still be visible in the record used to interpret the result.

Ask for the protocol version and a change log covering the plotted period. A reference measurement taken under one procedure should not be quietly compared with a later measurement taken under another. The reviewer needs to decide whether the sequence can be compared directly, requires qualification, or should be shown as separate segments.

Missing observations matter too. If several units leave the program, ask why and how the summary handles them. A curve showing only the surviving units can answer a narrower question than a headline suggesting that every tested unit achieved the same life. Reporting the sample disposition is more informative than giving a smooth average with no explanation.

A model output needs its own label

SAM's battery-results reference distinguishes quantities such as state of charge and remaining capacity associated with degradation. Its outputs describe the configured simulation. A percentage of charge currently stored is not the same quantity as the battery's remaining capacity relative to an original reference.

An export manufacturer using a model should identify the model, inputs, assumptions, and purpose of the forecast. A simulated replacement year is not a field-observed replacement year. A default degradation curve is not evidence that the supplier tested the offered pack.

Before reusing a modeled chart in sales material, ask the author to explain which inputs came from the offered product and which were assumptions. Preserve that distinction in the caption. If no product-specific validation has been completed, do not describe the graph as a validated lifetime result.

This also applies when a buyer requests a scenario outside the original analysis. Recalculating a spreadsheet for a different duty schedule may be useful, but it does not establish that the original model remains applicable to the new conditions.

Keep warranty and safety claims in separate records

A cycle-life result describes a performance observation or estimate under stated conditions. A warranty describes contractual coverage and its conditions. A safety assessment addresses another set of questions. None should be inferred solely from a high cycle count.

For a purchase review, place the requested performance evidence, warranty document, and applicable safety or conformity documentation in separate rows. Link them to the same product revision where appropriate. If the seller offers a service commitment, record its actual terms rather than deriving a guarantee from the aging curve.

The distinction prevents two common editorial errors: turning a test endpoint into a promised replacement date, and treating capacity retention as evidence that all safety requirements have been met. Resolving either question requires the relevant evidence and responsible review.

Write an English statement that retains its conditions

A defensible product page gives the reader enough information to request and interpret the supporting report. For a fictional publication, a draft might say: “The reported capacity-retention curve applies to the tested cell revision under the protocol identified in report B-17. Pack-level performance and the proposed operating schedule require separate review.” That sentence is an example of scope control, not a claim about an actual product.

Put the test object, endpoint, and key conditions near the result. Do not hide the fact that a number is modeled in a distant footnote while the title calls it measured. If a short comparison table cannot carry the necessary explanation, link a clearly labeled technical note from the same row.

Xindar's manufacturing export content work can help organize such English evidence for buyers and AI-assisted research. Content production does not conduct battery qualification, validate a life model, or authorize a warranty.

A buyer's acceptance exercise

Give a reviewer one advertised cycle-life claim and the evidence offered to support it. Ask them to identify the tested object, baseline denominator, endpoint, operating conditions, population, and whether the result was measured or modeled. An unanswered field is an evidence request, not permission to invent a conventional value.

Then ask the reviewer to reproduce one published percentage from the underlying values. The calculation should use the stated baseline and units. If the marketing figure uses a different denominator from the technical report, resolve the discrepancy before publication.

Finally, ask whether the proposed use matches the evidence's scope. A reviewer can correctly reconstruct the test and still conclude that it does not answer the buyer's application question. That is a useful outcome: the next experiment or technical clarification can be specified instead of substituting a larger headline number.

Questions about battery-life claims

Is 80% capacity always the definition of end of life?

No universal threshold is established here. The 80% threshold in the worked example is a fictional project choice. State the criterion applicable to the actual product, test, contract, and application.

Can I convert a cycle count directly into years?

The division gives a schedule-based arithmetic estimate under its assumptions. It is not a complete aging prediction or a warranty. Calendar conditions and the actual operating history still require review.

Does a cell result establish the same life for a pack?

Do not assume it. Identify the object tested and obtain the evidence needed for the offered configuration. A cell graph should not be relabeled as a pack test without support.

Does a successful simulation count as a completed lifetime test?

No. A simulation and a physical test are different evidence types. Report the model and assumptions, and describe any validation separately.

Source and scope notes

The official SAM help pages support the distinctions between calendar and cycling degradation, relevant model inputs, and differently defined output quantities. They do not approve a particular product or establish a universal procurement threshold. All capacities, thresholds, report identifiers, and review exercises in this article are fictional editorial examples.