Machine learning to reduce leakage at the Örby field
The Örby field and the Örby drinking water treatment plant, together constitute a technically complex system that contributes to securing the distribution of drinking water.
The drinking water production at the Örby drinking water treatment plant involves a system where incoming drinking water from Ringsjöverket is infiltrated into a closed geological formation in the Örby field. After approximately three weeks of residence time, the water is abstracted and pumped into the Örby drinking water treatment plant for further distribution.
Unfortunately, there are disadvantages to having a natural reservoir, and the biggest one is leakage. On an annual basis, this amounts to 3000-4000 m3 per day, equivalent to the daily water consumption for approximately 25,000 people. The leakage is influenced by uncertain forecasts for drinking water consumption and operational strategies. The challenge of balancing water intake and extraction from a groundwater reservoir is not unique to the Örby field.
The leakage is affected by uncertain forecasts for drinking water consumption and operational strategies. The system is difficult to describe using traditional calculation methods due to time delays in the system, complex reservoir conditions, human factors, and variations in water extraction. Therefore, the project aims to address the issue using non-traditional calculation methods that employ locally implemented machine learning. It will also tackle the challenge of using machine learning in an industry that highly values IT security.
For optimization calculations of groundwater reservoirs, there are currently no “ready-made” algorithms that have been trained with historical data from similar questions. The project will create and evaluate the type of algorithm (or group of algorithms) that can best be used based on data knowledge, domain expertise, and machine learning/data science proficiency.
If the leakage of drinking water from the Örby field can be reduced, it will decrease the carbon footprint and energy consumption for the region’s water distribution.
