Abstract
A proposed permanent magnetic suspension using flux path control has quasi-zero power characteristic since its unique structure realizes transmitting the suspended object’s gravity the system frame, and the motor just drives the permanent magnet to rotate and suffers a small regular torque caused of the system magnetic potential. The small regular torque will deteriorate the system’s response characteristics and suspension stability. This paper optimizes the system structure for improving the quasi-zero power characteristic. Firstly, the structure and the suspension principle of the former system is introduced, and the rotational torque on the motor’s shaft was measured using an experimental prototype. Secondly, a symmetric offset optimal structure was proposed through analyzing the experimental results of rotational torque, and the torque on the motor’s shaft was calculated by using a FEM model. The simulation results indicate that the rotational torque was reduced to about 15% of the former structure using the optimized structure with a magnetic separation iron plate. Finally, the experimental prototype with the optimized structure was manufactured, and the measurement experiment for the torque was carried out. The experiment results verified the validity of the simulation results and the optimizations.
Keywords
Introduction
Permanent magnetic suspension technology has the characteristics of compact structure, large magnetic density and no heat, which has been widely studied by domestic and foreign scholars [1–6], and has a wide application prospect in the fields of mechanical industry, aerospace, vehicle [7,8], electric motors and generators [9–11] and so on. At present, the production and processing of semiconductor, medical equipment, food and ultra-high precision equipment are required in the dust-free workshop, so magnetic suspension technology plays an important role in the dust-free workshop. In the field of dust-free transmission, a novel permanent magnetic suspension system has been proposed [12–14], which can effectively solve the difficulties of traditional manufacturing and greatly promote the development of magnetic suspension technology in dust-free transmission field.
The system is a permanent magnetic suspension system using variable flux path control method, which changes the magnetic flux through the suspended matter to control the magnitude of the suspension force by changing the angle of rotation of the permanent magnet in the system. But due to the structural characteristics of the device, the load torque is generated on the motor shaft, which has quasi-zero power characteristic [15].
In this paper, the authors reveal that because the former device is affected by the periodic magnetic potential, the permanent magnet has rotational torque. After analyzing the factors that affect the zero power of the device, a symmetric offset optimal structure is proposed, and the three-dimensional finite element analysis model is established. The finite element software is used to analyze the transient magnetic field, and the structure of the device is adjusted according to the simulation data. The authors build an experimental platform for measuring torque to verify zero power characteristic. Experimental results show that the method of active elimination using the waveform characteristics of the sinusoid can decrease the rotational torque of the permanent magnet and reduce the input current of the motor. Consequently, the purpose of reducing energy consumption and the quasi-zero power characteristic of the system is achieved.
Quasi-zero power characteristic
Structure of the former device
Figure 1 shows a dust-free transportation system using permanent magnetic suspension system, it mainly consists of four permanent magnetic suspension systems and a motion driven mechanism. The system’s attractiveness can realize the suspension between the carrier and the rail, and the driving force of the carrier is provided with the motion-driven mechanism. And this paper optimizes the structure of the former device for the permanent magnetic suspension system. Figure 2 shows the structure of the former device for the permanent magnetic suspension system. This prototype mainly consists of a disk permanent magnet, a servo motor containing a gear reducer and an encoder, a pair of F-type permalloy, a suspended object and an eddy current sensor. The diameter of the radially magnetized magnet is 30 mm and the thickness is 10 mm. The servo motor behind the magnet drives the magnet to rotate, the encoder measures the rotation angle of the magnet, and the position of the suspended object is measured by the eddy current sensor.

A dust-free transportation system using a permanent magnetic suspension system.

Structure of device.
Figure 3(a) shows that the magnetic poles of the magnet are aligned in the vertical direction and the N pole is at the upper side and the S pole is in the lower side. In this case, the facing angle of the N pole and S pole to each core are same, so all magnetic flux comes from the N pole and is absorbed into the S pole through each core respectively. There is no flux flowing through the suspension object, so zero attractive force generates between the cores and the levitated object. However, Fig. 3(b) shows the magnet rotated a certain angle, the facing angle of the N pole becomes bigger than the S pole in the right core, and that is reversed in the left core. Since that, the flux from the N pole in the right core is more than that in the left core. Some of the flux in the right core flow through the suspension object to the left core and is absorbed by the S pole. Consequently, there is some flux flowing through the suspension object, and the attractive force is generated. The flux flowing through the suspension object becomes more as the rotated angle is larger until the rotated angle reaches 90 degrees.

Principle of variable flux path control mechanism.
Figure 4 shows the measurement device for the rotational magnet. The rotational torque of magnet was measured with strain gauges, when varied the rotational angle of the permanent magnet and the air gap between iron cores and the suspended object. Two pieces of strain gauges for measuring torques were pasted on the side of the connector between the rotary motor and the permanent magnet.
Figure 5 shows the experiment results of the rotational torque when the permanent magnet was rotated at 10 degrees as one step in one revolution, and the displacement of the suspension object was from 2 mm to 8 mm and increased at 1 mm as one step. The rotational torque is determined by the angle of the permanent magnet and displacement of the suspended object. The magnitude of the suspension force is determined by the amount of the magnetic flux passes through a pair of F-shape permalloy. The gravity of the suspended object is directly transmitted to the frame through a pair of F-shape permalloy, and the permanent magnet and the motor do not support the gravity of the suspended object. But the permanent magnet produces rotational torque during the rotation, and in order to offset the rotational torque, the motor needs to input a small current. Consequently, the former device has quasi-zero power characteristics.

The measurement device for the rotational of the magnet.

Rotational torque of the permanent magnet.
Principle of optimality
The design of the structure optimization is based on the characteristics of the torque on the motor shaft as shown in Fig. 6. The torque value is sinusoid when the permanent magnet rotates under the drive of the servo motor. According to the waveform characteristics of the sinusoid, the torque between the two permanent magnets and two pairs of F-shape permalloy offset each other to eliminate drive power consumption of the motor. This paper proposes a symmetric offset optimal structure, two radially magnetized permanent magnets are mounted on both sides of the frame, and the magnetic pole of two permanent magnets is staggered 90 degrees.

Torque of different angles.
As shown in Fig. 7, the permanent magnetic ring is fixed on the main shaft, which is connected with the servo motor through the rigid coupling, and a F-shape permalloy is placed on both sides of the permanent magnet ring; there is a ferromagnetic plate between two aluminum frame; one end of the suspension object is fixed through the bearing, the other end of the suspension object is swing; the fixed object is fixed to the frame; an eddy current sensor is placed above the suspended object. The symmetrical structure can better avoid unnecessary influence factors, and try to eliminate the interference of external factors. Consequently, the results of experimental and finite element simulation are more accurate.

Structure diagram of the device.
Simulation analysis conditions
The diameter of the permanent magnetic ring is 40 mm, the material is NdFeB30; the distance between the two permanent magnetic ring is 38 mm; the model of F-shape permalloy is 1J85; the thickness of the ferromagnetic plate is 1 mm, the material is iron; the distance from the suspended objects to the F-shape permalloy is 5 mm; the distance between permanent magnet and F-shape permalloy is 2 mm. The simulation speed is 5°∕s, the total simulation time is 72 s.
Finite element simulation model
According to the research questions, the finite element model of the magnetic driving device using permanent magnetic suspension is assumed to ignore the leakage of the device and the effect of friction on the device.
As shown in Fig. 8, the finite element model of the import analysis software. Figure 8(a) is a finite element model diagram of the former structure. Figure 8(b) is a finite element model diagram based on the former structure, two radially magnetized permanent magnets are mounted on both sides of the frame, and the magnetic pole of two permanent magnets are staggered 90 degrees. Figure 8(c) is a finite element model diagram based on the optimized structure, there is a magnetic separation iron plate between two permanent magnets.

FEM analysis model.
The finite element method (FEM) is used to compare the rational torque of the former device and the optimized device. As shown in Fig. 9, it can be seen that the rotational torque of the permanent magnet is changed by a sine wave, and the maximum torque of the former structure is about 67 mN ⋅ m. The rotational torque of the optimized structure is reduced, but the phase of the torque diagram is changed and the torque of each permanent magnet is not a sine wave. When the distance of two permanent magnets is increased, the phase of the torque diagram is returned to normal and the torque of each permanent magnet is a sine wave. Consequently, there is placed a magnetic separation iron plate to block the influence between two permanent magnets. The rotational torque of the optimized structure with a magnetic separation iron plate is obviously reduced, and it can be clearly seen that the influence between two permanent magnets has been obviously eliminated, the maximum torque of the magnets is about 10 mN ⋅ m, which can be ignored. Therefore the symmetric offset optimal structure has basically reached the design requirements of zero power characteristic, and reduced the power consumption efficiency. But due to the influence of magnetic leakage, the phase of the torque diagram is also changed.

Comparison results of rotational torque.
The improved device and the measurement device
In this experiment, the servo motor drives the permanent magnet, which selects EC—max 30 servo motor produced by Sachseln, Switzerland. Its the parameters: rated voltage is 12 V, the moment of inertia is 21.9 g ⋅ cm2, the rated rotary speed is 6590 rpm, the maximum speed is 7980 rpm. The displacement sensor selects the EX-V eddy current displacement sensor produced by Kean company in Japan. The permanent magnet material is NdFeB30, and the F-shape ferromagnetic material is permalloy. The improved experimental prototype is displayed at Fig. 10.

Experimental prototype.
For verifying zero power characteristics of the permanent magnetic suspension system, the torque on the spindle is measured with the torque sensor, when varied the rotational angle of the permanent magnet. The torque sensor is placed between the motor and the permanent magnet and connected by two rigid couplings. The air gap length is 2 mm between the F-shape permalloy and the magnet, and the distance from the suspended object to the F-shape permalloy is 5 mm. The measurement device is shown in Fig. 11.

Measurement device for rotational torque.
The contrast results of the rotational torque are shown in Fig. 12. As shown in Fig. 12(a), the measured results are similar to the FEM analysis results for the former structure. But the measured results are smaller than the simulated results, because the magnetic flux leakage is taken into account in the simulation and is compensated. The contrast results of the rotational torque for the optimized structure are shown in Fig. 12(b), the simulated results are closed to 0 mN ⋅ m, but the maximum torque of the measured results reaches 30 mN ⋅ m. The bearing produces friction in the course of movement because the spindle is fixed on the frame by bearings, so which affects the experimental results. And the cable of the static torque sensor also affects the experimental results. Because the two magnets still have some influence, the phase of the torque diagram is also a little changed. The torque is reduced by 40% after the device is optimized, and the optimized effect is achieved.

Rotational torque of the permanent magnet.
For optimizing the quasi-zero power characteristic of a permanent magnetic suspension system, this paper proposes a symmetric offset optimal structure. And the characteristic of the optimized structure is analyzed by the finite element analysis software and verified by the experiments. The simulation and experimental results show that two permanent magnets in the optimized device have an important effect on the zero power characteristic of a permanent magnetic suspension system, the magnetic separation iron plate blocks the influence between two permanent magnets, and the rotational torque has reached the optimized effect by adopting active elimination method. Finally the experiment results verified the validity of the simulation results and the feasibility of the optimized structure. But the problem of the friction and the cable has affected the results of the experiment, the authors will improve the experimental device to reduce the friction force in the course of bearing rotation.
Footnotes
Acknowledgements
This research is supported by National Natural Science Fund of China (Grant No. 51105257, No. 51310105025), and the colleges and universities of Liaoning Province outstanding young scholars growth plan (Grant No. LQJ2014012), China Postdoctoral Science Foundation (Grant No. 2015M571327).
