Answers are grounded in this paper and cite their sources.
Regenerative braking control needs a loss model of the drive as a function of speed and deceleration. However, for a commercial in-wheel drive, often only battery-side (direct current, DC) data exist. This paper analyses the identifiability of such a model and builds an identification procedure on it. DC-side data determine the electrical parameters only up to a common scale factor when the modulation limit of the controller is unknown. For a 5 kW hub motor, the torque constant measured from the open-circuit voltage fixes this scale. Five tests check the manufacturer’s load test report against the model, and three low-speed bench events give the constant loss as 70 ± 20 W. The identified set comprises a constant drag torque of 2.71 ± 0.40 N·m and an equivalent iron loss resistance of 11.4 Ω per phase. With every constant taken at its bound, the predicted loss varies by ±16 to ±32 W, a combined bound band and not a coverage interval. On the road, the cut-off speed, below which braking returns no energy, falls to 4.1 km/h at 1.0 m/s2, so energy is recovered almost to standstill. Bench cut-off speeds are consistent with the model but not independent of the identification.
No takes yet. Share an insight, caveat, or question.
Long et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: