Hackathon challenge

Bemas, the Belgian organisation for exchange of knowledge and good practices of maintenance and asset management organises a hackathon as part of the asset performance 4.0 conference

UPDATE: The Hackathon Challenge is now closed. The three best entries and the winner will be presented during the asset performance 4.0 conference from September 15-17.


Pump failure prediction for smarter predictive maintenance

Aquafin is responsible for the transport and treatment of household wastewater in Flanders. We manage more than 2500 pumping stations as part of our operations. These installations pump raw untreated sewage from one location to another, e.g. under under a road. In case of pump failure, risks are flooding of streets and discharge of raw untreated waste water to rivers, lakes, …


About the challenge

Aquafin wants to know whether there is any indicative value in existing data that warns for upcoming pump failures.The successful application is an algorithm or data pipeline to detect upcoming pump failures or a visualisation that focusses the attention of asset engineers on possibly malfunctioning pumps. If solution is succesful, a better maintenance would be possible and a reduced risk of sewage discharge to clean surface water. Hence, as part of the Bemas hackathon, you are invited to work on a dataset of four pumping stations.


Provided input

The dataset comprises logs of relevant online measurements as well as a description of the maintenance performed on the assets during a period of 6 to 12 months. The dataset, shared in csv as well as in the parquet file format, can be downloaded below.


Target audience

Data scientists, A.I. Machine Learning, asset management specialists.


Kris De Gussem

Data scientist

pumping station Wervik