Research On Fire Monitoring System For Hazardous Stereoscopic Warehouse Based On Raman Scattering

The stereoscopic warehouse has become a major trend in the storage of goods.Because of the corrosion,toxic,flammable and explosive properties of dangerous chemicals,the stereoscopic warehouse which storing hazardous chemicals will cause great loss of safe property if there is a fire accident.Therefore,it is imperative to design a fire warning system for stereoscopic warehouse.According to the requirements of national standard,temperature is used as the criterion of fire warning,a fire warning system based on Raman scattering is designed for hazardous goods.The system includes temperature sensing fiber,data acquisition module and PC.The temperature sensing optical fiber of 62.5/125um is wound on the shelf.The data acquisition module collects the signal light produced by the sensing optical fiber.After the signal is filtered and photoelectric converted,the information is transferred to the PC.PC demodulates the collected data,designs the alarm strategy and completes the fire location and early warning task.In view of the attenuation of the optical scattering signal,a two-direction loop temperature demodulation method is proposed.The laser head inject the pulse signals from both ends of the fiber.The system collects the forward and back forward scatter signals generated by each point on the optical fiber,and brings into the modified temperature demodulation formula,the attenuation function related to the distance in the classical temperature demodulation formula becomes constant.Combined with the dynamic mean de-noising method to further suppress the noise of the APD circuit,The method can reduce the temperature error from 1 to 0.5?.In view of the random noise in the system,a wavelet segmentation de-noising method is proposed.The data from the temperature measurement structure are divided into the normal temperature section and the high temperature section.The classical wavelet de-noising method is adopted in the normal temperature section.The puseudo-Gibbs phenomena is suppressed by the translation invariant wavelet processing at the high temperature section.An improved threshold function is designed to further optimize the wavelet de-noising effect.Compared with classical wavelet de-noising algorithm,SNR of segmented wavelet de-noising in 10-30dB is about 2dB higher than that by using classical wavelet de-noising.