This thesis focuses on the development of new axial flux brushless doubly fed reluctance machine having the higher torque to volume density compared to its radial flux counterpart. In particular, the rotor pole design
which is very important for such machines has been focused and analyzed to increase the effectiveness and advantages of the proposed AF-BDFRM. Furthermore, sample prototype of 1.0 HP was selected for testing
considering the manufacturing cost and available testing facilities to verify
the proper working of the novel AF-BDFRM.
The proposed machine is most suitable for applications that require speed
control over a limited range like in wind turbines and compressors. As it
utilizes only slip power which reduces the rating of the converter to be used
with control winding. Therefore, the overall cost of the system goes down, and it becomes a viable option for such variable speed applications.
Firstly, the design concept including the basic design criterion in single and
dual air gap machines is explained. Then, the axial flux design consideration
was taken into account to design the proposed AF-BDFRM. Then the design
procedure including the design, analysis and optimization method for the
rotor pole of proposed machine are discussed in detail.
The proposed machine model has two stators and one salient pole reluctance rotor. The salient pole rotor is made of only iron and winding free. The
absence of permanent magnets, ability to control power factor and stable mechanical structure of the machine make it overall cheaper and more rugged.
The absence of copper losses in the rotor also increases the overall efficiency of the proposed machine compared to doubly fed induction machine. Proposed machine can operate in sub-synchronous, synchronous or super-synchronous
mode depending on whether a negative sequence frequency, DC or positive sequence frequency is applied to the control winding, respectively. The unaligned placement of two stators suppresses the cogging torque and torque
ripple in the proposed machine.
Torque to volume ratio of the AF-BDFRM is higher than that of a radial flux BDFRM because torque is directly proportional to D^3 and not to D^2L.
However, its air gap length is double than RF-BDFRM which will cause
more leakage
flux in the proposed machine but the surface current density
also can be increased in the proposed design as it has two stators whereas
RF-BDFRM has only one stator, and its surface current density has to be
limited to K = K_1 + K_2.
A transient 3D FEM analysis was done to analyze the performance of the
proposed AF-BDFRM and the performance was compared with the already
developed RF-BDFRM. The rotor pole shape was modified to a diamond
shape to reduce cogging torque and torque ripple. Transient FEA results
show that the proposed machine exhibits 91% higher torque density when compared with the already developed RF-BDFRM whereas the weight of the
proposed machine is 22.5% lighter. However, it was observed that the rotor
pole shape could be further improved and optimized for higher torque
density and lower torque ripple.
For the optimization of the rotor pole in the proposed machine, the objective function and design variables were selected. Inner overhang, outer
overhang, and rotor span were selected as the design variables. Rotor pole
span was chosen as a design variable to increase the lower reluctance path
for the
flux. Then, the Latin Hypercube (LHC) sampling method was used
to design the experiments. After that, the Kriging method was used to approximate the objective function and a genetic algorithm (GA) was utilized
to find optimal values for the selected design variables. A basic sinusoidal
shape instead of the previous diamond shape was selected as the starting
point for optimization of the rotor pole.
The back EMF increased from 378 V for the former diamond
shape to 453 V for optimized sinusoidal shape. The output torque performance was also greater for the optimized rotor pole corresponding to an
increased torque density of the machine. Furthermore, the torque ripple
was observed 16% for the optimized sinusoidal shape which was 26.6% for the
diamond shape. Torque output, which was previously 323 Nm, increased to
371 Nm.
To verify the FEA and the proper working of the novel AF-BDFRM, a 1.0
HP prototype was fabricated for experimentation. Because this prototype was needed only to verify the proper working of the proposed AF-BDFRM, therefore, considering the manufacturing cost and the available laboratory facilities, a 1.0 HP prototype is selected to perform experiment. A basic sinusoidal shaped rotor pole with no overhang and a 30o mechanical span was selected. This 1.0 HP prototype was
built and tested in the laboratory.
The rotor and stator poles were chosen to fulfill the speed requirement
and to reduce the unbalanced magnetic pull (UMP) which is necessary to
consider for the mechanical stability of axial flux machines. Stacking factor and the coil area of the constructed prototype was found to be lowered
during manufacturing which reduced the flux carrying capability of the iron
core and linkage capacity of winding coils. Due to this reason, a difference of 10% was observed between the simulated and experimentally induced
back EMF. However, Experimental results verified the proper working of the
novel AF-BDFRM and its smooth functioning over a range of variable speed.