Automated Methods to Sustain & Improve Control System Performance

51st ISA Power Industry Division Symposium18th Annual Joint ISA POWID/EPRI

Controls & Instrumentation ConferenceJune 8-13, 2008, Scottsdale, Arizona 1

George Buckbee, P.E.

Automated Methods to Sustain & ImproveControl System Performance

Copyright 2010 ISA. All Rights Reserved.

Distributed with permission of author(s) by ISA 2010 Presented at ISA Automation Week 2010; http://www.isa.org

ISA Automation Week 20102#

Agenda

• What is Control System Performance?

• Why Does it Matter?• How to Measure It?

• Case Study• Guidance for Capturing, Tracking, and Managing




What is Control System Performance?

• What is the Purpose of the Control System?– Reduce Variability– Stabilize the Process– Reduce Impact of Upsets– Operate Efficiently

– At Peak Efficiency !

– Improve Quality– Reduce Waste– Safety– Environmental




What is Control System Performance?

• How Well Does the Control System Do Its Job?

Measure Decide Action

Noise

Variability Tuning

Controller Design

Hysteresis

StictionNon-Linearity

Operator Changes

Loop notIn Normal

Disturbances

Valve atLimit

Interactions




Typical Control System Performance

Typical Statistics• 10% - 90% Control Loops in Manual• 30% Control Valves with Problems• 30% Loops-Tuning is completely wrong• 40% Loops Oscillating• Improper PID Loop Configuration

– 30% DCS Systems– 95% PID on PLC’s

• 85% Sub-Optimal Tuning• 75% Control Loops increase variability




The Unfortunate Truth:

• Most Plants Do Not Measure These Things, so…• Most Plants Do Not Know How Well They Are Doing !

AND….

• Most Plants Do Not Know When Something Goes Wrong




What is Performance Supervision?

Supervising and Managing the Performance of a System:

– Paying attention to the right things– Assessing the Situation– Prioritizing According to Economics– Taking the right corrective Actions– Tracking Results




Why Does it Matter?

• These Metrics:– Significant Oscillations– Variability– Service Factor– Opportunity Gap– Controllability– Valve Travel– …

• Directly Impact:– Energy Cost– Quality– Safety/Environmental– Cost– Waste, Recycle– Maintenance Costs– …




How to Measure Control System Performance?

• Service Factor– Time in Normal Mode– Valve at Limit– PV at Limit– Within x% of Setpoint

• Controller Tuning Indices– Harris Index– IAE– Integrated Error, Not Abs




How to Measure Control System Performance?

• Significant Oscillation

• Valve Travel




Case Study #1

• Combined-Cycle Plant• 1000 Control Loops• 2 Control Engineers




Case Study 1, Continued

Performance Measurement

• 20% Valves have mechanical problems– Stiction– Leaking

• 15% individual control loops not tuned properly

• 51% control loops oscillating– Investigate Significant Oscillations

• >50% control loops interact




Case Study #1

Methods & Metrics to Target Improvements

• Treemap• Biggest Payback (Pareto)• Valve Travel & Reversals• Variance / Variability• Oscillation Significance




Case Study #1

Step 1 - Targeted Variability Reduction




Case Study #1

Step 2 – Dynamic Response

Reduced Risk of Trip




Case Study #1

Business Impact

• Stabilized Operation Under All Conditions:– Steady Load– Load Increase– Load Decrease

• Reduced Risk of Trip• 2MW increase in Steam Turbine output




Case Study #2

• Gas Plant

• Located in Kuwait

• Major Process Difficulties– Traced to Steam System




Case Study #2Gas Plant, Kuwait

• 15% of control loops were oscillating.

• 33% had controller output at limit.

• 27% had excessive valve travel.

• 20% involved high variability.




Case Study #2Continued

• Gas Plant, Kuwait– Process Upsets Due to Steam System Upset– Steam Upset from Excess O2 Monitor– Stack Upset from Pressure Relief

Steam header

Separations Plant

O2




Case Study #2KPI Comparison Before and After

Key Performance Before After

Avg.% to Threshold 72.06 47.32

Variability – KPIVariability - %TVariability - %E

4.423147.4%

73.71%

2.59386.42%43.21%

Avg.absolute Error – KPI Avg.absolute Error - %TAvg.absolute Error - %E

1.33826.76%13.38%

0.532510.65%

5.325%

Output Std.Deviation -KPIOutput Std.Deviation - %TOutput Std.Deviation - %E

0.8383 16.77% 8.383%

0.3103 6.206 3.103%

Valve Travel – KPIValve Travel - %TValve Travel -%E

34550 138.2%

69.1%

344813.79% 6.895%

Valve Reversal - KPIValve Reversal - %TValve Reversal - %E

25220 50.43% 25.22%

6726 13.45% 6.726%




Guidance for Capturing Control System Performance Data

• Capture Data in Real-Time• Metrics Should be:

– Measurable– Meaningful– Actionable

• Diagnostics Built on These Metrics– Instrument & Valve Failures– Loss of Control– Cause of Oscillation

• Historize, Track, Drill Down, etc.




Guidance for Tracking & Managing Control System Performance Data

• Match Business Objectives with Metrics• Cover Each Unit Fully

– Most Upsets Come from Unexpected Places

• Integrate the Information Into Business Processes• Follow-Up• Document the Benefits




Questions & Answers



Automated Methods to Sustain & Improve Control System Performance

Business

abs isa automation week

results isa automation

wrong isa automation

variabilityisa automation

managingisa automation

permission of authors

control valves

individual control loops