| Electrical fire accidents have been increasing continuously in recent years.Electrical causes have become the main culprit of fire accidents,which cause serious casualties and property losses.Although the country actively popularizes electrical fire prevention education,electrical fire often has strong uncertainty,which is difficult to prevent,and the prevention of electrical fire has also attracted the attention of the state and the market.For electrical fire,in addition to popularizing the awareness of fire protection such as electricity safety,the electrical fire pre-warning system is the most reliable prevention means at present.Improving the accuracy,comprehensiveness and reliability of the system is conducive to reducing the number of electrical fire accidents and reducing losses.Firstly,the research background and significance expounded in this paper,introduces the research status and development process at home and abroad,summarizes the main characteristics and development trend of the electrical fire warning system,and the promotion and application of intelligent algorithm in the early warning algorithm.Secondly,the basic structure of the electrical fire warning system is determined based on the Internet of things.Based on the analysis of the causes of electrical fire,the harmonic current,residual current,resistance residual current,insulation conductivity,voltage current,temperature and other electrical signal quantities are used as the input warning parameters of the early warning system.The accuracy of early warning is improved by several high correlation parameters,and the warning accuracy is improved according to the system The whole system structure is modular and the hardware circuits of each part are designed.In order to further improve the accuracy of early warning,intelligent algorithm is introduced as the basis of fire discrimination.The single classification support vector machine(OCSVM)algorithm is used to establish the fire prediction model.The prediction model optimized by QPSO is verified by MATLAB simulation to improve the accuracy of the pre measurement;meanwhile,the integration algorithm is used to enlarge the hidden danger for the risk of the slow transformation of residual current,improve the system reliability.Finally,the software design of each part is completed based on the hardware framework.The single classification support vector machine and integration algorithm are written into the main program as fire discrimination means,and the monitoring and management system is built by C # language.Combining hardware system and monitoring system,the experimental platform is built.Through the experimental platform simulating various working conditions,the parameters collection and early warning alarm test are carried out for the early warning system,which meets the national standards and meets the design requirements.It is proved that the system is reliable,stable and accurate,and has certain practical significance. |
