#127 Interview with Alexander Chaloupka of sensXPERT: New AI-generated information out of measurement data from manufacturing
13.06.2024 40 min Staffel 5 Episode 84
Zusammenfassung & Show Notes
Ilkay Özkisaoglu, Co-Founder of Composites Lounge asks Dr. Alexander Chaloupka, CTO of sensXPERT - Optimizing Plastics Manufacturing what the moment was that it became clear to him that material characterization could be taken a step further?
Alexander, during his tenure at Fraunhofer, was diving deep into material characterization with different equipment and different technologies. He says "... I was looking into different parameters of material".
This is when he created a vision that there must be a way of creating the same information within the manufacturing, based on different sensor technologies, to really bridge the gap and to bring the two worlds together.
Listen in this episode what challenges he had to overcome with the mindset of many experts "... based on the results I have generated at the same time, I could see there is no match between what the experts have told me that it's not possible and the experience I have gained and the knowledge I have gained so far, and that drove me forward"
Which is why he paved his own way.
You will hear more about "specific algorithms to generate new information out of measuring or measurement data from the manufacturing" and "...that was when it was clear to me that I shouldn't just listen to the experts that are there that have done things so many years."
Join this truly inspiring podcast episode, where many of you, working in the engineering space will surely relate to.
Also, please listen to the product that was presented during JEC World 2024 by Alexander. We thank the JEC Group for the stage recording.
Any questions?
Let us or Alexander know!
YouTube episode: https://youtu.be/y0eH7jQ8mtc
Alexander, during his tenure at Fraunhofer, was diving deep into material characterization with different equipment and different technologies. He says "... I was looking into different parameters of material".
This is when he created a vision that there must be a way of creating the same information within the manufacturing, based on different sensor technologies, to really bridge the gap and to bring the two worlds together.
Listen in this episode what challenges he had to overcome with the mindset of many experts "... based on the results I have generated at the same time, I could see there is no match between what the experts have told me that it's not possible and the experience I have gained and the knowledge I have gained so far, and that drove me forward"
Which is why he paved his own way.
You will hear more about "specific algorithms to generate new information out of measuring or measurement data from the manufacturing" and "...that was when it was clear to me that I shouldn't just listen to the experts that are there that have done things so many years."
Join this truly inspiring podcast episode, where many of you, working in the engineering space will surely relate to.
Also, please listen to the product that was presented during JEC World 2024 by Alexander. We thank the JEC Group for the stage recording.
Any questions?
Let us or Alexander know!
YouTube episode: https://youtu.be/y0eH7jQ8mtc
Transkript
Ilkay Özkisaoglu, Composites Lounge: So wonderful good afternoon, dear LinkedIn community and composites experts out there. And this is now our interview session here at sensXPERT, a deep tech company out of the German Bavarian culture. And we are looking to process optimization together with Alexander Chaloupka. And Alexander is already a long term employee, and we will dive with him deep into the matter. First of all, Alexander, tell us shortly about your personal background and how did you come to sensXPERT to this position?
Alexander Chaloupka, CTO sensXPERT: So I'm physicist from education. So I started in Augsburg at that time the Fraunhofer ICT functional lightweight group. So the project group of the Fraunhofer ICT was located in Augsburg. And at the end of my studies, I joined directly into the Fraunhofer project group. And I was working with composites and plastics in general, was working on transferring laboratory data into the manufacturing. And yeah, that was my target at the end to bridge the gap in between the two worlds that belong together. But in general, they are separated from each other because we have no chance to really transfer laboratory data, material cards, data sheet data into the real production environment because there are so many deviations. And that also brought me in contact with NETZSCH Analyzing & Testing at that time. So I created a vision that there must be a way of gaining material characterization data in the manufacturing environment to deal with all the deviations and to create more robust processes. And that time NETZSCH Analyzing & Testing formed a future strategy that was called beyond thermal analysis. So they wanted to create something new out of the lab using the knowledge they already had. And yeah, that was the time I got in contact with NETZSCH Analyzing & Testing. And after joining in 2016, I had several jobs there in R&D, in sales, in global business field management, until we founded the NETZSCH Process Intelligence. So the company behind the sensXPERT technology in summer 2021.
Ilkay Özkisaoglu, Composites Lounge: So thank you, Alexander, for the introduction of yourself. Very impressive CV that you have. Can you tell us about your Digital Mold solution? What can we imagine on that one in 2 to 3 sentences?
Alexander Chaloupka, CTO sensXPERT: So I was always questioning myself, what is the benefit of a great technology if our customers are overwhelmed by the data and they are not able to translate it into valuable information? So we made the gap analysis in the past, and we pitched in front of our NETZSCH-internal shark tank, where we ask for money to create the minimum viable product in around the year 2020. And that brought us on the way with several customers that we created the sensXPERT Digital Mold, um, product in our minds, where we are capable of dealing with material deviations and even process deviations that affect the material behavior. So the sensXPERT Digital Mold technology is a package that brings together three worlds of data. So material science from the laboratory. So pure reaction kinetics in an ideal state process information based on sensor data from the material and process information based on process parameters from the machine and with AI models, machine learning models, we bridge the gap in between the different worlds that belong together but are separated today. And we have formed one package that can create the benefit out of these data sets.
Ilkay Özkisaoglu, Composites Lounge: So Alexander, this is very interesting about the values of your solutions. But let me ask you, in the materials industry, why is it that so many companies require destruction-freetesting but still don't use it? What are the hurdles in the destruction-free testing?
Alexander Chaloupka, CTO sensXPERT: Yeah, people always want to touch and feel what they are doing and they want to test something they can really understand. Destructive testing has shown that we can correlate the material behavior to the application range, while having the information of the mechanical performance of the material. So far, there is no commercially available solution that has shown that a non-destructive testing technology can gain the same information, like destructive testing our customers do today. We have done so with in the aviation industry, with our technology, where we have related, especially the chemical degree of cure and the mechanical performance of the material and the components that are resulting from manufacturing. Based on the information we've seen within the manufacturing, based on our technology.
Ilkay Özkisaoglu, Composites Lounge: Alexander, what was the moment it was clear to you that you can take material characterization a step further? What was the moment that really kicked this in?
Alexander Chaloupka, CTO sensXPERT: Yeah, Ilkay. So during my time at Fraunhofer, I was diving deep into material characterization with different equipment and different technologies. So I was looking into different parameters of material. There I created a vision that there must be a way of creating the same information within the manufacturing, based on different sensor technologies, to really bridge the gap and to bring the two worlds together. That time I was challenged by the problems and the issues we have faced within the manufacturing while transferring the lab data into the manufacturing. And I got in contact with a lot of experts in the field. So from the sensor side, from the material characterization side. And a lot of these experts have told me, oh, it's not really possible. So you will fail in doing so. In addition, I created a lot of knowledge. I was really diving deep into the the material characterization information and the material performance and how polymers work and act. And based on the results I have generated at the same time, I could see there is no match between what the experts have told me that it's not possible and the experience I have gained and the knowledge I have gained so far, and that that drove me forward. So I paved the way. And I was then working very, very strict day and night on specific algorithms to generate new information out of measuring or measurement data from the manufacturing. And that showed me that I was right and I should definitely follow this path. Yeah. And that was when it was clear to me that I shouldn't just listen to the experts that are there that have done things so many, many years. So I should really follow my way and use the information we could gain that time and pave the way that in the, in the manufacturing, we will be capable in the future of getting the the needed information and the desired information to really adapt processes in real time.
Ilkay Özkisaoglu, Composites Lounge: So let's come back to the JEC World that was conducted last month in Paris. And I've seen in the programme that you were invited to hold a speech. Let our community know, Alexander, what was the speech about and what were the key takeaways for your community there - the audience? or let's say, what had they in mind? What interested them a lot?
Alexander Chaloupka, CTO sensXPERT: Yeah, yeah. At JEC World this year I had a speech and I presented to the participants the comparison between state of the art technology that we use to translate material information into the manufacturing environment and the real material behavior we could measure within the manufacturing environment. So I showcased an aviation case where it got obvious that what we do today with reaction kinetics based on laboratory data and ideal state, does not match to what we could measure within the manufacturing environment based on our technology. And what I could see in the eyes of the participants was that with what I've shown them, I confirmed that gut feeling because they had already the feeling how the material behaves within the manufacturing. And what's the gap between lab and manufacturing. So I could really see big eyes that were rolling around where what I've showed them confirmed their gut feeling in what is really happening within the manufacturing. And that was the big takeaway for me. And I also think for the participants during my talk at JEC World this year.
Ilkay Özkisaoglu, Composites Lounge: So, Alexander, which extraordinary value do you offer the composites and injection molding industry? That is something that interests, of course, our community.
Alexander Chaloupka, CTO sensXPERT: Yeah. So with sensXPERT Digital Mold, we create unique insights into the material behavior that was hidden so far. So therefore we can translate the data we measure into valuable information that brings us into the position that we can adapt the processes of our customers based on material and process deviations. So therefore, with sensXPERT Digital Mold, we are in the position at the moment to deal with the deviations that can occur, but ensuring the quality of the components of our customers.
Ilkay Özkisaoglu, Composites Lounge: Now we are keen to learn from you your market focus. For which customer segments do you offer this value?
Alexander Chaloupka, CTO sensXPERT: Yeah. So for composites, it's where the biggest challenge is, where historically the most hidden potential is in. So that's definitely aviation and aerospace. And what we see as an upcoming market where we see huge potential is the hydrogen vessel market, where a lot of investments are done at the moment. And we also have seen and we have observed and monitored that there are some challenges our customers have where with our technology, we can definitely help them to overcome them. For the plastics industry historically grown, we come from the thermal setting field where we feel very, very well and where we know exactly what to do, but we are paving the way more and more into even the thermoplastic market, where definitely for the future and the trials we have done with customers together already, we see a huge potential in helping the plastics industry, dealing with the deviations that will occur with the EU Green Deal law in the back that we need to know, or to use more and more recycled material in the thermoplastic field with the upcoming challenges there.
Ilkay Özkisaoglu, Composites Lounge: Finally, what is your call to action? What should the customers now do after they've seen this video?
Alexander Chaloupka, CTO sensXPERT: Yeah, I suggest rethink the current status of your manufacturing. So look at your manufacturing from a different perspective. Um, believe your gut feeling. Where do you see potential that is unused today? And even think about new technologies that will show up in the future, especially like you see here in the background, our sensXPERT cloud environment, not just we offer a cloud environment, even the machine manufacturers offer cloud environments, and it will be more and more widespread in the future. So I see the future is every parameter will be controllable from the cloud in a few years. But what I can suggest, as I mentioned already at the beginning, rethink your manufacturing. Look to your manufacturing from a different perspective. And if you identify that our solution is in the range, you have identified how your potential can be used. Then let's stick together our minds and let's try to find the best solution for your manufacturing environment for the future.
Ilkay Özkisaoglu, Composites Lounge: So that is really interesting stuff that you are doing here Alexander, with your company. Really. Congratulations on that. Thank you so much for these insights. And now I'm keen to talk to Cornelia Bayer, your CEO, and discuss how you guys move the needle further.
Alexander Chaloupka, CTO sensXPERT: Great. Thank you. Thank you for being here. You're welcome. It's great to meet you. Okay. Welcome, everyone. Welcome to this first day of the JEC 2024. My name is Alex. I'm managing director of NETZSCH Process Intelligence. We are acting under the brand name sensXPERT, focusing on optimization and efficiency increase in plastics and composite manufacturing. Today, it's my pleasure to take you on a journey where material and the material behavior stands in the middle of the entire manufacturing process, because if we understand how the material behaves, how are the deviations within the material, we can adapt the process and we can act at the highest efficiency level during production at any time. So. Who are we? We are the NETZSCH Process Intelligence acting under the brand name sensXPERT. We are a corporate venture existing since two and a half years in a German bigger corporate: The NETZSCH Group. That is maybe in most cases relevant for you or already known in the field of material characterization in the laboratory environment. So standard material characteristics like mechanical performance, differential scanning calorimetry, that is maybe where, you know, NETZSCH from. I by myself I was with this company and now with the solution of sensXPERT we bridge the gap between laboratory material characteristics and the real material behavior during production, while combining the information in real time with machine learning models to know at every point in time exactly what happens within the mold and how can we readapt the process based on deviations in the material. So we are located directly in the manufacturing environment. We know that today customers are performing quality control before molding the components based on the pure resin or even the semi-finished products, and after the production at the already manufactured components. But there it's already too late to readapt the process and to optimize if we have been in danger of a scrap production, or if the process simply is too inefficient because we have safety times of 40 minutes, one hour, maybe one hour 30, based on the manufacturing process we have seen so far. How is the product set up and how does it work together? We have sensors characterizing the material in molds during the production of the components. Here, the sensors are widespread in different manufacturing environments from tape laying and then later on autoclave curing, resin transfer molding, high pressure resin transfer molding, injection molding, SMC - it doesn't matter. You can use the sensors in whatever process you're working with. The sensors are connected to our so-called edge device. This is the brain in the production environment. This is a device that is in connection with the machine to monitor the machine parameters for every single product produced, and to also have the material data in the cavity, how the material behaves in real time. On this edge device, we have machine learning models that are being executed in real time, taking the information from the machine parameters and the material behavior in the mold where we can have what our sensors do, the material characteristics. So really curing, crystallization, deviations in the material, even in the flow behavior and the viscosity. But we can also connect other sensors like temperature or pressure to get more information about the process. While the machine learning models are being executed in real time, we know at every point in time what is the performance of the material at the moment, what's the current state, and if we already meet our requirements from there, we can even look into the future to see. All right, we have a process of maybe four hours. After the few first minutes, we have enough data already to see. Are we on the right path for the material? Do we see optimization potential already? And we can predict and can even initialize other actions around the production environment? At the same time, when we evaluate the data on this edge device with the machine learning models, we send the data into the cloud environment. Why the cloud environment? The cloud environment is not just a storage system like we all usually have in mind, when we think about cloud. The cloud is so much more because when we start working with customers, the first benefit is created right away with our solution while bridging the gap between reaction kinetics from the lab and the real material behavior in the mold. But the cloud environment gives us the possibility to act with all the historic data already collected on the customer side to retrain our models from time to time if new information shows up. Because from the very early beginning, when working with the customers together, we cannot know all the failure modes that can occur. We can know some of them. We can even use the information and the knowledge of our customers to bring it into our models. But from time to time, there will be a new state of machine material behavior where new failure modes occur and we need to know them. Therefore, we have the cloud environment where we can give the feedback and the models will be retrained and then sent down to the edge device to act at the highest efficiency level on the shop floor level, where the edge devices connected to the machine. So keeping it short. The edge device connects to the machine and the sensors. We have all the characteristics of the process. We execute the machine learning models, readapt the process, even send signals to the machine when to open the mold, when the part is sufficiently cured or crystallized. And we sent it out the data into the cloud where our customers have access to all the data, the machine performance, whenever they want, wherever they are. We can even connect more than one machines, even at different manufacturing sites around the globe, to one cloud environment that you can really monitor what happens, especially when you have customers that have more than one manufacturing sites around the globe producing the same components. But the performance is different. You can see what are the differences. Every manufacturing site can learn from the other, while having all the data combined in one single source of truth. Now, I want to take you into a deep dive about the material behavior in mold. What data do we use in state of the art technologies, reaction kinetics simulations, even FEM simulations, and how the material really behave? Let's take a very well-known example here. Hexcel RTM6. A very well known material for many, many years in the aviation industry that is mainly used by Airbus in a lot of structures that they are using at the moment. Hexcel RTM6 is set to result in the same mechanical performance when it is stored at room temperature, after taking it out of the freezer for 1 or 14 days. So it is said that if you have stored the material for 14 days at room temperature, it will result in the same performance. And that's true. It will do. But state of the art technology uses data from the laboratory environment. So we characterize one batch of RTM6. We cannot take into account all the deviations that can occur during storing during pre pre curing in this case. But if we take state of the art technologies, then we would measure the temperature in the process. This will remain the same. We don't make any change in the manufacturing environment, and this would lead to the information that we would say. We have the same temperature profile, the same manufacturing conditions. So the material, if it is stored one day or 14 days at room temperature, will behave the same. Now the big question. True or not? Please raise your hands. Who of you is the true? Nobody. I pushed you in this direction. Be honest. Who of you would say? Yeah, that's state of the art. This is what we do today. So we would say it's behaving the same. Please raise your hands. Thank you. I can tell you it's not true. If you do that go one step back again. What did we do? Hexcel RTM6. The reference is a laboratory measurement one batch of material. We create the reaction kinetic model that just uses as input data the temperature in the manufacturing environment. But if the material deviates, we cannot take that into account with state of the art technologies. This is why reaction kinetic would say the same behavior. But what you can see on the upper picture, this is the deviation that really happens. Yes, the mechanical performance at the end is the same. But the behavior during production is totally different. You can see the viscosity behavior is different, so the older the material, the higher will be the minimum viscosity. During production, even the pre curing at this infiltration phase is visible more and more. And you can see that the older material cures faster. Because we have a lot of solvents within the resin that can degas during the procedure of storing it at room temperature. That is why the system cures faster, even the inhibitors blocking the reaction is a little bit more weak with the older material here, because the inhibitors will get destroyed from time to time. Our next example here. This is an example of our bulk molding compound. Even there in the bulk molding compound we usually have a high styrene content - again a solvent material. This can also degas. That means the older the material, the lower is this styrene content. And that also means the older the material, the faster the curing process will be. And even though the viscosity, the minimum viscosity is changing. And if we use different batches of material and we are aware that the viscosity is changing, the cure behavior is changing. We can adapt the process in a different way, because what we could see here from this picture, the older the material, one conclusion could be the faster I can produce my parts. But be careful. We see that the viscosity is changing as well, and I have to entirely fill my cavity to get a good part out of that. So I need to know how my material behaves, not to be in danger with all the material deviations that can occur that I produce scrap parts at the end. The next picture: Mixing ratio effects. For most of the epoxy resins that are out there today, it's already proven that a mixing ratio of +-2% in the hardener content results in the same mechanical performance. So I have an interval -2% up to +2% in the mixing ratio compared to the ideal state that I get the same mechanical performance of my components. But during production, we can see that the cure behavior is totally different. And here in this case. And that's quite exciting to see. Usually I would say if I have more hardener content, the system should cure faster, but it's not. It's somehow blocked. You can see that the reference mixture is the fastest one. And the one, the two ones with -2% and +2% of hardener content, these are slower in curing. Thermoplastics. Warpage. Stability in dimensions. These are the big topics here. Residual stresses. What you can see here is the cool down ramp of a thermoplastic material during molding. This is polypropylene in this case. Initial temperature in the molten state 240 degrees C. Mold temperature 40 degrees C. The big question is when would you open the mold to avoid residual stresses, warpage afterwards and dimensional stability of this part? What you can see if you follow from the right to the left. This peak you can see here. This is the solidification process of polypropylene. Crystallization and the freezing process of polypropylene. The best case would be to demold here where we have a linear behavior. But in terms of shorter cycle times, cost reduction, energy reduction I have to find a balance between stability, mechanical performance, a little bit of warpage, which I can deal with, where I can find a spot anywhere here where most of the crystallization has already been finished. But I can find the balance in between process efficiency and material behavior and stability in this case. So you always have to know what happens in the mold to really create efficient processes while ensuring the component quality. And this is what we stand for, what we are doing. But not to just bother you with examples from reality, even to show you some use cases with customers and at the end the financial benefit because these are the numbers at the end that count. Yes, we all want to act more sustainable and more efficient, but we are always looking and talking into and to companies where earning money at the end of the year is the factor that decides which way to go for. So the first use case, electronics encapsulation. What you see here on the picture is the new brain of a new e-motor series of one of the German OEM car manufacturers. The electronic circuit boards are getting encapsulated to be resistant against acids, humidity, mechanical-stable, shock resistant and what this customer is doing. You can find it in the web. Which customer this is so I can name it here. It's ZF Friedrichshafen, one of the biggest German suppliers in automotive. What they did in the past was a Six Sigma evaluation, a process setup of at least one and a half years from prototyping up to series manufacturing. They lost the information about the correlation and dependency between 40 different parameter variations in the process and the relationship to component costs and quality. This is where we stepped into. We showed them pretty fast the relationship between the different parameter variations and the effect on component quality and component costs. This made made it happen for us that we stepped into the series production, and now we control the process in this case where you can see. Sorry. ZF initially - you can see it here - manufactured the components with a cycle time of three minutes in a multi cavity mode. Three minutes was for ZF the balance between component quality and shortest possible cycle time, while knowing that they have a lot of safety times on top because the material deviates and they need to make sure that the scrap amount is as low as possible. But what you can clearly see here on this picture is the blue dots is the real material behavior in a series production environment. Here you can see an excerpt of 50 parts that have been characterized here and monitored. And the purple line. This is what you would expect if you characterize the material from one batch in the lab, and you try to predict the process behavior of the material. And you can clearly see, yes, there is a trend and it follows this trend. But we have a huge deviation in there. And we can even say three minutes initial cycle time. We even have parts that are finished at almost 2.5 minutes. So a lot of cycle time that is getting lost. And here we are talking about millions of parts every year that at the end results in an efficiency increase of more than 30%. If I can see what happens in the mold and I can react on that. Therefore, putting the 33% efficiency increase into numbers means we talk about more than €14 million every year that ZF was burning before using our system in this case. While just not having the look into the mold and understanding the material behavior in reality. Another example. Unsaturated polyester in an electronics environment. Usually everyone of you knows these circuit breakers. You all have them in your household. What do we talk about here? We talk about an initial cycle time of 22 seconds. So pretty fast for thermoset processing. And you can see the older the material and the picture, you know already because I showed you before when we had a deeper look into the material behavior, the older the material, the faster is the curing, but even the higher is the viscosity. And that means. We need to understand what goes on inside the process, to not produce scrap and to deal with all the material deviations that can occur. But even here, this is a very, very optimized manufacturing in one of the oldest and biggest companies in Germany, where even here we talk about 260,000 burnt money every year per manufacturing line. Now a very extreme use case. Aviation. You remember the example of RTM6 I showed you before? Here we are talking about a structure that is produced in a cascaded process. First resin transfer molding and then a post curing in the autoclave. What happens here? We have a material deviation in the RTM process and that is huge if you look to the scale on the left side. We talk about a variation in the degree of cure of 20% in this first reaction step. If you follow the entire curve then you follow the real manufacturing protocol this customer is using today. Yes. The manufacturing protocol works. You see it here at the end. All the parts result in the same level of degree of cure and glass transition temperature. But the way reaching that is totally different. And we have huge variations in there, where we can save them 90 minutes average in the cycle time just by having a look into the material behavior, dealing with the deviations of the material, and optimizing the manufacturing based on that. Putting this into numbers means more than €80 million every year for one component that is produced by this customer for the aviation industry per year. If we think about the European Union Green Deal Law, sustainability targets, CO2 footprint. I think this is pretty easy to translate into a CO2 footprint if I can save 1.5 hours per part throughout the entire year. So that means when we started in the past, the EU Green Deal law was not present. Yes, sustainability was a little topic, but efficiency in production, that was the big target of our customers. So in the past we started a few years ago with cycle time optimization, scrap reduction. But there is much more in it if you have the control over your process, if you know the material deviations, then we can directly steer our process even in a direction to reduce energy costs. To avoid burnt money to reduce the CO2 footprint. And that is how I want to end my presentation today. I thank you so much for being here. If you have further questions, you will find me the next half hour in the speaker corner right around the wall there. And you can find us here in Hall 6 at Booth T 52, where you can see the package, where you can get deeper insights and where we can further talk about customer use cases and your challenges you're facing. Thank you for being here. Enjoy the JEC 2024. Have a great day and hopefully a great week. Thank you. We would have two more minutes for question and answers. Otherwise I would invite you to join me outside this area. No questions? Yeah, there is one. I think you will get a mic.
Ilkay Özkisaoglu, Composites Lounge: Yeah. Um. Must you create a sustainable, uh, materials to have, uh, the proper sensors and so on?
Alexander Chaloupka, CTO sensXPERT: Yes. We create a materials database so you can understand the package that on this blue box, the edge device there is a materials database on it already about the material behavior in the laboratory. So really the physical chemical behavior that we can put into calculations. And this is what we match with the real production data the real material behavior during manufacturing. So we match these two worlds. And we can therefore calculate how the material behaves at the moment, what are the deviations and how will the material behave for the next five minutes, five hours, depending on the process we have there.
Ilkay Özkisaoglu, Composites Lounge: How many time to generate those, uh, production, uh, instruments.
Alexander Chaloupka, CTO sensXPERT: What takes longest is the discussion with your IT department. That's from experience. So that's really what takes longest the IT department. And that is what our package is based on. You will have the cloud connection. You will have access within your company to the system to see what happens in the mold and how the machine performs. That is from experience what takes most of the time. Um, if you decide to go that route and everything is prepared, then we can send you the package. If you order tomorrow morning, we will send it out tomorrow afternoon. You will have the package on site. It's easy to install, so we even have customers around the globe where we don't have been on site for the installation. We have been there before to see how the system is installed, how is the manufacturing environment. But then they can do the installation by themselves and then we can directly start to collect data, and usually within one week with all the data you have produced, we can already create the first machine learning model for you, where you will benefit from and then from time to time, in exchange with your experts on site, we will learn more about the failure modes that can occur, but you will have the jumpstart already after the first days or a few weeks latest.
Ilkay Özkisaoglu, Composites Lounge: Okay. Thank you.
Alexander Chaloupka, CTO sensXPERT: Thank you. Okay. Maybe. I think now we have reached the end. Exactly to this point, so I will invite you to the speaker's corner out there directly behind the wall. If you go out to the right side, I will be there in one minute and will be available for you. Thank you for being here.
Dieses Transkript wurde erstellt mit www.amberscript.com