| Dust-related enterprises are accompanied by generation of dust during their product process,which leads to dust explosion accidents sometimes.Recently,the Ministry of Emergency Management has issued relevant laws and rules on dust explosion prevention to further control explosion safety accidents.Dust-related enterprises are its main focus.Although the dust-related plant has been equipped with various types of dust removal equipment,there are problems such as low efficiency,untimely feedback,and inability to uniformly supervise.Besides,the safety supervision department cannot carry out inspection and supervision immediately.It is hard to respond according to the security incidents efficiently.This thesis studied the relevant theories of dust explosion protection.It analyzed the problems existing in the dust removal and explosion protection process of the plants.Then combined with the needs of actual application scenarios,it designed and implemented a dust removal and explosion avoidance system based on the Internet of Things.By the vertical division of the supervision area with the independence of different businesses,it forms a regional management.The information integration and standardized management of municipal projects to form a project management.It uses equipment information to ensure the control of the entire life cycle of sensor monitoring equipment management.Realize off-site supervision of problematic projects to form task management.It provides related personnel with global map statistics and monitoring data visualization service to form GIS global map.This system uses the current popular Spring Cloud architecture as the technical framework.It builds service components through Spring Boot,and independently develops and deploys each microservice component.So each module meets the characteristics of high cohesion and low coupling.And the system uses Nginx service,Tomcat cluster,Redis cache,MySQL cluster and other technologies to improve it’s concurrency,reliability,and availability.In the time characteristic view of the data collected by the dust and explosion avoidance system,and based on the time series analysis theory,this thesis carried out abnormal detection,prediction and early warning of plant working conditions through collecting data.The main tasks include: 1.Investigates the relevant time series model and select the prediction model used in this thesis.2.Preprocess the historical working data to make it meet the needs of the time series model.3.Decide the parameters for the model and evaluate the fit.4.Adjust the optimization parameters to improve the prediction effect.5.Starting from the time series similarity measurement,this thesis studied time series anomaly detection.And it mainly researched the similarity measurement algorithm DTW.It is based on the analysis and summary of previous research.It combined with the characteristics of the data to give the time series different weights.Then it made full use of the weights in the DTW algorithm of calculation of the similar distance.These weights could provide the upper and lower thresholds to stop the calculation,to make improvements and improve the calculation efficiency.Then based on the similar distance obtained by the improved DTW algorithm,this thesis carried out cluster analysis.So,the abnormality time sequence could be marked.And the result could be applied to the abnormal detection of working conditions to meet the needs of engineering application.This system has been operated stably for some factories in Jiangsu,Guangdong,Chongqing provinces and some other regions.It has realized automatic monitoring of the working conditions during the production process.It meets regulatory needs.Also,it improves the efficiency of supervision,provides reliable guarantees for people’s safety and property.It helps guarantee the production safety of enterprises. |
PDF Full Text Request