How can we use AI to optimise modelling? This was the central question that Dr Henning Oppel and Dr Benjamin Mewes, founders and managing directors of Okeanos, addressed as part of their doctorates at the Ruhr University Bochum (RUB). Oppel in the field of hydrology, i.e. flood forecasting; Mewes in the field of water management, i.e. with regard to distribution issues and resource planning. AI solutions in water management were "a novelty that had not yet arrived on the market at that time, 2016/2017", recalls Oppel. 2019 saw the joint spin-off. Since then, Okeanos has been developing digital applications in which engineering science and information technology converge and which - deliberately designed to be platform-independent - can be integrated into existing IT infrastructures. "In short, we develop AI-based software for the water industry. We try to solve problems."
One of these problems is early flood warning on third-order watercourses, i.e. smaller rivers and streams that are not the responsibility of the federal states or the federal government, but where local authorities have to take action. "As a rule, there is no monitoring and no gauges there. On the one hand, because it's quite expensive. Secondly, because gauges also have to be operated, which means there has to be someone who uses the information - which is then only valid for one spot - to make a classic flood forecast." The result: there is no early warning system. A shortcoming with dramatic consequences. "During the flood disaster in 2021, for example, enormous damage was caused to the tributaries of the Erft. There was no forecast, nobody was responsible. And that's exactly where we come in. And we are starting at a very early stage, because we are already measuring the development of floods with 'Netilion Flood Monitoring'."
The method developed by Okeanos combines data collection and AI-based forecasting. But what is special about it is that "we don't start the measurement in the water, we start by measuring the precipitation. We measure the soil moisture and then the water level. In other words, we measure at different points and in different areas in the catchment area of a stream. We measure that it is actually raining there now, and we see when something happens in the soil. And we can also see when the water reaches the stream. It is precisely this combination of data, which converges in our cloud and is analysed by our AI, that enables our solution to provide early warnings of flooding. And this is even possible in very small or very fast-acting bodies of water."
On behalf of the Emschergenossenschaft (EGLV), the Okeanos system monitors the Borbecker Mühlenbach in Essen, for example, and reports potential dangers directly to the EGLV flood control centre. "The catchment area of the stream is basically the car park in front of Messe Essen and the Grugapark. We now have 45 minutes' warning time here, and physically that's as far as we can go. It starts to rain, and three quarters of an hour later the water is down." 45 minutes lead time to clear the site in an emergency. A huge advantage in a situation where every minute counts. "We can distinguish between 'It's raining' and 'There's flooding'. And that goes beyond the performance of many other solutions. We can see the flood as it is forming, and that gives us a time advantage that other systems don't have."
‘I have yet to find a better location in Germany for water management than NRW, than the Ruhr region.’
Founder and Managing Director okeanosIn order to provide the most accurate data possible, the AI has to learn the characteristic hydrology of the area of application, meaning the algorithm has to be individually trained. "That takes about a year." The data is collected on site using IoT (Internet of Things) measurement technology. In other words, using self-sufficient devices that do not require power or an internet connection, but already have everything they need "on board" thanks to a rechargeable battery and SIM card. And they are virtually pocket-sized - "the water level sensors are just the size of a hand". This has important advantages: "No cables need to be laid for the installation, we can simply position the devices in a well concealed place. This ultimately also reduces the cost of the system, because as soon as something has to be built, especially around bodies of water, it becomes very expensive," he adds. And we don't want to cover the streams and rivers with concrete."
(Kopie 11)
Netilion Flood Monitoring has officially been on the market since the beginning of this year. Through various projects in pilot municipalities, the experts in Bochum have tried to incorporate as many different regional conditions as possible - from the mountainous landscape of the southern Black Forest with its natural streams to the urbanity of the Ruhr area. There is a great deal of interest in the development, says Oppel. "Especially in the catchment area of the Erft, in Ostwestfalen-Lippe, where flooding plays an increasingly important role, we are being met with open ears." The start-up's team has now grown to 17 people. Experts from the fields of environmental engineering, environmental sciences and computer science, whose agenda currently includes topics such as shortening the AI learning phase and differentiating the forecast, for example, which roads are likely to be affected and how high the flood risk will be.
The majority of the team studied at the RUB, the University of Duisburg-Essen or Dortmund University of Applied Sciences - and Oppel is proud that Okeanos, as a young company, is able to retain skilled workers in the region. This is partly due to the company structure - "I'm under 40 and I'm the oldest here" - but also due to the subject matter: "We offer an alternative to the traditional career paths for an engineer, particularly in the environmental sector." Last but not least, the Ruhr metropolis itself can also score points: "I have yet to find a better location in Germany for water management than NRW and the Ruhr region." The region has been dedicated to the topic of water since industrialisation and therefore has the corresponding entrepreneurial expertise. "The population density and the challenges here are relatively unique. The Ruhr region is therefore a very good place for people with new ideas.This has also helped us to establish such a good standing so quickly."
The fact that the state of NRW has transferred responsibility to different water boards is also helpful, so that tasks can be tackled "river by river". "This means that there isn't one single body that doesn't know where to start, as we see in other federal states, but there are separate competences in each case. And that alone gives us the scope to try out new things and experiment. After all, this is the only way to make progress on these important issues."
Cooperation in development
As a young start-up from the Ruhr region, Okeanos has an established, internationally active partner in the field of measuring devices and level technology at its side in the form of the long-established company Endress+Hauser.
You mind also find this interesting:
