Design of "Four Waters Returning to the Courtyard" Rainwater Optimization and Utilization System Based on Internet of Things and Edge Intelligence

The traditional “Four Waters Returning to the courtyard” in China has ecological wisdom in rainwater collection and discharge. However, in the efficient utilization of rainwater resources in modern cities, it is necessary to integrate digital technologies. This paper designs a “Four Waters Returning to the courtyard” rainwater optimization and utilization system based on the Internet of Things and edge intelligence. By deploying various types of sensors such as rainfall, water level, water quality, and environment etc. in the perception layer, combined with the intelligent data preprocessing and AI prediction models of edge nodes, real-time monitoring, quality analysis, and regulation optimization of rainwater can be achieved. The system adopts a hybrid communication structure of LoRaWAN and 5G, supports multi-protocol data upload (MQTT/CoAP), the platform layer integrates Hadoop + Spark for data processing, the model layer constructs a multi-variable LSTM network to predict rainfall trends, and adjusts the water storage and discharge strategies through a fuzzy control algorithm. The experiment was deployed in a reconstructed sample of the “Four Waters Returning to the Courtyard” in Hefei. The system collects about 1.2GB of data daily, the accuracy of the prediction model reaches 91.3%, and the rainwater recovery rate is increased by 26.7% after optimization. This research provides an effective path for the digital ecological transformation of traditional buildings and demonstrates the application potential of edge intelligence in green infrastructure.