Comment

01 May 2003

Upgrade or replace?

Continuous online diagnostic data becomes even more valuable as instruments age and become increasingly less reliable.

By Bud Adler

For years, technical journals have saturated their pages with articles describing the numerous benefits of implementing fieldbus strategies.

No one would deny that detailed diagnostic information would be great to have. But at what cost?

A wholesale upgrade to a full fieldbus strategy typically necessitates scrapping in-place analog and smart instruments and replacing them with more expensive fieldbus-capable versions.

Cost considerations, in most cases, make a major retrofit such as this out of the question. Most of us must continue to live with existing instrumentation. The irony is, continuous online diagnostic data becomes even more valuable as instruments age and become increasingly less reliable.

This is an especially valid concern when they are part of safety-critical applications. In such applications, an instrument malfunction could result in loss of life or limb or have serious economic impact due to lost or substandard production.

What many don't realize is that scores of plants already have in place all of the fieldbus diagnostic functionality needed. The more than 12,000,000 smart HART instruments installed in plants worldwide can provide a wealth of diagnostic and performance data. The trick is to unlock the smart transmitter's full capabilities.

The key is the use of HART loop monitoring techniques that are rapidly gaining momentum through continuing education and a new sensitivity to improved measurement reliability for safety applications.

DEVICE DRIVES LOOP RELIABILITY

Keeping detailed failure records for all devices in the plant and using this data to evaluate probability of failure best quantifies the reliability of a device. Because this data collection rarely takes place over a long period in typical plant operations, an alternative measure of reliability has come to the fore.

Using accumulated data from a number of different agencies, one may substitute generic (and usually very conservative) data. Some vendors, to provide more realistic availability data, have used third-party evaluation agencies to conduct a failure modes, effects, and diagnostic analysis on their products.

A result of this analysis is the calculation of the safe failure fraction (SFF) for the device. The SFF considers both inherently safe failure modes and dangerous failures.

It then considers which of the failures diagnostics identifies are dangerous and, of those, which can convert to safe failures. From a safe failure point of view, the higher the SFF, the more reliable the device.

Even safe failures are failures that may shut down a process. If they are due to failed sensors rather than the process itself, they are nuisance failures.

Most users are aware only of HART's advantage in providing remote configuration and the occasional loop check with a handheld configurator (HHC). While available in every compatible device, they leave the wealth of diagnostic and performance data untouched.

What most don't realize is the hidden information available from using smart HART technology offers tremendous opportunities to increase the effective diagnostic coverage of a field device.

To unlock valuable process and diagnostic information, a HART interface monitor permanently connects across the 4–20 mA signal wires, just as would happen when accessing the transmitter using an HHC.

Once installed, it continuously monitors process and diagnostic data available on the HART digital protocol's data string. Many users have found this to be an easy way to extract a second, third, or fourth measurement from a Coriolis meter or a multivariable transmitter. Others want valve stem position as a 4–20 mA value for control loop feedback.

The biggest growth area that has evolved for HART monitoring deals with vital diagnostic information that may extract from a field device. Intuitive use of this information dramatically increases the reliability of the loop containing this device.

DIGITAL SIGNAL SUPERIMPOSES

Timely maintenance circumvents most valve operational problems. Timely maintenance is rare. Reduced staffing, requirements for uninterrupted process operation, and reduced budgets are contributing factors that result in an "If it's not broken, don't fix it" mentality.

Valve vendors introduced smart positioners to their offering several years ago. These positioners provide dramatically improved valve and damper performance and offer a capability for continuous monitoring of the valve's operational integrity.

Remote readout of the performance parameters can actually diagnose developing situations before they become problems with potentially dangerous consequences. Alarming on low air pressure or excess friction in valve stems or damper operators offers significant benefits in reducing upsets and minimizing emissions.

Judicious use of this protocol has been the most successful method of achieving these goals. Smart HART positioners encode valuable performance and diagnostic information into a digital signal that superimposes on the 4–20 mA control signal.

MONITORING TOTAL TRAVEL

By monitoring this data back in the control room, operators immediately notice overall performance and abnormal conditions.

• Sticking/Jumpy valve or damper operation resulting from excess packing friction or shaft buildup
• Sluggish operation due to low air pressure that often results from clogged filters or piping leaks
• Excess temperature within the positioner that may indicate degraded performance and shortened life
• Actual stem/shaft position to allow monitoring of total travel (a monitor of packing wear), hysteresis (from excessive friction or low air pressure), excessive travel (worn packing with potential fugitive emissions), and travel vs. flow (an indication of trim wear)
• Up to seven status bits monitored to indicate any of several problem conditions identified by a change in state within the positioner

A HART interface monitor can identify a myriad of potential failure conditions and initiate actions to ensure more reliable operation. This increased diagnostic coverage effectively increases the SFF of the valve or damper installation.

The device can be a single-loop, stand-alone device that provides both relay status and analog signal interface between the data string and the distributed control system (DCS) or programmable logic controller (PLC). This monitoring function may also perform on a large scale by adding a HART front-end I/O subsystem and an appropriate software package to a host system: DCS, PLC, or PC.

The costs range from $600 for a single-loop approach on up to several hundred thousand dollars for a large system. For safety-certified PLC systems, a HART multiplexer is not included in the certification. Using external monitors with analog and discrete interface to the PLC is best.

PLC INITIATES PARTIAL CLOSE

Safety shutdown valves typically stay in the open position for months or years awaiting a command signal to operate. Many plants pay little attention to these valves outside of regularly scheduled maintenance during scheduled turnarounds.

The pressures of continuous production often stretch these intervals even longer. In some processes, buildup or corrosion on the moving parts of these valves or their actuators can prevent them from moving. They are stuck in the open position.

It is clear that for safety-critical applications, something must transpire to insure operability of the valve on demand.

Online partial stroke testing of the emergency shutdown (ESD) valves has been demonstrated to provide a dramatic improvement in availability. By regularly exercising the valve with partial stroking, it can move from its fully open seated position.

While this does not prove that the valve will, in fact, move all the way to its seat and close tightly, it offers far more confidence than doing nothing at all to test the integrity of the installation.

For clean gas, air, steam, or water flow, the probability of full closure is high. For service with suspended solids or polymer, positive seating may be questionable. For clean service, the full stroke test interval may be extended, while for dirty service, a more frequent, full-stroke test may be required.

In considering the diagnostic coverage provided, one must make allowances for the possibility that tight shutoff may not be achieved on the dirty or polymer service valves.

Partial stroke testing requires some method to verify valve stem travel. While limit switches have served over the years, they present several problems.

As a two-state device, it has no way to verify operational capability without a complete functional test. In addition, calibration/setup is expensive and time-consuming.

Using a HART positioner offers a reliable and self-diagnostic method to accurately determine stem position and verify actuator air pressure. Just as with the control valve scenario described above, the positioner communicates with the logic solver (PLC, PC, or DCS) using a 4–20 mA signal with data digitally encoded and superimposed on the 4–20 mA using frequency shift keying.

For this application, a HART I/O subsystem on the safety PLC is not useful due to regulatory reasons. The hardware and software configuration of the safety PLC are precisely documented and certified by a recognized authority as suitable for use in a safety-related system. Changes to the configuration require recertification prior to further use.

However, an interface monitor is an external device that mounts in the control room external to the PLC. It communicates through the standard I/O of the PLC with a 4–20 mA signal for stem position and relay status for stem position, verification, air pressure, and positioner diagnostic information.

The operator tests the performance test by changing the control signal from the PLC to initiate a partial close. For example, decreasing the signal from 20 mA to 18.4 mA should move the valve from 100% open to 90% open.

BACKUP COMMAND TO CLOSE

The interface monitor configures to close a relay upon reaching this 90% position, which serves as a verification of stem movement. The signal then reverts to 20 mA, and a second relay verifies that the stem has returned to the full open position.

Before beginning this test, the PLC can verify that there is adequate air pressure to reopen the valve using yet a third monitoring function that provides a relay warning signal for low air pressure.

A fourth relay advises of any diagnostic condition within the positioner. The PLC can monitor the valve stem movement vs. time to detect sticky operation that could indicate buildup on the stem or tight packing, providing enhanced diagnostic coverage.

Because most ESD valves are oversized, a 10% (or even 20%) partial stroke will have little effect on the process. Accordingly, this partial stroke test can happen on a regular basis to add a high level of diagnostic coverage to the loop and increase the interval for full stroke testing.

A leading valve positioner manufacturer has introduced a HART-enabled positioner that actually includes the partial stroke testing capability within its own onboard software. This precludes the requirement for any programming in the PLC.

The interface monitor provides annunciation of the stem movement back to the PLC to verify that the test has successfully performed. Additionally, for many applications, the orifice in the positioner is large enough to vent the actuator fast enough for emergency shutdown.

Other ESD valves must provide rapid response to dangerous process conditions and will require a large ported solenoid valve. For these applications, the solenoid operrated valve (SOV) is the emergency trip function because it will vent actuator pressure very quickly.

The 4–20 mA signal to the positioner can drop to 4 mA as a backup command to close the ESD, should the SOV fail to function. The somewhat slower closing is certainly better than no movement at all. This scheme reduces the average probability of failure on demand even further.

UTILIZE SIMPLE ANALOG CIRCUITRY

Limit alarms often serve in conjunction with a control system for redundancy or to locate shutdown logic close to the controls for the monitored equipment. For high integrity systems, especially in the nuclear power industry, use of a configurable limit alarm with digital circuitry is a cause for concern.

In fact, nuclear regulatory agencies have been requesting that any device incorporating software or firmware comply with IEC 61508 parts 1, 2, and 3. This is a tall order that has not yet arrived.

A related concern is that most pressure and temperature transmitters installing today are digital designs without these certifications. An important concern for using a HART-enabled transmitter in a safety-related application is the possibility that atechnician could unknowingly change the configuration or inadvertently leave the transmitter in the output fixed (manual mode) after a loop check.

While the transmitter knows of these conditions due to its internal diagnostics, it will not report the information unless asked. Connecting a handheld configurator to the loop could certainly identify these conditions . . . but how often would that happen?

A solution is to employ a simple analog circuitry limit alarm with its inherent high reliability and provide monitoring of the field device using a nonintrusive HART interface monitor that can be configured to monitor for these changes and provide a relay output to annunciate the problem.

The benefit is that standard transmitters, with all of their performance benefits, can operate and that the comfort factor for the user can be high because of this continuous monitoring of health and operating status. An added benefit is that the device also incorporates self-diagnostics that will annunciate any operational problem.

So while digital fieldbuses have many advantages for new installations or major upgrades, there are less expensive alternatives.

HART is a protocol that exists today in almost every process plant. Presently, it seems there is a lack of knowledge about it and a lack of commitment to use it, as plant engineers pine for digital solutions.

Using interface monitoring can provide benefits due to extracting additional measurement information from an existing transmitter with no additional pipe penetrations, installation, or cabling costs.

Also, one can reduce maintenance costs by using the remote analysis capabilities for valves, dampers, and transmitters.

Adding diagnostic coverage to a transmitter or a valve may be the difference between improving the availability of an existing installation and needing to purchase and install a redundant device to meet the required safety integrity level of the function. IT

Behind the byline

Bud Adler has a B.S. degree in electrical engineering and is a lifetime member of ISA. He sits on the ISA SP84 committee, Programmable Electronic System for Use in Safety Applications, and is a member of the newly formed Safety division. A frequent presenter at ISA Technical Conferences and local ISA sections, Adler works for Moore Industries. Write him at budadler@msn.com.

Related Links