Flow prediction of irrigation well pump based on experimental and big data analytics

Abstract Currently, China’s agricultural irrigation consumes a huge amount of water, and traditional agricultural irrigation methods lead to low irrigation efficiency and serious water resource waste. Agricultural irrigation water management is also difficult to achieve refined management due to the lack of accurate monitoring of water use data and information in various irrigation areas. To find a more convenient flow measurement method, this paper proposes using data-driven monitoring of water pump characteristic parameters to predict flow rate. Three big data-based methods for predicting the flow rate of irrigation well pumps were compared, including RBF neural network, Support Vector Machine (SVM), and Extreme Learning Machine (ELM). The results indicate that the Extreme Learning Machine (ELM) model not only has short time consumption but also high prediction accuracy, laying the foundation for the application of big data technology in water management and high-quality irrigation water management that can save water.