Westinghouse Electric Company, Arizona Public Service and Exelon have developed and qualified a technique that simplifies inspection of pressurised water reactor (PWR) U-tube-style steam generator tubes. The technique reduces the outage window required, lowers costs, reduces risk to inspection personnel, lowers the risk of human performance events and is likely transferable to other inspections. The team’s achievement won a 2017 Top Innovative Practice award from the US Nuclear Energy Institute.

Inspections of expensive, critical plant assets are necessary to protect against costly asset damage and minimise unplanned outages due to preventable equipment malfunctions. For steam generators, avoiding a tube leak or rupture is vital to continued uninterrupted operation. Eddy current examinations along the length of the tubes are conducted regularly during planned outages to inspect for various forms of steam generator tube degradation, including foreign-object-induced wear.

Inspecting the tubes is a considerable endeavour. Depending on design, steam generators have between 3000 and 16,000 tubes. Their inspection challenges plant outage schedules, outage risk assessments, radiation protection, personnel dose and overall outage costs. This is due not only to the size and number of tubes, but also the difficult-to-inspect areas within them – especially within U-tube style steam generators. These include a transition at the U-bend and from the expanded tube-in-tubesheet region to the unexpanded tube above the top of the tube region known as the top-of-tubesheet expansion transition, as shown in Figure 1.

Bringing out the best

The high-speed (up to 120 inches per second) bobbin probe is the one used to inspect most of the tubes’ length for foreign-object-induced wear. The new method qualifies it to adequately inspect the top-of-tubesheet expansion transition region for the presence of this wear.

The bobbin probe was formerly thought incapable of detecting foreign-object induced wear in this geometrically changing region, which meant that for each steam generator tube, an additional inspection using a probe capable of detecting wear in the expansion transition region was required to complete the tube’s inspection. Typically, a +POINTTM probe (rotating pancake coil probe) or X-PROBETM multiple coil array probe) is used to complete the inspection. Elimination of these additional inspections is a significant cost, time and dose savings.

Qualifying the use of the bobbin probe with more sophisticated analysis processes to reduce or eliminate these additional eddy current inspections has provided tremendous savings to the utilities.

If these methods are applied at other plants with U-tube style steam generators, potential savings on a per outage basis could include:

  • Cost savings of $200,000 to $500,000
  • In-generator time savings of up to 48 hours
  • Fewer eddy current analysts
  • Fewer inspection robots.

Conventional approach

To perform the tube inspections, a “probe pusher” is positioned on a platform near the steam generator manway opening. From here it pushes and retracts the probe using a mechanical drive system.

Inside the steam generator, an inspection robot guides the probe under each tube prior to that tube’s inspection cycle. The pusher and robot are connected via a flexible conduit. As it travels through the tube, the probe emits a magnetic field that passes through the tube wall. Any imperfection in the tube wall disrupts this magnetic field and a signal is observed by the analyst. These signals are recorded and analysed to determine tube integrity. Throughout each tube’s inspection, the probe head, cable, conduit and robot each experience stress from pushing and pulling. Used probes become radiological waste due to the radiological contamination transferred from the tubes’ inner surfaces to the probes.

Using supplemental probes for the top-of-tubesheet expansion transition area complicates this effort – from the steps required to change the probe setup, to increased stress on the probes and associated equipment, to increased dose to workers. The extra work represents a significant portion of the overall outage cost for the inspection, but typically yields few degradation indications. In addition, it increases the time required to complete the inspections, causes more frequent replacement of probes and associated equipment, results in higher costs for disposing of the additional contaminated solid waste products and increases the amount of time workers must spend in the high-radiation environment.

Revealing wear with bobbin probes

The standard bobbin probe is energised using multiple frequencies. These frequencies can be combined to improve the bobbin probe’s ability to detect specific types of degradation. For example, mixing high and low frequencies can improve the bobbin probe’s detection of axially oriented stress corrosion cracking at tube support plate intersections. Mixing high, mid and low frequencies (three-frequency mix) can improve the bobbin probe’s ability to detect volumetric degradation at geometric changes in the tube. However, standard mixing techniques are limited when it comes to refining the bobbin probe’s ability to detect volumetric degradation at the top-of-tubesheet expansion transition region.

To improve these outcomes, Westinghouse introduced two methods that can be applied singularly and independently, or combined, as the situation warrants.

They are implementation and analysis of a three-frequency mix channel or application of the Westinghouse Automated History Comparison software. This software compares bobbin probe examination data from two different outage inspection campaigns and extracts from them the signal that illustrates a change between the data sets. The process has been shown to result in improved volumetric degradation detection capabilities above the use of different mixing processes. Westinghouse supports these applications with its Real Time Auto Analysis software to measure the noise condition at the expansion transition region.

Quantifying detection capabilities

To show the improvement in the detection capability of the bobbin probe in the challenging transition region using these methods, the detection capabilities of the methods had to first be quantified.

For this, Westinghouse generated digital flaw samples by combining eddy current loose-part-wear scar signals from laboratory samples with tubesheet expansion transition eddy current signals from an operating unit using the Westinghouse Data Union Software (DUSTM). The software merges two digital signals – in this case a non-flawed expansion transition signal and a flaw signal (from the laboratory specimens). The resultant signal simulates in situ eddy current signals. Figure 2 illustrates this process for the three- frequency mix channel.

The resultant signals are analysed using either the three-frequency mix or the Westinghouse Automated History Comparison software to quantify the detection capabilities of each method. The volumetric degradation depth that is required to be reliably detected by either method is a function of the individual plant operating time between eddy current inspections, steam generator tube geometry, steam generator tube material properties (yield and ultimate strength) and individual plant performance criteria as defined by the Nuclear Energy Institute’s (NEI) NEI 97-06, “Steam Generator Program Guidelines.” Plants permitted to operate for an extended period between inspections must detect flaws of lesser depth (physical wall loss amount) than plants with shorter operating periods. 


Arizona Public Service’s Palo Verde 1 was the first to apply this method in April 2016. The method involved analysing the bobbin probe three-frequency mix to achieve improved volumetric degradation detection capabilities at the top-of-tubesheet expansion transition region. A study was performed in early 2016 by Westinghouse to show that the bobbin probe’s detection capability using this method was adequate to satisfy industry tube integrity requirements. The study was performed using three separate three-frequency mix noise conditions. For those locations with a mix noise that exceeded the threshold value defined by the largest noise used in the study, a supplemental probe examination was performed. Very few were needed.

Exelon’s Braidwood 1 was the second to apply this method. In this application, Westinghouse also applied its Automated History Comparison software. With this software, inspectors were able to show that foreign-object induced tube wear degradation depths which satisfied the Braidwood 1 operating period evaluation (for the time period between inspections) could be reliably detected.

An example of signal extraction using the Automated History Comparison software on one of the digitally generated specimens is shown in Figure 3. The software was qualified according to the standard defined in Appendix H of the Electric Power Research Institute’s “Steam Generator Management Program: Pressurized Water Reactor Steam Generator Examination Guidelines: Revision 7,” which provides requirements for examination plans and processes necessary to meet the performance criteria of NEI 97-06.

For each qualification, Westinghouse used the Data Union Software to combine donor flaws that represent several different flaw morphologies with various noise level conditions for the top-of-tubesheet expansion transition region. The process provided resultant signals that were used to determine a threshold level of the depth of degradation penetration into the tube walls. The detectable depths of degradation were sufficient to establish that tube integrity requirements were satisfied up to the next scheduled eddy current inspection. Without the Data Union Software, the qualification programme would have been too costly to pursue, since without this digital option it would have been necessary to prepare hundreds of physical specimens.

For both the Palo Verde and Braidwood applications, the depths of degradation required to be reliably detected were less than those required to be detected according to the Appendix H standard. Thus, the applied processes not only resulted in significant savings, but achieved flaw detection capabilities greater than required by the industry standard.


Both Braidwood and Palo Verde had very tight outage windows. Being able to fully inspect most tubes using the bobbin probe allowed the teams not only to meet the aggressive schedules, but to save many hours of the critical path outage time allotted.

During the Palo Verde 1 outage, using the bobbin probe to detect volumetric degradation at the top-of-tubesheet expansion transition region reduced the number of supplemental probe examinations from approximately 8000 (two steam generators with 4000 tubes each) to only four. Six additional supplemental test locations were conservatively included for a total supplemental test programme of only 10 locations. This saved the team from changing the bobbin coil probes to supplemental probes for reentry into the tubes an additional 7990 times, along with the increased dose and related potential human performance events.

Reducing the supplemental probe examinations so dramatically also saved the Palo Verde unit 24 hours of critical path outage time. That meant around 20 percent overall project schedule reduction and 85 percent savings in supplemental probe use. This allowed the utility to realise financial savings of about $500,000.

Dose reductions could not be measured during the outage because there were higher than usual anticipated dose rates in general at the plant’s work sites. However, based on traditional levels, the dose saving was anticipated to be roughly 50 milli-Roentgen equivalent man (mREM); use of this method helped to limit the extent of the higher than usual overall dose.

At Braidwood 1, the number of supplemental probe examinations was reduced from approximately 10,000 for all four steam generators combined, to 952. Reducing the estimated 10,000 supplemental probe examinations to 952 saved the Braidwood unit 49 hours (a little more than 12 hours per steam generator) of critical path time and required approximately 42 percent less probes. The outage time and supplemental probe reductions alone saved the utility about $200,000. In addition, without frequent probe changeouts, workers experienced less radiation exposure – by 70mREM – and less probe use reduced the expense of processing contaminated solid waste.

While the Braidwood experience did not have as dramatic a reduction in supplemental probe use, with improved outage planning and analyst training programmes for analysing Automated History Comparison extraction results, it is likely that fewer supplemental probe inspections will be required for future campaigns at Exelon’s Byron 1 and Braidwood 1.

Additional Benefits

Reducing the many changes from the bobbin probe to other probe types ripples to many other associated improvements that benefit plant personnel and reduce overall costs to the utility.

These include enhancing productivity for analysts, who no longer need to switch reading data types to interpret results, and reducing the potential for errors during interpretation. Decreasing the physical effort to change the probes lessens not only dose but also the potential for safety events.

Another benefit stems from use of Westinghouse’s Real Time Auto Analysis software, which monitors for noise. An increasing trend in noise at the top-of-tubesheet expansion transition may indicate bobbin probe failure. This early warning can reduce the potential for missed wear indications. It may also be used to adjust the setup of the probe conduit to remove unintended influence on probe performance, resulting in improved overall eddy current inspection efficacy.

The benefits of qualifying this process for industry use can contribute to the reduction of maintenance costs for PWRs and U-tube-style steam generators worldwide. The successful application of the Westinghouse Automated History Comparison software shows promise that the higher-speed bobbin probe may be able to adequately inspect other regions of the tubes where inspections typically employ the slower X-PROBE or +POINT probes. In addition, this process could be applied to other eddy current inspections performed in the plant.

Based on the results of these initial applications, the power industry can reap significant savings with this enhanced ability to interrogate and analyse bobbin probe data. 

About the authors

By Westinghouse Electric Company Fellow Engineers William Cullen and Qui Le, Arizona Public Service Palo Verde Senior Consulting Engineer Douglas Hansen, Exelon Corporate Engineering Manager Harry L. Smith and Senior Staff Engineer C. Lee Friant