How deep a battery cycles, and how full it waits, shape its life as much as how often it works.
Two habits quietly decide much of a battery's life: how deeply it is cycled, and the charge level it sits at between cycles. Deep cycles stress the electrodes; resting near full speeds up calendar ageing. Neither is forbidden, and both are sometimes worth it, but each has a cost. This guide explains what the research says about depth of discharge (DoD) and state of charge (SoC), how equivalent full cycles work, and how an energy management system can use the same battery more gently without giving up the energy the site needs.
Depth of discharge is how much of the battery's capacity a cycle uses: a cycle from full to empty is 100% DoD; from 80% down to 30% is 50%. Because cycles come in all sizes, researchers count wear in equivalent full cycles: the total charge passed through the battery divided by its capacity. Sandia National Laboratories uses exactly this measure, so two half-depth cycles count as roughly one equivalent full cycle.
Equivalent full cycles are a fair way to count throughput, but they do not mean every cycle costs the same per unit of energy. That depends on how deep each cycle goes, and on the chemistry.
In Sandia's multi-year comparison of commercial LFP, NMC and NCA cells, the rate of capacity fade increased with depth of discharge for every cell tested. Cycling between 40% and 60% was gentlest, 20% to 80% was next, and full 0% to 100% cycles were harshest. The reason is mechanical: larger swings mean the graphite expands and contracts more, which builds stress and micro-cracks.
The size of the effect depends on the chemistry. Sandia found NMC and NCA cells showed a far more dramatic jump in fade from partial to full cycles than LFP, and that the state-of-charge range had little effect on LFP capacity in its tests. Even so, conditions mattered: within the manufacturers' own specified limits, the number of equivalent full cycles to 80% capacity varied by thousands depending on how the cells were used.
A TU Munich study of LFP cells cycled for nearly 900 days found that very shallow cycles in the middle of the charge range caused an unexpectedly strong, though partly reversible, capacity loss. The lesson is that moderate windows help and extremes at either end hurt; it is not that the shallowest possible cycle is always best.
Between cycles, a battery's charge level keeps working on it. Sandia's comparison found that, in calendar ageing studies, capacity fade consistently increased with the state of charge cells were stored at, irrespective of chemistry, because a full cell puts the graphite anode at a low potential that drives SEI growth. NMC and NCA cells faded particularly fast at 100%.
A TU Munich study that stored cells at 16 different charge levels found the effect comes in steps rather than a smooth slope. For the LFP cells tested, ageing was low below about 40%, medium from about 40% to 70%, and highest above about 70%, with no extra jump towards 100%. The authors' advice: avoid high states of charge during long rests. One LFP cell manufacturer's datasheet, likewise, recommends operating between 10% and 90% and storing cells at 30% to 50%.
Heat makes resting full worse. A WRI India working paper on batteries in Indian conditions identifies high ambient temperature combined with high state of charge as a driver of calendar ageing.
Calendar ageing
Wear from time, even when nothing is happening.
Driven by
Cycle ageing
Wear from use, every time energy goes in and out.
Driven by
One working day for a solar-shifting battery
Waits, full, through the hot afternoon. No energy is moving, but the calendar clock is running: how warm the cells are and how full they sit decide how fast.
Waiting: calendar ageingChargingDischarging
Put those two findings together and a simple principle emerges: a battery should spend as little time as practical sitting full and hot. On many sites, the default schedule does the opposite. Solar fills the battery by early afternoon, and it then sits at 100% through the hottest hours until the evening peak.
A health-aware schedule delivers the same evening energy but times the charge so the battery reaches its target shortly before it is needed, and works inside a sensible window rather than end to end. The chart below shows the difference.
All of this depends on knowing the state of charge accurately, and LFP makes that hard. Its voltage barely changes across most of the charge range: a 2024 study in the Journal of Energy Storage notes the voltage changes by only about 0.1 volts across the plateau between roughly 20% and 80%, and that hysteresis can shift it by a third of that. A small voltage error becomes a large SoC error.
In practice, that means LFP systems need regular recalibration, typically by reaching a full charge where the voltage rises steeply, and an EMS that does not trust a drifting estimate. An RWTH Aachen field study of home batteries notes the flatter LFP curve makes the end of charge and discharge harder to detect accurately.
Because the energy is the point. A battery bought to cover an evening peak, back up a factory or meet a tender's dispatch obligation has to deliver its rated energy when called on. Indian grid tenders, for example, require the developer to schedule a minimum dispatchable capacity every year, and a tender specification may set a minimum depth of discharge in its cycle-life requirement.
The right answer is a trade-off made every day: use a gentler window when the extra energy is worth little, and use the full range when the day's value pays for the extra wear. That is a decision for the EMS, not a fixed setting.
Ingro EMS decides the charging window every day. It charges in time for the discharge rather than hours early, keeps a working state-of-charge window, and reaches into the extremes only when the value of that extra energy covers the extra wear, while always protecting the reserve and dispatch obligations the site depends on.
Ingro Brain supplies the cost side of that decision: an estimate of how much a deeper cycle or a longer rest at full charge will age this battery, learned from the site's own data and from batteries of the same chemistry across every site Ingro runs.
Battery health analytics and AI forecasting: state of health, early warnings before failure, and dispatch tuned to make the battery last.
4 questions
Moderate cycles that avoid both extremes are generally gentlest; Sandia found cycling between 20% and 80% aged cells less than full 0% to 100% cycles. LFP is less sensitive to depth than NMC. The best window for a site depends on how much energy it needs each day and what that energy is worth.
An equivalent full cycle is the total energy or charge passed through a battery divided by its capacity. Two cycles at 50% depth count as roughly one equivalent full cycle. Warranties and cycle-life ratings are often expressed this way.
Resting at a high state of charge speeds up calendar ageing in all lithium-ion chemistries, and more so in heat. LFP is more tolerant than NMC, but research still shows faster ageing above roughly 70% for the LFP cells tested. Charging shortly before the energy is needed reduces the time spent full.
LFP's voltage stays almost flat across most of its charge range and shifts slightly depending on whether it was last charging or discharging. Small voltage errors therefore become large state-of-charge errors, so LFP systems rely on current counting and regular recalibration at full charge.
Ingro Brain watches every cell and Ingro EMS runs the battery with its health in mind, on new projects and on batteries already in service.