The University of Sheffield
F14.pdf (378.21 kB)

CCWI2017: F14 'Pattern Recognition in Residential End Uses of Water Using Artificial Neural Networks and Other Machine-Learning Techniques'

Download (378.21 kB)
journal contribution
posted on 2017-09-01, 15:03 authored by Juan Carlos Ibáñez Carranza, Roberto Díaz Morales, Jose Antonio Sánchez
Machine-Learning and other Artificial Intelligence techniques have nowadays many practical applications in engineering, science or everyday life. In the water industry, there is also a broad scope of potential applications. In this paper, it will be presented a system developed by Canal de Isabel II to identify residential use of water in its different appliances, based on records from precision water meters equipped with pulse emitters. Developed models are based on Support Vector Machines, and Artificial Neural Network paradigms. Training data sets for the models have been extracted from a sample of about 300 residential users in the Region of Madrid (Spain), monitored since 2008. In this time, more than 35 million of water use events have been registered and about 15 million hours of water consumption monitored. Machine-Learning techniques have proved to be an accurate and suitable method for automate this task that otherwise should require a huge number of man-hours of processing by operators.