Alexandra Schaefer
Troubleshooting in Murky Waters – Alexandra Schaefer on Her Path to a PhD
All photos: Dajana Dopatka / TH OWL
When Alexandra Schaefer sits at one of her two very different desks and looks out of the window, all she sees is water. Water in basins, in manholes, in sewages, in giant screws, in pumps and filters. Water that stands still, that is transported, that is forced through screens and grates, to which oxygen is added, in which bacteria are at work, and which is ultimately purified. She sees how it arrives as wastewater and, in the end, is discharged into the Lippe River, purified. So, quite a lot of water wherever you look.
We meet Alexandra for an interview at one of her two workplaces: the wastewater treatment plant of Paderborn Municipal Wastewater Utility (STEB). We’d like to talk with her about her dissertation topic and her work as a researcher. But first, we’ll discuss Alexandra’s doctoral position, which is shared between two partners – an unusual model.
Researching in tandem: one PhD – two partners
Alexandra doesn’t just work at the utility; she also has a second desk, located at TH OWL’s Höxter campus. There, she is pursuing her doctorate in the Department of Construction and Environment under the supervision of Professor Martin Oldenburg. She is completing her three-year PhD as part of a tandem doctoral position, an opportunity offered by PROFuture: This practice-based PhD model qualifies early-career researchers for professorships at universities of applied sciences. It also allows them to later choose between careers in industry or academia. In Alexandra’s case, the practice partner for this dual qualification pathway is STEB, with whom TH OWL has already carried out many joint projects, including DIGIWATER.
Learn more about the additional opportunities PROFuture offers for early-career researchers and future professors.
“I’m really lucky with my position because I can work in both worlds: practical work at STEB and research at TH OWL,” says Alexandra Schaefer. She completed her master’s degree in Environmental Engineering and Modeling before beginning to study data science. “Working with data—with such a wide variety and such a large volume of data—just fascinated me so much, even back then. I wanted to explore that in more depth during my studies. But then I got the chance to go to a partner university in China for the CCWater project and do a three-month internship there. During that time, I realized: In the area of sensor technology and soft sensor technology (explanation), there’s still room to grow. There’s real research potential there, especially in error detection,” says Alexandra, adding: “And that’s exactly what I’m working on now—how awesome is that!?”
Identifying errors, calculating replacement values, reducing effort
Alexandra works on optimising measurement technology at STEB. The system is used to monitor treatment stages and measures key parameters using various sensors, including oxygen levels, temperature, pH, and turbidity. This ensures that the wastewater treatment process runs efficiently and complies with legal limits. However, sensors frequently begin to produce inaccurate readings, for example due to drift effects (explanation). Such drifts are often difficult to detect externally and therefore remain unnoticed for extended periods.
Alexandra is developing an identification for faulty sensor data in wastewater treatment plants and to generate realistic replacement values. To achieve this, she is building a hybrid model that combines mathematical formulas with machine learning methods. The resulting data will be displayed in the STEB operation control centre and may eventually contribute to the development of a digital twin (explanation), which could be used for simulation and to support future decision-making processes.
By the end of her doctoral studies, Alexandra aims to provide STEB with a practical application that significantly reduces the operational burden of the wastewater treatment plant. If faulty sensors can be detected earlier, they can be serviced and cleaned in a targeted way. This reduces the workload for staff and saves energy, materials, and maintenance costs. In addition, a reliable data foundation provides important input for digital twins, simulations, and AI-supported optimisation.
“Interaction and Structure Are Very Important to Me”
In this interview, Alexandra talks about her personal motivation, her daily work routine, and how she navigates challenging phases during her doctoral studies.
What interests you most personally about your field?
I really enjoy diving deep into complex topics and constantly learning new things—including new skills. I taught myself programming because I need it for machine learning. I’ll also soon receive more in-depth training in Python (explanation). I’m not a software engineer or programmer, but I’m particularly interested in hybrid models and model development. It’s always fascinating when I can translate what I observe into a digital model on a computer.
Important side question here: Do you like playing Minecraft?
No, not at all! I’m more into adventure games.
What’s the first thing you do when you get to work in the morning?
I get coffee, check my emails, and structure my day. I also think it’s important to quickly look at the news—what’s happening in the city of Paderborn, but also at TH OWL. And I always enjoy catching up with my colleagues to hear what’s new.
What do you tend to work on at STEB, and what at the university?
My position is very comprehensive: I can work on my doctoral research at both locations. At STEB, I monitor the process control system and prepare reports. I also give tours here and manage the Instagram account—please follow us! At the university in Höxter, I have access to the university network, which is essential for my research, and I also get to teach, which I really enjoy. In both environments, I receive strong support, have dedicated contacts, and the freedom to develop my research in a structured way.
What are the major milestones in your doctoral studies?
Right now, I’m working on understanding which types of error messages occur at the wastewater treatment plant. I want to build a comprehensive overview, which means analysing, evaluating, and clustering all available data. Which errors occur with which sensors? I spend a lot of time working through archived data, and the STEB team is extremely supportive whenever I need specific datasets.
Another milestone is developing a validated error detection method for a single sensor—ensuring that the hybrid model works reliably in that case. After that, I can explore whether it can be transferred to other sensors.
When you’re struggling or feeling self-doubt: How do you get through difficult phases?
What really helps me is getting back into a clear routine. I also find it important to take breaks—sometimes a full day off on the weekend just for myself. It also helps a lot to talk things through when I hit a dead end. Another doctoral researcher, Katharina, is always a great person to talk to—we’ve actually become close friends. That support really helps me get through any difficult phase.
What’s next for you after your doctoral studies? Can you imagine staying in academia, maybe even becoming a professor?
Right now, I’m keeping my options open and don’t want to commit to a specific path yet. A doctorate is already demanding enough, and there is still a lot of uncertainty. I can definitely imagine staying in academia, because I really enjoy teaching.
What advice would you give to students who are unsure whether a doctorate is the right choice?
If you’re interested in research and the opportunity arises, you should take it and pursue a doctorate. I also believe that if you can get excited about a very specific niche—like measurement technology in my case—a doctorate can be a great fit. You definitely need the willingness to explore new topics and acquire skills outside your main field.
For those interested:
If you’d like to take a tour of the STEB wastewater treatment plant, you can request a guided tour. With a bit of luck, Alexandra will personally show you everything that’s required to ensure wastewater can re-enter the water cycle, and she’ll guide you through the facility with great enthusiasm and even more expertise.
And for those who, understandably, have now been inspired to pursue a doctorate themselves, the Graduiertenzentrum.OWL offers all the essential information and advisory services. It supports doctoral candidates at TH OWL with workshops, lectures, and numerous opportunities for networking within the research community.
Dear Alexandra, thank you for the interesting insights into your research work. Best of luck with your doctorate!
Glossary
is a virtual replica of a real plant or process. It reflects how the system operates in real time and can also be used to predict its future behavior. This makes it possible to detect problems early, optimize processes, and support better-informed decision-making without having to intervene directly in the physical system.
refers to the slow, often imperceptible change in a measured value, even though the actual physical quantity remains unchanged. This can be caused, for example, by aging, contamination, or changes in the sensor’s electronics. As a result, measurements gradually become less accurate over time without this being immediately noticeable.
is a high-level programming language used to develop computer applications, websites, and data analysis workflows. It is known for its clear syntax and readability, which makes it especially popular among beginners as well as in academic and research environments. With Python, users can automate tasks, analyze large datasets, and develop applications in fields such as artificial intelligence and machine learning.
refers to measurement methods in which values are not directly captured by a physical sensor but are instead calculated from other available measurement data. Models or algorithms are used to estimate variables that are difficult or impossible to measure directly. This approach makes it possible to obtain additional information without installing new physical measurement devices.
