Prof. Dr. Dieter William Joenssen

Professor Dieter William Joenssen holds a professorship in Industrial Data Science and is currently the Head of School of Mechanical Engineering and Materials Science at Aalen University. He obtained his Diploma degree in industrial engineering from the Technical University Ilmenau in 2010. At the same institution, he obtained his doctoral degree in 2015 for his work in machine learning methods for missing data prediction.

His group focuses on developing and applying machine learning methods and statistical modeling for industrial as well as pure research purposes.

Research Topics

Anomaly detection is used in predictive maintenance, when no clear connection between available sensor data and maintenance events can be drawn. This may be because time lags between events are significant, event data or expert opinions are not available. Extra challenges arise when time dimensions and transient processes are considered. Here data preprocessing plays a significant role in reliable knowledge extraction and increasing model stability.

Non-stationary timeseries data lies at the core of model based predictive maintenance. Here model accuracy and specificity benefit significantly from feature engineering and selection. Many possible transformations – from the time to frequency or wavelet domains – deliver key information for improving model performance. However selecting only a reduced set of key features is necessary for restricting model variance. Striking the balance when sample size is low remains a current challenge for many technical processes.