Design of a Household Consumption based Water Leak Detection System Utilizing Machine Learning Algorithm

Water is a limited resource; thus, every drop matters. Leaks in water distribution networks (WDNs) and residences contribute considerably to water utility losses. The goal of this research is to design and build a water leak-detecting system based on home use. Based on historical water consumption and experimental flow rate water consumption data, this research developed a Bidirectional LSTM (BiLSTM) machine learning (ML) approach for building a water leakage detection model. The BiLSTM model is compared against Gated recurrent units (GRU) and Autoregressive models on the following performance metrics: mean squared error (MSE), root mean squared error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE) (MAE). According to the data, the BiLSTM model performed better overall in water leak detection.