ELECTRIC ELEVATOR DRIVE WITH POSITION CONTROL

A Thesis Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering

by Peter James Ford B.S., Louisiana State University Agricultural and Mechanical College, 2009 May 2012

Dedicated to my parents

ii

Acknowledgements
I would like to thank those who have provided me the opportunity to write this thesis: My parents who have always provided the motivation and support I needed to pursue an education in Electrical Engineering. My advisor and mentor Dr. Ernest A. Mendrela, who has been a guiding light along my path to both my Bachelors and my Masters degree in Electrical Engineering. Finally, I would like to thank the faculty and staff of Louisiana State Universitys Department of Electrical and Computer engineering: Dr. Kemin Zhou, Dr. Alex Skavantos, and Dr.Shahab Mehraeen who took the time and care to serve on my defense committee and Mrs. Beth Cochran, the graduate program coordinator, has worked tirelessly to ensure my interest in the department were looked after.

iii

Table of Contents
Dedication........................................................................................................................................ii Acknowledgements..........................................................................................................................iii List of Tables...................................................................................................................................vi List of Figures..................................................................................................................................vii Abstract..............................................................................................................................................x Chapter 1: Introduction.....................................................................................................................1 1.1 Permanent-Magnet DC Motors........................................................................................1 1.2 Digital Control Systems...................................................................................................1 1.3 Elevators..........................................................................................................................2 1.4 Teaching Lab.....................................................................................................................2 1.5 Objectives of Thesis..........................................................................................................2 1.6 Brief Description of Chapters..........................................................................................3 Chapter 2: Electric Elevator Drive.....................................................................................................4 2.1 Commercial Electric Elevator Drive Systems..................................................................4 2.1.1 Kone System.....................................................................................................4 2.1.2 Otis System.......................................................................................................6 2.2 Laboratory Model of an Elevator with Permanent-Magnet DC Drive.............................7 2.3 Mathematical Model of Elevator Electric Drive..............................................................8 2.3.1 Motors Electrical Equations.............................................................................8 2.3.2 Mechanical Systems Motion Equations..............................................................9 Chapter 3: Determination of Parameters of Electric Elevator Drive.................................................13 3.1 Steady-State Parameters.................................................................................................13 3.2 Dynamic Parameters.......................................................................................................21 Chapter 4: Designing of Control System...........................................................................................28 4.1 Feedback Control of Electric Drive................................................................................28 4.2 Current Control Loop......................................................................................................31 4.3 Speed Control Loop........................................................................................................35 4.4 Position Control Loop....................................................................................................38 Chapter 5: Performance of Electric Elevator Drive with Position Control........................................40 5.1 Performance of Simulated Electric Elevator Drive with Position Control......................40 5.2 Performance of Experimental Electric Elevator Drive with Position Control.................54 5.3 Comments on Performance of Electric Elevator Drive with Position Control................66

iv

Chapter 6: Conclusions....................................................................................................................67 References.........................................................................................................................................69 Appendix A: Load_Simulation_Parameters.....................................................................................70 Appendix B: Plot_Simulation..........................................................................................................71 Appendix C: Plot_RealTime.............................................................................................................72 Vita.....................................................................................................................................................73

List of Tables
Table 3.1 Steady-state parameter determination data.....................................................................16 Table 3.2 Linearized data received from versus characteristics............................................18

Table 3.3 Average values of motor parameters obtained from steady-state characteristics..............19 Table 3.4 Steady-state friction determination data.......................................................................20 Table 3.5 Steady-state friction parameters.....................................................................................21 Table 3.6 Inductance of armature winding....................................................................................24 Table 3.7 Results of equivalent inertia Jeq calculations...................................................................26 Table 3.8 Data for determination of equivalent inertia .................................................................27

Table 4.1 Parameters for current control loop design.....................................................................35 Table 4.2 Torque control loop integral gain calculations................................................................35 Table 4.3 Parameters for speed control loop design........................................................................37 Table 4.4 Results of speed loop gain calculations...........................................................................37 Table 4.5 Position loop proportional gain calculations...................................................................39 Table 5.1 Simulated position response of electric elevator drive with no load and 1kg, 2kg load....................................................................................................................................................53 Table 5.2 Percent change in simulated position response of electric elevator drive under load......53 Table 5.3 Real-time position response of elevator electric drive with no load and 1kg, 2kg load....................................................................................................................................................65 Table 5.4 Percent change in real-time position response of electric elevator drive under load........66 Table 5.5 Performance of Elevator Electric Drive..........................................................................66

vi

List of Figures
Figure 2.1 Kone EcoDisc electric elevator drive[4] ........................................................................5 Figure 2.2 Kone Machine-Room-Less Elevator System [3,4] ............................................................5 Figure 2.3 Otis steel belt [7] .............................................................................................................6 Figure 2.4 OtisGen2 electric elevator drive [7]................................................................................6 Figure 2.5 Laboratory model of elevator driven by permanent-magnet DC motor.............................7 Figure 2.6 Equivalent circuit of armature of DC motor.....................................................................8 Figure 3.1 ( ) Characteristic14...........................................................................................14

Figure 3.2 MUT connected to load motor.......................................................................................14 Figure 3.3 MUT axially coupled to load motor (seen in laboratory)...............................................15 Figure 3.4 Speed versus armature current characteristics...............................................................18 Figure 3.5 Torque-speed characteristic of steady-state friction model............................................21 Figure 3.6 Current response with blocked rotor ( )...........................................................22

Figure 3.7 Current waveform for 3V step increase...........................................................................23 Figure 3.8 System speed response upon shutdown..........................................................................25 Figure 3.9 Speed response after power supply is disconnected from motors...................................26 Figure 4.1 Feedback control system of DC motor...........................................................................28 Figure 4.2 Cascaded control system................................................................................................29 Figure 4.3 PI Controller Diagram...................................................................................................30 Figure 4.4 Expanded current control loop[6] .................................................................................31 Figure 4.5 Reduced current control loop[6] ...................................................................................32 Figure 4.6 Simplified current control loop[6] ................................................................................32

vii

Figure 4.7 Simplified current control loop[6] ...................................................................................36 Figure 4.8 Position control loop block diagram................................................................................38 Figure 5.1 Simulink model of elevator electric drive with position control......................................41 Figure 5.2 Simulated DC machine subsystem...................................................................................41 Figure 5.3 Simulated current controller subsystem............................................................................42 Figure 5.4 Simulated speed controller subsystem...............................................................................43 Figure 5.5 Simulated position controller subsystem..........................................................................43 Figure 5.6 Simulated load torque subsystem for ascending elevator car............................................44 Figure 5.7 Simulated load torque subsystem for descending elevator car..........................................44 Figure 5.8 Simulated reference position of elevator electric drive with ascending car.....................45 Figure 5.9 Simulated reference position of elevator electric drive with descending car...................45 Figure 5.10 Simulated position response of elevator electric drive with ascending car.....................46 Figure 5.11 Simulated position response of elevator electric drive with descending car..................46 Figure 5.12 Simulated speed response of elevator electric drive with ascending car........................47 Figure 5.13 Simulated speed response of elevator electric drive with descending car....................47 Figure 5.14Simulated current response of elevator electric drive with ascending car........................48 Figure 5.15 Simulated current response of elevator electric drive with descending car..................48 Figure 5.16 Simulated torque response of elevator electric drive with ascending car.......................49 Figure 5.17 Simulated torque response of elevator electric drive with descending car.....................49 Figure 5.18 Simulated voltage response of elevator electric drive with ascending car.....................50 Figure 5.19 Simulated voltage response of elevator electric drive with descending car..................50 Figure 5.20 Simulated position response of elevator electric drive for ascending car carrying no load, 1kg load, and 2kg load.............................................................................................................52 viii

Figure 5.21 Simulated position response of elevator electric drive for descending car carrying no load, 1kg load, and 2kg load..............................................................................................................52 Figure 5.22 View of experimental elevator in lab.............................................................................54 Figure 5.23 Safety cut-off switch..................................................................................................55 Figure 5.24 View of elevator electric drive and drive pulley with drive belt...................................55 Figure 5.25 Elevator electric drives power processing unit.............................................................56 Figure 5.26 dSpace 1104 hardware.................................................................................................56 Figure 5.27 Real-time Simulink model for position control of elevator electric drive......................57 Figure 5.28 Real-time current controller subsystem........................................................................58 Figure 5.29 Real-time speed controller subsystem...........................................................................58 Figure 5.30 Real-time position controller subsystem........................................................................59 Figure 5.31 Real-time duty cycle generation subsystem.................................................................59 Figure 5.32 Real-time speed averaging subsystem..........................................................................59 Figure 5.33 Control desk control interface for position control of elevators electric drive..............60 Figure 5.34 Real-time position response of elevator electric drive with ascending car....................61 Figure 5.35 Real-time position response of elevator electric drive with descending car...................61 Figure 5.36 Real-time speed response of elevator electric drive with ascending car.........................62 Figure 5.37 Real-time speed response of elevator electric drive with descending car.......................62 Figure 5.38 Real-time current response of elevator electric drive with ascending car......................63 Figure 5.39 Real-time current response of elevator electric drive with descending car..................63 Figure 5.40 Real-time position response of elevator electric drive for ascending car carrying no load, 1kg load, and 2kg load............................................................................................................64 Figure 5.41 Real-time position response of elevator electric drive for descending car carrying no load, 1kg load, and 2kg load..............................................................................................................65 ix

Abstract
This thesis is a study on the implementation of an elevators position-controlled electric drive. The information contained within this paper serves as a framework to expand the usefulness of electric drives through the addition of digital control systems and switching power supplies. The tangible example of an elevator driven by a permanent-magnet DC motor is used for this paper so that students may relate to the work and apply it to their future projects. The tasks to be accomplished in order to achieve an electric elevator drive with position control are: determining the parameters of the permanent-magnet DC motor, designing a control system to direct the motor as desired, and verifying the performance of the system through use of computer simulations and experimental testing. The tests to derive the motor parameters as well as the theory behind the test are covered in depth before the design procedures for creating a cascaded control system are started. Computer simulations are conducted using the parameters and controllers which will be implemented in real-time before experimental testing in the lab begins. Conclusions are drawn about the performance of the position-controlled electric elevator drive based upon the simulation and experimental results. The implementation of an elevator driven by a permanent-magnet DC motor with position control is successful and provides an illustrative example to those who wish to apply electric drives to various mechanical systems.

Chapter 1: Introduction
This thesis will cover the design, simulation, and implementation of a permanent-magnet DC motor controlled by a digital control system to drive a suspension-system elevator in such a way that the material in this thesis can be used as a teaching tool and laboratory experiment for the LSU Course EE 4490 Advanced Speed Drives.

1.1 Permanent-Magnet DC Motors

The permanent-magnet DC motor is an ideal choice for driving mechanical systems where precise speed control is desired. Regulating the speed of a permanent-magnet DC motor is as simple as changing the magnitude of voltage applied to the motors armature circuit; to make the motor operate in the opposite direction the polarity of the supply voltage needs to only be reversed. A switching power supply and digital control system provided an efficient means for doing this.

1.2 Digital Control Systems

Digital control systems are composed of software and hardware components which are optimized to execute algorithms quickly and efficiently. These algorithms are designed to speed up complex operations which may involve mathematical calculations and logic operations as well as the conversion of analog signals into a digital format or digital signals into an analog format. These algorithms can be executed as optimized code on a microprocessor or they can be hardwired into the controllers architecture. In order to increase efficiency and obtain the quickest response time each digital control system is designed for the system it must regulate. Using a digital control system to dictate the position, speed, and current of an elevators electric drive increases the accuracy and precision with which the elevator system responds to position commands.

1.3 Elevators
Elevators are used to move loads from one height to another without disturbing the load itself. The prime mover of an elevator can either be a hydraulic tube beneath the elevator cart which rises as fluid is pumped in to or out of the cylinder or an elevator can be moved using a cable system and counterweight which is driven by an electric motor [8]. Suspension systems driven by electric motors are the ideal choice for applications having minimal space and great heights because these systems do not require infrastructure to be built beneath the floor of the elevator and are limited in height only by the ability of the electric drive to overcome static forces such as gravity and dynamic forces such as friction and air drag. For these reasons this thesis will focus on the suspension system driven by an electric motor.

1.4 Teaching Lab

This thesis is to be an example of implementing position, speed, and current control of a permanent-magnet DC motor to drive a mechanical system which can be used as a teaching tool in the LSU electrical engineering course on electric drives. Students will be able to apply the lessons and techniques contained within this paper to other applications of digitally controlled electric machines beyond the position controlled electric elevator drive.

1.5 Objectives of the Thesis

To implement position control of an elevators permanent-magnet DC motor drive To develop the experiment on position control of a permanent-magnet DC motor drive as teaching lab for the LSU electrical engineering course on electric machines

1.6 Brief Description of Chapters

Chapter 2 is a review of electrically driven elevators and will cover two designs currently being used in commercial applications in addition to covering the elevator model to be used in this thesis Chapter 3 concerns the determination of parameters for the elevators electric drive Chapter 4 covers the design of the digital control system which dictates the elevators electric drives response to position input Chapter 5 summarizes the performance of the elevators electric drive, both in computer simulation and experimental implementation. Chapter 6 gives conclusions based upon the analysis contained within the thesis and elaborates upon its potential use for education

Chapter 2: Electric Elevator Drive

The physical arrangement of a suspension-type elevator has a direct affect on the forces acting upon its propulsion system. Ideally, a gear system is not used so as to reduce mechanical losses and improve the drive systems response time; however, a drive system without gears requires a motor capable of accelerating the system from a standstill. To aid the drive system, a counterweight is employed to reduce the overall force needed to drive the system when it is under load. To further reduce the work which must be done by the prime mover, the support cable/belt is configured to spread the load evenly between the drive pulley and the anchor points. These techniques of reducing the load force on the drive system are applied to all commercial elevators of the suspension type regardless of their propulsion system leaving the electric drive to be the defining factor of each companys design.

2.1 Commercial Electric Elevator Drive Systems

Two major international manufactures of elevators using electric drives are Kone and Otis. Both companies use the same approach to suspension of car and counterweight but have developed different electric drive systems for propelling the elevator car and counterweight.

2.1.1 Kone System

The Kone EcoDisc elevator electric drive system is a machine-room-less design which places a permanent-magnet, synchronous motor in the hub of the drive pulley as seen in Fig. 2.1. Steel cables are passed over this drive pulley and suspended the car and counterweight; as shown in Fig. 2.2, these cables are anchored to the top of the elevator shaft and cable tension is provided by the weight of the hanging car and counter-weight. In this arrangement when the elevator car moves upwards the counterweight moves downwards with cable ends never changing position.

Figure 2.1 Kone EcoDisc electric elevator drive [4]

Figure 2.2 Kone Machine-Room-Less Elevator System [3,4] 5

2.1.2 Otis System

The OtisGen2 electric elevator drive system is a machine-room-less design which uses a flat steel belt (see Fig. 2.3) in place of woven steel cables to support the elevator car and counter weight [7]. The use of a belt allows for a drive pulley with a smaller diameter which reduces the total volume of the drive system (see Fig. 2.4). This smaller drive system is mounted at the top of the elevator shaft which eliminates the need for an equipment room to be built above the elevator shaft, hence the moniker machine-room-less. Like the Kone system, the OtisGen2 electric elevator drive system is powered by a permanent-magnet synchronous motor. Also like the Kone system, the Otis system has the elevator car and counterweight suspended from the drive pulley with the ends of the belt anchored to the top of the elevator shaft so that the suspended car and counterweight provide tension on the belt.

Figure 2.3 Otis steel belt [7]

Figure 2.4 OtisGen2 electric elevator drive [7] 6

2.2 Laboratory Model of an Elevator with Permanent-Magnet DC Drive

Figure 2.5 Laboratory model of elevator driven by permanent-magnet DC motor 7

The elevator electric drive that is proposed for this thesis is most similar to the Otis methodology. The scheme of proposed structure is shown in Fig. 2.5. A drive belt is looped over a drive pulley with the ends anchored to the top of the elevator shaft. The electric drive is rigidly attached to the drive pulley without the use of gearing. The car and counterweight, which ride on pulleys, provide belt tension ensuring the toothed belt and drive pulley do not slip. The elevator car and counter-weight are of equal weight so as to reduce load on the motor to only what is placed inside the cart.

2.3 Mathematical Model of Elevator Electric Drive

The equations governing the operation of the elevator can be attributed to the motors electrical circuit and to its mechanical system.

2.3.1 Motors Electrical Equations

The elevator is driven by a permanent-magnet DC motor. The equivalent circuit of the permanent-magnet DC motor is confined to the armature circuit which is illustrated in Fig. 2.6.

Figure 2.6 Equivalent circuit of armature of DC motor The motor voltage equation of the armature circuit is: (2.1) where: is the electro-motive force developed in the armatures winding is the motor constant is the angular speed of the motors shaft 8 (2.2)

is the voltage difference between the armature terminals is the resistance of the armature circuit is the current flowing through the armature circuit is the inductance of the armature circuit The electro-magnetic torque developed by the motor is expressed as: (2.3)

2.3.2 Mechanical Systems Motion Equations

In order to derive the motion equation that describes the elevators mechanical system it is assumed that the mass of the drive belt is ignored due to its material composition and length. In practice, the drive belt or drive cable is constructed from steel and its significant mass contributes to the load torque in a non-linear fashion depending upon the position of the car. The motion equation of the entire system from the motors perspective is: (2.4) where: is the motors moment of inertia is the angular speed of the rotor is the friction coefficient of the motor is the load torque placed on the motors shaft The load torque that is placed on the drive pulley which is mounted on the motors shaft is

expressed in Eqn. 2.5: (2.5) where: is the radius of the drive pulley 9

is the inertia of the motor pulley is the force exerted on the drive pulley If the elevator car is moving upwards the load force [ where: is the gravitational constant is the mass of the car is the mass of the counter-weight is the linear speed of the car Eqn. 2.6 was derived for the elevator mechanical system shown in Fig. 2.5 in which the force exerted by the accelerating car is split into two parts. One half acts on the belt attached firmly to the top ( is defined by Eqn. 2.6: ) ] (2.6)

of the elevator shaft and the other half acts on belt which is looped over the drive pulley. In a similar way the force exerted by the counter-weight is split into two parts, one half acting on the belt firmly attached to the top of the elevator shaft and one half acting on the belt which is looped over drive pulley. The belt which is driven by the drive pulley is moving with the speed great as the car speed . This is expressed as: (2.7) The car speed expressed in terms of motor speed is expressed as: = The mechanical power of the car moving upwards with constant speed is: (2.9) This power when written with respect to the pulling belt is: (2.8) which is twice as

10

(2.10) Replacing the belt speed with the car speed, the power is shown to be: (2.11) Eqn. 2.11 shows that the power of the moving car is equal to the power of the pulling belt. In Eqn. 2.6 the force is affected by the gravitational force and the inertia of the elevator car

, both of which are expressed as: ( ) (2.12) (2.13) Eqn. 2.12 and Eqn. 2.13 hold only when the car is moving upwards and the belt is flexible; under these conditions the motor is not affected by the counter-weight during acceleration. With this assumption Eqn. 2.6 can be inserted into Eqn. 2.5. The load torque ( ) is now expressed as: (2.14)

As mentioned, Eqn. 2.14 was derived for the elevator car moving upwards. When the elevator car accelerates moving downwards the counter-weights mass elevator cars mass should be considered while the when the elevator car is

should be ignored. The equation for load torque

moving downwards has the form: ( ) (2.15)

Substituting the load torque equation for when the elevator car is moving upwards (Eqn. 2.14) into the motors motion equation (Eqn. 2.4), the motion equation of the entire system is obtained as: ( or: (2.17) 11 ) ( ) (2.16)

where the equivalent moment of inertia

is: (2.18)

and the load torque due to gravity

is: ( ) (2.19)

With the equations describing the performance of the elevator and its electric drive known, attention can now be paid to the permanent-magnet DC motor itself. Chapter 3 extends the analysis of this thesis to the unique parameters of the electric drive which serves as the prime mover of the system.

12

Chapter 3: Determination of Parameters of Electric Elevator Drive

In order to correctly design position, speed, and current controllers for the elevators electric drive, the parameters of the permanent-magnet DC Motor must first be determined. Parameters can vary from unit to unit and so each motor must have its unique values empirically determined. The mechanical and electrical characteristics of the permanent-magnet DC Motor are delineated as steady-state or dynamic values.

3.1 Steady-State Parameters

Steady-state characteristics of a permanent-magnet DC Motor reflect losses due to friction found in the electric drive and the relationship between electro-magnetic torque developed by the motor and the current drawn by the motor. The steady-state parameters to be determined are: - armature resistance motor constant, relation between armature current and electro-magnetic torque - static friction torque - motor friction coefficient The relationship between the motors supply voltage and its speed is linear and given by Eqn. 3.1. (3.1) which can be rearrange as: (3.2) As seen in Fig. 3.1 this dependency can be expressed in y-intercept form as: (3.3) where: (3.4) (3.5) 13

Figure 3.1

( ) Characteristic

To determine the parameters of the permanent-magnet DC motor in steady state, the Motor-UnderTest (MUT) will be axially coupled to another DC motor which will act as a load on the MUT. This arrangement is shown in Fig. 3.2 and a picture of the connected motors is shown in Fig. 3.3.

Figure 3.2 MUT connected to load motor

14

Figure 3.3 MUT axially coupled to load motor (seen in laboratory)

With the load motors shaft rigidly attached to the shaft of the MUT, the supply voltages of the load motor and MUT are changed in incremental steps. The load torque placed on the MUT is equal to

the electro-magnetic torque developed by the load motor. At each step the speed, voltage, and current of the MUT are recorded until enough data points are collected to draw the speed-current characteristic [2]. The data points collected for 5 different MUT supply voltages are shown in Table 3.1 and the corresponding ( ) characteristics are shown in Fig. 3.3.

For Table 3.1 the data points collected are: VM voltage applied to MUT VL voltage applied to load motor IM current drawn by MUT IL current drawn by load motor M angular speed of MUT and load motor 15

Table 3.1 Steady-state parameter determination data [ ] 5 10

[ ]

[ ]

-6.00

-3.75

-2.00

-0.66

0.90

2.25

-11.75

-9.25

-7.20

-5.85

-4.25

-3.00

[ ]

-0.87

0.14

1.07

2.01

3.01

4.05

-1.02

0.03

1.06

1.91

2.92

3.87

65

49

40

32

20

12

141

123

109

101

89

80

[ ]

15

20

[ ]

[ ]

-17.20

-14.85

-12.75

-11.00

-9.60

-8.15

22.2 0

-20.00

-18.20

-16.50

-15.00

-13.40

[ ]

-1.45

-0.16

0.82

1.86

2.75

3.66

-1.28

-0.22

0.67

1.51

2.27

3.22

210

199

180

171

160

149

278

262

255

247

230

218

16

Table 3.1 Continued

[ ]

30

[ ]

[ ]

32.70

30.40

28.30

26.40

24.70

22.50

[ ]

-1.22

-0.29

0.68

1.56

2.43

3.35

409

394

379

369

358

347

The data collected during the steady-state parameter test is plotted in Fig. 3.4 in a speed versus armature current fashion. For each voltage applied to the MUT, the speed versus current

characteristic is linearized; the slope (m) and y-intercept (c) for each linearized characteristic is used to determine value of and for each voltage applied to the MUT. Rearranging Eqn. 3.4 and Eqn. 3.5, the applied to the MUT can be calculated as: (3.6) and the value of for each voltage applied to the MUT can be calculated as: (3.7)

for each voltage

17

450 400 350 Motor Speed [rad/s] 300 250 200 150 100 50 0 0 1 5V 2 3 Motor Current [A] 15V 20V 4 y = -12.314x + 208.95 y = -11.857x + 136.81 y = -10.286x + 62.048 5 6 y = -11.543x + 277.19 y = -12.229x + 406.57

10V

30V

Figure 3.4 Speed versus armature current characteristics Shown in Table 3.2 are the parameters derived from the Table 3.2 Linearized data received from [ ]
[

( ) characteristics shown Fig. 3.4. characteristics [ ]

versus

5 10 15 20 30

-10.286 -11.857 -12.314 -11.543 -12.229

62.048 136.81 208.95 277.19 406.57

0.081 0.073 0.072 0.072 0.074

0.83 0.87 0.89 0.83 0.90

The average values are calculated as: 18

(3.8)

where

is the number of

( ) characteristics as seen in Fig. 3.4.

The calculated average values obtained from the steady-state characteristics are shown in Table 3.3. Table 3.3 Average values of motor parameters obtained from steady-state characteristics [ ]

0.0744

-11.646 218.314

0.864

Friction in the permanent-magnet DC Motor consists of two components, one that is constant and one that is dependent on the speed of the motor. The first friction component is the static friction torque which is constant throughout the motors range of operation. The second friction

component is proportional to the motor speed [5]. The motor torque equation considering these two friction forces in steady-state operation is: (3.9) To determine the friction of the permanent-magnet DC Motor the motor will be operated without a load; unburdened the motor must only overcome static friction torque and the friction represented

by the friction coefficient . As an increasing voltage is applied to the motor terminals, the current through the motor and the speed of the motor shaft is recorded [2]. From the recorded current the torque produced by the motor can be calculated as: (3.10) where: 19

(3.11)

The data obtained from the no-load test is enclosed in Table 3.4. Table 3.4 Steady-state friction determination data [ ] [ ]

[A] 0.344 0.385 0.415 0.446 0.475 0.508 0.546 0.570

5 10 15 20 25 30 35 39

59 126 193 260 328 395 462 515

0.0256 0.0286 0.0309 0.0332 0.0353 0.0378 0.0406 0.0424

The correlation between electro-magnetic torque and motor shaft speed is expressed as: (3.12) and can be expressed in y-intercept form where: (3.13) (3.14)

The torque versus speed characteristic for the data contained in Table 3.4 is plotted in Fig. 3.5. The y-intercept data obtained from the no-load characteristic shown in Fig. 3.5 is contained in Table 3.5.

20

Motor Torque [N*m]

Torque vs Speed
0.06 0.04 0.02 0 0 100 200 300 400 Motor Speed [rad/s] 500 600 y = 4E-05x + 0.0237

Figure 3.5 Torque-speed characteristic of steady-state friction model Table 3.5 Steady-state friction parameters B [slope] 0.00004 [intercept] 0.0237 [ 0.00004 ] [ ]

0.0237

3.2 Dynamic Parameters

Permanent-magnet DC motor dynamic parameter armature inductance, is an element of the motors electrical circuit and is a measure of resistance to change in the current flowing through the motors armature circuit . The dynamic parameter rotor moment of inertia, is an element of .

the motors mechanical assembly and is a measure of the motors resistance to change in speed The dynamic equations for the permanent-magnet DC Motor are:

(3.13) (3.14) Armature current inductance and angular speed are not independent of one another so both the armature

and moment of inertia will change with a variance in current or speed. To more and , the variables 21 and are isolated from one another. With

easily determine the values of

the variables isolated, two experiments can be performed to determine the electrical and mechanical dynamic characteristics of the permanent-magnet DC motor. To determine the inductance of the motors armature circuit the rotor of the permanentmagnet DC motor is blocked so that it does not move ( ). A step voltage is applied to the

armature winding so that the current increases exponentially with time until reaching the limit of (see Fig. 3.6) [2].

Figure 3.6 Current response with blocked rotor ( When the motors voltage equation is:

(3.15) which can be modified to find current: ( ) where: 22 (3.16)

(3.17) The slope of the curve as measured at as: ( ) (3.18) is dependent on . This dependency can be expressed

An examination of the current response during the blocked rotor test will reveal the armature inductance to be: (3.19)

where the slope is: ( ) (3.20)

Figure 3.7 Current waveform for 3V step increase 23

The current curve obtained during the blocked rotor test is shown in Fig. 3.7. The results of the blocked rotor test are show in Table 3.6 and inductance Table 3.6 Inductance of armature winding ( [ ] [ ] [ ] ) [ ] 0.0107 is found according to Eqn 3.19.

To properly determine the moment of inertia for the permanent-magnet DC Motor the effects of electro-magnetic torque must be negated. Disconnecting the motors power supply will effectively eliminate current flow and reduce the corresponding electro-magnetic torque to zero. Before removing power from the system, the motor-under-test is brought up to the no-load speed by a second motor such that current through the MUT is 0 and friction losses are compensated for by the alternate motor [2]. At 0. The dynamic torque equation: (3.21) becomes: (3.22) Once power is removed from the system the motor speed reduces exponentially with time and is expressed as: (3.23) power is removed from both motors and both and become

( )

The speed function expressed by Eqn. 3.23 is illustrated in Fig. 3.8. 24

Figure 3.8 System speed response upon shutdown The friction model is an approximation because the static friction torque but dependent on speed [5]. Once the motor speed reaches 0, can rearrange the above equation as: (3.24) is not always constant

also reaches 0. Knowing this we

( ( )

( )

Examining the above equation shows that by shifting the exponential curve with a constant value, the inertia is dependent on the slope of the curve at ( ) The moment of inertia estimating the slope of the curve. (3.26) ( ) and : (3.25) and graphically

can now be determined knowing

where: ( ) 25 (3.27)

The speed curve recorded during the test is shown in Fig. 3.9. The inertia of two motors J is determined with Eq. 3.26 using measurements taken from Fig. 3.9; this value must be halved to find the inertia of one motor JM.

Figure 3.9 Speed response after power supply is disconnected from motors The results of the no-load speed test are enclosed in Table 3.7.

Table 3.7 Results of equivalent inertia Jeq calculations [ ]

[ 379

1.71 0.0002437 0.000122

26

The inertia of the motor and its load should be considered in the designing of the elevators digital control system. This combined inertia can be found using Eqn. 3.28. (3.28) where: is the inertia of motor is the radius of drive pulley is the mass of elevator car is the inertia of the drive pulley The inertia of the drive pulley is calculated as [5]: (3.29) The data for the permanent-magnet DC motor and attached drive system for which the equivalent inertia was calculated for is enclosed in Table 3.8 and the result of the calculation is contained

with Table 3.9. Table 3.8 Data for determination of equivalent inertia

[ ] 0.000122

[ ]

[ ]

[ ] 0.000741

0.025 1.673

0.00487

The position, speed, and current control loops can now be designed using the unique values attributed to the permanent-magnet DC motor which will be used for the elevator electric drive. The design of these cascaded control loops is covered in Chapter 4. 27

Chapter 4: Designing of Control System

The digital control system for implementation of position, speed, and current control of the elevators electric drive is composed of three cascaded loops which are coordinated to work together using information obtained from the motors position sensor and power supplys current output. Knowing the motors steady-state and dynamic parameters the individual control loops are each designed to perform a specific task using a specific input. Using Matlabs Simulink software package these separate loops are integrated into one digital control system which is then compiled into machine language, loaded on the dSpace DS1104 hardware, and then executed in real time.

4.1 Feedback Control of Electric Drive

Feedback controllers are implemented with the goal of precisely and quickly regulating a system based upon real-time input obtained from the system itself without the need for correction from any outside observer. A properly designed control system will bring the steady-state error of a process to zero in a short time frame with few oscillations and minimum overshoot.

Figure 4.1 Feedback control system of DC motor Fig. 4.1 is a block representation of the elevators electric drive and feedback control system. A Direct Current (DC) power source is connected to the motor via a Power Processing Unit (PPU). The speed of the motor is captured by sensor and fed back to the controller. The position of the 28

motor is obtained by integrating the speed of the drive over time and is also fed back to the controller. The current flow out of the PPU is monitored and reported back to the controller as well. With these three real-time inputs the controller can bring the motor to the desired position. A closer examination of the controller block reveals three nested loops (see Fig. 4.2). Two loops contain a Proportional-Integral (PI) controller which is well suited to the task of regulating the electric drives current and speed and one loop that contains a Proportional (P) controller which will adequately correct errors in the drives position.

Figure 4.2 Cascaded control system Shown in Fig. 4.3 is a diagram of a PI controller. The difference between the desired input X*(s) and the actual output X(s) is known as the error E(s). Two corrective values are generated from the error E(s). The value Vc,p(s) is the product of the error E(s) and the gain kp; this provides a swift and proportional correction to system disturbances. The value Vc,i(s) is the product of the error E(s) and the gain ki with the result being integrated over time; this provides a slow and gradual correction to counter system disturbances. The two corrective values, one to ensure the system adequately follows the desired reference value and one to reduce steady-state error, are summed and applied to the plant Gp(s). The processs corrected output is compared to the reference input and the error E(s) is recalculated, continuing the cycle indefinitely.

29

Figure 4.3 PI Controller Diagram The error E(s) which is fed into the PI controller can be expressed as: ( ) ( ) ( ) (4.1)

The proportional gain multiplies the error E(s) times a constant kp; a larger error will produce a larger correction Vc,p(s) and a small error will produce a smaller correction Vc,p(s) as demonstrated in equation 4.2: ( ) ( ) (4.2)

The proportional gain acts in a linear fashion and can produce a steady-state error in response to a step-change so an integral gain is added to the controller so as to compensate for the proportional gains deficiencies. By responding over time to the systems changing output the integral controller is able to gradually reduce the steady-state error to zero. The correction produced by the integral controller Vc,i is expressed as: ( ) (4.3)

A derivative gain is not used for the control of the elevators electric drive because a derivate gain is susceptible to noise in the measurement system and could cause unwanted disturbances.

30

The open-loop transfer function of the PI controller seen in Fig. 4.2 is:
( ) ( )

(4.4)

4.2 Current Control Loop

The current control loop is the innermost and fastest loop of the controller. The current control loop calculates the difference between the reference current and the current being supplied to the motor by the PPU. The reference current value is obtained from the output of the speed controller and the value of the current being supplied to the motor is obtained by sampling the PPU output in real-time using an analog-to-digital converter on the dSpace 1104 hardware. Fig. 4.4 is a diagram of the current loop in its entirety. In addition to the PI controller, Fig. 4.4 includes the transfer function blocks which represent the armature winding and a transfer block which represents the mechanical system. The gain kPWM represents the PPU. Also shown in the diagram are two feedback loops, one is the current feedback loop and the other loop which is dependent on the motors speed accounts for the electro-motive force (emf) which opposes the armature voltage ( ). To simplify the design considerations of the current loops PI controller, the

current loop is reduced.

Figure 4.4 Expanded current control loop [6] 31

Fig. 4.5 is a diagram of the reduced current-loop shown in Fig. 4.4. To ease design considerations the load torque is omitted and the induced back-emf ( ) is calculated in terms of the current

( ). The reduced current loop can be further simplified to stream-line the design of the current loops PI controller.

Figure 4.5 Reduced current control loop [6] Fig. 4.6 is a diagram of the simplified current-loop shown in Fig. 4.5. Design considerations can be further reduced by assuming the equivalent inertia is sufficiently large enough to neglect the

effects of the induced back-electro-motor force which is inversely dependent on said inertia. The transfer blocks representing the electro-magnetic torque of the motor and the speed of the motor can also be eliminated from the diagram as they are not included in the current feedback loop.

Figure 4.6 Simplified current control loop [6]

32

Fig. 4.6 shows the simplified current control loop of the permanent-magnet DC motor. The loop first determines the error between the reference current and the actual current flowing through the motor. This calculated error is used by the current control loops PI controller to determine how the PPUs duty ratio must be changed in order to achieve a steady-state error of zero; the transfer function of the current control loops PI controller is given by Eqn. 4.4. The modified control voltage Vc(s) determined by the PI controller is multiplied by the gain of the 4-quadrant Pulse-Width-Modulation converter to produce the voltage Va(s) which is applied to the motors armature circuit. In Fig.

4.4 the permanent-magnet DC motors armature circuit is represented by a simplified mathematical model and is expressed by the transfer function: ( )

(4.5)

The open-loop transfer function of the simplified current control loop shown in Fig. 4.4 is:

( )

(4.6)

When designing the current control loops PI controller, the gains kpI and kiI are selected so as to achieve a phase margin of 90 for the controller. This is done by taking the open-loop transfer function of the current control loop (Eqn. 4.6) and setting the zero of the PI controller to cancel the motor pole as is done in Eqn. 4.7: (4.7)

This leads to Eqn. 4.8: (4.8)

33

is the electrical time constant of motor which the time required for current to reach 63.2% of its final value at a fixed voltage level. The electrical time constant of the motor can be calculated as: (4.9) With the motors pole canceled by the controllers zero, Eqn. 4.6 is reduced to: ( ) where: (4.11) Also, when designing the current control loops PI controller the choice of crossover frequency is (4.10)

an important choice. The crossover frequency of the open-loop gain is chosen to be approximately one to two orders of magnitude less than the switching frequency of the PPU so that any noise from switching operations will not affect the current control loops performance. The cross over frequency of the current loop is expressed in radians per second as: (4.12) The magnitude of the open-loop transfer function is equal to unity at the crossover frequency | ( )| :

(4.13) can be calculated

Thus following from Eqn. 4.10 and Eqn. 4.13, the gain of the integral controller as:

(4.14) Knowing as: 34 , the gain of the proportional controller can be determined by rearranging Eqn. 4.8

(4.15) The data for the permanent-magnet DC motor and PPU for which the current control loop was designed for is enclosed in Table 4.1 [1]. Table 4.1 Parameters for current control loop design [ ] [ ] [ ] 40 [ ] [

0.864 0.0107 The results of the and

2000

238

1495.4

calculations are enclosed within Table 4.2

Table 4.2 Torque control loop integral gain calculations 32.3 0.40

4.3 Speed Control Loop

The cascaded control systems middle loop is the second fastest loop and dictates at what

speed the electric drive should be turning. The speed control loop calculates the difference between the reference speed and the actual speed of the drive which is obtained from a sensor mounted on the motors shaft. This calculated error in speed is manipulated by the speed loops PI controller to produce a reference current value for the current control loop. Choosing a crossover frequency for the speed control loop that is at least one order of magnitude less than the crossover frequency of the current control loop reduces design constraints and ensures the nested loops do not conflict with one another. For the purpose of designing the speed control loop the closed current-loop is assumed to be ideal and is represented as a unity gain.

35

Figure 4.7 Simplified current control loop The open-loop transfer function of the speed loop shown in Fig. 4.7 is: ( ) where: is the integral gain of the speed controller is proportional gain of the speed controller Combining terms, Eqn. 4.16 can be rearranged as: ( ) ( )
( )

(4.16)

(4.17)

An examination of Eqn. 4.17 shows that the open-loop transfer function contains a double pole at the origin. At low frequencies this double pole at the origin reduces the magnitude of the transfer function at a rate of -40db per decade while the phase angle is at -180. This effect can be mitigated by choosing a crossover frequency that is at least one order of magnitude smaller than the crossover frequency of the current control loop and by choosing a phase margin of 60. At the crossover frequency Eqn. 4.17 yields Eqn. 4.18 and Eqn. 4.19: ( ) ( )
( )

(4.18)

36

[ From Eqn. 4.19 the integral gain is:

)]

(4.19)

( which can be rearranged as: ( and the proportional gain is: ( which can be rearranged as:

(4.20)

(4.21)

(4.22)

(4.23)

The data for the permanent-magnet DC motor and attached elevator for which the speed control loop was designed for is enclosed in Table 4.3. Table 4.3 Parameters for speed control loop design

[ 0.0744 The results of the and

[] 60

[ 20

0.000741

125.644

calculations are enclosed within Table 4.4

Table 4.4 Results of speed loop gain calculations 78.639 1.084

37

4.4 Position Control Loop

The control systems outer loop is the slowest loop and ensures the elevators electric drive is at the desired set point. The position of the drives shaft and the connected elevator is determined by integrating the drives speed; this is the same speed value which is fed back to the speed control loop. From the integrated speed, the position control loop determines the error between the desired set point and the actual position of the drive and reduces this error in position by changing the value of the reference speed. The position control loops crossover frequency must be an order of magnitude less than that of the speed control loop which must be less than that of the current control loops crossover frequency. For the purpose of designing the position control loop the closed speedloop is assumed to be ideal and is represented as a unity gain. The position control loop is different from the speed and current loops in that it only has a proportional gain. The use of an integrator to obtain position from the speed data will reduce the steady-state error to zero for a step-change in the reference position.

Figure 4.8 Position control loop block diagram The open-loop transfer function of the position control loop shown in Figure 4.8 is: ( ) where is the proportional gain of the position controller. (4.24)

At the cross-over frequency the transfer function for the position controller is: 38

( Since from Eqn. 4.25

(4.25)

(4.26) The phase margin can be found as: (4.27) The cross-over frequency is selected to be 1 Hz, hence the value of the gain shown in Table 4.5 Table 4.5 Position loop proportional gain calculations

[Hz] [rad/s] 1 6.283 6.283

With the design of the cascaded control system accomplished, the performance of the electric elevator drive can be examined. Chapter 5 covers the initial computer simulation of the DSP control and permanent-magnet DC motor as well as the experimental testing and results obtained in the laboratory.

39

Chapter 5: Performance of Electric Elevator Drive with Position Control

Upon determining the parameters of the permanent-magnet DC motor and designing the cascaded control system, a model of the system is created in Matlabs Simulink package. This model is used to simulate and observe the expected behavior of the system. Once the desired response is obtained during simulation the elevator electric drive with position control can be implemented in real time. The Simulink model used for simulation is modified to accept input and output from the dSpace 1104 hardware and then compiled into machine code which can be loaded on to the dSpace 1104 hardware. Position commands are issued from the Control-Desk software and the drives response can be visually observed in addition to position data being collected by the dSpace 1104 hardware from the permanent-magnet DC motors position encoder. Actual performance of the elevators electric drive is compared against the simulated performance of the elevators electric drive and observations are made as to how adequate the position control of the elevators electric drive is. Each of these steps will be conducted by students in the course on electric machines so that they may experience the development of position control systems for electric drives.

5.1 Performance of Simulated Electric Elevator Drive with Position Control

For computer simulation the permanent-magnet DC motor parameters determined in Chapter 3 and the cascaded control system designed in Chapter 4 are used to build a Simulink model in Matlab which mimics the performance of the actual physical model.

40

Figure 5.1 Simulink model of elevator electric drive with position control Fig. 5.1 shows the cascaded control system and DC drive constructed in Matlabs Simulink allowing for simulation of the elevators electric drive position control. Contained within this model is a subsystem which emulates the permanent-magnet DC motor used to drive the system in addition to the position, speed, and current control subsystems; the load torque placed on the permanent-magnet DC motor is determined within its on subsystem. The position input is generated by a signal block which produces a step change in reference position. Also seen in Fig. 5.1 is an integrator which obtains position from the mechanical speed of the motor and also seen is the gain represents the gain of the PPU. The PPUs output voltage supplys voltage and the duty-ratio which

is the product of the DC power

generated by the current controller. The block diagram seen

in Fig. 5.2 is a graphical representation of Eqn. 5.1 and Eqn. 5.2.

Figure 5.2 Simulated DC machine subsystem 41

( )

( )

( )

(5.1) (5.2)

( )

( )

where: ( ) ( ) ( ) ( ) (5.3) (5.4)

The position, speed, and current control loops which were designed in Chapter 4 are each contained within their own subsystem. Each control subsystem calculates the error between the reference value and the actual value. Each control subsystem also contains the appropriate proportional and integral gains in addition to a saturation block which accounts for physical limitations such as the maximum current output of the power supply or the maximum speed at which the elevator can safely operate. Fig. 5.3 shows the current controller and Eqn. 5.6 describes its output. Fig. 5.4 shows the speed controller and Eqn. 5.7 describes it output. Fig. 5.5 show the position controller and Eqn. 5.8

describes it output. Eqns. 5.7, 5.8, and 5.9 notate the reference value by use of an asterisk (*).

)(

(5.5)

Figure 5.3 Simulated current controller subsystem

42

)(

(5.6)

Figure 5.4 Simulated speed controller subsystem

(5.7)

Figure 5.5 Simulated position controller subsystem The load torque placed on the drive is a function of the drives speed and the mass of the load placed in the elevator car; the mass of the elevator car and the mass of the counterweight are equal so only the mass of the object placed within the elevator car shall be considered the Load. According to Eqn. 5.9 and Eqn. 5.10 (which were derived in Chapter 2, see Eqn. 2.14 and Eqn. 2.15) the calculation of load torque for the upwards motion is different from that of the downwards motion, for this reason the load torque subsystem for the ascending elevator car is different from the load torque subsystem for the descending elevator car. For the upward motion, the loads mass is added to the total load torque as a function of increased car mass and increased load inertia. For the downward motion, the loads mass detracts from the total load torque and the loads inertia is not considered as 43

only the counter-weights inertia in considered. These similar yet different subsystems are illustrated in Fig. 5.6 and Fig. 5.7 with their corresponding equations.

(5.8)

Figure 5.6 Simulated load torque subsystem for ascending elevator car

(5.9)

Figure 5.7 Simulated load torque subsystem for descending elevator car

44

During simulation the position reference signal is generated using a signal builder block which produces a step change in desired height over time. This step input emulates the desired position change by a button press as elevators in the real world often have a button for each floor. Fig. 5.8 shows the elevator car rising from 0 to a height of 1 meter. Fig. 5.9 shows the elevator car descending from a height of 1 meter to 0.

Figure 5.8 Simulated reference position of elevator electric drive with ascending car

Figure 5.9 Simulated reference position of elevator electric drive with descending car The response of the elevator electric drive for the elevator car containing a 1kg load is shown in Fig. 5.10 through Fig 5.19 for both the ascending and descending motions. 45

Figure 5.10 Simulated position response of elevator electric drive with ascending car

Figure 5.11 Simulated position response of elevator electric drive with descending car 46

Figure 5.12 Simulated speed response of elevator electric drive with ascending car

Figure 5.13 Simulated speed response of elevator electric drive with descending car

47

Figure 5.14 Simulated current response of elevator electric drive with ascending car

Figure 5.15 Simulated current response of elevator electric drive with descending car 48

Figure 5.16 Simulated torque response of elevator electric drive with ascending car

Figure 5.17 Simulated torque response of elevator electric drive with descending car 49

Figure 5.18 Simulated voltage response of elevator electric drive with ascending car

Figure 5.19 Simulated voltage response of elevator electric drive with descending car 50

From Fig. 5.10 and Fig. 5.11 it is observed that the elevator is able to move a load a distance of one meter in under three seconds. The speed at which the elevator changes position is shown in Fig. 5.12 and Fig. 5.13. Starting from a stand-still, the motor must overcome inertia and cannot immediately meet the reference speed issued by the speed controller; instead the motor accelerates to and temporarily exceeds the speed limit before settling at its maximum reference value of 25 - this

temporary overshoot in speed causes a rise in the voltage applied to the motors terminal by the PPU (see Fig. 5.18 and Fig. 5.19) as it counters the back-emf . As the elevator car approaches the

desired position the reference speed decreases exponentially to zero and the elevator car comes smoothly to a stop. The current response to the step-change in position is immediate (see Fig. 5.14 and Fig. 5.15) as is the electro-magnetic torque produced by the motor (see Fig. 5.16 and Fig. 5.17). The PPUs ability to instantaneous switch on (see Fig. 5.18 and Fig. 5.19) ensures that current through the motor follows closely with the reference current value. To generate the maximum amount of torque at start-up, the reference current is at its maximum value until the motors speed matches the reference speed at which point current through the motor is reversed to obtain braking torque in order to prevent the elevator car from accelerating beyond the speed limit. During steadystate operation current remains nearly zero as only friction and gravitational forces must be overcome. As the elevator car approaches the desired set point the motor must once again overcome inertia but in the opposite way as before. To start reducing the speed of the car a braking torque is applied; this torque starts at its maximum value and diminishes exponentially as kinetic energy is bled off from the system. Current once again returns to zero as the car glides smoothly to a stop and the system remains at rest. This process is repeated three times for no load, 1kg load, and 2kg load; the position response under these conditions is shown in Fig. 5.20 and Fig. 5.21.

51

Figure 5.20 Simulated position response of elevator electric drive for ascending car carrying no load, 1kg load, and 2kg load.

Figure 5.21 Simulated position response of elevator electric drive for descending car carrying no load, 1kg load, and 2kg load 52

From Fig. 5.20 and Fig. 5.21 it is observed that the addition of a load extends the time during which the motor must accelerate the system from a stand-still, but once in motion the elevator car moves at the maximum allowed speed regardless of what it carries. It is also seen that the car with a load decelerates in less time than the car without a load. It is evident from the data listed in Table 5.1 that placing a load in the elevator car does increase the time it takes for the system to respond to a step change in desired height. It is also shown that the additional time needed is proportional to increased load that added to the car. Table 5.1 Simulated position response of electric elevator drive with no load and 1kg, 2kg load Response Time when Moving Response Time when Moving Upwards from 0 meters Downwards from 1 meters [s] [s] To 0.5 meters 1.0 meters 0.5 meters 0.0 meters Height: Load = 0.848 2.712 0.847 2.598 0 kg Load = 0.882 2.745 0.873 2.625 1 kg Load = 0.980 2.843 0.950 2.699 2 kg Table 5.2 gives the additional time needed for each load as a percentage of the response time for the car without a load. Table 5.2 Percent change in simulated position response of electric elevator drive under load Up Down To Height: 0.5 meters 1 meter 0.5 meters 0 meters 0.0% 0.0% 0.0% 0.0% No Load 4.0% 1.2% 3.1% 1% 1kg Load 15.6% 4.8% 12.2% 3.9% 2kg Load

53

Based upon the information contained in Table 5.2, the elevator electric drive and digital control system are able to compensate for loads which are placed in the elevator car and that the variations in response time induced by loads are reduced over greater distances of travel.

5.2 Performance of Experimental Electric Elevator Drive with Position Control

With the simulated performance of the elevators electric drives position control recorded, the lab experiment is conducted using an elevator model built by the author (see Fig. 5.22). The physical model differs from the simulation model in that the physical model possesses two safety cut-off switches in the event that the elevator car or counter-weight exceeds it range of travel. These safety switches are mounted on the bottom of the elevators cover and are connected to an analog-todigital converter on the dSpace 1104 hardware. A close-up of the safety switch is shown in Fig. 5.23

Figure 5.22 View of experimental elevator in lab 54

Figure 5.23 Safety cut-off switch The permanent-magnet DC motor which serves as the elevator electric drive is shown in Fig. 5.24 and is attached to the drive pulley with the drive belt looped over the pulley.

Figure 5.24 View of elevator electric drive and drive pulley with drive belt 55

Fig. 5.25 shows the PPU which provides power to the permanent-magnet DC motor and Fig. 5.26 shows the dSpace 1104 hardware which contains the analog-to-digital converters, position encoder input, and micro-processor which executes the compiled code [1].

Figure 5.25 Elevator electric drives power processing unit

Figure 5.26 dSpace 1104 hardware 56

The real-time cascaded control system which is compiled and loaded onto the dSpace 1104 hardware is shown in Fig. 5.26. The value of the current as measured by the analog-to-digital converter is passed to the current controller by the ADC block; the motors speed as measured by motors position encoder is fed into the control system by the position encoder block. The three control loops shown in Figs 5.27 through 5.29 each have a saturation block which accounts for system limits; the duty ratio generated by the current controller is limited to 1, the reference current generated by the speed controller is limited to the maximum value of current which the power supply can produce, and the reference speed generated by the position controller is limited to the maximum speed with which the elevator can safely operate. The integration block shown in Figs. 5.26, 5.27 and 5.28 can be reset to return the system to its initial conditions. Fig. 5.31 shows the duty cycle generation subsystem which creates two duty cycles from one. These two duty cycles are used to drive the pulse width modulation block which directs the PPUs output.

Figure 5.27 Real-time Simulink model for position control of elevator electric drive 57

Figure 5.28 Real-time current controller subsystem

Figure 5.29 Real-time speed controller subsystem

58

Figure 5.30 Real-time position controller subsystem

Figure 5.31 Real-time duty cycle generation subsystem

Figure 5.32 Real-time speed averaging subsystem

Before compiling the real-time model, the simulation parameters must be correctly set according to the values shown in Table 5.3. 59

Table 5.3 Real-time simulation parameters Start Time Stop Time Solver Options Fixed Step Size Mode 0.0 Inf Fixed Step ode1 0.0001 Auto The Control Desk screen which is shown in Fig. 5.32 is used to input the desired position and to observe the performance of the elevators electric drive. The instructions on creating the control desk interface and utilizing the interface are contained with the Variable Speed Drives course lab manual [2].

Figure 5.33 Control desk control interface for position control of elevators electric drive With the digital control system compiled and loaded into the dSpace 1104s microprocessor, the elevator car loaded with 1kg is directed via the Control Desk screen to ascend 1 meter while data is recorded. After the elevators car comes to a rest at a height of 1 meter, it is directed to return to the original height of 0 meters while data is again recorded. From the measurements made during these test the performance of the elevators electric drive can be visually displayed. Fig. 5.33 and Fig. 5.44 show the elevator cars movement upon receiving a change in desired position. The speed at which the motor operates while the car is in motion is shown in Fig. 5.35 and Fig. 5.36 while the current through the motor is shown in Fig. 5.37 and Fig. 5.38. 60

Figure 5.34 Real-time position response of elevator electric drive with ascending car

Figure 5.35 Real-time position response of elevator electric drive with descending car 61

Figure 5.36 Real-time speed response of elevator electric drive with ascending car

Figure 5.37 Real-time speed response of elevator electric drive with descending car 62

Figure 5.38 Real-time current response of elevator electric drive with ascending car

Figure 5.39 Real-time current response of elevator electric drive with descending car 63

Testing with the experimental elevator electric drive with position control shows that the elevator car can be moved to a desired set point in a smooth and accurate fashion (see Fig. 5.34 and Fig. 5.35). When moving upwards or downwards, the car accelerates rapidly to the steady-state limit of 25 and remains at this speed until the speed controller begins exponentially reducing the reference speed (see Fig. 5.35 and Fig. 5.36). The actual speed never reaches a steady-state error of less than 10% which induces rapid changes in the reference current as shown in Fig 5.38 and 5.39. The actual current drawn by the motor does change with the reference current but the armature windings inductance dampens the magnitude of change. The elevators electric drive system is tested with three variations in load weight, moving both upwards and downwards. The position response of the system for these three variations in load are plotted in Fig. 5.20 for the elevator car moving upwards and in Fig. 5.21 for the elevator car moving downwards.

Figure 5.40 Real-time position response of elevator electric drive for ascending car carrying no load, 1kg load, and 2kg load 64

Figure 5.41 Real-time position response of elevator electric drive for descending car carrying no load, 1kg load, and 2kg load The real-time results of the elevator electric drive show that the addition of a payload to the elevator car has greater affect on the ascending motion than on the descending motion. Despite the increase in inertia when moving upwards, the loaded car does not greatly affect the position response when ascending and descending. Table 5.3 Real-time position response of elevator electric drive with no load and 1kg, 2kg load Response Time when Moving Response Time when Moving Upwards from 0 meters Downwards from 1 meters [s] [s] To 0.5 meters 1.0 meters 0.5 meters 0.0 meters Height: Load = 0.815 2.654 0.807 2.600 0 kg Load = 0.834 2.661 0.813 2.609 1 kg Load = 0.865 2.670 0.814 2.618 2 kg 65

Table 5.4 Percent change in real-time position response of electric elevator drive under load Up Down 0.5 meters 1 meter 0.5 meters 0 meters 0.0% 0.0% 0.0% 0.0% No Load 2.3% 0.3% 7.4% 0.3% 1kg Load 6.1% 6.0% 8.7% 3.9% 2kg Load

5.3 Comments on Performance of Electric Elevator Drive with Position Control

The elevator electric drive with position control has been implemented successfully first as a computer simulation and then experimentally in the lab. The cumulative results of the two procedures are listed in Table 5.5. Table 5.5 Performance of Elevator Electric Drive Response Time when Moving Elevator Upwards Car [s] Load Simulation Experimental 2.712 2.654 0 Kg 2.745 2.661 1 Kg 2.843 2.670 2 Kg

Response Time when Moving Downwards [s] Simulation Experimental 2.598 2.600 2.625 2.609 2.699 2.618

The experimental performance of the electric elevator drive closely followed that of the simulated electric elevator drive. In computer simulation, the speed and current loops both had minimal overshoot with a quick settling time while the position control loop had no overshoot and arrived at the desired value as fast as the speed limiter would allow. During experimental testing, the position control loop had no overshoot and arrived at the desired height as fast the the speed limiter would allow; however, the speed and current control loops never reached steady-state variables due to inaccurate readings from the motors position encoder. Despite hardware issues, the experimental electric elevator drive was able to meet the desired set-point in less time than predicted by computer simulations. 66

Chapter 6: Conclusions
The implementation of position control of an elevators electric drive was analyzed in this thesis. Forces involved in the mechanical system were determined allowing for the calculation of torque generated by the motor and torque placed on the motor by the load. The electric circuit parameters and mechanical system parameters of the permanent-magnet DC motor were empirically derived using a series of measurements collected from experimental test conducted in the laboratory. Using these parameters the current, speed, and position control loops were designed for the elevators electric drive. The cascaded control system and permanent-magnet DC motor were modeled in Matlabs Simulink package and position control of the elevators electric drive was simulated for no load conditions and loaded conditions. With the simulated performance of the elevators electric drives ability to arrive at a desired position recorded, testing was conducted in the lab using the permanent-magnet DC motor and scale elevator constructed by the author to collect data on the elevator electric drive to arrive at the desired height in a timely and accurate manner. Using data collected from the simulations and testing, the conclusions from this analysis of position control of an elevators electric drive are: It is necessary to determine the electrical and mechanical parameters of an electric motor in order to correctly design control loops which direct the electric motor to the desired position.

The design of cascaded control loops which direct the electric drive to the desired position must consider the parameters of the drive and the system which it is driving; the individual loops of the cascaded control system must also be coordinated so as to not conflict with one another.

67

The computer simulation of the system is an approximation of the actual system and provides an expedient method by which the performance of the system can be predicted. Verifying simulation results with data collected from real-time experiments validates the model and theory on which it is built allowing for rapid development of new applications.

Applying the techniques found within this thesis, a student is able to design and implement digitally controlled electric drives for an unlimited number of applications. The chapter covering the determination of motor parameters is not inclusive of all types of electric drives and the chapter on designing a digital control system does not compensate for all variables a dynamic system may encounter but it does provide a blueprint for upgrading systems which may have older drives and new control systems or for replacing an existing mechanical with an electrical one. Performing the steps of determining motor parameters, designing cascaded control systems, and collecting data in the lab provides a tangible learning experience for students enrolled in LSUs Electrical Engineering course on electric drives.

68

References
[1] dSPACE. DS1104 R&D Controller Board Hardware Installation and Configuration. Users Manual, March 2004 [2] Gunda, Kiran K. Adjustable Speed Drives Laboratory Based on dSPACE Controller. MS Thesis. Louisiana State University, Baton Rouge, May 2008 [3] Hakala, Harri. Integration of Motor and Hoisting Machine Changes the Elevator Business. International Conference on Electrical Machines, Espoo, 2000, pp.1241-1243 [4] KONE. 2012 Architectual Planning Guide. Brochure, 2011 [5] Mendrela, Ernest A. Lecture notes on Variable Speed Drives. Louisiana State University, Baton Rouge, December 2009 [6] Mohan, Ned. Electric Drives: An Integrative Approach. Minneapolis, MN: MNPERE, 2003 [7] Otis. Gen2. Brochure, 2011 [8] Otis. About Elevators. Brochure, 2011

69

Appendix A: Load_Simulation_Parameters
Ra=0.864 La=0.0107 Ke=0.0744 Kt=0.0744 B=0.00004 %Friction coefficent J=0.000741 %equivalent inertia of entire system Jp=0.000487 %inertia of drive pulley r=0.025 %drive pulley radius in meters g=9.8 %gravitational constant Mc=1.67 %mass of car Mcw=1.67 %mass of counter-weight Kpwm=40 %Volts lim_Ia=5 %Amps lim_Va=1 %Duty Cycle lim_w=25 %radians/second Ki_I=32.3 Kp_I=0.40 Ki_w=78.639 Kp_w=1.084 Kp_p=6.283

70

Appendix B: Plot_Simulation
load Position P; load Ref_Position R_P; load Speed w; load Ref_Speed R_w; load Current Ia; load Ref_Current R_Ia; load Voltage Va; load Torque Tem; clf figure (1) plot (R_P,'r','linewidth',2) hold on plot (P,'--','linewidth',2),title('Position Control Loop'), xlabel('time [s]'),ylabel('Position [m]') figure (2) plot(R_P,'k','linewidth',2) hold on plot (P,':k','linewidth',2) hold on plot (R_w,'r','linewidth',2) hold on plot (w,'--','linewidth',2),title('Speed Control Loop'), xlabel('time [s]'),ylabel('Motor Speed [rad/s]') figure (3) plot(R_P,'k','linewidth',2) hold on plot (P,':k','linewidth',2) hold on plot (R_Ia,'r','linewidth',2) hold on plot (Ia,'--','linewidth',2),title('Current Control Loop'), xlabel('time [s]'),ylabel('Motor Current [A]') figure (4) plot(R_P,'k','linewidth',2) hold on plot (P,':k','linewidth',2) hold on plot (Tem,'--','linewidth',2),title('Torque'), xlabel('time [s]'),ylabel('Electro-Magnetic Torque [N*m]') figure (5) plot(R_P,'k','linewidth',2) hold on plot (P,':k','linewidth',2) hold on plot (Va,'--','linewidth',2),title('Armature Voltage'), xlabel('time [s]'),ylabel('Armature Voltage [V]') 71

Appendix C: Plot_RealTime
load Position Pos; load Speed w; load Current Ia; t = Pos (1,:); r = Pos (2,:); a = Pos (3,:); figure (1) plot (t,r,'linewidth',3) hold on plot (t,a,:,'linewidth',3),grid,title('Position Control Loop'), xlabel('Time [s]'),ylabel('Position [m]') figure (2) t = w (1,:); r = w (2,:); a = w (3,:); plot (t,r,'r','linewidth',1) hold on plot (t,a,:,'linewidth',2),grid,title('Speed Control Loop'), xlabel('time [s]'),ylabel('Motor Speed [rad/s]') figure (3) t = Ia (1,:); r = Ia (2,:); a = Ia (3,:); plot (t,r,'r','linewidth',2) hold on plot (t,a,:,'linewidth',2),grid,title('Current Control Loop'), xlabel('Time [s]'),ylabel('Current [A]')

72

Vita
Peter James Ford was born in West Monroe, Louisiana. He completed his Bachelor of Science in Electrical Engineering at Louisiana State University Agricultural and Mechanical College in 2009. In August 2010 Peter returned to LSU to pursue his graduate studies; he is a candidate for the degree of Master of Science in Electrical Engineering for May 2012.

73