When I am in an instrument and valve repair shop, I see many more control valves than instruments, particularly with the advances in sensor technology, transmitter intelligence, and asset management systems. Valves are mechanical devices and as such require more maintenance. Packings, seals, seats, and o-rings wear out. To ease maintenance, a control valve must be located so it can be readily and safely removed from the pipeline. However, there is much more to consider in locating a control valve.
With liquid streams, a change in control valve position causes an immediate change in liquid pressure and, in a full and pressurized pipeline, a pressure wave traveling at the speed of sound. In less than a second, the pressure imbalance from the wave provides a driving force that overcomes the liquid’s inertia and accelerates the liquid to a new velocity. Consequently, within a plant the location of a control valve does not appreciably affect the response of a process variable (PV) within the pipeline. However, if you are measuring PV response in a pipeline and the control valve is in a different pipeline, throttling a flow that is being added to the pipeline that is being measured, the distance of the valve from the measurement may cause a transportation delay. This piping and valve arrangement commonly occurs in the dilution and blending of streams.
There are some really bad control valve locations that show no understanding of the negative impact of deadtime by process and mechanical design engineers (see Tip #70). One of the worst is where there are several vessels between the control valve and the measurement. The residence time of the smaller vessels in series with the larger vessel becomes deadtime. Also bad is gravity flow. Now a change in control valve position starts a wave traveling slowly down the partially filled pipeline. For small flows in vertical runs, the flow is a falling film causing a large and unpredictable deadtime. The worst case of deadtime resulting from valve location is encountered in pH control. A control valve often throttles the reagent flow to a dip tube. The dip tube has a minimum size for structural integrity and normal mixing rules put the dip tube down near the impeller. Unfortunately, this creates a dip tube volume of 2 gallons. If the reagent flow is 1 gallon per hour, when the control valve opens, it takes 2 hours for the reagent to flush the process fluid out of the dip tube. When the control valve closes, the reagent will continue to migrate into the process for several hours. The solution is to have the control valve add the reagent to a high flow recirculation line, thereby reducing the injection delay to seconds.
To prevent flashing, control valves should be located so the fluid pressure in the vena contracta (that is, the narrowest opening in the flow path) does not drop below the vapor pressure of the fluid. If flashing cannot be avoided, the valve type and trim design should be selected to prevent cavitation in the valve or downstream equipment. Stan Weiner, my coauthor of the Control Talk column, recommended the flashing control valve be installed directly on an inlet nozzle near the top of a vessel so the collapsing of bubbles would occur in the vessel vapor space when cavitation could not be prevented.
Concept: Control valves do not require much in the way of straight runs and do not introduce appreciable delays within plants when they are in the same pipeline as the measurement. (Long-distance oil and gas pipelines are another story.) Valves adding flows to process equipment can cause large injection delays or bypass mixing in the equipment. Valve location on streams to equipment should not introduce excessive deadtime or noise. Valve location, type, and trim should minimize flashing and cavitation and provide safe and easy access for removal and repair. On-off valve locations should minimize the volume to the destination when flow is stopped to prevent totalization errors in charges.
Details: Control valves should be at floor level or accessible from platforms. Block, flush, and drain valves should be installed to enable them to be safely removed. Control valves should be located on the same equipment or pipeline as the measurement and downstream of flow measurements. Reagent control valves should be moved from dip tube to recirculation line injection to eliminate injection delays for pH control. On-off valves close to the point of injection should be added to provide isolation and to shut off the flow. For pH and reactor control, the volume between the on-off valve and the nozzle should be minimized by flanging the on-off valve to the nozzle or nozzle block valve for low reagent and reactant flows and high process sensitivity. Gravity flow piping is not recommended because of variable head and velocities, but if it is used, the control valve should be as close to the nozzle of the destination as possible.
Watch-outs: The location of the nozzle and dip tube entry points into a vessel must not result in the flow being injected close to an exit nozzle; thereby short circuiting inlet flow to outlet flow and bypassing the mixing in the vessel. Throttle valves should not be used as isolation valves, and isolation valves should not be used as throttle valves (see Tip #83). If an on-off valve for batch control is not close to the flowmeter, the pipeline inventory between the on-off valve and flowmeter can cause the charge to be significantly different than the batch setpoint. To minimize excess charge, the on-off valve should stroke as fast as necessary when the “close” command is given.
Exceptions: For Coriolis meters in liquid service with no possibility of flashing, the control valve can be located upstream of the flowmeter because a Coriolis meter is not sensitive to velocity profile.
Insight: Control valves can cause damage to piping from cavitation and poor control from injection delay and short circuiting.
Rule of thumb: Locate control valves to be maintainable, provide fast injection into mixing zones, and prevent flashing and cavitation.
Look for another tip next Friday.