Home

Combined_ETD_reviewd.11-01g

1. 71 n 2 5 o d N W 1 x H 9 a 2 8 v 5 6v SHYO 9 ly 4 2 4 sr 1 1 1 1 lt KON L 4 L L L L L L lt T 4amod 4ewOd peWw So tr w w 5 lt Auo nosy ss t duy yunys diq yunys z jwodubH 11 5s LU e duyyunyg lt 01 9 802 199 4E 4 1 damod yBiq din tunus se XXXX 4 4 1 6c 12 52 Ho vA vA vA tx vA e A A A x ed EX o 0 Ho m ZA zx ZA zx ZA zx ZA 2 JO 85 BION i LA LA LA LA f 4 cs cr 24 b 24 1 24 b a DIOAH ou H NTHO 24 ANHO AHUO i HVLOIGET 1 Buimeip ees B6E 1 i s 1 60 i 6ELOl 2 L E 10 ul peje20 4018 9 Ssed q H 7 i 4amod H Jamod pay 24 3 so wajsAs N uiejs s Io uoo 4 i H Re OWA OL n 4 5 o d N W 1 x H 9 a a 2 8 v High Power Lab M LVSGIAA EStop 1 AA al CLINTON PERDUE APPENDIX D DESIGN SCHEDULE Appendix D Design Schedule The following table was created early in the design process as a predecessor to a Gantt chart for the desi
2. 14 2 3 SCADA what it means and why it satisfies our needs 15 24 Elements of a SCADA 18 2 4 1 Sensors descuido 19 2 4 2 TITEL ACE CU 22 24 tabe 23 2 4 2 2 Organization and navigation 24 24L2 3 Messages dud A IS seo i EE 25 2 4 3 Communications kuu n ie tct te tet ieu bere ier f Ee ed cL ute 26 23 5 1 Direet WAT We 26 2 4 3 2 Networking the modern marvel seen 27 Zo 3 mend TICE WOE KS Su u Sa GSA 28 2403 22 Ethernet eod is uu pem ads u oom ante 28 2 4 4 TIC acs a seat rcs tu eo cct 29 2 4 4 1 Role of the Control System vs the Role of the Operator 30 2442 The Administrative ROLE xis 31 2 4 5 Data HQ 31 2 5 Redundancy and He 33 2 5 1 Fault detection the fail safe principle
3. 33 2 5 Protective am u aa attach i d eager 34 2 6 Wer 34 2 7 Summary of deSISn goals Eua rera a des 36 Chapter 3 Design and Implementation 37 3 1 n u la L a diode amne abs de I hisa us 37 3 2 PO Sy SUT eee CURED NUM 40 3 2 1 What VO esanian 40 lil CLINTON PERDUE CONTENTS 3 2 2 Sensor Bard Wake 1a oso diet aget yd eoe au ERR VUE 3 2 3 3 3 Communications systems Procedures eieo S au 3 4 1 Saline the LOOM 3 4 2 Energizing 480V wall outlets 3 4 3 Energizing MV supply Chapter 4 Testing and Verification 2 4 1 Hardware commissioning 4 1 1 field device to module AA shed Discrete THDUIS doro oen RR 25 Discrete outputs o ua dete a e t eden 4 1 1 3 Analog signals eite eene te inae na 4 1 2 network 4 1 3 REWeHM
4. Figure 3 1 System level states and their transitions Again please note that there is no automatic way to energize the system This is a manual only action The 480 VAC is controlled by a manual close circuit breaker with a remote shunt trip input which students may be allowed to operate The 4160 VAC switchgear has provisions to be closed remotely but these have been disabled to obviate the risk of an erroneous or unauthorized close command doing harm Energizing the 4160 VAC is the sole responsibility of the lab administrator who must physically unlock the utility room and manually close the breakers This will be explained further in section 3 4 Procedures The choice of a particular SCADA software package was constrained by the capital equipment already in place After brief reflection we realized only a GE product would interface with the drive effectively making the built in higher functions the automatic referencing data capture diagnostics etc available GE Fanuc s Proficy Machine Edition formerly Cimplicity is capable of providing the graphical user interface data logging alarm messaging and logic functions we require Proficy is also capable of communicating by ModBus IP which is necessary to make use of the advanced features of the switchgear 39 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION 3 2 system This section deals with the particulars of selecting I O devices th
5. Assuming 480 primary voltage the nominal secondary voltage can adjusted 5 by the primary taps The variable secondary voltage allows testing different voltage level equipments or it allows higher voltage equipments to be initially tested at a lower safer voltage or it allows higher voltage to be tested at high voltage low current or at low voltage high current Therefore the rate of secondary winding reconnection might be higher than expected The transformer secondary cabling should be rated for maximum current times a factor of at least 1 25 480 8 1 25 601 amp rms and for the maximum secondary voltage of 4160 11 rms Rated 480 11 rms primary line current is 1202 8 amp rms The installed transformer should have adequate clearance on all necessary sides to meet NEC requirements for cooling and to allow access for removal of panels necessary to make primary and secondary winding reconnection 109 CLINTON PERDUE APPENDIX E HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN Allowable transformer acoustic noise at rated load has been specified to 65 Since there are two identical transformers if each is operating at rated load the maximum acoustic noise for the two transformers is 65 3 68 dbA I had requested that the iron core line reactors not add more than 1 dbA to total at rated load Thus the maximum noise due to magnetics in transformer room is expected to be 69 dbA
6. Fig s 2a 2b 2c illustrate three high impedance ground fault detection schemes where voltage across or current through the grounding impedance is measured and compared to a threshold to detect a ground fault 105 CLINTON PERDUE APPENDIX HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN m 4 M AC Source or Motor Figure 2A In an Netwo i I lt 1 d Ground Voltage or Impedance Current Sensor Z i DC Source iL or Motor 106 CLINTON PERDUE APPENDIX HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN Fig 3 illustrates a ground fault detection scheme used in all the GE high voltage 2300 6600 VII rms current source power electronic equipment implemented by using already existing RC line filter and line voltage sensing equipment The neutral of the three phase wye connected RC filter is connected to ground The line voltage sensing uses a high ohm approximately 1 megohm resistor string attenuator connected from line to ground to reduce high line to neutral voltage to a low signal level of approximately 2 volts The three attenuated line voltages are summed and passed through a notch filter then the absolute value of the notch filter is compared to an adjustable ground fault trip level The purpose of the notch filter is to equalize the ground fault impedance trip level regardless of whet
7. 98 CLINTON PERDUE APPENDIX C E STOP INTERLOCK Appendix C E stop Interlock This is the interlock between the existing lab E stop the switchgear and the HV lab control Per convention black is existing blue is new to be added red is existing to be deleted 99 CLINTON PERDUE APPENDIX C E STOP INTERLOCK IVVLOSAT dug eem ON 405 WIO y sai die qeq amod UBIH TWH SYONILLIHM suoieuuco dui junus gina 30W 1584 ou sin 90 HOHVIN 6 sus vm eluBIIA e 5340 2 5 2 oS UE 5 S 09 d l 8 w 4 an Dd gt OI ES OF 2 IX tox G 2652 IDE Paz ID 58 58 8 gt at F I jas 15 i Ql m des 1 2 E i gt ua gt I2 Q 10 Q o P o E eer cau uama k au kaa ee aie 9 5 x x ao o a e o mac gt gt gt L gt gt gt gt 2 3t x x s E E T 25 N
8. 139 Login Administrator Gateway with switchgear 80 4b 1 130 Pass Gateway 139 IO1 Connects N A N A 139 IOI HUB 131 101 NIU and CCM Gateway Important Physical locations are subject to change Be sure to update this list whenever any of these locations find a new temporary or permanent home Red letters indicate a known temporary position The Router has 8 cable ports each with clear number of markings on the case At present 5 Ethernet cables are connected 79 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE Table B 2 Router port assignments Physical Port Device Connecting Additional Information Number Location 1 161 IOI 1 Stripe on Cable 2 161 IO2 2 Stripes on Cable 3 139 IOI 3 Stripes on Cable 4 5 6 7 Tethys HMI viewnode 8 Carme Master computer Runs logic HMI server B 2 2 Modbus serial network The lab switchgear in room 139 uses Modbus a serial network designed to imitate hardwired relay logic to communicate between the circuit breakers We can get this data into the control and HMI via a bridge device located in 139IO1 Modbus is a serial daisy chain network New drops are added by putting another unit in parallel with the existing ones on the four communications wires and relocating an appropriate termination resistor if the new drop is on the end of the chain The wires are color coded just match the pattern Addressi
9. Appendix VT High Power Lab User s Guide VT High Power Lab User s Guide Author Dhane Ross and William Gatune Revisions Clint Perdue 1 1 Date 16 September 2006 B 1 Introduction The High Power Lab facility is designed to supply loads in excess of 250 continuous power The lab is controlled using GE Fanuc Proficy formerly Cimplicity Machine Edition an industrial supervisory control and data acquisition SCADA software package This system consists of Proficy Control a soft PLC a programmable logic controller running on a generic PC instead of special hardware for logic functions and Proficy View a graphical operator screen engine Both of these run on PCs in the lab The control system applications logic is designed to determine whether or not it is safe for the lab to be energized and to shut down the power if there is a fault or if a person enters an energized area This software is connected to the lab space instrumentation using a distributed network of I O modules that communicate via Ethernet The purpose of this guide is first to document the specific configuration of lab equipment that is unique to this project and point users toward the pertinent OEM documentation Copies of all the relevant documentation have been assembled into a project directory ask the lab administrator for access Secondly this guide will help users configure the system for their needs If a user simply requires power the
10. The maximum transformer loss at rated load is to be lt 1 5 rated MVA Thus watts loss to room due to the two transformers is 30 KW It would be reasonable to expect that heat runs will need to be conducted on the Equipment Under Test Normally heats runs take about 4 hours for everything to stabilize It is recommended that Va Tech calculate the expected transformer room temperature assuming 30 KW transformer losses and existing air conditioning If forced air cooling is required total cooling and magnetic acoustic noise should be estimated and compared to what the environment will tolerate 110 CLINTON PERDUE REFERENCES References GE Fanuc Automation March 2003 VersaMax Modules Power Supplies and Carriers User s Manual GFK 1504K USA GE Industrial Systems 1999 Installation Guidance For Innovation Series Drive Systems GEH 6380 General Electric Company USA Schneider Electric Company June 2001 MICROLOGIC 5 0P and 6 0P Electronic Trip Units Instruction Bulletin 48049 137 02 Cedar Rapids USA Schneider Electric Company Nov 2005 Power Zone 4 Low Voltage Metal Enclosed Drawout Switchgear with Masterpact NW and NT Low Voltage Power Circuit Breakers Class 6037 Catalog 05 Cedar Rapids USA Schneider Electric Company June 2001 Power Zone 4 Low Voltage Metal Enclosed Drawout Switchgear with Masterpact NW and NT Low Voltage Power Circuit Breakers Class 6037 Instru
11. INTRODUCTION Chapter three will detail the choices made to meet these requirements It includes both hardware and software designs Chapter four will document verification and testing of the control system elements that have been commissioned to date Finally chapter five summarizes what has worked and what hasn t and gives guidance toward future work as the facility is adapted to meet emerging needs A user s guide is included as Appendix B 12 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND Chapter 2 Control System Requirements and Background This chapter is in two sections The first speaks to the specifications for the control system both in terms of enforcing safety protocols and of operating the facility and proposes solutions to satisfy them The second section is largely background information on the methods materials and technologies with which the control system is built It is included in hope of furthering the broader educational purpose of the high power lab Figure 2 1 Design scope 2 1 Lab Operational Requirements The High Power Lab equipment is impressively large and potent beyond the scale of what most people deal with regularly How then does one construct the lab so it is appropriate for an academic environment So that it is safe yet not so restricted as to be unusable All the equipment needs to be interlocked for safety purposes and integrated in terms of cont
12. Q7 Energize Breaker Q8 Energize Breaker B QX1 192 168 1 111 2 6 960X1 192 168 1 111 2 7 96QX1 192 168 1 111 2 8 161 102 Node 2 VersaMax I O 192 168 1 110 EGD Slot 1 IC 200 MDD 842 11 Door 161 12 Door 171 Plug 1 14 Plug 2 Q9 Q10 011 012 013 014 015 016 Slot 2 IC 200 MDL 930 Q1 Tower A Red Q2 Q3 Q4 Q5 06 07 Q8 Tower A Yellow Tower A Green Tower A Alarm Tower B Red Tower B Yellow Tower B Green Tower B Alarm 961X1 192 168 1 110 1 1 1 1 192 168 1 110 1 2 IX1 192 168 1 110 1 3 IX1 192 168 1 110 1 4 QX1 192 168 1 110 2 1 QX1 192 168 1 110 2 2 QX1 192 168 1 110 2 3 0 1 192 168 1 110 2 4 0 1 192 168 1 110 2 5 0 1 192 168 1 110 2 6 0 1 192 168 1 110 2 7 0 1 192 168 1 110 2 8 DURIP State Control_Debugger MasterCarme Ladder_Program JS 9 18 2006 5 35 30 PM Page 10 CLINTON PERDUE APPENDIX A LADDER PROGRAM 6 3 Data transfer script Moves data from the switchgear serial ModBus protocol devices to an internal variable as a background process This frees the main program from waiting for a response to a request for data from a ModBus device This runs every second Script Created Aug 30 2006 Description Internal BreakerA fdbk PLC DINTI1 Internal BreakerB fdbk PLC DINT2 76 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE
13. 35 v 4 1 O HVSwitch2 HVSwitch4 1 1 H Off On 36 LightControl 161B LIGHT TOWER PowerOnState 161B BreakerA Closed BreakerB Closed Tower A Green 37 14 A M O Off Off DURIP State Control_Debugger MasterCarme Ladder Program 9 8 2006 53528PM Page zJ CLINTON PERDUE APPENDIX A LADDER PROGRAM ThermalAmbient_139A Tower_A_Yellow 38 vt Safe_Temp_480plugs Vt ThermalAmbient 161B LA v On HVSwitch1 BreakerA Closed Vt Vt On Off HVSwitch2 BreakerB Closed V V Off Off PowerOnState 161B Tower Red 39 i Q TP 1 Tower A Red TP Tower A Alarm 40 E IN Q Q PT ET OFF AccessViolation480_161B vV I AccessViolationHV_161B 1 171 LIGHT TOWER PowerOnState 171 BreakerA Closed BreakerB_Closed Tower Green 41 V Vt Off Off State Control Debugger MasterCarme Ladder Program 9 8 2000 53528pM Page 7 E CLINTON PERDUE APPENDIX A LADDER PROGRAM ThermalAmbient_139A Tower_B_Yellow V 2 42 Safe Temp 480plugs V ThermalAmbient 171 Vt Off HVSwitch3 BreakerA Closed Vt Vt On Off HVSwitch4 Closed Vt V On Off PowerOnState 171 Tower Red 43 11 O H N Q O Tower_B_Red TP Tower_B_Alarm 44 AccessViolation480_171 iI AccessViolation HV 171 1 45 END DURIP State Control Debugger MasterCarme Ladder Program 9
14. 78 CLINTON PERDUE APPENDIX VT HIGH POWER LAB USER S GUIDE B 2 1 Ethernet Devices We employ an Ethernet network of hardwired devices which communicate through a router and also a hub for devices within the Switchgear Room room 131 Our private network has all been wired with red CAT 5e Ethernet cables everything else in the area is blue Below in Table B 1 is the network configuration information for the Ethernet devices Table B 1 Ethernet configuration data Device Purpose MAC IP Physical Additional Name Address Address Location Information ROUTER Provide lab 00 12 17 192 168 161B Near Login Admin LAN side Ethernet 4c da 08 1 1 door Pass Admin ROUTER Internet 128 173 Subnet Mask WAN side connection using 90 202 255 255 252 0 cloned MAC Carme Soft PLC 00 07 e9 192 168 161b Registered developer controller 7d 19 5e 1 100 window Station Tethys Designated HMI 00 07 e9 192 168 161 HMI node 7d 61 cb 1 102 Center bench 161 IO1 Cabinet Module 00 09 91 192 168 161B Far VersaMax NIU NIU Communications 00 4d 6b wall 161 IO2 Cabinet Module 00 09 91 192 168 161B Near VersaMax NIU NIU Communications 00 4d be 1 110 window 139101 Cabinet Module 00 09 91 192 168 139 beside VersaMax NIU NIU Communications 00 4d de 1 112 switchgear CCM Communicates 00 80 67 192 168
15. Human Machine Interface HMI and Man Machine Interface MMI have been used to mean the same thing CLINTON PERDUE CONTENTS Contents uL cH LI HU tM C MC ED AS It BG ii roni X c met iii lop asa sda ates a Meer eck ene Dedication ets INCIMIOW IES ASI ES vii Chapter I tFOdUCHOoT S s De ose Eyre pd 1 1 1 Our Hei nl eee 1 522 IRC UTS TAG HAUS dettes esto taSs ted 2 1 3 The prior art other testing facilities eed 3 1 3 1 Static dissipative loads iicet Pea CUT TS RESTE Re ena 3 1 3 2 Dynamics ystems u Qs u tute tou Eb soos dui 4 1 3 3 Static regenerative system Sansenoi a ee 6 1 4 The VT HIGH POWER LAB available 7 1 5 Objectives OF this tHe BIS UE 10 Chapter 2 Control System Requirements and Background 13 2 1 Lab Operational Requirements edes ea taa 13 2522 Proposed oo So Cmn
16. T CLINTON PERDUE CHAPTER INTRODUCTION Figure 1 2 The author in a former life working on an 11 MW 6 6 kV AC induction motor The cooling and lubrication systems are not installed but the cooling water lines and heat exchanger can be seen atop the unit to the right Total volume is about 60 m the rotor alone masses about 23 MT and the stator another 36 MT 1 3 3 Static regenerative systems Given modern power electronics the mechanical elements of a recirculating system are purely optional Line power can be rectified inverted and fed back into the supply Alternately the DC bus can be fed from a line rectifier and a test rectifier one controlling current the other voltage with a load inverter returning bus power to the test rectifier Both options are depicted in the following Figure 1 3 This latter topology has been used at the General Electric Company s works in Salem Va to test motor drives in the multi megawatt range CLINTON PERDUE CHAPTER INTRODUCTION E Figure 1 3 Two options for a static regenerative system The result of both of these topologies is a realistic static load with the option to test either rectifiers or inverters some benefit of recovered energy and less complications than a mechanical system Table 1 1 Testing topologies relative strengths Regenerative The relative strengths of the four main approache
17. The SCADA can not protect anything without proper feedback 13 Energize auxiliary sources 14 Clear everyone from the room and close the main doors There is no way at present for the SCADA to know if the room is truly empty be careful 15 Verify permission to energize status on HMI This ensures that there are no other system faults 16 Close the main 480VAC breaker 17 Close the branch feed breaker End 33 Professionals literally count their tools before and after working on high power equipment to make sure nothing has been left inside m CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION 3 4 3 Energizing MV supply 1 Safe the room All energy sources must be safe before entering the room even ones not directly associated with the circuit to be energized 2 If 480VAC connections are in use perform steps 1 12 of the 480VAC checklist Check everything and get it ready to energize 3 Open the knife switch if it is not already locked out In case there is current present this will contain the arc Stand to one side of the panel and look away when you operate the switch 4 Open the panel door 5 Verify that both the line and load side voltages are zero Be sure that what you think you turned off is really off 6 Remove safety grounds 7 Test for load side shorts between phases and to ground Bad connections bad insulation or other damaged components tools or debris left in equipment can all caus
18. true inputs 4 1 1 1 Discrete inputs Discrete inputs are the simplest of the simple in essence just a switch They are tested by opening and closing the current through them being read by an I O module If opening is impractical then a wire can be taken loose to break the circuit We did this with the transformer over temperature sensors which are inaccessible to check we had them identified correctly Start tracing problems by testing the mechanical hardware Is the switch actually opening closing mechanically the switch body may have moved Does it open close electrically Does it have proper incoming voltage If not trace back to the supply Is the signaling voltage correct Trace it to the input terminal Is there noise voltage on the signal especially transients when some other load is switched 4 1 1 2 Discrete outputs Discrete output problems can be localized to field or module very easily by shorting around the module terminals with a wire jumper There are two types of discrete output modules installed the isolated relay contact which is again just a switch and a current source sink which provides 24 VDC depending on how itis wired These can be tested with a multimeter right at the module though the load may need to be disconnected 55 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Discrete loads can be relay coils as with the signals to the E stop and switchgear or solid state dev
19. 11 Ligure Desen SCODG Ls 13 Figure 2 2 Example of a time power heating curve used for control 17 Figure 2 5 Elements of a SCADA 18 Figure 2 4 Example of a sequence of status and alarm 26 Figure 2 5 Typical SCADA for motor control 1 420 Le ioo ee Haeo bd eee 35 Figure 3 1 System level states and their transitions eene 39 Figure 2 2 a au aaa na ean 42 Figure 3 3 High voltage switch status sensor open and closed 43 Fig re 3 4 Ethernet back Bones coto tot ut opea t sse ioter additi ait 46 Figure 3 5 Hardwired interlocks iuuenes rta te e eel te 48 Figure 3 6 Complete Ethernet topology 49 Figure 3 7 Complete communications topology eese nennen 50 Figure 4 1 I O module and schematic showing three true Iputs 55 Figure 4 2 Historical data retrieved from the system log 58 Figure 4 5 Online VO MONTO eain uakis 59 Figure 4 4 Logic st
20. 4 6 is the control system reacting to an external fault as opposed to the operator error in the first test From the circuit breaker s point of view it hardly matters it must open Transformer 2 has been simulated overheating by breaking the fail safe thermal overload detection circuit just as would happen if the over temperature klixon opened This occurred at 22 21 42 The cursor shows the next historical data record one second later after the signal ThermalTransformer_2 has changed value to 0 There is a programmed thirty second lag to filter spurious events after which at 22 22 13 the signal Energize_Breaker_B goes false The breaker reports that it is indeed open at 22 22 14 62 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Y axis title 100 7 901 704 60 50 30 201 0 22 21 20 22 21 30 22 21 40 22 21 50 22 22 00 22 22 10 22 22 20 22 22 30 Time DOOR 161B 1 BREAKERB CLOSED 1 7 ENERGIZE BREAKER 1 75 ENERGIZE BREAKER 1 5 HVSWITCH 1 THERMALAMBIENT 139A 1 THERMALAMBIENT_161B 1 75 JBREAKERA CLOSED 1 THERMALTRANSFORMER 2 0 THERMALTRANSFORMER 1 1 7 HVSWITCHI 1 Sunday April 09 2006 22 21 43 Figure 4 6 Transformer fault test CB 1 Testing for user added fault conditions as in support of a particular project would look much the sam
21. CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND In other cases critical signals such as E Stop are hardwired in order to increase their reliability The philosophy is simple is better Electronics especially computers should be avoided in the chain of communications because they can fail in unpredictable ways 25 the olden days direct wiring was the most common way to make logical connections between equipment To signal between devices A and B A would energize a relay coil to close an isolated set of contacts signaling an output This would close a circuit probably at 110 VAC wired from device B Often this would energize another relay in B which would in turn close a low voltage circuit to drive an internal logical indication This sort of connection is the lowest common denominator default way to signal between devices Sometimes it makes sense for very simple connections but is counter productive from a centralized control perspective these signals are not generally available for operator display or logging There are better ways to make logical connections which require less space power and are generally cheaper 2 4 3 2 Networking the modern marvel Networked communications is one of the greatest advantages modern control system designers have over their forebears In a modern system small clusters of I O devices are connected to network interface points At this point some form of interface
22. Control_ 161102 au99 vsd M View Runtime 21 View Network Figure 4 3 Screen capture of control system editor in online monitor mode showing I O status on the upper right for several devices Figure 4 4 shows another perspective with the logic editor online True inputs are highlighted which makes debugging easier 59 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION DURIP State Control_Debugger CIMPLICITY Machine Edition Ladder_Program Main View MasterCarme Fie Edit Search Project Target Variables Tools Window Help xu gt O EA six me India Value High Power Lab Power distribution control Wit MasterCarme StandbySetup ld MasterCarme StandbyStatus 0 il srART MasterCarme StandbyStatusLog 0 0 0 0 0 fia MasterCarme StandbySync lo Check on status of HV BREAKERS not the wall switches the ones in room 139 Wir MasterCarme Status 394437 NT MasterCarme STCompileBool EQ BreakerA_Closed 2 EN 0 MasterCarme STCompileInt 0 a MasterCarme STCompileReal umn WI j StopPending Nit StopScans ID 14 MasterCarme SystemError WI MasterCarme TargetScan 1000 WI MasterCarme TRUE On V MasterCarme UserName EQ BreakerB_Closed NI MasterCarme Version 0392564
23. DC recirculating test 171 DC test 171 AC test 171 2 AC tests 161 161 AC test Workstations Workstations with windows with windows 161B 171 161B Unit Under 171 High Power 1 High Power 1 Test High Power 2 DC Unit Under Unit Under a Test D To equipment room 139A To equipment room 139A 161B AC DC recirculating test 171 2 AC tests Workstations 161B 171 High Power 1 Unit Under Test 7 Unit Under HV 161B AC test 161B DC test 171 2 AC tests Workstations with windows 161B Unit Under High Power 1 Test 171 D AC Unit Under Test HV H sWwH SW 2_ Test DC 51 To equipment room 139A To equipment room 139A Figure 1 5 Lab floor plan for various load side power distribution topologies 1 5 Objectives of this thesis A convenient way to plan for a major project such as the high power lab is to break the specifications and requirements into three groups capital equipment physical layout and control Figure 1 6 These elements are all interdependent and it may take several iterations to achieve an integrated design 10 CLINTON PERDUE CHAPTER INTRODUCTION Figure 1 6 Project elements The selection and placement of the switchgear transformers fuses reactors high voltage cablin
24. H On On Safe_Temp_BreakerA 2 29 O DURIP State Control_Debugger MasterCarme Ladder_Program 9 18 2006 5 35 29 PM Page 4 70 CLINTON PERDUE APPENDIX LADDER PROGRAM Determine if a High Temperature is because of Switchgear Breaker B 30 second lag TON_2 TOF_2 ThermalTransformer_2 ThermalAmbient 139A ThermalAmbient 161B ThermalAmbient 171 TON TOF 30 l 1 H 1 H 1 IN O IN Q On On On Off HVSwitch2 HVSwitch4 1 H iF Off On Safe_Temp_BreakerB 30 Determine if a High Temperature is because of 480V outlets 30 second lag TON 3 TOF 3 ThermalAmbient 161 ThermalAmbient 171 TON TOF Safe Temp 480plugs 31 1 1 N Q N Q O On off PT ET PT ET Safe_Plug161 Safe_Plug171 l On 32 Energize outputs Determine if OK to energize 480 VAC outlets False will trip the breaker in room 161 and require a manual reset AccessViolation480 161B AccessViolation480 171 Safe Temp 480plugs Energize 480 33 V4 V E O Determine if OK to energize Switchgear Breaker A top False will trip the breaker and require a manual reset AccessViolationHV 161B AccessViolationHV 171 Safe Temp BreakerA Energize Breaker A 34 n OO HVSwitch1 HVSwitch3 On On Determine if OK to energize Switchgear Breaker B bottom False will trip the breaker and require a manual reset AccessViolationHV 161B AccessViolationHV 171 Safe Temp BreakerB Energize Breaker B
25. ThemalAmbient_161B 1X1 192 168 1 111 1 8 EE Graphical Paneis 19 ThemalAmbient 171 4X1 192 168 1 111 1 9 O8 Languages pa 6 Logging Groups H 5 RootLoggingGroup n T au Program 114 ThemalAmbient_1394 41X1 192 168 1 111 1 14 On S Es bes d 115 ThermalTrensformer_1 4 1 192 168 1 111 1 15 On Subroutines 116 ThemalTransformer_2 1X1 192 168 1 111 1 16 tr 71 J Qn 19 e 120 en Ethemet 1 0 Driver 9 2 IEC Addressing Method Driver Node Slot Tem 9 o 125 125 9 128 e 125 130 e 23 Jesse D Mastertarme MasterCarme On line Inspector maes Cono lel Variable Name Address Value X Transferring Index HTML 41 of files transferred a 24 Skipping up to date files jet Master Came PLC_DINT1 3961 16800000040 Transferring NewBuild 43 of files transferred Master Came PLC_DINT2 _ 300661 1680000 0050 Transferring index HTML rd of files transferred Skipping up to date file Transferring Index HTML 584 of files transferred Skipping up to date files Download completed Download Complete 0 error s 4 warning s Tip Press F4 to cycle through warnings and errors is EN Static K Auto f VIT Build Import Messages Reports References 3 Ln8 Col25 Online Administrator LOCAL DURIP State
26. assimilate data from diverse inputs and may take a wide range of actions When choosing equipment and software issues of migration and obsolescence coding languages and expandability adaptability for future needs all require consideration It is important to avoid implementing something for which there is no technical support no spare parts a large learning curve and no way to update it The standard answer for implementing a control system is the programmable logic controller or PLC These machines are optimized as controllers in many ways the bookkeeping tasks of managing memory I O handling interfacing to multiple networks 29 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND timing etc are already done PLCs are also typically closely integrated to the I O they need so there is a minimum of delay in the system The PLC environment leaves the programmer free to code the function they care about without worrying they are going to break something else with a coding error equipment maybe but the controller will not crash Often coding can be done at a very high level such as with a graphical control block language drag your mouse to connect the regulator block between the sensor and actuator blocks This makes them easy to use and fast to set 23 PLC vendors also supply tool suites for trending data logging and other necessities These tools have become so pow
27. avoid overloading the user 2 4 2 1 Graphic design Good graphics design will always put the most important information forward in a consistent predictable manner The graphics require a certain degree of abstraction but the display should also mimic the appearance and layout of the real equipment The relationship between them should be quite clear and intuitive On a given screen many of the objects will be active Some will depict control buttons which will take operational actions directly start stop select a mode etc Other objects will set anumerical value These objects can look like anything pointers to be dragged along a slider bar knobs switches Often the choice of input device depends on required precision and speed to make a change which in turn depends on the equipment and process involved Sometimes it is better to click and hold a vernier adjust button while looking at something else other times this is too slow and typing in a number is better Having both methods available and others besides is both possible and reasonable Status display objects will use text numbers color and position or animation to distinguish various states and bring especially important things to the operator s attention Changes in appearance should mean something definitive not be interpolations adrift between CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND known states A word of caution
28. chose various devices to add including a windows station PLC drivers and Modules 2 To add a service Panel Module a After reaching add target click GE Fanuc Remote I O gt VersaMax Ethernet 3 You must designate a The Power Supply Model b Network Interface Unit Model i After picking the module you must configure the network setup Be careful when specifying information in the Produced and Consumed Exchange tabs 1 Network Tab Specifies network connection settings a Designate the IP address b Subnet Mask 255 255 255 0 c Gateway IP 192 168 1 1 d Mode i EGD For use with GE Devices Such as the NIU ModBus TCP IP if you use a generic device 2 Produced Exchange Tab The Produced Exchange refers to the information that is generated by the module that will be exchanged with something else the PLC a Exchange ID Always remains 1 b Consumer Type Designates how the device s that consume the produced exchange will receive the information i IP Address Set the IP address of the computer that will accept the produced exchange H Group ID By specifying a group number you can allow information to be sent to all devices associated with that group c Producer Period ms How often the module will produce information 94 CLINTON PERDUE APPENDIX B USERS GUIDE 3 Consumed Exchange Tab The Consumed Exchange refers to the information that is received by the module a Exchange ID One would guess
29. e 5 gt m N e X z z 4 N i i o o 4 o o 5 e e m lt lt gt Ow Igi 12 j IS o o 24 X ge al 0 5 SV el aS gt 1o 9 0 o e T NM oW 81 e Ac 99 Y 309 ue I SNR 12 J NS ae 4 9 oF oe u FC x P FS 5 t oo RS m lt 2 dx J co Sc BS 98 ac S2 ogg t 58 lt e E ae a 5 5 ge Ne Ons lt ig nos ES o FEES QN e a 85S gt s3 lt mo 4 oe 5 24 SS j 258 gt S5 x 885 8 ee 285 E 9 S a lt es 83 5 lt 8 amp B S 5 E od o4 Ro M wow S bb Hook dE k SR SOS S 6 S 100 APPENDIX C E STOP INTERLOCK CLINTON PERDUE 99 ev uv Sv enpieg 90 HOHVIWG sss Ev SNONILIHM qe jemod uBiu dina uoe je 6342 amod VOVOS 10 wv 9053 dO1S 3 ze se ee lez uz sz ez n st e 60 20 50 o
30. it is the group from which the module receives its instructions from including all devices that are part of the group In this Lab we only want the master computer to send instructions to the modules i Avoid using default values for anything b Producer ID Refers to the source that the information the module consumes originates from In this setup it s the master computer s IP address i Note yes We know it doesn t seem right to designate the Master computer s IP address for two different information directions both times That s the way it is c Group ID Designate the Group that the produced information will be a part of c To Add Subsequent Modules i Add Carrier Base 1 IC200CHS002 I O Carrier Box Style a Chose a Module B 4 2 Some Basic Procedures The interface Adding Devices Cimplicity has various functions that make it attractive It is capable of taking in various signal types if it knows what type of data to accept produce signals produce view nodes and regulate other properties Like the modules in the cabinets the GUI of Cimplicity imitates the screw terminals of each module or even for virtual devices such as a modbus device To add a module view the Control I O Drivers section and Ethernet I O sublisting From the Ethernet IO screen you made add a Node Each node is a different physical device or in the case of a Modbus connection a different section of registers Communicat
31. limited by the moderate impedance this fault can be detected and the drive shut down before a second ground fault occurs which can result in large fault current and significant subsequential damage All GE high voltage 2300 6600 rms power electronics equipment is ohmic isolated by a transformer from the 60 Hz three phase source which is typically 4160 VII rms The transformer delta or wye secondary is connected three phase three wire to the power electronics equipment that is if the transformer secondary is connected in a wye the transformer neutral is not connected to the power electronics The two types of power electronics manufactured at GE which are most similar to the intended Va Tech high voltage test setup each have a source power converter and a load power inverter with energy storage means either a DC link filter capacitor or a DC link reactor separating the two power converters If the energy storage between the two power converters is a DC link reactor the equipment is referred to as current source equipment If the energy storage between the two power converters is a DC link capacitor the equipment is referred to as voltage source equipment For either the current source or voltage source equipment the purpose of the high impedance ground fault circuit is to protect against subsequential damage due to the occurrence of a ground faults in either the transformer secondary the cabling from transformer secondary to source
32. long distance communications with cheap off the shelf hardware from a multitude of vendors It is also very easy to make connections 28 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND As discussed previously most physical I O devices need some sort of interface module Small collocated groupings of driver modules all within a single I O cabinet for example will be networked serially to an Ethernet interface bridge module This in turn provides an Ethernet connection to the operator interface data logging and logical elements If there is a drawback to using Ethernet for control it is that the format has become too common Everyone wants to plug in a PC to the control network check their email and stream in music This is a drain on network resources and a security risk and these activities should not be allowed The SCADA system must have logical elements be they clockwork relays or computers which oversee the plant and take supervisory control actions Some of this function is delegated to off the shelf devices such as circuit breakers and self regulating valves with limited programmability suited to their very narrow scope A circuit breaker for example may have adjustable trip settings but will not care about the state of other equipment Programming it may only require a screwdriver Of more interest here are the general purpose logical elements those which
33. module adjusts voltage levels provides isolation encodes decodes I O to logical messages and acts as a network transceiver to share the data with controller s located remotely All this is no small task and it always incurs a time penalty Even so it is worthwhile because a network connection can replace hundreds of wires with just a single pair If they have to cover any great distance then the savings in the cost of the wire and the cableways and the There are no computers involved in the operation of nuclear power plants Period All signals are considered to be safety critical Lecture and tour of the North Anna Nuclear Power Station by engineer Douglas Struckmeyer 1990 The switching system for the New York City subway still relies on electromechanical relays because they open by gravity under all fault conditions Postman Monroe Robust Relic Letter IEEE Spectrum Jul 2005 Vol 42 Number 7 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND installation labor quickly pays for the networking equipment Furthermore if the connections must pass through a confined area a network connection may be the only practical method By virtue of being digital signals network communications can benefit from error correction techniques to achieve better fidelity than analog signals in the same channel Indeed sometimes a digital signal over a fiber optic link is the only practical way to est
34. of power system complex voltages and currents and publish this data via ModBus and network bridge to the control system The breakers themselves can be controlled remotely either by network commands or through traditional discrete hardware inputs 85 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE We have disabled the remote close capability closing is a manual only administrator prerogative but have connected their remote trip request discrete hardware input to a set of normally closed relay contacts that are held open by an energize OK output from the control If this output goes away for any reason including control failure the relay will deenergize and trip the breaker see the E stop wiring elementaries Each trip circuit pulls over 9 amps at 24 VDC for a few milliseconds to actually trip the main circuit breaker We have placed ride through capacitors in the switchgear upper equipment bay to buffer this load There is a similar power analyzer on the second floor feeder to the lab and its data is also available to the control system but has not yet been utilized There is also a remote trippable circuit breaker located in room 161 for the 480 VAC feed to the lab bays The trip control for this is similar in principle to that of the other units but it does not have a power analyzer and students are allowed to close it for themselves The rest of this section deals exclusively with the previous
35. to an end move the terminating resistor to the new end of the chain T CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION Square D switchgear Safety I O E stop power interlocks Hardwired Figure 3 5 Hardwired interlocks From chapter 2 the reader will recall that I O in general is connected to Ethernet enabled communications modules bridges in their own right between serially connected signal conditioning hardware and a network interface logical unit The complete Ethernet topology is shown in Figure 3 6 All fixed devices are configured with static IP addresses The addresses are used at compile time to set up runtime messages between devices so changing a device s address will lead to dangerously garbled communications In general there is no call for a permanent device to change address 48 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION Square D switchgear Safety EGD E stop doors cap Modbus IP bank transformers E e ea S ree cooling connections Hardwired U m Test equipment T Hardwired modules lica Data loggers scopes and custom sensors GE Public signal Modbus and Ethernet Ethernet PC runnin PC with GE drive HMI viewer 1 External configuration tools soft PLC HMI viewer 2 HMI Figure 3 6 Complete Ethernet topology Finally
36. to ensure safety then that state must be monitored in such a way that a complete circuit has to be maintained in order to signal a safe condition This means that protective sensors will be configured to open to signify a condition that is not intrinsically safe temperature too high lubricant or coolant level too low door open etc Thus a broken wire will not go undetected but will be interpreted as a shut down command eas CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND In the very most critical cases redundant elements may be used think of having multiple smoke detectors in your house To be really really careful redundant sensors should share no common elements not even power supply or cableways Commonly sensors will have complimentary outputs The normally closed contact will be used for the fault detection and the normally open contact can be wired to a networked input point This second path is not itself fail safe but provides supporting evidence for the primary signal Control logic can now watch for discrepancies This protects against failure in the sensor s supporting wiring and gives the control system the opportunity to take further action if the initial response has failed 2 5 2 Protective actions Lacking proof that a safe condition exists protective equipment ideally needs to act without a further supply of energy If this is not possible then steps need to be take
37. waits for the AC current to be near zero to minimize arcing and then opens the breaker This is not a fail safe operation because external energy is required to open the breaker and power the logic The situation is improved by use of a backup emergency power supply but the equipment can still fail to open Should the switchgear fail eventually a line fuse or other protective device will have to open Approximately a second after the initial problem develops the startled operator manages to hit the E stop button This mechanically opens circuits the control has already deenergized This action accomplishes the same thing as the automation albeit later and with potentially more damage to equipment This aspect of the system is also fail safe A broken wire or failure of the control network would also lead to an E stop action in the switchgear and motor control relay 2 7 Summary of design goals The controls for the high power lab will be akin to those found in modern industrial settings The I O system will be distributed about the lab space for ease of access with minimal wire runs and support sufficient instrumentation to observe the critical system states Wiring and other construction practices will accommodate the expected high EMI environment and be fail safe The operator interface will conform to known best practice ergonomics and provide diagnostic and troubleshooting aids 36 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEM
38. 00MCM Shield Grounded Both Ends High Power Lab 1 Room 161B r Sas det MV DC bus i U 1 0 5 0 KV I 1000 200 A GE Drive C 2000 uF 2 3 KV H lac 500A 1 Cable Phase 8KV 500MCM Shield Grounded Both Ends Figure 1 4 power distribution system elementary diagram High Power Lab 2 Room 171 CLINTON PERDUE CHAPTER INTRODUCTION Variable transformers allow distribution of between 480 and 4 160 VAC to any of four utility connections A GE Innovation Series drive located in the east bay can provide four quadrant AC DC conversion using 3 300 VDC 1 200 ADC IGBTs The lab is divided into two experimental bays with a substation in an additional equipment room Figure 1 5 illustrates this floor plan and anticipated load side lab configurations All operations are conducted from an outer room on the safe side of armored glass windows The circulating power loop which allows the lab to operate at 1 MVA is available in either bay If the loop is not used the facility is limited to 250 kVA from the utility feed There is Space to introduce moderate sized electromechanical elements should the need arise The drive transformers and switchgear are all air cooled Additional cooling is available from a 3 diameter water line CLINTON PERDUE CHAPTER INTRODUCTION Power topology options 171 AC
39. 18 2006 5 35 29 PM Page 8 STI CLINTON PERDUE APPENDIX A LADDER PROGRAM 6 2 I O listing Ethernet VO Driver ID amp 1 161 101 Node 1 VersaMax I O IP 192 168 1 111 EGD Slot 1 200 MDD 841 H 12 13 14 I5 16 17 l8 I9 HVSwitch1 HVSwitch2 HVSwitch3 HVSwitch4 Plug 4 Plug 5 Plug IL 3 4 ThermalAmbient_161B ThermalAmbient_171 110 111 112 113 114 115 16 ThermalAmbient 139A ThermalTransformer_1 ThermalTransformer 2 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 AM Q1 Q1 Q2 Q2 Q2 Q2 Q2 02 Q2 Q2 8 9 0 1 2 3 4 5 6 ro 961X1 192 168 1 111 1 1 961X1 192 168 1 111 1 2 9261X1 192 168 1 111 1 3 61X1 192 168 1 111 1 4 1 1 192 168 1 111 1 5 561X1 192 168 1 111 1 6 561X1 192 168 1 111 1 7 961X1 192 168 1 111 1 8 561X1 192 168 1 111 1 9 9261X1 192 168 1 111 1 14 1 1 192 168 1 111 1 15 1 1 192 168 1 111 1 16 DURIP State Control_Debugger MasterCarme Ladder_Program 9 18 2006 5 35 30 PM 74 CLINTON PERDUE APPENDIX LADDER PROGRAM 028 029 030 931 Q32 AQ1 AQ2 AQ3 AQ4 AQS AQ6 AQ8 AQ9 AQ10 AQ1 AQ1 AQ1 AQ1 AQ1 AQ1 AQ1 AQ1 AQ1 1 2 3 4 5 6 8 9 020 Slot 2 IC 200 MDL 930 01 Q2 Q3 Q4 Q5 Q6 Energize 480
40. 28 11 03 Drive at reduced power Status message 13 30 19 50 Alarm Temperature at Supervisor s limit Alarm message Diagnostic message Healthy message 13 30 19 72 Drive trip Over temperature Figure 2 4 Example of a sequence of status and alarm messages These messages typically have variable fields which are populated at the time the message is created something like Device X reports Y Sometimes these messages are also linked to diagnostics or help documents that suggest corrective actions Se The core concept of SCADA systems is the coordination of equipment Clearly there can be no coordination without communications between devices The control system depends on good flow of data The definition of good depends largely on what is being communicated For some signals a dedicated hardwired connection makes sense Others signals are well suited to network communications in one form or another The most critical require redundant channels and error correction logic In the following we examine the various types of connections typically found in a system 2 4 3 1 Direct wiring Signals with exceptional bandwidth requirements are directly wired between source and user An oscilloscope probe is a fine example of this as is a pulse tachometer feeding motor position data back to a drive Burdening these signals with a network overhead would severely degrade them and limit their utility 26 CLINTON PERDUE
41. 8 3 EN Q Nir MasterCarme WatchdogTime 5000 N Of MasterCarme WCLScan 7580 Wir MasterCarme WCLStatus B 99 Ni MasterCarme GNCardIOStatus 0 0 0 0 0 0 0 0 0 MI MasterCarme AccessViolation480_161B lotr Wt MasterCarme Access Violation 480_171 Off INT MasterCarme AccessViolationHV_161B orr 4 Room_161B MasterCarme AccessViolationHV_171 orr R00M161B 4160 V disconnect switches IN MasterCarme BreakerA Closed 1680000 0000 Sate_HV_161B MasterCarme BreakerB Close 5 MasterCarme BreakerB_Status 1680000 0000 On ott Wir MasterCarme BrkA_SWG_Vab E QA QUON ROOMIGIB 480 V Outlet Off Check NG MasterCarme BrkA_SWG_Vavg_LL 0 Ni MasterCarme BrkA SWG Vavg LN 1 Pog Safe 161 lt a jean gt On 0 On gon Utities t 2 mo J Al manual dscormects reported opentsadelott by sensor inputs 210 Safe HV 1618 Safe Phugl6l Disconnects_open_161B Target 7 H ts ot On Ott Name Occupied State Reset with door closed transition Type i Description Master PC fieemied tate 1 118 fieemied State 1RTR Variable Address Transferring Index HTML 41 of files transferred MasterCame PLC DI x rCame PLC D TATED Static Auto 4 Skipping up to date files Transferring NewBuild 43 of files transferred nf es tr red TIL Build Import Messages k Reports Refer
42. B VT HIGH POWER LAB USER S GUIDE B 2 5 Sensors Most of the sensors in the lab have a pair of complimentary contacts one set is open when the sensor is in its default quiescent normal state and the other is closed We have elected to use the set which is normally open so that the sensor when actuated indicates by a closed circuit that it detects a safe condition If the input circuit should fail at any point a fault will be indicated making the system fail safe Any new system interlocks should be set up the same way Table B 4 below lists the sensor types and their safe outputs Table B 4 Sensor outputs Sensor State Output High Voltage Switch Off 1 480 Voltage Switch Off 1 Door Sensor Closed 1 Ambient Thermal Sensors Cool ok 1 Transformer Sensors Cool ok 1 B 3 Switchgear One of the most important components of the system is the switchgear located in room 139 There are two units one on each distribution leg feeding the lab bays see Figure B 4 Both are Square D Power Zone 4 models with MICROLOGIC 5 0P electronic trip units We refer to them improperly as the high voltage switchgear even though they operate at only 480VAC because they control the current that becomes the 4160VAC feed to the labs and 4160 is really classified as medium voltage by utility people In brief these are circuit breakers with built in power analyzers They can determine up to the 40 harmonic
43. CANON auci titre aa quce di 4 2 Software commissioning 4 2 1 Variable assignments 4 2 2 Logical t sts Chapter 5 Results Conclusions and Guidance 5 1 Weaknesses in the design 52 SUCCESSES oeil edad ede ad 5 3 m ani at Appendix Ladder Program eoe Appendix B VT High Power Lab User s Guide Appendix C E stop 20 oe Appendix D Design Appendix E High Voltage Grounding lv Meese 54 tt 58 CLINTON PERDUE FIGURES AND TABLES Figures Figure 1 1 Types of power electronics testing 3 Figure 1 2 An 11MW 6 6kV AC induction motor eese 6 Figure 1 3 Two options for a static regenerative system 7 Figure 1 4 power distribution system elementary diagram 8 Figure 1 5 Lab floor plan for various load side power distribution topologies 10 Fig re 1 6 Project ele mefits aee dae iode eo
44. DESIGN AND CONSTRUCTION OF CONTROLS FOR A KV MVA CLASS POWER ELECTRONICS TESTING FACILITY Clinton L Perdue Thesis submitted to the Faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in Electrical Engineering APPROVED Dr Fei Fred Wang Chairman Dr Rolando Burgos Dr R Krishnan September 8 2006 Blacksburg Virginia Keywords Power Electronics Test Facility Recirculating Power Ring Supervisory Control and Data Acquisition SCADA Copyright 2006 Clinton L Perdue DESIGN AND CONSTRUCTION OF CONTROLS FOR A KV MVA CLASS POWER ELECTRONICS TESTING FACILITY By Clinton L Perdue Chairman Dr Fred Wang Electrical Engineering Abstract In order to facilitate research and testing of kV MVA class power electronics systems Virginia Tech has constructed the High Power lab facility This lab supports testing of equipment operating at up to 1 3 MW with maximum supply ratings of 4 160 V or 480 A depending on how the system is configured When operated as a recirculating power ring the system will make minimal demands on utilities An industrial supervisory control and data acquisition SCADA system will be used to control the facility In this paper we will detail the lab design and give insight to the decisions behind it with an aim toward helping the reader in their own similar effort The terms
45. DESIGN SCHEDULE zs a a a panels Checkout NENNEN sensors verify that logical signals conform 3 Software setup o Functional system P usta router firewall for all control equipment ul Stem semen montor VO tools states Create signal data Each I O point must be affiliated with a logical signal 34 Write HMI screens 22 04 Operator s displays and controls Write ss CU C C C CC i 103 CLINTON PERDUE APPENDIX HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN Appendix E High Voltage Grounding and Substation Design The following document comes from other contributors at GE who consulted on this project I have included it for background and to help maintain a cohesive record even though it is outside my scope I am not sure who to attribute this to but I thank them for their insites High Impedance Ground Fault Detection Discussion The ground fault protection GE Drive System s 2300 6600 VII rms power electronics equipment and that envisioned for the Va Tech high power lab is of the high impedance type that is there is a single point in the power electronics which is tied to cabinet frame through a moderate impedance The cabinet frame is then tied to earth ground One purpose for the high impedance ground technique is that on a first ground fault the fault current will be
46. Disconnects_open_161B 9 P Off Disconnects open 161B EM N ott Door 1618 Vt On ROOM 161B HV power A HVSwitch1 BreakerA Closed HV A on 161B 10 Vt d H O On Off DURIP State Control_Debugger MasterCarme Ladder_Program 9 18 2006 5 35 29 PM Page 2 68 CLINTON PERDUE APPENDIX LADDER PROGRAM 161 HV power ON HVSwitch2 BreakerB_Closed HV_B_on_161B 11 Vt 1 H O Off Off Room 161B Any HV power on HV_A_on_161B HV_on_161B 12 l O HV_B_on_161B 1 Room 161B Power ON state True when it is possible for 4160V or 480V outlet to be energized HV on 161 PowerOnState 161 13 il Safe_Plug161 LA v On Room 161B ACCESS VIOLATION Power on with someone in the room OccupiedState 161B HV on 161B AccessViolationHV 161B C n QU Off Room 161B ACCESS VIOLATION Power on with someone in the room OccupiedState_161B Safe_Plug161 AccessViolation480_161B 15 1 Vt C Off On 16 Room 171 171 4160 V disconnect switches HVSwitch3 HVSwitch4 Safe HV 171 17 a F l F n pP F a lt an rT OP Om LAQ On On ROOM171 480 V Outlet Off Check Plug 4 Plug 5 Plug IL 3 4 Safe Plug171 18 1 O On On On All manual disconnects reported open safe off by sensor inputs Safe HV 171 Safe Plug171 Disconnects open 171 19 1 1 Q Oc
47. ENTATION Chapter 3 Design and Implementation This chapter deals with the implementation of the control system for the High Power lab We will detail the design choices made for this application and illustrate how they satisfy both the operational requirements and the design philosophies outlined in chapter 2 In some cases the design effort was an exercise in hacking in taking existing things and making them work together by any means necessary This is a practical reality and a warning revamp installations are always difficult One manifestation of this in the VT lab is that all the disconnect knife switches are inside the lab bays instead of in a safe area This means the potential will always exist to close an energized switch while in proximity to the experimental equipment Secondly the switchgear and circuit breakers are not fail safe and cannot be modified to be so in that they A do not require an external signal to remain closed and B require external control power to open Lastly there is no way to lock out the 480VAC breakers We have addressed these issues as best we can in the planning of the software instrumentation and procedures Since the design requirements are formulated as high level statements a top down approach seems appropriate for explaining our efforts We begin then with the design of automation Later sections will deal with planning of instrumentation and communications network topology to get nec
48. Reading generic source data directly into the program WARNING what follows is an example based on our first efforts to get switchgear data into the control program It works but has serious consequences for execution time This method is no 88 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE longer recommended for communicating with the switchgear but is included because it may be useful for connecting to some other device The reasoning will be made clear in the next section Example Showing detailed instructions of how to read the time and date from the switchgear Time is a good dynamic easily verified signal to use for diagnostics a From inside the control project development software double click at the Control I O drivers then add a node b At the properties window the node is configured according to the setup we have This is shown in the next steps c The node type is a Generic device since we are communicating to the switchgear The switchgear communicates directly to the computer without using a control box and hence it does not use the VersaMax NIU d Select Modbus TCP for the protocol because that is the protocol for the switchgear All the other devices are made by GE and hence use GE s unique protocol known as EGD e The IP address of the switchgear has been configured at the router settings to be 192 168 1 130 f Enter 1 for the number of slots The number of slots designa
49. Students have designed and implemented their own graphical HMI screens for remote equipment monitoring and control There is an alarm system to annunciate events of interest and a data logger to keep a record The system supports a variety of coding languages for user functions which increases ease of use Importantly it is possible for users to play with graphics and controls without compromising the administrative software The system is secure both physically and electronically and it can not be energized accidentally The E stop interlocks are fail safe Loss of either external E stop or internal system integrity will trip the power as will failures of the controller or network Operations procedures have been drafted for the use of the facility We have compiled a body of documentation for future users and written project planning aids including commissioning guidelines In short we have demonstrated that this is a viable control system containing all the elements necessary to meet present and future needs At a minimum the control system will protect people and the power distribution hardware The infrastructure exists to support additional I O and logic that will be required to provide custom protection and instrumentation analysis for future research projects 65 CLINTON PERDUE CHAPTER 5 RESULTS CONCLUSIONS AND GUIDANCE 5 3 Future The possibility for future enhancements is basically unlimited by the syste
50. TROL SYSTEM REQUIREMENTS AND BACKGROUND possible to do basic frequency domain and statistical analysis and to export the raw data for processing in other packages MATLAB does not do well on bitmap images it needs numbers Operators of course deal in real time data They may want a short term trend display but the real concern is what is happening right now With this in mind most data acquisition and analysis functions are not going to be prominent on the operator interface screens They may not even be accessible directly why add the distraction 2 5 Redundancy and fail safe This topic warrants its own section even though we have touched on it in other places The design of a control system has to account for the things that will inevitably go wrong When this happens the control system must do whatever is possible to keep people safe This means detecting a fault and then de energizing the system in a safe orderly way 2 5 1 Fault detection the fail safe principle In some cases there are conditions that are so critical to safe operations they merit inclusion in a hard wired fault string This is a serial connection of indications that must be maintained true for the equipment in question to operate Primary among these are the Emergency Stop switches The operative attitude in fault detection is Guilty until proven innocent Sometimes twice Practically speaking if there is a state that must be maintained
51. UND It is also important to remember that computers are fallible Cosmic rays cause random changes to memory and or destroy components This can cause things to turn on all by themselves A computer must never be trusted to protect life and limb This must be emphasized in lock out tag out safety training for anyone working with any equipment Always put physical protections i e grounding in place before working on equipment connected to the utility or that could store energy Remember at some point a rodent will A chew the insulation off critical wiring or B decide to urinate on the high voltage bus bar In either case the control logic will not save you 2 4 4 2 The Administrative Role The role of the lab administrator is to make the control operator proof This means maintaining physical security and supervision of anything that is or could be energized The administrator will have to enforce lock out tag out procedures and supervise pre energization checks From a software perspective the administrative code will be that which sets the outer limits of the lab s use and behavior Users may write software as they require but it will never be able to supersede the administrative limits User code will issue requests for state changes the administrative code will vet these and issue commands to the actual hardware Data processing in this context is the collecting archiving and presentation o
52. ablish communications at all Lastly networked systems can be much more energy efficient both in the communications themselves and in power distribution to the I O loads Choosing what type of network to use can be challenging especially when multiple platforms and vendors are involved Generally speaking the first step in forging components of various provenance into a coherent system is to get as many of them as possible to communicate via some form of network 2 4 3 2 1 Serial networks Dedicated private serial networks are very common for applications with a small fixed number of devices Most test equipment for example typically has a PC style serial connection for data transfer There are many more varieties of serial networks ModBus ProfiBus Genius hp GPIB CAN bus USB and others which accommodate more devices Some are short haul only others can span hundreds of feet In general these connections work well Some are proprietary and very expensive Others were devised to operate on now obsolete hardware making them slow or otherwise limited The biggest problem tends to be shortage of connection ports Configuration can also be troublesome and require a bit of knowledgeable tweaking 2 4 3 2 2 Ethernet Ethernet has become a sort of lingua franca most modern instrumentation and control equipment either connects directly by Ethernet or can be adapted to do so by a bridge device Ethernet supports huge bandwidth and
53. accomplished we observed the dummy states faithfully reported to the HMI A word of caution One of the author s former colleagues tripped a very large utility substation by measuring voltage on a multimeter configured for current This should not happen here but you never know 56 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Satisfied they would communicate we mounted the modules in the field and repeated the dummy I O tests this time checking the newly installed network This lead to much practice re doing Ethernet connectors In the case of the switchgear we monitored the date and time status messages from the trip units to prove we were reading good data If an I O problem persists after the field wiring is checked verify that the wire has not come loose in the module terminal by pulling on the wire loose wiring terminals are the most common connection problem Above the individual signal wire termination the chain of network connectivity goes from module internal values through network bridge router PC and finally PLC software configuration Failures anywhere along this chain will affect multiple signals so looking for commonalities between malfunctioning signals is a good troubleshooting technique 4 1 3 Re verification Verifying that a system once commissioned still works just requires exercising all the devices and seeing them behave properly Think of it as a pre flight inspectio
54. are commissioned by safety obsessed professionals and then left alone not poked and prodded as an experimental setup will be Even so there are catastrophic failures The following operating procedures have been drafted with this in mind They will need modification as the lab equipment changes especially as energy storage elements and drive systems are added There are shortcomings in any mechanical layout or control system so it is imperative that lab users understand what procedural safeguards are in place and abide by them A major 50 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION deficiency of the existing system is that there is no way to lock out the 480V AC outside the room the 4160V AC can be locked out at the switchgear This could lead to someone closing a local disconnect that is hot courtesy of someone else s carelessness Procedure 3 4 1 Safing the room must always be followed even when only making minor changes such as moving probes or changing low voltage connections People get killed by dropping tools or loose screws or by touching the wrong thing accidentally Procedures 3 4 2 and 3 4 3 for energizing the room give the opportunity to check for faults If a system has been operating normally and has had only minor changes such as moving a probe then it can probably just be switched on again 3 4 1 Safing the room 1 Set all loads to off state 2 Open the appropriate supply co
55. atus screen shot iot aliens tats ates 60 Figure 4 5 Access violation testing 2 62 Figure 4 6 Transformer fault test eese 63 Tables Table 1 1 Testing topologies relative 7 Table 3 1 Possible communications 45 CLINTON PERDUE DEDICATIONS Dedications Thank you to Susan who got me through this Thank you also to all the GE field service employees who added a practical dimension to my education Finally thank you to the students faculty and staff who listened and put in resources to get this done CLINTON PERDUE ACKNOWLEDGEMENTS Acknowledgements Many thanks to TMEIC GE the successor of GE Industrial Systems for the donation of equipment and to the staff there for their time and encouragement Thanks also to the staff of GE Supply in Salem Va and of GE Fanuc Automation in Charlottesville Va for their assistance both in gathering the materials for this effort and troubleshooting during commissioning vil CLINTON PERDUE CHAPTER INTRODUCTION Chapter 1 Introduction 1 1 Our Motivations Power electronics have reshaped our world in the past fifty years by making electrical energy conversion equipment vastly smaller cheaper more capable and more efficient The ubiquitous cell phone is a good exampl
56. connections we began by planning an Ethernet backbone for the lab figure Figure 3 4 A router was the first necessity providing security and privacy by keeping external traffic off of the control network We have installed an 8 port Linksys model for this purpose Category 6e cable was used throughout because it is shielded and we expect a great deal of EMI Crimp on RJ 11 plugs were installed as needed Even though the proper crimp tool was used to attach the plugs they have been a repeated point of failure Patch panels with insulation displacement sockets would probably Work better Any PC will of course be able to use Ethernet so connecting a soft PLC and HMI viewers was no problem The GE drive supports a direct Ethernet connection via its ACL coprocessor using a proprietary EGD protocol This is inconvenient but still feasible No test equipment is connected at the moment but Ethernet connections on such are commonplace 46 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION The Square D switchgear is connected to Ethernet via a bridge The switchgear uses a 4 conductor serial Modbus internally with two drops for cradle communications modules on the network These modules interface to the trip units CPUs Communications to the switchgear are addressed by the Ethernet IP address of the bridge the Modbus drop address of the CCM and finally the register address of interest User software can query the switchgear di
57. converter ground faults in the source converter the DC link the load converter the load cables and the load or load transformer secondary if that be the case such as for the pump back test configuration envisioned for the Va Tech facility 104 CLINTON PERDUE APPENDIX E HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN Discussion of Several High Impedance Ground Fault Detection Approaches Fig 1 illustrates a ground fault detection approach based on sensing unreturned current The source neutral is connected to ground through some impedance which could be zero Fig 1 shows all three line currents going through one common CT but each line could have its own CT and the three current summed electronically or otherwise In the absence of a ground fault in the LOAD downstream of the CT s the currents should sum to zero It is envisioned this technique will be used to used to detect a ground fault in the transformer 480 rms primary winding assuming the 480 house feed to the Square D or equivalent low voltage switchgear is a three phase wye with neutral tied to ground Is is also assumed the transformer primary ground fault detection will be derived from the three line CT signals in the low voltage switchgear and will be summed and compared to a threshold in the low voltage switchgear electronics Note the cabling from the house feed to the low voltage switchgear is not included in this ground fault protection zone
58. ctable Note that just because a particular sensor is healthy that is no guarantee it is still physically in position to take a meaningful measurement Whenever possible crucial states should be monitored by three independent sensors From the multiple inputs a robust state estimation can be achieved Often it is better if the sensors are measuring different kinds of quantities so they are not all fooled by a single phenomenon For example a valve position reading may corroborate a temperature to out vote an alarm from an erroneous pressure reading which may be due to vibration There is always a temptation to choose a fancy sensor but often a simple and cheap one will do the job For example when measuring ambient temperature or coolant flow is an analog value really necessary or will a discrete threshold detection suffice Sometimes not motor windings and bearings are typically fitted with analog thermocouples so that rate of rise can be monitored Often multiple discrete sensors can be used to provide a distributed robust sampling which may be a better design Environmental factors are just as important in actuator selection as they are in choosing sensors The control system must also monitor actuator health Certain types solenoids motors can be monitored for short open circuit conditions Others such as servo actuated valves which regulate a position necessarily incorporate sensors in their control mechanism This se
59. cted You can not test for zero volts measured as a sensor fault state either because zero is a legitimate value In the end we decided to positively measure the mechanical position of the disconnect switches A position sensor with a wobble stick a flexible spring coil actuator is mounted inside each of the manual power disconnect switches These are placed so that when the disconnects internal cam snaps to the open position it deflects the actuator and the sensor switch closes 42 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION The disconnects cam moves quite violently in order to break any arc that may be formed if it is operated under load but a wobble stick can withstand this Figure 3 3 High voltage switch status sensor open and closed Knowing that the lab is vacant is a more difficult proposition We have mounted position sensors with roller tipped rocker arm actuators on the inside of the labs door lintels A closed circuit indicates the labs doors are closed We have not found a satisfactory way to detect people in the room directly Acoustic motion sensors mounted on the ceiling might work but cooling fans will be quite loud and could interfere Infrared sensors would be hopelessly blinded by heat sources in the room and vision systems seem overly complex People counters at the door such as photo gates would be very fallible unless there was a turnstile or other single person passag
60. ction Bulletin Cedar Rapids IA USA 111 CLINTON PERDUE VITA Vita Clinton Lee Perdue was born in Loudon Virginia on September 20 1970 He grew up on a small farm in Northern Virginia developing an interest in the sciences through applied experience Clinton matriculated to Virginia Polytechnic Institute and State University Virginia Tech in August of 1988 as an engineering student and member of the university s Corps of Cadets After receiving a BSEE in 1992 with concentrations in controls and microprocessors Clinton accepted a position with General Electric Industrial Systems in Salem Virginia Over the next ten years this led to extended tours designing and commissioning control systems for heavy industrial facilities in The Peoples Republic of China Brasil Canada The Netherlands and domestically in the USA Clinton has served in the post of Royal Bard to Princess Susan of the Bunnies since 1992 He was made Knight Order of the Shivering Duck in 2002 and been decorated with the Furtive Hare in 2005 Clinton returned to Virginia Tech for a class in 2000 and to seek an MSEE in 2002 In 2005 he took a full time position with Panaphase Technologies in Blacksburg Virginia developing drives and controls for switched reluctance motors 112
61. ctions 1 When adding instructions you may click on an empty space of rung and type in the command if you know it or select it from the list of instructions If you know the instruction name you only need to have the empty rung highlighted and begin typing 97 CLINTON PERDUE APPENDIX B USERS GUIDE 2 After the command is added the next dialogue box that appears is the variable selector which you use to associate a variable with the instruction Again if you know the variable you may type it in or begin typing it and it will match letters with variables on the list as you type B 4 3 2 Basic Useful Instructions NO Normally Open contact NC Normally Closed contact Out Output Relay TON Time delayed turn ON Time filter on rising edge Note that name of instruction instance is a scratch variable for the time elapsed don t reuse by mistake TOF Time delayed turn OFF Time filter on falling edge as above TP Time pulse Output is true for time duration after rising edge of input Again instance name is the scratch time elapsed variable B 4 3 3 Contact and coil types a b d P Positive edge triggered pulse input goes true output pulses true for one scan N Negative edge triggered pulse input goes false output pulses true for one scan S Set output lt name gt lt name gt will latch true R Reset output lt name gt lt name gt will latch false
62. cts An electronic copy of this document is located in the DURIP system directory B 2 4 Wiring Note that field wiring is shielded Power to the field device comes from the terminal strip to the left of the modules as close as possible to where the return signal conductor is routed The cable shield wire is tied to earth ground to the left side as well keeping the shield intact up to where the other conductors part The best resource for good wiring practice we have available is the instructions GE publishes for installing their drives and control systems Installation Guidance For Innovation Series Drive Systems GEH 6380 1999 General Electric Company USA Chapters 4 5 6 7 and 9 are particularly useful Some general points to keep in mind e Keep power and signal wires as far apart as possible and cross them only at 90 degree angles and never wrap one around the other e Cable shields must be tied to ground but at one point only usually the end near the I O module is most convenient e Conductors must be sized for the expected current they will carry so wire for relays and solenoids will be larger than for sensors The AWG wire size is the log of the resistive power lost for a given current Three sizes larger is 3dB more loss or twice the power so then twice the resistance If you get too much voltage drop due to wire resistance in a 16 wire 13 will cut the drop in half 84 CLINTON PERDUE APPENDIX
63. cts The contact will remain open without any given power When power or a signal is applied the device closes Normally Closed NC contact The contact will remain closed without a signal or power and closes in the presence of a signal or power In the ladder logic program the NO and NC contacts are represented by a capacitor like symbol and the same with a strike through it respectively Regardless of what the actual sensor is you may still chose either a NO or NC instruction in the ladder logic For example the door sensor physically is a NO switch Assume the variable assigned to the sensor is called Door When placing an instruction in a rung of the ladder logic despite Doorl originating from a NO sensor I can still chose either a NO or NC instruction to place in the rung 96 CLINTON PERDUE APPENDIX B USERS GUIDE Tt is important to understand that ladder logic is just a software language designed to look like physical relay logic What you write does not necessarily exist physically The instructions you place are generic and the variable that controls them may or may not physically exist Bottom line The ladder logic instructions do not model the physical switches associated with the controlling variable The variables assigned to an instruction simply control that instruction that you place All that is important with these variables are their value and how it changes your instructions you are not modeling
64. cupied State Reset with door closed transition OccupiedState 171 Door 171 OccupiedState 171 201 1 R DURIP State Control_Debugger MasterCarme Ladder_Program 9 18 2006 5 35 29 PM Page 3 69 CLINTON PERDUE APPENDIX LADDER PROGRAM Occupied State Latch true with any change of disconnect status or with door open TPulse_2 Disconnects_open_171 TP OccupiedState_171 21 IN Q S PT ET Disconnects open 171 mN Door_171 LT NsxV ROOM171 HV power AON HVSwitch3 BreakerA_Closed HV_A_on_171 22 Vt T Q On off ROOM171 HV power B ON HVSwitch4 BreakerB_Closed HV_B_on_171 23 Vt 1 Q On Off Room 171 Any HV power on HV A on 171 HV 171 24 HV_B_on_171 1 Room 171 Power ON state True when it is possible for any 4160V or 480V outlet to be energized 171 PowerOnState_171 25 1 O Safe_Plug171 gt Room 171 ACCESS VIOLATION Power on with someone in the room OccupiedState 171 HV on 171 AccessViolationHV 171 26 N 11 O Room 171 ACCESS VIOLATION Power on with someone in the room OccupiedState 171 Safe_Plug171 AccessViolation480 171 27 1 O 28 SwitchgearTemperatureChecks Determine if a High Temperature is because of Switchgear Breaker A 30 second lag TON 1 TOF 1 ThermalTransformer 1 ThermalAmbient 139A ThermalAmbient_161B ThermalAmbient_171 TON TOF 29 hl IN Q IN QT T IWE On On On Off PT ET PT ET HVSwitch1 HVSwitch3 1 1
65. d 1 5 Lab floor plan may be helpful at this point We need to know e If the lab bay doors are open or closed e If the 480 4160 VAC distribution knife switches are open or closed e If the high power distribution breakers are open or closed e Ifthe lab bay ambient temperature is acceptable e If the substation ambient temperature is acceptable 27 Please see VersaMax Modules Power Supplies and Carriers User s Manual GFK 1504K or later revision GE Fanuc Automation for a complete discussion of all the elements of a VersaMax system and available modules 40 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION If the transformers temperatures are acceptable If there is an E stop command from the lab These allow us to determine as certainly as possible that the lab is vacant or occupied and that the power is on or off Additionally we would like to know e Voltage and current at the switchgear e Current from the second floor substation These will be useful in determining time heating and other fault states The system drives the following outputs e Warning lights 3 and siren at lab bay doors e Permission to energize high power switchgear e Permission to energize 480 VAC distribution Figure 3 2 I O locations places these devices on the lab floor plan Please note that I O cabinets with interface modules have been placed near clusters of I O devices This keeps wire runs to d
66. dition but our controls do not consider these factors so the test is realistic Initially the High Voltage disconnect switch feed from CB 115 open so the room is safe as far as this breaker is concerned The switch is closed at 22 51 40 shown by the signal HV Switch going false The breaker should open at this point because a person must be in the room to close the switch ergo the room is occupied The breaker does not open in this case revealing a bug which was soon remedied We are reminded again testing is important Continuing with the test analysis then two seconds prior to the cursor position the door to room 161B is opened and Door 161B goes false This time the access violation 15 recognized and the output Energize Breaker A which permits the breaker to stay closed goes false immediately The breaker confirms that it is open within one second When the room door closes again at 22 52 00 the access violation condition is resolved but the breaker remains open Recall that the breaker is behind a locked door and must be manually reset 61 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Y axis title 22 51 40 22 51 50 22 52 00 22 52 10 22 52 20 22 52 30 22 52 40 Time DOOR_161B 2 HVSWITCHi 3 BREAKERA CLOSED 4 ENERGIZE BREAKER 5 Sunday April 09 2006 22 51 56 Figure 4 5 Access violation testing of CB 1 The second test Figure
67. ductors as a rule should be no more than two wire sizes smaller that the supply wiring Anything less is not a ground it is a fuse adjustments are to be made to 480VAC connections lock tag them out The guard on the main breaker may be relieved once this is done End 3 4 2 Energizing 480VAC wall outlets 1 Safe the room All energy sources must be safe before entering the room even ones not directly associated with the circuit to be energized 2 Post a guard on the main breaker There is no way to lock out the main breaker so a person must be delegated to watch it 3 Open the knife switch if it is not already locked out In case there is current present this will contain the arc Stand to one side of the panel and look away when you operate the switch 4 Open the panel door 5 Verify that both the line and load side voltages are zero Be sure that what you think you turned off is really off 6 Remove safety grounds 7 Test for load side shorts between phases and to ground Bad connections bad insulation or other damaged components tools or debris left in equipment can all cause explosions when equipment is energized 8 Shut the panel 9 Double check that the main breaker is still off 10 Remove locks tags from wall switch Everyone who has been working on this circuit has to personally remove their own safety lock out 11 Close the knife switch 12 Verify that proper status is reflected on the HMI
68. e 63 CLINTON PERDUE CHAPTER 5 RESULTS CONCLUSIONS AND GUIDANCE Chapter 5 Results Conclusions and Guidance As with any major construction project building the high power lab has had things that went well and things that didn t A unique aspect of building a laboratory space is that it is never really done At this stage we have a proof of concept for each critical feature of the lab and a rational framework for adapting the facility to meet the needs of future projects 5 1 Weaknesses in the design The first phase of our effort was to identify specifications for the lab and characteristics of the equipment As a result of this analysis we have identified three features of the lab which make its safe operation problematical Overcoming these issues is possible and we address them again here First the SCADA system is unable to directly positively observe that the lab bay is unoccupied We have overcome this by designing operating procedures and hardware interlocks that require a person to always have the final decision to turn on the power Secondly there is a danger of closing an energized disconnect switch The control has successfully negated this problem by interlocking the doorways with the power enabling outputs If procedure is followed the room will be empty before there is any power available and if someone enters a necessary step to close a disconnect then the feeders will trip One exception
69. e Appendix C for elementary diagrams EStop1 AA and LVSGIAA as well as the switchgear datasheets for additional information As stated previously the switchgear itself has a sophisticated power analyzer built into the trip unit This provides breaker status as well as voltage current and other data via a ModBus network 3 2 3 Outputs The system outputs include the permission to energize signal detailed above as well as commandis to light trees mounted beside the two lab bay doors The system states Safe Operating and Fault are represented by the lights As noted before though colors mean different things to different people so a clear explanatory sign is required for this context 3 3 Communications systems We have already touched on the communications requirements of some key equipment in the lab The following section details the complete communications topology for the lab As each element of the system was considered we had to identify what its possible modes of communication were and choose the most practical connections to make The results are tabulated below Installed indicates that the equipment on hand or in general has such a port Available means that such connections are possible usually requiring a special piece of 44 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION interface equipment PC card bridge etc Connections that are unavailable or extremely impractical are mar
70. e a static inflexible construct it would be much less useful 66 CLINTON PERDUE APPENDIX A LADDER PROGRAM Appendix A Ladder Program This is the ladder logic program and I O map extracted from the lab controller along with a data transfer script that runs as a background process on the controller 67 CLINTON PERDUE APPENDIX A LADDER PROGRAM 6 1 Ladder Program 1 START Check on status of HV BREAKERS not the wall switches the ones in room 139 AND EQ BreakerA Closed 2 EN DN EN Q O Internal_BreakerA_fdbk A 1 TempINT1 A 64 0 0 1B 1 EQ BreakerB Closed 3 EN DN EN Q O Internal_BreakerB_fdbk A C TempINT2 TemplNT2 A ff 80 0 0 1 8 1 4 Room 1618 ROOM161B 4160 V disconnect switches HVSwitch1 HVSwitch2 Safe HV 161B 5 H 11 Q On ort Off ROOM161B 480 V Outlet Off Check Plug 1 Plug 2 Safe Plug161 e 3 E Q On On On All manual disconnects reported open safe off by sensor inputs Safe HV 161B Safe Plug161 Disconnects open 161 7 E O Off On Off Occupied State Reset with door closed transition OccupiedState_161B Door_161B OccupiedState_161B 8 E R Off On Off Occupied State Latch true with any change of disconnect status or with door open TPulse_1 TP OccupiedState_161B LS S MM S 5000 0 Off
71. e explosions when equipment is energized 8 Shut the panel 9 Double check that the main breaker is still off 10 Remove locks tags from wall switch Everyone who has been working on this circuit has to personally remove their own safety lock out 11 Close the knife switch 12 Verify that proper status is reflected on the HMI The SCADA can not protect anything without proper feedback 13 Energize auxiliary sources 14 Clear everyone from the room and close the main doors There is no way at present for the SCADA to know if the room is truly empty be careful 15 Verify permission to energize status on HMI This ensures that there are no other system faults 16 Lab administrator will go to room 139 and close the MV switchgear 17 Close any 480VAC circuit breakers as required End 2523 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Chapter 4 Testing and Verification This chapter will document the commissioning of the control system hardware and software This is how we tested the new equipment and software so this chapter can serve as a guide for others re verifying system integrity after it has been modified The fastest way to find faults in a causal chain is with a binary search Cut the chain in half and test a segment If that segment is OK then the fault must be in the other segment and you have just avoided testing half of the nodes Cut the faulted segment s in half and repeat An ob
72. e of this on a low power scale The casual observer will have noted that both handsets and chargers are continuously getting smaller and lighter and batteries last longer Part of the vision of The Center for Power Electronics Systems at Virginia Tech is to realize similar improvements at the high power scale in large electrical power utilities Much of the research and design work that goes into developing a new electronic device or control system depends on computer modeling Today there exist many simulator platforms MATLAB SABER and others with numerical models for the well understood aspects of conventional power systems Why then is CPES interested in building a facility to physically test power electronics In a word research Investigating new concepts in device physics manufacturing control schemes and the like will require not just using computer models but creating new ones Those models in turn need to be validated by comparing their predictions to real world test results Some of this validation can begin on low power or physical scale models but truly new work will ultimately require a full size full power prototype The goal of the HIGH POWER LAB project is to build the attendant test facilities required for novel work in utility scale power electronics To this end the Office of Naval Research ONR has granted Virginia Tech funding under the DURIP program to build the high power testing lab 2 Wolfram Resea
73. e script is to create a application script that runs periodically Application script will run regardless of which panel is currently being displayed In the script you would do the assignments of the PLC Access variables to the internals variables ie InternalVar1 1 91 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE InternalVar2 PLCVar2 You can setup the script to go other other way Internal variables being using in the LDPC to the PLC Access variables ie PLCVar3 InternalVar3 PLCVar4 InternalVar4 The one problem with this is that the communications will only go one way If you try and send variables back and forth between two variables you will run into problems GE Fanuc Tech Support Q One final question What is the recommended way to accomplish bi directional communications A If you are talking about bidirectional communications through scripting It can be done but there are a few things that you need to be careful about If the data only flows one way in some variables and the other way in the other variables There shouldn t be a problem You just have you script written like InternalVar1 1 PLCVar2 InternalVar2 But if you want the data to go both ways on the same tag there may be some problems If the value on both sides change at the same time you will need to determine which side to write from You also don t want to continually write
74. eir placement and connections A bill of materials is included in appendix 2 Choice of the SCADA software package further constrained us to use GE I O products Others devices are available but using them would forfeit the benefits of using an integrated family of products Configuring communications with brand X as we discovered with the ModBus quickly becomes an exercise in low level programming and data formatting GE offers several lines of I O products We have chosen the VersaMax system because it is modular expandable and relatively inexpensive The Ethernet bridge unit and power supply clamp to a standard 35mm DIN rail Carrier modules also clamp to the rail and plug serially into the side of the bridge unit These provide serial network and power connections to the bridge as well as screw terminals for field wiring to I O devices Finally the actual I O interface module plugs into the carrier Interface modules support between four and thirty two channels with input or output analog or discrete source or sinking current or voltage signals operating from 12 VDC up to 240 VAC There are also special functions such as PWM output high speed counter input and direct interface to RTDs and thermocouples 3 2 1 What I O From the discussion of the automation design we can begin to decide what sensors are required and where they must be placed Refering back to figures 1 4 Power distribution system elementary diagram an
75. ences 7 114 D amp Done 1n7 Col 11 Online Administrator LOCAL DURIP State Control_ 161102 au99 vsd M Ey io status doc Micros iy VIEWSCRIPT txt Mi View Network View Runtime Figure 4 4 Screen capture of control system editor in online monitor mode showing I O status as used in logic Bold green highlighting indicates that the signal is active It is worth reiterating that fail safe signals show a positive feedback when in a healthy condition and a null when the system is inoperable Invariably the sense of some input will be reversed showing true for inoperative false for OK and go undetected until software checkout The temptation will be to reverse the logic at this point but this must not be done Not unhealthy is not equivalent to definitely healthy Go and change the wiring 4 2 2 Logical tests As with any other software the time comes to do live testing on the system by setting up realistic scenarios and observing how the system reacts The following Figure 4 5 illustrates two different access violation conditions where a person may be in proximity to energized equipment The line side utility feed to the switchgear 60 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION is not energized and the breakers are racked out to the logic test position where they operate but do not make contact with the main power buss This is a safe con
76. erated by one or more smaller AC DC MG sets CLINTON PERDUE CHAPTER INTRODUCTION complexity These support systems significantly complicate the system controls as their status become critical states that must be monitored when the system is operating As a point of reference see Figure 1 2 below for a view of the sort of large machines we are considering This machine has three lubrication systems a liquid heat exchanger a forced air system and twelve thermocouples mounted in the windings and bearings The gearbox not visible has an additional two lubrication systems liquid cooling and more thermocouples Any load would be similarly complex Obviously the space and mechanical support requirements are not trivial either Far from being unusual this is one of four 11 MW and five 5 MW units all in a row at this site a fairly typical industrial facility the really big machines are found in mines and on ships Once again we see the need for a large scale testing facility for power electronics and controls 7 This particular machine has suffered a burned stator winding one of 144 probably caused by a void in the insulation The drive system was probably capable of compensating for the bad winding but such an algorithm had never been tested on this scale and the customer was disinclined to host a science project Consequently the motor manufacturer is replacing this unit at a cost of approximately US 2 000 000 00
77. erful that sometimes they can replace the PLC itself the tools become a soft PLC A soft PLC has decided advantages over a hardware one If the computer the PLC software runs on breaks or simply becomes obsolete it is cheap and easy to replace This is an important consideration because industrial customers often find themselves paying huge premiums for a last of its kind piece of control hardware Also any tools to configure and troubleshoot the control will not be marooned on an obsolete computer OS kept solely to support the controller 2 4 4 1 The Role of the Control System vs the Role of the Operator There is a temptation with any supervisory system to make it the responsible party in any possible fault situation It shouldn t let the operator do that is a common refrain This sounds fair enough but remember you can make the control bullet proof but not operator proof The control system can observe advise and impose some degree of order but if people behave recklessly especially by defeating safety interlocks then nothing is going to stop them from causing a disaster 23 It can be quite literally child s play LEGO produces a simple PLC for automating toy robots There is actually a conventional hardware PLC add on board integrated into the lab s GE drive We not using it because if it breaks we may never find a replacement 230s CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGRO
78. essary information to the automation system Lastly we deal with design of operating procedures 3 1 Automation design The very highest edict of the specification is that the lab is to be safe to work in The first step is to decide what this means Our solution is to with a high degree of abstraction describe the lab s overall condition in terms of a finite state machine with three basic states e Operating when the power is on and the system is operating normally e Fault any time there is a condition that should inhibit energizing the system e Safe when the power is reported by the control system to be physically disconnected and the controls are operating normally CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION This state machine determines whether the lab utility connections may be or remain closed or if the system automation must open trip them The inputs which drive the transitions between system Safe Operating and Fault states are Power On Off Room Vacant Occupied and Control Healthy Fault Since these are themselves abstractions the resultants of many low level inputs and other factors we term them mid level states These are described as follows with the system state results illustrated in figure 3 1 and associated ladder logic code in Appendix A For Lab Occupied If the door is open the lab space is considered to be occupied Additionally any transition of a sensor in
79. etwork for the bulk of control and data communications Safety critical signals will be carried by discrete hardwired circuits e Logical elements in the system will continuously evaluate the state of the system and permit or prohibit the power distribution equipment from feeding power to the load Animated graphical status displays will indicate the state of the facility and illustrate the necessary conditions to operate the power system 14 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND system monitor will periodically poll the system internal states and log them It will be possible to trigger external data collection devices in a coordinated way e Useacommercially available system of small inexpensive network to I O transducer modules designed to be mounted near to their loads The individual transducers can be added or replaced as required Design enclosures with spare capacity for future needs and place them to minimize wire runs to I O devices e Wherever possible name brand equipment will be used so that the OEMs are more likely to offer future product support Preference will be given to products that are not tied to a particular generation of hardware or operating system and that have a large installed base e Selection will also favor tools and protocols that are well documented have technical support available and are industry standards if not open source e Safety and user
80. evices as short as possible To serve the I O needs detailed above we require discrete 24 VDC inputs and isolated relay contact outputs These have been provided along with a judicious number of spare points in consideration of future applications Additionally we have selected modules which may serve as PWM drivers or high speed counters Please see appendix 5 for wiring diagrams 41 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION s reer Physical layout Control output 138 161 Alarm ight tree Medium Power Lab High Power Lab 139A High Power Equipment Room Elevator Figure 3 2 I O locations 3 2 2 Sensor hardware The sensors we have selected are all metal bodied discrete mechanical types These were chosen over other types in an effort to improve immunity to EMI and because they tend to be physically robust They are also simple to use and explain to people they are either on or off no confusion about the meaning of an analog value is possible Analog signals are certainly supported by this system but generally speaking discrete measurements are sufficient to indicate if a situation is within safe limits At first there was a push to measure voltage at each supply outlet and equate a zero voltage measurement with a safe condition This violates fail safe philosophy because zero volts measured can simply mean that the sensor is broken or disconne
81. eway This would be unacceptable because it would get in the way of moving equipment and slow emergency egress In the end it comes down to trusting in good operating procedure rather than infallible sensor schemes The ambient temperature sensors are mounted near the ceiling where heat is expected to collect These are a simple bimetallic strip type with an adjustable trip point set at 55 deg C This should give a generous margin for operation yet trip before wiring insulation really begins 43 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION to suffer Finally there are other similar temperature sensors imbedded in the transformers by the manufacturer Their setpoint is unknown but presumably appropriate to the machine The lab at large has an E stop system based on hardware relays CRLV CRMV CRHV which stand for Control Relay Low Medium High Voltage This is a misnomer as the lab is divided into areas by power Our lab is part of CRHV s bailiwick CRHV had a spare contact which has been used to signal its status to the high power lab trip circuits by energizing a secondary HVOK1 relay also depends a permission to energize output from the control system to close or remain so If either of these conditions fails HVOK1 opens thereby opening HVOK2 and HVOK3 which in turn send trip commands to the high power breakers Loss of HVOK1 will also send a trip command to the 480 VAC main breaker Se
82. f the status of the equipment over time The time span can vary from a few milliseconds during a transient test to days or weeks for statistical trend analysis One of the bigger challenges is to collect 5 During one modest solar storm soon after the VT supercomputer came on line approximately 11 such events happened in its main memory 26 The type of IGBTs used in the GE drive have been seen to fail spectacularly due to the impact of an energetic neutron Borovina D L et al Neutron Induced Failure Tests of 3300 V IGBTs for the Spallation Neutron Source Accelerator Proceedings of the IEEE Particle Accelerator Conference 2003 ZA CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND enough of the right kind of data so that unexpected relationships can be sussed out after the fact This is again an advantage of having all possible equipment communicating on a common network place a data recorder on the net to eavesdrop on and store whatever comes by Part of making use of this data is knowing when things happened as opposed to when they were recorded All the elements in the system will typically be referenced to a stable common local clock This in turn will be synced to a high quality time master such as an atomic clock which is publicly available on the web This will allow meaningful time stamping of events A further understanding of latency can be gained by recording reference events through mu
83. frmo one side to the other ie InternalVar1 PLCVar1 If there is a new value in InternalVar1 it will be overwritten by the previous value PLCVar1 So you will have to determine when a value changes and only to write to the other side when it changes to avoid overwritting a new value on the other end GE Fanuc Tech Support 02 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE Table B 6 Key registers in the switchgear Register Number of Description Additional Information Number Registers 679 4 Shows Date Registerl Register2 ana ume Bytel Month 1 12 Bytel Year 0 199 Byte2 Day 1 31 PINN Mask Bits 14 amp 15 Byte2 Hour 0 23 on the Month Day Register3 Bytel Minutes 0 59 Byte2 Seconds 0 59 661 1 Breaker Bit 0 Position On Closed Status Off Open 661 1 Breaker Bit8 Disconnected Foston Bit9 Connected Bit10 test position 661 1 Breaker Bit 2 Tripped On tripped Trip Status Off not tripped 93 CLINTON PERDUE APPENDIX B USERS GUIDE B 4 Operating Cimplicity B 4 1 Connection Setup For every device on the network one must add that device to the Cimplicity Project this includes the controller this computer the view node and every other module on the network Adding and Configuring Network Modules 1 Right click in the project window on your project s name and Add Target a From there you will be able to
84. functions will be segregated so that the users can not compromise the safety functions This may require physical separation of the I O and control platforms lockable circuit enclosures and reference code templates Potentially dangerous actions will require a man in the loop e The system network s will either be isolated or reside behind a router firewall Additionally the lab is expected to be a high EMI environment so all circuit enclosures and wiring will have to be shielded This also leads us to favor I O devices which do not rely on field effects 2 3 SCADA what it means and why it satisfies our needs The implementation of many of these ideas is commonly referred to as a Supervisory Control And Data Acquisition SCADA system This is an umbrella term for integrated instrumentation operator interface and automation systems This is a core element of our proposed solution so at this point we give a brief explanation of how such a system works In the following sections we examine in detail the elements of aSCADA system and the design choices that were made in this case 55 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND The concepts certainly are not new since the advent of the industrial revolution people have placed recording devices on instruments and connected signaling and control devices from remote equipment to central control stations This term then describes a mature set of p
85. g and disconnects and the AC DC converter is beyond the scope of this thesis some interesting points of the design discussion are included as Appendix E All that work had been done and the equipment installed prior to this part of the project This thesis will document the design and implementation of a control system for the High Power Lab Our objective is first to identify the needs of the users and characterize the capabilities of the existing equipment Following this we must identify technologies and methods which may be brought to bear on the problem and plan a solution Detailed designs for networked and discrete hardwired communications selection and placement of sensors I O connection boxes and cabling graphical user interfaces and a system security plan are produced Finally the system is implemented and all major functions of the control are demonstrated concluding with an operational plan and user s guide The structure of this thesis is as follows Chapter one has oriented us to the motivations and methods of our lab and familiarized us with its major hardware and physical layout Chapter two will delve more into the operational requirements of the lab It will cover what we want the control system to do both in terms of enforcing safety protocols and of conducting experiments This is also where the philosophy of the hardware selection and the control design used in the lab is explained 11 CLINTON PERDUE CHAPTER
86. gn and construction Deliverables amp Milestones Panna Understand system List operator inputs Master I O list spreadsheet in project directory architecture amp physical interlocks system requirements control outputs and data Decide what I O is feedback Specify needed bandwidth sample rate and accuracy Layout rooms locate Identify device mounting Map of rooms devices positions cluster wiring into local groups Select I O devices Criteria range sensitivity sensors outputs rate Select I O modules Criteria protocol B comm diagnostics cost Transducers buss communications Layout I O panels Po onstruction drawings for I O Physical mounting Wall mountings din rails couplings holes Place power 110VAC 24VDC E net communications other networks breakers fuses grounding Place terminals Have connection point for Wiring elementary drawings modules each wire route wiring in panel Pay attention to voltage Physical wiring layout for construction levels shielding grounding Cable schedule low of conductors cables OM termination locations separation shielding OM OM Conduit schedule Protection as required BOM low level 5 7 Bill Of Materials for p r each panel CONSTRUCTION Assemble panels Install equipment internal low level wiring 102 CLINTON PERDUE APPENDIX D
87. he couldn t see them over his belly Recollections of accident report sent to other nuclear stations worldwide by General Electric engineer Christian Duplaa who was working at a French nuclear power facility at the time 15 The Apollo 16 lunar landing was almost scrubbed because of a mechanical fault in the abort switch which caused it to falsely indicate that it had been pushed Tn the minutes after the attack on The Pentagon on 9 11 2001 engineers working on the new environmental control system were able to reconfigure an undamaged control station to shut down air handlers in the part of the building that was burning thereby preventing the fire from spreading The primary controls had been destroyed Smoonian D Smart buildings Saving the Pentagon IEEE Spectrum Magazine IEEE On page s 18 23 Volume 40 Issue 8 Aug 2003 2902 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND then every control without some special requirement to be discrete tends to become a screen graphic There are three major things to consider when designing a graphical control station e Graphics must be designed for appropriate density and clarity of information Don t add distractions or fluff Emphasize the important and limit opportunities for misinterpretation e Organization of controls and navigation between groupings Make the right things available e Alarms and other messaging Make these punctual meaningful and
88. her the ground fault occurs in the source DC link or load and makes use of the fact that the ripple voltage amplitude and frequency on the summated line voltage signal signal will vary depending on where in the equipment the ground fault has occurred Fig 4 further expands on the equivalent circuit of this detection scheme and Fig 5 shows the detailed schematic 107 CLINTON PERDUE APPENDIX E HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN Fig 5 Detailed ground fault detector schematic A different ground fault detection scheme in used in GE s high voltage 2300 6600 V11 rms voltage source equipment This equipment uses either IGBT or IGCT based three level PWM source and load converters separated by a center tapped DC link capacitor The center tap of this DC link capacitor is connected to ground through a moderate impedance of approximately 40 ohms If a ground fault occurs anywhere in the system secondary of the transformer s source or load converter DC link or source or load AC line cables voltage will appear on this grounding resistor and when greater than a preset threshold annunciates a ground fault detected signal which can be used to shut down the equipment For redundancy in conjunction with DC link ground resistor the three phase AC line currents in the source converter are summed and when this summation signal is greater than a preset reference a ground fault signal is annunciated that a ground faul
89. hilosophies for doing the things we need in a facility like the high power lab The Supervisory aspect of such a system means the system designers set operational limits on the equipment which may not be exceeded by the operators The users operators may also set their own limits on the equipment for the system to automatically enforce A good example of supervision would be a pair of limits on a time power heating curve model used to govern the operation of a machine The higher less restrictive limit could be set by lab administration to protect the distribution equipment The lower tighter limit could be set by a user to protect some experimental device perhaps by reducing a reference Figure 2 2 This layer of supervision should not to be confused with the system fail safe protections It is only a regulation a software enforced limitation on operation that allows the system to continue running The fail safe protections like the E stop buttons always force the system toward a safe shutdown condition These concepts will be explored in more detail in section 2 5 Redundancy and fail safe One of the very early applications of a computer to a SCADA system was the automation of the Norfolk and Southern railway switching yard at Roanoke Virginia by the General Electric Company A tiny computer with only a few kilobytes of memory maintained records of the position of cars and switches throughout the yard Recollections of in
90. ices as in the light trees In either case the impedance can be checked at the module end to see if the load is as expected 4 1 1 3 Analog signals We do not have any analog signals in the lab safety system yet but commissioning troubleshooting techniques for them are covered here for completeness sake There really is not too much difference with discretes continuity must be maintained just the same The biggest difference with analog signals is that checking them requires corroborating field measurements with the state of the I O module Use the HMI to set or monitor the module s state so you know what to expect to measure External precision current voltage sources can be used to signal either a load or input module to see if they behave properly Outputs can be validated directly with a multimeter The most difficult case will be checking an analog sensor because the state it is sensing must be known 4 1 2 I O network connectivity The second half of getting I O to work properly is communicating with and configuring the interface modules At an early stage in this project we set the network communications modules on a convenient workbench hooked in dummy inputs and went through the manufacturer s network configuration procedure The HMI has an online monitoring mode that let us verify that we had communications We then were able over the network to configure the modules to expect appropriate field connections This
91. icture Screen objects may also open special purpose sub windows help diagnostics or navigate to another set of controls entirely This makes the navigation more intuitive and allows for ample alternate paths to retrieve desired information It is important to consider the balance between using pop up windows which by necessity obscure part of the parent screen and jumping to an entirely different screen you then have to navigate back from Again it bears considering the task at hand 2 4 2 3 Messages and Alarms The system logic can be configured to recognize important events and log them for later analysis Some of these events are critical enough so that the operator needs to be alerted to them even if they are not related to things represented in the current screen view To accommodate this graphic displays commonly reserve an area of the screen for displaying messages Status messages may be classified and color coded according to the type and severity of information presented A message window will accommodate a small number of such messages usually the most recent couple in each category with older ones scrolling away The previous time power heating example Figure 2 2 may generate a sequence of messages like the following in Figure 2 4 25535 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND Screen graphic message type 13 24 07 22 Drive start Status message Warning message 13
92. ides wire termination points for external devices Carriers can be configured to connect to any module type by way of turn dials The modules present are an isolated relay output module marked red and a 24 VDC Discrete I O module yellow 82 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE y INNER EDGE OF Ci OF BOX Extents of accessible box area 22 22 AC DC Network Interface Module CPU buss Sas power supply signal iN DC distribution Shield earthing points power v out ical I O Cabinet INNER EDGE OF BOX Networking Signal conversion Figure B 2 Typical I O box with signal path shown l a a w t HJ a EI a a LI a LI a ju L 2 x 2 it H a Figure B 3 typical I O cabinet Currently Installed Modules e Power 24vdc Power Supply IC200PWR002C e NIU VersaMax Ethernet NIU 200 1 001 82 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE Carrier Base Box Style I 0 Carrier IC200CHS002 e Mixed I O 24VDC Discrete High Speed Counter Module IC200MDD841 e Mixed I O24 VDC Discrete Module ICLOOMDD842 Relay Output Module IC200MDL930 Further details about these items can be found in the VersaMax Modules Power Supplies and Carriers User s Manual GFK 1504K March 2003 GE Fanuc Automation Programmable Control Produ
93. ion Checklist 95 CLINTON PERDUE APPENDIX B USERS GUIDE e Message Time Out During normal communication this is the time the computer expects to receive a message reply before it decides that the device is no longer communicating properly e Reconnect Time Out The time allowed for the reconnecting device to reply to the reconnect request e Time Between Reconnect The duration of the pause between trying to reconnect The controller will always ask to reconnect B 4 3 Ladder Logic Ladder logic is an instruction language based on the flow of an electrical signal across various rungs Each rung is a row in which the signal begins from the left side and travels to the right side The flow of logic is from top to bottom and executes each rung in that order The logical flow is similar to assuming a basic current or signal that originates on the left and the instructions you insert into the rung act like switches closing or opening the circuit At the end of arung you may output the final value to a variable In essence the ladder logic simulates a series of relay contacts which affect an output signal which could also be seen as another relay elsewhere To understand the instructions knowledge of normally open and normally closed relays is required The term normally describes the state of the device while there is no power applied to relay Below is a brief description of these operations Normally Open NO conta
94. is useful for design and evaluation of mechanical equipment motors generators gearboxes couplings and the electrical controls invariably called on to compensate for their shortcomings A useful sub class of mechanical load systems would be motor generator MG sets In these arrangements a DC motor turns an AC generator to create a variable voltage and frequency 1 The advantage in a testing facility would be that a MG set could provide a degree of isolation from the world at large The logical extension of having a generator in the system is recirculating power back into the utility supply Besides saving on the power bill this allows investigations of both loads and generators along with their interactions with the power grid The disadvantage of a mechanical system is that the mechanical element is always there even when it is not of interest Beyond simple cost and space requirements these components will always cause more losses and noise Bearings brushes and the like must be maintained Large machines often require forced lubrication and cooling leading to another level of n further reference to THE MATHWORKS advertising copy despite having a very good model they still built a real machine to test Ibid 1 In the days before electronic AC drives MG sets were used to power AC motors which had too much inertia to start directly from the utility supply In truly ancient times the DC might itself be gen
95. ix C table 17 pp 69 75 Recall that the trip units communicate via ModBus serial and so has its own drop on that network but share a single Ethernet IP address when communicating with the control system Therefore addressing trip unit registers requires using the proper unit ID see below Table B 5 Switchgear identification Unit ID Breaker Associated Associated name position transformer HV disconnects 45 A top 1 1 3 47 B bottom 2 2 4 B 3 1 1 Register Configuration When configuring the master computer to read the registers of the switchgear it is important to note the following e Note 1 on Switchgear Offset The slot information configures the offset and the length of the register in bytes not registers This is important because the slot information displays the entered Register Offset in terms of the total bytes Example Register offset 32 Register Size 16 Bits The reported offset next to the slot number will be 64 because 32 registers corresponds to 32 registers at 2 bytes 8 8 bits each Note 2 Switchgear Offset The register offset declares the shift in read registers The first terminal to be read from this offset starts one register afterwards So if you are attempting to read register 671 you need to set the register offset to 670 Possibly because the register offset only provides the base and the register count starts after that point B 3 1 2 Method 1
96. ked N A Table 3 1 Possible communications links 9 Available communications modes Equipment Ethernet Hardwired Data logger Oscilloscope E stop system Equipment To be determined under test To the greatest degree practical all wiring connections were made in accordance with installation guidance instructions published by GE Industrial Systems This document is a rather exhaustive treatise on wiring methods and materials and is highly recommended 78 Instillation Guidance for Innovation Series Drive Systems publication GEH 6380 issued 30 June 1999 1999 General Electric Company USA 45 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION All sensors are connected with AWG 18 solid Cu wire with shield Some use a cable with a single pair of conductors and others are bundled into larger cables In a given instance the choice was based mostly on cost in some cases multiple small strands are cheaper than a single larger cable University regulations prohibit students from making connections to permanently installed equipment at other than 24 VDC so this was used for all the sensors Data loggers scopes and custom sensors Modbus and Lab Public EGD Ethernet Ethernet PC runnin PC with GE drive HMI 22 HMI viewer 1 External configuration tools soft PLC HMI viewer 2 HMI Figure 3 4 Ethernet backbone Given the prior discussion of the advantages of Ethernet
97. larger E ambient multiconductor cable 139 IOI Adjacentto To control SquareD Gateway VersaMax NIU the HV communications switchgear CCM 4 Power Supply Fuses and with the switchgear Thermal Sensors beside the and gather j IC200MDD842 i Transformer 1 2 139 switchgear information ambient IC200MDL930 on the wall concerning 139A Power Logic EGX100 81 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE B 2 3 2 Cabinet Construction In each cabinet there are devices that allow access to the surrounding environment The cabinets themselves are NEMA 1 steel enclosures grounded to provide shielding from radiated EMI in the lab bays The general layout of the I O cabinets is as seen in Figure B 2 with a real example in Figure B 3 The following appear as a series of black boxes Network Interface Unit NIU The network communication module uses an Ethernet connection to distribute data from subordinate modules and essentially interface with them It must be programmed to have a specific IP address to operate properly e Power Supply The power supply is connected to the surface of the NIU and appears as a thinner black box with power wiring and ground wires connecting to it e Modules Each module has different specific functions according to the specifications it was purchased for The modules must sit inside of a base or carrier which connects to the NIU through a backplane and prov
98. ltiple channels and comparing when they appear to occur An example of this temporal interferometry would be using a marker pulse on a motor s tachometer to trigger both an oscilloscope recording external data and an internal recording of the motor drive s states which then get uploaded to the common data recorder Seeing the same event from two perspectives allows one to put all the other data collected from the two sources into context However the two recordings in this example are likely to be fairly dense and of different formats Getting them into the common archive probably will not happen in real time Doing so at all may require manual processing A more reasonable goal would be for an outside recorder to simply grab the data and store it in a way that reliably correlates the two files with low bandwidth real time data from other sources This could involve yet a third recording of the marker pulse via a discrete input module The system data recorder will collect a broad but modest resolution view of the plant to add context to specially triggered high resolution recordings of interesting events With good planning the conditions surrounding an experiment or a fault condition can be reconstructed accurately To analyze these events data recordings must be searchable at will show me variables X Y and Z between then and now Once the interesting events are isolated it must be CLINTON PERDUE CHAPTER 2 CON
99. m design The SCADA software supports 75 I O points at present but this is only a licensing restriction It can expand to thousands as can the I O hardware Since we use a soft PLC we can always upgrade the computer as processing speed or other requirements dictate What follows is a wish list of features that could be added with little additional investment First much more could be done with the HMI to graphically illustrate the health of the I O and give diagnostic assistance to users Next we have access to data from the second floor substation via Ethernet and should consider adding this to our displays and interlocks We can also set up the HMI to publish data to remote users outside the lab Lastly we can use existing I O to coordinate data capture by external devices and integrate the result with our existing data log Since the lab already has water cooling installed it would be reasonable to instrument the flow as an interlock to energize a liquid cooled load We could also monitor temperature and perform valve control as required The existing lab protections would benefit from having redundant complimentary sensors to improve reliability Secure power for the trip circuits is important to add Lastly the mechanical fittings for lock out grounding could be improved with ball fittings for clamping to This may seem like a long list but that is a good thing because it indicates we have possibilities available If the lab wer
100. n Flaps Check Rudder Check It is easier than commissioning because there may not be any faults or at least probably only one in a given channel In practice this means the signal from physical connection all the way through to logical representation can be checked at once One must still go and move each physical actuator doors disconnects but one can then just check the system log to see that each transitioned when and only when it was tested An example in the following Figure 4 2 shows how automatically recorded data can be retrieved from the system log Ert CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Y axis title 100 12 30 00 12 35 00 12 40 00 12 45 00 12 50 00 12 55 00 13 00 00 Time DOOR 161B 7 DOOR 171 Sunday April 09 2006 Figure 4 2 Historical data retrieved from the system log showing traffic into one lab bay These are binary signals so the Y axis scale is arbitrary and has been set for readability The lab administrator may well require this log satisfying a checklist before allowing the lab to be energized Obviously a friend watching an HMI status screen which displayed the checklist progress and indicated which equipment to test next would be helpful 4 2 Software commissioning Having succeeded in getting all the I O signals into the system is a major step It still remains to be seen that the data is received and interpreted p
101. n to ensure that energy is available installing a battery backup system for example Sometimes simply turning something off when there is a problem is not the best approach Coordinated controlled shut downs are usually safer especially when there is stored energy to be dissipated Careful design will separate the indications and controls for these situations which may still be fail safe and redundant from the truly last ditch effort of E stop controls 2 6 Acase study The following is a hypothetical case study of a motor drive system undergoing a series of faults This will illustrate how the control system reacts and where redundant or fail safe elements help 224 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND SCADA controller Aux Control network Qolo DA 1 D A Emergency ny aS aces rumu a na U Logic NECS Switchgear ARD gt O O Hardwired Drive control Motor control N relays Figure 2 5 Typical SCADA for motor control A drive feeding a motor through a motor control relay suffers a shorted IGBT in the output stage The motor current increases beyond limits so the drive s logic removes the hardwired close command to the motor rela
102. neophyte screen designers tend to clutter screens with distractions This is not a video game Colors and behavior of screen objects also need to follow a consistent schema Ina classic case of dueling standards sometimes Stop buttons are red and Start green like traffic signals and other times they are reversed to green for Stop safe and red for Start unsafe Adding to the confusion in some schemes a control is illuminated when it is available and in others illuminated to signify when the machine is already in that state To make matters worse colorblindness is a serious consideration Clearly then since people s associations with colors lights etc can not be taken for granted There need to be other cues For example controls need to be clearly marked with verbiage like Push to Stop and Now ON 2 4 2 2 Organization and navigation In the discrete world you have to physically go to where a control or indicator is located Oftentimes this requires jumping back and forth between control stations or coordinating actions between operators who may not know what the other is doing Things are easier in the graphical control world Graphical screens potentially have access to and control of any state available to the control network This makes it practical to group controls in support of a given situation or procedure and regroup them differently on another screen for another purpose For example an I O check
103. ng this data is covered in the switchgear section B 2 3 Wiring Cabinets As of Rev 1 1 the I O modules are located in three wiring cabinets 161101 161102 and 139101 Note that the cabinets are named according to their physical location and function For example 1611O1 is located in room 161 is an I O cabinet with electronics versus junction box JB with only wire connections and is box number 1 in that area 80 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE B 2 3 1 Functions of Cabinets Table B 3 Cabinet functionality Cabinet Physical Purpose Devices Internal Designation Location Controlled Connected Hardware 161 IOI Far wall of Acts as a pull box HV Switch 1 2 3 4 VersaMax NIU 1618 for SOIBIHBHI QUOS 480 Outlets 3 4 5 4 Power Supply with room 171 the all 171 Switchgear the high IC200MDD841 voltage switches Thermal Sensors IC200MDL930 and deviceson 171 161B ambient the far side of the temporary functions room 161 IO2 Beside Monitors Doors 480 Door 161B and 171 VersaMax NIU window in outlets in 161B 480v Outlets 1 2 Power S pnl 161B Controls both light PPY trees IC200MDD842 IC200MDL930 161_1 1 Beside door Mounting and Room 161B light tree None in 161 connections for light tree 171 On side of Brings together the 480 Outlets 3 4 5 4 None HVSwitch4 four 480 Switches all 171 enclosure and thermal sensor Thermal Sensor 171 into a
104. nnections by selecting manual trip from HMI 3 Verify that appropriate supply connections are open Visually verify that 480VAC main breaker in room 161 171 is open Check switchgear status indication for 4160VAC supply on HMI is open 4 If 480VAC is used manually open appropriate branch feed circuit breaker 5 If stored energy is present in the experiment bay wait for it to dissipate Electrical codes state that capacitors should be configured to bleed off to 50 VDC within five minutes but don t trust this to actually happen check 6 If 480VAC supply is connected post a guard on the main breaker An energy source that can not be locked out must be guarded by a person It is now safe to enter the experiment bay unless there are energy storage devices in use 3 4 1a No energy storage or alternate sources present No large capacitors no other power connections 120 208 VAC etc 7 Latch doors open 8 If any adjustments are to be made to the 480VAC connections lock tag them out The guard on the main breaker may be relieved once this is done 3 4 1b Energy storage present Put on proper protective gear glasses flame retardant smock hard hat e Latch doors open e Disconnect alternate energy sources e Use a voltage level appropriate hot stick and voltage tester to check that storage devices are discharged 51 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION Install safety grounds Ground con
105. nsor can be used to verify the output is probably working Another example is typical of motor control centers the large power relay that connects a motor to a source through a main or M contact often has a mechanically integrated but electrically isolated auxiliary AX or just X contact which signals the state of the main contact back to the control system This geometric increase in I O complexity is one of the reasons mechanical linkages were not used in the high power lab Presently the lab s only moving output parts are cooling fans 20 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND and a few signal level relays The cooling fans incidentally have auxiliary motor contacts to indicate they are operating There are other considerations for I O devices in the High power lab The interface components must be modular so that failed or damaged components can be replaced easily A variety of I O loads must be supported including analog PWM high speed high voltage isolation The system must be expandable to accommodate future needs Wiring connections should be kept short especially in light of our expected high EMI conditions so the I O system should be distributed Devices themselves should be EMI tolerant In other words simple and metal bodied Contact free proximity sensors probably won t work Cost Some sensors will have to be shock vibration tolerant Generally High po
106. nt some practical useable principles for setting up a good operator interface without delving into the physiology and psychology more than necessary There will always be call for some discrete devices the big red E stop button for one thing fire alarms for another Discrete devices have some advantages such as fixed location and function Sometimes they are easier to use as when wearing gloves offer better performance video gamers use joysticks for a reason or have security features such as key locks Discrete devices do have drawbacks though Operator station I O has the same health checking requirements as field I O Also compared to graphical controls discrete devices are more expensive hard to reconfigure require more power and get covered up or pushed accidentally Graphical screens have the advantage of being infinitely configurable The same indication can be presented in many different ways regrouped with others to suit the current situation or even the operator s taste Changes can be made very quickly In modern systems An industrial quality illuminated pushbutton can cost 30 more five position joystick can be hundreds of dollars Interface hardware mountings and wiring add many hundreds more 5 The famous accident at the Three Mile Island power station was exacerbated by the fact that the operator failed to observe some of the critical warning indicators Reportedly he was so fat that
107. ons have to be available and changes between them need to be fairly simple The operating environment including operator controls equipment interlocks and data recording also needs to facilitate this by being flexible and easy to use There is a common sense and ethical if not legal requirement to avoid endangering people which is not casually accomplished with equipment of this scale This requirement is doubly important because this is an academic institution not an industrial setting with its attendant safety codes and inspections Operating procedures need to be developed to keep people out of harm s way while working in the lab The automatic portion of the lab control system must be as fail safe as possible in the face of both equipment faults and operator errors HIGH POWER LAB students report prospective employers react quite favorably to their experience with an industrial SCADA system I O hardware networks and PLC programming B P CHAPTER INTRODUCTION CLINTON PERDUE 1 3 The prior art other testing facilities Testing facilities can generally be broken into four categories depending on whether or not they have moving parts static dynamic and if they primarily dissipate or recirculate energy dissipative regenerative 4 Dynamic Static dissipative dissipative Dynamic regenerative Static regenerative Figure 1 1 Types of
108. out screen may have a simple listing of sensor states corresponding to items on a check list while a logic diagnostic screen would have a subset of the I O but add some internal logic states Given a system with dozens or hundreds of special purpose screens finding the one you need can be a challenge We are all familiar with menu and icon navigation good operator Controls on Unites States Navy vessels use one scheme controls at nuclear power stations have been standardized on the other The latter is often staffed by people who have been trained on the former and the potential for error is sobering 9 Colorblindness was discovered as a result of people misreading traffic signals This is why signals are now always red on top green on bottom 20 The author once observed as two operators using discrete local controls on widely separated pieces of equipment managed to push a twenty ton coil of steel over a ledge and drop it three stories It bounced 24 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND screens generally build on this philosophy Dedicated buttons serve as shortcuts to overviews of major equipment DC source for example or functional groupings cooling The depictions of devices themselves may also serve as icons which lead to successively more detailed sub screens dealing with a particular system you want to know more about the status of something Then click on its p
109. p The switchgear is composed of 16 bit registers therefore select 16 bit q Now you are ready to run the program Refer to Table 3 2 showing key registers in order to understand how to read the data coming from the registers B 3 1 3 Method 2 Reading serial source data via a buffer with a script In brief what GE Tech Support recommends is periodically transferring serial sourced data to from a buffer variable using an applications script which runs periodically in the background The ladder or other process control logic will use the current value of the buffer instead of waiting for an update We have tried this and it seems to work but we do not have a final version as of this revision The conversation with GE is below Q Scan time for the ladder program is hundreds of milliseconds why A The issue is the Ethernet to Modubs bridge The Modbus TCP IP communication is SYNCHRONOUS and is a request response protocol When you put serial on the I O system you make the scan times HUGE the logic scan waits for the message response CLP Use the View Modbus driver to read this data and pass it to the controller GE Fanuc Tech Support 90 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE Q Is the the View Modbus driver just another element of PC control that can turn on or do we need to run some other program view runtime on the same PC or another PC i e a viewnode The View modbus driver is pa
110. power electronics testing facilities 1 3 1 Static dissipative loads In the most straightforward class of systems power from the grid is processed in some way by the test equipment and then dissipated Reactive load elements may change the power factor as they temporarily store energy but eventually it all is dissipated This can be a very realistic setup for simulating heating lighting or some other dissipative load that has no mechanical dynamics The early stages of design for a dynamic load may also benefit from using a simple well behaved load because it will not complicate the analysis CPES has several large air and water cooled resistor banks for this purpose CLINTON PERDUE CHAPTER INTRODUCTION 1 3 2 Dynamic systems A system with an actual mechanical load is superior in some ways to one with electrical dummy loads for testing electronic power conversion equipment Motor drives for example must deal with mechanical realities such as rotor radial displacement vibration and acoustic noise There are integrated design tools which can model all these phenomena simultaneously but the problem tends to be exceedingly complicated Running a real machine to validate the model is still quite important This class of system also provides an opportunity for collaboration with mechanically minded colleagues because mechanical dynamics can be investigated just as easily as electrical ones A mixed facility
111. r functions conscious The prior example of the system heating limit is autonomic This is a built in protection set up by the system administrator and standing guard True the operator may tune the function by setting a limit but in extremis the function acts without operator intervention Alternately conscious controls would include user initiated actions These could be pushing a control button setting a numerical reference for a regulator or running a user authored program to take a sequence of such actions Common examples would be a schedule of as ae CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND references in support of performance and transient testing simulation of faults injection or suppression noise etc There will of course be a need to measure the results of all this At present Data Acquisition in the labs is handled on a small local scale by individual instruments scopes meters etc A coordinated integrated Data Acquisition system will allow data to be collected across the facility both by permanent instruments and by ones added on for a particular project Some data will be collected continuously other only when triggered Ideally everything will be assimilated into a coherent collection to allow for a holistic analysis A major use for this sort of data collection is forensic fault analysis Often finding the root cause of a fault involves noting some suppo
112. rch the publishers of MATLAB have placed a full page advertisement in IEEE Spectrum usually on the back cover touting model based design nearly continuously for years IEEE Spectrum Magazine IEEE On rear cover Volume 43 Issue 8 August 2006 The Power Systems Toolbox in MATLAB is evidence of this It was written by Hydro Quebec in the course of their work CLINTON PERDUE CHAPTER INTRODUCTION The HIGH POWER LAB will also serve the more general educational purpose of allowing students to gain practical experience with a real world system They will encounter practicalities they may not have otherwise considered For example safety precautions and oversight become serious when operating in a realm where the safety of people and facilities can not be taken for granted Finally it is hoped that the presence of large power equipment will make an impression on students who might otherwise have no concept of the scale of these systems inhabit 1 2 Requirements To be useful this facility must provide a power distribution system comparable to the environment where the other technologies developed at CPES are intended to operate This calls for distribution level 24004160 AC voltages DC links at 3 3 kV or greater and at least 100 kVA continuous power The facility must also be able to host multiple projects simultaneously These projects do not necessarily need to be powered at once but a variety of connecti
113. rectly or go through the HMI which does the bookkeeping for us Square D recommends that the power supply for the communications be independent of the controlled line voltage to avoid ground loops presumably Further instructions including wiring diagrams and data addressing 29 30 31 can be found in the manufacturer s datasheets and in the Users Manual Appendix B Modbus is a daisy chain serial network so adding drops is straightforward provided the total number of drops nor the total length of cable is excessive Termination resistors are required at both ends The switchgear also has hardwired connections to E stop and to safety I O which consists of interlocks to loads typically permission to energize signals Figure 3 5 2 Power Zone4 Instruction Bulletin Bulletin No 80298 002 05 December 2003 1999 2003 Schneider Electric 8821 Garners Ferry Road Columbia SC 29209 USA 1 888 SquareD 1 888 778 2733 www us SquareD com Power Zone4 Catalog 05 2005 Schneider Electric Ibid 3 Instruction Bulletin MICROLOGIC 5 0P and 6 0P Electronic Trip Units v 7 522 Bulletin No 48049 137 02 06 01 2000 2001 Schneider Electric Square D Company PO Box 3069 3700 Sixth St SW Cedar Rapids IA 52406 3069 USA 1 888 SquareD 1 888 778 2733 www SquareD com There are four wires and a shield ground which are color coded for identification Connect a new device s wires in series with existing wiring If adding
114. rent in the on state These characteristics again form bounds to certify that the sensor is probably functional Making these health checks can be quite a problem requiring sensors to check the sensors that watch the sensors ad infinitum Commercial I O systems typically have an interface module that monitors whatever field device is attached to it and reports data and status to the next higher level of control That level then checks whether the interface module is acting as it should This secondary status is added to the package and the whole lot is reported further upstream a process that repeats with each device that handles the data until it is acted on Often the quality of the measurement is just as important as the data returned If a signal related to some important state is unhealthy then it is necessary to assume the worst from a control perspective The control system must compensate for this blind spot by making decisions based on other data operating in a modified mode so that the missing data is not crucial or shutting down entirely until things are fixed never a popular option Obviously some things are not worth the trouble to check it depends on what is at stake For safety 419 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND critical signals this entire structure is eschewed in favor of hardwired mechanical interlocks because their state transitions are more simple and predi
115. rol and data collection so useful work can be done As is evident from the physical layout and electrical distribution topology presented in the previous chapter doing so will require simultaneous operation and monitoring of widely distributed equipment eld CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND From these general considerations we derive the following list of specifications for the control system It must e Communicate data and commands from remote dangerous locations to a safe centralized one e Guide the operators in taking correct actions and react automatically to dangerous situations e Facilitate data collection Additionally by virtue of being a university facility the control system will have to be reconfigured periodically to host new experimental equipment The system must be flexible in that e It must be scalable and modular able to accommodate new needs e It must depend as little as possible on components that will become obsolete and hard to maintain e Interfaces must be open and intuitive to minimize the user learning curve System security is also an issue e Supervisory constraints and system interlocks must be reasonably robust and readily verifiable e Communications must be private and secure 2 2 Proposed solutions To address the above specifications the author proposes the following e Usea system of distributed I O devices connected by a digital data n
116. roperly and that logic functions correctly 4 2 1 Variable assignments The correct logical meaning must be assigned to each received signal As built wiring elementaries are a crucial check at this stage showing what is really connected to each I O point 588 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION The true test though comes in watching a live HMI view as devices are actuated one by one Figure 4 3 below is a screen capture of the HMI online monitor showing status of a group of inputs DURIP State Control_Debugger CIMPLICITY Machine Edition Control 1 0 BB Edit Search Project Target Variables Tools Window Help m x Em uz 4 2 2 S gt 2 Ps g gt x mis mee x ab gt 2 m J 2d xd Driver Node Address Value n a Y 2 Ethemet 1 0 Driver 08 1 161101 Node 1 VersaMax I O IP 192 168 1 111 EGD 9 RootAlarmGroup B Slot 1 C 200 MDD 841 Word Alarms 11 HVSwtch1 1X1 192 168 1 111 1 1 On Bit Alarms 12 HVSwitch2 1X1 192 168 1 111 1 2 OR JJ Application Scripts 13 HVSwitch3 24 1 192 1681 111 13 On VERET M HVSwtch4 4 1 192 168 1 110 14 On 28 1 0 Drivers 15 Plug_4 41X1 192 168 1 111 1 5 Ethernet 1 0 I6 Plug 5 1X1 192 168 1 111 1 6 eu Watch Lists 17 Plug 1L 3 4 24 1 192 168 1 111 1 7 On B 8 Global Functi 18
117. rt of the Control IO setup that you have configured It would be running with the PC Control side of the software GE Fanuc Tech Support Q Ok we are getting somewhere We found the help for setting up PLC access with a Modicon TCP IP driver and think it is configured with the proper modbus drop and address We also have a signal which was previously coming from a generic Ethernet driver under control I O drivers associated with the new driver However we get Error 324 at build time it seems that the ladder logic can not receive data directly from a PLC source Question then how can we get the PLC source data into the ladder program gather that may have to set up a script running in the background or set internal values as an action on a screen object A You are not able to use a variable connected to the PLC Access Variable in the LDPC portion of the product You are correct about being able to use a script to bring the PLC Access variables into the LDPC logic You can create a application script that will periodically copy the PLC Access variables into an internal variable You will then be able to use these internal variables in your ladder logic GE Fanuc Tech Support Q About the script OK that is what thought Now how do do it Can you please send me an example script and tell me how to get it to run It would be best if it did not depend on a particular screen or object being visible A The best way to do th
118. s are compared in Table 1 above In some situations there may be other factors to consider but given these a static recirculating system is the best value 1 4 The VT HIGH POWER LAB available facilities With the above analysis in mind it is not surprising that CPES has opted for a static recirculating system when laying out the high power lab The line side power distribution topology is shown in Figure 1 4 CLINTON PERDUE CHAPTER INTRODUCTION 4 Cables Phase 8KV 500MCM Shield Grounded Both Ends Substation Transformer U 480V 1200 L 0 2 mH amp up 4 Cables Phase 8KV 500MCM Shield Grounded Both Ends U 0 48 KV 1 1200A Us 1 2 2 08 2 4 3 17 4 16 kV 480 278 240 139 139 A S 1 MVA X 7 5 1 Cable Phase 8KV 500MCM Shield Grounded Both Ends 1 Cable Phase 8KV 500MCM Shield Grounded Both Ends 139A Second Floor Switchgear _ Cabinet Extension High Power Lab 480VAC I om Line Distribution Panel 1 Cable Phase 600V 500MCM Unshielded 1600 First Floor Re circulating Power Switchgear Equipment Room 4 Cables Phase 8KV 500MCM Shield Grounded Both Ends U 480V 12 1200 A L 0 2 mH amp up 4 Cables Phase 8KV 500MCM Shield Grounded Both Ends 0 48 KV 1200 A Us 1 2 2 08 2 4 3 17 4 16 kV l2 480 278 240 139 139 A S 1 MVA X 7 5 1 Cable Phase 8KV 5
119. sedly unrelated event that occurred such as a reference change to another piece of equipment just before the device in question changed state in some way For this to work the collection needs to be long term continuous broad in scope and reasonably frequent 2 4 Elements of a SCADA system Physically implementing a SCADA system requires sensors and actuators an operator interface communications logic and data processing functions Figure 2 3 Elements of a SCADA system 18 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND A control system of description is of little use without good feedback Achieving this requires proper selection and placement of devices with due concern for their environment One must consider the range of temperature shock and vibration voltage radiated EMI contamination and other abuses a sensor will be exposed to Often a sensor can give data about its health in addition to what it is supposed to be measuring For example an analog transducer may return a 4 to 20 mA signal to encode for a range of detected conditions i e levels temperatures positions etc If such a signal ever reads outside the 4 20 mA range then the wiring may be broken shorted or the sensor itself damaged Discrete valued solid state sensors can also be interrogated for health They usually have a minimum leakage current even in the off state and a maximum cur
120. serial links and special purpose wiring complete the system communications topology Figure 3 7 This is a maximum case build out showing all likely connections With this scheme the facility is both observable either directly or in analog and controllable Communication delays are quantifiable and data integrity can be verified It remains to good practice to control the system well but the necessary resources are available 49 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION saa Soe Switchgear Safety 1 0 EGD Bridge E stop doors cap Modbus IP bank transformers E uide I S cooling connections Hardwired U T cup I Test equipment T U Hardwired modules Data loggers scopes and custom sensors Hardwired custom serial and ethernet Custom control and data links wiring I I I GE Serial i TEES Lab Public link e EGD Ethernet Ethernet PC running HMI viewer 1 PC with GE drive HMI sapien HMI viewer 1 pod log configuration tools soft PLC HMI viewer 2 Wiring Key Hardwired Custom wiring Ethernet Serial network Figure 3 7 Complete communications topology 3 4 Procedures Industrial settings commonly allow for equipment to be energized automatically but there are several reasons why this is not wise in an academic laboratory setting Chief among these is that industrial systems
121. side the lab space which could reasonably be the result of a manual operation latches the occupied state true until a rising edge of the door closed indication occurs For Power On The power is assumed to be On unless there is a positive indication from a disconnect device for each circuit feeding the lab area that every circuit is open Any equipment in the lab capable of storing a dangerous amount of energy must be included in this decision For Control Healthy A positive Control Healthy output must be continuously asserted in order for equipment to be energized The control is assumed to be unhealthy unless there are good communications with all control elements and those elements all report that satisfactory conditions exist Thermal faults will include a lag appropriate to the thermal time constant of the equipment and be conditioned by which equipment is in operation 238 CLINTON PERDUE CHAPTER 3 DESIGN AND IMPLEMENTATION MAJOR STATES TRANSITIONS AND INDICATIONS SYSTEM STATES MANUALLY SWITCH ON Power on off Room vacant occupied Control healthy fault RESOLVE FAULT POWER OFF POWER OFF CONTROL OVERAGE OR FAILURE Color codes for warning lamp Red KEEP OUT Yellow Caution Power commanded off but check Green Power sources credibly ACCESS VIOLATION P gt R report open CONTROL OVERAGE OR FAILURE
122. structor David Ryan General Electric Corporate Training Center Salem Virginia during a lecture on the DC2000 motor drive control system August 1993 A time power heating curve is a simple model of the thermal mass in a machine integrating power lost in the machine minus cooling over time to yield current temperature It tends to take a saw tooth form with a fast rise and slow decay 16 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND Time Power heating curve with limits Modeled Power into the temperature system 2 E Figure 2 2 Example of a time power heating curve used for control In terms of Control the SCADA system is the operators primary means of interaction with the equipment The system provides a display typically on a PC monitor of plant status feedbacks and measurements and allows control of operational set points for all the various pieces of equipment The information on display at an operator station is collected by sensors scattered throughout the installation and likewise operator control inputs are distributed to the remote equipment These remote devices can be miles from the control station Operator interface design principles will be discussed in depth in section 2 4 2 The domain of system automation spans both supervisory and control functions By analogy the supervisory tasks could be termed autonomic and the othe
123. system can operate as is The lab administrator will enforce the safety procedures and checklists and can explain the default operator environment zs CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE In many cases users will wish to add custom interlocks to the system The system has provisions for this and we will describe them here The control also supports custom user programming for referencing data collection and other tasks as desired B 2 Communications The following section details how the various elements of the lab communicate Figure B 1 shows topology including Ethernet serial networks and hardwired discrete signals Current High Power Lab Communications Topology Serial Modbus Square D switchgear Safety 1 0 A E stop doors cap Modbus IP bank transformers cooling connections 4p Hardwired Equipment Under Test equipment Data loggers scopes and custom sensors ee Hardwired custom serial and ethernet Custom control and data links wiring Test Modbus and Lab Public Ethernet Ethernet HMI viewer 1 Tethys Wiring Key Second floor switchgear Incoming V 1 Q monitor Line side protection and control Hardwired Ethernet Custom wiring Serial network Other users Figure B 1 Communications topology Master computer Carme is in red
124. t has occurred somewhere upstream of the current sensing such as in the source transformer secondary or in the cabling from transformer secondary to source converter In similar fashion in conjunction with the same ground resistor the summation of the thr phas load converter line currents is compared to a preset threshold which when exceeded annunciates a ground fault signal upstream of the current sensors that is in the load transformer secondary or cabling from transformer secondary to load converter Fig 6 illustrates the power hardware used in the implementation of the ground fault detector in GE s 6600 11 rms IGCT voltage source power electronics equipment 108 CLINTON PERDUE APPENDIX HIGH VOLTAGE GROUNDING AND SUBSTATION DESIGN CAPACTIORS FF rom tetanus EJ ms uw 57 me lz J a sarta Urn Em amm NE Tl i 1 Fig 6 Illustration of ground fault detection hardware used in GE voltage source power electronics equipment This ground fault detection scheme exists in the 2300 IGBT source converter that GE donated to Va Tech 1MVA Transformer Discussion For flexibility the transformer secondary is reconnectable to the following options Vll rms Rated RMS Current MVA available in this 40 deg C ambient connection 1200 480 8 1 2078 277 6 1 2400 240 4 1 3174 138 8 2 76 4160 138 8 1
125. tes the number of modules In our case since it s the switchgear the slot is used to set a base to start reading the registers and we only want one base so enter 1 g Message time out is explained later in the Operating Cimplicity section for now just input 1000 h Reconnect time out is explained later in the operating simplicity section for now just input 2000 i Time between reconnect is explained later in the operating simplicity section for now just input 1000 Input 45 for unit ID This is how the switchgear has just been configured Unit 45 represents breaker A of the switchgear and unit 47 represents breaker B 89 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE k Enter 678 for input register offset Register 679 is the one to be read since it stores the date and time information See note 2 above to know why we enter the register offset to be 678 instead of 679 1 Leave output register offset to be zero because we are reading data not writing m The slot now has to be configured Double click on it Note that the offset on the slot will seem to be off but note 1 above explains n Atthe properties make sure it s on analog input o Click on the tab for module properties Number of terminals corresponds to the number of registers 4 in this case for the date and time Since the last register is showing milliseconds and that is in no ones particular interest just enter 3
126. the physical switches Each NO or NC contact changes states depending on the variable assigned to it Variables in Cimplicity can take one of two states although the actual labeling of the states may vary For instance the following values you may choose for a Boolean variable are all the same but with different names on off 1 0 true false I O Open Closed Regardless of which designation you chose the NO or NC contact changes state so long as the variable associated with it changes to true 1 I on or open as if the relays had power applied to them Traditionally at the end of each rung an output relay is placed so that whatever value the signal takes true on 1 etc whatever form you choose may be outputted to another variable You may place as many output relays in any location as you see fit however You may place instructions in parallel by click dragging an empty segment of rung and pulling the dotted line that comes from it across to the other side of one or more instructions AND ing and OR ing Since ladder logic operates like a circuit you can put two NO or NC contacts in series or in parallel If both contacts become closed NO contact variable 1 NC contact variable 0 then the signal is allowed to pass through The AND is written by having two NO contacts A NAND is written by putting two NC contacts in series By placing two NO contacts in parallel you ll write an OR B 4 3 1 Adding instru
127. to this is if one leg of the 4160 is in use and someone closes the switch for that same leg in the other bay The only prevention for this is an administrative padlock on the switch again a procedural point The final problem is that all the incoming circuit breakers require energy to open Control power is typically from a separate supply meaning that it is vulnerable to interruption while the main power remains on regardless of what the control system commands There is not much that can be done about this now Adding a UPS would help but is still not a guarantee 64 CLINTON PERDUE CHAPTER 5 RESULTS CONCLUSIONS AND GUIDANCE These issues are all serious but not insurmountable They require more care in drafting operating procedures including more exhaustive pre energization checks and more attention from the lab administrator Attention to safety procedures and a healthy paranoia are highly recommended just because something is supposed to be off is no proof that it is 5 2 Successes The preceding thoughts aside the construction project has been quite successful We have identified the requirements of the users and the capabilities of the equipment The communications I O and commercial SCADA package we have implemented give us the capability to meet those needs and make the most of our resources We have demonstrated automation functions that enforce operations interlocks that adapt to the equipment in use
128. units unless specifically stated otherwise 86 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE Substation Transformer Second Floor Switchgear Cabinet Extension High Power Lab 480VAC Distribution Panel U 480V 1 1200 A L 0 2 mH amp up U 0 48KV 02 1 2 2 08 2 4 3 17 4 16 KV h 1200A 12 480 278 240 139 139 A 5 1 MVA X 7 5 U 480V 1 1600 First Floor Re circulating Power Switchgear U 480V 121200 L 0 2 mH amp up Equipment Room 139A U 0 48 KV Us 1 2 2 08 2 4 3 17 4 16 KV 1200 1 480 278 240 139 139 5 1 7 5 Innovation Drive ac 2 3 KV c 500 A High Power Lab 1 Room 161B MV DC bus U 1 0 5 0 KV 1 1000 200A C 2000 uF L 8mH ri 1 Experimental Device Being Tested Figure B 4 power distribution elementary B 3 1 Reading Switchgear Register Information Each trip unit has an internal controlling computer that keeps vital information about the status of the system and the power passing through it Getting to this data is a bit of a basic 87 CLINTON PERDUE APPENDIX B VT HIGH POWER LAB USER S GUIDE programming exercise requiring reading and interpreting specific registers Addresses and further instructions can be found in MICROLOGIC 5 0P and 6 0P Electronic Trip Units Instruction Bulletin 48049 137 02 Append
129. vious first cut in this system is the dividing line between hardware and software We begin then with how we tested the hardware 4 1 Hardware commissioning This is not complicated but does require strict attention to detail Wires have to go to the correct terminals jumpers and switches must likewise be correct proper network terminations are required etc Wiring elementary diagrams produced during the project planning stage are a guide to where connections should be made These must be updated to as built status to reflect changes made during construction and must show every single wire connection point to be any use in troubleshooting later problems The hardware has its own division point at the network level AII the field devices the actual sensors and actuators are on one side and the network connectivity is on the other 4 1 1 I O field device to module Testing I O devices is usually just an exercise in applied linear networks Signals have to start somewhere trace the power from source through the device and back to ground Observe that the correct things happen along the way Success is unambiguous Actuating a sensor makes an indicator light up on an input module Figure 4 1 or a working output makes something go klunk 54 CLINTON PERDUE CHAPTER 4 TESTING AND VERIFICATION Status indicators for three inputs Figure 4 1 I O module and schematic showing three
130. wer lab devices are not required to be explosion proof nor concerned with dust moisture temperature extremes caustics or radiation Some care is required in designing mounting enclosures and cabling Everything should be shielded Cable runs need to be easily identified Anything related to safety interlocks needs to be physically secure Enclosure boxes need to be lockable Dr Porche when designing the Volkswagen Beetle was quoted as saying part that does not exist can not fail to explain his draconian simplification of the design We hope he would approve of the high power lab This is actually a calculated risk on the part of the drive manufacturer What they really know is that the motor has been selected on not that it is actually running P Explosion proof indicates that a device will not be a source of ignition in volatile environments 21 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND Operator stations where one finds all the buttons switches dials gauges and indicator lights required to run the equipment Modern ones include CCTV and computer terminals usually with a graphical control interface Things are always evolving in this area sometimes due to changes in common skills and preferences among users Other changes are driven by research in human factors engineering and the ergonomics of operator interface design At this point we will prese
131. y The motor relay close command energizes a solenoid that pushes the contacts together With power removed an opposing spring opens the contacts Hence this relay is a fail safe device The main motor contact does open after some hundreds of milliseconds and its auxiliary instrumentation contact closes simultaneously This auxiliary contact is wired to a control I O point informing the control system of the motor relay s state This is a redundant fail safe status signal The motor relay is undersized and unable to interrupt the full short circuit current When it opens an arc is formed maintaining a closed circuit This happens to some degree even in properly sized equipment eroding contact surfaces before the arc is snuffed Consequently regular inspection and maintenance of power contactors is crucial At this point the SCADA logic could note that the drive is still reporting a non zero output current over the control network yet the motor auxiliary contact is closed This is not normal The control could then decide to signal the switchgear to open the utility feed breaker by removing the energize command Without the feedback from the motor auxiliary contact 235 CLINTON PERDUE CHAPTER 2 CONTROL SYSTEM REQUIREMENTS AND BACKGROUND the SCADA would have to wait for some other fault indication from the drive Such an indication may come first anyway another good case of redundancy The switchgear logic

Combined_ETD_reviewd.11-01g

Contents

Download Pdf Manuals

Related Search

Related Contents