Partial Stroke Testing

Valves are expected to operate even if they haven’t been operated for awhile. To ensure proper operation in emergency shutdown (ESD) service, valves must be actuated. To do so, most processes must be interrupted—inconvenient at best, costly at worst, but certainly not as costly as a valve actuation failure in a safety instrumented function (SIF).

By Mark T. Hoske, Control Engineering August 1, 2008

Sidebars: ONLINE: More safety knowledge 7 misunderstood points on safety and partial stroke valve testing

Valves are expected to operate even if they haven’t been operated for awhile. To ensure proper operation in emergency shutdown (ESD) service, valves must be actuated. To do so, most processes must be interrupted—inconvenient at best, costly at worst, but certainly not as costly as a valve actuation failure in a safety instrumented function (SIF).

Integrating partial stroke testing (PST), if processes allow, can stretch duration between tests by one to three years and reduce the probability of failure on demand, said Sandro Esposito, global product marketing manager for Dresser Masoneilan. Esposito spoke on the subject at the Yokogawa Corp. of America 2008 User Conference & Technology Fair in April.

Depending on the process, documenting and doing the system integration for that kind of change can save millions of dollars in hardware and maintenance costs while increasing process efficiency, Esposito said. Because of the change in system design and integration of varied practices, people may be unwilling to consider partial stroke testing. Attitudes may change with greater understanding of IEC 61508, IEC 61511, and ANSI/ISA-TR96.05.01, Partial Stroke Testing of Automated Block Valves, recently approved by ISA and registered with the American National Standards Institute.

When Esposito hears of resistance to new ideas, he quotes Albert Einstein, who said: “The world we have created is a product of our thinking; it cannot be changed without changing our thinking.”

Actuation of the process isolation elements (the on/off valve that brings the plant to a safe state) represents the weakest link in the ESD system, Esposito said.

Sources of error are many, including trim wear (valve leakage), loose or seized trim, eroded valve seat, broken shaft or stem, bearings worn or contaminated, bonnet or packing leakage, damaged or leaky actuator diaphragm, broken actuator springs, leaking piston seals, clogged venting or filling ports, damaged mechanical feedback, and air contaminations to accessories (such as solenoid and booster).

Partial stroke testing, or PST, can improve the risk reduction factor and decrease the probability of failure on demand (PFD). In many processes, such testing can be done without disturbing the process.

Testing advantages

Esposito said a fully enabled system that allows smart ESDs can:

Lower testing cost;

Decrease probability of human errors due to manual testing;

Automate compliance documentation, ensuring compliance with safety standards;

Improve availability of ESD valve;

Increase duration between shutdowns; and

Automatically document a safety event.

Comparing risk between annual full stroke testing and partial stroke testing four times a year shows that safety of the integrated system may increase from SIL2 to SIL3. In addition, Esposito said, a smart design may reduce the number of required valves for the safety integrated function.

The graph above shows how partial stroke testing can increase safety, as well as the interval between partial stroke testing from one to two years.

Comparison, full and partial

Without the SIL increase shown in the graph, Esposito said PST could stretch the duration of full stroke testing to three years, as the following analysis shows.

Full-stroke testing assumptions:

Diagnostic coverage (DC) factor is 100% and time interval (TI) is 1x / yr.

PFD avg = [(DC)(ld)(TI/2)]= (1)(ld)(1/2)= 0.5(ld)

PST assumptions : DC is 70% for partial stroke and 100% for full stroke; TI is 6x for partial and 1x/3 yr for full.

PFD avg =[(DC)(λ d )(TI/2)] p + [(1-DC)(λ d )(TI/2)] F

=[(0.7)(λ d )({1/6}/2)]+[(1-0.7)(λ d )(3/2)]

=0.05 (λ d )+0.45(λ d )= 0.5 (λ d )

The results of the two analyses are equal. A more traditional system relies heavily on mechanical, pneumatic, and electro-pneumatic elements, Esposito explains, while a smart field device has an embedded microprocessor and sensors to increase reliability and provide device health information. A smart ESD is better than solenoids because: it provides information rather than data; can throttle rather than deliver just on/off pressure; and the pneumatic train is always under control versus saturation. Also, the PST is independent from shutdown mode; with a solenoid, it may not be apparent if it’s stuck open or closed. Information is bi-directional, allowing accessories testing, automatic documentation, and continuous self-testing.

Complexities in the solenoid andemergency shutdown valve contribute to making them the weakest link in the safety instrumented function.

Health, diagnostics

A smart PST can reliably move an on/off valve between shutdowns. It has remote communications for reporting health and tests according to the IEC 61508 safety standard. Such communications improve ESD valve integrity and confirm availability.

After a PST, the system flags the software, saying that a new signature is available. While 30% of travel may be possible with some systems, depending on the process, 20% of full range of movement is a typical target and generally a good number, Esposito said. System diagnostics can uncover things such as a clogged vent or tubing, an air leak, pneumatic difficulties, a stuck valve, broken springs, or ancillary difficulties, such as stuck booster.

Automated partial stroke testing can increase safety and extend time between full stroke tests. Depending on the process, this can save a boatload fo money.

Knowing the equipment health can help improve safety instrumented functions. Integrity relates to information of critical components and response time to components’ needs. While partial stroke testing can extend the interval between full stroke tests, it cannot fully replace them. “Diagnostic coverage is 60% to 70% because partial movement doesn’t test the valve seat, for instance,” Esposito said.

Wiring options include traditional discrete, discrete with a multiplexer, analog I/O modules with or without a multiplexer, and analog with HART I/O modules.

Analog safety demand (ASD) wiring is cost effective, using one pair of wires for safety function, diagnostics, and PST. The shutdown signature event also can be captured during a trip to document behavior of the ESD valve upon safety demand and provide the required full proof test signature. (See PFD calculation.)

The device also should allow efficient access to ESD information and allow integration to the system. Esposito recommends that PST and data collection be automated, so someone doesn’t forget (a test can be stopped if needed). For more insight, see the “7 misunderstood points.…” Esposito suggests looking at the wider picture in any implementation of partial stroke testing, beyond the technologies involved and into the processes. He cites Aristotle to drive home the point: “We are what we repeatedly do. Excellence, therefore, is not an act but a habit.”

ONLINE extra: More safety knowledge

Coming up in November, a Control Engineering article examines process safety and sensors. In the meantime, take a look at recent resources.

-A June 2008 online article provided links to process safety whitepapers .

-May article: Safety Sensors Rise to New Heights

-April Webcast: The Top 5 Things You Need to Know About Process Safety

-A March article: Tale of 2 Applications tells how safety controllers deliver safety system integration for a fiery roller coaster and a beverage can installation .