May 12, 2005

Department of Electrical and Computer Engineering

University of Idaho

Sponsors & Mentors

Dr. Herb Hess

Dr. Brian Johnson

Instructor

Dr. Brian Johnson

Flywheel Team

Gavin Abo

Nate Stout

Nathan Thomas

TABLE OF CONTENTS

TABLE OF FIGURES AND TABLES ii

Abstract iii

1. Project Description 1

1.1 Background Information 1

Objectives 1

1.2 Significance of the project 1

1.3 Methods 1

2 Project Description 2

2.1 Objectives 2

2.2 Significance of Project 2

2.3 General Work Plan 3

2.3.1 Data and Characterization 3

2.3.2 Output Regulation 3

2.3.3 Inverting the DC Signal from the Fuel Cell 4

2.3.4 Simulation and Interfacing 4

2.4 Methods and Procedures 4

2.4.1 Data and Characterization 4

2.4.2 Output Regulation 5

2.4.3 Inverting the DC Signal from the Fuel Cell 5

2.4.4 Simulation and Interfacing 6

2.5 Additional Considerations 6

2.6 Technical Advisor 7

3 Bibliography 7

4 Credentials 8

5 Time Schedule 8

6 Budget 9

7 Safety 10

APPENDIX A: Credentials A1


List of Figures and Tables

Figure 1 – Status organizational chart…………………………………………………….4

Figure 2 – Light fixture positioning……………………………………………………….8

Figure 3 – Camera connections…………………………………………………………..10

Figure 4 – Effects of polarization and filtration………………………………………….15

Figure C1 – Bitmap 1 image……………………………………………………………..21

Figure D1 - Pixel summing technique (Flow Chart)…………………………………….22

Figure D2 - Window-shopping technique (Flow Chart)…………………………………23

Figure D3 - Window checking (Flow Chart)…………………………………………….24

Figure D4 - Subtraction of scan row from window (Flow Chart).………………………25

Figure D5 - Counting of separate defects (Flow Chart)…………………………………26

Table 1 – Budget analysis………………………………………………………………..16

Table C1 – RGB analysis………………………………………………………………...20

Table C2 – Analysis time tests…………………………………………………………...20

Table C3 – Defect Locations…………………………………………………………….21

ii

Abstract

Title: Interfacing a Flywheel to the Analog Model Power System

Authors: Gavin Abo, Nate Stout, and Nathan Thomas

Date: May 13, 2005

Department of Electrical and Computer Engineering

University of Idaho

Background

In the mid-1990s, the University of Idaho acquired the Analog Model Power System (AMPS) from Idaho Power for educational and research use [1]. Idaho Power is an electric utility provider to about 883,000 people in southern Idaho and eastern Oregon [2]. The AMPS system was originally constructed by Idaho Power to test relays and breakers for equipment and system protection. In addition, the AMPS system was used to model part of Idaho Power’s own transmission and distribution system. Over the years, the University of Idaho has made several modifications to the donated system to incorporate the following:

1. A fault matrix, in which three faults can be placed on the system either simultaneously or in an evolving manner.

2. The ability to load impedance faults.

3. SEL (Schweitzer Engineering Laboratories) relays for system protection.

The AMPS is currently located in the basement (room G10) of the Buchanan Engineering Laboratory (BEL) on the University campus in Moscow, Idaho. It has been, and still is a valuable tool for students and researchers by providing insight into the workings of a power transmission system. The capacity of this system is continually being increased by the addition of subsystems such as the current topic of interest, a flywheel voltage sag correction system.

A flywheel is well known for efficient mechanical energy storage in its rotating momentum, which can then be applied to a generation source that converts mechanical energy into an electrical energy output to a system. Thus, a flywheel is expected to be a practical alternative power source for inline (series) voltage sag correction for the AMPS. The energy storage system should be useful in keeping the voltage on the system within given tolerances and protecting equipment that is required to operate within a very narrow voltage range. Furthermore, it could provide a model for observation and analysis, which would be valuable as an educational and research tool.

Satish Samineni, a past graduate student at the University of Idaho, started the flywheel sag correction project. He began the project with a model simulation that showed that the project was feasible using PSCAD computer simulation. Our project is to actually build and implement in hardware the model simulation that Satish completed for his Master of Science [4]. However, some modifications to his design must be made since the simulation design was for a shipboard power system rather than the AMPS system [5].

Problem Statement

The AMPS currently does not have the ability to correct for voltage sags on the model system.

Objectives

To further improve the capabilities of the AMPS, a flywheel voltage sag correction system will be interfaced to AMPS to automatically correct for voltage sags.

Constraints

The system will only have to account for balanced sags (equal voltage drops on all three phases), however it must be able to be upgraded for unbalanced sag correction in future modifications. It will also have to be able to run continuously, correcting for sags when they occur and keeping energy in the flywheel the rest of the time. The system must be capable of bi-directional power flow in for switching from running the motor from AMPS to putting voltage on AMPS to correct for a voltage sag.

The system will have a maximum voltage it can provide to the AMPS. Because our flywheel will be losing energy when correcting sags, it will not be able to correct for a sag indefinitely. After correcting for a sag, the flywheel will have to be brought up to rated speed again. This will limit how fast we can correct for cascaded sags. There are no specific size or weight requirements that have to be met.

Functional Specifications

The system will interface a flywheel to the AMPS. It will automatically correct for voltage sags that can be initiated in the AMPS. This will further facilitate the learning experience for students and enable them to experiment with different technologies in the power industry.

Solution

The system will require two AC/DC converters. One will act as a rectifier and the other as an inverter depending on which way the power is flowing. Each converter requires six Integrated Bipolar Gate Transistors (IGBT’s) and a Digital Signal Processor (DSP). Tier Electronics will probably provide the converters.

Figure 1 - Signal Flow for a Detected Sag

The system will start and keep the flywheel spinning. It will use space vector Pulse Width Modulation (PWM) to control the motor while for spinning the flywheel. When a sag is detected, the flow of power will be reversed, and the flywheel will be turning the motor as a generator for the duration of the sag. During this time, the converters will be using sine wave PWM to control how much voltage we put on the AMPS. When the sag is over, the system will return the default setting of the motor spinning the flywheel.

Status

Currently the simulation by Satish of the entire sag correction system is designed and working. Team Hydrofly also has working simulations of space vector modulation and the sag detector. Capacitor shorting bars are also designed, tested, and working in order to make sure the capacitors are not holding a charge while being worked on, as seen in Picture 3 in Appendix B.

Method of Solution

The system will interface the flywheel to the AMPS using two AC/DC converters using IGBT’s with anti-parallel diodes as seen in Figure 4 in Appendix A. Each converter will have a DSP that will control the modulation and monitor system voltages and currents closest to the side it is in control of. When correcting for a sag, the correction voltage will be kept in phase with the AMPS voltage by using a phase lock loop (PLL).

The system will use two forms of modulation, depending on which way the power needs to be flowing. Each form of modulation is used to control the voltage output to the lines. Space vector will be used for providing voltage to the motor because it uses the DC bus more efficiently, gives fewer harmonics, and it is specifically used in variable speed drive applications. The speed of the motor needs to be varied to control the energy flow into the flywheel. Sine wave modulation will be used to control the correction voltage because it has a higher switching frequency and it can be operated independently on each phase. This is important for future modification of this design.

When the system is spinning the flywheel, board 2 will be a rectifier and board 1 will be an inverter (Figures 5&6 in Appendix A). Board 1 will be using space vector PWM to control the motor. This scheme will be using different combinations of eight basic vectors, spaced 60 degrees apart, to form our desired vector (Figure 2 below).

The DSP’s will have to be reprogrammed to work for the design. Board 1’s DSP will be programmed with space vector modulation. Board 2’s DSP will be programmed with sine wave modulation. The sine wave will be generated with a look-up table and the triangle wave will be generated with a counter. A phase lock loop will be used to keep the injected voltage needs in phase with the AMPS voltage.

Theoretical Basis

A voltage sag is a short term drop in voltage. A drop of only 10% can cause sensitive loads to misoperate or shut down completely (thesis). Process and fabrication plants take a lot of time to restart after shutting down completely. They lose production time and therefore lose money.

Loads that draw large starting currents being connected into the system or electrical faults are the most common causes of voltage sags (thesis). A flywheel can put voltage sags into a system so that these critical loads never see the sag.

A flywheel is well known for efficient mechanical energy storage in its rotating momentum, which can then be applied to a generation source that converts mechanical energy to electrical energy output to a system. Thus, a flywheel is expected to be a practical alternative power source for inline (series) voltage sag correction for the AMPS. The energy storage system should be useful in keeping the voltage on the system within given tolerances and protecting equipment that is required to operate within a very narrow voltage range. Furthermore, it could provide a model for observation and analysis, which would be valuable as an educational and research tool.

A flywheel stores an amount of energy proportional to the moment of inertia and the rotational speed squared.

(1.1)

The moment of inertia for our flywheel is:

(1.2)

Once the flywheel is spinning, very little energy is required to keep it spinning. Space vector modulation is used to control the speed of the induction motor, which in turn controls the energy going to the flywheel.

Space vector modulation involves making an abc to dq0 transformation using the following matrix where Sas, Sbs, and Scs are the three-phase vector, and Sqs, Sds, S0s is the two phase equivalent.

(1.3)

Different combinations of the eight vectors created by switching are used to make the desired vector as seen in Figure 2. This desired vector represents the three-phase voltage we need on the motor.

Figure 2: Space Vector PWM Vector Diagram

Sine wave modulation involves overlapping a sine wave with a triangle wave and running them through a comparator. Whenever the sine wave has a higher value than the triangle wave, the comparator outputs a logic one. If the triangle wave has a higher value than the sine wave, the comparator outputs a logic zero. This makes a variable width square wave, as seen in Figure 3 below, with an underlying sine component.

Figure 3: PWM


Test Plan

‮ Measure the DC bus voltage with a multimeter and compare to board measurement.

‮ Verify switching sequence with an oscilloscope.

‮ Measure the flywheel speed with a tachometer and compare to position encoder measurement.

‮ Measure the frequency of SVPWM and SPWM with a frequency counter or oscilloscope.

‮ Calculate the energy of the flywheel using measured data.

‮ Verify the maximum sag correction duration of 1.5 s.

‮ Verify that 37 % sag is corrected for its duration to 0.95 per unit with an oscilloscope.

‮ Verify a sag response within 2 cycles with an oscilloscope.

‮ Verify that a 4 sample per cycle rate can initiate a sag correction within set response time.

‮ Verify functionality of the Phase Lock Loop (PLL).

‮ Display results from converter (likely using HyperTerminal through RS232 communication).

‮ Verify functionality of the sensors (LEMS).

8

Appendices

Appendix A: Figures

Appendix B: Pictures

Appendix C: Specifications

Appendix D: Bill of Materials

Appendix E: Parts Ordered

Appendix F: Individual Reports

Appendix A: Figures

Figure 4: Overall Design Schematic

A-1

Appendix A: Figures

Figure 5: AMPS Side Switch Diagram

Figure 6: IM SideSwitch Diagram

A-2

Appendix B: Pictures

Picture 1: AMPS System Board

Picture 2: Flywheel

Picture 3: Capacitor Shorting Bars

Picture 4: IM Nameplate Data

B-2

Appendix C: Specifications

Table I: Design Specifications

The AMPS / 3 Phase, 208 V, 60 Hz, 5 kVA
Series Transformers / 240V/240V, 7.5 kVA
LC Filters / (10mH, 20 μF)?
DC Bus Voltage / 450 V max
DC Bus Capacitance / 2 x 250V 1000 μF (grounded between the 2)
Flywheel Inertia / 0.911 kg-m2
Induction Machine Ratings / 208 V, 32.6 A, 60 Hz, 10 hp, 4 pole
SVPWM Switching Frequency / 1 kHz
SPWM Switching Frequency / 10.8 kHz
Maximum Sag Correction Duration / 1.5 s
Maximum Magnitude of Sag Correction / 37% (or 63% of rated) @ 0.95 per unit
Sag Correction Response Time / Within 2 cycle
Magnitude Sag Correction Tolerance / Within 0.95pu ± 0.05pu of rated
2 Tier Converters / 6 IGBTs 75A, JTAG (software not included), etc.
DSP Program Language / C with inline ASM from TI
Sample Rate for Voltage Correction / 4 samples per cycle
Flywheel Speed Sensor / Position Encoder

C-1

Appendix D: Bill of Materials

Bill of Materials
Quantity / Item / Manufacture / Unit Price / SubTotal
2 / AC/DC Converters / Tier Electronics / $1,700 / $3,400
3 / Single Phase Transformer / Hammond Power Solutions / $80 / $240
1 / Design Poster/Report Binding / UI Commons Copy Center / $30 / $30
1 / DSP Software / Texas Instruments / $250 / $250
6 / Voltage Transducer / Digi-Key Corporation / $37 / $222
6 / Current Sensor / Digi-Key Corporation / $21 / $126
2 / 1000 microF Capacitors / Digi-Key Corporation / $8 / $16
1 / Miscellaneous / Unknown / $541 / $541
Total / $4,825

D-1