| With the deepening of electrification,people’s life is increasingly inseparable from the support of electricity.However,the number of building electrical fires has remained high for a long time,causing serious property losses and casualties.Therefore,it is of great significance to realize the reliable monitoring of building electrical fire.Due to various limitations of traditional fire monitoring,the rate of false alarm and missed alarm is high,and the reliability needs to be improved.With the rapid development of new generation information technology such as Internet of things and artificial intelligence,it is possible to monitor building electrical fire quickly and effectively.On the basis of in-depth research,this thesis defines the relationship between electrical fire and arc,and realizes the reliable monitoring of building electrical fire by identifying the changes of arc state before electrical fire under the architecture of Internet of things.Firstly,on the basis of literature review,the characteristics of electrical fires in different types of buildings are analyzed.Starting with the structure and composition of the arc,the basic process of arc discharge is given,the Mayr and Cassie mathematical models of arc are derived,and the two arc generation mechanisms of electric breakdown and thermal breakdown are analyzed and compared.A simulated arc fault test-bed was built,and the simulated experiments were carried out on three types of typical loads: resistive load,resistive-inductive load and mixed load.The relevant data were collected,and the experimental data set was obtained at the same time.Secondly,the theoretical principles of the identification method and comparison method used in this thesis are systematically introduced.The trigger reporting mechanism based on wavelet transform and the arc identification method based on modal decomposition and feature matching are used to connect the wavelet transform,modal decomposition and feature matching in theory and logic in two stages.Data acquisition and processing are realized in the arc monitoring module,wavelet features of electrical fire are extracted to trigger and report,possible fire data are uploaded,and modal decomposition and feature matching are carried out at the platform end to realize accurate identification.The detailed process of using simulation experiments to verify the identification method is described in detail,and the key charts of the identification process are given.Through the comparison with the comparison method,the evaluation results are given from the two indicators of timeliness and accuracy,which verifies the effectiveness of the method used in this thesis.Thirdly,the electrical fire monitoring system of the Internet of things building is developed,The hardware sampling flow and sampling circuit are given,the design and implementation of functional modules such as realtime monitoring,selection and query,report statistics and alarm linkage are completed,the construction and deployment of cloud database are completed,the network communication from the arc monitoring module to the edge gateway and then to the cloud is realized,the interaction of system information and data is realized,and a visual platform based on B/S architecture is built,The remote monitoring of building electrical fire is realized.Based on the two-stage arc recognition method,aiming at the problem that the electrical fire data is difficult to obtain,the platform takes advantage of the advantages of Intelligent cloud fire protection and adopts the construction method based on the vertical and horizontal framework fire feature database to realize the preliminary establishment of the building electrical fire feature database,enrich the acquisition methods of building electrical fire data,and prepare for the continuous enrichment of the subsequent feature database and the continuous improvement of the recognition algorithm.Finally,the research work is summarized item by item,and the deficiency and development prospect of building electrical fire monitoring are prospected. |
