The Electric Fire Intelligent Algorithm Based On Arc Model Simulation Analysis

In recent years, the electrical fire accident in the nationwide, and people brought great harm to the society and influence of electrical fire more and more get people’s attention, in order to reduce the happening of the electrical fire, the first time found that electrical fire hidden danger, to establish a perfect arc fault detection system is particularly important.At first, this paper analyze the arc fault, able to display the information characteristics of arc fault, at the time of the arc light, electricity and other physical phenomena and arc fault current information to be able to give the arc fault messages, but the light, the physical phenomenon of electric is affected by the environment is large, it’s easy to have a miscalculation. So the fault arc voltage and current waveforms to more reasonable and accurate characterization of the arc fault. And then analyzes the existing arc fault model, through a large number of simulation analysis for Cassie arc model, get the variation, features of the fault arc current voltage waveform, this paper proposes a electrical fire intelligent algorithm based on arc Cassie model, application of artificial neural network algorithm, establish the arc fault current waveform and arc fault in a single cycle of the relationship between calorific value. In wisdom type electric safety management platform reading on the bus current waveform, the calorific value analysis to calculate the corresponding arc fault, and according to the arc to judge whether the calorific value of electrical fires. Through a lot of simulation experiments, through neural network training, in 1000 set of test data, only two groups appear a mistake, that can be predicted by arc fault current waveform of electrical fire, the first time given electrical fire hidden trouble and ruled out in time. At the same time we also adopted the wavelet transform to decompose, arc fault current waveform by comparing with normal working time of current waveform, obtained under different load, by comparing the details of the different components, can judge the fault arc.Finally, we through the comparison of the neural network algorithm and wavelet analysis algorithm, wavelet analysis can think fast process of generating arc fault early warning, but because it is difficult to through the severity of the wavelet to identify fault, will not be able to judge whether the fault arc will lead to electrical fire, and its environmental impact is bigger, the miscarriage rate is higher. And neural network algorithm, although there have been a mistake, but its accuracy is higher, and directly determine the arc fault will cause electrical fire, can be more efficient to screening of arc fault, reduce the occurrence of electrical fire.