As a top-rated AI development company, we built a machine learning water leakage detection system to stop leaks in toilet tanks and prevent property damage. During development, the team had to deal with fitting a heavy ML model into a device with limited memory. They also worked on keeping it connected to the internet at all times for it to work well. However, in the end, we have achieved the client’s goal.
About Client
Challenge
Our Solution to Get their Desired Results
It’s a battery-powered device that floats on top of water tanks. It uses AI and machine learning models to keep an eye out for leaks. Once connected to a mobile network, the device quickly detects any changes in the water tank that suggest a leak. It checks how severe the leak is and alerts the user in real time, using energy-saving methods to extend its battery life.
Data Collection
Employed microchips with accelerometers and gyroscopes to gather data, initiating the first step towards intelligent leak detection.
Model Training
The clean feature vectors, prepared through pre-processing, were analysed by a Convolutional Neural Network (CNN) model to predict leaks with minimal false alarms.
Device Specifications
Featuring an ESP32 board for reliable WiFi and Bluetooth connectivity and an IMU sensor for detecting water movements. Programmed with Arduino and powered by a battery, it promises up to 24 months of leak monitoring.
What it can do
Informed Management
The system can sort leaks by how severe they are, so property managers can act quickly and correctly.
Operational Efficiency
Significantly reduces the time and resources traditionally spent on traditional leak detection.
Property Protection
Offers a robust defence against water damage, particularly in large-scale facilities like hotels, where manual monitoring is impractical.