Design And Implementation Of Dust Explosion-Proof Monitoring System Based On Neural Networks

For factory production,excessive dust concentration can easily cause workers to suffer from lung diseases,and even cause dust explosion disasters,threatening the life safety of workers,and causing huge economic losses,so accurate monitoring of dust concentration is crucial.However,the current technology based on threshold alarm not only does not have realtime,but also cannot carry out unified management of dust monitoring and treatment and subsequent accountability.Therefore,a factory dust concentration monitoring method combined with deep learning is proposed,and a dust explosion-proof monitoring system is designed and implemented.The main research work of this paper is as follows:(1)The research on dust concentration prediction algorithm was completed.Firstly,in order to obtain the main factors affecting dust concentration,the data were preprocessed and analyzed in correlation,and the data with a large correlation with dust concentration was obtained by calculating the Pearson correlation coefficient,which was used as an important source for setting input parameters in the model construction stage.Then,a dust concentration prediction model based on LSTM long short-term memory neural network is established,and models with prediction steps of 1 hour and 3 hours are designed and the experimental results are compared.Finally,the existing model is optimized and improved by using the Attention Mechanism mechanism,and on this basis,the better activation function and optimization method are explored,and the best model based on LSTM optimized by Attention Mechanism is established,which effectively improves the accuracy,stability and convergence speed of neural network dust concentration prediction,and obtains a more ideal prediction effect.(2)Design and realize the dust explosion-proof monitoring system.According to the prediction results of the model output,alarm and provide an information management platform for managers,the back-end part is developed using Spring Boot and Spring Cloud microservices framework,the database storage uses My SQL and Redis databases,and the front-end part uses Html,CSS and Java Script technologies and Thymeleaf template rendering engine technology to finally complete the design and development of the system.Making enterprise information management and accident accountability more intelligent is of great significance to production safety.