Best-in-class maintenance Jim Shields, product marketing manager at Fluke (www.fluke.com), says a best-in-class maintenance strategy for control valves can reduce costs by both minimizing the number of valves pulled physically from processing and minimizing failure risks. Here’s his advice. To establish such a practice, the baseline condition of the valve at a known good state needs to be documented. Ideally the documentation occurs when the valve is commissioned or after it is overhauled. The technician records the signature of the valve in the ideal state, plotting the output mA or percentage os span signal versus applied input signal, and stores this information with the time, tag number of the valve, and the date the activity is performed. Calibration management software can be used to manage this information. Once the baseline performance signatures of the valves are recorded, a maintenance interval for testing the performance of the valves needs to be established. Using existing maintenance intervals is a starting point. If there is no established maintenance interval, the service location of the valve needs to be evaluated. Rough service applications dictate a shorter maintenance interval than light duty service, for example. An interval of six months to a year to start (unless the service location is very hard on the valves) is a reasonable starting point. Some devices installed in safety and shut-down systems need to be checked every three months regardless of service location. Once baseline valve signature data are recorded, the valves need to be tested at the defined intervals and the signatures recorded. The signatures can be compared to the baseline signatures to determine changes in their performance. If the output response curve has developed a nonlinear signature or has aberrations in the curves, the valve may be developing excessive stiction or hysteresis that may require it to be them to be removed for service. |
When assessing control valves and their performance, you need to understand the different types of valves and what they can be tested for, explains Jim Shields, product marketing manager at Fluke (www.fluke.com). “For open/close shutoff valves without analog control the tests are pretty simple,” he explains. “Do the valves open and close? When open, do they open all the way? When they close, do they close completely? Testing is mostly observational — looking at the valve and watching cause and effect in the process during the cycle.”
Control valves are a different beast altogether, says Shields. “These valves open and close proportionally and vary the degree of travel depending on the percent of span of the 4-20 mA signal applied to them,” he explains. “Observing the valve’s position, as reported on the visual travel indicator, gives the technician or operator a rough indication of percent of travel for a particular setting when in operation but does not provide any assurance of how the valve will operate under dynamic and changing conditions.”
The most sophisticated valve performance tests require removing the valve and testing its performance on a “valve prover,” continues Shields. “This is an expensive test device, out of range for most instrument shops. The valve prover is often only used by valve manufacturers in testing the valve when shipped or by highly qualified field service engineers. It offers a very complete test, but the tool is not feasible for most instrument shops.”
So, what is the technician to use for testing a control valve? What is a meaningful test that can be used as a baseline? “Since most valves use a 4-20 mA input signal, any test tool with an mA output signal can provide the input mA signal to drive the control valve across its operating range,” says Shields. “When applying a 3.8 mA input signal to a normally closed valve, the valve should be hard closed. It should remain closed at 4.0 mA and move slightly off its seat at 4.2 mA. At the other end of its operation, at 19.8 mA, it should be nearly full open. At 20.0 mA it should be fully open and hard open at 20.2 mA, resting on the travel stop. Tests of this nature will determine if the valve is opening and closing correctly but still fall short of testing the valve across the entire range where it provides control.”
The tools available for testing valves depend upon the technology of the valves and the control systems, says Jeff VonAhnen, senior FieldVue marketing and support engineer, Fisher Valves, Emerson Process Management (www.emersonprocess.com). “Tools available for analog-only instruments will be different from the tools available for a fully digital Foundation Fieldbus or Profibus plant network,” he says. “Tools for HART communicating valve instruments are a combination of analog tools and digital software tools. For analog valve instruments all adjustments must be made at the valve. For digital-based valve instruments, adjustments and readings made via the communication tools can occur anywhere along the control wires and not only at the valve.”
For wired HART communications, the speed of communications is based on how many field devices get serviced by a single HART modem, says VonAhnen. “If a personal computer running a software diagnostic program has a HART modem adapter, then communications through that modem can be dedicated to a single field device at the fastest communication rate,” he explains. “But if the computer is connected to a HART multiplexer with a single HART modem shared between 16 field devices, then the HART communications can take longer. If a single HART message exchange occurs with each field device, then each field device gets access to the multiplexer’s HART modem only 1/16th of the time. If all the communication is focused on only one of the 16 field devices, such as if a valve signature diagnostic is being performed on one of the valves, and then it gets all the communications from the HART modem, and the other 15 devices must wait for their next communication opportunity.”
For HART communications, the average time to complete a request from the master software program followed by the response from the field device is about 0.5 seconds, explains VonAhnen. “A computer communicating via a HART modem adapter to a single field device is having a message exchange every half-second,” he says. “But on a fully populated 16-channel HART multiplexer, if a single message exchange were sent to each of the 16 field devices, then even though the HART modem is communicating every 0.5 seconds, each individual field device would receive a HART message every 16 x 0.5 seconds, or once every eight seconds.”
Many valves, including those that are “smart,” have a feedback element built in that sends as output the actual position as a percentage of open/close, says Shields. “This output can be a 4-20 mA signal or a digital HART variable that represents 0 to 100 percent of control valve operating span,” he explains. “Applying a varying mA signal, while simultaneously monitoring the output mA or percentage of span signal, gives a technician a means to see whether a control valve is operating correctly over its range. By recording simultaneously the applied mA signal and the output mA signal or PV percentage of span, the valve’s performance can be documented. This documented test and result is often called a valve’s signature. The output should smoothly mirror the applied mA input signal. Any deviation from the applied signal is a potential indication of aberrant behavior by the valve.”
A valve-signature diagnostic is a test routine for a control-valve assembly where a software-based test system controls the setpoint commanded to the valve assembly and sweeps it through the valve’s range of motion while gathering measurements from the assembly of the commanded input signal, the elapsed time of the test, the valve’s stem travel, and other parameters such as the pressures being applied to the control valve’s actuator, explains VonAhnen. “These data points are then plotted in relationship to each other to show the reaction of the motion of valve assembly to the input signals,” he says. “Based on the reactions of the valve assembly, conclusions can be made about the mechanical condition of the valve assembly and perhaps the need for further disassembly or repair of the valve.”
Since the valve signature test is in response to an test input signal, this test must be done at times and situations where the motion of the valve assembly will not upset the production process, cautions VonAhnen. “These times are usually before the plant is commissioned and operational or if the valve assembly has been pulled from the pipeline,” he explains. “The key to the success of this program is the continuous and frequent monitoring of the operational valve for the earliest signs of valve degradation. The sooner the one affected valve out of the total population of thousands of valves can be identified, the sooner maintenance can examine the condition of the valve to make the decision either to keep running or to pull the valve in order to avoid unexpected process outages. Typical degrading conditions for control valve assemblies are leaks in pneumatic components, increased friction between rubbing parts, and changes in the level of command signals required to stimulate a given response. Good items to monitor from the valve assembly are parameters or alert conditions that reflect these conditions and relationships.”
Read Mike Bacidore's monthly column, Tactics and Practices.
Valve signature testing According to Jeff VonAhnen, senior FieldVue marketing and support engineer, Fisher Valves, Emerson Process Management (www.emersonprocess.com), the best practice for use of valve diagnostics are as follows:
|