Siemens & AUDI AG – Digital building technology: Control system recognizes potentials in plant operation

Hello Marc, hello Sebastian, welcome to the Industrial IoT Use Case Podcast and a warm hello to you too Rainer, here today representing Audi. Before I get straight to the content of your daily tasks and challenges at the site, I would start with the Siemens round. Marc, I’ll turn the floor over to you for a brief introduction of yourself and your field.

Marc

Thank you very much for the invitation. I am happy to be here. My name is Marc Richter and I work for Siemens Smart Infrastructure as Head of Digital Service Sales. My background is in traditional industrial automation. Here, a few keywords: WinCC, sematic controls. That’s the field I started in, too, and then found my way into building technology in 2009. There I worked in sales at Securitytechnik. That is: access control systems, video surveillance, intrusion detection technologies, building management system. Then via the path of building automation, i.e. HVAC (heating, ventilation, air conditioning) and instrumentation and control (measurement and control technology), I finally got into this position two years ago and that is of course very exciting because the topic of digitalization does not stop in building technology. Because with data analytics and cloud computing, of course, a whole new set of possibilities await us in terms of how we can meet the demands that come to a building.

Siemens Smart Infrastructure started like this: The world of the smart grid as well as the buildings, at the point where this meets, we talk about the so-called grid edge, i.e. the connection from the building to the smart grid. Now I have the smart building on one side and the smart grid on the other. We know that the ever-increasing use of renewable energies will bring about an enormous change in our power grids, and that buildings will also play a very important role, because buildings will change from being pure consumers to prosumers: The building is changing from a pure consumer to a so-called prosumer. So you get electricity in certain times, but you also produce electricity through cogeneration plants, through PV(photovoltaic) plants and many other possibilities. This also includes storage systems. This is changing a great deal, but also in the smart building sector. There are things like indoor navigation, there are completely new applications with which the user of the building, for example the tenant, interacts in a completely new way. One would like to know, for example, “What is there to eat?”, “I would like to book a workstation”, “I need to move around the building”. To be able to navigate in the building, I can do it via classic navigation. But there are also things like data analytics, because energy efficiency is playing an increasingly important role. The building thus becomes more and more intelligent. The issue of demographic change also plays a very important role. The topic of employer branding has a very important impact, so more than just the place where you work. On the other hand the operation of the building is becoming more complex. At the same time, however, there is less and less specialist staff and less specialist know-how available, so that digital assistants are very much in demand. It is precisely this whole construct that Smart Infrastructure, i.e. the topic of Smart Grid, Grid Edge and Smart Buildings, is concerned with. These are the three main key points where Smart Infrastructure can be classified.

Sebastian, I would just hand over to you. Could you also briefly introduce yourself, briefly say something about yourself and in which area you work at Siemens Smart Infrastructure?

Sebastian

My name is Sebastian Witt. I have been with Siemens for seven to eight years. At that time, I started with a dual study program in the field of energy and supply engineering, i.e. classic mechanical engineering, and then also did a master’s degree in the field of energy engineering. For a long time I have been working in the field of power engineering as a power engineer and energy technician for a large number of customers from the commercial segment, but also from the public or industrial environment. In the process, I developed and established concepts on how we can sustainably save energy, increase energy efficiency and enhance production reliability. And always in the area of smart infrastructure, meaning the entire energy supply, the whole topic of plant automation, and how we can optimize that, so to speak. I then moved into my new role as a Digital Service Specialist just under two years ago. That means: I do the technical clarification with the customers and the colleagues on how we can integrate the data into our cloud platform, how we can create connectivity, and how we can visualize this data and create corresponding added value. Always under the certain aspects of energy efficiency, plant performance and what you used to do in person at the customer’s site, so to speak, now happens from the cloud platform.

Rainer, then actually only you remain in the round. I would ask you to also briefly introduce yourself as to what exactly your responsibilities are and maybe you can already tell us a bit about your site and what your company is working with.

Rainer

I’m happy to do that. My name is Rainer Walter. I work here at Audi in Neckarsulm and am responsible for planning and operating the energy supply, building services and infrastructure – both for the site and for external properties. My professional career began classically with an apprenticeship as an energy system electronics technician. That means I practically started from scratch with electronics. I then went on to study precision engineering, comparable today to mechatronics with a focus on measurement and control technology. So also again in the direction of electrical engineering, automation technology, and then started after graduation in the field of automation technology with planning, design, programming and commissioning of systems. Subsequently, I moved into the area of production plant maintenance, again with a focus on plant automation. And today, here at Audi’s Neckarsulm site, I am responsible for plant infrastructure and energy supply, an area with around 90 employees, with whom we safely supply the entire plant here.

Marc, maybe a quick question for you to start with. I’d like to situate the listeners a bit on the topic of buildings. Can you describe what tasks a building has today from your point of view? Just to get an idea of the whole thing, including your customers, and what the tasks are classically.

Marc

When you hear about building technology, you probably first think of a classic office building. And think: Hey, the building is changing, it’s getting more and more technology to make people feel safer and more comfortable in it. This includes fire safety, security, access, it must not be too hot, not too cold, not too noisy. It needs healthy indoor air, the humidity and oxygen content must be right. Of course, I can apply the whole thing to a wide variety of objects. These are hotels, these are hospitals, these are public buildings, museums, you name it. Building technology also plays a very important role in the production environment. I like to call this production-relevant building technology. In other words, indoor air quality is very, very important in the manufacturing environment of the pharmaceutical and food industries, for example. In the case of pharmaceuticals, you can imagine that there must not be any high fluctuations in terms of temperature or humidity. “Clean room” is the keyword there. But I have the same thing in the food industry environment. If I imagine a butcher shop like that, which always has to be refrigerated, I also have a very, very important aspect. Or when I look at paint shops, even there air currents must not be too strong, but it still has to be dried. This building technology has a variety of tasks and in the production environment it is also becoming increasingly important because the requirement for this technology also continues to increase.

Requirement is exactly the keyword. I would jump straight into the practical side of things and ask Audi about this. But first, what are the top three challenges that you face or that you deal with on a daily basis in this area?

Marc

At the end of the day, I would almost say that the same issues always drive us. First, of course, there is the issue of cost efficiency. This means that operating costs, e.g. for maintenance, are to be reduced. At the same time, of course, it is also a question of plant availability. So you want to have fewer disruptions and downtimes. This is also a very important driving force. But then we also have a point – which I believe we cannot ignore in Germany – and that is the loss of know-how. Plants are becoming increasingly complex, they are becoming more and more intertwined, and they are being networked with many other trades and systems. At that point, it gets exciting, because a lot of overarching knowledge is required. And now, of course, we also have demographic change in Germany. And when I look at the technical operator groups of our customers, I have to ask myself: What actually happens when the baby boomer cohorts retire there at some point? Do you actually manage to maintain this knowledge? I think that is a very important driving force for the topic of digitalization, which is also taking hold there. The topic of digitalization offers us completely new possibilities in the area of plant availability. But also to ensure the issue of cost efficiency in a variety of ways. And now there’s something else – Sebastian has just alluded to it – and that’s the issue of energy efficiency. Carbon Neutrality, CO2 neutrality – of course, that’s coming up enormously now. And that’s where you also demand that the buildings make their contribution, because of course they also consume a lot of energy.

From your point of view, so to speak, it’s cost efficiency, operational and energy efficiency, sustainability, plant availability and the loss of know-how due to complex networking. I’m looking in your direction now, Rainer. You said you were responsible for building services and infrastructure at your site. Can you tell us a little bit about your site in Neckarsulm, what buildings you have on site and also what requirements come with that in practice?

Rainer

Here at the Neckarsulm site, we have a wide range of buildings. We have both our production buildings, which include the paint shop, body shop and assembly departments, but also administrative buildings, workshops and office buildings, and of course our technical development buildings. In total, we have an area of 1.25 million square meters here, on which our 70 buildings are located, all of which we supply with energy. The whole thing always according to the requirements, from the respective plants or technicians, who are located in the individual buildings. What also makes our situation here a little more exciting is that we are limited in our expansion here at the Neckarsulm site by the Neckar River on one side and the city of Neckarsulm on the other. This means that our buildings have increased significantly in height in recent years. We now have buildings that are up to 50 meters high, and this of course means that the demands on energy supply and the legal building requirements are constantly increasing.

What are the classic tasks in your team? That is, what must be ensured on a daily basis during ongoing operations, for example?

Rainer

The classic task or the original task for us is, of course, to ensure that the site is supplied with energy and media in the right quality and quantity. That is, electricity, water, ventilation, refrigeration, gases and this provision under the applicable laws and regulations. It is a matter of monitoring the plant to ensure security of supply and, of course, also of recognizing potential, where we can improve what. All these aspects from the point of view of efficient use, both in terms of energy demand and the economy of the plant.

That definitely sounds complex and like an exciting task. What are typical challenges you encounter in the process?

Rainer

The challenge is to provide media of the right quality. Quality means, for example: Keeping the correct air temperature constant in a test stand or in a measuring room and, of course, also providing the correct humidity in the room so that the measurement protocols or test stand processes run in the correct framework conditions.

To get a virtual picture: What kind of systems are there on site? Are these classically ventilation systems or how do you have to imagine the infrastructure on site?

Rainer

We have a wide range of ventilation systems here. We have 100,000 cubic meters of ventilation system here to ventilate the production buildings. However, we also have small ventilation systems for office areas or ceiling-mounted air-conditioning units that we look after here, and which we have also connected to our building control system accordingly.

A quick question: What role does energy efficiency play for you?

Rainer

Energy efficiency has been of absolute importance here since 1995. That’s when we started introducing an energy management system. We also have a target to reduce our energy consumption by three percent annually, so not absolute energy consumption, but to define measures that represent three percent of annual energy consumption, how to reduce energy consumption and thereby increase energy efficiency.

And in this context, probably the issue of sustainability, which certainly also plays a role for such a large site?

Rainer

Yes, the issue of sustainability also plays a major role. We have set ourselves the goal of making all our sites CO2-neutral by 2025. To this end, we also have a program called Mission Zero, in which the topics of decarbonization, water consumption, emissions, and biodiversity are bundled and corresponding projects are set up and measures defined that contribute to achieving the goals.

You have just mentioned the building management system or the building control system. What can you already monitor with this today?

Rainer

Today, we have a building control system in which you can manage about 70,000 data points, have more than 1,000 systems connected, and control the plant. Control means we run the plant based on production. So as the production runs, so our plant is switched on or off accordingly. This is done via the building control system and, in addition to this, there is an energy demand reporting system in which the user in the plant reports his demands, which are outside the normal production program, to us. According to the energy demand report, the system is then operated by us via the building management system, via the control technology.

Such a system would then be a ventilation system, for example?

Rainer

Yes, there are many ventilation systems. Due to their size, the ventilation systems naturally also represent a high proportion of energy consumption. And here, of course, how the production runs and a precisely controlled use of the ventilation system are enormously important in order to save energy accordingly and to be more energy efficient. What we also monitor is our weekend energy usage. We also have these in the monitor so that we can see that we are also switching off all consumption as much as possible at the weekend in order to achieve the lowest possible energy consumption. We do regular master weekends for this purpose. That is, we determine the minimum energy consumption in a building over the weekend. We set this value as a target and it is then monitored on a weekly basis and sent to the relevant areas to check whether we have reached the target or why we have not reached it. What measures need to be implemented to achieve the goal?

And these ventilation systems are then also already monitored accordingly via your building management system and alarms for limit values etc. are probably already stored there, right?

Rainer

Yes, we have also stored threshold values in order to recognize in advance when a plant is running in a certain limit range and then to be able to intervene in time. A very simple example is a filter change of a ventilation system. This is monitored via the differential pressure in the ventilation system and, depending on the differential pressure, the filters are then changed in good time.

You are really very innovative and structured. You also said earlier that there are 70,000 data points in the building management system, which are already available to run the energy analyses. You had talked about a power demand system that also monitors on weekends. Here, yes, we are still classically at the plant manager level for controlling and monitoring the facilities at your site. We’re not necessarily talking about IoT there yet. This is then the next step, so to speak, in which cross-trade or cross-site networking takes place and added value is leveraged there. I would also like to talk about that now. Sebastian, maybe asked in your direction. Now we are talking about a wide variety of data. From your point of view, what is the classic data from these various systems that are now also moving in the direction of the cloud, for example? And what technology assets do I even need to have in order to get a digital solution like this off the ground? And can I process all the data there? How does it work?

Sebastian

Basically, we can process all data from any plant, including from all trades. So we are not trade-specific, but can go beyond individual trades. We can do this using a specific example, for example, a special ventilation system that supplies a production area. There is a humidity sensor, there are humidity setpoints, control signals from valves, from fans, from dampers. In order to understand the entire interaction of the individual units in this overall system, we need the data from each unit, which then also includes the pressure setpoint, pressure sensors, but also the entire topic of air quality and volume flows. If we now also think of the cleanroom issue, for example. We can take all this data, so to speak, process it and then integrate it into the cloud platform.

Now you just said process in the cloud platform: What intelligence do I have to bring into these individual systems or into the cloud in order to leverage these various potentials that we are also discussing here right now? What intelligence is in there?

Sebastian

Once we have the data points up on the cloud platform, we don’t go and check just one incident based on that data or try to identify just that one incident, but we look at it holistically. With the help of all this data – in the case of a ventilation system, there are about 50 different data sets or data series – we check precisely the topics that Marc has already mentioned, and also continuously the system performance, the energy efficiency. For us, energy efficiency is a continuous malfunction, so to speak. We continuously check whether the plant could not have run better or whether the plant performance is really 100 percent as expected by the production behind it. In addition, there is the topic of predictive maintenance, which are, so to speak, the three aspects that we always check with this data. We have integrated the relevant know-how into the cloud platform with rules over several years, so that we can continuously monitor the three criteria on the ventilation system from day 1 with these several hundred rules accordingly. This starts, for example, with a simple setpoint violation, that production conditions are not met, and then it is also checked under which conditions they are met and under which they are not met. What does the interaction of the supply air temperature, the temperature brought into the room, have to do with the heater, with the cooler, with the outside temperature? When does this setpoint violation occur? Or also, for example, a continuous deterioration of a plant, because a heat recovery wheel slowly clogs up, due to dirt, perhaps also from the production area, and the efficiency becomes worse and worse. At what point does it really make sense to do a cleaning and track this deterioration? The last example is the subject of vibration behavior. Perhaps we have continuously complied with the conditions of production. So the setpoints all run through very well, are all theoretically good. We don’t get complaints from production management, but simply see an increased demand for energy or that equipment is breaking down faster. This can be attributed to oscillatory behavior, for example, because we continuously open and close the valve and, so to speak, always overdrive it a bit, then understeer it again, and the system is perhaps too fast or too slow in some parts. We don’t get that if we only put up one or two data points. But then we take all the data and we can compare that with the optimized twin of this plant and continuously check these disturbances in energy efficiency, maintenance and user behavior, i.e. the production conditions.

Thank you for the examples. I think we have also understood quite well here the advantage of the interaction or comparison of the data, even across trades, in the cloud. Now, of course, the question is once again for you, Rainer: What do you expect from the topic of IoT for your location or what are the advantages from your point of view? Maybe looking towards the future or even today.

Rainer

Not only do I expect something from IoT, but we are also using it to make it easier to access data from different systems. For us, an important criterion is the number of vehicles, for example, because our key figure is energy consumption per vehicle produced. And of course, the presentation and calculation of such a key figure with access to multiple data in the Internet of Things or the entire data world in general makes the whole thing easier through the evaluations and the calculation of forecasts.

There are probably a lot of different trades that come together. You also said at the beginning how many different buildings you have on site. That is probably also an exciting topic across the board, to network them with each other and perhaps also to leverage synergies and added value there, isn’t it?

Rainer

Yes, it is very important for us to be able to access control systems in production as well, because a lot of data is stored there. For example, we are already using the possibility of linking the production control system with the building control system in order to operate our systems based on a shift-break calendar in production or to control the building control systems based on the start-up times of systems.

Last question: How do you see the future of your buildings, if you now look five years, maybe even ten years into the future? And what technologies do you think will be increasingly used here?

Rainer

I believe that in the future, room air quality-controlled operation will play a significant role, i.e. that we will control our ventilation systems in the buildings depending on the quality of the air, for example via CO2 sensors.

I am also convinced that IoT, as I have already mentioned, really highlights the added value through this overarching intelligence and through easy access across system boundaries. Another exciting aspect is that the quality of indoor air can play an active control role here. Marc, now another question for you. We had now talked about a wide variety of systems. Can you explain what the difference is between what is already done today via the building management system compared to IoT and the intelligence that is brought in here? What exactly is the difference here? How does it work?

Marc

The exciting thing is that the systems are often connected to a building control system or an alarm management system. What I mostly monitor there are classic thresholds, malfunctions and alarms. So I would say, if the temperature exceeds the value, please give me a message, classify it as a malfunction, because that’s not optimal, not the way I actually want to run the system. That’s when I have to react. I can do exactly the same with any value. Malfunctions, classic threshold violations lead to an alarm, to an event. And that’s what I react to. But let’s look one level lower. Now, in a data analytics, you would go here and say, are there certain patterns that I can make into rules? So can I look at the plant – very generalized, so that it’s not just right for one customer, but fits many customers – and are there patterns and rules there that I can monitor permanently? And that’s exactly the part we chose. Classically, if you had such a predictive maintenance topic or use case coming up, or if you had a use case, you would go here and say: Okay, I have to have the sensors, these sensors, those sensors, I still need the value, and in addition I still have to record an event of the malfunction or the downtime. And if I have enough data, then I can use certain analyses to figure out what the pattern is. And that pattern I would then check back and take a set point and when that’s reached, I would check that. We are talking here about fault detection and diagnostic approaches. And then I would check exactly these control parameters. That’s what we did. In some cases, we have gone so far as to say that we do not want to regulate a single device, but rather the entire system functionality of the many complex individual devices. In the meantime, as I said, there are several hundred rules and these rule modules are constantly growing and thus we are checking the entire system.

If I now stay with this ventilation example – what do I have to bring along as a customer for these rules? How do I get that expertise in there?

Marc

Actually, we bring the expert knowledge. Through a large number of customers, we know what is considered universal. Let me give you an example, a very simple rule: If the ventilation system has been running at 100 percent for 24 hours, then something is wrong. Now I can also make this rule dependent on other parameters of temperature curves or production conditions. Normally you would say, yeah sure, if the ventilation system reaches a threshold of 100 percent, give me an alarm. But that’s not true, because the ventilation system should – no, it even has to – be able to run at 100 percent, but it also has to switch down again. Only when a time factor is added, as in the case of 24 hours, then something is wrong. And then, of course, it’s a matter of responding. But that’s what they’re trying to narrow down more and more. And this rule, which is universally valid, applies not only to customer A, but also to customer B and C. And so, bit by bit, rules grow. And then the customers keep coming up to us and saying: But I have something of my own that I want to test. Either they check it themselves, there’s the possibility to find out by visual data analysis, does something correlate, are there any irregularities or maybe even regularities that I don’t want to have? Or you say, I want to have my own rule. I have a special condition here, I would like to check it now. And then, of course, we can build up this rule and bring it in on a customer-specific basis. Quite often, of course, things come up where we find that this is not just applicable for one customer, but in general. And then this rule is basically made available, is available in the general rule set, and can actually be applied to any customer immediately as well. That’s actually what’s exciting about it. And that’s why this set of rules is constantly growing and getting bigger. Of course, you also have to look at the plants – not every rule will always apply, but at least we have a very, very large rule set and every customer, every use case at the customer leads us to improve the system further.

The bottom line is that it’s like a catalog of measures that I can use to say, I want to save on operating costs, I want to scale my plant availability, and I also want to improve energy efficiency – and all that in cooperation with the customer, right?

Marc

Yes, and what’s really exciting is that the customer often comes to us and says: I have an issue here, but I can’t figure out what the problem is. And then, without further ado, you can also set up a rule, check that, and in the end determine what the cause of that is. For example, we’ve also had heat recovery systems where someone had an indicator and said: I feel like it’s not really working, but I don’t find a problem either. And we ultimately found out with rules that it really didn’t run as it should. And during the in-depth analysis, it was simply found that it was not properly designed at all. The catalog of measures then states: Dear customer, for an additional investment you get energy-saving measures and ultimately the right control functions.

I would like to go back to the individual trades. It is now also going so far as to go into other stakeholders who may also take on a service function to this building there. Are there approaches you’re seeing there, too? So to work with externals and also provide this data across the board? Are you thinking along the same lines?

Sebastian

By having a cloud platform, of course, we have two options. Once, of course, on the import interface, that we can import a wide variety of trades there. Either because we have the same trade decentralized twice, but a decentralization, or several locations or different trades up to other data interfaces perhaps from the German Weather Service, SAP systems, etc.. Because it is a cloud platform, we also have the option of making the data available to others. That means companies have already pulled that data back out of the cloud platform via the API interface and made it available to users in mobile apps, for example, or even on other websites. All this in a prepared form, in a qualitatively higher form, because all this was checked and put together in one place and a data plausibility was made. For appropriate reporting, this data is mirrored back into an SAP system or maintenance tools, for example, to provide fault messages and alarms. Of course, the cloud platform offers this in perfection through the two API interfaces – one for import, one for export.

Ultimately, that is also the essential advantage of ensuring that the whole thing is truly cross-location and having open interfaces there. At the very beginning, we talked about the different building tasks, and now we want to close the circle a little bit and come towards the end. Marc, once again in your direction: You are now active with different customers. When will this system be used in general? Maybe on a different use case as well, how are you guys going about that exactly?

Marc

The application of such a monitoring system can be found in many different environments. Everything that has to do with security in the building, it is important to react in time, to react quickly, to react in an orderly manner when a certain event has occurred. And by that, we don’t mean the actual security case in the sense of – someone wants to get in there who’s not allowed to get in there – but actually the technology that has a malfunction or a problem. And then there must be an organized, fast response. I always say so nicely: whenever there could be a critical event where you need to react 24/7, around the clock, in an organized way, then this system can be applied there. And so you can say, there it’s always a matter of looking at the applications, what ultimately is the problem and that’s where I apply that.

Thank you very much. I found this use case, which we discussed in detail today, really exciting. Thanks again to you, Rainer, it was really exciting to see in practice what the challenges are here. Thank you so much for your input and I look forward to next time.