A fixed charge profile is tuned for a new cell and becomes progressively wrong as that cell ages: capacity fades, internal resistance rises, the voltage peak arrives earlier and the same current runs hotter. Ageing-aware adaptive charging closes this gap by having the charger continuously identify the cell it actually holds - estimating resistance, capacity and thermal behaviour from each charge - and adjusting current, stage transitions and termination so the intended stress level is maintained through life rather than imposed once at the factory. This paper develops the adaptive-charging concept, shows which observables support online identification, how current is re-planned from updated parameters, and why adaptation is one of the most practical routes to extending realised NiMH service life.
As established across the degradation group, age raises ohmic and charge-transfer resistance (Paper 17), shrinks capacity and the negative reserve (Papers 21, 23), and slows recombination; a timer sized for a new, full-capacity cell therefore overcharges the smaller aged cell more each cycle, a fixed current generates more I-squared-R heat in the higher-resistance cell, and a fixed -delta-V threshold meets an earlier, altered peak. The profile does not change, but the cell does, and the mismatch accumulates.
Adaptive charging treats the profile as a function of an online-updated cell model rather than a constant, so 'the same intended treatment' means different currents and times at age zero and age five-hundred cycles.
Several age signatures are measurable without extra hardware: internal resistance from charge-current step response or short pulses; the timing and magnitude of the voltage peak and dT/dt inflection; the charge accepted between full-charge anchors (an estimate of retained capacity, Paper 16); and the temperature reached for a given current. Trending these across cycles at matched conditions - same current, comparable ambient - separates ageing from the reversible effects of temperature and SOC.
A charger that logs every charge builds a per-battery history: resistance trend, shrinking accepted charge and earlier peaks converge into a state-of-health estimate without laboratory instrumentation, exactly the data a fixed-profile charger throws away.
With updated resistance and capacity, the controller re-plans: current is derated to hold predicted peak temperature and pressure proxy at their design levels (compensating age much as it compensates cold and heat, Papers 26, 27); stage-transition states of charge and the safety timer are rescaled to retained capacity so an aged cell is not overcharged by a new-cell clock; and -delta-V/thermal thresholds are confirmed against the evolving peak shape rather than an absolute millivolt assumption.
The effect is to keep each cell at a roughly constant stress trajectory through life: a new cell charges quickly because it can; an aged cell is charged more gently precisely when it needs protection, recovering service life that a fixed aggressive profile would have consumed early.
The most immediately implementable adaptation is thermal closed-loop: adjust current in real time to hold a target temperature-rise rate, charging faster when the cell stays cool and backing off when it heats - effectively making the 1 C/min dT/dt criterion (Paper 7) a continuous current governor rather than a stop switch. Pressure-adaptive versions demonstrated in traction research cap pressure near a setpoint (the 6.8 atm example of Paper 3) and modulate current to stay beneath it, achieving high fill (around 93 percent SOC in that study) without exceeding the gas-handling limit.
These controllers use the cell's own response as the feedback signal, so they are inherently robust to lot variation and ageing: whatever the underlying parameters, current is whatever keeps the measured thermal/pressure trajectory on target.
Adaptive and multi-stage intelligent schemes (including the ANFIS-MSCCC controllers of Paper 20) report higher efficiency and lower thermal stress than fixed current, and the physical logic - holding stress constant as resistance grows - predicts improved cycle retention. Limits are real: online identification needs current-stable measurement windows and enough cycle history, safety must never depend on the adaptive layer (hard limits remain, Paper 40), and a faulty sensor must trigger conservative fallback rather than aggressive 'adaptation' to a false reading.
The first figure contrasts fixed and adapted current trajectories for a new versus aged cell; the second sequences the identify-plan-act adaptive loop executed each charge.
Weijiang supplies the age-trend characterisation - how resistance, peak timing and accepted charge evolve over cycles for each grade - that lets partners calibrate adaptive algorithms and validate them against real ageing rather than assumed models. The next frontier paper asks a stronger question: can an electrochemical model compute the mathematically optimal charge trajectory, minimising time and degradation simultaneously?
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.
