Design and efficiency optimization of an axial gap permanent magnet sychronous machine adjustable speed drive system

In all areas of adjustable speed drive applications, there is a great need for increased performance, power density and reliability.

A method of design, analysis and performance optimization of an axial gap permanent magnet synchronous machine adjustable speed drive system is discussed in this dissertation. Initially, a brief analysis of the design parameters and their contribution to the drive system performance is discussed. The equations for the machine performance parameters are developed in terms of input design parameters. The single objective optimization (scalar optimization) procedure for performance parameters enhancement, in terms of machine efficiency and specific power has been presented. A multi-objective optimization (vector optimization) procedure, which optimizes the specific power and the efficiency is also discussed for a 3200 rpm, 100 Hp, 216 Volts axial gap brushless de machine. The feasibility frontier curve, which represents the set of optimum design points under the given constraints, is obtained by changing the weighting factors associated with the multi-objective function.

Next, the analysis and the efficiency optimization of the vector controlled axial gap permanent magnet brushless de drive system is discussed, using an actively clamped resonant de link inverter. The resonant frequency at which the inverter losses are minimum is obtained using a scalar optimization procedure. In the next chapter, an analytical solution for the no-load flux density distribution within the machine is obtained using three dimensional field analysis. The validity of the results is confirmed by comparing with the actual field measurement of the machine.

Finally, a Digital Signal Processor (DSP) based position sensor elimination method with an on-line parameter identification scheme is presented for a vector controlled permanent magnet synchronous machine drive system. This scheme accounts for the machine parameter variations and the core losses, and eliminates the shaft mounted encoder, resulting in a reliable, compact and a rugged drive system. A special inverter switching scheme at the machine start-up is also proposed and designed to enable the estimation algorithm to be used in the entire speed range, including the zero speed. The validity of the speed and position estimation algorithm is verified through simulation studies using SIMNON, a PC-based simulation language.