Faculty of Pure and Applied Mathematics, Wrocław University of Science and Technology is pleased to highlight a significant new collaborative research initiative: “Advanced signal processing techniques for cyclostationary modelling in Gaussian and non-Gaussian noisy environment.” Funded under the prestigious WEAVE-UNISONO call by the National Science Centre (NCN), Poland and the Research Foundation – Flanders (FWO), Belgium, this project represents a strategic partnership between two leading institutions in the field of industrial mathematics and diagnostics.
Project overview
Modern industrial machinery operates under increasingly complex conditions, where traditional signal processing often falls short. This project focuses on the detection and estimation of cyclic sources, the “heartbeat” of rotating machinery within environments plagued by both standard (Gaussian) and impulsive, heavy-tailed (non-Gaussian) noise. The primary goal is to develop robust mathematical frameworks and optimized algorithms capable of identifying structural faults (such as bearing or gear damage) long before they lead to catastrophic failure.
Key Research Pillars
- Cyclostationary Modelling: Leveraging the periodic nature of statistical properties in signals to isolate fault signatures.
- Non-Gaussian Robustness: Creating filters and estimators that remain accurate even when signals are corrupted by high-energy “spikes” or impulsive interference.
- Algorithm Optimization: Ensuring that advanced mathematical models are computationally efficient enough for real-world industrial applications.
- Validation: Rigorous testing of these techniques in practical fault identification scenarios to bridge the gap between theory and industry.
Collaborative Leadership
The project brings together a wealth of expertise from Poland and Belgium:
- Wrocław University of Science and Technology (Poland): led by prof. Agnieszka Wyłomańska (Faculty of Pure and Applied Mathematics), a specialist in non-Gaussian statistical signal processing and prof. Radosław Zimroz (Faculty of Geoengineering, Mining and Geology), an expert in vibration-based condition monitoring.
- Katholieke Universiteit Leuven (Belgium): led by prof. Konstantinos Gryllias, an expert in mechanical diagnostics and vibration monitoring.
Project Details
- Funding Agencies: NCN, Poland; FWO Belgium.
- Duration: February 2, 2026 – January 31, 2030.
This four-year endeavor promises to push the boundaries of predictive maintenance, offering the industry more reliable tools for monitoring critical infrastructure. We look forward to the innovative mathematical solutions and diagnostic breakthroughs this international team will undoubtedly produce.
By Agnieszka Wylomańska and Justyna Witulska (Wrocław University of Science and Technology).