July 2009

Your best bet in control valves

By Hans Bauman

Control valves may be the most important, but sometimes the most neglected, part of a control loop. The reason is usually the instrument engineer's unfamiliarity with the facets, terminologies, and areas of engineering disciplines, such as fluid mechanics, metallurgy, noise control, and piping and vessel design that can be involved depending on the severity of service conditions.

Any control loop usually consists of a sensor of the process condition, a transmitter, and a controller that compares the process variable the transmitter receives with the set point, (the desired process condition). The controller, in turn, sends a corrective signal to the final control element, the last part of the loop and the muscle of the process control system. While the sensors of the process variables are the eyes, and the controller the brain, the final control element represents the hands of the control loop. This makes it the most important, alas sometimes the least understood, part of an automatic control system. This comes about, in part, due to our strong attachment to electronic systems and computers, causing some neglect in the proper understanding and proper use of the all important hardware.

Control valves are the most common type of final control elements; but other types include:

  • Devices that regulate (throttle) electric energy such as silicon-controlled rectifiers
  • Variable speed drives
  • Feeders, pumps, and belt drives
  • Dampers

Some of these devices perform functions similar to control valves and could see use as an alternative. In order to control the pH level, you could use a variable stroke-type metering pump to inject acid into wastewater, instead of using a control valve lined with polytetrafluorethylene.

What then is a control valve? This is a difficult question since there is considerable overlap with other types of valves. A valve operating strictly in the on-off mode (such as the hydronic solenoid valve in your home heating system) could be replaced by a simple ball valve operated by a pneumatic cylinder, a type we usually refer to as an automated valve.

The distinction between automated and control valves is the ability of the latter to modulate or assume an infinite number of throttling travel positions during normal control service. Physically, there are three basic components of a control valve:

  • The valve body subassembly. This is the working part and, in itself, a pressure vessel.
  • The actuator. This is the device that positions the throttling element inside the valve body.
  • Accessories. These are positioners, I/P transducers, limit switches, handwheels, air sets, position sensors, solenoid valves, and travel stops.

There is no such thing as the ideal control valve, but we can attempt to develop a workable compromise. The closest thing to an ideal valve should have a constant gain throughout the flow range, such as a linear installed flow characteristic, no dead time with packing tightened, and a time constant that is different from that of the process by at least a factor of three.

Besides the obvious, such as good quality workmanship, correct selection of materials, and noise emission, you should pay special attention to two areas:

  • Low dead band of the actuator/valve combination (with tight packing)
  • Tight shutoff, in cases of single-seated globe valves and some rotary valves (if required)

The prime concern of an operator of a process control loop is to have a stable loop. Nothing makes people more nervous than a lot of red ink and scattered lines on a strip of paper from a recorder. The final control element will influence the stability of a loop more than all the other control elements combined.

The biggest culprit here is dead time. This is the time it takes for the controller to vary the output signal sufficiently to make the actuator and the valve move to a new position. The dead time, TDv, is the time it takes for the pneumatic actuator to change the pressure in order to move to a different travel position. It is most commonly related to the dead band of the actuator/valve combination, or, in case a positioner is used, the dead band of the valve divided by the open loop gain of the positioner plus the positioner's dead band (dead band keeps the valve from responding instantly when the signal changes, which, in turn, causes dead time). The valve itself should never have a dead band of more than 5% of span, that is, less than 0.6 psi for a 3-to-15 psi signal span or 0.8 mA for a 4-20 mA signal. The positioner/valve combination should have no more than 0.5% of signal span. Ignoring process dynamics, a positioner may, therefore, improve matters by an order of magnitude.
In order not to box yourself in price-wise, try not to specify what valve type to quote, but rather what features do not work in your given application. If you have control crude oil, specify no cage-guided trim is allowed (the solid particles may jam the plug guide). If you handle hazardous fluids, specify no threaded-end connections. Specify tight shutoff only when necessary; it may cost extra money. Only you know the corrosive nature of your process fluid best; therefore, specify stainless steel, alloy, or plastic as body or trim material. For installation in process plants, you may want to refrain from using all plastic valve bodies. Remember, anything a maintenance person can use as a step-ladder, he will.

When using a plastic-lined metal valve body, make sure the surrounding housing is pressure-tight, and a secondary stem packing and telltale connection are provided to protect you against liner or plastic bellows failure. Bellows made of TFE have particularly limited life cycles and are permeable to chlorine.

SOURCE: Control Valve Primer: A User's Guide, Fourth Edition, by Hans Bauman, ISA, 2009.