Analysis and Procedures for Water Pipeline Leakage Using Three-Axis Accelerometer Sensors: ADXL335 and MMA7361

Abstract

Previous studies have proposed vibration of the water pipeline using accelerometer sensors as a mechanism to detect leaks. Nevertheless, they only relied on a single x-axis data, no single study investigates all the axes of the sensors. Therefore, this paper aims to investigate a vibration technique to detect leakage on plastic water pipeline using wireless accelerometer sensors, namely, 6DOF accelerometer sensor (i.e., MPU6050) and breakout accelerometer sensors (i.e., ADXL335 and MMA7361) across the x-, y-, and z-axis. A 25-mm diameter of acrylonitrile butadiene styrene (ABS) pipe with a length of approximately 10 m was developed as the water pipeline testbed. All of the accelerometer sensors measured the vibration on the water pipeline across the x-, y-, and z-axis, over ZigBee networks. The vibration signals were then compared between the three sensors and analyzed by extracting the signal features in time and frequency domains. The sensors were examined based on three different cases, which are no pipe leakage, a 1-mm leak, and a 3-mm leak. All three sensors demonstrated a significant difference between no leak and leak conditions when the water pressure is in the range of 0.6– $1.2~kgf/cm^{2}$ for both time and frequency domains. For different leak size cases, ADXL335 can distinguish the 1-mm and 3-mm holes from the three-axis data. On the contrary, MPU6050 can identify similar leak size cases only from the y- and z-axis data. Overall, ADXL335 has the best performance compared with MPU6050 and MMA7361 in detecting water pipeline leakage, which includes the sizes of the leaks. Based on the empirical results, this paper finally proposes a procedure analysis for each sensor on improving the accuracy of water pipeline leakage detection.