State of Health SOH

What it is (precise)

Capacity-based SOH is only meaningful against a stated reference: capacity measured at a specified C-rate and temperature (commonly around 0.25C to 0.5C at 25 °C), over a defined voltage window, compared to the rated BOL value under the same conditions.

The basis also matters — a cell-level DC number is not the same as usable AC energy at the Point of Interconnection, which additionally reflects PCS, transformer and auxiliary losses. In operation, the Battery Management System estimates SOH continuously by tracking coulombs in and out, anchoring against full-range charge events, and watching each Cell's impedance signature; contractual SOH is then verified by discrete capacity tests.

Two mechanisms drive the number down. Calendar aging is passive degradation from time, temperature and resting SOC — a battery fades even sitting idle. Cycle aging is loss of cyclable lithium and active material from throughput, scaling with Depth of Discharge, C-rate and temperature per Cycle.

Because a utility-scale plant contains hundreds of thousands of series-parallel cells that age unevenly, the weakest strings tend to set the effective SOH of a Rack or container; this is why cell balancing, Module-level telemetry and string-level diagnostics matter so much, and why fleet SOH is a distribution, not a single clean number.

Why it matters in a real grid-scale project

SOH is the single number that ties the asset to its contracts. A typical supply agreement or tolling/offtake structure guarantees that the system will still deliver a minimum usable energy at the POI for a set number of years. To honor that, projects are deliberately overbuilt at day one — often 5-15% extra nameplate, sometimes more for heavy cycling duty — so that even at end-of-warranty SOH the contracted MWh is still met. Augmentation, meaning adding fresh racks or containers in later years, is planned against the projected SOH curve and is a major line item in the financial model.

Getting SOH wrong has direct commercial consequences. Optimistic degradation assumptions lead to under-sizing, missed capacity or availability payments, and liquidated damages; pessimistic assumptions waste capital on oversizing and augmentation that never pays back.

The measurement methodology is therefore negotiated into the contract: how capacity is tested (a controlled discharge under specified temperature and C-rate), how often, on what basis (DC or AC), and who pays if the result falls short. Round-trip efficiency decline and SOH are related but not identical — rising internal impedance degrades both, but a system can meet its RTE guarantee while missing its capacity guarantee, and vice versa.

Typical values and standards

Industry practice treats roughly 80% capacity-based SOH as the common nominal end-of-warranty threshold for stationary Lithium Iron Phosphate systems, though many assets remain operable well below that and contracts increasingly specify retained MWh year by year rather than one EOL number.

Annual capacity fade for well-managed LFP grid systems is typically on the order of 1-3% in early life, tapering as the fade curve flattens; the trajectory depends heavily on thermal management, resting SOC discipline and cycling duty. Vendor datasheets commonly quote 6,000-10,000+ cycles to roughly 80% retention for grid-duty LFP, always at a stated temperature, DoD and C-rate.

Chemistry shapes the curve. LFP, dominant in stationary BESS, degrades gradually and predictably, which is a core reason it displaced NMC in grid applications despite lower energy density; NMC generally shows a steeper fade under high-throughput duty.

Sodium-ion, the emerging stationary alternative, is marketed partly on cycle life and wide-temperature tolerance, but long-duration field SOH data at grid scale is still thin. Whatever the chemistry, the vendor's warranted degradation curve — not a generic literature value — is the authoritative sizing input, and prudent owners validate it against measured fleet data over time.

No single global standard sets an SOH limit; the number lives in the supply and performance contract. The surrounding framework, however, is codified: IEC 62620 provides performance and capacity test methods for industrial lithium cells; IEC 62619 and UL 1973 cover safety of industrial cells and stationary battery packs/racks respectively; UL 9540 is the safety certification for the complete ESS product; UL 9540A is the fire and thermal-runaway propagation test method whose data feeds NFPA 855, the installation standard.

These documents do not prescribe an SOH threshold, but aged, high-impedance cells run hotter, so SOH tracking is part of responsible operation and end-of-life planning under them.

How it shows up in specs, studies and contracts

On a cell or product datasheet, SOH appears indirectly as cycle-life tables (cycles to a stated retention, at a stated DoD, C-rate and temperature) and calendar-life claims. In the supply agreement it appears directly as a capacity maintenance guarantee: a year-by-year table of minimum retained energy, usually conditioned on throughput caps (MWh or equivalent full cycles per year), an operating temperature envelope and an SOC operating window.

Exceeding those usage conditions is the standard route by which vendors escape warranty claims, so the operating profile assumed in the revenue model must match the profile the warranty permits.

In execution, SOH shows up as test events: a commissioning capacity test that establishes the BOL baseline, then periodic (often annual or biennial) retests under the contractual procedure.

Read the test protocol as carefully as the guarantee itself — allowed temperature band, rest periods, charge procedure, measurement point (DC terminals versus AC meter) and how auxiliary consumption is treated can each move the result by whole percentage points. Interconnection and capacity-market filings also embed SOH: a resource that qualifies for capacity payments must show it can sustain its committed MW and duration as the fleet ages.

Practical questions to ask on any project: What is the SOH basis — DC or usable AC at the POI? How is a Cycle defined for the throughput cap, and at what reference DoD? What does the year-20 (or contract-end) retained-energy table say, and does the sizing overbuild cover it? Is the BMS-reported SOH contractually meaningful, or only the witnessed capacity test? Who owns augmentation risk — is it a vendor capacity-maintenance obligation or an owner cost? Answers to these determine whether an 80%-at-year-X promise is bankable or decorative.

Common pitfalls

The most frequent confusion is treating SOC and SOH interchangeably; the second is mixing bases. A vendor's 95% SOH figure at DC cell terminals, a BMS-displayed value, and retained usable AC energy at the POI are three different numbers, and warranty disputes regularly hinge on which one the contract actually references. Similarly, nameplate, usable and contracted energy diverge over life: displayed SOC windows are often narrowed or re-mapped as the battery ages, which can mask capacity fade from casual observation while the underlying absolute capacity keeps shrinking.

BMS-estimated SOH also drifts. Coulomb-counting accumulates error, and systems that rarely traverse a wide SOC range give the estimator few anchor points, so the online figure can diverge from tested reality by several percent between capacity tests. Finally, do not read 80% SOH as a physical cliff: it is a contractual and planning convention, not the point where cells fail. The real end-of-life decision weighs augmentation cost, energy value, footprint and safety posture — many LFP fleets are economically worth operating, augmenting or repurposing well past the nominal threshold.