The next generation of smarter valves

By Béla Lipták

SOME INTELLIGENT POSITIONERS are capable of fieldbus interaction, wireless communication by radio, can provide valve testing, self-diagnostics and predictive valve maintenance (PVM). Yet, in the area of valve performance testing, most software packages operate by introducing an artificial offset into the control signal and the plant operators are reluctant to accept the resulting upset. Such packages are often used only during plant shut downs. In short, the full capability for online testing is yet to be developed.

PVM and predictive diagnostics in the future should not require ramping the valve through a range of travel, but should evaluate the friction in the valve under normal operation. This way, when changes or abnormalities occur in the friction plot (valve signature), the data historian in the positioner can diagnose if for example seating or galling problems are occurring in the trim. 

Valves are one of the higher maintenance components of the process. Operating time based scheduled maintenance can be premature, can come too late and sometimes the unnecessary maintenance itself can cause problems. Therefore, it is desirable to monitor actuator pressure against valve travel, zero and span, so that one would know immediately if the stuffing box is too tight or leaking, the valve stem is corroded, the actuator spring or the stem bellows are damaged, the valve is leaking, etc.    

In the next decade, much improvement and change is expected in the area of smart and self diagnosing control valves. This is necessary, because the effects of throttling, such as the wearing of the trim affects the performance of the whole control loop. Other factors that need to be detected and compensated for include hysteresis caused by packing friction and air leakage in the actuator.

Valve Characteristics and Rangeability
In order for a control loop to be stable, it should be tuned (the gain of the controller adjusted) so that the product of the gains of all the loop components will equal about 0.5. Therefore, the stability of the control loop also depends on the characteristics of the valve. If the control valve is non-linear (if the gain of the valve varies with the opening of the valve), the loop will become unstable when the valve opening changes from the opening where it was when the loop was tuned. 

To eliminate this effect, the loop must be compensated for the gain characteristics of the valve, but such compensation is only possible, if the valve characteristics are accurately known (click the Download Now button below for a pdf version of the figure mentioned here). For this reason, it is desirable that control valve users ask the manufacturers to accurately determine, publish and guarantee the characteristics of their valves. For example, a valve should only be called “linear”, if throughout its stroke, the gain of the valve (Gv) equals the maximum flow through the valve divided by the valve stroke in percentage (Fmax/100%).

Rangeability is another valve performance characteristic that needs to be more accurately determined and uniformly defined for control valves. The rangeability of a control valve should be defined as the ratio of those maximum and minimum valve capacity coefficients (Cvs), at which the valve gains are still within 1% of the specified. Therefore, manufacturers should publish the stroking range (minimum and maximum percentages of valve openings) within which it is guaranteed that the valve gain characteristics will not deviate from the published by more than 1%.

Smarter Valves
A traditional valve positioner serves only the purpose of maintaining a valve an opening, that corresponds to the control signal. Digital valve controllers can also collect and analyze valve position data, valve operating characteristics and valve performance trending. They can also provide two-way digital communication to enable diagnostics of the entire valve assembly ( see http://www.metsoautomation.com/ and http://www.emersonprocess.com/fisher/products/fieldvue/dvc/index.html for details).

The potentials of smart valves are likely to be further increased and better appreciated in the next decade. Smart valve should be able to measure their own:

• Upstream pressure
• Downstream pressure
• Temperature 
• Valve opening (stem position)
• Actuator air pressure

Smart valves should also be able to eliminate conversion errors (D/A and A/D) and should be able to update their inputs about 10 times per second. In addition, they should be provided with filters to remove errors caused by turbulence.

Valves Can Also Measure the Flow
A control valve is a variable area flow meter with a variable pressure drop. Just as the float position in a rotameter (a constant ?P flowmeter) can indicate the flow, so the stem position of a control valve can do the same, if the ?P is detected. Naturally, in order to do this, the accurate knowledge of the valve characteristics and of the process properties of the flowing fluid must be provided (click the Download Now button below for a pdf version of the figure mentioned here).

In order for the smart valves of the future to be able to accurately measure their own flow, they must be provided with sufficient intelligence to identify the sizing equation, which is applicable. Therefore, they will have to be able to detect laminar flow condition in viscous, choking condition in cavitating liquid or sonic flow in gas flow applications. Naturally, the other requirement is to be able to accurately measure the variables that are required to for solving the applicable equations. The required measurements include valve stem position, inlet, outlet and vena contracta  pressures, flowing temperature, etc.       

Yet, the potential advantages of such smart valves much outweigh the required investment of time and money. The savings include the elimination of both the initial cost of purchasing, installing and the energy cost of operating a separate flow sensor. An added advantage can be the increased rangeability of the flow measurement obtained from the valve. This is because the rangeability of traditional flow detectors is usually between 3:1 to 10:1, while control valves can provide rangeabilities of 25:1 to over 100:1, because of their variable area orifices.

Béla Lipták, PE, process control consultant, is also editor of the Instrument Engineers' Handbook and is seeking new co-authors for the forthcoming new edition of that multi-volume work.