A Pipe Leak Detection System Using Pressure and Flow Sensors for Residential Water Pipes Using The Sugeno Fuzzy Method
Abstract
The water loss rate in Indonesia remains at a critical range of 20–45%. The fundamental challenge in addressing this issue at the household scale is the computational difficulty of distinguishing between pressure fluctuations caused by normal water consumption patterns and pressure drops caused by physical pipe damage. Therefore, this study designed a smart pipe leak detection system prototype based on the Internet of Things (IoT) using the Zero-Order Sugeno Fuzzy Logic method. The system utilized an ESP32 microcontroller as a local processing unit based on edge computing, integrated with a water flow sensor (YF-S201) and a pressure sensor to monitor fluid dynamics in real time. The acquired data were classified into three linguistic sets—Normal, Small Leak, and Large Leak—to control a relay actuator as an automatic mitigation mechanism. Based on the system testing results, the instrument measurement error rates were 2.12% for the flow sensor and 2.19% for the pressure sensor. The Sugeno Fuzzy Logic algorithm achieved an inference accuracy rate of 100% in classifying pipe condition status compared with theoretical mathematical calculations. Furthermore, the IoT architecture recorded an emergency notification transmission latency of 1.20 seconds to the Blynk dashboard. The implementation of decentralized control through the microcontroller demonstrated the capability to reduce false alarm risks and provide reliable automatic water supply shutoff mitigation without depending on the stability of cloud server connectivity.
