Battery Degradation Modeling for Solar: Inputs, Forecasts and Warranties

Answer: Battery degradation modeling estimates changes in capacity and power capability from time, operating conditions and cycling history. A useful solar-storage model separates calendar and cycling effects, defines its capacity reference and uses data relevant to the installed battery. Forecasts need uncertainty and operating assumptions; a warranty threshold, cycle count or chemistry label is not a site-specific lifespan prediction.

A battery that meets a load at commissioning may provide a different usable energy window years later. That matters for backup duration, energy shifting and replacement planning. The task is to make those changes traceable to evidence rather than attach a standard annual fade percentage to every solar battery.

Define what is degrading before choosing a model

Capacity fade reduces the charge or energy available under specified test conditions. Resistance growth can affect voltage response, heat and deliverable power. They are related aspects of aging, but a single capacity percentage does not describe every operating limit.

Write down the measurement boundary: cell or complete battery, Ah or kWh, nominal or usable capacity, DC battery terminals or AC output. Also record the reference temperature, discharge rate, voltage limits and permitted SOC window. Changing a control reserve can reduce delivered energy without demonstrating permanent cell degradation.

For capacity-based State of Health, use a consistent reference:

Capacity SoH = measured capacity ÷ reference capacity × 100.

For example, 160 Ah measured against a 200 Ah reference is 80% capacity SoH if both measurements use comparable conditions. It does not prove that the battery has 20% of its service life left, nor that AC backup energy is exactly 80% of its commissioning value.

Operational SoH estimates and controlled capacity measurements have different limitations. NLR describes both diagnostics and model-based estimation in its battery lifespan research. Treat a BMS estimate as an estimate with a method and confidence range, rather than a laboratory capacity test.

Calendar aging and cycling aging need separate inputs

Calendar aging occurs with elapsed time, including periods without useful charging or discharging. Temperature and SOC history can affect the relevant mechanisms. Cycling aging depends on the operating history, including depth of discharge, current and temperature. Their relative importance is specific to the cells and duty cycle.

A solar self-consumption battery, a standby backup battery and a frequently dispatched commercial battery can have different histories even when their nameplate energy is identical. Preserve those histories in the model instead of assigning a chemistry-wide annual loss.

Useful input records include:

• Time at different battery temperatures, including thermal management behaviour.
• SOC trajectory and time spent near operating boundaries.
• Charge and discharge current, event depth and duration.
• Delivered energy and the meter location used to measure it.
• Commissioning settings, later reserve changes, outages and replacements.

NLR’s BLAST modeling framework uses operating and thermal histories to evaluate battery aging. Its models are parameterized from data; the existence of a modeling framework does not establish an accurate forecast for an arbitrary battery.

For load-to-current calculations, use the separate battery C-rate selection guide. Current-based C-rate and an energy-based power ratio need consistent capacity and voltage assumptions.

Select a model supported by relevant data

An empirical model fits observed degradation to operating variables. A semi-empirical model adds terms intended to represent particular mechanisms. More detailed electrochemical models describe internal behaviour but still require suitable parameters and validation.

Choose based on available evidence and the question being answered. A project screening exercise may use documented sensitivity scenarios. A contractual capacity forecast needs much stronger support for the exact cells, pack controls and operating conditions.

Ask the model supplier which cell or product was tested, over which temperature/SOC/current ranges, for how long and with what validation data. Extrapolation beyond that envelope should be visible in the report. Check whether the model can represent accelerated aging or a later-life change in degradation rate; a straight line is an assumption, not a universal aging curve.

A published temperature-dependent LFP study fitted mechanisms to a specified Sony cell and experimental conditions. It is useful evidence for model construction, but its fitted parameters do not describe every finished LFP solar battery.

Where the Arrhenius equation fits

A commonly used temperature term has the form:

k = A × exp(−Ea ÷ (R × T))

Here, T is absolute temperature in kelvin, R is the gas constant, Ea is an activation-energy parameter and A is a fitted prefactor. Units must be consistent.

This term can represent the temperature dependence of a selected mechanism within a model. It does not establish that every 10°C increase doubles battery degradation. Fitted parameters, other mechanisms, SOC and the model’s valid temperature range still matter. Do not infer a cooling benefit or replacement year from the equation alone.

Count throughput consistently

A partial discharge event is not necessarily one equivalent full cycle. Declare the convention and capacity reference used for equivalent full cycles (EFC).

Consider a hypothetical battery with a fixed 10 kWh reference energy. It delivers 7 kWh at its DC terminals on each of 300 events:

Discharged energy = 300 × 7 = 2,100 kWh.

Discharge-based EFC = 2,100 ÷ 10 = 210.

There were 300 events, but 210 discharge-based equivalent full cycles. Counting both charged and discharged energy requires a corresponding two-way convention; simply dividing the combined throughput by 10 kWh would double-count this idealized example. Real charge energy also includes losses.

EFC is a useful summary, not a complete degradation model. Two histories with equal throughput can differ in depth, current, temperature and idle SOC. A model using cycle-depth counting needs the underlying trajectory, not just an annual energy total.

Use the selected software’s aging method correctly

Read the model documentation before importing a curve or combining calendar and cycle results. In SAM, the cycle/calendar table option uses the lower remaining-capacity result; other available internal models combine their effects differently. Adding two independently assumed annual fade percentages can therefore misrepresent the selected model. See SAM battery life documentation.

Record the software version, aging option, curve source, capacity reference and replacement settings with the exported results. Re-run dispatch where declining available capacity changes charging, discharge depth or curtailed energy. Retaining the original dispatch while changing only an annual capacity multiplier may miss those interactions.

For a practical workflow:

1. Define the required service: backup, self-consumption, tariff shifting or another duty.
2. Establish compatible battery/inverter limits and the initial usable energy boundary.
3. Build the load, solar, dispatch and thermal assumptions for the evaluation period.
4. Select and document a degradation model with relevant validation.
5. Review capacity, power, delivered energy and operating-limit results separately.
6. Test uncertainties and alternative dispatch or replacement assumptions.
7. Preserve the input files and compare subsequent operating records against the forecast.

Translate retention into useful service, with explicit assumptions

Suppose a sensitivity scenario assumes that a future usable DC energy window is 80% of an initial 10 kWh window. This is an assumed retention scenario, not a measured SoH result or a validated aging forecast.

Future usable DC energy = 10 × 0.80 = 8 kWh.

With an assumed constant 93% conversion efficiency and a constant 2 kW AC load:

Illustrative runtime = 8 × 0.93 ÷ 2 = 3.72 hours.

The same arithmetic gives 4.65 hours for the initial 10 kWh window. These examples exclude auxiliary loads and assume the usable window already reflects the permitted reserve. Actual runtime also depends on conversion behaviour, load changes and whether the battery/inverter can sustain the required power at the relevant SOC and temperature.

Use residential battery sizing to establish the initial load and energy requirement. This aging calculation addresses how that defined service may change over time.

Keep warranties separate from forecasts

A performance warranty defines contractual conditions and remedies. It is not a predicted capacity trajectory or a universal service-life endpoint.

For example, BYD’s Europe HVS/HVM limited warranty V1.2 specifies 60% retained usable energy for ten years from warranty start or the listed minimum throughput, whichever occurs first, subject to its conditions. For HVS 10.2, that document lists 10.24 kWh usable energy and 30.82 MWh minimum throughput. It defines throughput as recorded output energy and specifies measurement conditions.

That example applies to the named document and product. Obtain the warranty applicable to the actual sale, region and configuration; check exclusions and remedies. Do not turn its threshold into an expected year-by-year retention curve, an automatic replacement promise or another manufacturer’s warranty.

Carry uncertainty into financial and customer reports

Distinguish a central forecast from an optimistic or adverse sensitivity. Calling a scenario P90 requires a defined probability distribution, evaluation period and metric; an arbitrary extra fade factor does not establish that probability. Solar-resource uncertainty and battery-aging uncertainty also require separate treatment.

Evaluate replacement or augmentation using actual installed scope: compatible equipment, labour, controls, commissioning, downtime and disposal where applicable. A cell or pack price is not a complete project replacement cost. Confirm whether mixed-age equipment and expanded capacity are supported before modeling an addition.

A useful project handoff states:

• Initial and future capacity boundaries, service requirements and operating assumptions.
• Model identity, data source, calibration limits and uncertainty treatment.
• Expected delivered energy and power limits, rather than capacity alone.
• Replacement or augmentation assumptions with quote dates and scope.
• The applicable warranty document, reported separately from forecast results.
• Monitoring records needed to compare operation with the model.

Explain these assumptions alongside the broader solar design workflow. Avoid promising product-specific degradation automation unless its current documentation demonstrates the exact capability.

Frequently asked questions

How do you model battery degradation in solar systems?

Define the capacity boundary and service, then use operating histories and a model validated for relevant cells and conditions. Document calendar and cycling treatment, uncertainty and replacement assumptions. Revisit dispatch when available capacity changes; a standard annual percentage is only an assumption.

What is the Arrhenius equation for battery aging?

The Arrhenius form k = A × exp(−Ea ÷ (R × T)) can represent a selected mechanism’s temperature dependence. Temperature is in kelvin and parameters must be fitted with consistent units. It does not establish a universal rule that every 10°C increase doubles degradation.

Which battery chemistry degrades slowest in solar storage?

A chemistry label alone cannot rank installed service life. Compare exact cell or system evidence, operating conditions, controls and warranty scope. Laboratory cycle results and finished-system calendar aging are different inputs, and should not be substituted for one another.

How does temperature affect battery degradation in solar?

Temperature can affect aging mechanisms and operating limits, but the relationship depends on cells, SOC, current history and the valid model range. Use actual battery temperature and thermal behaviour where available. Do not assign a universal annual fade or cooling benefit.

What is State of Health (SoH) in solar batteries?

Capacity-based SoH compares measured capacity with a defined reference under comparable conditions. A 160 Ah result against a 200 Ah reference is 80%. That figure does not by itself establish remaining life, AC backup duration or available power.