PID Control of Valve Sticktion and Backlash - Greg McMillan Deminar Series
Upcoming SlideShare
Loading in...5
×

Like this? Share it with your network

Share

PID Control of Valve Sticktion and Backlash - Greg McMillan Deminar Series

  • 6,242 views
Uploaded on

Greg McMillan shows the limit of PID control options to help with sticky valves and valves with excessive deadband in this live deminar (demo/seminar)....

Greg McMillan shows the limit of PID control options to help with sticky valves and valves with excessive deadband in this live deminar (demo/seminar).

Recorded video at: http://www.screencast.com/t/OTJjYjE1Nz

More in: Self Improvement
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
6,242
On Slideshare
5,981
From Embeds
261
Number of Embeds
4

Actions

Shares
Downloads
281
Comments
0
Likes
0

Embeds 261

http://www.emersonprocessxperts.com 247
http://www.slideshare.net 11
http://www.linkedin.com 2
http://translate.googleusercontent.com 1

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. Interactive Opportunity Assessment Demo and Seminar (Deminar) Series for Web Labs – PID Control of Valve Sticktion and Backlash April 21, 2010 Sponsored by Emerson, Experitec, and Mynah Created by Greg McMillan and Jack Ahlers
  • 2. Welcome
    • Gregory K. McMillan
      • Greg is a retired Senior Fellow from Solutia/Monsanto and an ISA Fellow. Presently, Greg contracts as a consultant in DeltaV R&D via CDI Process & Industrial. Greg received the ISA “Kermit Fischer Environmental” Award for pH control in 1991, the Control Magazine “Engineer of the Year” Award for the Process Industry in 1994, was inducted into the Control “Process Automation Hall of Fame” in 2001, and was honored by InTech Magazine in 2003 as one of the most influential innovators in automation. Greg is the author of numerous books on process control, his most recent being Essentials of Modern Measurements and Final Elements for the Process Industry. Greg has been the monthly “Control Talk” columnist for Control magazine since 2002. Greg’s expertise is available on the web site: http://www.modelingandcontrol.com/
  • 3. The Latest on Valve Response and Smart Positioners Royalties are donated to the University of Texas Research Campus for Energy and Environmental Resources for Development of Wireless Instrumentation and Control
  • 4. Why?
    • My personal calling is to develop and communicate a deeper understanding of instrumentation, valves, and process control. My 20 books, 50+ articles and papers, and 100+ “Control Talk” Columns are a reflection of this quest.
    • Hopefully the unconventional cartoons, “Top Ten” lists, and musings provide some insights and breaks down barriers, and opens up minds as only humor can do.
    • However all of these literary means are static. I am ready to take the web site, which is an attempt to provide a more dynamic sharing of knowledge, to the next level.
    • My dream is that users would have worldwide access to an interactive learning tool that offers the control room experience to discover, innovate, and generate (DIG) process control improvements (PCI) to make plants more efficient and productive.
    • The interactive tool would be virtual plant labs of generic loops and common unit operations available anywhere anytime over the web featuring:
      • A non system specific interface by means of control room type operator graphics
      • The power of the DCS and in particular the PID controller
      • Performance metrics (e.g. integrated and peak errors for generic loops, number of communications for wireless measurements, accumulated travel for valves, and COGS and profit for specific unit operations)
  • 5. Top Ten Benefits of Membership in Adorable Automator’s Association
    • (10) Free admission to the Adorable Automator’s Hall of Fame - the hottest ticket in town
    • (9) Discounted admission to the Cinemax Film “Famous Control Scenes” - 3D trend charts
    • (8) The automator’s alternative to “Fruit of the Loom” underwear – “Geek in the Room”
    • (7) “Fuzzy is Cute” buttons - fuzzy logic memorabilia
    • (6) Dating service - “let us automate your pleasures”
    • (5) Simulation services - “almost as good as the real thing”
    • (4) Calendar - “Adorable Automators of the Month”
    • (3) Souvenir T-Shirts - “My boss went to the ISA executive meeting and all I got was this shirt”
    • (2) Hardhats that double as extreme bull riding helmets
    • And the Number 1 Benefit:
  • 6. Top Ten Benefits of Membership in Adorable Automator’s Association
    • (1) Travel discounts to exotic plants - ride elephants in Thailand
    Sheldon Lloyd – Developer of I/P and Positioner - Retired VP of Technology for Fisher Controls See Control Magazine May 2007 Control Talk “The Best of the Best – Part 5 http://www.controlglobal.com/articles/2007/127.html
  • 7. Loop Lab02 Demo 1
    • Objective – Show access to cascade loop lab setup and how to make load upsets to see PID response
    • Activities:
      • Show access to Cascade Loop Lab02 user interface
      • Show access to PID faceplate and detail
      • Show access to “Process History View” trend chart
      • Make load change to show cascade response by putting primary PID momentarily in manual and changing its PID output (e.g. 50% to 60%)
      • Show access to valve, process, and measurement parameters for sensitivity, noise, and dynamics
      • Put secondary PID in AUTO mode
      • Make load change to show secondary response by putting secondary PID momentarily in manual and changing its output (e.g. 50% to 60%)
  • 8. Question
    • What oscillation cannot be eliminated by PID controller tuning?
    • What is the most common cause of a sustained oscillation?
    • What response problems are not seen in open loop tuning tests?
    • What creates a deadtime inversely proportional to controller gain?
    • What problem became rampant in the 1980s?
    • What is the biggest hidden menace to loop performance?
    • Answer : ?
  • 9. Question
    • What oscillation cannot be eliminated by PID controller tuning?
    • What is the most common cause of a sustained oscillation?
    • What response problems are not seen in open loop tuning tests?
    • What creates a deadtime inversely proportional to controller gain?
    • What problem became rampant in the 1980s?
    • What is the biggest hidden menace to loop performance?
    • Answer : Valve Backlash and Sticktion
  • 10. Loop Lab02 Demo 2
    • Objective – Compare base case to scenario with 5% stick-slip on response of Secondary PID in AUTO mode
    • Activities:
      • First look at load response of PID for base case
      • Keep the secondary PID in AUTO mode
      • Increase valve stick-slip to 5%
      • Make load change to show secondary response by putting secondary PID momentarily in manual and changing its output (e.g. 50% to 60%)
  • 11. Valve Backlash (Deadband) and Sticktion (Stick-Slip) dead band Deadband Stick-Slip is worse near closed position Signal (%) 0 Stroke (%) Digital positioner will force valve shut at 0% signal Pneumatic positioner requires a negative % signal to close valve Valve deadband and stick-slip is greatest near the closed position Deadband is 5% - 50% without a positioner ! Deadband is the change in signal to reverse direction of stroke - principal sources are links and mechanisms for rotary valve actuation Stick-slip is the smallest possible change in valve stroke – principal sources are friction in packing and friction in closure element (trim) seal (can be seen as shaft windup and breakaway in rotary valves) Sensitivity limits associated with piston actuator and pneumatic positioners create a staircase response similar to stick-slip
  • 12. Loop Lab02 Demo 3
    • Objective – Compare slow versus fast tuning effect on Secondary PID AUTO Mode Response to 5% Stick-Slip
    • Activities:
      • First look at limit cycle in secondary PID for 5% stick-slip for fast tuning
      • Increase secondary PID reset time (e.g. 2 to 10 sec) to show the effect of slow tuning
  • 13. Direct Connection of Piston Actuator to Rotary Valve Less backlash but wear of piston O-ring seal from piston pitch is concern See “ Best Practices for Valve Performance” slide for best solution to minimize backlash and stick-slip
  • 14. Link-Arm Connection of Piston Actuator to Rotary Valve Significant backlash from link pin points 1 and 2 See “ Best Practices for Valve Performance” slide for best solution to minimize backlash and stick-slip
  • 15. Rack and Pinion Connection of Piston Actuator to Rotary Valve Stick-slip from rack and gear teeth - particularly bad for worn teeth See “ Best Practices for Valve Performance” slide for best solution to minimize backlash and stick-slip
  • 16. Scotch-Yoke Connection of Piston Actuator to Rotary Valve Lots of backlash from slot See “ Best Practices for Valve Performance” slide for best solution to minimize backlash and stick-slip
  • 17. Loop Lab02 Demo 4
    • Objective – See the effect of secondary PID I-deadband on limit cycle in PID AUTO mode for 5% stick-slip
    • Activities:
      • First look at limit cycle for 5% stick-slip for slow tuning
      • Return secondary PID to original reset time (e.g. 10 to 2 sec)
      • Set secondary PID I-deadband = 4%
      • Discuss function and estimation of I-deadband setting
  • 18. Installed Characteristic (Linear) Valve pressure drop ratio (  P R ) for installed characteristic: Characteristic 1:  P R  0.5 Characteristic 2:  P R  0.25 Characteristic 3:  P R  0.125 Characteristic 4:  P R  0.0625
  • 19. Installed Characteristic (Equal Percentage) Valve pressure drop ratio (  P R ) for installed characteristic: Characteristic 1:  P R  0.5 Characteristic 2:  P R  0.25 Characteristic 3:  P R  0.125 Characteristic 4:  P R  0.0625
  • 20. Loop Lab02 Demo 5
    • Objective – See the effect of secondary PID I-Deadband on limit cycle in PID CAS mode for 5% stick-slip
    • Activities:
      • First look at elimination of limit cycle by secondary PID I-deadband for 5% stick-slip for AUTO mode
      • Put secondary PID in CAS Mode
  • 21. Real Valve Rangeability Minimum fractional flow coefficient for an equal percentage trim and stick-slip: Minimum controllable fractional flow for installed characteristic and stick-slip: C xmin  minimum flow coefficient expressed as a fraction of maximum (dimensionless)  P r  valve pressure drop ratio (dimensionless) Q xmin  minimum flow expressed as a fraction of the maximum (dimensionless) R v  rangeability of control valve (dimensionless) R  range of the equal percentage characteristic (e.g. 50 - 200) X vmin  maximum valve stroke (%) S v  stick-slip near closed position (%)
  • 22. Loop Lab02 Demo 6
    • Objective – See effect of I-deadband in both primary and secondary on PID CAS Mode for 5% stick-slip
    • Activities:
      • First look at creation of limit cycle in secondary PID for 5% stick-slip for CAS mode
      • Set primary PID I-deadband = 4%
  • 23. Effect of Tuning on Limit Cycle from Stick-Slip in Self-Regulating Processes A o = S s  K o K o =  K mv  K pv  K cv T o = 4  T i  [1  ( K o  K c )  1] A o = oscillation amplitude (%) S s = valve stick-slip (%) K o = open loop gain (%/%) (more commonly known as process gain) K mv = manipulated variable gain (valve gain) (e.g. kg/sec per %) K pv = process variable gain (unit operation gain) (e.g. o C per kg/sec) K cv = controlled variable gain (measurement gain) (e.g. % per o C) T i = controller integral time (sec) T o = oscillation period (%) http://www.controldesign.com/articles/2003/164.html “ What’s Your Flow Control Valve Telling You?”, Control Design, May 2003
  • 24. Loop Lab02 Demo 7
    • Objective – See effect of valve deadband on CAS Mode with no I-deadband in either PID for 5% valve deadband
    • Activities:
      • First look at cascade loop response with 5% valve stick-slip with existing I-deadband in both PID for CAS Mode
      • Set valve stick-slip = 0% and valve deadband = 5%
      • Set I-Deadband = 0% in both PID
  • 25. Effect of Step Size on Small Valve Response Results show sensitivity limitation of piston actuators and pneumatic positioners
  • 26. Loop Lab02 Demo 8
    • Objective – See effect of valve deadband on AUTO Mode Response with no I-deadband in either PID
    • Activities:
      • First look at cascade loop response with 5% valve deadband with the I-deadband = 0% in both PID for CAS Mode
      • Put secondary PID in AUTO mode
  • 27. Best Practices for Valve Performance
    • Actuator, valve, and positioner package from a control valve manufacturer
    • Digital positioner tuned for valve package and application
    • Diaphragm actuators where application permits (large valves and high pressure drops may require piston actuators)
    • Sliding stem (globe) valves where size and fluid permit (large flows and slurries may require rotary valves)
      • Best rotary is low seal friction segmented V-ball or contoured butterfly with short splined shaft connection and rotary positioner feedback and actuation (original valve design is for throttling (process control) rather than on-off or isolation service (discrete control)
    • Low packing, sealing, seating, and actuator friction
    • Booster(s) on positioner output(s) for large valves on fast loops (e.g., compressor anti-surge control)
    • Valve sizing for a throttle range that provides good linearity [4]:
      • 5% to 75% (sliding stem globe)
      • 10 o to 60 o (segmented V-ball)
      • 25 o to 45 o (conventional butterfly)
      • 5 o to 65 o (contoured and toothed butterfly)
    • Online diagnostics and step response tests for small changes in signal
    • Dynamic reset limiting (FRSPID_OPTS) using digital positioner feedback
  • 28. Loop Lab02 Demo 9
    • Objective – See effect of valve deadband on AUTO mode response of high gain PID for an integrating process
    • Activities:
      • First look at single loop response with 5% valve deadband with the I-deadband = 0% in both PID for AUTO Mode
      • Change to secondary process type to integrating . To prevent instability, the order of following steps is important:
        • Check secondary PID is in AUTO mode
        • Set the Secondary Process Lag 2 = 20 sec
        • Set the Secondary PID reset time = 100 sec
        • Set the Secondary PID gain = 5
        • Set the Secondary Process Type = Integrating
  • 29. Top Ten Signs Your Control Valve is an On-Off Valve in Disguise
    • (10) Valve body looks suspiciously like the block valve next to it
    • (9) Actuator looks suspiciously like the one on the SIS valve
    • (8) Process engineer is seen going to lunch with piping valve supplier
    • (7) The valve deal is a steal
    • (6) Your flow is on-off
    • (5) Positioner measures actuator shaft instead of ball or disk stem position for feedback
    • (4) Oscillation amplitude exceeds largest historian compression setting
    • (3) 360 o feedback in your loop becomes 360 0 feedback in your performance review
    • (2) The valve package is nicknamed “Sloppy Joe”
    • (1) No valve leakage even with a controller output of 40%
  • 30. Loop Lab02 Demo 10
    • Objective – See effect of valve deadband on AUTO mode response of low gain PID for an integrating process
    • Activities:
      • First look at single loop response with 5% valve deadband of high gain PID for an integrating process
      • Change secondary PID gain from 5 to 1
  • 31. Effect of Tuning on Limit Cycle from Valve Deadband in Integrating Processes A o = DB  K c T o = 5  T i  [ 1  2  (K c ) 0.5 ] A o = oscillation amplitude (%) DB = valve deadband (%) K c = controller gain T i = controller integral time (sec) T o = oscillation period (%) http://www.controldesign.com/articles/2003/164.html “ What’s Your Flow Control Valve Telling You?”, Control Design, May 2003
  • 32. Summary
    • Valve Stick-Slip:
      • Adds deadtime inversely proportional to PID output rate of change
      • Creates a limit cycle (persistent constant amplitude oscillation) in any PID with integral action
      • Slower tuning increases period but has no affect on amplitude
      • I-Deadband must be added to both secondary and primary PID to eliminate limit cycle in cascade control system
    • Valve Deadband:
      • Adds deadtime inversely proportional to PID output rate of change
      • No limit cycle for a single loop on a self-regulating process
      • Creates a limit cycle in cascade loops
      • Creates a limit cycle for integrating process (e.g. level and batch temperature) and runaway process (e.g. Exothermic reactor)
      • Limit cycle amplitude & period are inversely proportional to PID gain
  • 33. PID Solutions
    • For self-regulating processes
      • I-deadband eliminates limit cycle but creates offset
    • For all types of processes
      • Deadband and stick-slip compensator or lead-lag on PID output can reduce deadtime and increase frequency of limit cycle but increases valve travel (wears out packing and seals)
    Best solution is control valve selected per “ Best Practices for Valve Performance”
  • 34. Summary of Demos Demo Sec PID Mode Sec PID Tuning Sec PID I-deadband Prim PID I-deadband Valve Stick- Slip Valve Dead-band Process Type 1 CAS FAST 0% 0% 0% 0% Self-Reg 2 AUTO FAST 0% 0% 5% 0% Self-Reg 3 AUTO SLOW 0% 0% 5% 0% Self-Reg 4 AUTO FAST 4% 0% 5% 0% Self-Reg 5 CAS FAST 4% 0% 5% 0% Self-Reg 6 CAS FAST 4% 4% 5% 0% Self-Reg 7 CAS FAST 0% 0% 0% 5% Self-Reg 8 AUTO FAST 0% 0% 0% 5% Self-Reg 9 AUTO HI GAIN 0% 0% 0% 5% Integ 10 AUTO LO GAIN 0% 0% 0% 5% Integ
  • 35. Help Us Improve These Deminars! WouldYouRecommend.Us/105679s21/
  • 36. Join Us May 12, Wed 1:00 CDT
    • PID Control of Slow Valves and Secondary Loops
    • Look for a recording of Today’s Deminar later this week at:
    • www.ModelingAndControl.com
    • www.EmersonProcessXperts.com
  • 37. QUESTIONS?