APPLICATIONS OF STEPPER MOTORS

APPLICATIONS OF STEPPER MOTORS
 
CHAPTER ONE
INTRODUCTION
1.1  BACKGROUND STUDY OF THE PROJECT
    Stepper motors are used in a variety of applications, including high and low propulsion technology, computer peripherals, machine tools, robotics, etc. The interest in this system has been steadily increasing requirements for accuracy and repeatability while at the same time placing ever tighter demands on the maximum and constancy of speed as well as position resolution. However it has a non-linear and coupled dynamic structure so we could use different control schemes to make the stepper more competitive to use in different levels of application. Open loop control will provide a satisfactory solution under limited conditions. But for high performance dynamic operation this will not give satisfactory results. So we need to find more sophisticated control methods to make the performance of stepper motors much more competitive. We can do this by using newer techniques for drive control using fast semiconductor power switches and powerful microcontrollers made for motor application.
 
 
 
 
 
 
 
 
 
1
 
This work is mainly focused on finding optimized control techniques using simulation and implementation in a PIC16F877 microcontroller using “micro-stepping” drive control. Further, use of a state estimator instead of a costly speed sensor will reduce the cost and increase the reliability of the overall system. By measuring two winding currents and using an extended Kalman filter, one can obtain speed and position estimation.
 
 
 
In recent years there have been successful applications of different types of control theory in the areas of robotics, aircraft control, and motor torque and speed control and estimation using powerful microcontrollers. Complex applications can be implemented by the use of different control methods in standard practice. In conventional methodologies, one has to use different kinds of sensors or transducers to make controllers more robust. But sensors cost a lot, may not give enough resolution, and may have a high failure rate. Using some kind of observer or filter one can get the information of non-measured states. So to achieve better solutions in low cost applications one can use the proper model of the system, information of easily measured parameters and some kind of observer or filter to get the information of non-measured parameters. The two basic groups of observers depend upon the control methodology we use.
Open loop observers can be based on Current model
Voltage model
 
Closed loop observers can be based on Full-order observer
 
Model reference adaptive systems Kalman filter techniques
Adaptive observers based on voltage and current Neural network flux and speed estimators
The Kalman filter technique is one of the good methods employed to identify the speed and rotor-flux based on measured quantities such as current or voltage. This approach is based on the system model and mathematical model describing stepper motor dynamics. Parameter deviation and measurement disturbance are taken into consideration by initializing covariance matrices to proper values. It has good dynamic behavior and disturbance resistance against measurement and process errors. It can also work well in the standstill position. For nonlinear motor model we can used an extended Kalman filter (EKF). Reduced order models are also proposed to shorten and speed up the complex EKF algorithm.
 
 
To implement a filter in practical application is a very complex problem, which involves math routines to calculate in real time. One can easily achieve these by microcontrollers or DSPs with high mathematical power. The computing power of microcontrollers allows users to make a shift from hardware to software by software modeling and simulation in real time. This approach can be accomplished by development and implementation of advanced control algorithms.
 
Significant research advances have been recently made and numerous methods been developed to control motor position and speed without sensors. The literature consistently
 
shows that this was not easy to achieve and only some authors succeeded in fully implementing sensorless control in real time. The others studied in detail different estimation techniques but this work was mostly theoretical and sometimes accompanied by experimental results.
 
 
Estimation technique based on mutual magnetic coupling between winding was proposed by Arefeen, Ehseni and Lipo [ARE94]. This technique was borrowed from the rotor position techniques used for the switched reluctance machine. It was used to identify the rotor angular position at the zero crossing of the phase currents, which were controlled by a simple hysteresis regulator. An advantage of the method was that it was unaffected by stator resistance, while shortcomings were low estimation update rate and required knowledge of angular velocity and the use of lookup tables.
 
 
An experimental sensorless torque vector controller was designed and built by Lagerquist, Boldea and Miller [LAG94]. Speed estimates were derived from the flux phasor position and thus no speed sensor was needed. The flux magnitude was controlled at the knee point value on the maximum torque per ampere characteristic in order to minimize copper losses at full load. Despite this, it was shown both theoretically and experimentally that the machine efficiency was generally less than under the condition of the maximum torque per ampere control strategy, particularly at light loads.
 
 
Some other representative approaches have also been considered to estimate the speed without a shaft sensor by comparing their sensitivity to parameter variations, their ability
 
to handle the load on the motor and their speed tracking capability. Adaptive methods considered the speed as an unknown constant parameter and used techniques to estimate the speed. Bodsen and Chiasson approached the problem from the parameter identification point of view [BOD93]. The idea is to consider the speed as an unknown constant parameter and to find the value of speed that best fits measured data to the dynamic equation of the motor. This approach assumes the parameters are known and fixed with time.
 
 
Domenico Casadei, Giovanni Serra and Angelo Tani proposed the technique to estimate the speed and flux of the motor without assuming the speed is slowly varying compared to electrical variables [DOM03]. Their approach uses polar coordinates of the flux rather than Cartesian. They developed complementary flux estimator when the sign of the speed is unknown which uses the ratio of stator and rotor electrical frequency to find estimator convergence conditions.
 
 
 
As we know stepper motors are used in a variety of applications, including high and low propulsion technology, computer peripherals, machine tools, robotics, etc. The interest for this system has been steadily increasing requirements for accuracy and repeatability, while at the same time placing ever tighter demands on the maximum speed and the constancy of speed and reduced overall cost of the system. But position and velocity sensors used in different types of application cost around $100 to $10,000 each. When we require more than one sensor in a particular application that makes the overall system
 
cost increase by a lot. In the United States the sensor market for every year is around $6.1 billion and increasing with an average annual growth rate (AAGR) of 4.6%. So after 5 years it will become around $7.6 billion [GLO04]. So our aim is to find some technique where we can estimate the states or parameters of the system using the information of the measured outputs. To estimate the states we can use the Kalman filter technique with a proper mathematical model of the whole system. We can implement the mathematical model with a Kalman filter routine in a cheap microcontroller to estimate the states online. Using this type of technique we can replace this costly sensor by a cheaper micro- controller in any industrial application. Here, our plan is to estimate the velocity and position of a stepper motor by measuring the winding currents using the PIC16F877 microcontroller. By doing that we can get away from the use of costly encoders.
 
 
 
This research work considers the development of sensorless control of permanent magnet stepper motor using MATLAB® first by simulating the motor equation and than implementing in a PIC16F877 microcontroller for real time control. Chapter II presents the problem formulation. It describes methods or procedures used as well as obtained results. Chapter III presents an overview of various kinds of stepper motors, their advantages and various control techniques. Chapter IV presents an introduction to the Kalman filter, different techniques for implementation, and their advantages and disadvantages. Chapter V presents how to implement sensorless control in a PIC16F877 microcontroller with peripheral hardware. First it describes how to implement the “micro- stepping” technique to drive the motor in software using C. Then it discusses how to
 
implement the extended Kalman technique with P or PD control using peripheral hardware with the PIC16F877 and its software algorithm. Chapter VI presents the results obtained by implementing the extended Kalman technique with P or PD control using MATLAB® simulation. It finally states the various advantages of this type of technique and proposed future work. Chapter VII presents some conclusions and suggestions for future research.