Home

Industrial Electrical Engineering and Automation

1. Vacuum machine tank bar x 10 3 Turbine brake temp C Ti white 48 T2red 48 50 48 M ARE a il 44 42 40 Temp C T 466 T2 41 3 Kaeser 39 5 Water 180 Mic 40 30 2 10 PI101 80 pr102 7 3 P1103 21 PI104 0 24 10 F E pi SS CV 103 open 101 1104 Impure gas storage pressure bar 153 6 Temperature purifier K 11102 127 T1104 83 3 1704 160 150 140 Purifier pressure bar P1150 2 37 P1151 12 22 MO lE Differential pressure purifier bar 0 15 0 25 5 02 0 15 SS M M Volume machine tank LHe liter 270 CV111 open 26 40 ure K i Ww Www T 3m z E zj h HE 2005 5 150 ad i i stop 3 gt um 25 100 S88 EEEES 10 sli mM Figure 3 5 Front Panel of plotter 12 Chapter 3 Method and implementation program description 3 5 4 Communication and libraries The communication between the control program control run on cRIO and UI is done via shared variables The library networked variables is available to any device on the local network and UI reads specific variables and displays the data to the user UI can also write to some variables in the library which are used to provide instructions to the control program The library single process variables contains another set of variables with similar names but is only available to VIs on cRIO This is where the control program reads instructions from and writes i
2. Virtual Instrument A LabVIEW program or sub program consisting of a Front Panel with inputs and outputs and a Block Diagram with logic 49
3. The definition and graphical representation of the inputs and outputs of a VI in LabVIEW The inputs and outputs are represented by buttons meters etc as in a physical operator panel Division of Industrial Electrical Engineering and Automation at LTH Describes a process where the system has constant enthalpy Describes a process where the system has constant entropy Describes a process where the system has constant pressure Describes a process where the system has constant volume Describes a process where the system has constant temperature The programming environment used to make programs for NI controllers Liquid Nitrogen Cooling Liquid nitrogen is used to help cool the helium gas The nitrogen cools some of the first heat exchangers the helium gas passes after compression This increases the production rate 48 liquefaction LTH NI TCF 20 triple point VI Appendix F Glossary The process where a solid or gas is converted to liquid form In this report exclusively the conversion of gases to liquid by cooling The Faculty of Engineering at Lund University Lunds Tekniska H gskola National Instruments An American corporation founded in 1976 Currently active in many fields including electronics The original machine or plant in this report referred to as the Liquefier and the Purifier The temperature and pressure for a substance where the three different phases solid liquid and gas coexist
4. The new system will be implemented on CompactRIO 9022 using LabVIEW Most of the old control system based on analogue circuit boards and the entire PLC system are to be replaced The new system should mimic the behaviour of the old to the extent possible The replacement must be done such that the normal production is not interrupted 1 3 Outline of the project The first step of the project is to gather information about the current system Available documentation includes primarily the manual of the plant but also short film sequences and notes by the Kryolab staff The work is to be carried out at Kryolab so help and guidance regarding the functionality of the plant will be available To be able to study the behaviour of the current control program and the general process behaviour programs for basic plotting and logging should be implemented The gathered information will be used to build a basic control program with a possibility to provide manually controlled input data and to force a desired state of the logic This is needed to test the behaviour of the control program before it is connected to the real process Control will gradually be moved over to the new system after testing For this reason it must be possible to turn on and off outputs separately in the control program To be able to control as many parameters in the control program as possible during the testing and development the cRIO can publish a graphical interface directly conn
5. Appendix E presents not only features of the system but also guidelines for the work This chapter is mainly about the features and behaviour of the system and how it was tested Typically each new control signal was verified first with a set of different manual inputs then on real data comparing outputs from the two systems and lastly with connection to the real process All these steps have taken some time and although the final tuning with connection to the real system was more time consuming the actual work involved has been as much in the earlier tests The examples of testing presented in this report are all from the phase when there was a connection to the real process This is because it is more interesting when it can be shown how the process actually reacts and not only what the control system is attempting to do error out status code LLIN lower limit Purifier lower limit fo J32 95 source LIN higher limit Purifier higher limit a 975 gs P Pmach Cooling water Pbuff Vmach Pstorage SIMULATED MEASUREMENT SIGNALS go 9 go 2800 gd REAL MEASUREMENT SIGNALS Tl Speed trip Alarm T1 close CV103 T1 activate stop PI buffer TS109 Vacuum CV103 4800 o 4825 Jasso PI buffer TS109 Vacuum Cv103 i D r r y y y Js 5 J m A0 T2 Speed trip Alarm T2 close CV103 T2 activate sto 5 71087 fas 1313 13 32793 100 545 PI machine Ts110 vin nm A ALT e EA T5110 cni 103 ref fo go do Jio etivate TurbinActivste Alam Actvate V Activate Compr
6. Lunds Universitet Ett nytt styrsystem f r en anl ggning f r forvatskning av helium har utvecklats I anl ggningen renas heliumgas f r att sedan kylas i flera steg till dess att den kondenserar Den befintliga anl ggningen r mycket lik de som byggs idag men systemet f r styrning och vervakning r f r ldrat och Kryolab nskade uppdatera detta Det nya systemet r programmerat i LabVIEW och k rs p en CompactRIO fr n National Instruments Det ers tter det ursprungliga systemet best ende av analoga kort och ett kompletterande mjukvarubaserat system I arbetets f rsta fas har maskinens beteende studerats genom loggning och plottning ven befintlig litteratur i form av manualer ritningar och sparad data fr n tidigare k rningar har utnyttjats Koden har utvecklats iterativt och varje del av koden har testats innan den anv nts p den verkliga anl ggningen F r att verifiera programmets beteende i olika situationer har ett anv ndargr nssnitt med m jlighet att manuellt ange sensorv rden utvecklats Nya funktioner har testats och lagts till successivt Det nya systemet presenterar betydligt mer data n tidigare och ger en helt annan verblick ver processen Det nya systemet erbjuder ocks kade m jligheter att manuellt justera vissa ventiler och att f r ndra gr nsv rden och b rv rden som anv nds i regleringen Ett nytt alarmsystem har implementerats och start och stoppsekvenserna har ytterligare automatiserats
7. Provides buttons to start and Liquefier Purifier and stop Liquefier Purifier and LIN cooling LIN coaling Evaluates the state depending Provides buttons that can be on the collected data from the used to manually control some e system valves single process Variables on Sets the outputs of the cRIO Library to control the process Used for internal Executes orders from Ul communication an cRIO plotter vi Displays data collected from the system in diagram form for CommunicationHighPriority Overview subi Facilitates the comparison of Only communication regarding behaviour fram one cycle ta target e g if all execution on the next networked variables cRIO should be aborted Gives an indication of a Library current trend e g in production Used for communication rate aver the network MI p CommunicationLowPriority i subl egger Handles most communication I between control and UI Collects data and saves it in a compact format shared variables Makes it possible to go back longer in time to find when a problem started Makes it possible to compare the machines behaviour to narmal values previously saved Autagenerated library Variables for all ICs on CRIO Variables can only be read Figure 3 1 Sketch of the program structure 3 5 1 Ul The F
8. aee Resta tagnesects oe V ea aea TAa EENE E 20 3 8 Alarms and automatic shutdown sssresrserserrerreesrerrrsrerrrsresrrsrrsrrsrrssrsrrsrrn enne en nne 21 35 9 Transformatoren i eer ee eee aeterne nai nade i ee 21 4 Testing and evalua on ue epit te ALE eed eta see 22 4 Examples Of tes ng eere e EO RE erre eee ebsites 23 4 1 1 Improvement CV 111 before and after 23 4 1 2 Varying conditions start up from cold and warm sess 24 2 Results itio sU e ed S eie reise lebe eese cedtele ee lu e Dee be bees 27 6 Discussion and conclusions nissun iniaiaiai a i ea iia 29 T sBUUUre work erenn 31 References M 33 Appendix A Kryolab eet eei he eese eese deett 34 Appendix B Helm oie Ret eed tete eet e bees eie Pee eee HE tea 35 Appendix C Cryogenics and refrigeration cycles esee 36 Appendix D Helium liquefaction plant esee 40 Appendix E System sp cification erer menai eii e o aaae nennen nennen nennen s nk f sta 45 Appendix FE Glossary oen e e RE bed dice He eut eet ine tete eere 48 IV Chapter 1 Introduction 1 Introduction 1 1 Background This master thesis in Industrial Electrical Engineering and Automation forms the last part of a degree in Mechanical Engineering with a specialisation towards Mechatronics at LTH the Faculty of Engineering at Lund University It has been carried out at Kryolab
9. and by graphs The plots make it easier to see patterns and deviances early The operator has information about all output values valve positions and the current state of the logic The user interface also provides increased possibilities for the operator to control the process and affect logic and set points in a way that was previously not possible The new automatic start and stop sequences and the remote control are working well No problems have been detected with cRIO and the control of the plant List of requirements Restructuring and refining the original list of requirements a revised list was created during the analysis phase and it has been used for system evaluation and acceptance Both the original requirements and the revised list can be found in Appendix E Almost all of the 42 criteria listed have been fulfilled These requirements are not only focused on functionality but also in some cases instructions on how things should be done In some cases things have been done differently than originally intended and in a couple of cases 12 17 this leaves room for further improvement In three cases the criteria have not been met because of lower priority and lack of time 16 20 41 All requirements that have been met differently than planned or that have not been fulfilled because of time constraints are specifically commented in the following sections with their reference number Overall limitations The integration of the new contr
10. eee peepee eens P rattar ra qoa Y 30843743308 arg 11 36 37 13 00 00 14 00 37 14 33 00 16 00 00 18 32 20 Time Time RH AIM turbine temperatures BB 79 Turbine temperatures g ete 5 50 5 50 5 40 5 40 5 p 30 20 20 S recs TRC AEE EE RE A A ARA T a ROSOR 11 36 37 13 00 00 14 00 37 14 33 00 16 00 00 17 00 00 18 32 20 Time im m CV 103 HB d cv 103 HB 110 75 s 5 50 a a 5 S764 Oe IP EEEUENRUEEONNEORORURUN EEA D E g D 75133 Tana un LL LLL aN 11 36 37 13 00 00 14 00 37 14 33 00 16 00 00 17 00 00 18 32 20 Time Time Figure 4 3 Measurements during the start sequence of Turbine speed and temperatures and the opening of the inlet valvet to the turbines CV103 Plots to the left show the behaviour with warm gas in the system and plots to the right colder gas in the system with additional use of Liquid Nitrogen Cooling In the case with warmer gas in the system the first turbine white does not reach its full speed while the control valve CV103 is being opened With colder gas and LIN cooling both turbines reach their full speed during the opening of CV103 To allow time for the turbines to warm up the opening is paused when the turbines are allowed to reach optimum speed There is also a strange shape of the curve for CV103 between 15 10 and 15 25 this is to slow down the increase in pressure and consequently speed and is necessary because of the LIN cooling During the testing of
11. fact the same as helium 4 nuclei The use of helium for buoyancy in air ships is no longer a large market but helium is used for many other purposes such as helium neon lasers dilution of oxygen for deep sea diving pressurization of rocket fuel in balloons as a protective gas in arc welding and silicon wafer manufacturing As no other coolant can be used to reach as low temperatures as helium it is also to a great extent used for cooling purposes in high energy accelerators MRI scanners and all sorts of cryogenic applications 4 5 As helium is a natural gas by product the extraction of natural gas greatly influences the amount of available helium But the helium content also varies greatly between different geographical regions and not all natural gas fields produce helium Most of the helium produced today is produced in the United States where gas fields have been found with helium contents ranging up to 7 In general concentrations above 0 3 are of commercial interest 6 The Federal Helium Program was initiated in 1925 and was supposed to ensure supplies of helium for defence research and medical purposes The military particularly valued it for its use in buoyant aircraft Over the years a large amount of helium was stored in the Federal Helium Reserve In 1996 it was decided in the Helium Privatisation Act that the helium should be sold off to repay the debts of the program as federal use of helium was low The past five years approxi
12. form This is very helpful to view trends and detect similarities in and deviances from the normal cycle pattern All measurement data can be stored in log files This can be used to see patterns over longer periods of time or plot a specific measurement value retrospectively Alarms and warnings Threshold values for warnings and emergency stop have been identified In many cases a possibility to change them at runtime has also been implemented Suitable stop sequences have been identified implemented and tested for all possible states as well as a fast shutdown for critical situations The external emergency shutdown of the power to the compressors is still in use Normally the routine for handling of speeding turbines or the fast shutdown is sufficient but this should be kept as an extra safety precaution The vast possibilities to influence the control means that it is possible to e g set a threshold value for controlled shutdown such that it is not triggered in time This external emergency cut of power is a safety precaution in case of an incorrect behaviour of the control whether introduced by an operator or hardware malfunction Remote operator control It was planned to used existing features in LabVIEW to expose the user interface over internet 40 42 This has been used and it worked very well to run the plant from home At this point remote desktop is used to fulfil the same need as the current LabVIEW options for password protection
13. handling and the storage is described in Section D 3 How the Purifier works is described in D 4 and the Liquefier and surrounding units in D 2 Buffer tank E Figure D 1 Overview of the plant Liquefier Machine tank D 2 Liquefier Figure D 2 shows a simplified view of the piping of the Purifier and the Liquefier This part of the plant is located inside or close to the cold box The purpose of the cold box is to isolate the piping from the surrounding air The vacuum inside the cold box is maintained by a pump to minimise the heat transfer between the piping and the surroundings Inside the cold box a series of heat exchangers cools down the pressurised gas They are depicted as simple boxes in Figure D 2 If liquid nitrogen pre cooling is used this is done in one of the first heat exchangers Apart from the possible LIN cooling the refrigeration required is provided by the turbines acting on a side stream of gas which is returned to the low pressure stream of the cold box after expansion A brake is applied to the turbines by two gas compressors mounted on the respective shafts The part of the helium which goes through all the heat exchangers is approximately 7 K when it reaches the J T valve CV111 As it is expanded in the valve to 1 3 bar it becomes a mixture of liquid and gas at 4 2 K The liquid stays in the machine tank and the gas goes back to the low pressure stream of the cold box The low pressure return stream is
14. is blown to atmosphere In the second step it is cooled to approximately 30 K causing any remaining impurities to freeze on the finned tubes of the heat exchanger After expansion through valve CV162 the gas is joined by a stream of cool helium provided by the Liquefier through CV175 It is the stream through CV175 that cools the Purifier The cold helium is used in the heat exchangers to cool the incoming dirty gas before it enters the low pressure line via valve MV 169 Low pressure High pressure Gas from storage Turbines Pure He from high pressure side ofliquefier gt D lt MV173 CV162 CV175 Machine tank Figure D 5 Gas flow through the Purifier when gas from the storage is being purified When the Purifier is running it adds more gas to the system than the Liquefier can liquefy causing the buffer tank to fill When the Purifier is blocked with solid impurities it needs to be warmed up purged and then cooled down again During this time the pressure in the buffer tank will decrease To warm up the Purifier warm helium gas from the high pressure stream of the Liquefier is used see Figure D 6 It enters through valve MV172 and returns to the low pressure side via valve MV174 This will cause solid impurities to melt and vaporise Before the Purifier is cooled again with gas from the Liquefier through CV175 the remaining impurities are blown to the balloon gas bag with help of gas from the high pressure side see Figur
15. is an existing system that turns of the power to the compressors when the turbines are speeding it is important that the turbines are not allowed to speed up so much that they reach that limit Not only the values but also the actions needed to reduce unwanted behaviour must be identified 3 1 6 Remote operator control This interface must be made available over the Internet There is built in functionality in LabVIEW for this Any Front Panel can be published by the host computer to be able to use the panel it is necessary that the connecting unit has LabVIEW Runtime Engine Web Browser Plugin installed The panel can also be presented simply as a picture in which case the plugin is not needed 3 2 Choice of state machines Going through the documentation for TCF 20 it became evident that although the regulation of some valves only depended on specific measurement values most of the control actions where influenced by the state of the Purifier The addition of a control structure to automatically open and close the valves that where previously manually controlled and related to starting and stopping the machine meant that almost all control actions where now placed in a context and related to the state of the machine Additionally these states followed each other in a fixed order with few exceptions It seemed reasonable to place the logic for the control actions in a state machine or other similar structure In some cases this would not have be
16. on each other and the ideal gas law defines a relationship between them pV nRT where R is a constant p pressure V volume T temperature and n number of moles of the molecule This is valid only for the theoretical ideal gases and is a good approximation for many gases at normal temperatures and pressures It does not hold for real gases especially not in cryo temperatures This means that when a gas undergoes isenthalpic expansion pV constant its temperature will change If the temperature increases or decreases depends on the gas pressure and temperature In the property region inside the Joule Thomson inversion curve the expansion will cause the gas to cool see Figure C 1 This is called the Joule Thomson effect and is used in a Joule Thomson or Linde cycle 36 Appendix C Cryogenics and refrigeration cycles inversion curve cooling region Figure C 1 Joule Thomson inversion curve C 2 2 Basic refrigeration cycles One of the most well known examples of a reversible cycle in thermodynamics is the Carnot cycle The theoretical Carnot heat engine operates on this cycle using the temperature difference between two reservoirs to create work The Carnot factor defines the maximum amount of work that can be extracted from a process given the temperature levels Similarly the needed minimum work input of the Carnot refrigerator which is operating on the reverse can be estimated In the Ts temperature entropy diagram
17. open state A valve that is normally open is depicted as dark green in its passive open state and red in its active closed state Control valve diodes are normally grey indicating a valve in a static position A green diode means that the motor is working to further open the valve and a red that the valve is closing Control valve Magnetic Valve Magnetic Valve passive opening open true closed False closed true open False closing Figure 3 4 The diode types at the UI Front Panel 11 Chapter 3 Method and implementation program description 3 5 2 Logger The VI called logger is run on the PC and must be started separately It is entirely passive and stores all data provided on the network by cRIO The collection of log files is not necessary but can help to trace when a specific problem first occurred or compare data over longer periods of time The Front Panel of logger only provides buttons to turn the logging on or off To view or analyse the data a separate program is needed that can import the files It is possible to choose the file format in logger The default is to store the data in a LabVIEW Measurement File lvm file a plain text file that is arranged in rows and columns 2 This format was chosen as it can be opened even in a very simple text editor but to plot or analyse the data the use of Excel or Matlab would be a better option The data can also be saved in more compact binary formats these ar
18. panel with additional detail added 30 UI contains LED representing the control signals to the valves 31 UI contains numeric values representing key pressure and temperature readings 32 Start and stop buttons to control Liquefier Purifier and LIN cooling implemented 33 Buttons to control all the control valves manually 46 Appendix E System specification Additional data display 34 A separate display shows measurement readings in graph form 35 Key readings output signals and state information are continuously stored to log file Alarms and warnings 36 Warning events are triggered at certain levels of key measurements 37 At critical events an emergency stop must be performed 38 Threshold values for warnings and emergency stop must be decided 39 Safe emergency stop sequences must be decided for each state and system Remote operator control 40 Existing LabVIEW feature is used to expose user interface over internet 41 Email notifications on warnings is a lower prio feature 42 Ordinary HTTP password protection is a sufficient security level 47 Appendix F Glossary Appendix F Glossary backplane Block Diagram cRIO CompactRIO dewar enthalpy entropy FPGA Front Panel IEA isenthalpic isenthropic isobar isochor isoterm LabVIEW LIN cooling The backplane is located in the chassis of the CompactRIO All modules are connected to the backplane bus over which the communicatio
19. single thread This is to ensure all deadlines are met To guarantee this it is essential that the program is actually run as a standalone application and not presenting an interface on the host computer If the Front Panel is needed to tune parameters or in other ways change things in the program deterministic behaviour cannot be guaranteed Under normal circumstances this is not a problem as the capacity of the hardware is much higher than the program actually requires The presentation that the operator sees on the screen of the host computer is updated at a slower rate than the control and is not guaranteed to be updated exactly at regular intervals as other processes on the host computer might pre empt the program This adds to the initially very small delay caused by the Ethernet communication 3 8 Alarms and automatic shutdown The most critical part of the plant is the turbines and the turbine speed must always be monitored This is done to ensure that they do not spin too fast when the machine is running but also to ensure that they start rotating in a good way when the machine is started Build up of frozen impurities on the turbines can prevent them from spinning correctly and lose pieces can damage the turbines The system of warnings presents a warning to the operator when the turbine speed is becoming too high and shuts down the plant when necessary to avoid damage to the turbines Other cases where warnings are presented include if th
20. that the possible cooling capacity is high enough to brake the turbine at the new faster speed when necessary If the pressure is increased the turbine speed will increase and it will not be possible to brake When the machine is started from room temperature the first turbine will not reach enough speed to be regulated to optimum speed during the start up sequence instead it will be braking fully still as the inlet valve reaches its fully open position If the machine is started from a colder state or LIN cooling is used the first turbine will reach optimum speed during the opening of the inlet valve just like the second turbine When this happens it is important to allow time for the turbine to warm up meaning that the opening of the inlet valve should be paused again The plots to the left show the behaviour of the inlet valve and the turbines when the machine is started with gas at room temperature approximately 292 K see Figure 4 3 To the right is an example of when the machine and the gas inside it are colder approximately 230 K and LIN cooling is used Please note that the time scale is not the same on the two plots time markings on the x axis are set 5 minutes apart on the first and 10 minutes apart on the second 24 Turbine speed 5000 T 4000 3000 d ba ku Turbine speed Chapter 4 Testing and evaluation HE 9 5000 4000 3000 DI zl 2000 F 2000 amp 1000 amp 1000 F
21. the cool and warm start up a special case appeared that the code can not handle The tests showed that in the case of a start with very cool gas in the system as in the special case where the machine is started immediately after it has been stopped the first turbine will start spinning before the second one see Figure 4 4 The code does not support this special condition the control monitors the speed of only the second turbine in the beginning This means that there is a risk that the first turbine will not have enough time for warm up as the opening of the inlet valve and thus increase in pressure will continue This is not a common situation and usually only occurs after some kind of problem or adjustment requiring a stop For special situations like these the start of the plant should be monitored but even so it could be worthwhile to change the code for this special situation to make it easier for the operator It is not complicated to change but there has not been enough time for it during this project When the machine is started from such a cool state the production starts while the start up sequence is still running The gas is cold enough to condense when it is expanded through 25 Chapter 4 Testing and evaluation the control valve to the machine tank This is the reason why CV111 opens so early even if it the system has not yet reached its full cooling capacity In the case with the warm plant the production has still not started al
22. the cryogenic facility of Lund University in cooperation with IEA the division of Industrial Electrical Engineering and Automation at LTH Kryolab is located at the Department of Physics and produces liquid nitrogen and helium used mainly in research at the university and elsewhere in the region When possible used helium is returned as gas to be liquefied and redistributed Helium s scarcity on earth makes it a valuable resource to recover The plant for purification and liquefaction of helium gas is from 1985 and is physically very similar to plants constructed today and uses the same principles but the control system needs to be updated The plant is currently experiencing regular stops caused by malfunctions in the control system and unforeseen blockages 1 2 Task The task is to replace the control system of a helium liquefier plant A list of the original requirements can be found in Appendix E and is summarised below The new system should be easier to understand and program have a high reliability and include some additional features The new control program should have a higher level of automation provide increased possibilities to adjust control parameters and provide a clearer overview of the state of the machine This will hopefully increase the capacity and reduce the number of stops In addition to the original control system based on analogue circuit boards there is also a PLC from SattControl implementing additional functionality
23. the states in the Purifier logic An overview of these states and how they are related is found last in this section see Figure 3 11 17 Off When the Purifier is turned off all inlet valves are closed see Figure 3 8 The only connection between the Liquefier and the Purifier is from the secondary side of the Purifier to the low pressure side of the Liquefier via valve MV169 This prevents pressure from building up as the gas in the Purifier expands when it is warmed up to room temperature by the surroundings On the primary side valve MV163 is open to the gas bag keeping the pressure low Standby The Purifier is turned on by pushing the designated button on the Front Panel of the VI named UT This causes the control program to automatically enter Chapter 3 Method and implementation program description P storage 95 1 compressor 2 HP vus D B PI 101 e I e MV160 PI 100 ad s d gas bag PI 150 1 Z ws Qe to atmosphere TIS 102 123 i TIC 104 T 90 4 MV167 9 ur not connected 7 enm MV108 P a HP Cv162 Gua SF a at Mv173 e ED Alarms and Warnings 0 100 I 104 0 27 34 TIS 11 Figure 3 8 Off the state called Standby It remains in Standby until the pressure in the buffer tank reaches a set limit normally equivalent to about half full or the state is forced to shift by the operator Purge 1 AS the Purifier is triggered to start clean helium gas is blown throu
24. this case is not exactly the same which makes it a little easier to see when the new system is connected The temperature at TI106 is dependent on the gas flow which increases when valve CV111 is opened The plots see Figure 4 2 show the measured values of how open the valve is and the temperature An optimum working point lays somewhere around 12 K The value of interest is the temperature any difference in the position of CV111 is meant to adjust the temperature to a given set point The markings on the time axis are set one minute apart The plots show first the old analogue control and then around 13 44 the new controller and at 13 54 a change in its set point from 12 to 13 K This was the first attempt to take over the control of this valve and after this the code was improved step by step to try to find a good set point and reduce the oscillations Motor speed can not be controlled so the final solution opens and closes the valve in steps allowing time for the temperature to stabilise This is a common solution for most of the control valves though the duration of and distance between the current pulses differs 23 Chapter 4 Testing and evaluation CV 111 80 0083 D 3 pad AAA ANA ae 2 s MHPPEUE ET pere CHERRIES Sy T e opeje ne 13 30 30 13 40 00 13 50 00 14 00 36 12 02 01 12 02 01 12 02 01 12 02 01 Time TI 106 o 20 7143 i TUS E 15 Pa AL X un peepee Ree 13 36 29 13 50 00 14 00 36 12 02 01 12 0
25. tutorial 4139 en Kryolab webpage Kryolab Accessed Feb 20 2012 at http kryolab fysik lu se Jones L Atkins P Chemistry Molecules Matter and Change Fourth Edition W H Freeman and Company 2002 Mineral Commodity Summaries Helium webpage U S Geological Survey Jan 2012 Accessed 13 Nov 2012 at http minerals usgs gov minerals pubs commodity helium mcs 201 2 heliu pdf Broadhead R F Helium in New Mexico geological distribution resource demand and exploration possibilities New Mexico Geology November 2005 Volume 27 Number 4 Kelly T D Matos G R Helium statistics from Historical statistics for mineral and material commodities in the United States webpage U S Geological Survey 2011 Accessed Nov 13 2012 at http pubs usgs gov ds 2005 140 ds140 heliu pdf The Federal Helium Program webpage Bureau of Land Management Accessed 13 Nov 2012 at http www blm gov nm st en prog energy helium federal helium program html Cryogenics webpage Encyclopedia Brittanica Accessed Nov 6 2012 at http www britannica com EBchecked topic 144945 cryogenics Wagner U Refrigeration webpage CERN Document Server Accessed Oct 22 2012 at http cds cern ch record 808372 files p295 pdf Cengel Y A Bowles M A Thermodynamics an Engineering Approach Fifth Edition in SI Units McGraw Hill 2006 User Manual TCF 20 Linde Kryotechnik AG 1985 33 Appendix A Kryolab Appendix A Kryolab Kry
26. used to cool down the incoming gas from the high pressure side 40 Appendix D Helium liquefaction plant Low pressure High pressure 4 Gas from storage cv103 To gas bag 4 DH MIS wn C ES To air f Turbines MV167 Pure He from sump high pressure side of liquefier gt x PE MV173 CVi82 CV175 cvi12 zx Xon 4 Machine tank Figure D 2 Left Simplified view of the piping of Purifier and Liquefier Right The cold box It is important to keep the pressure at the high pressure inlet at the correct level see Figure D 2 This is achieved by the pressure control panel see Figure D 3 Both the gas from the high pressure side and the low pressure side passes through it The pressure regulation is achieved with the help of a buffer tank containing helium gas at a pressure of 4 8 bar The Purifier is providing gas on the low pressure side and when there is a surplus in the system part of the gas is stored in the buffer tank to keep the pressure steady on the high pressure side The gas in the buffer tank is then used to maintain the pressure when the amount of gas in the system is insufficient part of the gas is continuously leaving the system through the liquefaction The compression and consequently work needed for the cooling is done by screw compressors Cooling of these main compressors and the braking compressors used for the turbines is provided through the district cooling syst
27. 2 01 12 02 01 Time Figure 4 2 Inlet valve to the machine tank CV111 and temperature at turbine outlet TI106 The first parts of the plots show the behaviour with the old control and the later parts the first attempt to control the valve with cRIO 4 1 2 Varying conditions start up from cold and warm The process behaviour at start up is strongly influenced by temperature The most critical parts of the plant are the turbines and when the machine is started from cold the turbines gain speed very quickly even at low inlet pressures As a low temperature can cause the turbines to spin too fast during start up it was important to test the start sequence starting from different temperatures The increase in pressure on the turbines and the time allowed for turbine warm up have been chosen such that they provide a wide safety margin for most cases However if the machine is started from very cold as would be the case if it has been stopped and then immediately is started again it is important to keep an eye on the turbines The regulation of the turbines is dependent on their speed rather than the cooling capacity When a turbine reaches a specific speed during start up it is regulated to its optimum working point This decrease in braking of the turbine means that the cooling decreases At this point it is essential that the pressure on the inlet is kept stable long enough for the temperature on the cooling side of the turbine to rise so
28. CODEN LUTEDX TEIE 5315 1 55 2015 New System for Control and Monitoring of a Helium Liquefaction Plant Implemented in LabVIEW C O JUS U a O dum 2 lt O v e C U U E e EN m U pm JE U 9 EN Tove Mattsson Division of Industrial Electrical Engineering and Automation Faculty of Engineering Lund University Industrial New system for control and monitoring of a helium liquefaction plant implemented in LabVIEW Master s thesis at the division of Industrial Electrical Engineering and Automation Faculty of Engineering LTH Lund University Tove Mattsson 2015 Supervisor Kryolab Leif Magnusson Supervisor IEA Gunnar Lindstedt Examiner Ulf Jeppsson Abstract This master s thesis report describes the development of a new system for control and monitoring of a helium liquefaction plant which has been carried out at Kryolab Lund University At the plant helium gas is purified and then cooled in several steps until it condensates The plant is very similar to the ones constructed today but Kryolab wished to replace the outdated control system The new system has been programmed in LabVIEW and is run on a CompactRIO from National Instruments It replaces the original analogue boards and a complementary software based system In the early stages of the project the behaviour of the process was studied Programs for plotting and logging were implemented for this purpose and used
29. O ANKAN d arr pe N RER ARSA 5 3 Method and implementation program description esee 6 351 Revised requirements e RERO ee a ee eee 6 3 1 1 Overall limitauons urere te C E eee 6 3 1 2 Hardware and signals z etie em eiecit ete ede need 6 3 1 3 Control program software on CRIO esee enne 6 3 1 4 User interface software on host computer esee 6 3 1 5 Alarms and Warnings nti decederet eene eder tree ite deter ede ea 7 3 1 6 Remote operator control esee eene enne enren rennen nen 7 3 2 Choice ot state machines ect eee ee te getto ete Std erae N 7 3 3 Study of the system with plotter and logger esee 7 3 4 Incremental replacement of control system parts eese 8 3 5 Program structute oU St ceed Qe D HER Edit Ur p tee I RR et 8 3 5 1 DEP 9 3 5 2 LO0gB8etc suce SD HR SR Be IRE E e ee TRUE 12 3 5 3 du oreert ea EA sates weld ccs test devas esac NN RE Ro RAA RA 12 3 5 4 Communication and libraries cece ceeceseceseeeeeeeeeecaaecaeceseceaeesseeeseeeseeenneeees 13 3 5 5 latge t ines upper eet ser Gee Ug e Ite dte Leta hee ee 13 3 6 X Control program logic uer tee eee miren ie e en Re eren 13 3 6 1 Liquefier logic description of states sese 14 3 6 2 Purifier logic description of states sssressrsressrsrsssrersrsresrrsrrssresrsrrsrrsrrrsrrr rna 17 SEE niic ect isks antes eves tease rarte a
30. Projektet har som helhet varit mycket lyckat Den f rb ttrade verblicken underl ttar s v l identifieringen av potentiella problem som orsaken till dem De kade m jligheterna att p verka styrningen har gjort det m jligt att k ra anl ggningen under sv rare omst ndigheter n tidigare M jligheten att vervaka maskinen fr n en dator eller telefon utanf r Kryolab har gjort det m jligt att i h gre utstr ckning undvika n dstopp och minskat behovet av att ta sig till Kryolab utanf r planerad arbetstid Det nya systemet ger operat ren b ttre insyn i hur regleringen fungerar och f rhoppningen r att detta ska bidra till en kontinuerlig utveckling och f rb ttring av styrsystemet II Acknowledgements First of all I would like to thank Leif Magnusson at Kryolab for giving me the opportunity to do this work introducing me to the problem and for the help with setting up relevant requirements for the new control system His knowledge about the plant has been very valuable and he has been very helpful doing much of the new electrical wiring needed My supervisor Gunnar Lindstedt at the Division of Industrial Electrical Engineering and Automation IEA put me in contact with this project and encouraged me to learn LabVIEW Getachew Darge also at IEA gave good feedback on my ideas for the simple analogue circuitry that I built Examiner Ulf Jeppsson gave useful feedback on the presentation of the project and on this report Help wi
31. The inert gas neon is generally considered too expensive Nitrogen is along with helium most widely used Helium is exceptional as it has no triple point at all It may solidify only at pressures above 2 5 MPa The lambda point of helium often found in tables instead of the triple point refers to the point where two kinds of liquid helium one of which contains a superfluid part and gas coexist rather than solid liquid and gas The fact that helium evaporates at such an extremely low temperature makes helium the most useful and popular option for cryogenic purposes Nitrogen does not have as useful characteristics but is still widely used when it is possible due to the lower cost 10 C 2 Thermodynamical theory of the plant To understand the theory behind the Liquefier some knowledge in thermodynamics is essential This chapter briefly describes some basic concepts particularly useful for this purpose These descriptions are intended as reminders rather than exhaustive explanations Some useful expressions can be found in the list in Appendix F C 2 1 The ideal gas law and the Joule Thomson effect Enthalpy is the total amount of energy in a thermodynamic system If no energy is added and no work is done on a system the enthalpy should remain constant How much energy a system with gas has depends on how much gas there is number of molecules which volume it occupies and which pressure and temperature it is kept at These entities are dependent
32. able Slightly increased loop times will still yield a good over all behaviour as long as both are controlled at regular intervals as one is dependent of the other Any other activity where determinism is required must consequently also be integrated in this VI This is the case with e g the Alarm system It is desirable that it can affect the logic with regular intervals and must not be starved 13 Chapter 3 Method and implementation program description The control logic is as previously stated based on state machines Any particular state does not necessarily represent fixed control signals rather it describes where in a process the machine is to which values measurement signals should be controlled and the direction of gas flow 3 6 1 Liquefier logic description of states The different liquefier states normally follow each other in the order they are described in this section There is also an illustration showing how they are interconnected see Figure 3 6 STANDBY Standby is a state suited only for short term shutdown It is assumed that the operator has ensured that certain requirements are met before the system is started This includes correct valve positions on manual valves and that the system is filled with pure helium START 0 prepare Control valves are regulated to their start positions CV103 fully closed CV111 fully closed CV175 fully closed CV120 and CV121 80 90 open START 1 open buffer The valves to the buffer ta
33. adjust as many parameters as possible was so appreciated that it is still normally used So far the cRIO always had so much extra computing capacity that it could present its Front Panel on the host computer without causing any disturbances of the control It has been ensured that the control program in cRIO resumes normal control if contact is lost with the host computer If contact is lost with the host computer the possibility to use the Front Panel of control is lost even as the connection is re established again The VI UZ will still work though so the behaviour can still be affected yet at a higher level the level that was originally intended Even if this works and the extra capacity of cRIO makes it unlikely that the control should not meet its deadlines it is important to note that determinism cannot be guaranteed unless the program is uploaded to cRIO as a standalone application It has not been a goal of this project to evaluate LabVIEW but some of the experiences of the programming environment in this particular project are summarized below This is from a beginner s point of view Previous programming experience consisted primarily of projects in Java and other textual languages The experience of more graphical approaches was small and limited to the graphical languages defined in IEC 61131 3 for programmable logic controllers and the graphical extension of Matlab called Simulink 29 Chapter 6 Discussion and conclusions LabVIEW
34. along with other available resources such as manuals drawings and collected data from the plant The code has been developed iteratively and every part of it has been tested before it was used to control the real plant To verify the program behaviour in different situations a user interface with the possibility to set sensor values manually has been developed New functions have been tested and added successively The new system presents more data and a better overview of the process than previously available Increased possibility to manually control some valves and to adjust trigger and set points used in the control has also been added A new alarm system has been implemented and the start and stop sequences have been further automated The project has been very successful The improved overview aids in identifying potential problems and their causes It has proved possible to run the plant under more difficult circumstances with the extended ability to adjust the control system The opportunity to supervise and control the plant from a computer or telephone outside Kryolab has reduced the need to go to Kryolab outside regular working hours and made it possible to avoid some emergency stops The new system provides the operator with improved insight in how the control works The aspiration is that this will facilitate further development and improvement of the system Sammanfattning Denna rapport beskriver ett examensarbete utf rt pa Kryolab pa
35. and the activity there Facts about helium can be found in Appendix B and a brief description of the field of Cryogenics and some basic refrigeration cycles and how they compare to the one used in the plant in Appendix C The functionality of the machine TCF 20 for which the control system is to be replaced is described in Appendix D as is the rest of the helium liquefaction plant 2 2 Plant overview The returned gas that is to be liquefied is compressed and put in storage see Figure 2 1 The gas contains impurities essentially air which needs to be removed In the Purifier the gas is cooled until all impurities condensate or solidify and the pure helium can be transferred to a buffer tank As impurities build up in the Purifier it needs to be regularly heated and purged with pure helium gas The Liquefier is supplied with pure helium gas from the buffer In the Liquefier the helium is cooled this is described in Section 2 3 The liquid helium is collected in the machine tank from where it is transferred to other vessels for delivery Buffer tank aes Figure 2 1 Overview of the plant Liquefier Machine tank 2 3 Liquefaction process The Helium gas is cooled in a cycle where it is alternately compressed by screw compressors and cooled during expansion through turbines The work needed for the cooling is provided by the compressors High pressure gas enters the Liquefier from the compressors and gas with low pressure
36. arily someone who could provide solutions but a person to discuss questions with and explain problems to The NI support was excellent when it came to basic LabVIEW questions and the manuals and the Kryolab staff provided much information regarding the plant and how it worked But often a combined knowledge of both the plant and the programming environment was needed to understand the challenges It is a benefit when solving problems to have more than one point of view and multiple sources of ideas For the implementation time on the other hand it would probably not have made as big a difference since the timing was much related to opportunities for testing The response from Kryolab on the project has been very positive All of the important criteria set up by the Kryolab staff at the beginning have been met and the new control has made it possible to handle increasing amounts of return gas This leads to the conclusion that it has been a successful project 30 Chapter 7 Future work 7 Future work It has been the aim to build this control system in such a way that it is easily expanded and some changes are already prepared for Code already written but not tested includes regulation of control valves CV 120 and CV 121 which control turbine speed and the magnetic valve MV 108 providing a bypass for the gas when the machine is too warm to open the valve to the machine tank At this point the analogue boards provide an excellent and robust regula
37. ates 16 Chapter 3 Method and implementation program description 3 6 2 Purifier logic description of states The Purifier provides the Liquefier with pure helium gas Before the return gas enters the Purifier it has been dried of any water vapour In the Purifier the gas is cooled down in a heat exchanger causing any impurities e g nitrogen to condense or freeze on the wall The impurities fall into the sump either immediately or when the Purifier is reheated during regeneration Regeneration is triggered when the buffer tank is full or when too much impurity has built up on the walls The state machine for the Purifier is based on the normal cycle of the Purifier which is described in Appendix D about the machine TCF 20 In the user interface described in Section 3 5 1 the Purifier can be found to the left The valves controlled by the purifier logic are MV160 MV163 MV167 MV173 CV162 MV 168 CV175 MV169 MV172 and MV 174 see Figure 3 7 V 404 cooling water compressor temp pressure control panel 16 1 37 0 lt lt line drier ul E L EIE y Mv160 MV169 PI100 0 0 PI101 9 1 P storage 91 9 compressor gas bag to atmosphere TIS 102 49 1 i 27 9 TIC 104 MV167 not connected My108 TIS 106 PI 151 E sump L cv162 CV 175 T HP 12 6 MV173 e e a Figure 3 7 On the Front Panel of the UJ vi the Purifier is depicted on the left The rest of this section contains the description of
38. bly the most important side effect of this project The increased possibilities to adjust the control also mean that it is possible to run the machine under more difficult circumstances In some cases this can be very good and the opportunity to use remote control of the plant can for example be used to prevent the need for an emergency stop However when the plant is run under abnormal conditions it will have to be monitored continuously It is important to be aware of this and not use this possibility more than necessary Even if the remote control ability means that whoever is responsible for the machine can monitor it around the clock it was meant to reduce work for the operator not to create a situation where the plant controls the operator When the automatic control system can not handle operation it should be seen as an indication of maintenance need That the control system should be run separately and deterministically was one of the most important requirements However in reality it is often not A list of all the parameters that should be possible to control manually was established early in the project But during the construction of the VI control it was necessary to implement a Front Panel to make it possible to tune and adjust far more parameters in the control during runtime Since the control parameters have been chosen and tested the Front Panel of control should not be needed anymore during normal operation but the opportunity to
39. closed STOP 8 venting 24 hours are allowed for the system to vent any remaining impurities to the balloon via CV119 STOP 9 prepare standby Control valves are regulated to resting positions CV103 open CV111 closed CV175 closed 15 Chapter 3 Method and implementation program description STANDBY amp amp CY 403 100 STOP 10 prepare standby START D prepare 2 min R amp CV 11 lt 1 TOP amp amp CV 178 lt 1 amp amp CV 103 0 STOP 9 ti Nott M 1h 1gs START 2 compressor 1 STOP 8 venting is 105 START 8 compressor 2 FI 101 3 6 0 bar STOP 7 clase buffer START 4 open CV 103 to 0 7 bar 105 PI 102 F 0 7 bar amp amp SIC 100 gt 200 rps amp amp SIC 10 gt 200 rps STOP 6 compressor 1 1 8 START 5 open Cv 103 to 1 bar i STOP 5 compressor 2 PI 102 1 bar amp amp SIC 100 gt 1000 rps amp amp SIC 101 1000 rps gt START 6 open CY 103 to T2 3200 ros STOP 4 close v 118 5 min STOP 3 open Cv 111 PI 101 4 1 5 bar STOP 2 clase v 100 Cv 109 0 STOP 1 close CY 103 SIC 101 3 3200 rps START 7 static Cv 103 T2 warm up 10 min e START 6 open CY 103 ta T 3200 rps 1 Cv 103 100 CV 1034100 se sic 10d gt 3200 rps START 7 static Cv 103 T1 warm up 10 min STOP START 8 open CM 103 fully or CV 103400 RUNNING Figure 3 6 Liquefier st
40. cription Cooldown 2 The cooling of the Purifier continues and as it grows colder the pressure on the primary side will drop When it reaches below 30 K it is so cold that any impurities will condensate on its walls and fall into the sump Purify Purify Reset and Pre Purify At this point valve MV160 from the storage of dirty gas will open As long as the temperatures are maintained low enough MV168 CV162 is open This allows dirty gas to enter on the primary side where any impurities will remain either frozen or in liquid form and move on to the secondary side where it is used to cool new incoming gas before it enters the Liquefier on the low pressure side As the gas in the storage is room temperature it will cause the temperature reading to jump when the valve is first opened MV168 CV162 then closes allowing more time on the primary side so the gas is properly cooled MV168 CV162 will open and close a few times in the beginning before the walls of the heat exchanger are thoroughly cooled not only the gas in it and the valve remains in an open position When the Purifier has been running for a while CV175 will gradually close more and more By the time the walls of the heat exchanger are cooled the cleaned gas itself will be cool enough to cool the incoming dirty gas with little help from the gas cooled in the Liquefier The reason for the three states is that it is important that the temperatures have stabilised before regulation of CV175 star
41. do not suit the needs in this specific case A very simple test program has been made to see that it is possible to send emails with notifications of warnings 41 but this is not a working feature in the project as it was a lower prio feature 28 Chapter 6 Discussion and conclusions 6 Discussion and conclusions The most important requirements set up in the beginning of the project have been met and the system has now been in use for more than six months The plan for the work has been followed and the iterative process and stepwise replacement of the old system worked very well The staff at Kryolab has been involved in the design and evaluation of the new control system This has been a key to make the final version not only a competent control software but also extendible and adjustable according to future needs The new automatic start and stop sequences are working well facilitating the work of the operator All of the parts in the old systems that where no longer functioning properly have been replaced and the current system is much more stable and reliable The improved possibilities to visualize collected data have provided a lot of information and an opportunity to understand the process better Variations in behaviour over the cycle can be followed and it is much easier to see deviations from normal behaviour now The increased knowledge and understanding of the process that the visualisation has brought to everyone involved is proba
42. e D 7 43 Appendix D Helium liquefaction plant Low pressure High pressure Gas from storage l l l V100 Turbines Pure He from high pressure side a_n of liquefier P lt MV173 CV162 CV175 Machine tank Figure D 6 Gas flow through the Purifier during regeneration when the heat exchanger is warmed to remove any frozen impurities Low pressure High pressure Gas from storage t I CV103 Turbines Pure He from high pressure side of liquefier CV175 MV173 CV162 Machine tank Figure D 7 Gas flow through the Purifier during Purge when any remaining impurities are blown to atmosphere after regeneration 44 Appendix E System specification Appendix E System specification List of original requirements The overall goal is to extend or replace the existing semi automatic control program interfacing the already operational helium liquefier machine as well as provide improved remote monitoring functions The system will be implemented in LabVIEW on cRIO Approximately 30 40 different voltage signals analogue and digital should be sampled and transformed from a voltage to pressure temperature rotational speed etc and presented on a user interface on the host computer These values should be used to determine the action in the control program Several of these signals will have one or more threshold limits that should activate warnings or activate the stop program Importa
43. e Purifier and the liquid nitrogen cooling The Liquefier can run separately without the Purifier if it is provided with pure helium from another source It can also run without the extra cooling this normally gives a production rate high enough to liquefy all return gas The control program controls the Liquefier and Purifier The state of the control program can be described as a combination of the states of the Liquefier and Purifier An early thought was to merge these states into one state machine as the Liquefier normally is in a steady running mode when the Purifier operates The main reason for instead keeping the state logic separate and making one state machine each was the scalability The later addition of another state machine for the LIN cooling was facilitated this way Like the Purifier it runs independently but will shut down if it detects that the Liquefier for any reason performs a stop sequence Even though the Purifier operates largely independently all control logic is placed in the same VI The main reason for this design choice is how the program is compiled before it is loaded to the cRIO In order to guarantee a deterministic behaviour the code is compiled into one single thread Two different loops or VIs can not both be guaranteed to meet their deadlines only one can have the highest priority Separating the Liquefier and Purifier would mean that the control of either of them would have priority over the other which is undesir
44. e also LabVIEW specific file formats but can be read in Matlab Excel etc sometimes requiring the installation of an add in 3 5 3 Plotter The diagrams provided on the Front Panel of plotter see Figure 3 5 shows some key measurement values over a shorter period of time enough to be able to see what has happened during a night and to get an overview of a few cycles This makes it possible to spot trends like regular problems with blockage in the Purifier which are often caused by high levels of impurities in the return gas It also makes it easier to estimate the rate at which the storage is emptied and the production rate Temporary variations caused by filling of dewars and drop in production rate when the Purifier uses part of the cool gas during cool down can easily be seen and taken into account The plots are also useful to monitor turbine behaviour if the machine is started from an unusual state e g very cool which can cause the turbines to spin too fast File Edit View Project Operate Tools Window Help gt een Pressure Buffer Tank bar 7 7 Temperature liquefier K T1106 124 T1111 345 Pressure machine tank bar 0 18 Turbine speed rps TL white 4629 T2 red 3766 AVA VATANA TAVAVAVA VAN Fig 5000 0 5 4000 3000 rere erepti rper n p 02 18 04 00 06 00 08 00 10 00 12 00 14 18 BIBIBIRHRINIRERHIRERIREREERERERER EEE 02 18 04 00 06 00 0800 10 00 1200 1418 0248 04 00 0600 0800 1000 1200 14 18
45. e cooling is not working properly causing compressors or turbines to overheat and if the pressure across the Purifier is too high which is normally caused by a blockage The trigger points have originally been chosen taking documentation and the recorded behaviour of the process into account During the months that the plant has been running with the new control system many of the values have been adjusted through simple trial and error to new values to ensure that the warnings and alarms are not issued with unnecessary regularity and that they really are active when necessary 3 9 Transformation The transformations of measurement values from voltages to other units are generally simple In most cases they are done by linear equations and give good measurement values close to the working points of the machine and less good for extreme temperatures or pressures For some temperature sensors more complex approximations have been used in combination with an offset to compensate for the drifting over time In some cases the transformation has been chosen such that the values are in accordance with the ones measured by the old system and have not been further validated To get good and accurate measurement values which are truly correct and not approximate and based on previous systems a recalibration is in some cases necessary and has not been done 21 Chapter 4 Testing and evaluation 4 Testing and evaluation The revised list of requirements in
46. ected to the control program on a separate computer during development and testing cRIO itself does not have any screen Chapter 1 Introduction A real standalone user interface run on a host computer must then be implemented The control program must be independent of this user interface The user interface should present information and possibilities to transfer information to the control program to influence the process Possibilities for remote control of this user interface should be added Documentation of the work and some kind of user manual will be necessary 1 4 Report overview This chapter contains information about the task as well as some background information and the basic outline of the work More information about Kryolab can be found in Appendix A Chapter 2 contains background theory needed to understand the report It describes the old control system the new hardware and LabVIEW that is used for the programming There is also a brief description of the plant and the liquefaction process Further information about helium can be found in Appendix B and a brief description of the field of Cryogenics and basic refrigeration cycles in Appendix C The functionality of the machine TCF 20 for which the control system is to be replaced is described in Appendix D as is the rest of the helium liquefaction plant Chapter 3 describes the implementation of the new control system Section 3 1 gives an overview of the requirements set up f
47. ed It is important that any added code is functional from the beginning The machine is stopped at fairly regular intervals but normally only for a short amount of time It is important to use each such opportunity to make updates to the control system replacing and testing limited parts each time The new control system is to replace and extend an already operational control system the parts that serve as replacement should in all essentials work as the old system The logic set points etc should initially be the same as in the old system but can be revised to increase performance as the new system is tested 3 1 2 Hardware and signals It is decided that the system should be built with LabVIEW 2011 and CompactRIO 9022 Hardware modules for the IO already available are NI9205 0 10V DC for inputs and NI9477 60V sink and NI9476 24V source for the outputs The measurement signals that are to be read are direct current voltages most of them with a maximum value of 5 or 10 V but some ranging to 24 V There are also some signals with alternating current and voltages up to 220 V If at all possible these will not be measured and used High voltage increases the risk of personal injuries and should not be integrated in a system that is to a high extent open and easily accessible to the operator 3 1 3 Control program software on cRIO The main analogue control program is now controlling the action of the purifier and all the valves of the liquefie
48. em that supplies cold water Figure D 3 Left Pressure control panel Middle Buffer tank Right Screw compressors 41 Appendix D Helium liquefaction plant D 3 Return gas handling Return gas arriving in pipes from research labs throughout the university is measured by flow meters and collected in a big gas bag or balloon see Figure D 4 Before it is compressed into the storage which holds gas to a maximum of 200 bar it is dried of any water vapour Gas that is collected from the plant itself e g the gas going through valve MV 163 see Figure D 2 during the purging process of the Purifier is also collected in the gas bag and handled the same way Remaining return gas comes in gas bottles that can be connected either instead of the regular storage or in parallel with it such that gas from both sources is fed to the Purifier through valve MV 160 Figure D 4 Upper left The gas flow meters measuring return gas Upper right The gas bag Lower left Gas compressor Lower right The storage of incoming gas 42 Appendix D Helium liquefaction plant D 4 Purifier The dirty gas enters via a pressure regulator through valve MV160 see Figure D 5 The box on the left hand side represents a heat exchanger where the incoming gas is cooled in two steps In the first step it is cooled to 70 K If the impurity level is above 1 5 the impurities will condensate and the liquid is collected in the vessel called the sump from which it
49. en necessary as the behaviour of the process itself would ensure that a specific value triggering an action could not be reached during wrong state The choice to connect almost all actions to the state machine was made partly because sensors sometimes can give an erroneous value but also because it was pointless using computing power for comparisons that provided everything worked would always provide the same result The choice to make multiple state machines was not an obvious one but it was decided to use separate state machines for the Liquefier and Purifier It would have been possible to merge these into one without the problem of a high increase in number of states as the Liquefier basically has one running state and all the rest are connected to the start up and shut down of the system The Purifier could be limited to be active only when the Liquefier is in its stable running state However the separation into two state machines represents the real structure and mimics the original behaviour of the control system better The Liquefier can be run without the Purifier if it is provided with clean helium this possibility was kept even if it is not normally used The modularisation also facilitates a future extension of the system If other parts are to be connected as would later be the case with the LIN cooling they would also have there own separate state machines possibly with the states of the Liquefier and Purifier as inputs 3 3 Study
50. erature and pressure are so low the gas would be represented inside the Joule Thomson curve see Figure C 1 This is the cooling used in a Joule Thomson or Linde cycle It is not possible to liquefy helium though most other gases in a pure Joule Thomson cycle At ambient temperature we are far outside the cooling region in the Tp diagram in Figure C 1 The helium must first be pre cooled It is the combined use of the turbines and the throttling that cools the gas This combined cycle is also more efficient than the pure Joule Thomson cycle The combined cycle described here is often referred to as a Claude cycle 8 11 39 Appendix D Helium liquefaction plant Appendix D Helium liquefaction plant D 1 Overview This appendix contains a description of the Helium Liquefaction Plant which parts it consists of and the helium s path through it A more basic description of the functionality of the plant can be found in Chapter 2 More information about the theory behind the Liquefier and a basic description of the refrigeration cycle is available in Appendix C The information on which this appendix is based has been gathered from the manual of TCF 20 12 from Kryolabs web page 2 and from the Kryolab staff The Helium Liquefaction plant is used to collect incoming gas clean it from any impurities liquefy it and transfer it to vessels for delivery Figure D 1 shows a simplified view of the helium s path through the system The return gas
51. es do not affect the control system in any way as it is built to work separately from the host computer This independency was one of the most important requirements To regain access to the plots and options for manual control a restart of LabVIEW on the host computer is necessary after which connection can be re established with cRIO Data collection from the crashes on the host computer could help explain these problems At this point the logging is limited to the collected data from the plant More information about what is happening on the host computer and on cRIO is needed to determine how to solve this problem Most importantly it would be advisable for any operator of the machine to strive for a basic knowledge of LabVIEW and the program structure as the program is prepared with the explicit goal of making code that can be altered and further evolved by the operator It has 31 Chapter 7 Future work been the intention to make it as open and easily understandable as possible Still some basic understanding of how changes in one part of the code will affect the rest is a necessity 32 References References 1 10 11 12 Using NI CompactRIO Scan Mode with NI LabVIEW Software webpage National Instruments Dec 19 2011 Accessed Jan 12 2012 at http www ni com white paper 7338 en Specification for the LabVIEW Measurement File lvm webpage National Instruments Jul 07 2010 Accessed Jan 12 2012 at http www ni com
52. essor joacox 47 6056 16 6651 2 00695 Prol SIC 100 cvi CV 103 100 Alarm Stop close CV103 Automatic Stopand Buffer Alarm Stop P1101 a CV175 CV 103 10096 i f 2 f i gis Jo 2 24 Jo C 8 07028 4647 45 5 31572 7 90466 P1102 SIC 101 mao CV 103 0 102 SIC 101 102 CV 103 0 je de 212 oto 130565 5 125 968 Jo oog71 103 Pstorage T1106 CV 111 ref slow CV175 103 storage T1106 CV111 ref 28 Ho Cm fo E 4 7 2 05785 155053 12 5839 214976 P1104 Us107 T1104 CV 175 ref 104 us107 104 CV175 ref Ao z Jo Hss do 0 22497 421209 92 018 212843 LUE usigzref TU Vmach Liquifier State Purifier State Usi07 ref T Vmach giis Jo 45 7 Jo RUNNING Standby 1 30563 6 16463 340273 280 851 P1150 P balloon MV 110 CV 175 ref2 rtris 150 balloon MV 110 CV175 ref2 fo Ao Z 1 48143 617926 7 96864 10 102 J p J 5 hi v Goto next skip purifier state How fast CV103 i Turbine Turbine Compressor Cooling water o x EM Turbinel Turbine 2 jo Jo Jo Jo V iow fast CV103 close regenerate Immediately enter 37 2328 107412 39521 A gy regenerate How many seconds between eset i i i 0 5 s of active valve activate MV Liquefier SRA E pu LIQUEFIER Cx 2 OPEN CV103 CLOS CV175 open PID coolt Y moon Popre comp 1 comp CV activate E bute ER a comedi aD comedia p kp ko Ko Kp close A r z 01 2112 A oi doi Foi Jo v100 v404 v8 0 25 50 75 100 7 m E t
53. f the measurements where really correct and calibration should be done on a regular basis The transformations used are approximations and provide the most reliable data within the natural working range of the plant It would be interesting to redo the calibrations and maybe also develop more complex transformations that give more accurate measurement values It would make it easier to compare the plant with others or with future data The current solution for the filtering of signals is very simple using only an average value over a number of samples This is enough to give the intended behaviour of the control but could be improved to provide more accurate information to the operator Continued development of the logging and warning systems could also be useful including the use of e mails or SMS to communicate errors and potential problems as well as tools to easier diagnose incorrect system behaviour from the logged data There have been occasional disturbances in the communication between the host computer and cRIO as well as a few crashes of LabVIEW on the host computer providing the user interface It is hard to determine to which extent this is caused by this project as the same problem has also occurred earlier with some simple programs used to log the behaviour of other systems at Kryolab It is likely that this is related to LabVIEW or other activities carried out by the PC rather than the code for the user interface of this project These crash
54. gh the primary side of the Purifier for 5 seconds to blow any remaining impurities from the last cycle to the gas bag where return gas is collected before it is compressed to storage After this the valve to the gas bag MV 163 is closed Cooldown 1 To cool down the Purifier cool helium gas from the Liquefier is let in through valve CV175 see Figure 3 9 This gas goes through the secondary side and back to the low pressure side The secondary and primary sides are still kept separate by valve MV168 CV162 To avoid pressure drop on the primary side as the Purifier is cooled down valve MV173 remains open for clean helium gas from the high pressure side of the Liquefier to the primary side of the Purifier keeping the pressure constant Purge 2a and 2b As the temperature reaches below 80 K LP v118 3 JE PI100 0 P storage line drier j d PI101 9 1 compressor gas bag PI150 9 1 to atmosphere TIS 102 106 Jj ss TIC 104 y MV167 9 Prial not connected Mv108 sump D HP Cv162 CV 175 ick a ee F MV173 d DENDENETT Alarms and Warnings 0 100 1104 EM Figure 3 9 Cooldown 1 on the secondary side the primary side is purged with clean helium gas from the high pressure side of the Liquefier again MV163 is then closed and after another five seconds the inlet of clean helium gas MV173 too 18 Chapter 3 Method and implementation program des
55. important signals can be used Control program software on cRIO 16 Each probe will require specific calibration and conversion in SW 17 SW filter or regulator may be needed to stabilize measurements 18 SW regulator may be needed for output signal control 19 Measurements read from AD converter represent pressure temperature etc 20 SW frequency sampling needed to read turbine speed 21 Timers required to measure time 22 Start and stop sequences should be based on the instructions for manual start and stop found in the manual for TCF 20 and if needed the behaviour of the current semi automatic system on the SATT control 23 The logic controlling the purifier should be based on the old analogue control system 24 Logic for LIN cooling must be decided based on current operator experience 25 When documentation and behaviour differs the current behaviour is rule 26 The control program should be possible to run with direct contact to the host computer presenting a Front Panel there that can be used to change process parameters in the control program at runtime Run this way the program is not required to be deterministic 27 The control program must be possible to run as a standalone application on cRIO in which case it must be deterministic User interface software on host computer 28 The control SW is standalone and functions well without UI 29 The software user interface should mimic the current analogue operator
56. in Figure C 2 the work is equal to the enclosed area 1 2 3 4 logT s const v const p const 4 isentropic isochoric esu Figure C 2 Ts diagram for various cycle processes T is temperature s is entropy Carnot cycle 1 2 3 4 Stirling cycle 1 2 3 4 Ericsson cycle 1 2 3 4 Brayton cycle 1 2 3 4 Other theoretical alternatives to the Carnot cycle that are also very efficient could be the Stirling or Ericsson cycles During the temperature increase the entropy is allowed to increase 37 Appendix C Cryogenics and refrigeration cycles and instead either the volume or the pressure is kept constant The three basic cryogenic refrigeration cycles with isothermal expansion and compression discussed here the Carnot 1 2 3 4 Stirling 1 2 3 4 and Ericsson 1 2 3 4 cycles are all ideal cycles with good theoretical efficiencies see Figure C 2 They all enclose a small area which should indicate that a small amount of work is needed to cool gas in this way But the isotherm constant temperature transitions are very difficult to approximate The Brayton cycle 1 2 3 4 makes use of isobaric compression and expansion like the Ericsson cycle and isentropic expansion and compression like the Carnot cycle This way no isothermal expansion and compression is needed It has significantly lower efficiency by definition than the other cycles this
57. is represented by the larger area enclosed in the Ts diagram But it is also much easier to approximate than the other cycles It can be said to form part of the basis for the cycle used in the Liquefier but some alterations are still needed C 2 3 Liquefier refrigeration cycle log T compression two phase area S Figure C 3 Ts diagram for the Liquefier TCF 20 Figure C 3 depicts the Ts diagram for the Liquefier where a cycle very close to the Brayton cycle can be identified 1 2 5 6 The cooling 1 3 and heating 5 6 is done at constant pressure The gas compressors are in their nature isentropic 6 1 even though some cooling is applied to limit the temperature And the expansion through the turbines 2 3 4 57 is at least ideally isentropic though in reality polytropic The work exerted on the turbines causes the gas to cool The purpose of the turbines and consequently the Brayton cycle is to reach low enough temperatures to be able to use the phenomenon previously described the Joule Thomson effect In the Liquefier helium gas at low enough temperature and pressure is expanded in a valve 3 4 becoming so cooled that it enters the two phase area the liquid is collected in the 38 Appendix C Cryogenics and refrigeration cycles machine tank and the gas is used to cool incoming gas through a series of heat exchangers on the way back to the compressors This cooling during expansion is due to the fact that the temp
58. l actions and the control should not rely on any information from the operator during normal operation All control actions are placed in one VI which is given the highest priority The handling of communication over Ethernet with the host computer is taken care of by other VIs with lower priorities The cRIO is fast compared to the cycle times needed to control the plant generally only approximately 13 of the processor capacity is used This means that even if the VIs handling the communication has a lower priority it is unlikely that cRIO will have difficulties to maintain the communication During the development it is possible to be online and show the Front Panels of cRIOs VIs on the host computer This is very useful to monitor and influence the control and fine tune the parameters used but determinism cannot be guaranteed Chapter 3 Method and implementation program description 3 Method and implementation program description 3 1 Revised requirements From the original criteria specifying the new system and with the environment in which it would be done taken into consideration a set of demands and tasks where formulated to guide the work These are presented in the list of revised requirements in Appendix E Some of these criteria and the reasons for them are described in the following sections 3 1 1 Overall limitations Since the machine is operational and running for the most part the possibilities to test the code will be limit
59. mately 75 of the world production came from the U S and a great part of it from the Federal Helium Reserve Figures from the U S Geological Survey indicates that about 30 of the helium used worldwide is extracted from this resource Other production sites around the world currently running include fields in Algeria Russia Poland and Quatar The demand for helium is increasing and the fact that helium is produced at a very limited number of production and refinement sites and that the production is so dependent on the demand of natural gas means that the availability of helium on the world market has been unstable in recent years 4 7 8 35 Appendix C Cryogenics and refrigeration cycles Appendix C Cryogenics and refrigeration cycles C 1 Cryogenic fluids Cryogenics is the production and application of low temperature phenomena It is a utility enabling applications such as liquefaction separation storage and transport of gases and the altering of material and fluid properties by reduced temperature The temperature range in question has been defined as ranging from absolute zero to 123 K 9 In a cryogenic system a cryogenic fluid is used as a working and cooling agent These fluids must have a low triple point Hydrogen neon nitrogen and oxygen all have triple points below 100 K Oxygen is difficult and dangerous to work with because of the high reactivity and the use of hydrogen has diminished over the years for the same reason
60. most two hours after full pressure has been achieved on the turbines Turbine speed a 5000 D a 0 m1 LI I I I LI LU n 15 30 49 15 40 49 15 50 49 16 00 49 16 09 14 12 05 04 12 05 04 12 05 04 12 05 04 12 05 04 Time CV 103 D DI 2 2 E d I I I I LI 15 30 49 15 40 49 15 50 49 16 00 49 16 09 14 12 05 04 12 05 04 12 05 04 12 05 04 12 05 04 Time CV 111 Li DI 2 E X ar ES RS US TR TCU T 15 30 49 15 40 49 15 50 49 16 00 49 16 09 14 12 05 04 12 05 04 12 05 04 12 05 04 12 05 04 Time Figure 4 4 Measurements during the start sequence of Turbine speed the opening of the inlet valvet to the turbines CV103 and the valve to the machine tank CV111 Plots show the behaviour with very cold gas in the system 26 Chapter 5 Results 5 Results The overall goal of the project was to extend or replace the existing semi automatic control program interfacing the already operational helium liquefaction machine as well as provide improved remote monitoring functions The system has now been in use for over six months and has been functioning as planned This is the best indicator of a successful project Almost all of the old control system is replaced the main difference is the addition of an extensive user interface Interesting signals that were previously internal to the control system are now presented both in their context on the UI Front Panel
61. mp oe when the temperature is so high that all 9 g 1 impurities have melted From the sump Mv173 9 the impurities mainly oxygen and ee oe has I EE 1 nitrogen are blown to atmosphere Figure 3 10 Regenerate 19 Chapter 3 Method and implementation program description gt 110 K TIS 102 110K Controle Purge 2a Cooldown 2 Purge 2b Figure 3 11 Purifier states 3 7 Timing The loop time is set to 200 ms which gives a considerably slower control than the old analogue continuous one The process itself is stable and slow in its responses so this is fast enough The hardware can easily handle much faster regulation but it is not necessary in this case The response times can in some cases be a second or more creating risks for overcompensating and much regulation is done even slower than 5 times per second when opening a valve for example it is done little by little to ensure the process has time to stabilise not to overcompensate for small disturbances In the case of control valve CV 175 which is controlling the flow of cooling gas from the Liquefier to the Purifier this is done with a PID regulator calculating a desired valve position rather than just seeing that the valve should open more or close more 20 Chapter 3 Method and implementation program description The control program is run on the cRIO and is built separate from the network communication and is compiled into a
62. n is carried out Apart from the bus the backplane of cRIO also contains an FPGA for fast calculations and signal processing The logic of a VI in LabVIEW Inputs and outputs are connected to function blocks and subprograms with wires The wires represent the data flow in the program A program is executed starting at the inputs and following the wires to the outputs The execution is completely dependent on the dataflow the place in which function blocks are placed in the Block Diagram has no significance only how they are interconnected The code is inherently parallel An embedded control and acquisition system from NI It consists of a chassis a controller and modules Vacuum flask used to store e g liquid helium Size ranges from 20 500 litres at the Kryolab facility It is like a large thermos with double walls with near vacuum in between for insulation A systems internal energy or thermodynamic potential denoted h where h U pV and U is internal energy p pressure and V volume The unit used for enthalpy is Joule A systems disorder denoted s The unit is Joule Kelvin For a gas the entropy is related to the amount of kinetic energy the molecules have i e the temperature of the gas Field programmable Gate Array An integrated circuit that can be configured for different tasks In the CompactRIO the FPGA can be configured to execute part of the program in dedicated hardware to make it more efficient than the normal CPU
63. n the screen of the host computer via Ethernet This makes it possible to individually turn on and off all of Figure 4 1 Front Panel of control 22 Chapter 4 Testing and evaluation the control signals to try different set points and vary how fast valves open and close among other things It is also possible to choose if inputs should be taken from the machine or entered manually or by another VI To show that the old and new system behaves similarly it is important to ensure similar test conditions Small variations like build ups of frozen impurities or leaking gaskets influence the behaviour Even normal variations during production like the pressure in the storage for return gas impacts the measurement values The original plan was to run the two systems at the same occasion thereby under similar conditions and compare their behaviour In reality comparison between the old system and the final solution has not been possible during the same test run The main reason for this is that parts of the old system where no longer functional at the time of final verification In most cases however testing and comparison have been possible regarding individual features Even though the original test plan was not followed the testing has been enough to ensure a similar behaviour part by part and a continued development into a fully functioning system It would have been easier to show that the testing has been systematically done if all the data f
64. n u oe D i n e cum o ec e achik 0 025 10 025 0 025 ds os ox or Led OPEN CVi CLOS l A3 20588 Td Td Td j Automatic T106 CV111 I y 1 c A CV activate E f D D 0 0037 0 0037 10 007 20 0037 o Mv 108 MV108 cvi oo 38 5 J y 0 037 cvi I ake chai Meu V UPUM RUN TU eo Lok inkopptag e ox o ioo zz range2 range3 range4 range A A 0 6 62921 i i 2 Automat y y Numeric Numeric Numeric Numeric activate MV Purifier OPEN CV175 CLOSE yi CD E z E i f a 10 CV activat CV eee 2 m js o5 gn Tu PURIFIER ee a a 2 POE 2 lanual n D b b Mv MVE js E Mmvaso MiSo mvaa mvs v El vE s W o E dead band activate actiste activate overran ay ATs 20599 o 0 25 50 75 100 tomatic CV120 y g RE ox setpoint temp 2 setpoint temp 3 setpoint temp 4 setpoint temp OPEN CVI20 CLOSE El A A E e A e revu ee amem 40 40 28 28 28 27 p MVISP MV168 illo Ca ES Manual 0 123 45 67 8 910 y d y Tv ok e m ma 9 Lol 3 AT ARS desdband2 _dedbend3 desbandi dexiband Automatic CV121 I g A A al A Mv 169 MY159 72 F4v 172 OPEN CVI21 GOS y 1 yg 1 gi 1 3 08 emn Ma S PES Pe 2 ESS setpoint val ene e HO localAbortExe To be able to do all these tests the VI named control has a Front Panel with a lot of different options see Figure 4 1 When control is run as a standalone application on cRIO this Front Panel is not drawn anywhere However control can present the Front Panel o
65. nformation to UJ This is to ensure that control is not affected by network communication This is instead taken care of by the two Communication VIs They copy information between the libraries No variables are used for two way communication they are either written by UI and read by control or vice versa The communication VIs use a subVI to determine if the connection is active or not called ConnectionActive This is to ensure that the manual mode is discontinued if cRIO accidently looses contact with the host computer to minimise the risk that settings that have not been tested remain active when the operator can no longer control them 3 5 5 Target The VI named target starts all VIs run on cRIO CommunicationHighPriority CommunicationLowPriority and control The communication VIs are responsible for the contact with the host computer and are run as lower priority threads The most important VI is control which also has the highest scheduling priority Out of all the VIs control is the most complex one It keeps track of the state of the process and what is to be achieved with the help of state machines It is responsible for the evaluation of all the data collected about the process and the outputs Essentially control is the complete control system A description of the logic and states used in control is available in the next section 3 6 Control program logic The machine essentially consists of the Liquefier with supporting systems e g th
66. nk and the outlet valve from the low pressure side of the Liquefier to the compressors are opened START 2 compressor 1 The compressors are started by a current signal during one second V404 is opened START 3 compressor 2 Immediately after compressor start V100 begins to open and pressure starts building up at the Liquefier inlet PI101 START 4 open CV 103 to 0 7 bar When the pressure at PI101 reaches 8 bar the inlet valve to the turbines CV103 is slowly opened until the inlet pressure on the first turbine PI102 reaches 0 7 bar At this point the program goes no further until the turbines are spinning SIC100 and SIC101 are higher than 200 rps or the operator forces the program to continue START 5 open CV 103 to 1 bar This is the next security level where it is controlled that the turbines are spinning properly at one bar turbines should be reaching at least 1000 rps If the turbines do not spin properly the operator can take manual control or activate the stop program START 6 open CV 103 to T2 3200 rps When it is ensured that the turbines are spinning properly the inlet pressure is increased until the second turbine reaches the point where it is automatically controlled approximately 3000 rps The braking is reduced and the turbine speed is stabilised to the interval 4300 4500 rps START 7 static CV 103 T2 warm up The reduced braking and cooling by CV121 will cause an increase in turbine temperature which should be allowed to stabili
67. nt parameters should be possible for the operator to adjust during run time Collected data should be gathered in log files and presented in a graphical format Examples of values to measure and present level in machine tank pressure in machine tank pressure in storage pressure in buffer tank pressures PI 100 PI 101 PI 102 PI 103 PI 104 PI 150 PI 151 speed of turbines temperature of turbines temperatures TI 102 TI 104 TI 106 TI 111 temperature in compressors level in sump positions of valves CV 103 CV 111 CV 162 CV 175 vacuum pressure in cold box estimated production of LHe per hour A start program making the start process automatic A stop program making the stop process automatic A controller for the purifier controlling all valves associated with it in a predefined sequence and keeping track of the state Control of the valve connecting Liquefier and Purifier CV 175 making TIS 102 approach 28K Control of the flow to the machine tank via valve CV111 to keep TIS 106 at 12 K Control of MV 108 to enable bypass if TIS 111 is higher than 20 K Control of MV 110 to enable the connection of Liquid Nitrogen Cooling TIS 108 should aspire a predefined value Control of CV120 and CV121 to regulate turbine speed A user interface in LabVIEW from which the Liquefier can be supervised controlled started stopped etc Control should be possible from both Windows and Mac OS and from a local hos
68. of the system with plotter and logger Before the construction of the new system could begin the behaviour of the process and the old control programs were studied This was done with the help of two very simple VIs used Chapter 3 Method and implementation program description to plot and log measurements from the machine To be able to follow the variations in temperatures and pressures throughout the process caused by the current state of the Purifier and to get an idea of how exact measurement values that would be possible to obtain would prove very useful The data from these first VIs and the available documentation for the original control program where used to choose the set points that the new control program tries to reach and the conditions that would trigger specific actions These values have since been revised to improve the performance but these first measurements gave good starting points 3 4 Incremental replacement of control system parts The first very basic version of a control program had direct contact with the host computer At this point the structure of the logic including which states and conditions were to be used was finished only on paper In the first version a limited number of these states where used in the program and it processed few measurement values and governed even fewer output signals To test the responses from this control program a panel was added in which measurement data could be provided manually incl
69. ol system and the replacement of the old control step by step has been carried out successfully Testing outside the machine was done with the help of created input data for extreme conditions and by running the new system in parallel with the old one but without connection to the process to ensure similar behaviour All adjustments of the plant have been carried out during the ordinary planned stops Sometimes this has imposed limitations on the work and adjustments in the planned work sequence of the project but it has not caused any disturbances of the production Hardware and signals The choice of hardware was a prerequisite for this project and it has functioned as intended Almost all of the initial requirements and ideas have been carried out as planned with exception for requirements 11 and 12 The need for an AD converter module for measurements up to 24 V 11 was solved differently than planned with a simple home made analogue board using voltage division The ambition to avoid integrating 220V signals 12 into a system that contains only low voltage signals and is so easily accessible to the operator was only a limitation regarding the measurement of valve positions for the valves controlling the speed of the turbines CV120 and CV121 These two valves may never be fully closed as there should always be some brake applied At this point the control of these two valves is implemented but not tested or connected The signal showing when the
70. olab is located at the Department of Physics and provides liquid helium and nitrogen for coolant purposes Liquid nitrogen is extracted from the air through liquefaction and fractional distillation The abundance of nitrogen in air makes it comparatively cheap and with a boiling point at 77 K it is a good cooling agent For some applications this is not enough and helium is used Helium has the lowest boiling point of all elements at 4 22 K It also has a very low chemical reactivity as it is a noble gas These unique properties make it particularly useful in cryogenics As helium is not so easily accessible it is important to try to reuse it when possible The liquid gases are used mainly for experiments where cooling is needed There are many different options on how to supply the cooling When the fluid is used to cool a sample it can be evaporated to the atmosphere or recovered In the case with nitrogen which is comparatively cheap and originally extracted from the atmosphere there is little need for recovery Helium however is a limited resource and often accounts for a significant part of the cost for an experiment thus it should be the aim to recover the gas when possible particularly if larger amounts are used If the helium can be recovered it can be liquefied at the Kryolab plant and reused Most of the helium is used in closed cryostats from which the boiled off helium gas is collected The gas used by the helium liquefier plant primarily come
71. or the new system and explains some of them The full list of requirements can be found in Appendix E The basic structure for the logic is implemented as state machines the reasons for this design choice is described in 3 2 Sections 3 3 3 4 describe the process to implement the new control system and 3 5 3 6 provide an overview of programs and the implemented logic Section 3 7 describes the demands on the control regarding scan cycle times Section 3 8 describes the new alarm system and automatic shutdown that has been implemented The transformation of measurements and its limited accuracy is discussed in Section 3 9 The testing has been done continuously during the project and Chapter 4 provides examples of tests Chapter 5 contains the results comparing the system to the specification in Appendix E The project and its results are discussed in Chapter 6 and suggestions for future work can be found in Chapter 7 Chapter 2 Theory and preconditions 2 Theory and preconditions 2 1 Background This chapter describes the preconditions for the new control system In the early stages of the project time was spent on studying the plant and the liquefaction process as well as the existing control system and the hardware already chosen for the new implementation The following sections should be sufficient to understand this report Additional background theory is kept in appendices Appendix A contains a short description of the purpose of Kryolab
72. r when the machine is running It is well documented in the manual for the original machine TCF 20 The semiautomatic start and stop programs that should be replaced are governed by a PLC from SATT control and have been more difficult to retrieve code and documentation for The values for setpoints and conditions should be chosen as the ones used now at least initially It is important to measure these values as they might have been modified and are not necessarily the same as in the existing documentation 3 1 4 User interface software on host computer A user interface similar to the old operator panel should be used to present information about the current state of the process The current operator panel shows a simplified view of the piping of the system but gives a good overview Keeping this design as a basis for the new user interface will make it easier for the operator to know where to look for the information needed Chapter 3 Method and implementation program description A number of measurements control signals and the current state of the logic should be presented on this operator panel Buttons to start and stop the machine and manually regulate the control valves should also be implemented 3 1 5 Alarms and warnings The system for alarms and warnings is completely new so there is no previous system indicating reasonable threshold values This means that these must be chosen based on experience of normal process values There
73. returns see Figure 2 2 On the way to the machine tank the incoming high pressure gas is cooled in a series of heat exchangers After the first heat exchangers part of the incoming gas is led to the turbines and part continues through another series of heat exchangers to the machine tank Chapter 2 Theory and preconditions The part of the helium which goes through all the heat exchangers is approximately 7 K when it reaches the throttling valve Joule Thomson valve leading to the machine tank As the gas is expanded through this valve to 1 3 bars it becomes a mixture of liquid and gas at 4 2 K The liquid stays in the machine tank and the gas goes back to the low pressure return stream that passes through the heat exchangers The refrigeration required to keep the heat exchangers cool is provided by the use of turbines The part of the gas that is directed here exerts work on the turbines which causes it to cool This side stream of gas returning from the turbines is cold and added to the low pressure return stream The low pressure return stream consisting of a mixture of gas from the turbines and from the machine tank is used to cool down the incoming gas from the high pressure side Compressors High pressure Low pressure Heat exchangers g Turbines Throttling valve Machine tank Figure 2 2 Simplified view of the Liquefier 2 4 Current control systems The machine used at Kryolab today to liquefy heli
74. rom the tests during the development had been saved In some cases the old system stopped working before the new one was implemented In the case of the control of CV175 strange or even wrong behaviour in the old control was a reason for making that part of the implementation a priority In this case a comparison with the old system has not been possible and the behaviour has been compared to old documentation of the system Even if only a part of the tests during the development have been saved there is still a lot of documentation of different tests and problems that showed up over time All of that material is not included in this report but some examples are given in this chapter All of the plots are from the Front Panel of the plotter VI When comparing these plots it can be seen that the visualisation has undergone small changes in appearance over time This was to ensure that the most significant plots for the current tests could fit on the screen at the same time Only the most relevant plots for each example are presented rather than the full Front Panel 4 1 Examples of testing 4 1 1 Improvement CV 111 before and after This is one example of how it could look when a new control signal was connected In this case it was attempted to reduce the large variations in how open the valve to the machine tank was hoping that a more stable temperature would help increase production rate This means that the behaviour of the old and new system in
75. ront Panel of UI provides an intuitive user interface The design is based on the physical user panel of the TCF 20 see Figure 3 2 mainly for two reasons Firstly the similarities make it easier for the operator to understand and know where to look for different information Secondly the original panel shows how the different parts of the machine are connected and how the gas flow is moving in a good way The illustration of the machine pipes and valves is obviously very simplified but gives a good overview The U7 Front Panel see Figure 3 3 provides much more detailed information than the old control panel as the number of diodes and displays have been increased Chapter 3 Method and implementation program description Figure 3 2 Original user panel of TCF 20 There are also some elements not present on the old control panel that have been added This includes the boxes to the right see Figure 3 3 providing information about the state of the control system With some basic knowledge of the process it means that it is easy to see what the control system is trying to achieve how the gas is flowing and which temperatures should be reached Some elements have been added to the original sketch as a way of presenting information about parts of the system located a bit further away like the compressors storage buffer tank etc without loosing the general feel for how they are connected to the system The user interface also provides an opport
76. s from permanent pipes from research labs throughout the university Helium gas has very small molecules so it is hard to avoid leaks There are also customers who do not recover the helium gas Currently about 60 of the delivered helium is returned to Kryolab as gas Kryolab also purchases liquid gases to account for the helium that boils off and when demand is higher than the production capacity of nitrogen 3 34 Appendix B Helium Appendix B Helium Helium has the lowest boiling point of all elements at 4 22 K and is not found in solid state unless pressurized It also has a very low chemical reactivity as it is a noble gas These unique properties make it particularly useful in cryogenics Helium is not so easily accessible on earth even though it is the second most abundant element in the universe outnumbered only by hydrogen The low weight means that a large fraction of helium atoms evaporated from the cloud of earth and dust that would eventually form Earth Helium atoms can easily reach high speed and escape from the atmosphere so the volume percent of helium gas in the atmosphere is very low a few ppm and helium is not commercially obtained by distillation of liquid air Most of the helium used today is found in natural gas from which it is extracted using fractional distillation This helium has been created by radioactive decay of larger atoms and collected under rock formations over long periods of time alpha particles are in
77. se during five minutes when the inlet pressure is kept steady START 8 open CV 103 to T1 3200 rps The valve to the turbine inlet is further opened until turbine 1 is regulated to its optimum speed or the valve is fully open If the valve is fully opened the running state is immediately entered START 9 static CV 103 T1 warm up This state allows time for turbine one to warm up START 10 open CV 103 fully Once turbine one has been warmed up CV103 continues to open slowly until it is fully open When this is done the machine enters state RUNNING 14 Chapter 3 Method and implementation program description RUNNING At this point mainly MV108 and CV111 are controlled to stabilise the temperatures at TI111 and T1106 The set points should be chosen such that the production rate of liquid helium is as high as possible Set points can be manipulated in control STOP 1 closing CV 103 The inlet valve to the turbines is closed slowly as the pressure decreases so does turbine speed until the turbines eventually come to a halt STOP 2 close V 100 The inlet valve to the Liquefier is closed STOP 3 open CV 111 The valve to the machine tank is opened and cold gas can leave the system to the buffer during a five minute period STOP 4 close V 118 The outlet valve from the Liquefier to the turbines is closed STOP 5 compressor 2 The compressors are stopped STOP 6 compressor 1 Valve 404 is closed STOP 7 close buffer The valves to the buffer are
78. t with direct contact and over Internet with remote log in A locally placed panel should contain functions such as manual on off start stop of the Liquefier start stop of the Purifier start stop of compressors and V 404 manual control of CV 103 CV 111 CV 175 45 Appendix E System specification Revised requirements The original set of requirements was restructured and refined during the analysis phase and eventually became this list This list of final requirements was the one used for system evaluation and acceptance Overall limitations 1 Production should not be halted apart from normal stops 2 Replace parts of the control system incrementally 3 New parts must have similar characteristics to those replaced 4 Ensure testing outside the machine before integration 5 Revise control parameters as an optional improvement Hardware and signals 6 CompactRIO 9022 with additional IO modules is selected controller 7 LabVIEW 2011 is the controller software platform 8 cRIO can be connected to existing LAN with wired IP connection 9 f suitable the IO modules should be wired directly to the machine 10 AD converter module available for measurements max 10 V DC 11 AD converter module for measurements max 24 V DC is needed 12 Avoid if possible to use any of the 220 V AC signals 13 Need to drive 24 V output signal 14 Need to sink 24 V output signal 15 There are limited number of IO ports and only the most
79. th LabVIEW and National Instruments hardware was provided by Payman Tehrani at NI support centre I came in contact with Eva Andersson at Studieverkstaden far too late in the writing process but she has never the less given me new insights about writing and much appreciated support I would also like to thank Anita Brodin for proofreading particularly the material about thermodynamics And last but not least a special thank you to Jens Gulin for his patience and encouragement during this project and willingness to discuss everything related to it HI Table of contents T Introduction seo etd e ette diet eee che d eder ut te E alae toad at 1 L1 Background neue eet eie Fog dit sne ie cesses HER de eevee 1 12 lt Waskei keine ie i ore e dee doe Seo ber e en 1 133 Outline of the projects anas eere ere Ote bench gi egere edu 1 VA Report OVETVIEW uen ee eee head eel ee ee eet tete esi ee stewed 2 2 Theory and preconditiohs 6 epe epe eret ritiene Ree cot e o neu e eee Rr eene gucci 3 2 l Background eie itt e teet tie n ter hac ERES lates tas UE EE A ERU neben d 3 2 2 Pl ntoverview eeu e eed E RR e e RU es 3 2 3 Laquefaction DEOC6SS a eee ciere reet rece Du tree e esie NNE 3 2 44 Current control Systems oe eR e ae Re I a ERU Po Re eS 4 2 5 CompactRlO 2 crece ect eere ro e esee eee etre e leto 4 2 6 LabVIEW eee eb et ete e Eee FU e cele toes o be De e ed EE 5 2 5 Realtime inepti REHE iu TAR NINE e R NA A
80. this list will be empty EGG Ele Edit view Project Operate Tools Window Help TOR 2 a Purifier Pagei Page2 Paces cooling water 16 1 compressor temp 37 0 W118 P storage 91 9 compressor y 5 Puget PI 100 wat Cooldown 1 ruez cooldown 2 O Purity LJ Regenerate standby MV169 PI 101 3 MV160 Mv172 29 gas bag F PI 150 ET to atmosphere Liquefier warr START RUNNING sroe not connected Mv108 Alarms and Warnings Liquefier is running Cv 12 manual y Cv 111 m rate P machine tank 0 315 V machine tank 531 amp Server c83 249 221 213 bredband comhem se a pressure control panel B I 91 buffer TIS 106 12 6 d Ot El ii MV CV 119 i LIN cooling to gas bag auro Edi SIC 100 Speed Turbine 1 2000 4000 6000 Cv 103 PI 102 Dae T5 109 Temperature Turbine 1 D Teeter dar agera o 20 40 60 80 100 SIC 101 Speed Turbine 2 1 PI103 0 2000 4000 6000 2 03 TS 110 Temperature Turbine 2 Peeepeeegeespenupered 0 20 40 60 80 100 zoo ve Em lt 1 Figure 3 3 Front Panel of UI vi Diodes indicating the position of a solenoid valve are red when closed and green when open see Figure 3 3 and 3 4 A valve that is normally closed is depicted as dark red in its passive closed state and light green in its active
81. tion of these valves but should it be necessary to replace these boards in the future the activation of this code could be useful It is however not tested and built according to the description in the user manual for TCF 20 in particular the set point values for different actions must be chosen in accordance with the behaviour of this specific machine The structure was designed to allow an easy addition of logic for LIN cooling and a basic control for LIN cooling has been added after the project Further development of this to optimise the production rate could be of interest To facilitate the use of Liquid Nitrogen for pre cooling the installation of a permanent pipe from the Nitrogen Liquefier tank could be highly useful Worth noting though is that at this point in time the plant has no difficulties liquefying all return gas even without this increase in efficiency Continued work to increase the fraction of the helium sold which is returned as return gas is advisable Many of the control parameters in control can be further optimised The system is running and working well at this point but the behaviour can obviously always be improved As it has been the goal to provide measurements comparable to the old system the original transformations have in many cases been copied and recalibrations have not been done This makes it easier for the operator to recognise patterns and deviances based on experience In a longer perspective it would be better i
82. ts The flow of cool gas should not be restricted because of temporary low temperatures it should be kept steady until the Purifier is genuinely cold In Pre Purify CV175 is consequently kept open and the state is time limited to 30 s Should the temperatures during this time get to high Purify Reset is entered from where it automatically goes back to Pre Purify Essentially these two extra states function like a timer where Purify Reset is just passed through resetting the time temperatures have been stable Purify is left when the buffer of helium gas is full when there is too much build up of impurities in the heat exchanger or when the operator turns of the Liquefier or in some other way performs an action that forces the state machine 92 4 118 3 3 V10C the heat exchanger is warmed to make any frozen impurities melt and fall into e MvV160 MV169 PI100 0 0 PI101 9 1 the sump At this point MV168 CV162 E zs Regenerate and Controlled Stop When no more gas should be purified is closed and warm gas is taken from the sste Li 201 high pressure side of the Liquefier and ism win run through the secondary side in the l reverse direction via valves MV172 and mee m MV174 see Figure 3 10 As the Purifier is warmed the gas on the wie e oo ME primary side will expand and is led to ag PI 151 115 0 not connected i MVv108 the gas bag via MV163 This is aborted u
83. uding a switch that decided if real live data or these manual inputs should be used Possibilities to force the state machine into a desired state were also added Using this it was possible to test the program s reaction to different measurement signals and check that the behaviour was correct before the wire with the old control signal was replaced by a signal from cRIO The new piece of logic implemented could then be monitored from the plotter VI checking that the overall behaviour was stable and similar to the old control or better In this way the control of the different relays and motors was gradually transferred to the new control unit in an iterative process Each new control signal was verified first with a set of different inputs manually then on real data comparing outputs from the two systems and lastly with connection to the real process 3 5 Program structure The program consists of several VIs and a couple of libraries providing variables used as a way of transmitting data from one VI to another The following sections describe each of the VIs and their purpose The sketch is intended to give an overview of the program structure see Figure 3 1 Chapter 3 Method and implementation program description PC cRIO ul target MI MI Shows the current state of the control program control Displays measurement values subi collected from the system Keeps track of the state of
84. um return gas is from 1985 The process for cooling helium is the same in newer machines but the control system is outdated and is in some respects not working properly The system is well documented in the manual and drawings of the analogue control circuitry are available During the nineties a second control system was added It was implemented on a PLC from SattControl making the previously entirely manual start and stop sequences semiautomatic Data from this system is fed to a computer which presents key data on a web page The new control system replaces this system and many of the analogue boards 2 5 CompactRIO The use of CompactRIO from National Instruments was a precondition and the selection process was not included in this project Some comments on the experiences with CompactRIO can be found in Chapter 6 and 7 Chapter 2 Theory and preconditions The CompactRIO 9022 is modular system for control and data acquisition The controller has a real time processor which makes it possible to run deterministic programs The IO modules come in many different versions and are connected to the controller via a backplane The core of this backplane is a reconfigurable hardware an FPGA which can be used for signal processing and control It has direct contact with each IO module and provides contact with them in LabVIEW Real Time This can be configured automatically activating scan mode 1 The use of the FPGA for local control can be valuable
85. unity for the operator to gain some manual control over the system The control valves can be regulated manually instead of by the control system however it is not recommended unless the process for some reason shows unexpected behaviour The build up of frozen impurities or aging of gaskets causing small leaks are two examples of things that can cause a change in process behaviour This will cause a reduction in production rate as the control system parameters have been chosen during other circumstances In such a case a fixed valve position provided by the operator might be a better choice than the one provided by the control system and make it possible for the machine to keep producing but at a slower rate than normal until a stop for service is possible The possibility to manually control valves during run time also increases the possibility to identify problems while the machine is running and has made it simpler to find out if unexpected behaviour by a specific valve was caused by a relay the valve itself or maybe an intermittent cable and if it was necessary to shut down the machine to fix the problem or not In the upper right corner is a box related to the Liquid Nitrogen pre cooling It is possible to select the state of the cooling using a drop down list There is also a little box 10 Chapter 3 Method and implementation program description displaying a list of alarms and warnings placed at the bottom left Under normal conditions
86. valves have reached their closed position is a 220V signal and needed for the control At this point the old analogue board is still in use 27 Chapter 5 Results Control program software on cRIO The manual start and stop sequences have been automated in accordance with the descriptions in the manual for the original plant This has facilitated the work with the plant Logic for LIN cooling has been added based on operator experience The current control program is completely standalone and not in any way dependent on the user interface presented on a separate computer If contact is lost with the user interface during manual control of one or more parameters automatic control is resumed Frequency sampling of turbine speed 20 has not been implemented The sampling rates used in the rest of the project are too low but it should be possible to implement this functionality on the backplane FPGA It was decided to continue the usage of the old analogue board that converts the speed to a voltage This kind of unit is neither expensive nor hard to find and time was better invested in other parts of the project User interface software on host computer The new user interface is similar in appearance to the original operator terminal but provides much more detailed information and significantly improved possibilities to both supervise and control the process Additional data display All key measurements are continuously presented in graph
87. was a good environment to work in although it took some time finding the most suitable language features and determining which program design patterns that where best for this language It is very easy to get started and construction of simple implementations is intuitive The problem is that those implementations do not always scale well A lot of time was spent in the early stages of the project trying to figure out a good format for the program to make it easily extendable and also testing different kinds of functionality to see what was possible to do and what was not This exploration of possibilities was necessary to do before the basic structure of the implementation could be chosen To invest some time in the basic courses early on is probably good for beginners A wider knowledge of the available functions and to have someone to discuss format with would have been beneficial There are many things that are already implemented but it is not so easy to find if you do not know exactly which subVI or example you are looking for Better previous knowledge of LabVIEW would have speeded up the process considerably The discussion board on the NI website is an excellent source of information most problems that you might experience have probably already been posted there by somebody else usually with good replies and solutions One of the main drawbacks of being only one student in the project was the lack of someone to discuss problems with Not necess
88. when high speed is important As the controlled process is fairly slow and the computing power more than enough it has not been necessary to use it Communication with the host computer is done via Ethernet The host is the computer used to deploy the program on cRIO and to show the user interface as cRIO has no display 2 6 LabVIEW The programming is done in LabVIEW which is a programming environment for a graphical programming language The code makes it easy to follow the dataflow which is represented as wires All LabVIEW code is inherently parallel and the order of execution is dependent on the dataflow A LabVIEW program is called a Virtual Instrument a VI Each VI has a Front Panel and a Block Diagram The Front Panel displays a graphical representation of inputs and outputs e g buttons and meters that can be used by an operator A VI can also be used as a subVI within another VI that then controls it via the inputs and outputs In the Block Diagram most of the programming is done by connecting different function nodes and subVIs with wires 2 7 Real time It is important to create a reliable control of the plant The user interface will be implemented and run on a Windows computer as cRIO has no display but the actual control must be implemented as a standalone application on cRIO In this case it means that it must work regardless of network connection The communication with the user interface can not be allowed to delay the contro

Industrial Electrical Engineering and Automation

Contents

Download Pdf Manuals

Related Search

Related Contents