Challenges description

 

Fiware4Water Challenges is closed. The competition challenged data scientists and data analysts to show how real time data could be used to support better water management services through the data generated by digital meters and providing feedback to consumers and water utilities.

 The Fiware4Water Challenges aimed at facilitating training for data scientists, providing with the necessary tools that a data scientist needs in order to train a machine learning model. The aim is to improve the management of water data, detecting anomalies in water quality, sensors out of range to evaluate their impact and possible actions.The evaluation phase has now started. The winners will be known on the 5 August 2021

 

Challenge 1: Sofiyska voda AD

Proposed by: Sofia, Sofiyska voda AD

Summary: Sofiyska voda AD searches for a solution that can combine both data from sensors’ measurements in the sewerage network (level meters, low cost temperature sensors data for Combined sewer overflow monitoring etc.) and rain gauge data, to be able to perform a cross-analysis of the measured parameters knowing if they happen in dry weather or during rain events. The prospect of the solution will be a real-time control of the behaviour of the sewer system.

Expected Outcomes:

  • A predictive model capable of predicting if an overflow is happening in a specific time.
  • The predictive model should predict as fast as possible the overflow.

 

 

Challenge 2: Challenge 2 – CAP

Proposed by: CAP

Summary: The challenge proposed aims to find a concrete solution to explore real-time the quality of sensor data. Data quality assessment should be obtained employing periodic laboratory analyses as reference.

Expected Outcomes:The participants can work in a clustering solution to profile the different correlations observed between the real data and the laboratory data. Therefore, this task has an analytical implication (analysis of the multiple profiles and conclusions), it is expected to get an analytical report of the different profiles/clusters found and to justify the predictions of their algorithm/algorithms.

 

 

Challenge 3 – South West - United Kingdom

Proposed by: South West Water

Summary: The objective of the competition is to calculate the short-term and long-term water network events in order to evaluate possible bursts pipe, leakage or any change in the customer water consumption to be considered. It is very important to define task forces to resolve water leakage or analyse any possible improvement in the water network distribution.

Expected Outcomes:

  • Short-term events such as large pipe bursts which normally correspond to short, steep spikes in the flow profile typically lasting hours or a few days. 
  • Longer-term trends which normally correspond to underlying leakage from small, unrepaired bursts or changes in customer water consumption such as the completion of a new housing development. This will typically last weeks to months.