The Design And Implementation Of Hidden Hazard Analysis System Based On Natural Language Processing

Baihetan Hydropower Station is the second largest hydropower station in China after the Three Gorges Hydropower Station.The safety production problems that have arisen in its construction have also received attention.Establishing a correct safety awareness is the focus of all employees of the hydropower station and one of the measures to ensure the safe production of hydropower stations.With the advancement of society and the development of science and technology,in addition to production,safety issues are also the main problem solved by hydropower stations.Minimizing the safety hazards during the normal operation of the hydropower station is a commitment to the safety of employees and the basis for the normal and smooth operation of hydropower stations.At present,it is necessary to mine and analyze a large number of on-site construction safety hazard data,so as to effectively avoid non-safety factors.Natural language processing is an important direction in the field of artificial intelligence.It is a science that integrates linguistics,computer science,and mathematics.It mainly studies various methods to realize communication between humans and computers in natural language.Nowadays,deep learning research oriented to natural language has made great achievements in the fields of text categorization,sentiment analysis,grammar analysis,machine translation,etc.,which provides feasibility for data mining based on natural language.This hidden danger analysis system,based on the Python language development,uses the mainstream open source machine learning framework TensorFlow and the deep learning representative algorithm-Convolutional Neural Network(CNN).The system mainly includes three functional modules:high-frequency vocabulary display,typical hidden danger identification and hidden danger attribute classification.The high-frequency vocabulary display is used to display the high-frequency words counted in a large number of hidden danger records through the word cloud map;the typical hidden danger identification can determine whether the hidden danger record is a typical hidden danger,and the typical hidden danger is a hidden danger with high attention.Hidden dangers in order to make repairs early;the classification of hidden danger attributes is to classify the hidden danger records more intelligently and quickly into the corresponding hidden danger attributes,so that it is easier to deal with hidden dangers later.The designed system basically achieved the expected functions and was used in theconstruction safety management of Baihetan Hydropower Station.