Modular 3-D Lumped Parameter Thermal Network for Thermal Estimation of Permanent Magnet Synchronous Motors in Electro-Mechanical Braking Systems

With the development of electric intelligent vehicles, drive-by-wire systems are gradually popularized, the thermal reliability of electro-mechanical brake (EMB) motors, as the key actuators of drive-by-wire systems, is closely related to vehicle safety. However, the blocked-rotation operating condition of EMB motors typically involves applying unequal DC currents to the three-phase windings to achieve braking. This results in uneven winding heating, and conventional thermal models may fail to accurately represent the heat transfer dynamics arising from such tangential thermal imbalance, creating risks of motor local overheating. This paper is focused on thermal performance analysis of EMB motors under blocked-run conditions. First, a single-tooth lumped-parameter thermal network module (tooth module) incorporating external interfaces was developed. This module was adapted using experimental and simulation data, with its internal thermal parameters identified via a particle swarm optimization algorithm, thereby enabling encapsulation into a directly applicable thermal network base module. Subsequently, based on the target motor's mechanical structure, the minimal motor topology was derived. Multiple tooth modules were interconnected through shared nodes and their interfaces according to this minimal structure, establishing a modular three-dimensional lumped-parameter thermal network (modular 3-D LPTN) that accounted for longitudinal, axial, and tangential thermal pathways. Finally, weighted PSO parameter optimization was implemented for the modular 3-D LPTN of the minimal motor. This approach completed PSO optimization of thermal parameters spanning from individual tooth module to the complete motor assembly, realizing a stepwise optimization process from local to global scales. The feasibility was confirmed through simulation analysis. Experiments were conducted to predict the EMB motor's thermal performance under three-phase inhomogeneous heating conditions using the proposed modular 3-D LPTN. The predicted temperatures were compared with those from two existing LPTN, demonstrating the effectiveness of the proposed modular 3-D LPTN for temperature prediction.