| Electrical cabinet fires often occur,once they occur,they will cause serious losses.Therefore,the early warning technology and real-time prevention and control technology of electrical cabinet fires have become a research hotspot.Taking the low-voltage electrical cabinet as the research object,this paper aims to study a set of on-line monitoring system that can realize the early warning of fire in the low-voltage electrical cabinet.The main research work of this paper is as follows:（1）By analyzing the characteristics of fire development and the main fire causes in the electrical cabinet,the multi-sensor fusion scheme is adopted to monitor the four characteristic parameters of CO concentration,CO2 concentration,temperature and residual current in the low-voltage electrical cabinet in real time,and the early warning model is modeled around these four characteristic parameters.By analyzing the actual requirements of the system,a fire early warning experimental system based on multi data acquisition node-Cloud Server-remote real-time monitoring client is designed and completed.（2）By analyzing the characteristic that the fire of electrical cabinet evolves continuously with time,the LSTM algorithm suitable for processing time series data is selected to predict the fire probability.The algorithm can evaluate the current fire probability according to the input time series data of CO,CO2 and temperature.In order to improve the accuracy of fire early warning,the residual current is also introduced in the modeling of early warning model.The residual current is combined with the existing fire probability for fuzzy reasoning to jointly determine the fire early warning level.（3）According to the overall requirements of the system,the hardware selection is completed,and the hardware circuit is designed around stm32f407 Series MCU.By comparing different communication schemes and embedded software architecture,the communication scheme based on HTTP and the embedded software architecture with real-time operating system are selected,and the embedded software design is completed based on RT thread real-time operating system.（4）The design scheme of remote monitoring software based on C/S architecture is selected,and Alibaba cloud lightweight server is used to carry and calculate the main data.The hot standby of the server and the data sharing function between the servers are designed,which solves the problem that the server sends control instructions to the data acquisition node.The overall structure of the host computer monitoring client is planned,and the functions of the monitoring client are realized based on c#language in vs2017 development environment.The electrical cabinet fire early warning simulation experimental platform is built.Through the experimental platform,the experiments verify the integrity of the software and hardware functions of the early warning system designed in this paper,which lays a research foundation for further application research. |
