State of charge (SoC)

Definition (what it is)

State of charge (SoC) is a dimensionless measure of how much charge remains in a rechargeable battery relative to its maximum available (usable) charge at the current state of ageing and operating conditions. It is typically expressed as a percentage from 0% (empty) to 100% (full). In formula form: SoC = Q / Qmax × 100%, where Q is the remaining charge and Qmax is the present, usable capacity (which changes with temperature, rate, and ageing). The complementary metric depth of discharge (DoD) is DoD = 100% − SoC.

Key characteristics

  • Not directly measurable: SoC is inferred from electrical and thermal signals plus models; it is a state estimate, not a physical sensor reading.
  • Reference capacity matters: The denominator (Qmax) may be the initial rated capacity, a temperature-compensated capacity, or an adaptive value reflecting degradation. Many systems show user-facing “0–100%” over a limited usable window (e.g., true 10–90%) to protect longevity.
  • Charge- vs energy-basis: SoC is charge-based. Because pack voltage varies with SoC, temperature, and current, the remaining energy may be better captured by state of energy (SoE), which some applications use for range or runtime predictions.
  • Complementary states: SoC interacts with state of health (SoH, capacity/power fade over life) and state of power (SoP, instantaneous power capability at a given SoC and temperature).

Estimation methods (typical approaches, often combined)

  • Coulomb counting: Time-integration of current in and out of the cell, corrected for coulombic efficiency; accurate short term but subject to sensor offset and drift, requiring periodic recalibration.
  • Open-circuit voltage (OCV) correlation: Mapping rested cell voltage to SoC using chemistry-specific OCV–SoC curves; accurate after sufficient rest, but limited by hysteresis and long relaxation times (especially challenging for flat-OCV chemistries like LFP).
  • Model-based observers: Equivalent-circuit or electrochemical models (e.g., Thevenin/RC networks, PNGV) with state observers (e.g., extended/unscented Kalman filters, sliding-mode, particle filters) fusing current, voltage, and temperature to correct drift and account for dynamics and hysteresis.
  • Data-driven and hybrid methods: Machine learning or physics-informed ML trained on historical and field data to adapt to ageing and diverse operating conditions; often combined with physical observers.
  • Auxiliary techniques: Impedance-based estimation, incremental capacity/differential voltage analysis, and end-of-charge/end-of-discharge detection to refine SoC and capacity.

Influencing factors and error sources

  • Temperature (affects capacity, internal resistance, and OCV), C‑rate/load dynamics, cell chemistry (OCV slope and hysteresis), ageing mechanisms (loss of lithium inventory and active material), self-discharge and parasitic loads, current-sensor accuracy/offset, and pack imbalance.
  • At very high or very low SoC, available power and accessible capacity are limited by voltage and resistance constraints, complicating estimation and control.

Calibration and referencing

  • Realignment (“gauge reset”) is achieved via:
    • Rested OCV points (after low-current or idle periods),
    • Full or near-full charge/discharge events,
    • Diagnostic routines that refine Qmax and hysteresis parameters.
  • Many systems enforce a usable SoC window (e.g., 20–80% daily) to extend life; the displayed SoC may be normalized to that window for user clarity.

Pack- and system-level considerations

  • Cell dispersion: Variations in cell capacity/impedance cause cell SoC spread; weakest cells limit pack performance.
  • Balancing: Passive (bleed resistors) or active (DC/DC energy shuttling) balancing aligns cell SoC to prevent over/under-voltage at the pack limits.
  • Sensing and hardware: Precision current shunts or Hall sensors, high-resolution voltage and temperature sensing, and robust ADCs and timing are essential for low-drift estimation.

Relevance and applications

  • Electric vehicles (EVs): SoC underpins range estimation, navigation-based energy planning, power/torque limits, regenerative braking, fast-charging control, thermal preconditioning, and safety interlocks to prevent overcharge/over-discharge.
  • Stationary/storage and grid services: SoC drives dispatch, reserves, and degradation-aware scheduling for UPS, home batteries, microgrids, and utility-scale systems, including vehicle-to-grid/home (V2G/V2H).
  • Consumer and industrial devices: SoC informs runtime prediction, charge indication, and lifecycle management in laptops, phones, tools, robots, and drones.
  • Safety and reliability: Operating and charging limits based on SoC mitigate lithium plating, copper dissolution, and thermal events; diagnostics and plausibility checks (functional safety) monitor estimator health.

Accuracy notes

  • With good sensors and calibration, steady-state SoC errors of roughly 1–3% are achievable; under highly dynamic loads, temperature swings, or flat OCV regions, errors can be larger unless robust observers and periodic recalibration are used.

Related terms and distinctions

  • Depth of discharge (DoD): DoD = 100% − SoC.
  • State of energy (SoE): Remaining energy fraction; often more relevant for runtime/range than charge-based SoC.
  • State of health (SoH): Capacity/power degradation relative to new; affects the Qmax used in SoC.
  • State of power (SoP): Instantaneous power capability at a given SoC and temperature.
  • Remaining useful life (RUL), state of function (SoF): Lifecycle and functional capability estimates.
  • Colloquial: battery level, fuel gauge. Not to be confused with SoH (degradation), which is a different concept.

Chemistry and design context

  • SoC applies to all rechargeable chemistries (lithium-ion, LFP, NMC/NCA, lead–acid, NiMH, etc.), but OCV–SoC profile, hysteresis, rate sensitivity, and usable windows are chemistry- and design-dependent.
  • EV cells (e.g., NMC/NCA with graphite or silicon-graphite anodes; LFP with graphite) exhibit distinct OCV curves and thermal behavior, which drive estimator tuning and calibration.

Communication and standards (examples)

  • SoC is commonly communicated over in-vehicle networks (e.g., CAN) and to chargers/energy management systems via higher-level protocols. In EV contexts, SoC exchange is supported in standards and ecosystems such as ISO 15118 (vehicle–charger communication) and OCPP (charger–backend). Functional safety practices (e.g., ISO 26262) guide diagnostics, plausibility checks, and fault handling for SoC estimation.