Determination of the convective heat transfer coefficient in large electrical machines by a new simulation strategy

Abstract

Purpose

The purpose of this paper is to present a new computational fluid dynamics model for large electrical machines to simulate the heat transfer at specific components to the appropriate ventilation method. The most damageable parts for overheating in generators are the end winding bars, pole windings and stator ducts.

Design/methodology/approach

The reduced model introduced is basically derived from the state-of-the-art pole section model (PSM) and enables faster computations for heat transfer and cooling simulations of electrical machines. The fundamentals of the two methods and the grid generation are described. Two PSMs and four different reduced models are presented and compared among each other to tune the reduced model.

Findings

As a topic of outstanding interest in large hydro generators, the heat transfer at the end winding bars is solved with the aid of the reduced model. This slot sector model (SSM) has been validated and the computation time has been reduced enormously in comparison to the state-of-the-art PSM.

Research limitations/implications

The heat transfer has been carried out only for the end winding region of large hydro generators. The effect of the reduced model on the pole sections and stator ducts has not been investigated. Nevertheless, the reduced model is also valid for large motors.

Practical implications

This reduced model can finally be used for parametric studies with different cooling schemes and boundary conditions in the design process.

Originality/value

The comparison of various SSMs to PSMs shows an acceptable accuracy of the reduced model in combination with a rather low computation time. Due to modeling one slot only, the MFR-MP approach is an adequate and fast analyzing method for this kind of model structure.

Citation

Klomberg, S., Farnleitner, E., Kastner, G. and Bíró, O. (2015), "Determination of the convective heat transfer coefficient in large electrical machines by a new simulation strategy", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 34 No. 4, pp. 1335-1348. https://doi.org/10.1108/COMPEL-02-2015-0070