Deodhar, Rajesh Pranay
The Flux-MMF diagram technique and its applications in analysis and comparative evaluation of electrical machines.
PhD thesis, University of Glasgow.
Full text available as:
The thesis describes a new technique, called the flux-MMF diagram technique, for analysis and comparative evaluation of electrical machines. The technique has evolved from the principle of virtual work, and the -i diagram, used commonly in designing switched reluctance machines and relays. Several applications of this technique are demonstrated in the thesis, supported by experimental validation. These are, the prediction of electromagnetic and cogging torque ripple, modelling of the effect of skew on torque and torque ripple, modelling of the variation of torque constant due to saturation, and comparative evaluation of different types of electrical machines.
The thesis shows that the technique can be applied successfully in analysis of a wide variety of electrical machines. These include conventional machines such as the DC commutator, PM brushless AC, Interior PM, and the synchronous reluctance machine; as well as non-conventional machines such as the switched reluctance, PM brushless DC, and the doubly-salient PM machine.
The technique has been implemented in a finite-element software, with the help of a link program which links the FE software with the dimensioning or sizing software, such as PC-BDC, produced by the SPEED Laboratory. The link program serves as a vital means of shortening the time it takes to analyse a new design in an FE software, by several orders of magnitude.
The thesis also describes a new brushless doubly-salient permanent-magnet machine, called the flux-reversal machine. The design and fabrication process, and the experimental results are presented for a prototype single-phase, high-speed flux-reversal generator. The performance analysis of the prototype based on the flux-MMF diagram technique is included, and this validates its capability in analysing new and non-conventional machines, which cannot be analysed using the classical means.
Actions (login required)