An analytical thermal model applied to a Tooth Coil Permanent Magnet synchronous machine

Abstract

This paper analyses the thermal behaviour of a Tooth Coil Permanent Magnet synchronous machine. The thermal network is obtained by interconnecting particular circuit elements: they derive from the specific analytical solution of the thermal field in proper subsystems of the electrical machine. This implies the computation of the axial temperature mean value of the nodes; then, it is possible to compute analyticaUy the temperature in each axial point. The network numerical resolution is validated by an experimental test on a small rating machine.

Analysis and comparison of different diode rectifiers solutions in grid connected WECS employing modular PMSGs

“…The single PMSG losses components have been calculated as follows: − the winding losses are evaluated on the basis of the simulated time-dependent current waveforms, thus taking into account also the losses due to the additional resistances and to the harmonic currents; − considering the generator low operating frequency, the iron losses are calculated simply with the classical model of hysteresis and eddy current, based on the lamination data sheets available from the steel manufacturer; − the PM losses are based on the model developed in [17]; − the mechanical losses are assumed as due to a constant torque due to friction in the bearings, multiplied by the operating rotation speed. As we can see, at first sight, all the solutions seem to give similar losses and efficiency.…”

Thermal modeling for the design and check of an axial flux PM motor

“…By inspecting (12)- (15), some further remarks arise: − R c disappeared: this corresponds to eliminate it from the thermal network, so neglecting the related heat delivered: as shown later, this network simplification implies a limited error; besides, it is a conservative approximation; − about the three loss sources, the PM loss P pm is usually small compared with the stator losses P s = P fe + P cu ; in fact, P pm must be small, otherwise the PMs could be demagnetized: usually, the simplest way to limit P pm is to adopt a suited segmented PM structure; − however, in (12)-(15) the terms proportional to P pm cannot be systematically neglected, depending on the weight of the involved resistive factors; in particular: P pm can be neglected in (12) compared with P fe + P cu , because these loss items appear directly added; P pm cannot be neglected in (13), because R t » R w ; P pm can be neglected in (14), because the corresponding temperature drop is small compared with the other temperature drop terms; P pm cannot be neglected in (15), because R p + R t » R w . Therefore, the final expressions for the approximated estimation (subscript "a") of the temperature rises are:…”

“…It is known that, in order to accurately model the heat flow in the HS, the actual water temperature profile should be considered along the cooling duct length [15]. However, thanks to the small value of the allowed temperature difference (T wo − T wi < 10 °C), the duct water average temperature can be estimated as:…”