01 February 2004

Dynamic simulation proves pressure control system

Flexibility in pipeline operations introduces significant performance demands on its entire pressure control system.

By Roger Shirt, Lawrence Neumeister, and Jack Broyles

Pressure control valves comprise 90% of the final control elements in Enbridge Pipeline Inc.'s (EPI's) 15,000-kilometer liquid petroleum pipeline network in North America. Responsive and robust line pressure control is essential for avoiding preventable line shutdowns due to pressure excursions and for minimizing periods of excessive throttling. Service conditions for the majority of existing pressure control valves have evolved since their installation. As well, the mechanical performance capability of the existing control valves is largely unknown.

Pressure control systems involve several components that influence overall system performance. These include mechanical performance of the installed valve assembly and actuator, hydraulic behavior of the control valve and surrounding piping, and control system tuning. Experience has shown that engineering, operating, and maintenance personnel often implement modifications without full appreciation of the degree of interaction amongst pressure control valve system components.

Here is a systematic, multistep approach to optimizing the overall performance of existing control valve installations in liquid pipeline service. A control valve installation that has undergone this process is certified.

An engineered approach to control valve optimization is necessary to achieve the best performance for the current service conditions.

Selecting oversize is wise

EPI operates a lengthy system of batched pipelines in North America that transport a broad range of liquid petroleum products. Ninety percent of its pump stations employ pressure control valves as the final control element. Pressure control valves in mainline service are typically greater than 12 inches. Due to the considerable capital investment associated with each control valve and actuator installation, a strong incentive exists to maintain and optimize their performance. Another unique factor in pipeline engineering arises from pumping energy costs making up a significant portion of operating expenditures. Historically, this has led Enbridge to select oversized control valves.

Responsive and robust line pressure control in pipeline operations is essential for avoiding preventable line shutdowns caused by pressure excursions. Flexibility in pipeline operations requires frequent changes in flow rates and delivery sites. These flexibility requirements, in addition to batch transport of widely differing products, introduce significant performance demands on the entire pressure control system.

Control valve commissioning usually consists of specification of service conditions, control valve sizing (by valve manufacturers), and field tuning of the controller during start-up. Closed-loop performance of the valve will typically have tested out over a restricted operating range during this process.

The performance of control valves has received considerable attention over the past decade. Some of the knowledge garnered from various industries, most notably pulp and paper, resides now in ISA standards. The central theme arising from this document is that good mechanical performance of the control valve is essential to ensure good process control.

Enbridge's experience is that several factors lead to pressure control system performance degradation including:

  • inconsistent (nonlinear) response at different flow rates
  • changing service conditions
  • undocumented changes to the control system configuration or tunings after original commissioning

In reviewing current pressure control system performance capabilities, one must assemble a complete picture of all factors. Enbridge has adopted a systematic approach to analyze and optimize the performance of control valves.

The five main steps in this process are:

  1. Conducting on-site control valve field tests.
  2. Performing a simulation analysis of the control valve and its hydraulic environment.
  3. Implementation and testing of settings derived from the field data and simulation analysis.
  4. Performing online bump tests to confirm that actual response matches predicted response.
  5. Sign-off of engineering and documentation package by a qualified person.

A control valve installation that has undergone this process is certified. In projects conducted to date, most effort was directed toward the first three steps of this process.

The first part of this article defines the control valve system and pressure control configuration encountered in liquid pipeline applications. Then, requirements for conducting the control valve field tests and hydraulic simulation analysis will play. The second part of the article provides results from studies conducted over the past two years in Enbridge. They are available online through ISA at www.isa.org/intech/feb2004/pressure.


Control valve system

In liquid petroleum pipeline operations, pressure control valves are most often encountered as:

  • pump station (suction and discharge) pressure control
  • delivery site (holding) pressure control
  • custody transfer flowmeter back-pressure control

Electrohydraulic actuators drive control valves in these applications.

Entech defines a control valve system to include the actuator, positioner, drive train, and valve body. Once installed, one cannot alter the mechanical capabilities of this system to any significant degree. But one must ascertain those capabilities to optimize overall control system performance. The main goal of the reengineering and certification program is to ensure the best possible performance of the installed pressure control system using configurable parameters. These parameters include actuator dead band, valve/actuator calibration, PID controller tunings, and the addition of nonlinear compensation in the controller configuration.

As to testing the control valve, the goals are to:

  • determine the mechanical performance capabilities of the control valve and actuator assembly
  • collect baseline information about the current pressure control system configuration and current service conditions
  • identify opportunities for improvements to either the control valve mechanical performance or pressure control system configuration

The control valve performance testing procedures adopted in the present study are in line with ANSI/ISA-75.25.02-2000, with differences highlighted in the discussion that follows. The procedures adopted in these investigations provide a consistent and repeatable method that generates performance indices virtually independently of the technician doing the testing. This independence is important due to the large geographical distribution of the control valves, because it allows the efficient use of technicians from different regions to do the same performance testing.

As well, with performance testing one can maintain a historical record on each valve. This record can help to optimize system performance, assist in the early detection of performance degradation, identify problem valves, and provide information for control system adjustments.

Each control valve installation undergoes both off-line testing and online data collection.

Off-line tests ignore process

These tests verify the control valve's range of motion and dynamic response by driving the actuator and valve assembly using a current source external to Enbridge's programmable logic controller system. This ensures that signal conditioning does not affect the valve dynamic response.

A temporarily installed linear position transducer for sliding stem valves or a rotational transducer for rotary valves measures the valve position. Fisher Controls' FlowScanner software was used to generate the external current source and record the collected data.

Off-line tests take place during a pipeline shutdown with the control valve in its process setting. Specific tests and desired results appear in summary form online at www.isa.org/intech/feb2004/pressure . Dynamic scan and full travel step tests are in addition to the tests outlined in ANSI/ISA-75.25.02-2000.

Testing the valve while the process is not operating allows one to program different sequences for the input signal. However, as discussed in ANSI/ISA-75.25.01-2000, off-line testing disregards process forces, which could introduce additional dynamics to the valve movement. It is worth noting that in contrast to the bench test procedure that the standard describes, control valves tested in this investigation have operated in a pressurized environment for possibly several years.

The most common in-situ corrective action made following analysis of data from these tests is recalibration of the valve to ensure it travels 0-100% open for the full range of the applied 4-20 mA signal. Severe problems with other valve performance measures resulting from the off-line tests require adjustment of settings (e.g., dead band in the actuator), valve/actuator maintenance, or replacement.

A separate period of field data gathering at each site collects dynamic data while the process is operating. The dynamic data serves to determine the installed valve capacity curve, calibrate the dynamic simulation model, and catalogue disturbances in line pressure. Together with information supplied by operations personnel, collected process data also provides a picture of current service conditions.

The sampling rate was 500 milliseconds, with data sets covering an approximately one-week span. Typical data collected includes station suction, case and discharge pressures, controller output, and valve stem position.

Technicians collected additional data from the process data historian system at a sampling rate of twenty seconds to provide information on operating pump units and pressures at upstream and downstream locations.


Hydraulic simulation analysis

After collecting and analyzing the field data, hydraulic simulation comes into play to reengineer the control valve system for the current and actual service conditions. A commercial simulation program can handle this chore.

Only a full dynamic analysis of the control valve system, which includes all nonlinear aspects present in the process and equipment, can generate settings that will optimize the installed performance. This type of analysis relies on a dynamic simulation program. Another advantage of employing a simulation modeling approach is that the entire operating envelope is subject to analysis. This envelope includes different numbers of operating pump units, variations in product physical properties, and valve positions. The boundaries of the envelope are the mechanical limitations of the system and required service conditions.

The main results of the simulation analysis are:

  • analysis of the open-loop process relative gain
  • determination of a controller output linear function that compensates for nonlinear behavior in the open-loop controller gain across the operating envelope
  • determination of initial controller tuning parameters
  • analysis of choked flow potential to identify potentially troublesome operating regions
  • determination of control valve pressure drop over the range of valve travel, which can confirm actuator torque capabilities

Once the linear function and controller tuning parameters have come to light, simulated performance tests take place comparing the existing control valve system configuration against the reengineered configuration.

One then implements optimal settings determined from analysis of the valve performance tests and simulation analysis into the system. This requires coordination with instrumentation, maintenance, and operations personnel. Verification tests of the pressure control system during normal process operation then take place.

An essential part of the certification process is for all personnel involved in the configuration, operation, and maintenance of the process control system to understand the range of interactions present amongst the control system settings. The dynamic simulator provides a useful platform for increasing awareness of these effects.

Goal of system stability

Maintaining pressure set points is critical for ensuring liquid pipeline stability. Pressure disturbances readily propagate between pump stations, potentially causing interplay between mainline pump station controllers. Tighter process control at individual pump stations or delivery sites reduces the risk of causing line shutdowns when operating close to pressure constraints.

Enbridge has hundreds of pressure control loops within its liquid pipeline network. A systematic program that is able to achieve incremental improvements in the performance of individual pressure control loops will derive significant benefits when applied to the wider system.

The reengineering and certification process laid out here provides such an approach to achieving the broader goal of system stability. To effectively optimize the performance of individual pressure control loops, knowledge of the mechanical and hydraulic characteristics of each system is imperative. The field data gathering and analysis methodology outlined above provide a comprehensive picture of current system capabilities, while accommodating the operational and geographical characteristics of pipeline operations. Simulation analysis allows characterization of process nonlinearity over the entire operating envelope. With this information, engineering staff can develop the most appropriate strategy for performance optimization. CP

Behind the byline

Roger Shirt is a member of ISA. He is president of R.W. Shirt Consulting, which specializes in control valve performance optimization for the pipeline industry. He has a Ph.D. in electrical engineering from the University of British Columbia. Roger has co-authored several papers on control valve optimization, simulation, and process control. Write him at rshirt@telus.net .

Larry Neumeister is also a member of ISA, and he is a registered professional engineer (P.Eng). He is a process control optimization specialist with Spartan Controls in Edmonton, Alberta. He holds a Bachelor of Science degree in chemical engineering and has twenty-four years of professional experience. Neumeister has taught courses and authored papers on advanced process control and optimization. Write him at neumeister.larry@spartan.ab.ca .

Jack Broyles is a member of ISA too. He is a control valve specialist at Enbridge Pipelines Inc., in Edmonton, Alberta. He is also a member of the ISA Control Valve Standard SP 75 Technical Committee. He has authored papers and presentations at the ISA Expo and the International Pipeline Conference. Write him at jack.broyles@enbridge.com. This article comes from their ISA EXPO technical paper Reengineering and Certification of In-situ Control Valves. Read the paper online at www.isa.org/intech/feb2004/pressure .

Don't trust that data: A case for checking the factory's numbers

A factory collected dynamic data for a period of two weeks during normal pipeline operations.

Using this data, engineers calculated the control valve capacity (Cv) curve.

In this chart see also the manufacturer's reported Cv curve, which deviates significantly from the actual and observed Cv curve.

In particular, a region exists between 30-50% valve travel area over which the calculated Cv is constant.

That is to say, no change in pressure or flow will occur with a change in valve position over this range.