Leakage Detection of Pipelines in Water Distribution Network Using Machine Learning Algorithm

EasyChair Preprint 9230

Abstract

Technology plays a vital role in our day today life. IoT along with Wireless Sensor Networks ( WSN) and small nodes that communicate by wireless links can be used to collect information of various operations, conditions and performance of any system / operation / task / environment that is far away from the controlling unit i.e. remotely controlled. One such remotely controlled scenario is water distribution networks ( WDN). Water distribution network is the source of water supply in many cities in the globe. Water is a pivotal natural resource which is broadly mishandled by human beings. A report says around 40% of clean water is wasted due to leakages in the WDN. According to McKenzie in South Africa (so called a water – stress country) due to its limited water resources, steadily increasing population and growing water demand loss / leakage of water plays an important role as its municipal water service is threatened with an appraised value which is more than R7 billion per annum. Thus it’s important to conserve water loses in the form of leaks in WDN. In this project, a network of wireless sensors is placed in pipelines at a fixed distance representing nodes of our model and pressure / flow at that point is monitored and transmitted at regular intervals. The transmitted signal is then processed and analysed using machine learning algorithms to detect the leak. The data thus obtained is stored in cloud for further reference.

Keyphrases: Cloud Storage, Internet of Things, Machine Learning Algorithms, Water leakage in Pipelines, Wireless Sensor Network

@booklet{EasyChair:9230,
  author    = {M Saravanabalaji and S Shakthi Raagavi and E Praveen and S Samritha},
  title     = {Leakage Detection of Pipelines in Water Distribution Network Using Machine Learning Algorithm},
  howpublished = {EasyChair Preprint 9230},
  year      = {EasyChair, 2022}}