화재 감지용 자외선 적외선 불꽃감지기의 특성 개선에 관한 연구

Abstract

This thesis focuses on the problem of fire detection which sense radiant energy in ultraviolet and infrared range. The flame which are caused from detects radiant energy from the flame of fire fuels containing carbonaceous material, liquid and gas fire, automobile fires in tunnel, a high ceiling building like a hanger, atrium. The flame detector can be classified are ultraviolet detector, infrared detector and UV/IR detector. In this research, we propose a method to reduce false detection. The fire detector consists of driving circuit and signal processing unit which includes microprocessor and filtering procedure. Ultraviolet sensor is called a UV Tron which uses gas multiplication effect on current discharge and photoelectric effect of metal. To have high sensibility and to gain proper output voltage, it has high sensibility. The infrared sensor makes use of black body radiation theory based on pyroelectric infrared sensor(PIR) As for the driving circuit, UV Tron using discharge effect requires high dc voltage to discharge start voltage. Driving circuit uses DC-DC convertor to obtain high voltage, because input voltage of only DC 24V±20% is used. To obtain high voltage of DC 330V pulse width modulation(PWM) method and small transformer are used. Driving circuit of PIR sensor make use of cascade operational amplifier to amplify tiny voltage in superconductor polarizing action. Output signal of UV and IR sensor include analog signal as well as digital signal. UV sensor output is digital pulse caused from alternating discharge. IR sensor output through amplifier is analog signal continuous in time. Signal processing of UV Tron has three stages. First, output signal accumulates in timer/counter port of a microprocessor. Next, the accumulated value is compared with the reference value. If sensor signal is over the reference value, a processor regards it as fire. In the signal processing of PIR sensor frequency analysis is used. First, PIR sensor signal is converted to digital signal using a analog-digital converter. Next, The converted data is analyzed in frequency domain using discrete fourier transform(DFT) and other methods. Last, if the frequency is in the range of 2～20Hz which corresponds to flame flicker frequency, the processor recognizes it as fire. The proposed driving circuit and signal processing methode is tested extensively. First test is fire detection test at distance and angle certificated KOFEIS and FILK approval. The result propers the prevention and extinction of fire technique degree, certificated operation of detector. Next test is false alarm test at source of lights, heater, arc welding, sun lay etc. The experimental results show the proposed method is stable fire detection and false alarm test for certification.This paper presents a wireless sensor network system in which weight measuring instrumentation process, data conversion process, and the series of signal controlling processes are seamlessly cooperative. 16 strain gauges are working as incoming sensors. Each output is amplified and filtered for proper analog signal processing. Several measuring instrumentation OP amps and general-purpose OP amps are used. 12 bits A/D converter transforms analog signal to digital bits. PIC microprocessor controls the operations of measuring strain gauges. RF RS232C module is adapted for wireless communication between the PIC microprocessor and TCP/IP hosting device. Service program is considered as seamless and stable monitoring operation tool over Internet. Database-based data handling system could bring firm basis for statistical data processing application. In order to overcome the time sequence problem among different device parts, a total communication system using handshaking method has been developed.