Control valves that aren’t working correctly can cause problems for your entire system. While some issues are easy to spot, others require a powerful diagnostic tool, like GE’s ValScope. This article provides a brief introduction to five common control valve problems you should watch out for.
Deadband is a range of input that doesn’t result in any output. In this range, the system is essentially “dead.”
For a control valve, deadband refers to a range of controller signal that fails to trigger any activity of the valve. Deadband happens when the valve needs to change direction. To compensate for deadband, the controller must send additional output, which can cause the valve to overshoot its target position. This in turn causes the process to overshoot its setpoint.
There are several consequences of high deadband:
- Increased dead time
- More errors from load disturbances
- Higher possibility of rupturing pressure relief discs or vessels
- More oscillations in the control loop
- Increased control variance
Process Industry Forum provides a good, simple definition of hysteresis: “the difference between the valve position on the upstroke and its position on the downstroke at any given input signal.”
For example, imagine your controller output is 30% and the process variable, pressure, is 125 psig. When the controller output increases to 40%, the pressure increases to 150 psig. So far, so good. But then, when the controller output decreases back to 30%, the pressure only decreases to 130 psig.
About 1% to 3% hysteresis is normal, depending on if the valve has a positioner, but more than that can cause problems including:
- Process cycling around the setpoint
- Slower controller responses
- Increased control variance
Hysteresis can be caused by a faulty mechanical linkage, such as between the valve and the actuator. When this happens, it’s called backlash.
Stiction is a combination of the words stick and friction. Like deadband, stiction causes a valve not to move, but instead “stick” in a particular position. Moving the valve requires additional force, which, also like deadband, can lead to the valve overshooting its position and the process overshooting its setpoint. Then the valve gets stuck in the new position.
Stiction can be caused by several factors: sticky valve internals, undersized actuators, tight shutoff, media viscosity, and more. The consequence is that the controller will never reach the desired setpoint.
Control valves are designed to be linear. This means that the flow increases linearly with valve travel. Deadband, hysteresis, and stiction can all cause nonlinearity, which means the relationship between the flow and valve travel isn’t stable.
Nonlinearity affects the tuning of the controller and make it so that the controller responds optimally at only one operating point. The further the valve position from this operating point, the more variable the flow.
Over- or undersized valves
Improper sizing can make all of these problems worse, especially for smaller valves.
When control valves are oversized, a small change in valve position can have a big effect on flow. When control valves are undersized, process bottlenecks occur. ISA-75.01 provides control valve sizing equations.
Are your control valves causing you problems? We’re here to help. Learn more about GE’s ValScope-PRO, a diagnostic tool for testing and monitoring control valve performance.