Pipe Network Water Level Prediction Platform Coupled with SWMM and LSTM

In this paper, a pipe water level prediction platform is proposed to address the low predictability issue of traditional drainage systems in combating waterlogging. Based on the drainage pipe network, the study integrates real-time data collected through the Internet of Things (IoT) sensor system and predictive data from the Storm Water Management Model (SWMM) and Long Short-Term Memory (LSTM) neural network models to enhance the accuracy and speed of waterlogging forecasts. Additionally, a web-based and WeChat Mini Program application platform has been developed. The platform was tested and validated in a specific region in Huzhou. Experimental results demonstrate the platform’s high prediction accuracy and its potential for urban waterlogging risk monitoring and early warning, offering a novel approach to smart water management.