Cycle life is not a single number printed on a datasheet - it is the output of a protocol space with at least three axes: how fully (and how far beyond full) a cell is charged, at what current, and at what temperature. Interactions matter: a current that is harmless cool is damaging hot, and a small overcharge at 0.2C is not the same insult at 1C. This paper frames NiMH charge ageing as a design-of-experiments problem, shows how the axes and their interactions combine to set cycles-to-80-percent, explains how to read and extrapolate accelerated life testing without over-claiming, and turns the resulting map into a procedure for specifying a charge protocol that hits a defined life target rather than relying on generic rules of thumb.
The most controllable life lever is how far beyond full charge current continues. Stopping at the first valid termination minimises the recombination-regime exposure of Paper 21; every additional percent of overcharge at high current is disproportionately damaging because it is pure oxidation and heat. Charging only to 95-98 percent and accepting a small capacity trade can extend cycle life markedly in life-critical applications - a deliberate depth-of-charge limit analogous to the depth-of-discharge limits well known on the discharge side.
The design question is therefore not only 'how fast' but 'how full': a profile that terminates slightly early and never overcharges can outlast one that reaches 100 percent with a sloppy tail, even at identical bulk current.
Rate's effect (Paper 11) is concentrated in the end band: moderate bulk current is comparatively benign, while high current above the oxygen knee is strongly damaging. The rate-depth interaction means life is optimised not by a single current but by a trajectory - high in the bulk where acceptance is high, stepped down as the knee is passed (Paper 2) - which achieves fast total time while keeping the damaging high-rate/high-SOC combination brief.
Reading life tests requires separating these regimes; a test that compares constant 1C to constant 0.2C overstates the penalty of a well-designed multi-stage profile that is 1C only in the benign bulk.
Temperature multiplies the other two axes through Arrhenius-like corrosion kinetics and the falling oxygen-evolution potential: the same overcharge at 45 C ages a cell several times faster than at 20 C, while charging cold raises pressure risk through slow recombination. The protocol space is therefore genuinely three-dimensional with interactions - cool, moderate-rate, no-overcharge being the longevity corner and hot, high-rate, overcharged being the short-life corner.
This is why a single 'cycle life' figure must always carry its test temperature and protocol; numbers measured at 20 C / 0.2C / minimal overcharge are not transferable to a sealed enclosure running 1C to a timer at 40 C.
Accelerated testing raises temperature and/or overcharge to compress calendar time, using an acceleration model (often Arrhenius for the thermal component) to project field life; credibility requires multiple stress levels to fit the acceleration law rather than a single stressed point, matched control groups, and defined endpoints (typically 80 percent of initial capacity plus a resistance ceiling). Projection should be bounded - mechanisms can change at extreme stress, so tests far outside the intended envelope risk accelerating failure modes the field will never see.
Factorial designs that vary depth, rate and temperature together reveal the interactions - e.g. that temperature amplifies rate damage - that one-factor testing misses, and they directly populate the life map used for specification.
The specification procedure is: set the life target (cycles to 80 percent), bound the enclosure's thermal resistance and hence achievable temperature, choose the bulk current from charge-time needs, then walk the current down through the end band so modelled overcharge and peak temperature stay inside the contours that meet the life target; verify with a focused accelerated test at the chosen corner and a real-time confirmation group. The first figure visualises the qualitative life contours over rate and temperature; the second decomposes relative life contribution by protocol factor.
Such a map converts arguments about 'gentle versus fast' charging into measurable trade-offs against a stated objective.
Weijiang supplies cycle-life curves across defined charge depths, rates and temperatures, plus acceleration parameters for projection, so customers can locate their operating point on a measured map rather than extrapolating from a single datasheet number, and can co-design multi-stage profiles that meet a stated cycles-to-80 target. The final ageing-group paper turns to the very beginning of life - formation charging in manufacturing - and shows how the first charges a cell ever receives shape everything that follows.
Weijiang Power designs and manufactures nickel-metal hydride cells, matched packs and charging-ready configurations for consumer, industrial, medical and mobility customers, and supports partners with charge-protocol guidance, IEC 61951-2 performance files, IEC 62133-1 safety evidence and charger co-validation. Share your cell format, charge rate, thermal envelope and cycle target and our engineers will specify a cell-and-charge combination that protects both runtime and service life. Review the range on the products page.
