Finite Element Modelling of a Field-Sensed Magnetic Suspended System for Accurate Proximity Measurement Based on a Sensor Fusion Algorithm with Unscented Kalman Filter.

Amor Chowdhury,Andrej Sarjaš

doi:10.3390/s16091504

Abstract

The presented paper describes accurate distance measurement for a field-sensed magnetic suspension system. The proximity measurement is based on a Hall effect sensor. The proximity sensor is installed directly on the lower surface of the electro-magnet, which means that it is very sensitive to external magnetic influences and disturbances. External disturbances interfere with the information signal and reduce the usability and reliability of the proximity measurements and, consequently, the whole application operation. A sensor fusion algorithm is deployed for the aforementioned reasons. The sensor fusion algorithm is based on the Unscented Kalman Filter, where a nonlinear dynamic model was derived with the Finite Element Modelling approach. The advantage of such modelling is a more accurate dynamic model parameter estimation, especially in the case when the real structure, materials and dimensions of the real-time application are known. The novelty of the paper is the design of a compact electro-magnetic actuator with a built-in low cost proximity sensor for accurate proximity measurement of the magnetic object. The paper successively presents a modelling procedure with the finite element method, design and parameter settings of a sensor fusion algorithm with Unscented Kalman Filter and, finally, the implementation procedure and results of real-time operation.

Highlights

Linear proximity sensors (LSPs) with mid- and low-range measurement capabilities are devices that are used widely in many industrial and non-industrial applications
The backstepping is a nonlinear recursivecontroller design technique, synthesised from the derived model in simplified second order nonlinear model in was which stabilises the origin of the system in strict feedback form [44,45]
The sensor fusion algorithm with UKF improved the accuracy of distance measurement and system states estimation drastically, which are used in feedback control

Summary

Introduction

Linear proximity sensors (LSPs) with mid- and low-range measurement capabilities are devices that are used widely in many industrial and non-industrial applications. They are mostly used to determine the displacement, direction of movement, orientation, speed, etc., of a measured object. The LSP exploits different physical principles of operation, where capacitive, inductive, ultrasonic, optical and magnetic phenomena are the most commonly deployed in a sensing operation Many of these physical phenomena, especially accurate optical technology and the ultrasonic principle, require complex pre- and post-processing operations which, unintentionally, result in a high price and relatively large dimensions of the measuring unit. In the time of high expansion and pervasive sensing technologies, especially in the field of miniature sensors, as well as a highly efficient processing unit of relatively small dimensions and price, they offer many applicable solutions which can effectively replace many complex and expensive

Results

Discussion

Conclusion