What does the SAP Cloud Platform have to do with flying robots and mechanical inspections of wind turbines? We talk to SAP solution architect Helge Hackbarth to find out.
Mr. Hackbarth, as a SAP Cloud Platform specialist, you are currently involved in a project called “automated inspection of offshore wind parks.” What is that project all about?
We want to fully automate the mechanical inspection of an offshore wind park’s turbines using drones. The drones fly to the turbines, capture close-up images of the rotor blades, return to their starting point and deliver the data they have collected. The whole process is fully automated through intelligent interaction between the drone and the control station. All the human operators have to do is give the order for which wind turbine the drone should fly to and monitor the mission on a dashboard. Of course, the operator can manually intervene in the process at any time. The drone is controlled centrally via the SAP Cloud Platform, in which we build a digital twin of the drone that includes all of its technical properties.
A drone mission sounds like a lot of work. What is required of the drone to ensure the mission is a success?
Because the drones make a kind of solo flight and come very close to the turbine blades, it has to be possible to position them with precision down to the centimeter. Another reason that’s important is because the photos have to be completely sharp and in high resolution so that the AI systems can analyze them correctly. Of course, the drones also have to be resistant to weather conditions at sea, as they will be deployed far from the coast.
So far, so good. What happens after this technical photo shoot – how do the drones process the images?
The drones themselves do not process the images. They take around 2,000 photos per flight and transmit them to the SAP Cloud when they return. The AI that has been implemented in this cloud and trained on a large number of images analyses the photos and classifies them based on the type of damage they depict, such as erosion or cracks, as well as the extent of such damage. The SAP Cloud Platform pays a key role in subsequent data analyses, as its SAP HANA in-memory database allows for real-time analysis. The AI then generates a damage report that includes all key information about the position of the damage and the accompanying photographs. Humans make the final decision on whether or not a repair is needed.
That sounds very complex. How many people are preparing this project to ensure that everything goes smoothly when it is put into practice?
There are currently three of us. One colleague is coordinating the project and takes care of communication with the customer. I ensure that all the technology used complies with the requirements, and I am responsible for automating the mission – in other words for how the images will be generated and for ensuring that the data makes its way into the cloud securely. The third person in our group is an expert on automated image evaluation. He ensures that the AI system correctly analyzes the photos in the SAP Cloud.
“The SAP Cloud Platform allows for real-time data analysis through its SAP HANA in-memory database.”Helge Hackbarth, SAP solution architect
AI seems to be an important part of the overall solution. When it comes to image evaluation, how does it outperform humans?
When it comes to capturing and analyzing images, AI still can’t quite beat a human. At the moment, an AI needs to see a huge number of examples in order to be able to learn something independently. Humans go through a learning phase too, of course, but they are currently still better at recognizing objects, even in an unfamiliar context.
AI’s strength is in evaluating images within a clearly defined context such as this wind turbine project. Before the drone begins undertaking productive missions, we will train the AI system using a large volume of photos that have been taken by prototype drones at wind turbines and categorized by humans. That way, we can teach it to recognize damage. The advantages of using AI are clear: Unlike humans, it doesn’t get tired, even if it has to go through tens of thousands of images. AI’s fields of application are expanding exponentially. I think we are only seeing the beginning of what AI will be capable of in the future.
In addition to AI, the cloud is a key component of the project. Why is Lufthansa Industry Solutions using the SAP Cloud Platform to evaluate the photos?
The German software developer SAP, based in Baden-Württemberg, takes data security very seriously and guarantees data protection in accordance with European guidelines. Besides that, the SAP Cloud Platform is more than just a development platform of the kind that other providers offer. What makes it special is its array of additional technical and business services, a programming model that provides sensible existing guidelines, and straightforward options for integration with back-end systems, particularly with a digital core such as S/4HANA. Lufthansa Industry Solutions and SAP are connected by a strong bond of partnership. Our decades of cooperation have given us countless high-quality contacts to SAP. In addition, our drone project is part of the global SAP Co-Innovation Lab network in which SAP supports innovative ideas by making its own expertise available, and we have a direct line to their in-house development department.
Back to the drones – did you work with unmanned aerial vehicles before this project or is this completely new territory for you?
I developed a passion for aircraft as a child. I built my first model aircraft when I was a teenager. Drones have become very powerful at a technical level in recent years, and as a “maker” I can manipulate them in almost any way imaginable. I also develop the software largely by myself to adapt the drones and their payload to specific applications. That is why I am all the more excited that drones are increasingly a part of my working life too.
About Helge Hackbarth
Helge Hackbarth is an SAP solution architect based at the Lufthansa Industry Solutions location in Hamburg. After obtaining a diploma in mathematical economics, he initially gained experience with SAP at a consulting firm before switching to the Lufthansa Group in 1997.