Application of knowledge graph in smart irrigation district management decision making

Area of Science:

• Agricultural Engineering
• Water Resource Management
• Artificial Intelligence in Agriculture

Background:

• Traditional irrigation and drainage management in China faces challenges including complex processes, fragmented information systems, and limited intelligent decision-making support.
• Existing systems struggle with data integration and resource sharing, hindering efficient management of smart irrigation districts.
• The Yellow River diversion irrigation areas, specifically in Henan Province, require improved management strategies.

Purpose of the Study:

• To develop a refined intelligent management decision-making system for smart irrigation districts.
• To address issues of data dispersion and insufficient auxiliary decision-making in traditional irrigation management.
• To propose a novel concept for the informationization construction of China's irrigation areas.

Main Methods:

• Construction of a knowledge graph using data from 28 Yellow River diversion irrigation districts in Henan Province and Zhaokou Irrigation District project reports.
• Application of the BERT+BiLSTM+CRF model for intelligent entity recognition (irrigation projects, problem events) in inspection log texts.
• Utilizing entity alignment technology to link inspection text data with the knowledge graph and employing graph retrieval for decision-making solution generation.

Main Results:

• Successfully built a knowledge graph for managing issues in Yellow River diversion irrigation areas.
• Demonstrated the reliability of the proposed problem management decision-making scheme through validation with specific irrigation district examples and model evaluation.
• Facilitated information data integration and enabled visual management of intelligent irrigation zones.

Conclusions:

• The integrated approach of knowledge graphs and advanced AI models offers a reliable solution for intelligent irrigation district management.
• The developed system effectively addresses data integration challenges and enhances decision-making capabilities.
• This research presents a novel and effective framework for the informationization and smart management of China's irrigation areas.