Estimate Water Level in Pump Station Based on Sound

As a leading water technology company with 23,000 employees operating in over 150 countries, Xylem is at the forefront of addressing the world's most critical water challenges. We invite passionate individuals to join our team, dedicated to exceeding customer expectations through innovative and sustainable solutions.

Job Description
A wastewater pump station is designed to transport wastewater from households and industries to a treatment plant. A pump station holds a tank that is filled with wastewater. Once the tank reaches a certain limit, a pump will start, and wastewater is pumped to the next station.
In the pump station tank, there is normally a pressure sensor that measures the water level and triggers when the pump should start and stop. This sensor can sometimes break for various reasons. Then a backup system is needed to estimate the water level.

This thesis work is to investigate if sound, captured from a microphone, can estimate the water level in the tank.

The Goal of the thesis is to use sound clips to estimate water level.

Extent
• Extract sound and water level data from a pump station.
• Build algorithms that fulfills the above-mentioned goal. Algorithms to explore are Machine Learning models for example different types of neural networks such as CNN and RNNs.

Scope

Master thesis for one student.

Contact:

Martin Wessman, martin.wessman@xylem.com