|

Smart Water Leak Detection Based on AIoT

Smart Water Leak Detection Based on AIoT

Smart Water Leak Detection product is an IoT-based smart solution designed to monitor water flow and automatically detect leaks in piping systems. This product combines flow sensors, pressure sensors, and artificial intelligence algorithms to detect flow anomalies that indicate leaks. Equipped with cloud connectivity and a monitoring dashboard, this product provides real-time notifications to users to prevent further damage and save water usage.

🌐 Project Domain 

This project falls under the IoT‑based smart water management domain, focusing on real‑time detection and localization of water leaks in PDAM pipelines. By integrating flow sensors, pressure sensors, and an ESP32 microcontroller with embedded AI, it provides accurate leak detection, seamless connectivity to AWS IoT, and remote monitoring capabilities through the Blynk app, making it ideal for supporting PDAM operators and field technicians.

🧑‍🤝‍🧑 Meet Our Team

  • 👨‍💼 Project Leader : Rihan Hidayat (225150301111009)
  • 🖥️Embedded System : Muhamad Aditya Sanjaya (225150301111011)
  • 🤖 AI Engineer : Rifky Akhsanul Hadi (225150207111012 )
  • 🔧Hardware Engineer :  Reyhan Rashid Rizqullah (225150300111047)
  • ☁️Cloud System : Afifah Maulidiah (225150300111033 )

❗ Problem Statements

Water leaks in pipe distribution systems, both in household and industrial environments, are often not detected early. This can lead to waste of water resources, damage to infrastructure, and increase operational costs. Conventional methods for detecting leaks, such as manual inspections or periodic meter readings, are not effective in providing a quick response to sudden or hidden leaks. Therefore, a water leak detection system is needed that is automatic, real-time, and able to provide early notification to users, so that preventive measures can be taken immediately to minimize the negative impacts of the leak.

🎯 Goals

  • Develop a system that can detect and find pipe leaks automatically.
  • Reducing water waste due to undetected leaks.
  • Assists in preventing further damage to pipelines and supports data-driven preventive maintenance.
  • Embedding machine learning in a smart water leak detection system

💡 Solution Statements

To overcome the problem of water leaks that often go undetected and cause waste and damage to infrastructure, a system called Smart Water Leak Detection based on the Internet of Things (IoT) was developed. This system automatically monitors water flow and pressure using integrated sensors, and applies artificial intelligence algorithms to detect leaks in real time. Data is sent to a cloud platform and displayed via a monitoring dashboard, allowing users to receive early notifications and take quick action before the damage becomes more severe.

🧰 Prerequisites – Component Preparation

⚡ Hardware

  • ESP32 Development Board: Main controller with Wi‑Fi connectivity and processing capabilities for sensor data and AI inference
  • Flow Sensors (YF‑S201): Measure water flow rate before and after potential leak points
  • Water Pressure Sensor (G1/4 1.2 Mpa): Monitor pressure changes caused by potential leaks
  • Power Supply (5V / Power Bank): Provides stable and continuous power for ESP32 and sensors
  • Resistors (15kΩ & 10kΩ): Serve as a voltage divider to adjust sensor output for ESP32 analog input range
  • Prototype Pipe Simulation Kit: Includes PVC pipes, valves, and fittings for realistic leak detection testing

💻 Software & Services

  • Arduino IDE: Development environment for ESP32 programming and sensor data processing
  • Google Colab + TensorFlow Lite for Microcontrollers: Model training and conversion for AI inference on ESP32
  • AWS IoT Core: Enables secure cloud connectivity, data storage, and remote access
  • Blynk Mobile App: Provides remote monitoring, real‑time status, and leak detection alerts to end-users

🧩 Schematic

🎬 Demo


Run water through the pipeline, simulate leak conditions by opening the leak valves, and observe the ESP32 detecting anomalies. The Blynk app displays flow and pressure data in real‑time and triggers instant leak alerts when anomalies occur.

You can watch a demo of the Smart Water Leak Detection System here

🔬 Evaluation:

  • Model Inference Accuracy

    • Leak Detection Model (FNN): Achieved an F1‑score of approximately 88.8% for multi‑class leak detection, reliably distinguishing between:

      1. No leak (normal conditions)

      2. Leak at point 1

      3. Leak at point 2

      4. Leaks at both points (1 and 2)

  • Inference and System Response Time

    • Local Model Inference (ESP32): ~1 second average for sensor data acquisition, feature scaling, AI inference, and leak detection.

    • Blynk Notification Response Time: Under 2 seconds from detection, allowing quick alerting and action.

    • AWS IoT Message Delivery: Approximately 1–2 seconds, providing near‑real‑time cloud availability for remote monitoring.

  • User Experience and Monitoring

    • The system provides seamless real‑time alerts through the Blynk app, making it easy for PDAM operators and field technicians to respond quickly.

    • Alerts and status data are reliably transmitted to AWS IoT, making long‑term trend analysis and leak detection possible.

✅ Conclusion 

This project delivers a smart, reliable, and efficient solution for detecting and locating water leaks in PDAM pipeline networks. By combining real‑time data from flow and pressure sensors with a Feedforward Neural Network embedded in an ESP32 microcontroller, the system can accurately identify both the presence and location of leaks. Its seamless integration with AWS IoT and the Blynk platform allows for remote monitoring and instant notifications, making it an invaluable tool for PDAM operators and field technicians. Future enhancements could focus on increasing detection sensitivity through ensemble learning methods, introducing data buffering to maintain reliability during connectivity interruptions, and designing IP‑rated enclosures for improved durability in challenging environments.

📲 Contact us!

Our Git: https://github.com/RifkyHadi7/water-leak-detection

Our Linkedin :

 

 

Similar Posts