Cycle aging

What it is (precise)

One full equivalent cycle is discharged energy equal to the rated energy — one complete charge-discharge cycle — regardless of how the swings are distributed. A battery management or EMS layer counts FEC either by dividing cumulative discharged energy by rated energy or by rainflow counting, an algorithm that decomposes irregular partial swings into equivalent full and half cycles.

Every FEC consumes a small, permanent slice of capacity. The dominant physical mechanism in LFP cells is loss of cyclable lithium to the growing solid-electrolyte interphase (SEI) on the graphite anode, accelerated by mechanical stress from particle expansion and contraction during lithiation; lithium plating at high charge rates or low temperature and active-material cracking add to it.

The observable results are Capacity fade in usable Ah and a slow rise in DC internal resistance, which also erodes Round-trip efficiency over life as I2R losses grow. Cycle aging is distinguished from Calendar aging, which proceeds even when the system sits idle and is driven by time, temperature, and average state of charge.

A real grid-scale battery experiences both simultaneously, and a credible degradation model superposes the two. The relative weight depends on duty: a daily-cycling solar-shifting asset is cycle-dominated, while a rarely dispatched capacity or backup resource may be calendar-dominated.

Cycle aging is not linear over the whole life. Early life typically shows a fast initial fade of a percent or two as the SEI stabilizes, followed by a long, roughly square-root-of-throughput or near-linear regime, and in some cells an accelerating "knee" late in life where plating and electrolyte depletion compound. Warranty degradation tables are drawn to stay above the expected curve including that knee, which is why year-15 or year-20 guaranteed values look conservative relative to a naive linear extrapolation of the first years.

Why it matters in a real grid-scale project

Capacity guarantees are written against usable energy, referenced at a stated measurement point — the DC block, the BESS AC terminals, or the point of interconnection net of conversion losses, depending on the contract — not against cell nameplate.

The supplier's degradation curve sets how that guaranteed energy declines year over year, and the gap between guaranteed and required capacity is closed by Augmentation — adding racks or containers in later years — or by oversizing on day one, commonly 5-15% of DC energy. Cycle aging is therefore a direct capital and contractual driver: a tolling agreement that demands 400 MWh at year 15 forces the developer to model exactly how many FEC per year the offtaker's dispatch will consume.

Operating choices feed back into the aging rate. Tight throughput limits, moderate C-rates (a 4-hour system cycles at only 0.25C, one reason long-duration LFP ages slowly), and avoiding full 0-100% swings extend life, but they can conflict with the revenue the dispatch strategy is chasing in energy arbitrage or ancillary markets.

Most warranties cap annual equivalent cycles — often around 365 FEC per year for a one-cycle-per-day application, with 500-730 FEC per year available at a price for heavier duty — plus a lifetime throughput cap in MWh. Exceed either and the Capacity warranty can be voided regardless of measured health, so the EMS and the trading strategy must be cycle-aware in real time.

Typical values and standards

End of Life for stationary BESS is conventionally set at ~70% of beginning-of-life capacity, though the exact threshold is contractual. Modern LFP cells are typically rated for roughly 6,000-10,000 FEC to that threshold, with recent 300-500+ Ah stationary cells advertising 10,000-15,000 cycles under favorable test conditions — an average fade in the stable, post-break-in regime on the order of only 0.002-0.005% per FEC.

NMC, used mainly where energy density or footprint dominates, generally tolerates roughly 3,000-5,000 FEC and is more sensitive to high state of charge and temperature. By contrast, EV batteries are designed for closer to 1,000-3,000 cycles — several times below stationary cycle-life ratings — because a car never accumulates grid-scale throughput.

Vendor cycle-life figures are always conditional on a specific test point — commonly 25 °C, 0.5C charge and discharge, 90-100% depth of discharge, cycled continuously without rest. Real projects run at different temperatures, partial DoD, and with idle periods, so the datasheet number transfers only through a degradation model, never directly.

On the standards side, IEC 62933-2-1 covers performance parameters and test methods for electrical energy storage systems, and IEC 61427-2 addresses on-grid duty-cycle testing; acceptance follows the project's contractual capacity-test procedure built on such guidance. These are separate from the safety regime — UL 9540 system certification, UL 9540A fire-propagation test data, NFPA 855 installation requirements — which governs whether the system can be sited, not how it ages.

How it shows up in specs, studies and contracts

On a cell or DC-block datasheet, look for the cycle-life claim and, critically, its footnote: the DoD, temperature, C-rate, and end-condition it was measured at. Ask the vendor for the underlying cycling data and for curves at more than one condition, because a single 6,000-cycle number at 25 °C and 0.5C says little about a site in a hot climate running two cycles a day.

In the supply contract, cycle aging appears as the degradation table in the warranty annex: guaranteed capacity by year for a stated cycling profile, with FEC-per-year and cumulative-MWh caps, temperature envelopes, and SOC dwell restrictions as conditions precedent.

In operations, cycle aging is tracked through periodic capacity tests — typically annual or biennial full charge-discharge tests at an agreed C-rate and temperature window — compared against the warranty table; testing downtime itself interacts with the Availability guarantee, so test windows are negotiated.

In revenue modeling, the cycling assumption drives LCOS and augmentation timing: an arbitrage case at 365 FEC per year consumes warranty cycles at exactly the standard cap, while frequency-response duty may register far fewer FEC despite constant activity, because shallow swings rainflow-count to fractions of a cycle. The practical questions on any project: how are FEC counted, what dispatch profile underlies the degradation curve, and who pays if actual duty deviates.

Common pitfalls

The most common error is comparing cycle counts across incompatible bases. A 10,000-cycle claim at 80% DoD to a 60% EOL threshold is not better than an 8,000-cycle claim at 100% DoD to 70%: 10,000 × 0.8 is only 8,000 FEC-equivalent, and to a laxer end condition — so normalize DoD, end condition, and temperature before comparing bids.

A second trap is conflating cycle count with FEC: two 50% swings equal one FEC, so a battery "cycled twice a day" at half depth ages far more like a once-a-day full cycler than twice as fast — often somewhat gentler, in fact, since shallow swings typically do less damage per FEC, which is why DoD must be normalized when comparing claims.

Third, accelerated lab tests cycle back-to-back with no calendar time; a 20-year field life adds calendar aging on top, so field fade is always worse than the cycling curve alone. Finally, the usable SOC window the BMS displays already hides margin — degradation is measured against contracted usable energy, and the reference point (DC terminals versus point of interconnection) must be stated in the test procedure, or the warranty capacity test is unenforceable.