| Urban railway system has gradually palyed an important role in realizing city trafficsustainable development as a large capacity public transport mode. And it is anoteworthy problem that urban railway system brings about underground metal structurecorrosion nearby the railway, therefore, this dissertation researches optical-fiber sensingmethod to monitor metro stray current corrosion, which is a key subject in theconstruction and operation of urban rail transit and which is also important theoreticallyand practically for healthy development of urban rail transit.Firstly, the formation of metro stray current and corrosion mechanism is analyzedand the essential rule is researched of underground metal structure corrosion under theaction of stray current, which provides basis for monitoring metro stray current. Anddifferent corrosion environment on the influence of metal structure corrosion rate isresearched through stray current corrosion experiment. Moreover, on the basis ofanalyzing existing stray current corrosion monitoring method and its existent problem,the neotype stray current monitoring method is proposed in view of optical-fiber sensingand it turns out that this method is viable to monitor directly stray current corrosion.Secondly, the sensing characteristic of optical-fiber current sensor is researched andvarious error of affecting accuracy of system measurement is analyzed. Then, integratedtheoretical model of double-solenoid optical-fiber current system is established makinguse of Jones matrix based on Faraday effect. Aiming at the Verdet constant measuring oflow birefringent optical-fiber, a fairly accurate Verdet constant is obtained throughadjusting the location of analyzer. Furthermore, when different current value is inflowedto the optical-fiber current sensing system, the correlativity of the current input andsensing system output is established and the errors are analyzed.Thirdly, the refraction law of optical wave transmission in metal is analyzed takingadvantage of maxwell’s equations, and the amplitude reflection coefficient andtransmission coefficient of the interface is gained between optical-fiber and metal cladutilizing Fresnel formula, through which the transmission characteristic of metal cladoptical-fiber can be received. Moreover, Fe-C alloy film is prepared using chemicalplating and electroplating. Then, the coating and clad of the optical-fiber is wiped off byphysical method and chemical process respectively. Afterwards, when preparing thechemical nickel-plating, the orthogonal experiment is used to discuss the best ingredientand tecnology condition of chemical nickel-plating after the sensitization andradioactivation of optical-fiber surface. Afterwards, selecting appropriate cathode-current density and plating time to prepare Fe-C alloy film which can satisfy the sensingperformance requirement, the grain uniform and smooth Fe-C alloy film surface isobtained. Lastly, the consitituent and architectual feature of Fe-C alloy film is analyzsein microcosm, and the result declares that the film quality of Fe-C alloy film prepared byelectroplating satisfies basically the sensing performance requirement.Fourthly, fractal geometry dimension can be as an important characteristic parameerto describe the corrosion morphology of optical-fiber corrosion sensor. Throughexperiment research, the optimum thickness of Fe-C alloy film is discussed to be fit formetal corrosion monitoring. Next, statistical quality control method is used to evaluatethe performance of optical-fiber corrosion sensor. Meanwhile, through synchronouscorrosion experiment between sensors and metal structure, the quadratic polynomialfunction is fitting to obtain the relationship between the output light power ofoptical-fiber corrosion sensor with certain film thichness and the errors are analyzed.Finally, beacause the optical-fiber corrosion sensor loses efficacy frequently prior tometal structure, the metal corrosion condition must be forecasted. In view of the randomprocess of metal structure, the corrosion condition prediction model can be establishedon the basis of markoff process random theory. Exponential function relationship isfitting between the maximum pit depth and corrosion time, and the matrix of transitionprobaility of homogeneous markoff process is acquired by means of statistical analysismethod. Then, through comparing with the measured value, the result testify that it isviable to forecast metal corrosion condition using markoff prediction model. Afterwards,the last group measuring data before losing efficacy of optical-fiber corrosion sensor isas original state to forecast the metal corrosion condition, which is consistent with thepredicting outcomes from exponential function and actual measuring result, whichexplains that markoff process can be used to forecast effectively the metal corrosioncondition. |
PDF Full Text Request