Early Stage Fire Warning of Substations Based on Concentration Prediction of Thermal Ionization Particle

Due to the imperfect fire monitoring and early warning measures, it is difficult to find out and take corresponding measures to deal with the fire in the early stage. A method of substation fire identification based on thermal ionization particle prediction is demonstrated in this paper. Firstly, the influence of environmental gas fluctuation on fire identification is eliminated, the collected data are extracted, and the particles of thermal ionization particle are screened out by distinguishing wavelength, light transmittance, and light spot dissipation speed. Then, a neural network model for predicting the concentration of thermal ionization particles is established, and the selected thermal ionization particle data is input into the neural network model to predict the variation curve of thermal ionization particles in the substation in the future. Taking the exponential change of thermal ionization particle with time as an important feature, it is judged whether the concentration change curve of thermal ionization particle is consistent with the exponential change of fire, and if so, a fire warning signal will be sent out to achieve the purpose of a timely warning.