Study On Signal Processing Scheme For Fiber-Optic Interferometric Sensors Based On Software Interrogation Method

Fiber-optic interferometric magnetic field sensor is related to optics, electronics, mechanical manufacture and signal processing technologies. Because of its high sensitivity, large dynamic range, good capability to electromagnetic immunization, it have been intensively studied.Fiber-optic interferometric magnetic field transducer has high sensitivity, which brings forward strict requirements to signal processing scheme. A well designed signal processing scheme should have high quality of recovery capability of sensed signal. However, the traditional hardware detection method lacks of agility and has some weakness such as bringing in electronic noise. Based on these considerations, this thesis introduces a kind of signal processing scheme, which is based on software to decrease noise and improve the performance of the sensor.Eliminating the system noises is the key issue of signal detection since those noises would cause the phase drift of the output signal. Fiber-optic interferometric magnetic field sensor noises include environment-induced noise and optical source noise, fiber-optic thermal noise, short noise of optical fiber and photodiode detector noise. After studying those noise resources, it has been found that thermal noise and short noise of optical fiber are the main factors on sensitivity(minimum detectable phase). In this thesis, an adaptive line enhancement is proposed to reduce the above two kinds of noises. The adaptive line enhancement is a method based on time delay and iteration and we use it to effectively eliminate the system wide band white noise such as optical thermal noise and shot noise. In the frequency domain, the adaptive line enhancer is a adjustable narrow band pass filter, which has advantage over hardware filter since it can adaptively track the sensor working frequency. The experimental results show that, the introducing of adaptive line enhancer can decrease the system background noise by 13dB. The interrogation linearity is improved and the fiber-optic sensor sensitivity reaches 5.22×10 ?3mV nT.Further more, software based signal processing method is convenient to change parameters depending on specific applications. With the trend of sensor networking, the software interrogation based signal processing system could be designed and used as a virtualized instrument, which can remotely perform the real time measurement in the future.