| In the context of rapid economic development,the incidence of fires is still high,among which electrical fires occupy the first place among various types of fires,and arcs are the main cause of them.At present,the fire safety of interior spaces in residential buildings in China is still a difficult problem,and the main reasons are divided into the following two categories: First,various electrical appliances may cause electrical failure and fire during use.Second,residential buildings tend to develop towards high-rise buildings,and it is difficult to escape in the event of a fire.Therefore,the establishment of a complete indoor electrical fire comprehensive protection system can not only ensure the electric safety and reliability,but also help reduce the probability of electrical fires,casualties and property loss.To this end,the research work carried out in this paper is as follows:(1)In order to monitor the use of household appliances,various types of load power consumption data are identified from the total household power load,an improved household appliance load decomposition algorithm based on an improved BP neural network is proposed.By using Simulink to build simulation models of multiple household appliances that run individually and simultaneously to obtain relevant waveforms.The adaptive differential algorithm was used to improve the BP neural network and it was applied to load identification under the condition of multiple electrical appliances.Compared with the unimproved BP neural network algorithm has better recognition accuracy.(2)Among family users,the main causes of arcing are the aging of equipment insulation,the existence of ablation points in switches,etc.When the arc is small or the location is hidden,it is easy to cause a fire.After the improvement of the arc model,based on the decomposition of the simultaneous operation state of multiple loads in the previous section,the loads are classified into resistive,capacitive,and inductive loads and run separately in different states.The fault arc is identified by extracting characteristic values.(3)For electrical appliances that generate fault arcs during operation,extract the fault current of the electrical appliances.The Fourier transform is used to obtain the harmonics,and the relationship between the harmonics and the heating value is established,and then the neural network is used to predict the heating value.Set a threshold according to the actual situation to determine whether the appliance has the possibility of fire,and perform a risk analysis on the appliance that may catch fire.(4)Considering the dangerous fire situation,a three-dimensional fire dynamic evacuation strategy based on intelligent algorithms is proposed.For a residential building on fire,first consider the individual evacuation situation in the building,model it in a three-dimensional environment,and improve the traditional ant colony algorithm to reduce the generation time of individual evacuation routes.Then consider the situation where multiple people are evacuated at the same time in the underground garage of the community.Based on the location of the fire source and the overall personnel evacuation situation,a dynamic evacuation strategy based on cellular automata is proposed to optimize the evacuation exit and reasonably configure the dynamic evacuation indicators in the space to make the overall evacuation process fast and intuitive.The results of the above simulation experiments show that the comprehensive electrical fire protection system can effectively check the household appliances with the possibility of fire in the house,reduce the probability of fire occurrence,and generate the best evacuation route to reduce casualties in critical situations. |
