 All right. Greetings everyone. My name is Matthew Homiak and I am the NRC's lead for the Extremely Low Probability of Rupture, or XLPR Probabilistic Fraction Mechanics Program. This is a joint venture between the NRC's Office of Nuclear Regulatory Research and the Electric Power Research Institute, or EPRI. Today I'm joined by my counterpart from EPRI, Craig Harrington, NRC and EPRI management, and also some key contributors to the program from the NRC, EPRI, and some of our technical support contractors. On behalf of everyone, I'm very pleased to have you all here today for our XLPR code pre-release event. The purpose of the meeting today is to present to you the XLPR version 2 Probabilistic Fraction Mechanics code and provide information on how you will soon be able to request a copy. This is an NRC Category 3 public meeting. That means the public is invited to participate by providing comments and asking questions throughout. Today we're using the WebEx platform to deliver this meeting. We have demonstrations and other dynamic content, so I would encourage everyone to participate through the WebEx application. You'll get the best experience that way. In WebEx, you can submit your questions and comments at any time in the chat window. If you don't see that on your display, hover over the bottom middle of the main WebEx display screen and click on the comment blue icon. There you should see the option to send message to all panelists, and that should be the default. At certain points, we may also open up the phone line to take questions and comments orally. After the meeting ends, we will issue a summary which will be available on NRC's public website. And any questions that we weren't able to get to today, we will respond to in the summary. And we're also recording the meeting for later viewing. So that covers my administrative items for the meeting today. And now on to a brief overview of our agenda. Here's a copy of our agenda today. We're currently in the introduction and opening remarks portion. Following that, we'll cover the XLPR program history and some perspectives. Then I'll provide an overview of the code itself and some of the features it has. We have an excellent demonstration for you on the code itself and some applications present and planned. Craig will provide an overview of the process for requesting a copy of the code. And we'll also announce some opportunities for future training webinars that we have planned. We'll follow that with the opportunity for general questions and answers. And then closing marks, and we will adjourn. With that, it's my pleasure to introduce Louise Lund and Kurt Edsinger to provide a few opening remarks from NRC in eprimanagement. Louise is the Director of the Division of Engineering and NRC's Office of Nuclear Regulatory Research. And Kurt is the Director Research and Development Materials and Advanced Nuclear at EPRI. With that, Louise, would you care to kick it off? Yes, thank you. Welcome, everyone. This is quite the virtual audience. Some of you may have followed the XLPR journey for years while others only more recently. The fact is that today marks a significant milestone in that journey where the fruits of our labor will be made available to regulators, industry, and researchers across the globe. From the NRC's perspective, XLPR version 2 fills a long-standing need to have a technically vetted probabilistic fracture mechanics code for piping integrity risk assessment. Also, NRC is currently embracing transformation to realize the staff's vision of a nuclear reform regulator. Although XLPR wasn't conceived with that specific vision in mind, it's already well positioned to support both aspects of it. We consider XLPR to be one of our flagship safety codes at the NRC, and today you will hear from NRC staff and the U.S. nuclear industry experts why that is. First and foremost, and I think my counterpart, Kurt, will agree, much of XLPR's value lies in the effective technical working relationships built over years among the NRC's office of nuclear regulatory research staff, EPRI, and their contractors. Some 80 experts were involved in the making of the XLPR version 2, but it wasn't enough to bring the best and brightest to the development of this code. We also built XLPR using modern software design concepts under the framework of a robust quality assurance program. Extensive verification and validation activities were central to this effort. Finally, we built XLPR to be flexible and transparent. As you'll soon see, flexibility is maximized through a modular design, customizable inputs, and the ability to extract a wide range of information for specific applications. As to transparency, the technical basis has been extensively documented, embedded, and many of the algorithms can be readily inspected. XLPR is certainly no black box. NRC and EPRI have developed the code and our intent all along was to share it. I'm pleased to see us at that point today. Our hope is that many of you will go on to become active users and contributors to its further development, thereby maximizing the benefit to all. With that, I would like to give Kurt the opportunity to provide some of EPRI's perspectives. Thanks, Luis, and good morning. The XLPR probabilistic fracture mechanics code has been developed by the U.S. NRC and EPRI, as Luis noted, to quantitatively evaluate piping rupture events. With a primary impetus on resolving leak before break and implications of primary water stress corrosion cracking, but always with an eye to many other possible applications. I agree with Luis. We've produced a tool that both industry and NRC understand and have confidence in, and that enables its usefulness in a range of applications. And the first application to leak before break, LBB, and PWFCC, is now underway. So today marks an important day. A large number of people will now have access to a tool to evaluate active degradation mechanisms and quantify the implications for a range of things that have up to now been assumptions. This has been a lot of years in the making. One of the things I'm most excited about seeing is what questions the new users are going to address. Questions that we may not have even thought of as being answered by this tool. So my sincere thanks to the team for all the hard work that went into this. And I look forward to the seminar today. Maybe Matt, back to you. Yes. Thank you, Curt and Luis, for those opening remarks. Okay. Now on to the business portion of our agenda here. And I'll have Craig lead us off from the next topic here, which is just to provide a brief history of the XLPR program and supporting him. There will be some additional perspectives from Dr. David Rudlin. He is the Senior Technical Advisor for STEAM Generators Integrity and Materials Inspection, the Division of New and Renewed Licenses, and NRC's Office of Nuclear Reactor Regulation. He is also my predecessor as NRC's XLPR program lead. And we also have Dr. Robert Tragoning. He's a Senior Level Advisor for Materials Engineering Issues, a Division of Engineering, NRC Office of Nuclear Regulatory Research. Craig? Well, good morning. Appreciate the opportunity to share a bit of XLPR history and our progress with the XLPR story began in late 2007 in the office, NRC Office of Research with Mr. Bob Hardies and Mr. Al Santos. Al and Bob were both actively involved with and were strong proponents of probabilistic applications within regulatory matters, and particularly with probabilistic fraction mechanics. PFM computational tools tend to be component and problem specific, but Bob and Al recognized that the underlying fraction mechanics was really agnostic to such differences, even as materials, the loads and stresses, the configuration complexities varied. They proposed development of a common flexible PFM platform with a modular structure to efficiently accommodate a wide range of pressure boundary integrity problems and components of various configurations that could be addressed by adding new modules. Jennifer Ewell provided management support to pursue this idea, which formally launched what became XLPR. The initial focus was piping and specifically the leak before break issue was associated with primary water stress corrosion cracking into similar metal welds. To make this ambitious, to tackle this ambitious scope, NRC partnered with EPRI to share resources and funding. But more importantly to combine the depth of technical knowledge that both organizations could bring to bear on the project. A series of planning meeting placed over the course of 2008, but they seemingly identified more questions than answers and the team ultimately decided that undertaking a pilot study was the essential first step. To answer in particular the three most critical questions that we identified of is overall goal realistically attainable, what would be the best computational approach and platform to use, and can NRC and EPRI successfully collaborate in this way. To investigate the computational platform issue, the pilot study developed two competing limited scope codes employing a common set of physical modules to carry the theme of a modular approach. In one of these the integration framework was programmed as a standalone software product while the other was developed within a commercially available simulation modeling environment. Building on successful completion of the pilot study and incorporating a number of lessons learned from that part of the project, we launched development of production version of XLPR in 2007. Consistent of the original idea and the pilot study, key goals were to ensure the final product was well vetted, appropriately represented the diversity of current knowledge in the relevant technical areas, and that it employed a distributed development approach. We wanted in particular to avoid an insular outcome and over reliance on the small set of helpers. To this end as shown on the screen, the project team involved on the order of 80 subject matter experts from NRC, EPRI, three national labs, and seven commercial companies. While not always efficient, this large team and the diversity it reflected was critical to ensure that the endless number of small and large technical decisions involved in this type of a project were thoroughly considered and the decision basis in each case was technically sound. NRC and EPRI provided joint management in sight of the project throughout. We established a virtual organization that was co-funded in roughly equal shares without commingling funds between the organizations, and this is the same way we manage the pilot study as well. NRC and EPRI separately funded and independently contributed staff through this project team, and the leadership roles within each of the different component elements shown on the slide were carefully balanced between NRC and EPRI contributors. We employed a collegial approach for decisions on all technical matters and almost all decisions were made by consensus. While the process was in place to address differences in professional opinion, it was never exercised. We incorporated the intent of an Appendix B QA program for software development from the inception to both ensure a well-documented, high-quality out-town, and to facilitate future use of the code by industry for licensing submittals on Appendix B. We incorporated program audits as well as both internal and external oversight groups to build additional confidence in our developmental process and in the final product. XLPR-V2.1 is the first step toward that goal of a generic probabilistic fracture mechanics tool for evaluating degradation of pressure boundary components. Although presently limited to piping configurations, it has set a very high bar in terms of flexibility and the range of options available to the user to consider almost any variable as given. We have distributed and to tailor its application to a diverse set of problems. However, to accomplish this degree of flexibility, we have also challenged some of the capabilities of the underlying platforms. So as with most new computational tools, we do have early opportunities for continued improvement. One final point is worth mentioning. Many of you have been wondering if we were ever going to release this code. Much of the recent delay has been an unfortunate artifact of our early singular focus on those technical goals, the project, and frankly, a lack of perspective amongst the team for the legal issues involved in co-mingled intellectual property rights in a living software product held between two very different organizations. We are extremely glad at this point to reach this milestone where those issues have been sorted out, and we're excited to now share this tool with the broader technical community. I'd like to shift now to just some brief perspectives on the project. It has been my privilege to manage EPRI's participation in this project and to work closely with my NRC counterpart and this team of subject matter experts. They have no idea how much I have learned from working with them. As noted in the opening remarks by Louise, a long-lasting hidden legacy of this project will be that regulatory and industrial representatives alike gain a greater appreciation and understanding of differing perspectives across the range of technical issues addressed over the course of our work. I'm convinced that such improved understanding benefits all participants and will pay dividends for years to come. PFM analyses and the underlying computational tools are complex. When applied in a regulatory setting, it is incumbent on the regulator to not only assess the analytical outcome, but to also understand how that outcome is obtained. With PFM applications, this can often shift the applicant's emphasis away from the analysis to focus on explaining and defending the computational tool. And this can become an impediment to acceptance of the actual application itself. Somewhat similar to the status of the favor code, the depth of knowledge of the basis and the inner workings of XLPR presently held by NRC holds the promise that greater focus by both the applicants and the regulator can be placed on how the code might be used instead of just trying to understand how the code itself works. With that, I will hand it to Dave Rudlin for perspectives from the NRC staff. Thanks. Thanks, Craig. Yeah, this is Dave Rudlin. You know, I have to say that I share a lot of Craig's perspectives and I thank him for the times that we've spent together over the years trying to get this code completed and released. I've been with the agency now about 12 years and XLPR was one of the first programs that I worked on when I started at the agency. And while the development of the code was not always smooth, it was by far the most collegial, cooperative, and honestly exciting programs that I've worked on since I've been at the agency. Personally, I want to thank both the NRC and every teams for working together and working together in such a professional way and in the development of a very technically challenging piece of software and technology. As we move forward, I really look forward to experiencing how we use this code, not only in research space but also in regulatory space. And I, of course, will look forward to how the code evolves and how our use evolves with that code. So again, thank I just want to personally, I just wanted to thank everybody for all the hard work and I'm glad that we that we made it to this to this point. So I guess I'll turn it over to Rob for a couple of comments. Yeah, thanks Dave. I appreciate the remarks that both you and Craig made and I certainly echo those remarks. But to me, the excitement and everything that we plan for with XLPR is really coming to fruition today and hopefully into the future because, you know, when our C and Epri have both been over our histories, we've developed a lot of code. So the fact that we that we developed a new code, there's nothing unique about that. But I think what is unique is we always develop this particular code with an eye to the future. And some of the unique things that Craig already touched on that I'm hoping will bear fruits for us in the future is the extensive documentation that we did up front. The fact that we relied on a wide broad distributed array of experts instead of just a few subject matter experts and a particular specialty. And again, the fact that we did extensive quality assurance measures and verification and validation efforts up front. So I hope when other users get this code, all this groundwork that we've provided will provide a really good platform and basis for future and ongoing development of this code. Because I think we all know software codes are living organisms and if they don't continue to be used and develop and evolve, then they will eventually die and not be used. So I think, given that we've all put this work in, it's incumbent upon us to make sure that its future will be established. And I believe, and I think this will be borne out, that the work that Craig and Dave and all the technical contributors have made have provided us with a good framework for future development and use of this particular code. So with that, I will turn it back over to Matt Homiak. Great. Thank you gentlemen. Was there any questions anyone has for Craig, Dave or Rob? If so, you could please submit them through the chat. Okay, we'll see if some other questions trickle in and we're just getting started here. I'll lead us into the next topic here, which is an overview of the code itself and some of the features it has. So XLP, of course, is a probabilistic fracture mechanics tool for nuclear power plant piping integrity risk analysis. And as you've heard, it's a joint effort between the NRC's Office of Nuclear Regulatory Research and the Electric Power Research Institute now in its second version. XLPR is capable of modeling degradation mechanisms such as stress corrosion cracking, model crack initiation, residual stress effects, and well as a variety of mitigations to combat stress corrosion cracking. For example, NRC and every staff have been using XLPR to risk conform emerging piping integrity issues using probabilistic approaches. So XLPR is a probabilistic fracture mechanics code, so what does that mean exactly? We'll take a look at, for example, structural integrity problem and illustrate the differences between deterministic and probabilistic fracture mechanics. So assessing the integrity of a particular component, we should look at the stress, consider material properties such as toughness and perhaps the crack growth and a crack size, and that will give us an estimate of the life of the particular component. Now, looking at the deterministic approach, what we'd probably do is take a conservatively high value for the stress, conservatively low value for the toughness, and a conservatively high value for the crack growth, and then we might also put on a large crack size. And so our prediction of life using that approach will be based on a single calculation plus margin giving us a conservative result. Now contrast that with the probabilistic approach. And then sample many values of stress, many values of toughness from probability distributions as well as crack growth and different crack sizes as well. We're going to aggregate multiple calculations, give us a probability failure over time. So two different approaches, the probabilistic approach is giving us much better insight into the actual life of that component with much more data. Let's take a look at how pipe is represented in XLPR. We have basically a configuration of a butt weld and three main components, which we call an XLPR left, pipe, right, pipe, and weld. Here the weld is shown in the red. And we can assign different material properties to all those components, and that would allow us to model a similar metal weld or a dissimilar metal weld in the case of a nickel alloy. In XLPR we can model two main degradation mechanisms, stress corrosion cracking and fatigue. We can model the effects of both of those at the same time. We can model crack initiation from both those mechanisms where we can seed initial flaws. And we can also model crack growth as well. A crack model in XLPR includes two different planar orientations circumferential and axial. And we can consider multiple cracks around the circumference of the pipe. And XLPR's core business is really to model the crack life cycle. We do that with surface cracks, which grow into transitioning through wall cracks. Those are cracks that could leak and potentially rupture. And then further growth would lead to an idealized through wall crack, also leaking and would have the potential to rupture as well. We can also model how the plant operates over time. We can do that with the different operating conditions, be it pressures, temperatures, stresses. We can also model different chemical species in the water chemistry. And we can also model different types of mitigation as well. And so XLPR would be taking all those things into account and splitting up the operating history of the plant. It's shown in the chart here based on several different time intervals on all the activities that might happen in the simulation. So for mitigation, we have a couple different types we can model. And these are based on the approaches that are currently being used in industry. We can model inlays, which is, for example, a PWCC resistant material on the inside of the weld, shown in purple here. We can also model overlays. That would be, again, a PWCC resistant material applied to the outside of the pipe. We can also model a mechanical stress improvement process that actually squeezes the pipe and puts it in compression on the inside diameter. So those are all different ways that we can slow the initiation and crack growth. We can also do that with different types of water chemistry as well. So we can simulate hydrogen, how it affects crack growth, zinc, how it affects crack initiation, or both of those. You can detect cracks in two ways in XLPR through in-service inspection modeling, primary ultrasonic testing, and as well as through leak rate detection. Okay, so that's a little bit about the main features that XLPR has in the models. I want to talk a little bit about how XLPR is actually designed now, the actual structure of the architecture of the code. So we start with a framework, and this is what's going to take all the user inputs. It's built in the commercial GoldSim Monte Carlo simulation engine, and this is going to control the simulation. It's going to step the different trials through time, and it's going to aggregate all the results at the end and show them to the user. Built around the framework, we have a variety of deterministic modules that model the various physical processes. These are all coded in Fortran, and it proceeds basically through the evolution of the crack life cycle. Essentially, we'll start with initiation, and then going clockwise, move on to growth. In the case of CERC cracks, we would model the combination of the two together using a coalescence module. We can then model the transition through a through-wall crack using the transition module, all along assessing the stability of the cracks using a variety of stability modules. We have a crack opening displacement module to determine how far open the crack is, and that's used in helping us calculate our leak rates. Finally, we have in-service inspection, which is used to detect cracks that may be there throughout the simulation. Combining it all together here with the framework, you see how it works is that the framework would be calling any one of these modules, and then the module will do its respective calculations, pass the information back to the framework for record keeping purposes, and then the framework would advance to the next module as appropriate. That's just a brief overview of how the code is actually structured. I will talk a little bit about how the code handles uncertainties. XLPR has a dual loop sampling structure, so uncertain variables can either be assigned to an aleatory loop or an epistemic loop, and you might assign an aleatory loop to an uncertain variable if the uncertainty was just due to natural variation. Epistemic would be due to lack of knowledge, and this is very useful if you're trying to understand which of the many input uncertainties might need to be reduced. XLPR supports a broad range of probability distributions. Some examples on the slide here. These are the typical ones we use, normal, log normal, uniform, discrete, and viable. XLPR sampling algorithms, we're doing the Monte Carlo simulation of several. Simple random sampling, of course, but we also have more advanced sampling options, such as Latin hypercube sampling, discrete probability distribution, and important sampling. Those advanced sampling options are there to basically make the simulation more efficient as compared to simple random sampling. Here's an illustration of what output from the code might look like. Like I said earlier, we're looking at some metric over time. The main ones would be crack, leak, and rupture, and we're seeing some results here from the code. All these gray lines are showing you an individual trial or realization, and so we can see here that your results could be as bad as the gray line on the very top of the plot, but most are down on the lower region. And then we can take statistics of those individual realizations as well to draw conclusions about the system. Crack, leak, and rupture probabilities are the main ones, but there are many other outputs from the code too, such as local frequencies and things like that. Okay, so we talked a little bit about XLPR and it's built under a quality assurance program. What was that exactly? It was select elements from the ASME and QA1 2008 and 2009 standards. Those are NRC endorsed ways for meeting NRC's own 10C FAR Part 50 PENXB quality assurance requirements. And then on top of that, we also followed international software engineering standards from IEEE. We use those as guidance too in the quality assurance program as well. XLPR has extensive technical documentation, software requirements, software design, testing, test plans, and test results for every module in the framework, and then everything together as well. We verified and validated XLPR extensively. The difference there, of course, is verification is just making sure that you fulfill all your requirements, whereas validation would be making sure that you're actually getting the result that you think you should. Over 4,000 verification tests were performed. That was quite an effort that we were all involved with. And we validated XLPR each physical model by itself, so all those modules that I showed you earlier, and then also all the models integrated together with the framework. The framework we validated against operating experience, finite element simulations, and benchmarked against other probabilistic fracture mechanics codes. We also had XLPR reviewed externally. We had what we called the External Review Board, which was a panel of international experts that assessed key aspects of XLPR during the development process. So I'd like to conclude here with a short summary of some of the capabilities that XLPR has. Like I said, it can model stress corrosion, cracking, thermal mechanical fatigue. We can model crack initiation from both those degradation mechanisms. We have ways to count for aleatorian epistemic uncertainties, perform leak rate calculations. We can consider residual stress effects in the simulation as well, both before and after mitigation. We simulate water chemistry impacts, effects of ultrasonic inspections. We can also model effects of the increased stresses from earthquakes in a probabilistic way, as well as a variety of stress material mitigation techniques, such as in-lays, on-lays, and overlays. So that concludes my presentation here on some of the features in XLPR. This is not to get into the details of any of these really, but just to provide an overview of what's in the code. We'll cover a little bit later some opportunities we'll have to do some more technical discussion on the features of the code. But let's see. I think we have a little bit ahead of schedule here, so we'll see if we can entertain some questions. Matt, there's been an active set of questions coming in. We're trying to respond to them as quickly as we can. Is there any you want me to try to field right now? This is Dave Broden. They're coming in very, very quickly, and we can't answer them as quickly as they're coming in. So I don't know how we want to try to handle this. Of course, I haven't been seeing any of them during my presentation here. Let me see if I can catch up a bit. But Dave, Craig, if you wanted to field any that you might be a little head on, please go ahead. I'm trying to read through them as quickly as I can. Here's a question says, regarding the in-service inspection input parameter, you can provide your own curves, input as a distribution. Yes, you have curves that can be used. You can provide your own curves. There is sampling applied to the POD inputs. There is great flexibility there as well. Yes, this is Dave. There's been several questions about CC and fatigue growth, and the two phenomena are calculated separately each time step, and the contributions are added together. I think I may have said that in the comments, but there are other questions about it, so I just wanted to make that comment verbally. But we also had someone asking about multiple cracks in the proximity rules and whether those proximity rules were applied conservatively, for example, high-moderate flow, and the answer to that is the user has the flexibility to determine those coalescence rules, actually. You can pick a best estimate, or you can skew it either high or low if you wanted to, depending on what you're trying to do. Matt, there's a question on slide 12. What is the definition of crack probability and leak probability on that slide? This is slide 12 here. Definitions of crack probability and leak probability. On the example here, we're looking at probability of a crack, so that would be using the initiation models, whether or not those models predicted the pipe to have a crack or not. Initiation isn't generally not very likely, so we're seeing most of the results there down in the lower orders of probability. Whereas at the top, you have a 1.0, and that would be a 100% chance of having a crack. In this particular slide, one of the key points of XLPR is it does generate some specific pre-canned results, but the data behind those results is accessible to the user, and that allows you to develop many different results. So the definition in many respects would be whatever you want it to be in terms of how you process the data. So the code doesn't just generate a few answers, and that's all you get. There's great flexibility to the user and how to pursue the process. There's several questions about materials and SIRC and axial flaws. Let me start with the SIRC and axial flaws. Both circumferential and axial flaws are modeled in any one analysis. However, they do not interact. What I mean by that is that the SIRC and axial flaws do not interact with each other. They both may influence the probability of leak or rupture, but they don't react with each other. In terms of materials, the input deck for XLPR is very flexible to put in whatever material properties that you'd like. However, if there is special material laws or special growth laws that aren't standard, then those aren't available right now in XLPR. There's a question that says, would it also be possible to access the results from different steps of the calculations, for example, results from just the stability module or transition module? And that's one of the tremendous benefits. It's also computationally and memory-wise challenging. And this code does accumulate all of that data, and as will be briefly mentioned and demonstrated in the next section of the agenda, you can reach in and grab that data to go calculate the additional results that may be important to your analysis. So the basic answer is yes. There was another question about proximity, and they wanted to know if the ASME rules are programmed in XLPR, and I think it was Matt that mentioned about the flexibility. And I believe you can use that flexibility to mimic the ASME proximity rules. It's not an option where you select a button and it becomes the ASME proximity rules, but the flexibility allows you to put in the distance between the cracks and where they interact so you can mimic the proximity rules in ASME. There's a question that says, is each user supposed to perform case-specific validation of the code, or as long as the code is used within the applicable range of input parameters, the user can assume the code is validated? That's a bit of a more complicated question, but fundamentally, we have extensive documentation of the range of validation that currently exists for XLPR from our development code, and it would certainly be incumbent upon the user of the code to either ensure that and be able to demonstrate that you are applying it within that range of validation, and if for some reason your application goes beyond that, then it would be incumbent on the user to address that difference. We've had another question here about the crack shape. Does the user need to assume a shape in the orientation first? And the answer is that the user can set the crack shape, actually, the aspect ratio that's an input that the user can control, essentially. And that's whether the crack is set up as an initial flaw or whether, you know, the size of the crack that's initiated after some period of time by the initiation module. So I think some questions here I've been seeing, like, you know, about inputs and things like that, and usually the answer in XLPR is, can I change it? And the answer is yes, the inputs for the large part are completely controlled by the user, and they're easy to change. There's a question. Oh, sorry, go ahead. All right, there's a question about the changing, the changes of the stresses due to a change in flexibility of the piping. And there was a lot of discussion technically about that. What the code does do it, it allows you to put in a factor of the amount that the thermal stresses would be reduced due to a crack. However, it doesn't give, however, the code or the manual doesn't give any guidance exactly on what number to pick. It leaves up to the user to determine how much of that particular load is displacement control loads are shed. Flexibility is there, but the user needs to know something about their piping system before they can decide what factor to use. There's a question about the non-US materials such as EDR, reactor type steels. And again, as Dave indicated, there's tremendous flexibility to put your own information in. However, you also have to consider what are the degradation mechanisms that you're concerned about and be sure that it primarily addresses now fatigue and PWSCC. And so you would have to understand those materials in the context of those degradation mechanisms. There's also a question about changes in toughness over time, thermal aging or radiation. That is not explicitly addressed presently in SLPR. There are maybe some tricks and ways that you could trick the code into addressing that to a limited degree. But it is one of the things that we have noted might be of use to add to the code. There's a question about crack opening time. And so the code works such that it calculates the crack opening displacement at every time step. There are different modeling assumptions for the flaws as they transition from a surface crack to a two wall crack. And the crack opening displacement is calculated at every time during that transition. I had to turn in this bit of SLPR as opposed to deterministic analyses. Time is considered. We look at the life cycle of a crack from initiation to an ultimate failure and all points in between and that data is available to reach into the results and extract for whatever analyses you might want to apply it to. There's another question, can the tool be used on pipe like vessel internal components, for example, core spray jet pump risers? And the answer is if you can resolve your problem into the basic but well geometry that I showed in the beginning of the presentation, then you can use SLPR to do that. Another question is a sample problem available. And the answer is yes. In the user's manual, there's some sample problems that the user can work through to get acquainted with the software. And we'll also be providing some additional training along with the code when we distribute it that we'll have even more sample problems to work through as well. There's a question about PWSEC versus IGSEC. And yeah, the code was originally centered PWSEC since the thought was the LBB issues in PWRs. However, the crack growth models are generic enough that one could use IGSEC models if the parameters are appropriate. You compared the deterministic and probabilistic approach, what failure probability would you typically associate with deterministic approaches and common codes like ASME section 11? That's probably too complicated a question to get into here, but I don't know Dave, unless you have some some thoughts, but I'm sorry, I was reading another question. Which one were you working on? It says you compared the deterministic and probabilistic approach. What failure probability would you typically associate with deterministic approaches in common codes like section 11? Yeah, that's a tough question. Whenever you use safety factors and bounding cases, you're pushing yourself to the almost unrealistic limits, right? So that's a hard question to answer. Does XLPR handle embedded flaws that can grow and transition into a surface flaw? I think the fundamental answer is no, we did not build embedded flaws into this. I don't know Dave, if you have any additional thoughts there. Nope, you are correct. I have a question on verification tests and they said, how did you perform 4,000 verification tests? So the verification tests were against the requirements that were set out for the software at the beginning of development, and so if there was a requirement it needed to be tested. Some of those requirements were tested statically, where for example, you might have just examined the source code to make sure the particular element was there. Other verification tests were done dynamically, which would have involved actually running the code and looking at the results and making sure that they were meeting the acceptance criteria. How did we do it? We had a bunch of people involved to run those tests and make sure everything was working properly. There's a question about normal operation versus faulted conditions. The code allows you to put in a normal operation that allows you to put in transients that can mimic fault conditions. And it also has the ability to put in your earthquake loading with the probability of occurrence on the earthquakes. Here Dave, this is a good one for you. Does XLPR account for how the moments change in the pipe system with the change in flexibility of the piping? We'll just use the unpack. Yeah, I think I tried to address that one earlier. Okay, I missed it. A factor that can be put on the loads to account for the reduction in the displacement controlled loads. However, the code doesn't give any particular guidance on what value to choose. It's up to the user to choose that number based on their piping system. Yeah, I apologize if we're not getting to all these questions. I'm trying to read them as fast as possible, but there are an awful lot of questions coming in. Yeah, I just responded in the chat. Also, we will, any questions that we can get to today, we will go ahead and respond to in the NRC meeting summary that we put out afterwards. We are winding down on the time for this particular topic, so I'd like to go ahead and move on. We'll circle back at the end of the meeting if we have more time here. So our next two agenda topics will treat in a combined way. So we're going to involve an actual demonstration of the code itself, and as well as some discussion about the different applications of the code. And supporting those two topics, we have Dr. Cedric Salabari. He's a mathematician at Engineering Mechanics Corporation of Columbus. He's been involved with XLPR since the beginning with version one. He was part of the team responsible for developing the computational framework. He's a great resource to have and he's trained me on how to use the code. We also have Marcus Burkhart. He's a senior engineer at Dominion Engineering. He supported a wide range of areas in both XLPR development and applications on the upper side. In addition, Nathan Glunt, he's a senior technical leader in the EPRI Materials Reliability Program. There, his focus is on applications of the code beyond leak before break. All right, thank you gentlemen for being here today. Cedric, would you please take it away? Thank you. So now we move to a different topic which will demonstrate how the code works and we will list some of the potential applications of XLPR. One important point is this is a demonstration. This is not a tutorial. So there are many, many things that we won't cover with this demonstration. What we plan to show you is how we set up the input via the XL file, which will be the mouse-direct version once your familiar with the code, or the simulator, which is not end-dome, kind of a wrapper around XL that simplifies when you first start with XLPR. Once we have been through set up a problem, we will show how to use golfing to run the code. And for this, we'll use a player, which is a pre-version of golfing. And finally, we'll present results from an already run case so you can see which outputs are available by default and how to access other kind of output. And we will illustrate this with an application for need before break or LBD problem. So the problem we'll look is this one, which looks at primary water stress, corrosion, cracking, so PWS issues, for reactor vessel outlet nozzle on a generic West India wood for loop pressurized water reactor. This is a generic case, so we don't represent a specific power plant. We have used average or representative value to cover any kind of plant from these categories. In the screen, you can see the global geometry. The LVON we consider is a 34 inches pipe, and the thickness is around 2.6 inches. In the operating condition, we consider the temperature is around 320 degrees C, and the pressure is a little shy of 50.5 megapascal. The simulation time will do is for 80 years with a one month time step, which represents about 960 times that total. And in this analysis, we look at the impact of in-service inspection and mechanical stress improvement process, which is one of our mitigation case. So now I will move to the Excel file. In Excel PR, when you want to modify something, all these inputs, you will do it from the Excel file. The Excel file is your input set, and the Excel file values will always have the same name that you can see on the top here. It's Excel PR dash 2.1 input set dot Excel SX. The data in the Excel file is separate into different tabs. And I will go through the tab now. The first tab, the one on the form of left, is the user option. So the user option is how you set up your problem. It's mostly a series of on-off switch where you decide what you want to do with the analysis. Do you want mitigation or not? Do you want to consider axial, circumferential crack, boss? So for instance, this one lets you decide which kind of direction you're interested in, and so on. You can select the mitigation, you can select the type of load or stress you use, you can define your transient, and so on. When you move next, you will add the properties. So now we talk about the inputs of the model, and all of these inputs can be both constant or distribution. This will cover the pipe geometry. It will cover the flow size, depending on if you have a flow due to fatigue, CWCC, or a initial flow. You will have the operating condition, and so on. Next to the property, you will have four tabs, left pipe, right pipe, weld, and mitigation. All of them cover the material property. So this is where you will define the material property for your left pipe, your right pipe, the weld, and if you consider inlay or overlay, the mitigation material is used for inlay and overlay. This one will be equivalent to the property except we cover the material properties such as yield strength, ultimate strength, and so on. Finally, here we have the WRS, weld residual stress, which defines the OOP, weld residual stress for the axial crack, and the axial weld residual stress for the third crack. It's slightly different from the other because the WRS is represented as a function through the thickness of the pipe. So it's a profile. This profile can be constant, or it can be also inserted with a distribution, and in this time the distribution will be only normal. And finally, the last part the user can change is if you want to have fatigue in your problem, you can define the transient. So the first tab here, you can define up to 20 transient where you can set the evaluation interpreter and pressure over time. And in the second tab for type 1, type 1 into type 3 transient, you can define the properties of your transient. When it starts here in this column, when it ends, the frequency, how many events you have per year, and every time you have an event, how many cycles you have. So these are the basic tabs the user can change when it's set up, the problem is interested in. You have three more tabs here that I will cover just a little later. These are informative tab. They are not changed by the user, but they give you additional information when you want to make a run. As we discussed, XRPR is a probabilistic code. That means that you can define the input as deterministic, or you can define them as probabilistic. The way we run this code is a Monte Carlo approach. So it means that once we put the uncertainty in the parameter, we'll have a distribution to define them. We'll sample and we'll just run the code deterministically as many times as decided by the sample size. In XRPR, any of the input we have from the property tab to the T-value tab can be uncertain or deterministic. I will take an example here for the ESPY, which is the effective full power year, which indicates how much of the year you can really use at full power with your power plant. Right now, the data source is set as constant in column E, and the deterministic value in column H, you see highlighted here, is set to 80. We mean that in our theoretical example, we consider that we use the plant at maximum capacity all the time. You may want to change the value, you can do this deterministically, but you also want to change with a distribution. In this case, if you click on the data source, you can select one of the two uncertainty loop we discussed, Matt talked about, which is the epistemic loop and the aleatory loop. As soon as you select one of them, you will see that the deterministic value is now great because it won't be used by the code. But on the side, now you have the probabilistic values that can be defined. And I highlighted it here. The first part is the distribution type, you select the distribution type and it represents all the distribution available in both things that are available in XRPR. In this example, it was normal, but you can change to whatever you want as an exponential distribution and so on. Now, one of the things is, depending on which distribution you use, you will have different parameters. In the Excel files that define as parameter 1 up to parameter 8, it's not very informative. And this is where you can use one of the tabs. And if we go to the end and look at the second tab from the right, which is called drop instruction. If you click on it, you will have the list of the distribution you can consider and which parameter corresponds to what. For instance, for the normal distribution here, the first parameter is the mean to set the standard deviation, which is the classical parameter for the normal distribution. And then because you can define this as a truncated, you can also define a minimum and maximum. So with Excel, you have control to all the input you want for the XRPR model. The only thing is, when you start to use it, it may be kind of intimidating, because if you go here for the property, you have a lot of properties. On the loads, on the probability of detection, if you have inspection and so on. If you go to material property, you also have a lot of properties, so it's easy to get lost. That's why Excel PR is also including the CMeditor when it's released. The CMeditor is a wrapper around the Excel files that will be able to read the Excel file and to write into the Excel file and give you a more graphical representation. So now I will open an example of simulation here. The same value will do it. Generally take about 20 seconds to open. During the opening of this, you will see when you run the CMeditor that you can open a simulation and just work as it was your simulation instead of working with Excel. You also have a database access. The database allow you to create a set of data for representative weld, also for representative material. And once you have populated this, you can reuse the material properties in a different position and same thing for the welds. So it helps you to create cases a lot faster than initially. So once it will be open, we'll see that the structure here is kind of the same as the Excel file except now it's graphical. You have access to the left pipe, the right pipe, the weld mitigation. Well, here we only consider an SIP so we don't have different mitigation material and the different property like operating conditions and so on. For every button here you have different menu on the side. You can click on it. If you want to modify here, it will be kind of the same as with the Excel file. The only difference here is there is a face safe. Right now we are in view mode so you cannot change anything. You need to click on edit so you don't change by mistake some value. Once you press edit, you will have access to the data source as we had in the Excel file. I can put this as epistemic now and you will have a defined distribution this time because it's graphical. You have directly the property here as mean standardization, mean and max. And if I change the distribution, you will have like for a triangular distribution. The first part tells you if it's a regular or log triangular distribution and then you can define the minimum and maximum. So up to now what we have seen is the Excel file that can help you set up the problem or the wrapper, the simulators that make simplify some of this implementation at the beginning. Once you have set up your problem, you can move to the Excel PR file and as I say here it's open for the player version. And the player or the full version of GoldSim will start with this dashboard. GoldSim purpose is to run the code. So the only controls you have is not in the input input that is already defined but just a sample size and the capability to run the code. And this is what this first part at the top does where you can change the sample size and you can run the code. And then to look at the result and work with the result, which is the bottom part that we will cover just in time later. When you want to run the code, you can select your sample size. So here is a button here on the top left corner where you can select the sample size. Right now it was five. And you can see here in the GoldSim run control that also helps you run. It says that we have run zero out of the five renovations during edit mode. If I go here and I change the number of renovation of the first line here to six. And now the sample size will be six and the controller tell us we have zero out of six renovation per form. When you want to run the model, you can use a controller with the play button, which is this triangle, or you can click on this icon here which says run Excel PR model. So whichever you run, it will be the same thing. The first thing GoldSim will do is take all your information from the Excel file and download into memory. So GoldSim will have up-to-date information of which case you want to run. And this is what it's doing right now with importing the data. If you also download in memory all the DLL and see if you are missing some of the DLL, you will have a message saying, oh, you're missing the DLL is not ready. Once GoldSim has everything into memory, and as I said, it should take about 20 seconds, then it will start to run the code. And you will see it here in the GoldSim run controller when it will state that realization one out of six will start. If you move through the controller at this time, you will see where you are in the realization. So for instance, when I move the mouse, we were just at the beginning with 0.75 years. Here, well, the calculation is fast enough so we don't have time to see, but here we can see that we are already on the fifth realization. Out of six, and after about 30 seconds, the case will be done. So after some export of results, GoldSim will just give you a message to say now the run is finished. And you can see some message warning in case there is a problem or error. Usually, you will have one message shaking some of the value and no fatal error, which means the code has run perfectly and we don't need to see the one made, so we'll press now. And now, since the code has been run, we can look at the result. Well, the problem we have set up as probability of event around 10 to minus three. So it means that it happened once, twice, three times, over a thousand case. So if I do just six realization, we won't see anything. So I have another case now where here we have run it for 10,000 realization. So here you see the sample size was 10,000. And if you look on the run controller, it says that we have run 10,000 out of the 10,000 realization. And it took about three hours and 40 minutes here to run this 10,000 realization. Now, we are not interested anymore with the sampling approach. We want to look at the result option. Since we continue to potential direction, you will have result for the actual crack, which are the top two button. And result for circumferential crack, which has a bottom two buttons. Our case didn't have actual cracks, so we only look at the circumferential crack. The bottom left button is for result directly. The bottom right button is the error dashboard to check is everything went well. What we recommend when we use Goal Seam is to check first the error. Some of the error will stop Goal Seam. And so you know what's going on. Some of the error may still be okay. They will do out top bound, but I can still calculate. So the cards will calculate, but it's important to see what's going on. Because if you have an error like this and you calculate, the result may be meaningless. So we'll go first to the error dashboard. The error dashboard will give you information for all the module called by Goal Seam. As we say, Goal Seam is just a wrapper of the XRPR. All the calculations are made in independent modules so they can be replaced in the future for SDLL. And all the input is defined within Excel. So for each of the modules, there will be a check to see if the module replied correctly. And the legend tells you what's going on. A check mark, green check mark say, okay, no error happened. The warning signs say, okay, there may be some warnings. The result may still be okay, but see how to note this. The stop sign means that one of the modules had a fatal error. And the red flag means you may have multiple error or fatal error. Finally, the gray box means, for this realization, nothing happened. We didn't call this module. In this case here, what you see is we have either a module not called or check mark. So everything went well. So we're good. We can go back to the global setting to look at the set result. Or we can go directly to the circle for and show results here. Okay, we have a question here. So I will respond here. What we use here is Gold Seam version 11.1, which is not the latest version of Gold Seam. Now Gold Seam is at the version 12.1. We are doing testing to see if XLPR still works with 12.1. But the official version we use for XLPR is 11.1. Now, whether you buy the full license of Gold Seam or not, you will be able to have access to both the 12 and the 11 version. So going back to the results, the result dashboard has also two parts. The first part are the general results. And generally, these are the ones you're interested in risk analysis. These are information on adverse events like a crack occurring, leakage occurring, rupture occurring, or locus and so on. And these will be presented as statistic probability just, you know, the mean over all the relation and tells you what's the risk, what's the probability of having such an event occurring. The second part is more crack specific results. You can see represent them as statistic, but they may be more informative when you want to look at specific realization and see what's going on. And one of the things of Gold Seam is, as Craig mentioned earlier, Gold Seam saved all the data. So we have access to all 10,000 realization here if we want to look at this. Let's start with the general results here. And the detection effect is probably the most informative output we have because it looks like a variety of value. When you open it for the first time, you will have all this value which are just the currencies because it's only for one realization for realization one. Of course, it's not informative. We discuss about this. We have probability around 10 to minus 3. If you look at just one out of 10,000 realization, what you will see is mostly zero. So you will go to the third tab here which says display realization and you will move to statistics. It's the first time you move. It will take about 20 seconds to open. I opened it early on so we can see it here. And here we have direct result from Gold Seam. We can look at the probability of having the crack was moved. The mouse here will see it's around 10 to the minus 3. The probability of leakage and rupture at 80 years are around 7 to the minus 4. And if we add inspection, leakage goes down to between 1 and 2, 10 to the minus 4. And rupture goes between 6 and 7, 10 to the minus 5. So these are the kind of results you can have with XRPA. In this example, the case run was with MSIP. So this is why the curve starts to flatten out around 40 years. Well, these results are just graphic representation. You may want to use the data and test them and present them with different codes. So you can go to the second icon here on the top which says state table. And then you will have a table of results. Once you're in the table of results, you can work as if you were in Excel. So that's the value control C, control V copy in whichever program you want to use. Now these results were kind of averaging risk value. And as we said, you may be interested sometime in looking into more details for a particular realization. So I will take here an example of looking at secret evolution with depth and inner affluence for one realization selected. So once again, realization one gave us nothing. But if I go to realization 18, so this time I don't change the display. I just change the number of the realization I want to see. And what you will ask is the evolution of a crack for realization 18, a crack occurred. It occurred at time 28, 29 years. But it starts to grow up to year 40 when you begin to have the mitigation, the MSIP mitigation, and the depth stop once you add the MSIP mitigation. Because it was probably adding the MSIP into a strongly conservative area where suddenly the load and stress were negative and nothing could happen. I can do the same thing with the inner affluence here where I plot directly as a result of realization 18, where this is the same story for the inner affluence direction growing. But the crack starts around 29 years. It grows here and at 40 it will still grow, but at a lower pace than it used to do before MSIP. So here is kind of a quick demonstration of XRPR, how it works, where you use Excel to set up your problem, use Go Sim to run your problem and have access to the output and you can extract the output. Now, you may be only interested to use XRPR as use as running and looking at the results. You may also be interested in looking at the model. And well, if you have the full version of Go Sim, you will have access to the full model and you can change whatever you want. If you have the player version, you cannot change the model, but you can still look at the model. If we go to the Go Sim run controller here and you look at the last button and you click on the button, you will be able to navigate through the model. There is an option which is navigation and you can go to any dashboard we have looked, but you can also go to the model route and see the structure of Go Sim. In this structure, you will have the classical view of an object-oriented program where you have the object defined and some container where you can put the object and the link between objects. We won't cover this now because it's a huge model with many layers. The only thing I want to show you, and I will move quickly here, I apologize, is that we have the same structures that Matt presented previously for the crack evolution. In our Go Sim model, we have the same module with crack initiation, then crack growth, then the possibility of coalescence for circumferential crack where transition from a surface crack to a swirl crack, the stability that can lead to leakage or that can lead to pipe rupture. And if you have a swirl crack, you can calculate the crack opening displacement, then the leak rate. And finally, you can also estimate the probability of detecting a crack with in-service inspection. So all this structure is available for you and you can look at every element and see what the value of this element was at any time for any realization. Okay, so let's go back now to the PowerPoint presentation. So we have looked at how we can run and look at results. Now, this is kind of the result we have after we extracted the value. Here, we have the impact of inspection and the impact of MSIP outline. If you look at the blue and the red curve, which are leak and rupture, and compare to the green and yellow curve, which are the same probability in service inspection, you can see that inspection gave us about one order of magnitude for leakage and two order of magnitude for rupture if you apply inspection. Now, the effect of MSIP is seen when you look at the curve of the same color. The plain line is without MSIP and the dashed line is with MSIP applied at 40 years. You can see here that the crack initiation becomes almost flat once you reach MSIP. And the same thing with a small lag for leakage rupture. And the only difference with MSIP was the difference in the WRS profile. And you can see here the difference in WRS profile. And we have also the reason why the other cracks stopped because we were in a strongly conservative zone for realization 18, so it stopped going through the depths. Now, XRPR has a certain number of outputs, as we said, that we can use directly, but it's not limited to these outputs. You can find intermediate results and make some additional analysis. And as we say that to begin with, we will demonstrate this for LBB application. The current LBB procedure follows the NRC standard review plan section 363. The way it works is you estimate the crack size required to have a 10 gallon per minute leak rate. You estimate the critical crack size that will lead to rupture. And then you can calculate a ratio between critical crack size and leakage crack size. And the acceptance criteria is set as this ratio equal to. So if you have a ratio greater than two, it's acceptable. This procedure is deterministic and static. Deterministic because we just use one value, static, because we don't need to grow the crack. We just need to calculate these two values. Uncertainty is not ignored here. It's considered, but it's represented with conservatism. One of the conservatism is on the first bullet. We use a 10 gallon per minute leak rate, which is 10 times the expected detectable leak rate of one GPM. The second one is use an acceptance criteria at row equal to two instead of one. It will be less, even more margin. So if we want to move this to the probabilistic LBB, what do we do? Well, the first thing is why do we want to do this? Well, the reason is the standard review plan was not applicable if there is a negative degradation mechanism such as PWCC. So that was the driving factor of going to a risk-based framework. And since it has been developed, we have an increase in knowledge, but also in computer capabilities. So we can do dynamic codes now. We can do probabilistic fracture mechanics. And we can include more mechanism such as surface fracture, impact of axial fracture, effect of mitigation, inspection, and so on. And one of the reasons we want to do the probabilistic approach is it can help us reduce the conservatism to go to a more realistic analysis. And this is what we'll see in this next slide where we calculate the LBB ratio probabilistically. So for this case, we had to run a larger sample size, 100,000 run because not all run gives you rupture. But since we have a larger sample size, we can estimate this LBB ratio exactly the same way for each realization. And once we have a set of realization with LBB ratio, we can sort them and create a distribution. We can do this for 10 GPM, and that's orange curve here. And we can do this for 1 GPM, which is a red curve. Our criterion threshold here of 2 is circling blue. And this is our limit. We don't want to go low. What we see is in both case, we are okay. But the difference is if you use a conservative 10 GPM ratio, your ratio vary between 4 and 6 with an average of around 4.4, 4.5. If you use a more realistic 1 GPM, your ratio vary between 8 and 12 and with an average of around 10. So this result quantifies how the conservative them can affect the answer. In this case, everything was okay. But in other case, you may be below the limit with 10 GPM and above the limit with 1 GPM. One of the things here of XRPR is, as we said, this LBB is not a direct result from XRPR. So to extract this, we needed to do some post-processing of XRPR. So now I will let my colleague Marcus Burkhardt from DEI to talk to you a little more of what can be done when it's not a direct result from XRPR, but you can still have access to it and what you can do. Thanks, Cedric. So Cedric touched on a few outputs that we considered in the collaborative NRC and EPRI typing system analysis that were not default outputs directly calculated within XRPR. So those included the LBB ratio, which is the ratio between the critical crack size and the crack size of the given leak rate, as well as the time between a given leak rate and a rupture. So rather than modifying the XRPR source code for these typing system analyses, it's decided to perform the needed post-processing outside of XRPR. So additionally, performing post-processing allowed us to combine results for multiple runs. Increasing the total number of realizations that could be considered in an individual analysis case and also allowing us to evaluate lower probability events. Next slide. So how did we do this? We developed a third-party software in Python to help automate the data extraction and post-processing, and that was broken into two different scripts. The first utilized an open-source Python package called PyOtagoey, and this was used to automate extraction of data from XRPR for every single realization for multiple outputs, with final outputs and also intermediate values calculated in XRPR, and it saved those text files for post-processing. The second script was then used to perform computations and to develop statistics for several outputs of interest, including default outputs, such as the time to rupture, as well as non-defile outputs calculated from intermediate values, such as the LBV ratio at the time. So this is just an example of showing you that it's possible to access intermediate values and also calculate additional outputs in addition to those provided in XRPR. Back to you, Cedric. Thank you, Marcus. So to conclude this demonstration, what we have showed you is except you are once developed to be flexible and applicable to different problems. You have a large number of outputs directly available, such as probability of first crack, first leak rupture, locas, where you can define the locas. You have also additional outputs that are available via intermediate results, and you can extract them with your party software or with some simple changes of code. The point is everything is safe, so it makes this kind of huge fine. That's why sometimes we have to run separate files, but everything is safe in those things so we can access to everything whenever we want. Now, this demonstration was done for LBV case, but XRPR is not limited to LBV, and we can use for a variety of different problems that we studied. And now I will let my colleague, Nathan Blant, from AAPRI, to talk about potential application of XRPR. Nathan? Thank you, Cedric. Now that everyone's seen a little bit about how the code works, I'm going to just very, very briefly discuss where we think XRPR could go next. Remember, this is just the start of XRPR, so you'll be hearing a lot more about it in the future. But as you heard, the initial focus was on leak before break issues that are associated with primary water stress, corrosion, cracking into somewhere metal welds. But as Cedric just said, I want to point out again, XRPR is not a leak before break code. XRPR is a probabilistic fracture mechanics code. It was designed with a modular structure and a flexibility to consider almost every single variable as distributed. And so there are already cases where we're looking into where else we can use this code beyond leak before break, as the code already is. And then we're also looking at how we can exercise that flexibility from the modular structure to take the problems beyond XRPR or beyond leak before break as well, excuse me. So the first and most obvious use of XRPR that we can see is how we can optimize inspections and repair replacement strategies. So as you've heard a lot about already, XRPR can take care of both primary water stress, corrosion, cracking, and the fatigue degradation mechanisms. Not only that, but you can also model in your mitigation. So that's also key. So, you know, right now it's set up excellently for piping welds, butt welds. But the modular structure lends itself well to expanding the use to different geometries. So this is something we're going to keep pursuing. Beyond that, if you do a big picture, you know, you can look at redefining design based on break size. You know, this is a big question to ask, but if we can use XRPR to prove that the probability of rupture for a large break loss of cool and accident, so large break loco was indeed extremely low. And you can combine that with PRA insights for a consequence analysis. Can we possibly change the manner in which we evaluate large break loco? Especially as a double into 18 break. So this is just just a big thought, you know, it's a big problem. It could have a lot of benefits, you know, reduced EQ impacts, increased containment, social margins, increased safety system response time margins. You know, this is something we, you know, love to look at. This is a long term project. So next slide please. Thank you. So when it comes to licensing of high burnout, high enrichment core designs, we have fuel out there. So we want to see if we can use XLPR and the insights about the probability of rupture that can be gained from the code to see how we can apply it to that new fuel. High energy line break population, you know, there's just some ongoing work already. You know, for class 1 components, especially the, the postulation of intermediate line break locations is, is based on stress limits or human usage factor for fatigue limits. Can XLPR be used to either, you know, provide us some insight into a better way of doing this or improve the way we evaluate high energy line breaks. And finally, the last one that I have here is balance of plant systems. So BOP, you know, we know BOP is important. Can we use XLPR to look at local frequencies and to that project on how we design, maintain, inspect those BOP systems. And only that could we actually apply XLPR to certain BOP systems that are affected by fatigue to lend itself to how we inspect those systems or how we design those systems. So this is very, very brief rundown. I just want to make sure everyone's aware that, you know, this is the first step in XLPR. There's going to be a lot more coming in the future. So stay tuned. Thank you. Thank you, Cedric and Nate and Marcus for the excellent presentation. Now we're going to go to a topic Craig's going to introduce here, which you've probably all been waiting for. How are you going to be able to get a copy of the code? I've heard several times. Let me see. Did I? Am I unmuted? Yes, Craig. You are. We can hear you. Okay. Thank you. I'm sorry. I forgot to do that. We've, we've referred several times to having developed the code within a software quality assurance environment and the resulting extensive documentary record for context. Now, when we refer to XLPR, we are broadly referring to the software elements, both the source and executable elements. The user manual, input databases, project summary, technical reports, training materials, and the software quality assurance records. So many elements when we, when we make this general reference to XLPR. So to be more specific, our initial priority is for potential users to have ready access to the code in order to become familiar with its capabilities and the potential applications. Therefore, this initial public release will include the executable software elements that are necessary to set up and run the code. The user manual, specific input databases, a series of supporting summary technical reports and various training resources. And for a portion of the prospective user community, it's likely that this will be all the information that is needed. Other users will require access to the QA pedigree detailed documentary records to support their use of the code in a nuclear QA environment. For XLPR, this constitutes in excess of 100 individual documents, several thousand pages, as well as, of course, the source code elements of the software. Management of a nuclear QA pedigree code requires rigorous maintenance of these supporting QA artifacts and a degree of ongoing two-way interaction with the user community. Therefore, we plan to provide access to this full set of documentation that constitutes the XLPR QA pedigree through membership in an XLPR user group. We will provide additional details regarding the establishment of this user group in coming months. So, right now, we're focused on the top half of this slide of what's being in this initial public release. The USNRC and EPRI have jointly agreed to an initial distribution plan that allows for distribution to US domestic entities and individuals, as well as to entities and individuals in a specific list of approved countries. These criteria are summarized on this slide, but will be more explicitly stated within the end user license agreement that all applicants will be required to review and accept prior to the granting, granted access to the code. The list of approved destinations presently includes this set of 32 countries. Requests from entities or for citizens from non-listed countries will require further review and will therefore be handled by exception. So, if your country is not listed here, that doesn't mean that you can't get the code, but it will have to be considered separately. So, distribution of the XLPR code is being managed through the EPRI distribution process, and it will be accessible starting from either nrc.gov or from EPRI.com, from our respective home pages of the two organizations. If you're starting from nrc.gov, you would navigate to the computer codes page, and that will provide relevant information about XLPR as well as having a link to the XLPR v2.1 abstract page on EPRI.com. Likewise, if you start from EPRI.com, navigating to the abstract page involves just a simple search, and I'll show a little bit more of both of those paths in the next slides. Once you get to the, on the EPRI.com, the EPRI, the XLPR v2.1 release abstract page will provide additional pertinent information about how to proceed. Your request will be screened, and if the applicant meets the access criteria, they will be presented in end user license agreement to read and agree to. By accepting its terms and conditions, the applicant will be testing that they and any other end user meet the XLPR access requirements that are explicitly stated in that license agreement. The applicant will then receive an email containing further instructions how to access the release package via a secure FTP site. Now, all these necessary links and web pages are not yet in place, so you'll have to be patient with us a little bit longer. We're working on getting all that in place, but the following example kind of showed you how this process should flow. If you start from nrc.gov, if you've been here, boy, you're familiar with this tab of options across the near pop. And under the about nrc tab, there's a couple of ways you can navigate from their drop down list or from this list on the left hand side. If you go to the research section, you'll see this display of research activities. And in particular, for our interests, the items of computer codes that provide general information and then obtaining the codes of how to access the various codes listed, including XLPR v2.1, which will be added there soon. This will take you then to the abstract page where you'll find the instructions to actually initiate the process. If you start from epri.com, the process will be not terribly different. I will point out that if you go to epri.com right now, it does not quite look like this. As it happens this coming month, we will be rolling out a different look to the epri.com home page and as well as subsequent screens. So I've tried to reflect what that should look like later in May in these examples. You'll find a search icon in the upper right corner of the epri.com home page. So that icon will open this search feature and you can type in the product ID when that's available to you. You can just type in XLPR and search. That will take you to a results page. And this obviously is not an XLPR results page, but it's what was shown in the beta version that we have available right now. So you would, from the search results, XLPR v.1 should be listed at the top and readily available for you to select. That would take you to the XLPR v.2.1 abstract page. And again, this is clearly not that. We do not have that posted yet. This is just an example of roughly what it will look like. It would identify information about the code, the release package, and in a box in this general area will have additional instructions for how to request the code. And that will require that you provide basic information and, of course, as noted before, that will be quickly vetted. You'll get a response, the end user license agreement and subsequent steps. So when will the XLPR code be released? This is our final testing. At this point, we've just completed several maintenance actions, including verification of the software. But now it is undergoing some independent functionality and usability testing. We are continuing to complete the necessary updates to key pieces of documentation and the training materials. At this point, with a clear line of sight to completion of all of those remaining tasks, we have set the distribution date as May 28, 2020. XLPR runs are computationally intensive and run times can easily extend into hours while a wide range of hardware platforms can run XLPR. Clearly, as this is no surprise to anyone on this call, more capable platforms certainly will offer runtime advantages. GoldSim is a Windows-based environment, and XLPR 2.1 has been tested in Windows version 10. And spreadsheet, as Cedric displayed, is an Excel spreadsheet. And so you need a current version of Excel. We have been testing version 2.1 with Excel 365. XLPR framework is a GoldSim model. The model files for both the GoldSim player version as well as the GoldSim Pro version will be provided in the release package. However, each user must obtain their own copy of either the GoldSim player or GoldSim Pro. The player can be downloaded from the GoldSim website at no cost. But user flexibility and exercising the features of XLPR is more limited. GoldSim Pro has a fee-based license, but it will allow the user to access all features of XLPR. And it's important to note once again, as Cedric pointed out, that XLPR was developed and fully tested with an earlier version of GoldSim. And presently, it is not forward compatible with the latest available version from GoldSim. So as noted here on the slide, it's GoldSim 1117 that is compatible with XLPR. In addition to the training resources that will be included with the release package, we do plan a series of several focused informational webinars in conjunction with the Distribution Go Live date that are intended to complement the training materials in the release package. And Matt will cover that in a little bit more detail shortly. And finally, watch for announcements later in the year regarding formation of the XLPR users group. And with that, I will hand it back to Matt. All right, Craig, thank you very much for telling everybody how to get a copy of the code. So I will go ahead and take us to the end of the presentation here. And I want to just talk a little bit about some other training that we're going to be doing in the future here. This is a general session today just to provide some information about the code, things like that. But we have some other opportunities available where you'll be able to learn more about the technical details. Okay, so we're going to hold a technical seminar on June 3rd. That's right around release time of the code. And then we'll get more into detail about different models that are in XLPR. And those kind of technical aspects there and how the code operates together with all those modules. And then after that, we'll have some additional user sessions planned every two weeks. And the format of these will be to demonstrate some of the key features of the code, provide some hints and tips. And then also to devote a good portion of the time to interacting with users and helping them use the code and answer any questions that they may have. The first of those will be on setting up the inputs. We're looking at the week of June 15th for that one. There we may cover some of the simulation options, setting distribution types and things like that. The next one would be on actual running the simulation slated for the week of June 29th. We'll talk about, you know, running Gold Tim, working with the global settings, dashboard, sampling options, errors, and debugging. And then the next one will be accessing the results for the week of July 13th. And we'll talk about, you know, navigating the results menus, et cetera, showed us some of the default results and as well as getting other information out of the code. These are going to be great opportunities to get involved more with learning about the code and using it. I definitely highly encourage you to participate and we'll have some announcements on those coming out after this meeting. Okay. We had slated some time for, at the end here, for questions and answers. And fortunately we are running out of time. However, we have been swanning to I think as many questions as we have capacity for in the chat here. As mentioned, we will take a copy of the chat and we will respond to the questions in the NRC meeting summary that's put out. So with that I'm going to move on to deliver a closed meeting. What we've done today is presented you with XLPR version 2 code as a state-of-the-art tool for piping integrity risk assessment. And a code that's developed and vetted by leading experts from government and industry. It's built using modern software design concept under a robust quality assurance program. And it's also a code that is transparent and flexible. We are very much looking forward to having you become part of the XLPR community. Please mark your calendars for that May 28th release date and also look for the future announcements like I said on the additional training webinars. Greg, did you have anything else you wanted to add here at the end? No, Matt. I think we've covered plenty of information for today. Okay, great. Thank you. I want to thank everyone for their time today and joining us. Hopefully everyone and their families are staying safe and healthy. After the meeting ends, whoever wants to provide feedback, welcome to do so. We'd like to understand your views about this meeting and potentially improve future NRC meetings as well. If you'd care to provide feedback, go to nrc.gov, click on the public meeting schedule. I showed recently how meetings, find this meeting, and then you can click and it'll be a feedback point. Just play the electronic thing that you can fill out and submit. Just a reminder, again, we'll be issuing a public meeting summary for this. We'll be making a video recording of this in terms of available as well. And finally, we have set up xlpr at nrc.gov and xlpr at epri.com, so you can communicate with us. Greg and I monitor those email addresses. If you have any questions at all, you can go ahead and use those resources to reach out to us. Once again, I want to thank everybody and I will look forward to seeing you and interacting with you more in some of our future events. Thank you very much.