USING WEIBULL ANALYSIS TO PREDICT VALVE REPLACEMENT

The five-year replacement interval for relief valves isn’t a practice that gets much scrutiny. It’s spelled out in IIAR Bulletin 110, and a provision establishing the need for a common interval appears in the Boiler and Pressure Vessel Code, making it a generally accepted “no-brainer” for much of the industrial refrigeration industry.

At its core, that five-year estimate for optimal replacement of valves is just that, an estimate. And it’s an estimate that raises questions about how relief valve replacement intervals can impact optimal safety, efficiency and economics in any facility.

Until recently, no real mechanism for quantifying those replacement intervals – and the potential benefits of extending or shortening them – has existed.

But a new analysis tool, called SRVcalc, may prove to be a powerful way to do just that, with long reaching implications for how the industry implements generally accepted practices that have been on the books for years.

SRVcalc is a tool intended to statistically analyze post-mortem relief valve test data to refine the replacement interval of relief valves for a given installation. SRVcalc is the byproduct of a research project funded by ARF and overseen by the IIAR Research Committee. The principal investigator for this project was Dr. Frederick T. Elder of Frederick T. Elder and Associates, Madison, WI.

“Like all IIAR research projects, SRVcalc is intended to fi ll in the gaps in our knowledge base,” said Bruce Nelson, IIAR Research Committee Chair and President of Colmac Coil Manufacturing Inc.

“Up to this point, we’ve really just depended on what is, essentially, a very general rule-of-thumb that specifies a five-year replacement interval regardless of facility, type of vessels, location or environment. SRVcalc gives contractors and end-users a way to acquire some good science to determine that frequency based on real test data from real valves in specific locations.”

The ability to finally quantify replacement intervals could pay off for end users and any facility looking to focus on safety and efficiency, said IIAR President, Dave Rule. “The potential advantage for this tool is that the five-year interval may be too soon [for some facilities], the possibility of extension is different for everyone. This is good because it’s a way to determine which direction an individual company might need to head. It might help you see opportunities to use a different grade of valve, or it may tell you you’re right where you need to be.”

“The status quo is always safety and the protection of vessels, it’s just a matter of how you go about managing that system,” said Rule. “It’s a benefit to understand the capabilities of the equipment offered in the industry at this point.”

Nevertheless, IIAR’s Nelson stressed that SRVcalc is in no way an attempt to do away with the accepted five-year replacement interval, but rather it provides a foundation for those looking to apply the second requirement for determining replacement as laid out in IIAR Bulletin 110.

According to Bulletin 110, Guidelines for Start-up, Inspection and Maintenance of Ammonia Mechanical Refrigerating Systems, “pressure relief valves should be replaced or recertified in accordance with one of three options: 1) Every five years from the date of installation; 2) If an alternative to the prescriptive replacement interval, i.e. fi ve years, can be developed based on documented in-service relief valve life for specific applications using industry accepted good practices of relief valve evaluation; or 3) [According to] the manufacturer’s recommendations on replacement frequency.”

“That second provision will be an attractive option for end-users, now that a way to conduct the kind of evaluation it specifies exists,” said Nasser Karimzadeh, Department Manager for Mechanical and Refrigeration Engineering for Design Group Facility Solutions.

“Of course, the number one issue is safety, but the second issue here is cost,” said Karimzadeh. “From an end-user’s perspective, [replacing valves on a less-than optimal interval] presents a rather substantial cost. If that interval is too frequent, you’re wasting money and taking away from other repair and maintenance functions. It’s best to have the proper cycle for replacement.”

That optimization of safety and efficiency are often the driving forces behind any effort to quantify a commonly used practice, and SRVcalc should be seen by IIAR’s members as an attempt to further those goals, said Eric Smith, IIAR Vice President and Technical Director.

“This technique allows us to quantify replacement intervals for the first time in the ammonia refrigeration industry,” he said. “Having the ability to finally drill down and get these specific measurements is valuable.”

Smith added that the real value in using SRVcalc, beyond the primary emphasis on improving safety, exists in three different arenas: finding potential cost savings, assessing environmental conditions, and maintaining flexibility in meeting regulatory requirements.

First, finding cost savings in facilities or companies that replace large numbers of relief valves is an obvious benefit. “When it comes to replacement, there’s expense for time, labor and documentation, and if you can extend intervals safely, then why wouldn’t you want to do that?” he said.

Second, the tool gives users the ability to assess valve life in unique environments, such as those in an area close to the ocean, for example, where salt in the air can impact an operating environment.

Third, the tool allows users the ability to retain flexible decision making, while still meeting regulatory requirements. “This is a way to substantiate why a facility may be following the prescriptive five-year relief valve replacement interval,” said Smith. “It allows facilities to pick their method and gives them flexibility in how they meet code and regulatory requirements.”

In order to generate data on relief valves, SRVcalc relies on a two step process. First, valves at the end of a specified use-period must be tested as they come out of service. ARF funded the first phase of this research that entailed developing appropriate facilities and protocols for the post-mortem testing of relief valves. The IIAR Research Committee provided oversight of this project and Dr. Todd Jekel along with Professor Douglas Reindl both of the University of Wisconsin Madison’s Industrial Refrigeration Consortium were the project’s principal investigators.

“As valves come out of service, they are put on a bench test rig to see how they function,” said Reindl. “Then, that data gets logged and input into statistical analysis software.”

The statistical analysis, the second part of the tool, allows a user to arrange data based on equipment type and generates optimal intervals for replacement.

Because data sets can often be small as batches of valves come out of service slowly an appropriate statistical analysis methodology was needed to predict the service life-expectancy of relief.. For that, Dr. Elder – who is an adjunct professor at the University of Wisconsin, Madison, and president of Frederick T. Elder and Associates – turned to a kind of statistical modeling called the “Weibull Analysis.”

“We needed to build a tool that could take into account and measure with accuracy the frailties of different environments and subsystems,” said Elder. “SRVcalc deals with small data sets and can take into account low failure rates. It deals with a lot of issues we have when we’re measuring the lifecycle of a valve, and generates accurate data over time.” (Read more about the technology behind this process in the technical paper section that appears at the end of this issue of the Condenser).

As SRVcalc was developed, “what really became apparent is that you can’t generalize with this type of data,” said IIAR’s Nelson. “This tool is intended for use by the specific end user in their specific facility. It makes measurements that depend on a range of unique variables, including climate, installation, etc. Relief valves exist in a variety of places, you can have them in a nice dry, clean engine room, or on a rooftop exposed to the elements. The analysis tool is intended to give a picture of relief valve life in local facilities.”

While SRVcalc may prove to be a valuable tool for some in the coming years, the impact it has on how the industry views relief valve replacement intervals may be longer term, said IIAR’s Smith. He added that if long term use of the tool is widespread enough, the industry may eventually have enough data to take another look at its five-year requirement, but that possibility is far off.

For now, SRVcalc simply exists as another engineering tool developed by the industry to help meet the unique needs of its members, said Nelson. “It’s giving the industry a tool to determine in a defendable way, a method for potentially saving money by finding the optimal interval for safety relief valve replacements on a facility by facility basis.” The IIAR Board of Directors is currently reviewing how it will distribute the tool, said Nelson, adding that SRVcalc will be made available to IIAR members sometime in the next twelve months.