Looking for Labs that Don’t Lie: Validating Lab Data for Fracture Toughness Testing with David Futch

{Intro} 0:03 Rhett On this episode of Pipeline Things, we find out what Chris's cholesterol and challenges with laboratory test data have in common. Believe it or not, both can lead you to a sense of panic at the end of the day. It's an episode with a return guest, and I think a lot of great content on the subject of the validity of fracture toughness tests. Thanks for joining us. 0:41 Rhett Boom. All right. Welcome to today's edition of the podcast. We continue our PPIM series with a return guest. But before I introduce that guest, I mean, Chris, sometimes it feels like we go back to the well we know for guests. But in this case, I really think the well water was so good. It's like you just want to keep drinking. 1:01 Chris That or we like the data that it gives us, that source of data. It's like we understand what it's going to give us. We can align with what it's going to give us. And so we can deal with the data it's going to give us. 1:08 Rhett Right, right, right. So, before I introduce him, though, which I am excited about him being back, Chris, it was like, what maybe two years ago you had a health scare. When I was thinking about the segue for this podcast, I was thinking about your health scare from like. Was it a year or two years ago? Do you remember how old it was? 1:22 Chris It was two and a half years ago. I fully remember because you know me. I like to run. I like to be somewhat fit. You make fun of me because I do juicing and masticated juices and all that stuff. I also like to drink. I'm an engineer. We like data. I had a little bit of a scare. I was like, what? I wasn't expecting this. 1:42 Rhett And so, elaborate a little bit, right? Would you get like a high cholesterol, or something is what I'm recalling? 1:53 Chris I've always been borderline cholesterol, like low 200 since I was like 18. I graduated high school and my mom was like; you need to start doing checkups. And so, I've always known, but I never really knew what that meant. Then my granddad had a heart attack, and I was like, oh, I need to pay attention to this. Scare mode. And so, they were like, run more, eat fiber. So, I run every day and all this fun stuff. And so, then I got a new PCP, primary care physician, and they were like, Chris, you're getting older. You've had this forever. You've probably got buildup. We need to start mitigating this. I'm going to start using our terminology. 2:20 Chris And they're like, you need to get on cholesterol medicine. And they're like, it's low dose. It's everyday. It's easy. And I was like, I don't know if I want to do this, so I do it. 2:24 Rhett Data said it, man. You had high cholesterol. Get on that medicine. 2:28 Chris And we had a track record of it, right? Time to medicate. 2:31 Chris And so here I go. And I failed because I should have done what the next doctor did, right? And so I'm on that for a while. And then I go for annual checkup, and then my liver enzymes are high. and they're like really high and she couldn't explain what it was like we're talking like 200 liver enzymes are supposed to be like around the 20s to 40s so I go to GI and the GI is like okay why are you on cholesterol medicine i was like oh because I've had borderline cholesterol for like 20 years he's like not a reason to be on it second one he's like did you work out the day before. 3:00 Chris I was like bro I work out almost every day he's like that's why your liver enzymes pop he goes when you run and you get all the lactic acid going around all this fun stuff your liver enzymes shoot up he goes and so he took me off the medicine Six months later, my liver enzymes are back to normal. Sends me to go get a calcium score on my heart. Zero. No calcium in my heart. None of my heart's blocked. None of my veins are blocked. No reason to have the medicine. And he's like, get off all the medicine. Total unnecessary health scare. 3:23 Chris For 12 months, I was pulling my hair out of my head. I was stressed out for nothing. Paying for medicine. Going to the wrong place. Not asking enough questions. 3:30 Rhett Preparing your will. 3:32 Chris It was. You remember all that. 3:37 Rhett I do, man. 3:39 Chris I prepared my will. That's right. 3:41 Rhett Oh, gosh. So, audience, that's a morbid beginning to the show. But the truth is, the reason it ties in so well to today's show is that I don't think sometimes we question enough of the data that we get back from labs. And today's episode is a perfect play on that. from, again, a return guest. 4:00 Rhett So I'm going to bring on Mr. David Futch. You guys have seen him before. We don't have a clever name for him like the dirt merchant. But he's joining us to talk about the PPIM paper that he wrote this year dealing with fracture toughness. Now, David, the audience is already familiar with you. I feel like you need no introduction, right? But is there anything you'd like to say to the audience on your return to the podcasting world? 4:26 David Appreciate you having me back. Oh. The ratings may plummet after this, but we'll see. 4:31 Rhett You know, we accept that risk. Well, I mean, you kind of have a golden ticket when you name your firstborn child after one of the hosts. Allegedly. Let the record stand that David's son is named for me. 4:42 Rhett So, David, now this paper, again, for the audience, this was a paper that you wrote in conjunction with MMT. But we're not here to talk about the MMT blade toughness meter, right? So, for audience, if you go pull that PPIM paper that David Futch is on, the driver was MMT's BMT technology. But the reason you got pulled in was a little bit different, right? So if you don't mind, can you share some of the background on that? Like, how did you get pulled into this publication? Where does Futch fit in? 5:15 David Yeah, so MMT developed a tool or is developing a tool, BTM. 5:18 Rhett Which for the audience, that's the blade toughness meter. 5:21 David Exactly. Yep. So obviously, anytime you develop something new, it needs to be rooted in data, right? You need data to show how well that tool is performing. And that's where we came in. We came in to, one, provide data, test and provide data, and two, to collect and analyze and provide data when needed. So that way, their tool could properly demonstrate performance. you know, blind and unblind. So we were sort of the third party on the outside looking in, providing data when they needed it to help the tool kind of get pushed along. 5:52 Rhett Absolutely. And so, in this particular case, as I understand it, just so the audience appreciates the... you're familiar with MMT's technology, you're probably familiar with, again, there's an episode you can go look where we had Simone Belmar on. And they focused on the development of yield and toughness from a field deployable unit. If you've been watching the industry, they've been developing a new technology. It's to quantify toughness that works on a different parameter. 6:15 Rhett And as part of that, like Fletch is describing, they were benchmarking. their blade toughness meter against a set of test data where they had toughness data 6:27 Chris Hold on and to be clear the test data came from a lab correct which means it has to be accurate 100 because that's what we do we send. 6:33 David it means it was tested at a lab 6:40 Rhett And so the reason you got pulled in is because the blade toughness meter the BTM wasn't agreeing with the lab toughness data correct? 6:48 David That's correct 6:50 Rhett all right yes so Futch, I'm going to turn it over to you. You came in, and what did you find when you started looking into the lab toughness data? 6:58 David Yeah, so the outputs of the BTM model was predicting essentially one slope. The outputs, if you looked at the data from the lab, was on a completely different slope. So that either means two things. One, the tool is completely incorrectly adjusting or quantifying whatever parameter they're trying to get to, right? Or that one of the two data sets is somewhat either partially or fully incorrect. So the logical place to start in that case was to review the test reports to try to figure out if there was a sort of smoking gun, right? 7:34 David Something that we can point to to say, hey, this number is slightly off. And that's then displaying a parameter that's shifting our slope. And then therefore, the two data sets are disagreeing with each other. 7:45 Rhett When you refer to these lab testing reports, right? So this is a standard CTOD report that operators would get if they ran a CTOD test anywhere, right? 7:49 David A J integral. 7:52 Rhett j integral test, yeah, thank you and that j integral test has uh the j value and then it has some other information that's presented on that report as well, right? 8:02 David It's a series of loads and unloads, right? And then that forms a curve. And then from that curve, there's a bunch of math that goes into determining validity, which we'll leave out of the conversation, but a bunch of math that goes into validity and also goes into how they calculate the K’s and the J’s and the things like that. 8:22 David And those then are often based upon or are based upon the material properties of whatever you're testing, your yield, your tensile combined to get a flow stress and a modulus. 8:31 Chris Right. So, is it easy to mess those up, those tests for J integrals? 8:38 David The yield and tensile and modulus? Yeah. Yeah, so those would come from a standard tensile test, right? They can be done wrong. because they're relying on a frame and an exonsometer. If the exonsometer is placed wrong, maybe not calibrated or slips, you would sort of see saw-toothing in the modulus of the upramp, which could give you an incorrect number. 8:57 David But the modulus for steel is a quantifiable, a known number. And so typically labs, when your modulus doesn't meet or doesn't fit, they would retest or go through some quality control procedure to take that data out. 9:14 Rhett Right. You jumped ahead a little bit on this one, but in your first review of the data, like what was one of the first things that you found when you took a look at the data? So obviously the setting is the BTM is not agreeing with the test data. Enter Futch. Futch reviews test data. What does Futch find in the review of the test data? 9:32 David So the modulus in those reports were low, right? 9:36 Rhett How low? Like you said normally still is 29.5. Thousand or 29,500 KSI or 30,000 KSI. 9:40 David MSI. 9:45 Rhett MSI. Well, okay. Thank you. Did you correct the host? This is the last time he's on this show. 9:52 Chris Extra points for him. Continue. 9:55 David Just units. All right. 9:57 Rhett That's 29 or 30, whatever units. Yeah. 10:00 David Yeah. So, in some cases it was half. It was a range. So different reports. So different steals or different pieces of pipe tested. had a different modulus number. So some were off by half, some were off by a little bit, some were off by more. 10:10 Rhett And just so we're clear, off by half, steel does not, that would be 15. Correct. 15 million PSI. Steel never has a modulus of 15 million PSI. Right. So what influence does that have on the final J curve? 10:25 David Yeah. So then the calculations, the math to get to the K or J would then be off. uh how much it's off depends on lots of other variables right but the modulus is an input to get there and if one of your inputs off then your output is logically off as well 10:43 Rhett Right which is your garbage in typical garbage out motto right correct so um in your experience i'm just curious this actually now because i'm curious i feel like the no question that most people ask when they get the j integral test back is what's my j how often do you see people diving in and reviewing the maybe other variables that impact that such as modulus? 11:04 David Usually people will review the validity criteria. There isn't a validity criteria for what's your modulus and is it in this range, right? Because it's an ASTM spec. So an ASTM spec can be applied to a bunch of different materials, right? But if enough of your validity is out, it could signal that there's an input issue or it could signal that the test was just bad. Right? I mean, there's two options there. But yeah, a lot of people will typically review the validity. In ASTM, there's a list of criteria that has to be, quote unquote, valid. And then the test itself would be qualified. And Modus isn't one of those, but Modus certainly feeds into several of those calculations. 11:49 Rhett So it's possible for the modulus to be 15 and have been input either incorrectly or calculated incorrectly, but not show up in the validity requirements. 11:57 David Correct. 12:01 Rhett You found that was your first error. What was the second issue that you found with the material test data? 12:09 David Was there a second one? 12:11 Chris Yeah, that's what I was like. I don't remember what the second one was. 12:13 Rhett There was a second one. It was the validity. I was going with the validity. 12:16 David Oh, yeah. Validity. Well, the... Yeah, so the validity is based on... 12:22 Rhett Just for the audience, I want y'all to know we do this live. 12:25 David Yeah, exactly. I didn't know where you were going with the second one. They used the wrong colors in the background of their report. I just didn't like them. No. 12:32 What are the ASTM requirements for colors? 12:35 David So ASTM 1820 has a list of, well, one, it prescribes how you do the test. Sample geometries, ratios of dimensions, things like that. And it also has then how you're supposed to load and unload the sample. So basically any lab could pick it up, read it. If you have the right equipment, theoretically you could comply with it, right? Once you get to the point of processing the data and looking at the numbers, they then have a list of validity criteria that you want to be above a certain range, like the number of unloads. You want to be above a certain number. Or your pre-crack or your straightness of your crack is between X and Y, between two numbers, some percentage of your dimension. 13:20 David And so those validity criteria's are all listed out in 1820. But 1820 was meant for very thick wall things, testing nuclear applications with really thick material, where in this case, we're testing pipeline steels. In some cases, could be as thin as 156 or 188 inch. And so not exactly apples to apples, which typically causes you to look at the validity criteria harder and consider what that might do going from something that was thick to something that's thin. 13:50 Rhett So what is thick in your opinion? Where does thick start when you start talking about J-Integral test? 13:59 David It depends on how long you make the sample. Okay. But probably an inch. 14:05 Rhett Wow. So that's, I mean, let's be real. That's substantially thicker than most of the pipeline material, right? 14:09 David Onshore. 14:12 Rhett Onshore, for sure. Yeah. And the influence of this, as the audience understands, right, is that the thicker the material, it prevents the crack from more rapidly tearing, or it forces the crack. Too more rapidly tear, I should say, right? Like you're more likely to get a true plane strain condition. Whenever you get a thinner material, you don't get that plane strain condition that the test was targeting. 14:35 Rhett And do you know what influence does that normally have? Does it typically make the numbers too high or too low? 14:40 Depends on where the curve you're looking, but typically your J or K would be overcompensated. It would be higher. 14:50 Rhett All right. So taken together, so you've got... data input issues you've got validity test issues what did that mean overall how do y'all compensate for this or what were your your recommendations for this what do you do. 15:00 David Yeah so back to the development of the BTM right, that that data set can't be used with great confidence right if you're trying to compare to a known value and the value has question marks then maybe you can't use that and so they went down the route of collecting additional data to sort of supplement into the model yeah 15:20 So making a tie to all the operators out there, I mean, the tie is clear, right? In this case, this prevented the proper validation of a technology. But this has much broader implications if you're using the J-intergral grader maybe to evaluate burst pressures for EMAT. Or any crack type. Any crack technology. And I think what's shocking to me, David, is that I don't know that many people are going as deep as looking at the modulus and querying. We usually just get the data. And use the data yeah right so 1553 Chris Yeah when we work together i remember making this an issue um that was when a lot we would start seeing a lot of testing for MPV right material property verification and people were getting a lot of tensile tests and they were cutting samples out and sending labs for testing and where they get toughness values that were really low in bond lines. And I remember talking to operators and saying, I was like, okay, well, what's your lab test qualification procedure? And what is your lab test results review procedure look like? And the first question, sometimes I get a response, but the second question, never. It was just like, we sent it to the lab. What do you mean? And I'm like, well, who reviews report? Like when you get any test, you do it. Like when you do an ILI, you check for where the essential variables met. 16:35 Chris You know, what did the onsite report look like? Did you meet, you know, mag levels or whatever? I was like, what'd you do for your material testing? And usually, it's kind of silent. And this is one of those examples where it could have had some pretty big implications on the BMT. 16:48 Rhett I want to leave a little bit of a cliffhanger for the audience before we go on break, which is before I come back. Some of you might be thinking like, oh, this is a guy who does lab testing, wants to make a big deal out of lab testing. I want to ask you what you think the error was in these tests. 10%, 20%, 30%, 50%. We come back, think, how much do you think these errors with modulus, these issues with plane strain had an impact on what the actual toughness value is? And we'll give you that answer when we come back. Break 17:30 Rhett All right, audience, welcome back. When we come back, you know, I told you guys I was going to brace y'all for what the variation of the data was, and I asked you to think about what that error was, 10, 15, 20 percent. 17:38 Chris But we all know what they were thinking. They were wondering why your two index fingers have green fingernail paint on them. That's what everybody's thinking. 17:44 Rhett True story. Shout out to my daughter, Peyton. Yes. It was her 11th birthday. 17:48 Chris That stuff won’t come off, will it? 17:49 Dude, she is knee deep into nails. And for her 11th birthday, she got a whole bunch of nail stuff. I told her she could paint two of dad's nails. And I tried to get D2I green. I don't know how close it was to D2I green. 18:01 Chris Close enough. 18:03 Rhett But yes, my 11-year-old gave me a full two-finger. Is this a pedicure? No, manicure. I don't know what this is. She gave me a two-finger nail job. That's what she gave me. And yeah, so shout out to my 11-year-old. So I leave it on because every time I see my green nails, it makes me laugh and think of my 11-year-old daughter. 18:20 Chris It's a good time. It reduces stress. 18:24 And it makes people ask questions when they see me on the camera. They're like, I can see them tracking my nail, and they want to ask, but they don't know if they can ask. And then they're like, well, I thought you maybe damaged your nail, so I didn't want to ask. It's really funny the reaction it gets from people. 18:39 Chris My kids are not doing that yet. They're all doing this. My two daughters are doing the stick-ons until they bought some fancy ones at a farmer's market, and they wouldn't come off, and they got really scared. They're like, Dad, these aren't coming off. They're not coming off. 18:49 Rhett Oh, there's removers for that. Yeah, we have all that. She's told me that she has some remover that will take it off. haven't decided I'm going to do it yet. 18:55 Chris Girl dad, sorry guys. 18:57 Rhett No, I got three of them, man. I'm prepped and ready to go. But this is my youngest one bringing that in. She's actually trying to charge her sisters for doing nails, which is really comical. And they're refusing. She started offering discounts. So, she's on her way to entrepreneurialism. 19:10 Rhett But okay, going back. You were thinking, audience, still waiting to ask that question. You know, Chris, I need to give you some time because I didn't know if your cholesterol would spike and we'd need to like do CPR on you in the middle of the show. 19:18 Chris No, we're good. 19:20 Rhett So the errors. In when upon retests in this publication were in many cases two to three times higher so I want you to think about that two to 19:32 Chris The toughness values right so the j integral values the result in j integral values 19:36 Rhett The one that was most shocking original test which the BTM was trying to compare against 69.4 ksi root inch retest 208 ksi root inch Right. Think about the implications of that if you're a pipeline operator and you're assessing a crack. Right. That's the difference between a crack that you expect to behave in a brittle manner and a crack that you completely expect to behave on the ductal side of the spectrum. Right. So these errors that we were talking with Futch about, they're potentially very consequential. 20:06 Chris And that difference wasn't just a Futch factor. Right. You didn't just go in there. 20:10 Rhett Yeah. You didn't even retest. Right. You just scratched it out and wrote 208 beside it. Is that, you know. 20:15 Chris Let's just recalculate. There's no testing required. We're just going to use the right modulus, right? 20:26 David Yeah. We tried that. So that was one attempt was can we use the calculations in the data to back into what it could have been? But there's other variables in there that it's not as easy as just, you know, playing math, right? 20:36 Rhett So a lot of these had to be retested. I want to know, practically speaking, I mean, look. The message that I'm hoping people on this call here is that just because you got it tested doesn't mean it's right. Just because you got a lab test report that says your toughness is 56 KSI root inch or 80 KSI root inch or 120 KSI root inch doesn't necessarily mean that that's right. What advice, guidance would you offer? If I get a lab test report, what should I do first? 21:11 David Yeah, so before you even do that, I would even go this. you know, before you trust a lab to do your testing, you probably want to go see the lab, right? You want to go validate. 21:20 Chris Field trip, 21:22 David Yeah. You want to go validate that they're operating to some sort of procedure, right? It's just not some guy in the back just doing whatever he wants. Some sort of procedure. They understand what you're trying to go for. For instance, if you're trying to hit an ERW bond line and they don't know what ERW is, are they going to hit the bond line with confidence? 21:40 Chris Or asking how they’re going to find it. 21:41 David Bingo. 21:45 Chris Did you etch that before you put the notch in there? Stuff like that, right? 21:47 You just rub your finger and find the scarf and call it good, right? But that's where I would start first, is have some sort of confidence in the lab that you've selected to partner with, send stuff to. Once you do that, then you send things. 22:02 Rhett So that means the lab cannot be in your garage, Chris, just so we understand. 22:06 Chris Yeah, I know. Hey, stop offing leading indicators. So the next question I have is like, but doesn't that make those customers difficult? You don't want to work with them, right? When they come and ask questions and say, hey, how do you do this? What standard are you using? Isn't that making them complicated and you prefer to go with the guys that don't ask questions? 22:23 David But it also helps the lab become better. 22:25 Chris That is true 22:28 David Because if you've been in a lab, labs typically do things one way because that's how the people that work there know how to do it. When you bring in outside opinions, that lab gets better. 22:38 Chris How do you feel? Is it only the operator or can consultants do that too? Are they viewed differently? No. We were recently on a project where we were asked to say, hey, we're going to go get some toughness data. I was like, okay, what are you going to do? And they were like, we need to do the bond line. And they're like, okay, what are you going to do? They're like, we're going to do CVNs. And I was like, how do you know they're going to nail the bond line? And they're like, well, because that's what the lab is responsible for doing. And I said, yeah, but you can ask questions. I was like, ask them how they're going to land it. And so they call me back and they go. Chris, they just said that they're really good at this. And I said, okay, that's awesome. I was like, ask them if they'll etch it and show you where the notch is relative to the bond line. They go, that's a great question. So then they called me back and said, they said that that's not really what they do. And I said, well, did they agree to it? He goes, no, they just said that's normally not what they do. And I was like, so what are you going to do? And they're like, I think we're going to ask for it. I was like, that sounds great. 23:30 David Yeah. So, you know, when you're typically, like a Sharpie, you're validating your notch, the curvature of the notch, the size of the notch, the spacing from the ends and all of that. Often the labs will also verify, not everyone, but some out of the batch to confirm that they hit a bond line. So typically with that, you would etch, you would scribe or mark your bond line, then you would notch, and then you verify that the notch is within certain distance of the bond line. It's not 0.00, right? You're not spot on it. There's a little bit of a top. there uh but yeah that's uh typically not done right where it's in a report but if it that's what's requested that's what you do yeah 24:08 Rhett so going back step one audit the lab you recommend it they visit all right so if we assume we did that check the box this lab looks legit they have solid procedures the floors were swept and clean they bought good lunch yeah there we go exactly um what's the next step when you get that report 24:24 David Yeah, so, then I would start looking, if we're only talking about fracture toughness or J's, I would start looking at the report itself. Look at the variables that went in, the yields, the tensils, the modulus, those things. Is it consistent with what the pipe you thought it was? If you think it was X60 and they're reporting it as X52 or X42, Maybe the pipe was low strength. That's a possibility. Or maybe there's some input that's wrong. So sort of checking the input parameters. 24:44 David And then I would probably look at the fracture face itself. Does the crack look stable? Are there other things interacting with it? What does it just generally look like? And then I would go back and probably look at the validity or the graph. Does the graph look clean? That obviously plays into your validity because you're not going to get your points if it's a whole bunch of pop-ins and things like that. Just the visual cues of the graph and the visual cues of the validity. Are you within a reasonable amount or does it look terrible? 25:25 Rhett And the validity sounds like it might be the most challenging one to get into, right? Because you've already alluded to some tests aren't going to make it within the validity requirements. Yeah, especially the really thin ones. The really thin ones, you're dealing with something that's very thin in itself. you typically side groove, so you're removing more of the wall thickness. And all of these criterias are percentages on your geometry or your thickness, right? And so if you're dealing with something that's 150 thick, 10% of that is a very small number. When you're dealing with something that's one inch thick, 10% is a much larger number. So your margin for error is much larger when the stuff is thicker, right? 26:05 David So with the thinner samples, you just have less margin. So that's where the... engineering thought goes in is what did my picture look like? What did my graph look like? And how far out am I actually am? If I'm really far out of that validity, maybe it's a bad test. 26:20 Chris Is it, I mean, it makes me, I think of a couple of things, David, and it's should an operator expect that if you go from one lab to another, that you're going to get this spread of two to three times different toughness values? because of all the complexities and i'll set the stage right i mean in previous talks we've talked about full scale like full size full sample sizes you know half sample size whatever it may be the orientation of the test was it flattened to get the test in. Now we're talking about what modulus was used these are a lot of to me it sounds like a lot of parameters that could change and if one lab does it one way another lab does it another way I mean, give us something to point to, to where it's like, what is an operator? How do you begin to deal with this? I guess is my point. Just use one lab. That way you have the consistent bias across the board. What do you do? 27:10 David From my perspective, an operator should always have multiple labs. 27:15 Rhett You might regret that statement. 27:20 David I run a lab and I'm out here saying it, but you always have to have multiple options. Yes, it's true. Anything you do, you've got to have multiple options. You have to have a bench that you can call on when things go south and you need a lot of work done. You always have to have the bench. With that being said, I think that's where you go visit them and you confirm that these two or three labs are all operating generally the same. If you wanted to go the route, you could always send them samples of the same one and sort of blind them against each other to see which one's performing. But if a lab's testing in the same general realm of things, right, right inputs, similar sample size, similar processing, you know, not flattening it, flattening, whatever they have to do, that the results should be within a few percentages. 28:02 Chris What is the lab qualification standard in the industry right now? most operators could point to and say, before I go and do a J-Integral or a CTOD with these guys, I want to ask them about how they got qualified through fill in the blank. Yeah. So, the easy one is show me your qualifications or your certifications, your calibrations on your equipment. 28:28 Rhett That's a given though. That should be given. 28:29 David Is the clip gauge calibrated? Is the machine calibrated? All of that. And then the next one is, do you have some sort of quality manual? Whether the quality manual is standalone for the lab or the quality manual is taken one step further into a governing body like 9001 or 17025 or something like that. But if they haven't gone to that route, do they have a quality manual that they practice, that they follow? And if they're following that, then that's something you review and you audit and then... When things go wrong or things are south, you can always check to see where it went south in the quality manual. 29:05 Rhett So, I think the last thought that I had is you mentioned, are we only talking about J data or toughness data? Is there other data that you would recommend to take a look at when you're trying to qualify whether or not, in this case, a toughness test was done correctly? 29:27 Chris Well, actually, I think it was the tensile test. That was done incorrectly on this one that led to incorrect data for the J calculations. 29:31 David In the most simplest form, a tensile typically is done in the hoop direction if your diameter is greater than six, six inches, right? So, eight and above, or it's actually seven, but there's really no seven inch pipeline. So eight and above, right? You're doing a transverse tensile. To get that tensile, you have to flatten. So the easiest thing to check is, do you have a procedure you flatten with? Or do you run it over with the forklift? Or do you just press in the middle until you get it generally flat? Those kind of questions. Typically, labs will have a procedure on how they flatten. 30:02 David That was the study of a lot of work maybe 20 years ago where a research institute went out and figured out how's the best way to flatten so we have consistent results. 30:12 Rhett Was it with a forklift? 30:15 David Surprise? No, it was not. You want to step in from the ends. 30:18 Chris Warm it up and then bend it. 30:22 David Yeah, exactly. Yeah, you usually step it in from the end, so the work is all at the tabs where you're gripping with the machine versus the middle where you're trying to take a reading. 30:30 Rhett Excellent. Well, where I was actually going with that is because the paper had some stuff in there was some stuff about Sharpies, right? Is there any benefit? Can you use Sharpies to any advantage if you have Sharpies and then you go back and get toughness data for an indication of whether or not your toughness data was good? 30:47 David Yeah. So, if you do have Sharpies, you can convert the Sharpies to a K and then you can say, hey, I'm getting this K here and I'm getting that K with a converted Sharpie. Are they generally in the same ballpark? 30:57 But same ballpark is wide with Sharpies. We know that, right? Like this isn't, hey, I converted my Sharpie and I got 50 KSI root inch and the test got 70. This is, you're looking for pretty big gaps if you do that, I assume. 31:09 Chris Like 60 to 200. 60 to 200, 31:13 David Yeah. Going what you said, 50 to 70 is probably within the realm of reasonable because the conversion to go from Sharpie to a K is conservative in itself. So instead of converting with a math equation, you're actually going straight to the number. So I expect that one to be higher. 31:30 Excellent. Well, Futch, thanks for joining us again. Do you have any parting words for the audience? 31:35 David I do not. I didn't expect that question to come. 31:40 Chris Hire more than one lab. 31:42 David Yeah, have more than one lab and make a call. 31:45 Rhett Don't have just one lab for sure. So, I want to say thanks for joining us. Audience, hopefully you enjoyed this conversation. I mean, I know for me, it's one of those things where a lot of times we just get data and we don't question that data enough. But just like in the case of Chris and thinking he's dying, you might think your pipe's dying in a reality. not in that case time to get a second opinion in walks fudge thanks for joining us we look forward to joining you again as we continue our conversation on PPIM papers and thanks for joining us on pipeline things. {Outro} 32:09 this episode was executively produced by Sarah Etier. Thanks to my lovely 11-year-old daughter for doing my fingernails and our return guest David Futch Thanks also to FORME marketing for the use of their location and filming this podcast, deuces.