Abstract
Frameless torque motors are attractive for humanoid and collaborative robot joints because the stator and rotor can be integrated directly into the load-bearing structure, thereby combining a compact axial envelope, low rotor inertia, and a hollow passage for cables or cooling. This study develops and applies a coupled electromagnetic, thermal, and structural workflow to an 18-slot/20-pole frameless permanent-magnet synchronous motor supplied from a 48 V DC bus and rated at 460 W, 1.75 N·m, and 2500 rpm. Calibrated two-dimensional finite-element models in Ansys Maxwell with 3D end-winding leakage and axial fringing corrections are combined with map-based and lumped-parameter thermal models in Ansys Motor-CAD and electromagnetic-force-wave evaluation in Ansys Mechanical. The numerical assessment includes the no-load field, air-gap flux density, back electromotive force and its harmonics, cogging torque, rated and overload operation, losses, efficiency maps, temperature- and current-dependent demagnetisation, transient and steady-state temperatures, and vibration-related force indicators. At the rated operating point, the predicted mean torque is 1.7518 N·m with 3.03% peak-to-peak ripple. The cogging-torque amplitude is 1.833 mN·m (0.105% of rated torque), while the phase back-EMF total harmonic distortion is 5.68% and is dominated by the third harmonic. The peak map efficiency is 90.91%; the rated-point map shaft efficiency is 86.61% (accounting for temperature-corrected copper loss and mechanical/windage losses at thermal equilibrium), whereas the cold electromagnetic-only efficiency at 20 °C is 87.6%. An overload capability point of 4.00 N·m is obtained at 30.63 A (with a transient peak torque reaching 4.03 N·m at 30.5 A), whereas the rated-load thermal model predicts a winding hot spot of 106 °C and a magnet temperature of 96 °C. Under continuous rated operating conditions, the maximum radial electromagnetic force density reaches 0.594 MPa (593,677 N/m²). The nearly unity saliency ratio (Lq/Ld ≈ 1.01) shows that the investigated rotor is a surface-mounted or shallow-inset PMSM and that its reluctance-torque contribution is negligible. These results identify local saturation, rated efficiency, and the second-order radial force mode as the main design constraints. A same-envelope IPM-SynRM rotor with multilayer flux barriers is consequently identified as a subsequent comparison case rather than presented as a validated result. All performance values reported here are numerical simulation predictions that establish a multi-physics design baseline, requiring prototype experimental validation.