Digital twin technologies in action
Real Conversations podcast | S5 E14 | July 13, 2023
Mirjana Scheele a deploy services expert holds a Master of Science in Telecommunication engineering from the University of Belgrade and joined Nokia in 1996 as an installation and commissioning engineer and has acquired, since then, broad international experience through various service roles in Belgium, France, China, and Germany. Mirjana is very passionate about digitalization in deployment services.
Digital twins have numerous surprising benefits for the deployment of networks. Mirjana Scheele uses her extensive experience to explain more.
Below is a transcript of this podcast. Some parts have been edited for clarity
Michael Hainsworth: Deploying 5G involves more than just climbing a tower and pointing at a transmitter. Artificial intelligence, machine learning algorithms, drones, and digital twins are just three key technologies that come together to reduce the time it takes to build a network by as much as 35%. As we leverage digitalization further, further reductions in cost and time are expected. Nobody knows this quite like Mirjana Scheele. She tells me it's been a 25-year-long journey to the top of the metaphorical and literal transmitter tower.
Mirjana Scheele: Deployment services start at a point in time when the site needs to be found and the telecom and radio equipment has to be installed and it is finished at the moment when this equipment is fully in operation and it's in commercial use.
So that is the world of deployment services where I am. I entered this world 27 years ago now, and it has really been a journey of meeting extraordinary people, visiting wonderful places, and learning things which I had never thought that I would have the opportunity to learn.
MH: Well, tell me about that. Because it must have been quite a journey.
MS: Yes, it is. It is a journey, and it started like this, "I'm going to be at the top of my professional career," working in level-three support. Because I have seen at that point in time, as I just graduated as an engineer from the university of Belgrade, from studying electrotechnics and the telecommunications specialization, that those people who are solving the problems what we are having in the field are really those who are on the top of the world. I managed, in less than four years, to end up there, and then recognized, "Well, what am I going to do next?"
And the journey continues. I have continued taking and assuming many different roles in different countries, traveling the world over these past 27 years, working on different continents.
MH: The digitalization you must have seen over the course of the last 27 years would be remarkable I can imagine. How do you define digitalization today?
MS: Digitalization can have many different meanings, actually depending on the context. For me, and then generally speaking, digitalization is a process of converting something into a digital form. In the telecom industry, and especially in deployment services, we are converting all types of information and pictures into a digital form, which is necessary for us in order to improve the business outcome, in order to assure the best customer experience, in order to increase operational efficiency, which enables us to grow simply and give us a competitive advantage.
More concretely looking into the digitalization of deployment services, if we think about why, it is so needed and what digitalization is bringing to us, we can see it brings us speed. We are doing things faster and with better quality. It helps us to reduce the costs and investments for customers. It is helping us to address environmental sustainability challenges and it is definitely helping us to improve all the activities and all which are related to health and safety. So that is the meaning of the digitalization of the deployment services for me.
MH: And for you, what role do digital twins play in your day?
MS: The digital twin from the Deploy Services point of view, is actually a digital 3D replica of the sites where our radio equipment is being installed. And it means a lot to us in this industry. It means that we are becoming able to do things very fast because this digital 3D replica is available anytime, anywhere and it's online. It helps us to open the world and gives us the opportunities to do things in a better way, in a more sustainable way, cheap way, and to do things faster.
MH: What's the power of digital twin technology when deploying radio access networks specifically?
MS: There are many opportunities where digital twin technologies can be utilized and many more ways which are still to be exploited and then explored. For example, when we are doing a site design and we need to do this site design in the most optimal way in order to use space most effectively. Then also to reduce the utilization of any installation material and investments which are associated with it. The role of the digital twin enables us to have the best site design and achieve the most optimal way to calculate the material for installation. Moreover, based on this digital site twin, we have the capability to create a so-called automatic bill of material (BOM).
First, we are measuring all the distances digitally. We are looking at everything that is necessary to be known and then we are automatically creating a bill of materials. This helps us to significantly decrease any kinds of mistakes in this process and to automate other activities in the deployment services in material management. It plays a role in better estimation of the efforts of the work, which is necessary to be done and, respectively, costs. Additionally, it is having a positive impact on the environment from a sustainability point of view because it's reducing unnecessary site visits by 25%.
Everything we need to know about the site, we have it in digital form, and we can look at any point in time. In the past, what has been happening, these site surveys are being done and the information about sites is being documented in the way that we send somebody who goes on the site and who collects this information on a piece of paper. And that's the way I even started doing my job 25 years ago. And it has been happening that when the person comes back to the office, exactly one detail which is really needed to make a particular decision is missing. And then if you don't have the status, then you need to send somebody to go on the site again.
And when you think about the distances between the sites and how remote some of our sites are, then we can imagine how difficult it is and how cost-consuming it is to do it and time-consuming and so on. So it is helping us to reduce the amount of the site visits by 25%, and then additionally also to be better prepared for any kind of work on the site from the health safety point of view, so we can assess better any kind of hazards in order to align with the field crews which are going to work on the site, prior to going on the sites, for the particular hazard which are typical for the site.
The opportunity is immense between these digital site twins and the digital site content of the field. Additionally, to mention, I mean, with these digital site twins, we can also calculate exactly what the positions of the antennas are from the Tilt plane and azimuth. We have algorithms which calculate, "What is the height of the antenna? What is the size of the antenna?" And so, it is also helping with the natural planning optimization. I probably need to stop here. I'm listing all the opportunities which have been made possible by the digital site twin, but we are not yet exploiting all of them in this industry. By far not, so they're really immense.
MH: I could imagine though that you are exploiting the benefit of having a digital twin once you've deployed it at a specific site down the road in the future for things like upgrades. You don't have to do the work all over again.
MS: Absolutely. We call this digital inventory of the site and this digital inventory; it's really becoming available for anybody in the lifecycle management of the site and the equipment of it for any kind of activities which are happening later on. It is much easier to prepare for the upgrade, much faster, it is much more optimal to have this content at any point of time available to anybody who is going to need it than it has been ever before. Absolutely.
MH: If you could reduce site visits by 25%, I can imagine you could shrink the time to market to get a site deployed. Can you quantify how much digital twins are shrinking in that time to market?
MS: Yes. Only by using this digital site twin, which is actually an outcome of digital site surveys, and then the site design. We can reduce the time of go-to-market for the total life cycle of the deployment by up to 10%. 5% is probably directly related to the reduction of the efforts and the speed in getting the site design and then associated efforts related to the permits. And then the other 5% is related to everything else, prevention of mistake correction, availability of the data at any point in time. It is only this digitalization of the site design, and non-utilization of the digital site twin on the site design, that is giving us the potential for a further reduction of the total deployment cycle of up to 10%.
MH: And what does that 10% translate into actual time? Are you shaving days off from installation? Months off an installation?
MS: If we talk about digital site design alone, we are shaving off days. We are probably currently estimating the direct reduction of around two weeks of work. That is this direct impact. And then there is this indirect impact of probably another two weeks of the work of the different parties. But overall digitalization is impacting the time of total development significantly more.
MH: So, sounds to me like it's a month shaved off the time to market and that's just one individual deployment. If you're rolling out the network wide, that could be a remarkable improvement in how long it takes to get a network up and running.
MS: Absolutely. The return on investment, for customers, having the ability to put the equipment they are buying into operation faster, is significant. For them, for every site, it’s like one month of additional revenue. Or getting revenue one month faster in comparison to the past. And this is only by the digitalization of the site design using digital site twins. There are even further improvements which are helping reduce the time required for the deployment, allowing the customers a faster return on investment and what they can do with equipment.
MH: And I think it's important as well to point out that we're not just talking about communication service providers in the fourth industrial revolution. We're talking about rolling out these kinds of networks for industrial customers and enterprise businesses with private network needs. They're going to have to go through this entire process as well and this is what you're helping to ensure; that they get their side up and running faster than in the past and, therefore, they're saving money and have the ability to make more money down the road as well. So, there's a profitability component to this.
MS: Absolutely. Perfectly said. It's not only cost reduction, but also about revenue, getting it faster, cost reducing it all together and then bringing the services to the end users faster, because everything is a positive benefit down the road. By doing these things, we are helping our customers to get the equipment fully into operation as fast as possible.
MH: Do you have a favorite example? Because artificial intelligence and machine learning are being leveraged, I'm just fascinated by how AI is being used today.
MS: I do have a couple of examples and maybe let me take one or two which I consider extremely impressive. These are examples of one of the biggest operators in the world and it's one of our biggest customers in India, actually. What we did with that customer, and we started a few years ago, was to significantly reduce the efforts and the time of the acceptance processes. So how did we do that? For every customer, when we are putting the equipment into operation, there is a certain checklist. From a quality point of view, we need to fulfill that and then what are the proof points that we need to give to the customers that this is being fulfilled so that they accept it and then they pay for us and, obviously, the equipment goes into operation. We sat down with this customer, we created this quality checklist, then we translated this quality checklist into the digital form. We trained our machine learning algorithm to go through these quality checkpoints, which are supposed to be documented through the pictures.
And we gave access to our field crews via a mobile phone application of our platform on which this machine learning is working. And this platform is called Nokia Delivery Platform for Deploy (NDPd). So that when our field workers, or our suppliers' field workers, are doing the work on the site, they directly verify, as soon as they finish installation work on the site, all these quality checkpoints. And then they run this algorithm on the pictures that they have taken, and then they get the information directly. "Is it okay or it's not?" Once they've finished this, when this algorithm has run through all the checkpoints, it is just being delivered to the customer. The accuracy of that one is over 98%. We have developed this over time. To that extent that the customer is not even looking at every single site, every detailed acceptance documentation, they're just looking at the result of the algorithm and then randomly checking one or two sites from the certain period in time to see if everything is really like what they have been expecting to see.
In comparison to the past, there is an unbelievable improvement in speed and quality. Everything is being done digitally, everything is being done automatically, so a process that used to take several weeks, sometimes even months, has been reduced to a few days. This has been done on a massive scale and is already in use. That's one example. And I have another very interesting example that goes in the other direction of the machine learning application, for health and safety.
In our industry, we are dealing with many different types of risks. We have people who work with heavy equipment, lifting cranes, they climb and when they're working on the rooftops, they drive long hours. We actually have any kind of risk that this industry can offer in the deployment services, and we pay immense attention to health and safety in order to make sure that we prevent any accidents to the maximum possible degree.
We have machine learning algorithms which help us to do that. For one really large operator in Indonesia, we are utilizing these algorithms in several different areas. For example, verification of personal protection equipment before people goes to do any kind of work on the site. Then even the verification of the fire extinguisher equipment on the site, and even the verification of the profile of the tires on the cars before the teams start to drive on the site. And this helps us to minimize any kind of incident to the maximum possible degree and we are developing this further in new cases and continuing to use it in the future on a massive scale for many other customers globally.
MH: I'm fascinated as well by how much your industry has turned to one of my pet projects, which is drone technology, and tied it into artificial intelligence and LiDAR, when designing the site itself. You're tying this all into the digital twin as well.
MS: That is all coming together, and it is bringing it into this digital site twin and how we are using this technology there. So maybe one of the use cases for the application of machine learning given is this creation of the site designs, or the digital site twin, or so-called machine learning for hardware recognition. When we create this digital site twin, first we do a digital site survey, which is good to do. It's not a mandatory precondition, but it's much better if it is done as a digital site survey. And in this digital site survey, we can do it either with drones or with drones and the LiDAR cameras or with the LiDAR cameras alone. Obviously, the advantage of doing it only with drones is that people don't need to climb, so then there is a huge health and safety advantage as well.
Sometimes, if, for whatever reason, we cannot use the drones, then we can use the LiDAR camera, but then climbing is necessary. Then we collect all the digital data, pictures and videos of the people who have been working on the site. They actually send it automatically, even on the time as soon as they're finishing on the site. Then this is exported from one platform to another one on which the machine learning algorithm for handset recognition is running. And then this algorithm automatically identifies which objects have been found on the site and it is creating this object in this digital site design automatically. Not all the objects, in all the third parties, on all the sites in this world we can recognize them this way. We have trained this algorithm for our own equipment, and we are bringing on top of each the equipment of the other most common third-party vendors.
If we don't have this equipment already being trained in the algorithms of other third parties, we still do have our center of excellence, people who are capable of identification of objects manually and they're putting it into this digital site design. This object hardware recognition probably requires something between 300 and 400 pictures of the certain object in order to recognize it with an extremely high level of accuracy, which is acceptable. This altogether is also helping us even to reduce further the time which is necessary for the creation of the site design based on the digital sites survey.
MH: It's amazing to think that you can bring all this technology together to do something as simple as figure out how long the cables need to be. Being able to establish a bill of materials and have it created automatically, I can imagine, contributes to reducing that time to market. How do you manage material tracking?
MS: The material tracking is a very interesting part. In the past and not that long ago. I mean also in the recent past. This has been one of the really big challenges in our services, and for us, and for the customer, because there are a lot of bits and pieces from the different vendors which are ending up on one site and have to be there in order to finish the proper installation of the site. Sometimes it's just enough that one cable is missing and then you can simply finish it, and then you need to go again, and you need to trace where it is, or you need to buy it and so on. This problem of this material tracking we have resolved by the utilization of so-called digital platforms as well. The digital component of our platform is NDPd, which we call Material Management.
Through this platform, we have access to all our warehouses and we understand, in a particular country and for the particular project, at any point of time, what we have in our warehouses, so that when the time is coming that we need to do some work on the site, we can access this inventory and decide, "Do we need to order the equipment or not? Or you just take it out of the warehouse." When we see that we have the equipment which is needed, we do the so-called ‘call-off’ of this equipment. For that particular site, and for that particular configuration, and then for the particular field team or supplier team which has to do work on the site. Then we either deliver this equipment to the site or we ask our field team to pick it up. Obviously, when the equipment is being picked up, it is scanned, and it is in our system automatically.
And equally, when it is being delivered on the site, there is this barcode reader capability. We scan it and we automatically upload it to the system where it is being delivered. And we are using this in some projects, and we have already seen the benefits of that. We have had a couple of cases where suddenly we recognize, "Well, the equipment is not on the site, the team cannot work," but we know that the equipment has been picked up by the team. And then the discussion started, and they said, "Hold on, let's just have a look at the tool. Where is it?" We looked at it, we found it and we could solve it in a matter of minutes instead of a matter of days like it has been in the past. This material tracking is not such a problem anymore thanks to digitalization today.
MH: There's that famous quote, "The future is here, it's just not evenly distributed." Are all customers ready to deploy digitalization and digital twins?
MS: Excellent question. And the very simple answer is no. There is much more complexity behind that simple answer. Every customer has a different story, and every customer has different challenges and different opportunities. They have to deal with their environment and their situation in a different way than what they do. Obviously, the interest in digitalization is everywhere. Every customer has a high interest in digitalization. Some of them are trying to do things alone, but the investments that larger companies are making in this digitalization in comparison to the investment that some of the smaller companies are capable of making are not really helping them to progress very fast.
Any change is actually coming with discomfort, even if positive changes. And this digitalization, although it has been strongly supported by everything that has happened over the last couple of years with COVID, is still bringing a certain level of discomfort. Then, it depends on the particular situation of the customer; the readiness for the change is higher or it's very high or it is even more. Still, it has to be a tailor-made approach. There is no one size fits all. It has to be an approach which brings interdisciplinary knowledge in order to make it really fit for the particular customer and be successfully implemented.
MH: You've had a 27-year journey to get to this point today and you've learned and seen a remarkable amount that we couldn't possibly capture in just our time together today. Tell me though about what you will see for the next 27 years. What's your vision for the next big thing?
MS: Well, 27 years is a very long period of time. I'm not sure if I'm going to be brave enough to think about that. Probably not. But for the next couple of years, probably three to five years on the horizon, what I see is utilization of the metaverse in the different areas. The metaverse, in the sense of the metaverse, is what we have already. Our digital site twin is nothing else than the metaverse of what we are doing in the industry. But the implication of the metaverse may be more for training, some installation work, more for the material testing and so on. So that is one of the areas which definitely is progressing very fast.
And another area, again, is interdisciplinary. I am getting much closer to the telecom industry and the civil engineering industry. Preparing everything that we have from the digital twins, which is Building Information Modeling (BIM), for building information modeling, which is massively in use in the civil engineering industry, and helping us through the utilization of the BIM to do more digitalization in the precincts arena in order to be much faster in responding, or even automatically responding on the Request for Quote (RFQ)s, utilizing all the digital data which we have.
Then further improvements in health and safety. What we are dreaming about is having absolutely no incidents of any kind in the work of the deployment. And that is our goal and objective, and we have some thoughts, and some developments and research which is going in that direction. "How are we going to make sure that we have the full visibility on the utilization of the personal protection equipment at any point of time, of anybody who is working on the site? And how are we going to make sure that we don't enable anybody to climb if it is not being properly connected?"
There are some interesting things that are in the research phase inside of Nokia there. And probably, if I look at the shorter term, which could be the period of three years, together we need to find a way to increase the penetration of digitalization. In this industry, particularly in the deployment services, we have much more today than the market is absorbing. The benefit of what we have, even if we don't do anything else, is going to be immense just by better penetration in the next three to five years. So that would be how I see the development in the next couple of years.
MH: Based upon our conversation today, if there was one key takeaway for the listener, what would it be for you?
MS: Digital twin metaverse digitalization is the only way forward for us in this industry to address current business and environmental challenges. I think this is the key takeaway and we are ready to help you. We are ready to help our operators to address the business challenges with digitalization, with what we have.