Leakage analysis of campus water supply system based on MLP neural network
Abstract Water supply system is an important part of campus public facilities, and the water supply pipeline will leak, not only increase the cost of water supply, but also cause water waste. This paper collects water consumption data of water meters at all levels in a school’s water supply system, establishes MLP multilayer perceptron neural network model, determines the water leakage rate according to the fluctuation of predicted value and actual value, so as to analyze the leakage situation of each school’s water supply system. When the fluctuation between the actual and predicted water consumption exceeds a certain threshold, water leakage occurs on that day. Through solving the model, the following conclusions are finally drawn: (1) the annual water leakage rate of the school is 10.74%, and the water leakage is 29131.418L. (2) The water leakage rate in the first quarter is the highest, and the water leakage in the second quarter is the highest.(3) The school aquaculture area is the most serious leakage phenomenon, and the maximum water leakage rate of each water meter node is more than 10%.