Design and Implementation of Magnetic Levitation Based Modular Omnidirectional Conveyor System

Abstract

Smart manufacturing is now required to be more intelligent and flexible to support mass customization. Magnetic levitation or maglev technology can increase the flexibility of modern manufacturing with its advantages of low friction and multi-directional motion. Two promising industrial applications of maglev are planar motors and precision stages. This research aims to develop a novel planar maglev system, Magnetic Levitation Floor, and implement it as an omnidirectional conveyor. The contributions of this research are the design and analysis of a novel modular planar electromagnet stator and permanent magnet mover, wrench–current decoupling enhancement using multivariate linear interpolation, and the development of a decoupling strategy for long-range motion. For maglev system development, the planar electromagnet stator is an array of square coil modules. The coil was designed to attain optimal levitation force with efficient power. The support frame has a modular structure for unlimited expansion and flexible layouts. Two planar maglev prototypes, 10-Coil Testbed and Maglev Floor Prototype, were built in addition to the Maglev Floor system for pilot testing. Next, a four-disc magnet mover was designed to achieve six-degrees-of-freedom motion. The magnet layout was optimized with analyses on condition number, maximum current, and power consumption. For real-time levitation control, the cross-coupling effect in the system is decoupled by a wrench–current decoupling matrix. The Lorentz force-based wrench matrices are pre-computed and stored in the lookup table. With discrete lookup data, however, the levitation performance can deteriorate when the mover levitated at pose without data, especially in rotations. In this work, a multivariate linear interpolation was implemented to improve wrench–current decoupling effectiveness between lookup data. For conveying applications, another limitation of the decoupling method using a lookup table is the storage usage which is directly proportional to the data range. To achieve long-range planar motion with efficient use of data storage, the decoupling strategy with small-range lookup data was developed in this work. The lookup data has the range of quarter coil area. The active coil set is selected from the mover location. Then, the wrench matrix is estimated from lookup data and mover pose relative to the active coil set. The feasibility of flexibility enhancement in smart manufacturing using the designed planar maglev system was evaluated through long-range and fine motion experiments. The multivariate linear interpolation significantly improved the levitation performance which suffered from discrete data switching. The decoupling strategy enabled the mover to traverse the levitation area. Experiment results demonstrated the capability of the system for real-world omnidirectional conveying applications.