母女情深!娜扎录节目妈妈出镜 对女儿又夸又抱

(soft techno music) - [Narrator] We're making great progress along the digital thread after all, we've finally have a printed part. But our work's not done yet because now we move into post-processing. For that, we'll travel across Youngstown to M7 Technologies. Before we get into that, we check in with scientist Fred Percy and M7 President, Michael Garvey about the plan and the results of our institute monitoring. - We got a part! - We do? - All right. - Let me show you some of the data. - Yeah. - We have the laser sensor in the build file. This is just a tiny little swath of the part. It's just one layer, one piece of one layer. The blue is the powder bed and the colorization shows you the height of what's above the powder bed. The red area here, the yellow and green, that's where the laser came by and actually welded the material. So that's a little bit higher than the powder bed. And that's what we're trying to do. We're trying to build this layer, layer upon layer. So this layer, if you look at the color mapping here on the right, it's only about 150 microns high. And what we're trying to determine, here is everything look good. Are there any spots that are missing? Are there any extra? But you can see right out here to the right there's this strange little anomaly, a little spike that's sticking up. That's not supposed to be there. And you obviously you look on the part you know, there shouldn't be anything over here. - Yep. - So, what is that? It's probably, it could be a little speck, it could be a little piece of metal that washed away a little bit. We know it's not important for us, it's not there on subsequent layers. - So in general, this is your clue to go look either during the build process to check-- - Correct. - To see if it continues to propagate through the part or maybe something that we want to check later during inspection? - Yes. - Okay. - Yes, definitely. - This thing has something like 1,626 layers to it and you're just pointing at, in this picture, you're just pointing at a small fraction of one layer. - One single layer from this sensor is millions of points. So you multiply that by your 1700 layers-- - You'd be talking about billions of points. - You're talking about billions of data points. Collecting all this data, you are talking Terabytes of data, potentially, per build. That's a lot of data. - Having a strategy for how we collect that data, where and how we process that data, how we store that data, what we store, how long we store it could all be-- - The accuracy of the data. - The accuracy of that data could all be really important dimensions of the problem. Is that a fair? - That's very fair. I think where we're going to see is the research that's going to be done identify exactly what needs to be monitored, then the trick is going to be how to capture that and distill it in real-time down to a more manageable size that can follow along with the part. - Interesting time, interesting times. - Very interesting. - Mike, where do we go now? I mean, this part, you know, great lookin' part, kind of cool lookin' but not what we need in order to have a functional part. So what's going to happen next year? - So, we're going to take this part that's on the build plate, we're going to take it into the workshop and we're going to cut the part off the build plate, and then we will mount this part off the build plate, as built, onto a machining center and we will do an initial scan of the part to determine the stock that is on the part and to make sure that we have plenty of stock in the areas that are required for finish. Then we will machine the part. And from machining the part, we will then scan the part and make sure that it conforms. (cheerful music) (machine whirring) - Okay, so we got a part. It's done. It looks like we expected it to look. What happens now? - What we want to do now, is verify that. - Verify the part and check that it actually dimensionally is what it's supposed to be. You know what? Why don't we do that in the Metrology Lab, it's a little noisy out here. We have this, it's on the network now, so we can open it up in the other room, and we'll take a look at that. So this is our Metrology Lab. This is where we make all the measurements, qualified parts, things like that. What's you're lookin' at here this very large machine is a C.M.M. machine, a Coordinate Measuring Machine. And this machine is used to make precision measurements on parts. This part in particular, because it's been topologically optimized, it doesn't really do well with a machine like this. This is made for very simple round surfaces, or flat surfaces. It would have a very hard time measuring the detail of these very sort of oddly-shaped struts in here. So, we've scanned it. We have a much higher density point cloud and we're going to head over to the workstation behind you and we're going to look at that point cloud. - So, we're going to access that data that we had on the machine, that we took on the machine, we're going to bump that up, as built, we're going to bump that up against as designed. - Exactly. - To make sure it fits. Can we take a look? - Yep, let's take a look. I first want to show you the 3-D model of the optimized part. This is going to show the design intent, exactly what this part should look like, theoretically. Now what I want to show you is the RAW point cloud data that we just took, just a few minutes ago, off that machine, the C.N.C. machine. - Okay. - This is the RAW data point, so XYZ, three-dimensional coordinates for all of these points. So that's what we have, here. So, first thing that we would do, naively, is to sort of look at them together. So, let me turn the other one one and you don't really see much, right? - They're basically-- - It looks like the model. - It looks like the model. They're basically on top of each other. For each point, so each point in the point cloud, how far is that away from the actual 3-D model, the design intent model? - Okay. - And what we'll do then, is we'll colorize this model based on what that deviation is. So I will turn that deviation on, right now. Activate it. And this is what we get. So what we're looking at, here, to explain, is the different colors show you how far that area is away from the actual design intent. So, how close are we? And if you look, the color is scaled over here on the right. It tells you red, it'll be really bad, purple would be bad in the opposite direction. Good data is kind of here in the middle of this green and blue. - Okay, which is mostly what we're lookin' at. - It's mostly what we're seeing, here. I can sort of visually rotate this around, scan this around. You know, you see in here, you got the green and the blue. But that's just, you know, green is a little bit positive, blue is a little bit negative. That's all good. - It strikes me that one of the big advantages that we've got here is as described, we've got this as-designed part, this perfect part in our minds and then we've got this as-built or as-manufactured part and through the digital thread, we're going to essentially be able to wed those things kind of like what we've done, here and have them reside together forever more. Is that a fair characterization in your mind? - Yeah, it's definitely a fair characterization. What we want to do is serialize each part. Give each part a serial number. - So, especially on an airplane, right? This will have a unique - Yeah, definitely. - Part number I want to know? - A part number and a serial number, too. So then you want to then keep this information around for the life of that part. - [Narrator] We've made good progress. We've gone from the original belt crank to this new topology-optimized part. It's been tested and validated all along the digital thread. But how does this all come together to provide real business value? For that answer, we'll had back to Albuquerque. (soft techno music)