AI-Based Acoustic Processor for Classifying Types of Water Leak
Budget: $30 – $250 USD
Objective:
Automate classification & verification of types of water leak in water distribution network with high accuracy
Background:
Local water operators with old-aged pipeline for their water distribution network experience water leakage that comes with a high cost. In Malaysia, water operators have embraced using a digitalized leak detection system, where multiple IoT-connected sensors are installed into pipelines that feed raw data into the cloud and presented in the dashboard for end-users in the local water operator.
Problem:
Even with the system installed, with mass of amount of data being feed from sensors, anytime a leakage is detected, a team needs to be send to the area to do verification of leakage. This consumes financial, time and energy cost to the local water operator to verify everytime.
Solution:
To create an AI processor that uses historical acoustic data from sensors to detect patterns of water leakage, and use the pattern as a highly accurate warning system that eliminates the need to do most of on-site verification
Available Inputs for AI:
> Historical Data of Acoustic Raw Data (6 months)
> Historical On-Site Verification records (6 months)
Expected Result:
Minimum Viable Product (MVP) / Prototype that completes the objective in 2 months form agreed date
What we're looking for:
Someone with completed past development projects that heavily uses Machine Learning (Artificial Neural Network (ANN), Random Forest (RF), Autoencoder Neural (AE) Network, etc)
About Rivil:
Rivil is a Malaysian Water Technology company with the mission to accelerate access to freshwater at affordable cost. ( www.rivil.co/ )
Automate classification & verification of types of water leak in water distribution network with high accuracy
Background:
Local water operators with old-aged pipeline for their water distribution network experience water leakage that comes with a high cost. In Malaysia, water operators have embraced using a digitalized leak detection system, where multiple IoT-connected sensors are installed into pipelines that feed raw data into the cloud and presented in the dashboard for end-users in the local water operator.
Problem:
Even with the system installed, with mass of amount of data being feed from sensors, anytime a leakage is detected, a team needs to be send to the area to do verification of leakage. This consumes financial, time and energy cost to the local water operator to verify everytime.
Solution:
To create an AI processor that uses historical acoustic data from sensors to detect patterns of water leakage, and use the pattern as a highly accurate warning system that eliminates the need to do most of on-site verification
Available Inputs for AI:
> Historical Data of Acoustic Raw Data (6 months)
> Historical On-Site Verification records (6 months)
Expected Result:
Minimum Viable Product (MVP) / Prototype that completes the objective in 2 months form agreed date
What we're looking for:
Someone with completed past development projects that heavily uses Machine Learning (Artificial Neural Network (ANN), Random Forest (RF), Autoencoder Neural (AE) Network, etc)
About Rivil:
Rivil is a Malaysian Water Technology company with the mission to accelerate access to freshwater at affordable cost. ( www.rivil.co/ )