 Hello, everyone. Thank you for joining us today for a virtual meeting on the extremely low probability of rupture or XLPR probabilistic fracture mechanics code models overview. My name is Matthew Homiak. I'm the NRC's lead in the Office of Nuclear Regulatory Research for the XLPR program. XLPR is a joint venture with the Electric Power Research Institute or EPRI. My counterpart from EPRI, Craig Harrington, is also with us today. Supporting the meeting are members from the NRC staff, EPRI staff, and some of our contractors. The purpose of the meeting today is to provide an overview of the models in the XLPR version 2 code. This is the following one to our April 23rd XLPR code pre-release meeting. At that meeting, we announced the pending public release of the code and provided an overview of the code request process. We also provided perspectives from NRC and EPRI senior management and from those involved in code development from the beginning. We highlighted some of the code features and provided a live demonstration. We also discussed some areas of code applications present in future and announced plans for future series of webinars surrounding the XLPR version 2 code release. And the meeting today is the next in that series. The meeting on April 23rd was recorded and you can watch a video of it on YouTube.com. There's also an NRC meeting summary. If anyone at the meeting here today missed the April 23rd meeting, please feel free to reach out to me and I'd be happy to point you to some of those materials so you can watch it. Whereas the April 23rd meeting was more high level, today we'll be getting into some more details and technical content about the code. The goal is to provide technical details about the models programmed into the code so that prospective users can gain better understanding of its capabilities. Our agenda here for today is we're currently in the introduction and opening remarks portion. And I'll use the remainder of that time to tell you a little bit more about the XLPR version 2 point release and what that includes. Then we'll provide a general overview of the code. And we'll cover each of the deterministic models one-to-one. Those include the loads and stresses, stress intensity factors, fatigue and stress crushing initiation models, crack growth and crack coalescence, crack transition, crack opening displacement, leak rate calculations, crack stability, and in-service inspection. We'll round that out with dedicated time for questions and answers. And then we'll have closing remarks and adjourn. This is an NRC category 3 public meeting, which means the public is invited to participate by providing comments and asking questions throughout. Today we're using the WebEx platform to deliver the meeting. And I'd encourage everyone to participate through WebEx so you can see the slide presentation. And this also will primarily plan to take your questions. So you'll get the best experience using WebEx. And WebEx, you can submit your questions and comments at any time using the Q&A feature. If you want to make sure that's displayed on your screen, the Q&A feature looks differently depending on whether you're using WebEx in your internet browser or through your desktop client. For the WebBrother, you can access the Q&A by clicking on the question mark bubble located between the speech bubble and the three dots that's shown on the left here in the screen. For the desktop client, you first click on the three dots and then select the Q&A button on the right of the screen. Either way you do it, you should display a Q&A box on the right side of the screen. And this is just a note that we're using Q&A today. This is different from the way we took questions at the April 23rd meeting where we're using the chat feature. So remember, today we're using Q&A feature for questions and answers. We received a lot of excellent questions from the April 23rd meeting. In fact, so many so that we weren't able to keep up with all of them. So in response today in the background, we've got a five-person team to help answer any questions you may have, so feel free to ask away. After the meeting, we'll issue a summary, which will be available on NRC's public website. And we'll also be recording this meeting, and we'll make it available on YouTube.com as well. The cover is my administration for the meeting. And now a little on the XLPR version 2.1 release. So as mentioned, we had the pre-release meeting on April 23rd. We announced that we would be releasing the code, version 2.1. We have a slight delay in releasing the version. There's some technical issues we're trying to grapple with now, but it's very soon to be publicly available. And we'll issue an announcement when that happens. It can be available for request again on the NRC public website and to the EPRES public websites. And I'll go through each of those processes quickly after this. The code is free for requesters, but they'll need to sign a end-user license agreement and also meet certain citizenship requirements. Two-minute credits is approved. Instruction will be sent to access the code through a secure file transfer site. Now on the NRC side, okay, on the NRC side, this is a NRC public website. You go to the research on the left-hand side there. And that will display the content here in the middle of the screen. You click on the link there, attaining the codes. XLPR will be listed at the bottom. The instructions will be provided for access to the code. On the EPRI side, go to EPRI.com. You search for XLPR on the EPRI website. It will display a list of results. And then you would look for XLPR version 2.1 and select it from the list and follow the instructions on a page that would look similar to this. So this is not XLPR page. This is an example of sort of what it'll look like. Okay, now a little bit about what you're going to get in XLPR version 2.1. This is a snapshot of the actual release package itself. I'm going to go from the top down here. The first one being a folder. It's going to contain some input databases, materials, residual stress profiles, things like that. The DLLs folder contains all the modules that the code uses. Those are what we'll be covering today. Actually, what the actual physical models that are programmed into each one of those modules. The next folder is the SIM editor. That's the GUI for the code. There's a folder on training materials. Talk about that in a little bit. The other folder is XLPR main. That's the main folder that contains the Oltim files and input set spreadsheet. That's what you'll be getting in XLPR version 2.1. So what's going to be made available in terms of training materials in the code? Available to everyone is going to be the video recording from this meeting today off the video recording and presentation materials from the April 23rd meeting. Some of the XLPR development related reports. We had some requests for information about the code. There's plenty documented in those. We're going to make those publicly available very soon. Also available to XLPR users will be the user manual as well as a couple of different training manuals. Here's a snapshot of one of these training manuals called the theory training manual. It has several different lessons here. Many different lessons actually covering the different modules in the code. It's basically a two and a half day version of what we're going to be giving you today. So there's plenty of additional details in there as well on the theory behind the code. You'll have some links to some pre-recorded session of that course also available again on YouTube. This, as I mentioned, is part of a series of webinars surrounding XLPR code release. Focus today is a high level overview of the XLPR and its underlying deterministic models. We're going to have several other seminars which I'll talk about a little bit more at the end of the thing on setting up the inputs, surrounding the simulations and accessing the results. Those are tentatively scheduled for July. What we're going to do with those is set those up when we start actually getting licensee users of the code. So those are going to be user-focused training sessions. Whereas the meeting today and the one on the 23rd is more providing general information about the code. Greg, would you like to add anything? Thank you, Matt. Just a couple of quick comments. I really want to thank everyone for your participation today. We do hope that you will find this information useful. XLPR is a complex tool. It will require investment of your time to really learn it and be able to apply it effectively. But we do expect that users will come to appreciate the flexibility that it offers for analyses that fall within its capabilities. I apologize for the slight delay here in the release of the code. We've had a lot of moving parts to get to this point. But then in addition at the end, this product is an unusual product to be released through EPRI.com. And that's created its own hurdles and challenges that we've been working through. So we're almost there. It actually sort of is out there on EPRI.com. It's just not everybody can get to it. It's not readily apparent yet. We're moving through some final last-minute technical details. So bear with us for another day or so. And I expect that we'll have right up. As Matt said, we will make sure that an announcement goes out, especially to this group that is participating on the call today. So with that, let me get out of the way and turn this back to Matt and our presenter. Thanks, Craig. So look for an announcement soon on XLPR version 2.1 release. We've had folks attending this a meeting. We have your email addresses, so you'll receive an announcement any day now. Okay. Now onto the meet of our meeting today. I'd like to introduce Marcus Burkhart. He'll be doing most of the presenting today. Marcus is a senior engineer at Demingian Engineering Incorporated. He's supporting the XLPR efforts under contract EPRI. Marcus has supported development of a number of the XLPR models and inputs. He's one of the more experienced XLPR users. And then backing up, we have a team of other people. And that's Greg and myself as the XLPR program managers. We'll be helping out with the Q&A. Part of that team is Dr. Cedric Selberry. He's a mathematician and engineering mechanics corporation. Contract to NRC. We also have Nathan Glint from the EPRI staff. And Marge Erikson from Phoenix Engineering Associates. She's under contract to EPRI. Marge was the lead of the models group, XLPR version 2 development. So she brings a lot of good information. And then we also have Giovanni Facco from the NRC staff. He's helping out with the WebEx today. If you're having any difficulties with your WebEx setup, feel free to reach out to him through the chat feature. Thanks again for everyone for being here today. Thank you so much for joining us. I'm going to go ahead and move forward to the meeting. Just a reminder to use the Q&A feature today. Okay. With that, Marcus, why don't you take it away? All right. Thank you, Matt. So now I will jump into the technical portion of today's presentation, starting out with an overview of XLPR. So there are several parts that come together to make the inputs need to be provided using either an XL input file or using SIM editor, which can be used to generate the inputs that go into the XL input file. The preprocessor is then run on the inputs in the input set to generate lookup tables for two preprocessors, Leopold and Tiffany. The Goldsum framework is then used to run the actual XLPR simulation. The framework manages all of the probabilistic aspects of XLPR, including data management, sampling, model evaluation, aleatory and epistemic looping, and generating output statistics. It then also provides a user-friendly, graphical, spectacular view of the outputs and of calculated intermediate variables. So as I mentioned on the prior slide, there are two ways to provide inputs to XLPR. The SIM editor top application provides a user-friendly way to generate the input set, whereas more advanced users may want to use the input spreadsheet, which is directly editable. In the XLPR 2.1 release, several input libraries are provided, which contain a sample input values for various volume signal stress profiles or material properties. The two preprocessors that are run from within the input set are Tiffany and Leopold. Tiffany stands for thermal stress intensity factors for any coolant history, and generates lookup tables for stress and stress intensity factor K values based on client transients. These lookup tables are utilized for the fatigue initiates and growth models. Leopold stands for weak analysis of piping and openings. The lookup tables calculated using Leopold provide calculated leak rates as a function of crackling and crack opening displacement. The XLPR framework is the center point of the analysis, handling information flow and linking the deterministic sub-models together. The pulse in itself is a graphical user interface based probabilistic programming software. It takes the dynamic simulation engine and embeds it within a monochrome probabilistic simulation framework. Various random sampling options are available, including simple random sampling, Latin hypercube sampling, and important sampling, which can be applied to specific portions of distributions for key inputs of interest. The framework also handles sampling of variables in one or two loops, either the epistemic loop, which is the outer loop, or the aleatory loop, which is the inner loop. The user can specify in which loop each variable is sampled. GoldSim also includes several tools for executing realizations in parallel, helping improve runtime. The framework tracks variables for all, or tracks values for all variables that are then passed between the modules. So in GoldSim, you have the option to save values for each of those variables. That is very useful for understanding the results that you're receiving from XFALT-1, as well as for debugging the sort of problem that you find in the model. And finally, through the framework, you're able to access the simulation results, and you can view those graphically or in a tabular form, and either in a statistics, or on a realization by realization. So what does a component model with an XFALT-1? The model is focused around a well connecting two components. In XFALT-R, those are similarly simply called left-pipe and right-pipe. And then for application of XFALT-R, it's a week before break, which were the primary locations of interest during development. That's kind of what the problem is focused around. So each of these parts or materials, left-pipe, right-pipe, and well, are considered to be distinct and separate inputs to be provided to XFALT-R, each of these materials. Not shown in the figure, but also included in XFALT-R is a mitigation material which can be used for modeling components that have inlays and overlays. The circumference of the component in XFALT-R is broken up into a number of distributes of units. For crack initiation, all flaws are models to initiate within the welds with up to one axial and one circumference of the wall within a given subunit. All axial cracks are modeled to exist in different planes and do not coalesce with each other. Once axial cracks reach the boundary of the welds, they can either be modeled to rest or to continue propagating into the base metal using the appropriate crack growth rate equations for that material. And so selecting which option occurs there is an input that's provided to the user. Circumferential cracks are all models to exist within the axial center line of the welds. As they're in the same plane, they may coalesce. So I'll be touching more on the XFALT-R coalescence models later in this presentation. For time evolution in XFALT-R, the user inputs a simulation duration. The time step is set to a default of one month, but that is a setting that can be changed by the user within the plane. There are certain inputs that are applied based on the operating periods and the mitigation status, which are all additional user inputs to XFALT-R. Up to three time-based operating periods can be modeled. For each operating period, the user can define separate inputs as either constant or distributed values for pressure, temperature, load, and stresses. The user can also set a time at which mitigation is applied. So this could either be chemical mitigation in which changes to zinc or hydrogen concentrations are applied, or it could also be physical mitigation, such as mechanical stress improvement, weld-over-life, or inlay-online, in which inputs for geometry, welding residual stress profiles, in-service inspection model parameters, and material properties can be updated. So I mentioned that either chemical or mechanical mitigation both can apply either one or both at the same time. For each mitigation technique, at a time at which the mitigation starts, is input SELT-R. So at a maximum, you could have a different mitigation start time for zinc, a different start time for hydrogen, and a different start time for the mechanical mitigation technique. For zinc mitigation, the zinc concentration is updated, which can have an impact on the calculated PWSEC. For hydrogen mitigation, the dissolved hydrogen concentration is updated, which can impact the PWSEC growth rate. Mechanical mitigations can make more significant changes as well. Mechanical stress improvement results in an update to the welding of these stress profiles, and an update to the probability of detection models. Weld overlay results in changes to the wall thickness, material properties, stress profiles, and probability of detection. Any present through-all cracks are also modeled to become surface cracks with steps equal to the original pipe wall thickness. So here are some general references that are available to the XLPR users that are useful for finding more general information. The XLPR user manual and the XLPR training manual through the code theory will be included within the XLPR release. And then the XLPR user manual and the XLPR training manual through the code theory will be included within the XLPR release. And then the computational framework report and the inputs group report will be made available separately at a later time. So now we'll jump into more detail on the deterministic models built into XLPR. So XLPR models both axial and circumferential cracks. And all cracks are assumed to initiate on the component inner diameter. Part through wall cracks is shown on the left are modeled as semi-elliptical surface flaws. After growing through wall cracks are then treated as transitioning through wall flaws or trapezoidal flaws which you can see in the center some of the figures. Unlike shown in the figure here they're treated as trapezoids rather than oxoids. And then after some period of through wall growth the transition through wall flaws eventually become idealized through wall flaws. With equal inner and outer diameter lengths for the case of axial cracks or equal inner and outer diameter angles for the case of circumferential cracks. XLPR was built to be modular based. So there are numerous sub-models built into XLPR. With the intent of if there are any future developments to these models they can be swapped out for updated models relatively easily. XLPR is focused on modeling the development of cracks within components such as the similar model piping bubbles. The sub-models that were developed for XLPR helped to use it. So first cracks are modeled to initiate. Once the cracks are initiated crack tip stress intensity factors are calculated which then feed into crack growth. For transients Tiffany calculates the corresponding delta K values which results in fatigue crack growth. As cracks grow they have the potential to coalesce if they are multiple circumferential cracks that are sufficiently close to each other. For transition through wall cracks the crack transition model calculates adjustments to the K and crack opening displacement solutions. Crack opening displacement is then calculated and fed into the week rate calculations performed by Leopold. Stability is evaluated to determine the cracks and results in rupture. And finally an in-service inspection model is included to model inspections and results in repairs. Each underlying model is implemented deterministically within the probabilistic aspects coming from the framework and the values and examples. All models were verified through a intuitive test and validated through comparison against laboratory or field data. So for the loads XLTR allows the user to input both normal operating loads and transient loads. For the normal operating loads included R crossword dead weight and thermal loads these can be defined for up to three operating periods. Group and axial welding residual stresses can also be input with different stress profiles applied before and after medication. Furthermore there is an input that can be used to adjust the value of the stresses at the inner diameter surface in both the hoop and axial directions. Three main types of transients are included in XLTR. Type 1 transients are temperature pressure time histories and are used to model transients such as heat up and cool down or other normal or upset thermal transients with temperature and pressure in the sun. Type 2 transients represent thermal stratification treatments. Stresses from this type of transients represent global bending and membrane stresses associated with stratified flow during a transient. Type 2 transients are always associated with type 1 transients. Type 3 transients are meant to model mechanical transients including membrane and global bending stress inputs. For example you can model an operating basis earthquake using a type 3 transient. Finally there are earthquake loads which model safe shutdown earthquakes and those are used to contribute to stability but do not contribute to crack initiation. So for transients up to 20 of them can be input into XLTR. For each transient the user can define the interval over which the transient may occur based on a start time and an end time along with a frequency with which the transient occurs. A front-back loading ratio specifies when within each interval the transient should occur either if it's loaded or if a smaller front-back loading value would have the transient occur earlier within that window whereas a greater value for that input would have the transient later. So for example a value of 0.5 would mean the transient occurs in the middle of the interval. And finally the number of cycles specifies how many load cycles are modeled for each transient when it occurs. The users can also input their own welding residual stress profiles in XLTR. These profiles can be defined for up to 26 points through the thickness of the component. Each point in the welding residual stress profile is input with an associated through-all range. Welding residual stress profiles in XLTR are modeled as axisymmetric so they do not vary around the circumference. XLTR users are able to define weld residual stress profiles both axial weld residual stress which are applied to the circumferential cracks and hoop weld residual stress which is applied to axial cracks. And then different profiles for hoop and axial weld residual stress can be applied both before and after medical medication. The WRS inputs can either be constant or normally distributed. For normally distributed input profiles the user can select a point-to-point correlation coefficient to help mitigate the sawtooth effects that would otherwise occur when sampling WRS at each point through the thickness of the component independently. So included in the XLTR release is a library that contains several examples of WRS profiles. So to develop these profiles as part of the XLTR development efforts for independent weld residual stress modeling experts we're all tasked with modeling the same components following the same set of general assumptions. Some modeling details such as their own codes, meshing or repair modeling details were left up to the modelers. The results were then combined to develop these example profiles. With modeled uncertainties for the distributed weld residual stress inputs derived from a variation between the profiles developed from the individual modelers. The provider with XLTR are profiles for a Westinghouse reactor pressure vessel in the hustle. A Westinghouse steam generator nozzle welds and a B&W reactor coolant pump nozzle weld. For each weld either a no repair a 15% through weld repair or a 50% through weld repair profile is provided. Where the repair depth is assumed to be constant all the way around the circumference. Then literature solutions were then applied to come up with post-mitigation weld residual stress profiles for weld overlays, mechanical stress improvement and inlays. So key references for more information on loading and weld residual stress inputs include the weld residual stress subgroup report as well as the weld residual stress portion of the theory training and also the user manual which has an appendix session focused on the stresses. So one of the elements of the XLTR code is the calculation of stress intensity factors model cracks to evaluate track growth. Rather than applying an influence coefficient based approach which requires a stress profile to be described by a polynomial up to this order the universal weight function method which does not require a polynomial that was implemented in XLTR. This allowed for more accurate case solutions when considering more complex welders of the stress profiles that are not easily defined using polynomials. In the weight function approach the product of the weight function and the stress profile are integrated over the crack depth to chain K. The stress intensity factor calculation models were developed in XLTR for calculating K for axial circumferential seminal of the surface cracks on the inner diameter which are implemented in the KPW module as well as axial circumferential idealized through all cracks which is implemented in the KTW module. The crack transition module which I'll discuss later provides adjustments to the stress intensity factor solutions for the transitioning trapezoidal through all cracks. So by assuming that the stress profile varies as a piecewise linear stress profile as is shown in the bottom right figure closed form solutions could be derived to evaluate the weight function interval that I showed on the prior slide. These closed form solutions are what is implemented in XLTR. For circumferential cracks the contribution on K from the global bending stress is also considered. It is noted that for through all cracks through all average welding residual stresses are considered when determining the stress intensity factor rather than applying the piecewise linear stress profile that is applied for part through all flaws. The figure on the bottom left shows a comparison of A values obtained using finite element models and using the universal weight function method demonstrating the accuracy of the approach. This has been confirmed for highly nonlinear weld residual stress. It is important to note that for surface cracks K is calculated at the surface and at the deepest points. For through all cracks only one K is averaged over the entire crack front. For transitioning through all cracks different adjustment factors are calculated on the inner and outer diameter resulting in two different Ks at the inner and outer distance. Tiffany is a standalone pre-processor in the input set that prepares cyclic stress intensity factor transient rise time and cyclic stress values for transients which are then used in Excel theater. These calculated values are used as inputs to keep crack initiation in both calculations. Changes in coolant temperature reduce stresses in the walls of the model piping. These are called radial gradients thermal stresses. Changes in coolant temperature lead to cyclic stresses which can then contribute to this. Tiffany is able to model the three transient types considered in Excel theater. Type 1 transients which are thermal transients type 2 transients which are the stratification transients and type 3 transients which are mechanical transients. As I mentioned before the type 1 transients are defined using a pressure temperature time history whereas the type 2 and type 3 transients have a membrane and banding substance that are input into Tiffany. So as Tiffany is a pre-processor its outputs are lookup tables for the cyclic stress intensity factors transient rise time and cyclic stress for each transient and they're defined as a function of the crack shape. These lookup tables are then read in and interpolated by the Excel theater framework in runtime. So here are a few references that are useful for more information on the stress intensity factor calculations. There's the model subgroup report for the case solutions as well as the model subgroup report for Tiffany. The theory manual had separate sessions on case solutions as well as on transient modeling and the user manual has appendix sections that further discuss the case solutions and the Tiffany pre-processing module. So in XLPR initiation is defined as the emergence of a flaw of engineering scale on the order of maybe half of a millimeter to several millimeters in depth. The simulation of micro-sized flaw growth and coalescence is not modeled in XLPR. Both PWSEC and fatigue initiation mechanisms are included and can be evaluated individually or both at the same time. However, if PWSEC and fatigue are considered model separately from each other, so the effects are not superimposed for the purposes of initiation and there's no correlation between the models. Additionally, an initial flaw model allows the user to model flaws existing at the beginning of the simulation time. Initiation model parameters were calibrated using a combination of field and laboratory data and so the models are semi-apirical. The approach was selected as initiation is difficult to describe for more mechanistic first principle standpoint. Earlier in the presentation I talked about the different time periods for loads or operating conditions as well as changes to welders or stress profiles associated with mechanical mitigation. So to account for these changes and factors which made influence initiation a minor rule type approach is applied to calculate cumulative damage over time. For each time period damage is calculated with initiation models and exceeds one. A component in XLPR is spatially discretized into end subunits around the circumference. All subunits are modeled to be the equal length. The first subunit is centered at zero radians at the top-dead center of the welds and subsequent subunits are coincident with the edge of the prior subunit. This goes all the way around the circumference with no overlap between the subunits. Each initiation model determines an initiation time with a discrete volume of material. So the initiation models are called for each crack orientation in each subunit at the start of the simulation. As I mentioned, PWCC and fatigue initiation are modeled separately. If both are considered, the initiation time for given subunit orientation is based on whichever of the two initiation mechanisms is modeled to occur first. When crack initiation is modeled to occur within a subunit the circumference location of the crack within the subunit is determined by uniformly sampling location within the bounds of the subunit. Axial location of a crack is always assumed to be within the center line of the welds. Once a crack initiates, other crack-specific properties such as initial length and initial depth are sampled by the framework. As I mentioned, the XLPR initiation models are semi-empirical containing functional dependencies for conditions that are known to have strong impacts on initiation. Three different PWCC initiation models are available on XLPR. Two direct models and a Weibull model. Each of these models contains failure time model parameters which are calibrated based on field data and effect model parameters which are calibrated based on laboratory data. Direct model 1 is based on the material index model which includes dependencies for surface stress. The temperature dependence is defined using an Arrhenius term and the surface stress dependence is captured using a power line. The model also includes a user-specified surface stress threshold below which PWCC does not occur. Direct model 2 adds effects of cold work on the SEC's effectability by considering mechanical properties defining the material's strength and strain-hardening response. Again, the temperature dependence and surface stress in cold work dependencies are related to the function of the material yield strength, ultimate strength, and elastic modulus. Direct model 2 also includes calculated stress thresholds which below which PWCC does not initiate or above which PWCC is modeled to initiate immediately. The Weibull model uses a Weibull failure time model an Arrhenius term for the temperature dependence and a power law term for the stress dependence. For the Weibull model, the stress threshold for initiation is fixed with a stress of zero. So any positive stress will result in initiation. All three PWCC initiation models include a zinc effect model which applies a factor of improvement when the zinc concentration is above a given threshold. Fatigue initiation models are also developed for the carbon and low-out heels, austenitic stainless heels, these models are trained and considered environmental effects such as temperature, sulfur content, dissolved oxygen content, and strain rate. For fatigue, it is noted that fatigue initiation models are focused on low cycle fatigue and do not consider damage accumulation due to high sensitivity. Some key references for more information on crack initiation models include the crack initiation model subgroup report as well as PWCC and fatigue crack initiation sections of the theory training and also a user manual appendix section on crack initiation. So in modeling the growth of cracks, XLPR considers the cracks to be one of three ideal spliced flush shapes. Either semi-elliptical surface cracks as shown on the left, transitioning cracks as shown in the center, or idealized through all cracks as shown on the right. The semi-elliptical surface cracks are models to grow in both length and depth directions. Transitioning cracks have different crack growth rates evaluating the ID and OD plots and idealized through all cracks assuming the same crack growth rate on both the ID and OD. Some factors which may impact crack growth are samples on a subunit by subunit basis. Adjustments are made in the framework to ensure cracks remain symmetric consistent with one of the three model idealized flush shapes. This is done by evaluating growth at one general surface tip and then using an average of the conditions at the two surface tips to evaluate the crack growth. The crack center location is then shifted to account for any asymmetric conditions at the crack surface. PWSEC and fatigue growth are assumed to be independent. So both mechanisms are modeled in XLPR at the same time. Calculated PWSEC and fatigue crack growth rates are added to obtain a total growth rate so in XLPR you have the option of modeling either only PWSEC, only fatigue or both mechanisms at the same time. Crack growth rate models include functional dependencies on predominant drivers of the growth such as stress intensity factors, temperature, water chemistry, material condition, or loading characteristics. Certain dependencies are not explicitly captured and are instead considering certain be associated with each model. These include effects on orientation with respect to the material and manufacturing, cold work or residual plastic deformation or differences in crack growth rates in your interfaces such as the heat affected zones or weld dilution zones. The PWSEC growth model is based on the models developed in MRP 55, MRP 115, and MRP 263 for alloy 600 and alloys 82, 182, and 132. These models are semi empirical as well and include theoretical dependencies with laboratory data used to develop the model parameters. Included dependencies in the crack growth rate model are a power law dependence on stress intensity factor with optional stress intensity factor threshold and Arrhenius temperature dependence as well as factors for the dissolved hydrogen concentration component to component such as heat to heat and within component such as within heat or within weld variability as well as a factor of improvement. PWSEC growth and other materials can be modeled using the custom growth model which implements the same model form as shown above but with the ability for the user to specify input values for each of the parameters. So for PWSEC the user has the option to include a correlation between the sampled initiation and growth model parameters to investigate the potential assumption that materials with earlier initiation times also have higher crack growth rates. Materials specific fatigue crack growth rate models are included for nickel based alloys, automatic stainless steels, and ferric steels. Like the PWSEC growth models, the fatigue models are semi empirical with parameters developed based on laboratory testing. The bases for the included model forms are Nuregg 6721 for mixed light based alloys, Codecase N809 for austenite stainless steels and Codecase N643-2 for ferric steels. All three models are based on Paris's law which is the ADN equals C delta K to the N. The nickel based alloy model then for fatigue in air also uses additional alloy type temperature and load ratio depending on their dependencies. The environmental fatigue model adds a rise time dependency to that. The austenite stainless steel crack growth rate model uses Paris law with additional dependencies for stainless steel class, temperature, load ratio, and rise time. The general ferric steel crack growth rate model uses Paris law with additional dependencies for load ratio, rise time, and sulfur content of the material. The model has different regimes determined by sulfur content and rise time to reflect different degrees of environmentally accelerated corrosion. The transient schedule over the component lifetime determines one fatigue crack growth is evaluated and is handled by the framework. So the coalescence model is a set of rule based conventions used for simulating the combination of two cracks based on their sizes, their shapes, and their locations. This simulates cases where two cracks close to each other may coalesce on the time scale faster than would be predicted by treating the growth of each of the two cracks individually either due to crack interaction or surface cracks or due to local ligament collapse or through all cracks. Coalescence is implemented as a rule based model with specific coalescence rules varying for different pair wise combinations of crack types which I'll go into in a little bit. In all cases, coalescence is modeled between two cracks at a time. The potential for three or more cracks to coalesce is extremely rare and in these situations XLPR would model two or more sequential instances of pair wise coalescence. And the user can investigate implications of this assumption by changing a coalescence direction input in XLPR from clockwise to counterclockwise to counterclockwise coalescence. All circumferential cracks are assumed to be coplanar and have the potential to coalesce with other circumferential cracks if the coalescence proximity rules are met which define how close the cracks have to be together in order for them to coalesce. For XLPR, axial cracks initiate in several circumferential subunits and are not modeled to interact. Therefore, circumferential cracks are also not assumed to interact with axial cracks. So this slide and the next slide will provide a summary of the coalescence rules showing the resulting flaw shape of the coalesced crack giving the two input flaw shapes. Additional rules exist for defining the coalescence distance which is the distance between two cracks before the coalesce as well as the resulting geometry of the coalesced crack including the inner half length, the outer half length, and if applicable the depth. So here you see that two surface cracks coalesce to form another surface crack, two transitioning cracks coalesce to form another transitioning crack, and two idealized through all cracks coalesce to form an idealized through all cracks. Then when you have two different crack types that coalesce, they form a transitioning through all cracks. So a surface crack and idealized through all cracks from a transitioning through all cracks. A surface crack and a transitioning crack form another transitioning crack. And a through all crack and a transitioning crack also form a transition. So key references for the crack growth and coalescence models include the crack growth and coalescence model subgroup report sections in the theory manual on PWSCC growth fatigue initiation growth and crack coalescence project sections on the crack growth model and crack coalescence model. So talked about transitioning cracks a little bit. When a surface crack grows through all, you need to consider what the initial through all crack size should be and determine what the appropriate crack length should be on the ID node. They're going directly from an idealized through all crack sorry, from an idealized surface crack to an idealized through all crack with an equivalent crack area at the time of leakage what results in inaccurately great predictions. So this motivated the implementation of the transitioning through all crack type and thus the crack transition model. The crack transition model calculates adjustments for the stress intensity factor and crack opening displacement solutions for both axial and circumferential transitioning through all cracks. So these adjustments are made by applying factors to the corresponding K and COD solutions for idealized through all cracks with the same ID following. Then the applied correction factors are developed based on finite element models. So again contrary to what's shown in the figures here, the transitioning through all cracks are modeled more trapezoids and not as ellipsoids. The framework determines the geometry of the initial transitioning through all crack as well as the transition through as well as when the crack transitions to be an idealized through all crack. So surface cracks are modeled to transition to a transitioning through all crack when the depth exceeds 95% through all. The initial inner outer half length is then defined as the minimum of the component thickness and the inner half length divided by four. The transitioning crack then grows using the K's calculated with the help of the stress intensity factor calculation module and also the crack transition module until the ratio of the inner half length to the outer half length becomes less than or equal to 1.05. At that point the crack then transforms to being an idealized through all crack and continues evolving in time as an idealized through all crack. The more information on the crack transition model subgroup report also in the crack transition section of the theory manual and the user manual appendix section of the crack transition. Crack opening this placement or COD is an input to the weak rate calculations. Models for estimating the COD for both circumferential and axial idealized through all cracks have been implemented in the XLPR. Both models are based on the TEFRE methodology for predicting crack opening dismissals where the elastic and plastic influence functions are fit to finite element results. The total COD is the sum of the elastic and plastic contribution. Now the XLPR models are an extension of the TEFRE solutions that they consider combined tension and bending load cases as well as extend the ranges over which the elastic and plastic influence functions are applicable. So that's a greater range of pipe mean radius thickness ratios and normalized crack lengths that are included. The COD models are implemented as lookup tables as a function of component geometry, crack length and strain hardening exponents. The COD models are then output COD values at the pipe od id and halfway through the pipe thickness. Circumferential COD includes contributions from tension bending and crack-face pressure loading contribution. Axial COD includes contributions from pipe internal pressure, weld residual stress, and crack-face pressure with the weld residual stress being applied through an effective pressure. So as I discussed previously the crack transition models calculate correction factors for the COD in the case of transitioning through all cracks. So the couple of figures that are below here provide an example of what a large or a small COD would look like. So the leak rate model determines the leak rate of fluid flow through a crack for a given crack size, pressure, temperature, and cracking mechanism. These calculations are performed by solving equations developed by Henry and Paus to represent fluid flow through a long pipe in which steam generation occurs resulting in two-phase flow. In order to more accurately adapt these equations to fluid flow through a crack, modifications to the equations are made to account for pressure losses due to entrance effects, phase acceleration, friction, flow path deviations, flow changes, and crack morphology effects. Surface roughness, the number of turns, and flow path length are all key crack morphology parameters. Leopold is available as standalone software and also as an XLPR preprocessor. The preprocessor is used to generate leak rate tables as a function of crack length and crack opening displacement for a set of bounding pressure and temperature combinations and for both PWSC and fatigue cracking mechanism. The XLPR user inputs crack morphology parameters for SSC fatigue as well as a minimum and maximum pressure and temperature limits to which the leak rate tables are generated. During runtime the framework then interblades these tables for a given pressure, temperature, thickness, inner crack length, and COD to determine the leak rate for a given crack during a given time step. XLPR also allows the user the option to apply uncertainty on leak rates that are low such as less than 4GPM. Leopold is based on the same thermohydraulic model of squirt but includes several improvements including modifications to account for crack morphology. The four fluid flow phases are implemented in Leopold to produce a smooth transition in the calculated flow rate between the single phase orifice flow for sub-cooled liquids and the Henry Fausk model for two phase choke flow and a tight crack regime. These flow regimes are all a function of the flow path length to the hydraulic diameter ratio. In addition to the two phase flow regime which is regime 1 and the orifice flow regime which is regime 4, two intermediate regimes are included. Regime 2 is a bridge regime for ratios of the flow path length to hydraulic diameter from 12 to 30 in which the tight crack model for regime 1 is used. Regime 3 is a transition regime for flow path length to hydraulic diameter ratios from 4.6 to 12 in which linear interpolation between the single phase orifice flow modeled in regime 4 and two phase choke flow modeled in regime 1 is applied. The leak rates are calculated assuming and idealized through all crack shape meaning that the crack opening areas on the ID are the same. But then adjustments are made in the framework to account for this assumption. By default, leak rate is calculated based on the inner diameter crack size. However, if the flow is in regime 1 the leak rates are then based on the ID and OD COD and then averaged. So key references for more information include the COD and leak rate model subgroup reports as well as the theory training sections on circumferential COD, axial COD and leak rate calculations. User manual also includes appendix sections on the COD and on the leak rate models. So XLPR includes several different stability models for circumferential cracks and axial cracks. The circumferential stability models are implemented in SCFail and TWSCFail for surface crack and through all cracks respectively. The axial crack models are axial SCFail and axial TWSCFail again for surface cracks and through all cracks respectively. When the XLPR stability models are called they generally output whether or not rupture occurs with surface cracks. The ratio of the current applied loads to the critical loads and then for through all cracks the ratio of the current crack size rupture due to seismic conditions is also considered an XLPR. If stability limits are exceeded due to seismic conditions then rupture is reported but the simulation for that realization continues. However, if rupture occurs due to normal operating plus transient loads then the rupture is reported and the time evolution for that realization. So for circumferential cracks if one or more surface crack exists a multiple net section collapse model is used to evaluate stability. This model is applicable to one or more circumferential cracks and evaluates stability under combined tension and bending loading. Surface cracks are models of constant depth and if any through all cracks are present they are also considered in the multiple net section collapse model of deep surface cracks with a depth of 99% of wall thickness. No EPFM model is implemented for surface cracks. Now if any through all cracks exist they are then also individually evaluated for stability using both net section collapse and EPFM models also subject to combined tension and bending loading. After evaluating stability using both net section collapse and EPFM models the solution that yields the smallest critical crack size is then used for the output from the crack stability module. It is noted that for crack stability all cracks are modeled as ideal is through all cracks. Stability of transitioning through all cracks is modeled using an equivalent ideal as through all crack with a crack half length equal to the average of the ID OD half length of the transitioning crack. Axial surface cracks have stability evaluated using a plastic collapse analysis with the cracks modeled as constant depth surface cracks. Axial through all cracks have stability evaluated under both limit load and EPFM models. Again the smallest critical crack size is then given the output from the axial crack stability module. Axial crack stability is evaluated on a per crack basis meaning that no interaction is considered between multiple axial cracks. Neither of the axial crack stability models include effects of wall residual stresses. It is assumed that axial cracks involves can be analyzed using the current axial crack stability models even though the models were developed for cracks in homogenous materials. It is unlikely that axial cracks will be limiting case for stability in dissimilar metal piping foot wells. These assumptions were considered to be appropriate. However it can still be considered when interpreting and applying the results. So for more information on the crack stability models we refer to the crack stability model subject report, the theory manual which has sections on surface crack stability and axial crack stability as well as the user manual panic section on crack stability. In-service inspections are also modeled in XLPL and contain two key parts an inspection model and an evaluation model. The inspection model is actually a detection curve as a function of crack depth to determine if a crack is detected during a given inspection. The evaluation model has been used to size the detected crack. Determining if that crack has a size greater than the threshold size will appear. Both models utilize crack depth as an independent variable. Other attributes of the cracks such as the length of the crack opening displacement may also have an impact on the sizing. However, these dependencies are part of the model uncertainty rather than being explicitly included. ISI model parameters recommended for use in XLPR in the model subgroup report were based on the APRI performance demonstration initiative or PDI program. And details of that are provided in MRP262. In XLPR the user has the ability to schedule in-service inspections using either a frequency or by inputting the specific time of the inspection. Different inspection inputs including inspection schedule TOD models and sizing models can be provided to XLPR before and after the application. The inspection model is modeled using a probability of detection curve describing probability of detection as a function of lava. XLPR represents TOD using logistic equation with the number of conventions built in to improve model flexibility to address small flaws and variations in performance between mock-up and field examination. The coefficients in the exponential term of the logistic equation are the main TOD model parameters. The logistic equation model form has a convenient property of varying between 0 and 1 with model parameters that can be varied and shipped to the overall curve. So this has been the model of choice for modeling probability of detection as a function of crack size when it's based on binary data such as either detection or no detection. Because the ASME code requirements do not require flaws less than 10% through all for vendor qualification the current PDI data do not quantify POD performance for small flaws. So XLPR includes options for treatment of POD for small flaws which can either be modeled by linear interpolating a POD between a lower bound POD and a small flaw depth threshold such as 10% through all or setting the POD in that depth range to 0. The figure on the right shows the linearly interpolated option. So a POD effectiveness factor input can also be used to scale down the nominal POD curve as a way to model differences between laboratory and field performance. And the figure on the right that effectiveness factor is shown by the dashed line. So the nominal form of the sizing model is a linear relationship between the measured flaw size and the actual flaw size and it has normally distributed uncertainty. Effectively the sizing model determines how much the estimated size of the flaw is either under or over predicted. The depth sizing error curve is also included and added to the nominal measured flaw size. The sizing model is combined with the repair threshold input to calculate a probability of repair as a function of flaw depth. Using a repair depth threshold of 0 would result in all detected flaws being modeled as being repaired. So the XLPR framework considers all modeled repairs as being perfect. So this means that the probability of additional initiation leakage or rupture due to a crack in that subunit is 0 for the remainder of the simulation for that specific realization. So key references for the in-service inspection models again include the in-service inspection model sub-unit report, the theory manual which contains two sessions on in-service inspection one for model parameter development and one for model implementation. And finally the user manual also has an appendix section on in-service inspection. So with that I'll pass it back to Matt. Thank you very much, Marcus. We have concluded the technical discussion here in the presentation. We've been keeping up with the questions and answers that have been coming in through the Q&A feature. I'd like to open it up now to just general questions and answers. We could take those still through the Q&A feature. Also if you'd like to ask a question verbally go ahead and raise your hand please. You can do that in WebEx a couple different ways again. You do this and we'll see you and we'll unmute you so you can ask a question. The steps that raise your hand are different again. If you're using the internet browser versus using the desktop client and instructions are shown here on the left there's the browser version. There you'll first click on the three dots at the bottom of your screen and follow that click on the raise hand button. Here in the desktop client you see on the right here the bottom of the screen you'll see a little hand symbol. Click on that and that will raise your hand as well. Again, general question and answers please feel free to ask them however you prefer. Here's a question from Peter asking, while residual stress uncertainty is perhaps better estimated by using uniform distribution. Is this possible to use as an input? I believe the answer to that one is no. Currently welding residual stresses you can either model as only constant or as uniformly distributed normally distributed sorry not uniformly distributed. Most other inputs you have a little bit more flexibility on the distributions that are applied but due to the added complications that come with the point-to-point correlation only normal distribution is available. Thank you Peter for your question. We have slide 34 let's go there real quick what is the meaning of damage? Does this mean leak or rupture? So this is in terms of tracking initiation and so it's basically a way of accounting for how close you get to the initiation time within a time step under a certain set of operating conditions. You can then use that to combine multiple time periods with multiple operating conditions resulting in a way to determine when cracks are modeled to initiate so damages in terms of initiation are not in terms of leak rupture in the context of slide 34. Joel also has a question regarding an earlier part of the presentation if welds are limited by code and type so ASME or AWS or is there any allowance for novel methods such as internal friction weld? So I don't think XLPR quite gets into those sorts of welding details on how the weld is actually manufactured. You just have the ability to model certain material properties associated with each weld which are mechanical material properties as well as material properties associated with fatigue as well as as well as fatigue and SCC growth. So Ben has a question regarding axial cracks and how they behave if they reach the boundary of the material so if they blow the weld into the base metal and if there is a step change in the crack workflow. So yes in that case you would have a step change in the crack workflow so the crack growth law that is applied is associated with where the location of each crack is so if an axial crack were to grow into the base metal then you have the option of continuing to model crack growth in the length direction using the crack growth models associated with. So Ben when a crack is initiated by fatigue can the crack propagate by PWSCC the answer is yes if you are modeling both PWSCC and fatigue mechanisms if fatigue happens to be the mechanism that results in initiation before PWSCC then yes the crack can also propagate by PWSCC. The third one now is the one from Youngsoon Park with the stability module. I think I don't know if I fully understand the question but I think it's asking for the two wall cracks stability model how do you model that for a transitioning crack with different ID and OD wall size and there the answer is you model the stability of an equivalent idealized to wall crack that has a half length that is equal to the average of the inner and outer half length and so you average those half lengths to get an equivalent idealized to wall crack and then model through wall crack stability for that equivalent crack. Hopefully that answered the question if not please submit another one and we will try to get better clarification for you. There is a question from Young what is the difference between thermal transient and stratification transient. So the thermal transient is basically the portion of a transient where you're modeling changes in pressure and temperature and how those changes in pressure and temperature within the pipe result in changes in stressors. Whereas the stratification transient is used to model situations where some sort of changes in temperature then results in changes in membrane and bending stressors within the pipe for that particular transient. So I guess as hopefully everyone is able to see over the course of this presentation XLPR gives the users lots of flexibility of what they want to model and in certain cases you have multiple models that you can use and we can switch between those models to see how the answer is. So it gives a lot of flexibility to the users. There's also a lot of complexity to all of the models as well which is why we're holding these seminars to try to give a better overview and also with emphasis there's a lot of material out there that you can also use to find more information. So this is just kind of a relatively high level overview with the theory training providing the two and a half day version and then the user manual and the various model subgroup reports then providing an even more depth version of that. We had a follow on question here from Yan about the transients heat up and cool down is similar to the stratification so in XLPR with transients you can model stratification with normal heat up and cool down type transients or you can just model the heat up and cool down aspect of the transient. So you get one or both I'm sorry you get either the thermal transients with stratification or without you can't ever have just with stratification. The third option of course is just a mechanical transient not associated with the heat up or cool down. Hope that clears things up a bit there. So Yan has another follow-up question if we can look at thermal fatigue the kind of mechanical weather. I think the way XLPR models transients and it's kind of looking more at low cycle fatigue and strain based models and so I think some modifications would be needed to be made to the models to properly model thermal fatigue. So when they're irradiation assisted in stress corrosion cracking can be modeled the answers no not currently and the response here that was IG SEC IA SEC probably a typo there. Yeah however if the model form matches the model form that it showed in the slide there. The user has the ability to input their own model parameters for each of those valleys so as long as you have some form that's a function of stress intensity factor to a power and has an erroneous term in it you can use your own model inputs to model other types of TSE. The question regarding weld metal and so in slide 18 indicates that normally we use butter metal on the base metal and then use weld metal and asking when inputting material properties which material properties to input for the weld metal or base metal and so in XLPR you're able to input material properties for the weld and then also for the base metal for both the left pipe and the right pipe so I would for the weld to be inputting weld metal properties and not butter metal or heat affected zone or dilution zone type material properties and yes internal pressure is one of the inputs to XLPR. Just to add on that you can have the pressure change throughout the course of the simulation up to three different times you can use different inputs for different plan operating periods so you can have for example a constant in operating period one you might want to have an operating period two and you can have a different distribution than an operating period but a flexibility there and that's true for temperature and the loads as well. So Johnson has another question regarding heat open cool down transients how the transients interact with the internal pressure so those transients are defined using a pressure, temperature and time history where the input values are actually the change in pressure and change in temperature from the normal operating pressure and temperature conditions so you can basically have a table that you can fill out with points in time and then in addition to the points in time you have pressure values and temperature values and by filling out that table that's how the transient is defined so all in all there are probably several hundred inputs in XLTR and this is to in order to give the user flexibility and also control for modeling all of these different complex models and so the next session that we will be focused on the input specifically and showing the user how to set up the inputs for a given run that will help provide more information there. I think we're caught up on the questions does anyone else have a question or needs clarification you can raise your hand and we can open the mic for you and we can have a discussion on it. So Joel was asking model allows for load changes during crack development changes from tensile loads to compressive loads the answer is yes it does for example in the different time periods you can have different loads associated with each which each of the three time periods that are considered and so say the first one could be tensile which results in crack initiation and then the second can be compressive in which case you wouldn't have any crack growths due to SEC during that time period and then the third you could then maybe have something that causes it to be tensile again so you continue I guess if you wanted to look at a certain aspect of that too you could also put in your own test profile to influence how the crack loads you know you can make them compressive or tensile that way I appreciate everyone's questions here today I know we've thrown a lot of technical details at you in this meeting but again we're just trying to make sure that we are explaining what kind of features the code has and a lot of that has to do with the models that are contained within it I think that's a question regarding the fatigue crack growth calculation if incremental damages are extrapolated by multiplying a certain amount of cycles or fatigue incremental damages updates cycle so for fatigue crack growth during each time period each one month time period if the transient is modeled within that specific time period then all growth due to that transient is modeled and basically the crack size is then increased accordingly during that same time step so Peter if you're a question on while residual stress is not including the leak rate calculation then the largest non-conservative estimate will be obtained just after this little crack has been created using the non-idealized geometry of the crack and you're wondering what our comments are on that and yeah you're right about that and that's just that's uncertainty that is included in the leak rate calculations and that's part of why in the framework these are the options of applying additional uncertainty on leak rates that are relatively well to try to capture some of that uncertainty. I also wanted to just add on to that is the NRC Office of Research is sponsoring a study right now on WRS effects on crack opening displacement which of course affects the leak rates so we'll be looking for the results from that to form any enhancements so PR. Also back on the WRS on the leak effects we do consider WRS in the axial crack opening displacement calculations. So John's saying for the next presentation Q&A works better than the raise and option I think we'll probably stick with that. Thank you for pointing that out John. Any final questions on the presentations today? Give it a moment here to see if any others trickle in. It looks like it to me. Marcus had a number of references here to all of the technical materials that are going to come with the code be that in the two and a half day theory training session or the user manual itself or as part of the technical reports that we're going to make available. Those get into much more detail on all these aspects and you're welcome to consult those to help answer some of your questions. Of course we'll have the follow on webinars too where you can bring some of those issues up as well. So move on to the closing remarks portion of the agenda now. Today's meeting we provided an overview of the models that were programmed in XLPR and today's meeting was a part of our webinar series surrounding the XLPR public release. Looking ahead we have some additional sessions planned. These will be every one to two weeks and while the session on the April 23rd and today was targeted for more of a general audience these next three events are going to be targeted for actual users who have access to the XLPR public release. So these dates are tentative right now. What we're going to do is wait and see. We have enough user community built up where it would make sense to have one of these sessions and that's when we're going to have schedule the first one. Here they're every two weeks and we've been talking about maybe just doing them every week so it might just be a series of three weeks sessions themselves will be the same. We will announce a firm schedule for those as soon as we think we have enough users to support them. So look for our announcements on those. Again these will be to demonstrate some of the key features of the code itself and provide hints and tips and again we'll devote a good portion of those sessions to interact with the users and answer any questions that they may have. The first one is going to be on setting up the inputs. So we'll talk there about setting up some of the simulation options and defining the distributions and working with the input spreadsheet and the sim editor. To prepare for that we would recommend but it's not required that you take a look at the ExcelPairInputSet file that's the spreadsheet in your release package that's file xlpr-2.1 inputset.xlsx is also module 3 of the xlpr-trn introduction that's another one of the training manuals that will be included with xlpr version 2.1 that's a hands-on training session a two and a half day training. For the video you can watch section 3 or module 3 rather in that focus on the inputs. And then I also recommend that you look at the user manual there's probably a variety of things there you might want to look at but section 3.3 covers the input specifically as well as the pen xb. So that's some reading you may want to do after you get the code and before the first webinar here we have for the users after the inputs we'll talk about running the simulation using gold sim and the settings there and debugging from the errors and things like that and then we'll follow that up with another session on accessing the results from the simulation extract news out of the code and then also setting up some additional outputs if you'd like to do that. So again these are great opportunities to get involved with learning about the code as a user and we highly encourage everyone to take advantage of that. I want to thank everyone for joining here today. Look for again there's an estimate of training webinars as well as for announcement on the actual xlpr code release. Greg did you want to add anything? Okay Greg just to interrupt me if you would like to. So again if you have any questions you can reach out to Greg or I we monitor these email addresses xlprnrc.gov and xlprabry.com Matt I I was speaking into a muted phone I have two mute buttons and they weren't synced. So yes we appreciate everyone's participation today the active interaction with questions and look forward to delivering the code to you very shortly and your participation in future training and use of the code and your feedback. Yeah thanks Greg that's well said. Okay all right that's all we have for today. I just want to remind everyone after the meeting ends if you would like to provide feedback and help us understand your views about this nrc meeting and potentially improve future nrc meetings. You can do that and just go to nrc.gov look for the public meeting schedule web page and click on show recently held meetings and you can find this meeting listed there. Click on that and then there'll be a meeting feedback form you can submit right there. Please do that if you feel like you can help us otherwise that's all we have today we'll send out a meeting summary and also re-watch this video of course on YouTube. Thanks very much everyone and we look forward to seeing you at the next webinar on the xlpr inputs. Thank you we'll adjourn.