Möchtest du unsere Inhalte auf Deutsch sehen?


Metallurgy and Heat Treatment Digital with ALD Expert


Click on the button to load the content from Spotify Player.

Load content

Listen to the IoT Use Case Podcast on Spotify.
Listen to the IoT Use Case Podcast on Spotify.
Listen to the IoT Use Case Podcast on other platforms.

IoT Use Case Podcast #133 - ALD Vacuum Technologies

In episode 133, Madeleine Mickeleit welcomes Frederic Schum, Product Area Manager for Digital Solutions at ALD Vacuum Technologies, a leading manufacturer of vacuum technology systems for metallurgy and heat treatment. ALD operates worldwide and is known for its innovative solutions in the melting, casting and heat treatment of metals.

Episode 133 at a glance (and click):

  • [14:13] Challenges, potentials and status quo – This is what the use case looks like in practice
  • [22:09] Solutions, offerings and services – A look at the technologies used

Podcast episode summary

ALD Vacuum Technologies is a company with 900 employees in 10 countries, active in two strategic business areas: vacuum metallurgy and vacuum heat treatment. Frederic Schum introduces its new “Digital Solutions” product area, which is managed under the ALD Expert brand and offers digitalization solutions for all ALD system segments. These solutions are highly specialized and designed to meet the specific requirements of the industry.

A particular focus is on three specific use cases that ALD has already successfully implemented:

  • Camera-based process monitoring and image recognition (module: AOS): Enables visual monitoring and anomaly detection to ensure the coating quality of turbine blades.
  • Process data visualization and monitoring of KPIs (module: combination): Visualizes and monitors key performance indicators to optimize plant control.
  • Melt review sheets and connection to IT systems: Improves documentation and traceability through melt review sheets and IT integration.

Frederic explains the challenges their customers face on a day-to-day basis, such as the obligation to provide proof of quality, susceptibility to errors and the need for traceability and quality assurance. He emphasizes that ALD Expert offers a modular platform that enables web-based access and an open system architecture that can also be integrated with existing customer systems.

A central aspect of the discussion is the technological implementation of the solutions. Frederic explains how ALD Expert is based on Docker to deploy applications in containers and how special Python scripts are used for data analysis. The data is stored in INFLUX DB and visualized with Grafana.

Podcast interview

Welcome, Frederic. It’s great that you’re here today. Yes, how are you doing anyway? Where are you right now?


Yes, hi. First of all, thank you very much for the invitation and thank you for letting me be part of IoT Use Case or you today. I’m doing fine so far. Thank you, I hope you too.

Thank you for asking. I’m doing great too. You, I just saw here on LinkedIn that you were traveling in China, right? What have you done there?


Exactly. Right now I’m here in Hanau, at the ALD headquarters. That’s about half an hour from Frankfurt. Next to the Wolfgang Industrial Park, which is known to many people through the Evonik company. SAXONIA also exists and the Fraunhofer Institute has now also built here. The whole thing is now also called Science Park. So it’s a super great infrastructure that has been created here in Hanau-Wolfgang. We are 550 people here at the site, so that’s really cool.

Great! Best wishes to our colleagues on site for now.


Yes, thank you. But you’re right, LinkedIn reveals a lot. I was in China last week and many sales colleagues, including our management, were there. We held the so-called ALD Symposium there. It lasted a week, two full days, lectures and workshops. And on the third day, we also visited our heat treatment center in China. And yes, it’s cool that you bring that up.

Fits in well with today’s podcast. On Wednesday morning, I also gave a presentation together with our CTO about our new digitalization portfolio and the ALD Expert umbrella brand. I also presented the architecture and various solutions and our Managing Director, Mr. Wittich, went into more detail about energy storage and new technologies that we have developed here as part of the ALD.

Yes, now you’ve given me the perfect transition. You’ve already mentioned the keyword, heat treatment. You have a lot to do with that. I think we’ll talk about that again in a moment. Perhaps we can start by introducing yourselves and talking a bit about you as a company and then move a bit towards digitalization. Reading about you online, you are the world’s leading manufacturer of vacuum technology systems. I’m sure you can tell us a bit about that. You have, I think, 900 employees, are represented in ten countries and are part of a listed group environment. Is that correct?


Yes, nicely summarized. Let me start from the back: Integrated in a corporate environment, that’s right. The parent company is called AMG, Advanced Metallurgical Group, which also has various strategic business areas or strategic business units. And the engineering division of the AMG Group is ALD Vacuum Technologies GmbH with its almost 900 employees, where I am sitting here today at the headquarters in Hanau. Leading manufacturer of vacuum technology systems also fits well. We have two specific strategic business areas: vacuum metallurgy and vacuum heat treatment.

Okay, that just means metallurgy for those who are not familiar with the industry. That means you melt, you cast, you deal with metals and you probably process the whole thing in liquid form. That’s the topic of metallurgy. Can you explain it like that?


Exactly, metallurgy means that we have a wide variety of product areas in both business areas that relate to the systems that we manufacture and sell accordingly. In metallurgy, we have melting and remelting plants. In other words, where metal and sometimes scrap is melted down under vacuum in several process steps and high-purity alloys are produced. We have systems for investment casting, for example, titanium is melted down and cast under vacuum. These are end products, small turbine blades. Example: We have systems where titanium turbine blades are subsequently coated with a wafer-thin ceramic layer. I’ll come back to this in more detail in the use case. These are so-called coating systems. Systems for powder production, i.e. powder atomization systems for downstream 3D printing. Yes, that’s roughly what you do in the field of metallurgy.

Yes, super exciting. Just to have an order of magnitude, how hot do your systems get? We’re already talking about a few thousand degrees, aren’t we?


Yes, absolutely right. And under vacuum.

That means in a vacuum, where the whole thing is then processed further?


Exactly, to ensure the purity in the process in the process.

Okay, I can imagine. You have now also mentioned turbine blades. These are precision components and also huge parts. It’s really exciting to see the environment you’re working in.


Yes, that is very exciting. We have a very large product portfolio and always have special systems. I just remembered a nice example. We even made a vacuum roasting system for the Bahlsen company several years ago. That means chips, which you can now get in the supermarket.

Okay, interesting. So it’s probably worth taking a look at your website. There are many different systems in all possible sizes and for different application scenarios. Today, however, the focus is on digital solutions, because you are not traditionally in mechanical engineering. But you also offer digital solutions. You’ve already mentioned what you said at your customer event or the symposium. Can you give us a little insight into what exactly you are doing in the Digital Solutions division and what your responsibilities are here? And then perhaps we can also talk a little about your customers. Maybe a little bit about your department first. What are you doing there exactly?


I’ve been with ALD for a good five and a half years now. This was also the birth of a new area. The division was originally called Automation and Industrial IT. When I joined the company, I was responsible for business development, helped set up the organizational area, built up the product and service portfolio and, of course, was also responsible for sales in order to implement the first customer projects. The aim of this department, which is really unique to ALD because we have been a machine and plant manufacturer for many years, was to set up a department responsible for supplying digitalization and automation solutions for all plant segments. And these new solutions are marketed by us under the ALD Expert brand.

It’s an exciting transformation that you’ve accompanied, isn’t it? But I mean, this is a completely new area, and probably completely new expertise that you have developed in-house. That was a huge change for you in terms of sales and organization, wasn’t it?


Absolutely and that’s why I’m so proud of it, it was a huge challenge. Let’s face it, this is a start-up founded by a conventional company, with advantages and disadvantages. We use existing processes and draw on all of our colleagues’ expertise, which is essential. But innovations and changes in the conservative market are always difficult. You know it from your history, but you have mastered it successfully. And now I’ll build another bridge to the digital solutions you mentioned. After the success was established and many customer projects were realized, we also decided, I can proudly say, to establish another sales division, in addition to the coating division for melting and remelting systems, a division for digitization and automation solutions. And that is the Digital Solutions division, which I have been responsible for in Sales for almost two years now.

Okay, and Digital Solutions, this also includes your product ALD Expert. Is that right?


Precisely, everything that goes in the direction of digitalization and automation is managed under this umbrella brand ALD Expert.

Just one last question. You have now mentioned coating. How many developers do you have in the team now?


The department now comprises just over ten developers. Frontend developers, backend developers, people who are responsible for DevOps. So really the definition of new-fangled jobs in the IT environment. And the interesting thing is that the people have a totally diverse background. We have a really diverse range of backgrounds, from PhDs in chemistry and physicists to IT specialists, and that’s what makes this new area so unique.

Yes, exciting. Do you also do data science and AI topics in-house? Or is that something you do with partners? Just to pick up right there.


Yes, we definitely do data science. We will come back to this later, when we are not only looking at data aggregation, but also data analysis. I’m always a bit wary of the term AI with our developers, they always put the brakes on me. So first of all, data analysis. At some point, we will be able to talk about machine learning, self-learning systems and perhaps AI at some point in the near future.

Exactly, statistical models are a challenge in themselves. Then perhaps a question about Digital Solutions. We are now also talking about what is your vision here now, especially now with your customers? Perhaps you can talk a little more about what added value you are now creating for your customers. Who are they anyway? Or which customers do you have there?


Gladly. Clearly, our USP, as I always say: solutions from the industry for the industry. What is the added value that we create? ALD, machine and plant manufacturer for decades. We have the personnel here who have built up and perfected system know-how and process expertise over the years. And we are now combining this with the digitalization expertise of our new, younger employees from the IT segment, and this is creating truly new digitalization and automation solutions for our existing systems. So really always specialized in our core area.

Okay, so does that mean it’s mainly your customers that you already know, or is it also new customer business that you’re targeting?


First and foremost, when we come back to the outlook later, it is our existing customers that we are focusing on. However, we may also find new customers where we have not yet managed to sell systems that are in our pipeline and perhaps, of course, open up new markets in a further step.

Cool, yes. It’s also exciting to perhaps extend this again. Exactly, now I always talk about specific practical use cases here in the podcast. When I think of your systems now, I think of data such as temperature, pressure, it’s probably also about cooling temperatures and so on. Can you give us a little insight into what use cases you are actually implementing with your customers? Every customer is probably a little different, everyone has different needs. But are there any recurring use cases where you go in and implement them? Do you have a few examples?


Yes, you put it beautifully. Every customer is different. We are an absolute custom plant manufacturer. Every customer is unique and has special requirements. But today, of course, we have several use cases, but I would like to briefly outline the three that occur most frequently. One is, for example, that we have supplied camera-based process monitoring for the process. And it is not only the detection and recording of the process that is interesting, but we have also written image recognition software into this use case that makes it possible to detect irregularities during this production process.

Okay, so this is streaming data from the camera data that is then sent to you, so to speak. Okay, interesting. Do you have any other use cases? just to understand the portfolio a little bit, where you are traveling, just a little bit teased?


Another use case, for example, is process data visualization. This means data aggregation and visualization of all systems at the entire production site, i.e. a central analysis of all machine and system data. And of course it goes one step further, the monitoring of so-called KPIs. So we are talking about condition monitoring. This is usually a correlation between different process parameters.

Okay, that means it’s also a use case that you have more frequently, so to speak, that customers generally want to monitor your systems. What is the condition? What are the company’s KPIs, so to speak?


Exactly. Condition monitoring of various process parameters, system components, simply to be able to predict failure scenarios in good time.

Okay, cool. Do you have another one?


Yes, Use Case 3, which is very often used in the melting and remelting area. We call them Melt Review Sheets. This means that we connect the system first. As a rule, there are also several systems. This is when the real added value of ALD Expert comes into play. With the higher-level IT system. Every one of our customers has an Enterprise Resource Planning (ERP) and a Manufacturing Execution System (MES). The first step here is to ensure bidirectional data transfer between the system and the higher-level system.

What’s more, so-called quality sheets are generated automatically. This means that process parameters are monitored and automatically entered and documented in a review sheet at the end. Did I have a deviation from the prescribed target values? Yes, no. Was my process parameter in the intended range? Yes, no.

Okay, interesting. In other words, it goes a bit in the direction of quality monitoring, but also the digital documentation of, yes, perhaps an obligation to provide evidence or simply deviations that you need to present internally.


Exactly right. In the end, I have created a tool where I can say that I had this and that batch at this time and on this date, with these and these process parameters.

[14:13] Challenges, potentials and status quo – This is what the use case looks like in practice

Okay, keyword obligation to provide evidence. Perhaps we can talk a bit more about the business case. Because that’s also an issue for customers now. I mean, I’ve probably had your system there for many years, if not decades. Now it’s all about developing added value with the customer. What is the business case for your customer? So what is he losing today in terms of time and money potential where he is perhaps now prepared to pay money for it or where he sees great added value for himself? Do you have any examples of this business case behind it?


Good question. I would say, let’s take one use case for this. I would like to pick the first use case, camera-based process monitoring. We have developed this solution for our coating systems. In these facilities, titanium turbine blades are coated with a wafer-thin ceramic layer in the micrometer range so that the turbine blades, later used in engines, can withstand a higher combustion temperature and can therefore be flown with less noise and less kerosene consumption. An electron beam gun is used to irradiate an ingot. An ingot has a cylindrical shape and is a material. In this case, for example, ceramics are bombarded with the electron beam. This vaporizes the ceramic and in this vaporization cloud the turbine blades are moved back and forth in a translational and rotational manner so that they receive a very homogeneous, thin layer of the ceramic. And why did I say that? We are talking about the aerospace industry here, we are talking about turbine blades, and then you can imagine the immense quality requirements that such a process entails.

In other words, it’s simply about the business case. Yes, the quality must also be right there. These are very high requirements. At the end of the day, if a turbine blade doesn’t meet the quality requirements in the test, that’s money down the drain, so to speak. Then this is a batch that I can no longer use.


Exactly, this is about the obligation to provide proof of quality. Such a destroyed batch costs well over 10,000 euros. If such a coating does not work properly, this means that the production process has to be stopped. This incurs considerable costs and means that the turbine blades made of high-quality titanium, which cannot simply be thrown away, have to be stripped of their existing coatings in a complex process before they can be recoated, and if you’re asking specifically about the business case, it’s really about not wasting employee capacity when I have plant downtimes or have to remove these coatings again. I want to reduce system downtimes and, of course, I don’t want any loss of quality and I want to have sustainable quality documentation.

What your customer had as a challenge here, looking at this coating, has probably been solved manually somewhere in the meantime and now you also want to have it digitally in front of you to some extent. This is probably the challenge that many of your customers are facing right now.


Just the right start. And then to expand again, an operator, a user, stands in front of a small sight glass, a so-called viewport, for the entire shift, which is a three-shift operation, and really has to look through it for eight hours and watch the evaporation process of the material. We are talking about so-called pools, where the material is vaporized.

Eight hours. Sorry, I have to process this first. Eight hours is really a lot. That’s the whole working day.


Once you’ve seen it, it’s an extreme challenge ergonomically and how do you manage to be fully alert for eight hours to recognize irregularities in the evaporation process and then he has to change the parameters of the electron beam gun and so on on the control panel so that the process runs smoothly. This is precisely why we developed the camera system, specially developed holders, specially selected cameras that record this entire evaporation process and display it on a huge screen above this so-called viewport.

What kind of data do you need to do that? D just mentioned it, these are ceramic layers, these are probably combustion temperatures. What data sets do you need to tackle this use case with customers now?


In concrete terms, all the data from the camera is needed to record the process alone. I talked about it again, so for now it’s just the recording of a process. We have taken the whole thing even further in this use case and have also programmed an algorithm that automatically detects these anomalies, which the operator should recognize with the naked eye, and displays them to the operator in this software on this large screen. This means that the operator no longer has to constantly look through this sight glass, but has the option of having these irregularities displayed on this large screen as a video.

Do you have some data that you can give as an example? I think you can then always understand a little more precisely what data you are taking from the controller or PLC.


This is where it gets interesting. Not only the process videos are interesting for the customer, but also the corresponding correlation to all process parameters, so that in the end the process videos can be linked to the current process data. And these data are, for example, the classics: temperature, pressure over time, the gas flow and also the gas pressure in this so-called coating chamber, which are very decisive.

You combine your knowledge with that of the customer, as the user knows the system in practical use. Of course, you know the system inside out from research and development. In other words, you probably work quite closely with the customer to identify precisely these parameters and the errors that can occur, right?


Absolutely right. Yes, exactly, as a rule the customer approaches us and we build the system around the desired process.

Many customers are already doing this with you. Many still want to implement it. What are the classic requirements, perhaps from a technical point of view, where customers tell you they need a solution? Are there any special technological requirements for your solution? In this case, it is probably ALD Expert. What do you need to bring with you so that you can set it up at the customer’s premises?


Yes, so in our language, in IT language, an on-premise solution is crucial.

Yes, so hosting on site, at the customer’s premises so to speak, right?


Exactly, we have highly sensitive, productive data and a cloud solution is not yet an option in the short to medium term. So always an on-premise solution. What the customer also demands from us is the accessibility of the solution. So the solution can still be web-based, which means that you can not only access our solutions directly at the system, but also from your office network or even from home via a secure VPN tunnel. Customers who buy systems from us are generally used to turnkey installations and expect the same from our digitalization and automation solutions. This means that we take care of the installation, commissioning and, of course, the training of our ALD Expert solution afterwards.

You have had service level agreements with the customer for years. In other words, that’s a bit on top of what is then agreed.


Exactly, so in addition to the system, we also have an after-sales area where the customer can book regular system maintenance with us. In the software business, it is also common for what you call service level agreements to be offered. I am also trying to do this more and more for our ALD Expert solution. But it’s still a bit of a difficult undertaking at the moment. The customer simply assumes that it’s the same as with a system: put the system in my yard. Software solution as well, should be put into operation once, should work and that the customer is willing to pay another annual fee for a software solution is not there at the moment, yes, not yet.

[22:09] Solutions, offerings and services – A look at the technologies used

If you now set up ALD Expert as a solution, what does it look like and what exactly can it do? Can you tell us a bit about how you implement this on site? With what scope of products, so to speak?


Gladly. So the basic prerequisite, I always say Layer 1, is to establish connectivity to the machines and systems at the customer’s production site. This means that the systems are integrated somewhere in the customer’s manufacturing network and we have to access the data somehow. This is usually done via direct access to the PLC of our system, i.e. to the control system or to the system computer that is responsible for operating the system. Layer 1 is the installation of our own platform, which we programmed ourselves, the so-called ALD Expert Grid. This platform runs either virtualized on the customer’s server. This is always the preferred solution. However, we can also install this solution on a small industrial PC, which is then usually placed in the electrical control cabinet of the system. So two variants. Either virtualized on the server or installed on a small industrial PC.

Okay, and you said one, which means zero was the connectivity. And then one thing is your ALD Expert Grid, as I believe you called it, i.e. the integration of the software on site.


Exactly, this is the programmed platform on which the databases run, the necessary concept for data security, role directories, i.e. where the user can then log in with their corresponding Windows credentials.

We have just talked about the business case. In other words, the customer ultimately wants his evaporation process in correlation with the camera data, but also with the process parameters. Do you have different modules where you solve this case, for example? So are these software modules that are customized according to the customer or how should we imagine it?


Exactly, layer 2 is then, as you say, the application level, the solution level, where we then implement our modular solutions on the platform, which are called process data visualization and condition monitoring. For the use case you just mentioned, I didn’t even mention the name for the camera-based system and image recognition system AOS earlier. This is short for Advanced Observation System. This is then the solution or a modular solution that is installed on the platform.

Okay, interesting. There are many companies, particularly in the mechanical engineering sector, that are entering the market with precisely such solutions. Of course, this is always very specific. You are now in a very specific core business where you have of course been experts for years and where the USP or unique selling point, as they say, is almost obvious because you also know the processes. You have the team and so on, you know your way around the analyses. But what makes this solution so special? Especially now in relation to the ALD Expert. Can you explain this again in comparison to existing solutions on the market? Perhaps there are no direct competitors. You know more about this area.


Gladly. So three clear statements at this point. One thing I mentioned earlier is the expertise we have built up over many years in terms of the system and the process, combined with the expertise in the area of digitalization and automation that our new IT department brings to the table, i.e. the area of automation and industrial IT, point 1. Point 2 is that we are not limited to ALD systems with our ALD Expert. We know that our customers also have machines and systems from other manufacturers. Sure, there is competition, but with ALD Expert we can establish connectivity to all machines and systems. This means that ALD Expert can be the central solution for an entire production site. Point 3, ALD Expert has an open architecture. This means that we can connect to existing customer systems. Most customers will have Enterprise Resource Planning systems or Manufacturing Execution Systems. And we can interact with these systems.

Yes, what you meant with Use Case 3 was, I think, the one with the “Melt Review Sheet”. For example, you also need data from the MES or ERP. In other words, you would then dock on and integrate this data.


Exactly. This is an important point for us, because who wants an encapsulated have a solution? It is important to push the production data into the higher-level systems. In return, the system also wants to receive the order digitally from the ERP system.

Very nice. We’ve already talked a bit about data recording. Finally, I would like to briefly discuss data processing, hosting and analysis. Perhaps just one more question. You have now said openness of architecture. What does that mean in detail about the technologies you use there? So when we talk about data processing or hosting, what technologies do you work with?


Data is usually recorded directly via the PLC or the system computer. For example, when we read out data from the system computer, this production data is already stored in the database, in a CSV file. The corresponding log file parser is then written, which reads the data from the CSV file. If we go directly to the PLC, for example if we need high-frequency data, then this can be done using a programmed OPC UA agent, for example. Of course, there are also ready-made open source solutions such as those from Kepware. The offers various software adapters for different PLCs. Why is this so important to us? We deliver systems worldwide. In the Asian market, the control system can be from Mitsubishi, in the USA it is usually Rockwell Automation. The automotive industry is very Siemens-heavy. We therefore need to adapt to these different issues.

And from there, the data then goes from the PLC into the ALD ExpertGrid, as you said. Is there any relevant information on how to make this application available for the software?


Exactly, so as I said, the ALD ExpertGrid, the platform that then receives all the data from the connected machine systems, is the linchpin of the modules. Each of our modules is equipped or structured with so-called microservices. These are self-programmed Python scripts, various plug-ins that make use of open source libraries. Yes, the architecture of ALD Expert is structured in such a way that we use Docker software to provide the application virtualized in so-called containers.

Earlier you talked about step two of the application, which is data analysis, evaluation and visualization. You probably also have a standard toolset that you work with to assemble different software depending on customer requirements.


Exactly, what comes to mind is that our developers use special tools for data analysis. Use Python scripts to establish a correlation between several process parameters. Data can be stored via an Influx database, especially for process data. Grafana dashboards are very popular for visualizing data. We rely heavily on open source technology, which is also state of the art today.

Yes, most people use this. And it’s also nice to see that you are building on this. Very nice. If you listen now and say, hey, Frederic, I have a similar issue, let’s talk. I’ll just link the contact details for you, Frederic, in the show notes. You are welcome to network on LinkedIn, where you can also discuss details or common starting points. Perhaps the last point for today. I am also always asked in the network about best practices, how has he or she implemented the use case, what are the pitfalls? Do you have any experience from your many years of project experience that you would like to share with the audience?


Absolutely, I had to go through the same experience. We said at the beginning that we are a start-up in a conservative industry. I have noticed that terms such as cloud or AI unsettle our customers. We need to use simple language that our users in the industry also use. It doesn’t matter whether the solution relies on Grafana dashboards or something else, the solution is crucial. The customer has certain requirements and we have to find a solution based on these requirements. This solution is then compiled from the modules in the ALD Expert solution portfolio and adapted to the customer’s requirements. The technological framework conditions are all in place today. We can draw on any technology and the challenge is simply to demonstrate the added value of the solution to the customer. I think that’s the biggest challenge in a new digitalization project.

Yes, definitely. Very nice. Thank you very much for the best practices from your side and from my side, thank you for taking the time today. As I said, I have a lot more questions, but that can be clarified afterwards. Many thanks from my side. These were very specific use cases. Of course, you are in a very specialized core business where you can talk very specifically about your systems. It was really exciting to get an insight into what your solutions are in practice. You have done a fantastic job of describing how this works. So thank you very much from my side. Thank you so much, Frederic, for joining us today and taking us into your use cases, your journey and the vision and objectives with your customers. That would give you the last word of the day.


Many thanks for the words of praise. Thank you also for having me on your podcast today, on IoT Use Case. I was very pleased to be able to give the audience an insight into the architecture of ALD Expert and our solution portfolio. I would be happy to answer any questions and I would also be happy to be invited to another session if necessary, because there are still many use cases and projects that I could report on.

I think so too. It cries out for a follow-up. Thank you very much for today, glad you could join us. Have a great rest of the week. Take care. Ciao.


Thank you very much, I wish you the same. Ciao, ciao.

Please do not hesitate to contact me if you have any questions.

Questions? Contact Madeleine Mickeleit

Ing. Madeleine Mickeleit

Host & General Manager
IoT Use Case Podcast