Study On Optical Fiber Sensing Theory And Method Of Stray Current In Urban Rail Transit System

An urban rail transit system is based on urban passenger transport, which helps to decrease the urban traffic pressure. The security operation of urban rail transit system is important for the urban daily travel and the healthy development of city economy. The stray current corrosion has been found on the buried metallic pipeline of urban rail transit system. However, the traditional evaluation method of stray current corrosion risk has the low long-term accuracy. Thus, this dissertation focuses on the stray current measurement of buried metallic pipeline based on the optical fiber sensing technology.The formation mechanism, corrosion mechanism and corrosion characteristic of stray current on the buried metallic pipeline of urban rail transit system are analyzed. On this basis, the traditional evaluation method of stray current corrosion risk is studied based on the polarization potential deviation measurement of buried metallic pipeline. And the fundamental causes of its low long-term accuracy are demonstrated based on the intrinsic property and measurement principle of the reference electrode. Moreover, the feasibility of corrosion risk evaluation method by measuring the stray current is illustrated, which is based on the magnitude and randomness of stray current.According to Faraday magneto-optical effect, polarization theory and Jones matrix theory, an optical fiber measurement system of stray current on buried metallic pipelines and its optical model are built. In addition, the S-D detection for the system output signal process is proposed. With the heterodyne detection as a reference, the S-D detection has positive effect on the system sensitivity improvement and the extinction ratio error suppression of the polarizer. Moreover, the effect of alignment error between two optical elements on the system output is illustrated. And the elimination method based on polarization controller modulation is proposed. Its feasibility is proven by the simulation results. The actual measurement system is set up and the sensitivity test is conducted, which shows that its sensitivity is about 0.0261/A. In addition, according to the test results in zero point, the modulation error of the polarization controller can be neglected in the measurement system with the S-D detection.According to elastic-optic effect and elastic mechanics theory, the linear birefringence of sensing fiber induced by the temperature variation is calculated. And the measurement method of geometric linear birefringence is proposed based on a Faraday rotator. On this basis, the output error mode of measurement system is built to demonstrate the effect of linear and circular birefringence of sensing fiber on the output error. Moreover, the cylinder twisted fiber is proposed to produce the circular birefringence. And the mapping relationship between the characteristic parameters of cylinder twisted fiber and the produced circular birefringence is built based on the geometric-optic rotation effect. Thus, for minimizing the output error, the characteristic parameters can be determined. Furthermore, the manufacture steps of cylinder twisted fiber are demonstrated in detail. And a temperature test is conducted to prove the effectivity of cylinder twisted fiber.The Allan variance calculation of the stochastic noise in the measurement system is determined by comparing its calculation accuracy and efficiency based on different subset division methods. Moreover, the statistical properties of stochastic noise are analyzed based on the quantitative relation between its Allan variance and power spectral density. Furthermore, the shortcomings in the tradition identification method of Allan variance theory(i.e., the least square method) are illustrated. And the novel identification method based on genetic algorithm is proposed. The identification accuracy is tested based on a series of the simulated stochastic noise. Finally, the genetic algorithm is applied to identify the stochastic noise in the measurement system.The important performance parameter test of the measurement system is conducted. And the test results are shown as follows: the repeatability is about 0.367%; the measurement accuracy is not greater than 0.4% at 4 A and 8 A; the measurement accuracy is not more than 0.35% at 12 A; the measurement accuracy is smaller than 0.3% at 16 A and 20 A. Moreover, if the output power of SLD source decreases, the negective effect of stochastic error will increase and the Verdet constant of sensing fibers will decrease, which will both decrease measurement accuracy.