Arduino Based Control of a New Magnetic Bearing

Summary

Magnetic bearings require accurate, real-time information on the position of the rotor for the control system to provide stable levitation. Displacement sensors can provide this information but can be costly, hard to install and may cause problems in environments where space is an issue. Solutions to remove displacement sensors and cut costs include self-sensing methods in which the position of the rotor is determined from measurements of system parameters such as current, voltage or inductance. Other significant costs come from the type of microprocessor used, therefore this study aims to use one of the cheapest on the market, the Arduino, to assess its capability in a scientific context. This project is an extension of the previous years work, who successfully achieved stable magnetic levitation of a ferromagnetic ring using an internal 8 pole stator and two eddy current sensors, controlled using an Arduino.

An overview of the use of magnetic levitation in an engineering context is provided, highlighting early work and its relevance to the project. The relevant literature is then discussed which explores the areas in which this project might add value to. Aims and objectives of the project are drawn from this including a methodological approach that is split up into four work packages consisting of mathematical modelling, simulation, physical implementation and experimental calibration and validation. A gaant chart is also provided to help appropriately plan and visualise a timeframe for completing this project.

INTRODUCTION

Conceptualised in early stories such as Johnathon Swift’s Gulliver’s Travels (1726) floating island of Leputa, magnetic levitation is an old concept and by now a well-established field. It has many useful applications in the modern engineering world, which will be introduced through a brief history of the subject matter followed by its most relevant application and the topic of this report, the magnetic bearing.

The first novel conception was a “levitating transmitting apparatus”, patented by Emile Bachelet (Bachelet, 1912) allowing for bodies to travel at high speed from one point to the next using electromagnetic suspension. This work was furthered between 1922-1934 by Hermann Kemper, who conceptualised and patented a “monorail with no wheels attached” (Kemper, 1934) which is more commonly known nowadays as the ‘maglev’ train. Presently, Germany and Japan are at the forefront of this technology, recently testing a new route in Japan enabling speeds of 600kph.

Although not as awe-inspiring, magnetic levitation has found a more widespread use in the application of bearings. The principle is to levitate a rotor within an electromagnetic stator using an active control system to determine its position. Passive control alone cannot maintain stability, proven by Samuel Earnshaw in 1842 (Earnshaw, 1842). Most commonly found in turbomachinery, the advantages over mechanical bearings are numerous, especially when required for oil-free applications. Non-contact operation reduces friction allowing for a greater mechanical efficiency and tuneable dynamic properties without the need to adjust the dimensions of the stator or rotor is advantageous (Polajzer, 1999; Habib and Inayat-Hussain, 2003).

MAGNETIC BEARINGS AND THE ACTIVE TOPICS OF RESEARCH

As mentioned previously, the use of magnetic levitation in bearings is well established (Chiba, 2005; Maslen, 2009). The literature on the topic grows rapidly each year with an increasing number of international journals being published across a range of subject areas including mechanical engineering, electrical engineering and applied mathematics. It is an interdisciplinary area of research consisting of technologies such as control systems, stator design and material, electronic amplifiers, and sensor and sensorless configurations. To compare all the literature would be a monumental task given the timeframe, therefore the scope of the literature review was confined to:

  1. The systems and control used for magnetic bearing levitation
  2. Stator design and layout
  3. The use of Arduinos in scientific research
  4. Exploring the areas of research that the undertaken project will add value to

To provide accurate control, the instantaneous position of the rotor or shaft is required. This can be accomplished using a variety of displacement sensors including optical, eddy current, acoustic and capacitive. The literature in these areas is well researched and understood (Chiba, 2005; Maslen, 2009). However, displacement sensors have their drawbacks. Good quality sensors are usually expensive, need space for installation, misalign over time and cause reliability issues because of the wiring involved (Mukhopadhyay, 2005). In some environments they simply cannot be used.

An alternative is to use self-sensing methods. The literature in this area is also well-researched. Vischer and Bleuler reported the self-sensing active magnetic bearing (AMB) that measures gap width based on the measurement of current within the amplifier (Vischer and Bleuler, 1993). Using pulse width modulation (PWM) to drive the system, the gap width can be extracted from coil current and voltage by a signal demodulation process (Tang, Zhu and Yu, 2013). Sivadasan proposed a method in which the inverse of the coil inductance was linearly related to the bearing gap (Sivadasan, 1996), however no experimental results were produced.

Different types of control systems have been implemented to control magnetic bearings. Proportional-Integral-Derivative (PID) control is most widely used because it is simple and provides good performance across a range of operating conditions (Petrov, 2002). However, it was found that PID controllers become ineffective when the machine is operating in highly non-linear ranges (Habib and Inayat-Hussain, 2003) and therefore more robust methods are required. Some of these methods include fuzzy logic control (Hung, 1995) which characterises the model using non-discrete inputs, sliding mode control where the parameters of the system adjust themselves to achieve stable behaviour (Tian and Nonami, 1996) and H∞ control which characterises the problem as a mathematical optimisation problem (Keogh, Mu and Burrows, 1995).

How they currently work

Alternative methods

Objective of this study

Why is it important

State of the art

Literature needs to be critiqued

Where your work will fit in (added value)

Justification for the methodology chosen

No harm in being explicit

AIMS AND OBJECTIVES (250)

This study aims to levitate a ferromagnetic ring using an internally mounted stator through sensorless operation controlled by an Arduino open source platform. This aim will be achieved through the following objectives:

  1. Operate the current test rig to gain a physical appreciation of the project
  2. Analyse how the current control system maintains stable levitation
  3. Design a new control system using Matlab/Simulink utilising sensorless operation
  4. Incorporate the control system onto an Arduino open source platform
  5. Calibrate the system experimentally to achieve stable levitation
  6. Demonstrate the advantages of the new system by comparing quantifiable data against the old system
  7. Assess the capabilities and limitations of an Arduino system in comparison to a dSPACE system

PROGRAMME AND METHODOLOGY

The programme of research will be implemented through four work packages (WP) outlined below. Details of the methodology used, their relation to the aims and objectives and a rough timeframe will be discussed.

WP1: Current system verification and analysis

The first phase of the project will concentrate on bringing the student up to speed with the background technical knowledge required to advance.

T1.1 Analysis of current test rig. Bath University Mechanical Engineering student Nick Rupp (2018) built a test rig to demonstrate magnetic levitation. The test rig comprises of a housing which holds an internally mounted 8 pole stator, an outer ferromagnetic ring and two eddy current sensors to measure the air gap. The electronics used to control the mechatronic system include an Arduino connected to a breadboard, amplifiers, power supply and electronic circuit boards. The main areas of interest are;

  1. The control system used and their parameters
  2. Coding of the Arduino and the use of C++ as a programming language
  3. Overall electronic circuit arrangement

Through operating the test rig, a greater understanding and appreciation for the project will be had. This will narrow down the scope of research required to what is relevant.

T1.2 Background research. Although writing this literature review has provided a good understanding of the current systems and methodologies used for magnetic bearing levitation, it does not cover any technical details. It is at this stage where relevant theoretical material will be read and understood.

WP 2: Design and development of sensorless control system

T2.1 Theoretical modelling of system. Using theknowledge from WP1, the mechatronic system can be characterised mathematically. Using this, a control system can be devised to stably levitate the magnetic ring without the use of displacement sensors.

T2.2 System modelling. This will be modelled in Matlab and Simulink to ensure that it works in theory.

WP 3: Implementation of control system onto Arduino

It was concluded in last years project that tuning control system parameters on an Arduino was a frustrating ordeal because of its slow interfacing with code and changes needing to be recompiled each iteration. Luckily, the department of Mechanical Engineering owns a dSPACE system. It allows for system parameters to be changed on the go with real-time data being displayed. It also couples well with Simulink, an integrated Matlab program that creates an interactive environment for coding using block diagrams to help visualise the simulation and analysis of models.

WP 4: Calibration and validation of experimental test rig

Data collection will be a vital part of the project, in terms of calibration, process validation and performance verification. The data gathered from the eddy current sensors can be used to calibrate the air gap measurement based on the system current and voltage.

References

Bachelet, E. (1912) ‘Levitating transmitting apparatus’. Available at: https://patents.google.com/patent/US1020942A/en.

Chiba, A. (2005) Magnetic bearings and bearingless drives. Amsterdam: London: Elsevier/Newnes.

Earnshaw, S. (1842) ‘On the nature of the molecular forces which regulate the constitution of the lumiferous ether’, Trans. Camb. Phil. Soc., (1), pp. 97–112.

Habib, M. K. and Inayat-Hussain, J. I. (2003) ‘Control of dual acting magnetic bearing actuator system using fuzzy logic’, in Proceedings of IEEE International Symposium on Computational Intelligence in Robotics and Automation, CIRA. doi: 10.1109/CIRA.2003.1222070.

Hung, J. Y. (1995) ‘Magnetic Bearing Control Using Fuzzy Logic’, IEEE Transactions on Industry Applications. doi: 10.1109/28.475746.

Kemper, H. (1934) Suspension railway with wheelless vehicles, which are guided along iron rails by means of magnetic fields.

Keogh, P. S., Mu, C. and Burrows, C. R. (1995) ‘Optimized Design of Vibration Controllers for Steady and Transient Excitation of Flexible Rotors’, Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science. doi: 10.1243/PIME_PROC_1995_209_139_02.

Maslen, E. H. (2009) Magnetic Bearings: Theory, Design, and Application to Rotating Machinery. Berlin, Heidelberg : Springer Berlin Heidelberg.

Mukhopadhyay, S. (2005) ‘Do we really need sensors? A Sensorless Magnetic Bearing Perspective’, 1st International Conference on Sensing …, pp. 425–431. Available at: http://www-ist.massey.ac.nz/conferences/icst05/proceedings/ICST2005-Papers/ICST_207.pdf.

Petrov, M. et al (2002) ‘Fuzzy PID control of nonlinear plants’, in Proceedings First International IEEE Symposium Intelligent Systems, pp. 30–35.

Polajzer, B. et al (1999) ‘Modeling and Control of Horizontal-Shaft Magnetic Bearing System’, IEEE International Symposium on Industrial Electronics, 3, pp. 1051–1055.

Sivadasan, K. K. (1996) ‘Analysis of self-sensing Active Magnetic Bearings working on inductance measurement principle’, IEEE Transactions on Magnetics. doi: 10.1109/20.486516.

Tang, M., Zhu, C. and Yu, J. (2013) ‘Self-sensing active magnetic bearing using real-time duty cycle’, Journal of Zhejiang University SCIENCE C, 14(8), pp. 600–611. doi: 10.1631/jzus.C1300023.

Tian, H. and Nonami, K. (1996) ‘Discrete-time sliding mode control of flexible rotor – Magnetic bearing systems’, International Journal of Robust and Nonlinear Control. doi: 10.1002/(SICI)1099-1239(199608)6:7<609::AID-RNC168>3.0.CO;2-T.

Vischer, D. and Bleuler, H. (1993) ‘Self-Sensing Active Magnetic Levitation’, IEEE Transactions on Magnetics. doi: 10.1109/20.250632.

Professor

You must be logged in to post a comment