EPREVAEngineer Paper
You are reading version 2 of 2 (5 Oct 2026).
Change note: Added the discrete BMS form (eq. 5), a limitations section and reference [3].

A first-order thermal model of a battery cell: derivation, parameters and a worked example

We derive the lumped thermal model used in many battery management systems, show how its two parameters set the steady-state temperature rise and the response time, and work through an illustrative cylindrical-cell example with Ohmic heating of 2 W. Steady state is reached after about five time constants; with an illustrative thermal resistance of 8 K/W and heat capacity of 70 J/K the cell settles 16 K above ambient with a time constant of 560 s. The model ignores internal temperature gradients and entropic heat, which we discuss as limitations.

Introduction

Cell temperature drives battery ageing and safety, but few cells carry an internal sensor. Battery management systems (BMS) therefore estimate temperature with simple models. This note derives the most common one, the first-order lumped thermal model, and works through an example.

Model

Treat the cell as a single body at uniform temperature TT with heat capacity CTC_T, exchanging heat with ambient air at TaT_a through a thermal resistance RTR_T. An energy balance gives

CTdTdt=Ta−TRT+QC_T \frac{dT}{dt} = \frac{T_a - T}{R_T} + Q
(1)

where QQ is the heat generated inside the cell. Following the energy balance of Bernardi et al.[1] and neglecting the reversible (entropic) term, the heat generated is the irreversible loss

Q=I (VOCV−V)≈I2R0Q = I\,(V_{OCV} - V) \approx I^2 R_0
(2)

with R0R_0 the cell's internal resistance.

Steady state and time constant

Setting dT/dt=0dT/dt = 0 in (1):

ΔTss=RTQ\Delta T_{ss} = R_T Q
(3)
Playground · Temperature rise from thermal resistance
solve for
Temperature rise
10 K
Plot temperature rise against
1,00000.01200 K/W (log)

For a step in QQ at t=0t = 0, the solution of (1) is

T(t)=Ta+RTQ(1−e−t/τ),τ=RTCTT(t) = T_a + R_T Q \left(1 - e^{-t/\tau}\right), \quad \tau = R_T C_T
(4)

Worked example

ParameterValue
Current I10 A
Internal resistance R_020 mΩ
Heat Q = I²R_02.0 W
Thermal resistance R_T8 K/W
Heat capacity C_T70 J/K
Time constant τ560 s
Steady-state rise ΔT16 K

Assumptions

Discrete form for a BMS

A BMS samples at a fixed step Δt\Delta t. Holding QQ constant over each step, (4) gives the exact recursion

T[k]=ϵ T[k−1]+(1−ϵ)(Ta+RTQ[k−1]),ϵ=e−Δt/τT[k] = \epsilon\, T[k-1] + (1-\epsilon)\left(T_a + R_T Q[k-1]\right), \quad \epsilon = e^{-\Delta t/\tau}
(5)

which costs one multiply-add per sample. The same structure appears in Chowdhury and Lu's one-shot identification method[3], where (5) is rearranged to be linear in its parameters so that they can be fitted by least squares.

Limitations

References

  1. D. Bernardi, E. Pawlikowski and J. Newman, "A General Energy Balance for Battery Systems," Journal of the Electrochemical Society, vol. 132, no. 1, pp. 5–12, 1985.
  2. C. Forgez, D. V. Do, G. Friedrich, M. Morcrette and C. Delacourt, "Thermal modeling of a cylindrical LiFePO4/graphite lithium-ion battery," Journal of Power Sources, vol. 195, no. 9, pp. 2961–2968, 2010.
  3. M. A. Chowdhury and Q. Lu, "Equivalent-Circuit Thermal Model for Batteries with One-Shot Parameter Identification," arXiv:2503.12616, 2025. arxiv.org/abs/2503.12616