Interview with John Nyholt, BP NDE Specialist

John Nyholt has primary responsibility for NDE consulting and troubleshooting for BP around the world in the refining, chemical and gas processing industries. We at Inspectioneering Journal thought it might be valuable to spend some time chatting about his background, the challenges he has faced recently, and what he feels are some of the biggest challenges ahead for the Inspectioneering community. We hope you enjoy the interview.

John, it is always interesting to hear about how someone got into the inspection profession. Will you tell our readers a little about your background and how you became a worldwide resource, providing NDE direction for BP, often at crucial times?

John Nyholt (JN): After earning my AAS degree in Non-destructive Testing from Moraine Valley College, I started my NDE career working at the McDonnell Douglas Corporation (now Boeing) in St. Louis, Mo. I worked for their fighter aircraft division from 1982 to 1991. The work was interesting and in a technologically innovative environment. In 1991, I went to work for Amoco (acquired by BP in 1999) at the Whiting, Indiana refinery. In 1994, I moved to the Amoco Research Center in Naperville, IL as an NDE specialist for the corporation. My office is now at the Westlake complex in Houston, TX. I also have an NDE lab offsite. As BP’s mergers and acquisitions over the last four years brought expanded territories and workloads, I don’t get to spend as much time in Houston as I would like.

In the earlier years of my current position, I created course content for and delivered internal training for routine NDE methods used within the corporation and handled special NDE applications as they arose. Some of these tasks were daunting, but were usually successful with the help of some excellent mentors in our central engineering team. Many are engineers at the top of their fields in metallurgy, corrosion, equipment design, and project management. We worked well collectively and were willing to answer late-night phone calls. This helped me tremendously in understanding the behavior, morphology, and anticipated locations of damage mechanisms. This type of understanding is very important in the effective development and application of NDE.

There is little time for me to conduct in-house NDE training now. My current duties are to tackle the tougher, non-routine jobs. I perform less of the routine day-to-day NDE consulting and deal more frequently with various NDE vendors and consultants involved with special NDE applications. As many of the applications involve emergency shutdowns or achieve continued plant operation through on-line NDE monitoring of equipment damage, some jobs become high profile. I’m prepared to cover a couple of those in today’s interview.

What is the first case that comes to mind?

JN: It has to do with high-pressure, heavy wall reactors nozzles at one of our gas processing facilities. 16 heavy wall reactors (molecular sieve dehydrators) cycle from 100 PSI to 1200 PSI, and from 90F to 700F three to four times per day. The vessel heads are carbon steel, 4.5 inches thick.

Internal cracks had developed in the nozzle to head welds and HAZ’s of the top and bottom nozzles and the top manways. Finite element analysis indicated the cracks/leaks are due to the combination of thermal shock and stresses. A risk analysis showed high potential consequences with knock-on effects if any of the reactors were to fail catastrophically. However the same study indicated that the cracks would leak before failure. The central engineering group elected to keep the vessel trains in service and use on-line ultrasonic monitoring as the best way forward.

The bottom nozzle is centered in the head. However the top head nozzle and manway were off-center by design, creating complicated weld geometries and difficulty for obtaining accurate and repeatable NDE crack detection and sizing. As with many pressure vessel nozzles, original weld flaws were present that complicated both the NDE and FFS assessments.

An NDE company offering robotic automated ultrasonic shear wave inspection initially performed AUT on one of the top nozzles. However the robotic scanner was too complicated to install and operate within the 45 to 60 minute window of opportunity for ambient temperature work. A crew of 3 (system operator, scanner technician, and helper) spent 4 days inspecting one nozzle at a cost of over $10,000. As there were 48 nozzles and manways to inspect for each inspection Sept/Oct, 2002 Volume 8/Issue 5 5 interval, AUT was considered cost prohibitive.

I was given R&D funding and asked to develop and personally perform a successful and cost effective NDE test. I had a miniaturized semi-automated nozzle scanner built that would ride on the nozzle neck and a 12-inch cantilevered scanning arm that controlled an encoded transducer holder that scanned the head to nozzle weld from the adjacent head. By semi-automated, I mean a manually manipulated 2-axis scanner that collects the same X and Y positional data and triggers the pulser-receiver in the same manner of a motorized scanner. The 6 x 3 inch scanner worked with a single-channel UT system the size of a small suitcase and a single 15-foot cable. A 2-gallon couplant jug provides automatic couplant feed by a hand pump and needle valve control. The primary ultrasonic probe for this application was a 2.25 MHz focused beam dual 45-degree shear probe. The nozzle or manway is first ultrasonically imaged, followed by manual UT while referencing the current and previous UT images for that particular nozzle or manway on a laptop computer. A key element of this test was its ability to differentiate between original fabrication defects service-induced cracking. A manual ID creeping wave probe helps identify indications with ID connectivity; focused shear wave probes help study crack morphology (shape) and verify linkages to original weld flaws.

With the new technique, a one-man UT crew (with one local laborer) was able to inspect 48 nozzles and manways in 10 twelve-hour days. Spreadsheet reporting of flaw types, sizes, and locations for each nozzle and manway took an additional two days. With everything working right, the average time to inspect a nozzle or manway was one hour, allowing all work to be completed within the ambient temperature time window. The low complexity of the system also avoided downtime typically experienced with fully automated scanners and couplant feed systems. The cost of the UT system: $30,000.

We now monitor these reactors on a semi-annual basis.

John, in general, how do you feel about the reliability of nozzle weld inspections to find flaws in the field, i.e. from the majority of service providers in the market?

JN: In the case we just discussed, we had tried both manual and automated UT vendors that were available locally and abroad. I was not very satisfied. However it is was not the NDE operator’s fault. Nozzle weld geometry, access to only one side, and non-relevant indications make this type of inspection difficult.

As training, certification, and qualification exams are major overhead cost burdens for NDE service companies in a highly competitive market, the typical ultrasonic shear wave inspector does not see much specialized training. On-stream nozzle inspection is only now being practiced routinely. However the industry’s ultrasonic shear wave training and qualification programs are still working on carbon steel butt welds, and struggling to find qualified UT operators at that. The minimal standards for developing NDE technicians has and will continue to force the end-users of NDE services to supplement training and hands-on qualification tests with programs such as the API UT Contractor Qualification exam. I would expect to see similar programs for other NDE methods in the near future.

What do you feel are some of our biggest challenges today?

JN: Again, the number one problem is finding NDE technicians that are qualified to do the work. Industry certification programs, such as ASNT, set minimum training requirements that don’t assure the levels of quality and expertise that (according to articles in their own and other industry publications), is now hard to find.

We as owner-users are partly responsible (i.e. wanting a cheap service). Both direct and indirect inspection costs, (NDE crew/ equipment and unit downtime) impact our profits. When we demand faster and cheaper inspections, NDE service companies respond by giving us what we asked for.

One casualty of this formula has been the entry-level NDE technician. Competitive pressures and high turnover rates have resulted in low starting pay and minimal investment in this person. Starting pay, raises, training expenses and capital investment in equipment all have to be kept low to hold down charge out and equipment rates. Subsequently, technician turnover rates are high, and investing in the technician’s future or new technology can be an unacceptable risk. Some NDE companies, to their credit, try to pay and hold their people to a higher standard, however many of them soon find themselves at a competitive disadvantage. As for the entry-level NDE technicians, we find that fewer and fewer people are selecting NDE as a career choice, making it difficult for an NDE company that wants to be world class find the right people.

Valuable insight and one our industry has been struggling with for a quite some time. I hope we are able to build the case for quality, soon. This grows more important each year as plant equipment is aging.

Back to examples. Any others you would like to share?

JN: I would think that the inspection of heavy wall hydrotreater rectors might be interesting. We inspected 3 of 12 of heavy wall reactors to meet a 20- year inspection requirement recommended by the manufacturer, Japanese Steel Works (JSW). We opted to perform 100% of the JSW inspection recommendations on the reactor which saw the most severe operating conditions and represented the highest potential risk, and perform a limited internal inspection on two sister vessels.

These heavy wall reactors have 100% austenitic (wide ribbon sub- arc) weld overlay (WOL) over a 12- inch thick carbon steel base metal. The original inspection plans called for a liquid dye penetrant inspection of all WOL at attachments welds, vessel weld seams, and any WOL area suspected of disbonding. Disbonding was to be detected by a zero-degree ultrasonic inspection from the vessel OD. All vessel shell welds and head to shell welds were to be 100% AUT shear wave inspected. Nozzle welds were to be 100% manual UT shear wave inspected.

Overall, the inspection plan was a good one, however it expanded the downtime of the unit and was quite expensive to implement. The NDE companies were prepared to do what they came to do, however they typically do not feel it is their responsibility to recommend better inspection alternatives. In-house consulting is typically needed to make that kind of recommendation to management. With so much of the NDE being performed from the vessel OD, scaffolding costs for both the ID and OD of the vessels was expected to be $750,000. If the NDE techniques could be performed from the already scaffolded vessel ID, considerable time and costs could have been saved.

• WOL Dye Penetrant Inspection: The WOL had an extremely hard .009-inch thick iron carbonate deposit that could only be removed by sand blasting. However residual sand from a sandblast could potentially damage the large pumps at the bottom of each reactor, and would have added two days to the shutdown. Phase sensitive surface eddy current testing (PS-ET) was recommended as an alternative surface test. However the surface roughness, and material permeability variations were a concern. We selected a 2-channel, hand held eddy current instrument with the ability to “mix out” material permeability variations via a frequency mix that proved capable of detecting cracks without sand blasting. A mock-up of the vessel wall was built and used for the qualification. ET results were instantaneous, further reducing unit downtime for inspection. A crew of 6 ET technicians completed the internal examination of all three vessels in 1-1/2 work shifts. No relevant indications were detected.
   
• WOL disbonding: If zero-degree ultrasonic inspection for WOL disbonding were performed from the vessel OD, considerable exterior scaffolding would have been required. The UT technician would also have been required to discriminate 0.375-inches of WOL from over 12-inches of metal sound path. The ID surface of the vessel was examined for the ability of the WOL to adequately transmit ultrasound. If the ultrasonic bond testing could be performed directly on the WOL, the technician could much more easily resolve a .375-inch deep disbond indication against a 12-inch back wall sound path. The ultrasonic UT examination to detect disbonding was performed from the interior with no relevant indications detected.
   
• UT-SW of the reactor welds: A UT technique for examining the welds from the ID and through the WOL was considered, however there was no time to develop and validate such a test (if one could be validated). A limited amount of scaffolding was erected for AUT and manual UT from the vessel OD. Also, it’s notable that the vendor performed manual UT on the vessel nozzles, as they did not offer AUT for their configuration. As stated earlier, nozzles subject to service- induced flaws should be ultrasonically imaged in order to detect, differentiate, and baseline original versus service induced flaws.

The surface eddy current test found yet another application in locating flushed shell welds. The weld crowns were originally ground smooth and were not visually evident. Acid etching would normally have been used to detect welds by their fusion zones, however the hand held eddy current instrument with a cross-axis coil was easily able to detect these welds by scanning a vertical line on the shell and detecting the local change in material conductivity at the weldment.

I heard about an instance of external chloride SCC in some stainless steel piping in a BP refinery on the Texas Gulf Coast, and that you came up with a great solution for finding it. It seems there was a lot of piping with cracking that could have been anywhere resulting in a unit shutdown that cost the refinery a lot of money for each day of downtime. I also understand the piping was insulated. Will you tell us about the problem, the application, the challenges, the NDE chosen, logistics and results?

JN: That would be a good one to tell. Perhaps many of your readers had already heard of it, as it was probably the largest surface inspection project in our industry’s history.

The refinery is right on the coast and had experienced many sources of chloride and aqueous water ingress into the calcium silicate insulation of five operating units with mostly austenitic stainless steel pipe. This resulted in the stainless steel pipe version of corrosion under insulation, i.e. chloride stress corrosion cracking.

A small diameter pipe on one of these units exhibited a hissing leak that was initially thought to be a steam leak from steam tracing. However an alert operator noticed that the leak was not the right color. The area was de-insulated, and visually examined to find a small hydrogen leak in the pipe adjacent to where a steam leak had occurred. The line was shut down and the section of pipe removed for analysis. Even though the leak area under microscopic magnification verified the presence of a crack, visible dye penetrant and ultrasonic shear wave could not detect it.

That’s when I’m called in. I remember that I was snowmobiling in Colorado with my two boys when I was surprised to see that my cell phone was able to receive a call high up in the mountains. My vacation came to an abrupt end as I was asked to be at the refinery by the next morning.

When I arrived, I was shown the samples and asked to come with an appropriate inspection technique for detecting and sizing chloride stress corrosion cracks. The plant inspection department had spot inspected various piping circuits with visual dye penetrant and found many indications large enough to be detected by that method. It soon became evident that several miles of stainless steel pipe would either have to be replaced or saved by inspection and repairs. As these types of pipes, fittings, and valves were not readily available, a massive inspection, grind out and repair effort was about to take place. The job lasted 4 months.

I was given an acceptance criterion of “no cracks of any size and depth”. As you would expect, that put tremendous pressure on the NDE technique and NDE crews. A scratch can look like a crack, right? Given the criteria and extent of work, phase sensitive eddy current testing was selected as the primary inspection method because dye penetrant testing was labor intensive, and needed surface polishing in order to find small, tight cracks. Ultrasonic shear wave testing using a 45-degree full V-path looked viable, however very small cracks could only be seen in the baseline noise of the instrument display and were lost altogether for some stress orientated cracks.

The next step was to search for as many level II eddy current technicians as we could find. That was difficult as it was tubing season and there are only about 60 level II techs in the US to begin with. We initially found about a dozen that showed up with a menagerie of instruments, probes, cables and probe adapters. By the second day of testing, many of these unique pieces of equipment started to fail, putting the $75/hour guy in the trailer waiting for repairs, and yes, we had to have a full time probe and cable repair guy for the duration of the project.

Ongoing R&D efforts tried to help the situation. We experimented with many hand-held instruments and probe designs, and eventually selected a single instrument, probe and cable, and set-up configuration for each technician. As equipment charges and rental fees mounted and some service providers refused to buy standardized equipment, we justified buying 18 instruments ourselves and training each technician on how to use them. More ET technicians joined the project until we reached a peak of 56 technicians from 12 companies working two twelve-hour shifts. Two large work trailers were brought in to house the technicians and equipment. We had ET technicians from petrochemical, nuclear, aircraft, and the military. Some were from Canada, South America, and Europe. Some took one look at the unit and saw the physical and mental stress involved with the work and immediately left the plant.

As all unit inspectors should, the BP inspection department began validation (10%) of the ET results with polishing, dye penetrant, and 100X pocket magnifiers. They found many missed cracks, and many of the indications found by ET were false. Further investigation revealed several sources of error:

• The austenitic stainless steel had a wide range of permeability.
• Many areas of the pipe were fully ferromagnetic,
• Iron oxide deposits were present
• The natural hematite (iron carbonate) layer on the pipe surface was slightly ferromagnetic.
• Cracks tended to start at a slight nick in the oxide layer, and spread subsurface from that point. This complicated the eddy current test as subsurface indications have a unique phase shift that can be misinterpreted as a permeability indication.

The ET technicians were being pressed to circle any indication, and we used a special ET crew to buff the area and look again. We eventually required 100% flapper wheel buffing of all surfaces, however much of the permeability and surface anomalies remained. The data validation crews continued to find missed cracks from our hand scanning miles of pipe with probes the size of a pencil. Each of these areas had to be re- tested and some technicians had to be “run-off” the job. Plant management commented that the cost and time of these inspections were giving them a reason to just order and replace all of the piping. It became apparent that it was my time in the barrel.

This struggle went on for weeks until we finally decided to use the two-channel frequency mix feature on these instruments. New probe designs were also improving. We successfully cancelled out much of the permeability and surface anomalies that caused false inductions. We were also able to stop pipe buffing surface preparation all together, and test the pipe in the as-stripped condition. The number of circled indications dropped with each proving out as a crack. Missed cracks were seemingly eliminated. A final probe design, delivered just a week before the end of the project, was the size of a credit card contoured to each pipe diameter, and had an active coil area of nearly a half an inch (five times the diameter of the previous surface testing probes). By the time the last operating unit was inspected, we were able to inspect 5,000 feet of pipe in 2 1/2 days.

This last challenge was a great example of real-world innovation and of being under tremendous pressure to perform. I guess the addage, “Necessity is the mother of invention” holds true again. Also, it must have taken a lot of coordination to make the project a success.  Thank you to John for taking the time to share some of his knowledge and experience with us today. I am sure our readers will find his insight both practical and valuable and we look forward to hearing from him again in the near future.