Modular Combined Radial Axial Flux Permanent Magnet Synchronous Machine
This was my Major Qualifying Project for WPI. I led a team of 3 to design and manufacture a high-torque, low speed motor for direct-drive robotics applications. This project originally started as designing an axial flux motor as a learning exercise; however, upon attempting to increase the slot count and winding factor with a toroidal stator, it became apparent that significant increases in torque density could be achieved with a hybrid flux configuration.
Ideation
In robotics applications, high torques are frequently sustained over short periods of time. As a result, the limiting factor of a motor is more commonly the saturation flux density of the soft magnetic material of the stator rather than heat dissipation. Thus, my hypothesis was that given the stator is going to saturate, the way to increase torque-density further is simply to maximize the surface area between the stator and rotor. This is done by placing permanent magnets on the axial and radial faces of a toroidal stator.
Stator Configuration
Permanent Magnet Configuration
It also follows that if increasing surface area is desirable, stacking parallel stator-rotor sections would increase the axial surface area as well, potentially increasing torque density further. In addition, this would provide a convenient way to modify the torque output of consecutive R joints on a robotics arm, just as an example, where each joint would require less torque than the previous. In addition, stacked configurations could provide redundancy, where if one stator section fails, the others may still be able to support the load for a short period of time, creating a more stable failure condition.
Single stator-rotor configuration
Double stator-rotor configuration
Triple stator-rotor configuration
Analytical and Electromagnetic FEA Design
An analytical design tool was built in MATLAB, and was used alongside EMWorks’ Electromagnetic FEA SOLIDWORKS integration to confirm my hypotheses and optimize the magnetic model of the motor. Overall, all of the data confirmed that increasing the air gap surface area for the same stator configuration increased torque output, and that a stacked hybrid flux topology could significantly outperform radial and axial flux motors in the 96 x 40mm form factor.
Compiled FEA data from EMWorks
Manufacturing
The obvious drawback of this design is complexity, mostly in the manufacturing of the parts. In an attempt to reduce this complexity, I chose to machine the rotors and stators out of 1018 steel, rather than pursue a laminated design which would be more typical for a motor like this. If there was a competition for the most pretty motor, I think I would be in the running.
Front and middle rotor, front and back stator, and back rotor sections after surface prep.
Results
However, I don’t think this version of the motor will be winning any awards for performance. The peak torque output was measured with an analog torque wrench and the torque constant was found to be around 0.12 Nm/A, which isn’t a significant improvement over a radial flux motor of a similar size. As far as I can tell, this discrepancy came from me vastly underestimating how poor of a magnetic material 1018 actually is, which wasn’t reflected in my analysis because electromagnetic FEA simulations don’t take into account the saturation of the magnetic materials. For the same amount of current, an identical stator made from a proper magnetic steel would produce 4-5 times the flux density, which is conveniently similar to the difference between simulated and measured torque output. Thus, next steps would be to redesign the stators to use a more traditional laminated silicon steel.
