| Contents: |
|
Poor control valve performance is the single biggest source of process variability. Inadequate valve performance can often result in process cycling, high process time constant and excessive deadtime. This compromises the control loop performance.
In many cases valve performance problems can be resolved with routine maintenance. In some cases the valve must be replaced. The productivity improvements resulting from valve repair/replacement on key loops usually far exceeds the repair costs.
Control valve maintenance falls into two categories: routine maintenance and response maintenance. Routine maintenance or preventative maintenance involves checking the control valves on a predefined schedule. Response or breakdown maintenance is required when valve performance has seriously degraded and the process is having a dramatic negative impact on process performance.
Breakdown maintenance must be avoided. A preventative maintenance program is the preferred method, since valve problems are detected before serious impact on the process has occurred. However, a preventative maintenance program is inefficient because all valves are checked regardless of their condition. This wastes valuable maintenance manpower during shutdown.
During preventative maintenance the obvious mechanical issues are attended to and then the valve is bench set and stroked. Checking the stroke typically involves operating the valve over its full range in 25% intervals. This approach does not adequately identify performance problems such as valve deadband. Additionally, the testing performed on the bench or in the line during shutdown does not reflect the valve performance under operating conditions. Process operating pressures can greatly impact the valves operation.
There are several key control valve performance factors that are not adequately evaluated during current preventative maintenance programs. These factors can severely compromise control performance and result in higher process variability.
Deadband is a result of backlash, hysteresis and stiction or a combination of the three.
Backlash and hysteresis are lost motion due to slop in valve linkages. This is evident when the controller direction is reversed and there is no process response.
Stiction or static friction is the friction the valve must overcome in order to move.
The positioner is a feedback controller with its own tuning perameters. Inappropriate positioner tuning may cause the valve to overshoot or cycle.
The majority of control valves are oversized due to conservative engineering practices. This has the effect of increasing the loop process gain, which tends to result in higher variability. A high process gain amplifies the impact of backlash, stiction and hysterisis.
A flow control loop has a process gain of 5%span/%output and a control valve with 1% stiction. The controller output must move greater than 1% in order to put the valve into motion. The valve will step by 1% causing a steady state flow change of 5%span. For a thick stock flow of 1000 gpm the minimum flow change would be 50 gpm. This is high process variability, which will affect process performance.
A detailed installed evaluation under normal operating conditions is the only way to truly understand the performance capabilities of a control valve. The evaluation is carried out by placing the controller in manual mode, making small steps in the controller output (typically 0.5 to 4%) and measuring the process response. This procedure is known as the open loop bump test. The valve performance is indicated by the process response.
Dynamic control valve evaluations involves small movements in the control valve, under normal operating conditions, while monitoring the response of the process variable.
There are some key requirements for this type of valve evaluation.
LAMBDA CONTROLS, 17 Erbsville Court (RR# 3), Waterloo, ON, Canada N2J 3Z4
Tel.: 519-884-6998, Fax: 519-884-1089
E-mail: inquiries@lambdacontrols.com, Web site: www.lambdacontrols.com