| There are a number of landslides that developed and revived on the bank of domestic and abroad reservoirs under the periodic fluctuations of water level. Compared with landslides that developed under other circumstances, the reservoir landslides cause more different forms of disasters and the area that affected is larger. Slope monitoring is one of the primary means to prevent the landslide hazards and learn the failure mechanism of slope. It is of significance both in theory and application to develop the monitoring method and do some research on the stability of slope.The bank slope is a complex multi-field system which includes seepage field, stress field, temperature field, deformation field, et.al. Conventional methods for slope monitoring have numerous drawbacks and are difficult to meet the requirements of fully obtaining、analyzing and evaluating the multi-field information during the evolvement process of landslide. The newly developed distributed fiber optic sensing technology (DFOS) which offers a number of attractive features can achieve continuous acquisition of multi-field information in slope that overcomes the drawbacks and break the limitations of point monitoring methods. In this paper, a series of experiments tests that based on three kinds of distributed and quasi-distributed optical fiber sensing technologies, FBG, BOTDR and ROTDR were carried out both inside and outside the laboratory to monitor the different fields of slope. A DFOS based monitoring system was established in Majiagou slope of Three Gorges in China. With the data that monitored within two years, a data mining technology based on association rules was used to analyze the relationship between different fields. Besides, the stability of the bank slope is analyzed with a numerical model and evaluated with catastrophe theory. The research results are presented as follows:(1) Three commercial DFOS systems, BOTDR BOTDA and BOFDA, were tested in the lab. The spatial resolution of PPP-BOTDA analyzer from the company Neubrex is better than the BOFDA analyzer from the company fibrisTerre GmbH, while the repeatability of them is nearly the same. Besides, there is a restrict relations between the spatial resolution and the repeatability for the devices that with the improvement of spatial resolution the measurement repeatability becomes worse. Their multiplied value is about 0.2MHz*m which corresponds to 4με*m. BOTDA and BOFDA can achieve refined measurement, however, both of them need loop setup which increase the risk of failure for the monitoring system. BOTDR with the ability of single-end detecting is more suitable for geological and geological disaster monitoring regardless that its performance is a little worse than BOTDA/BOFDA.(2) The experiment results shows when the length of fiber that attached on the surface of the structure is longer than 15.8cm, the length of low strain transfer of optical fiber is lower than 7cm. The length of sensing fiber installed on the surface of structures should be long enough. And the data that monitored around the ends of fiber should be excluded since the error is usually large.(3) The error estimation formula for DFOS based displacement meter was obtained. It shows that the displacement error is proportional to the square of length and inversely proportional to the diameter of the tube, which was well confirmed in the experiments. Besides, this formula can be used to analyze the displacement error of other beam structures that monitored with the similar DFOS based method.(4) A novel metal-reinforced optical cable was designed and improved for distributed strain sensing. Different kinds of temperature sensor, osmometer, liquid level gauge, reinforcement stress meter and earth pressure gauge that based on FBG were developed and tested. A DFOS based multi-field monitoring system was designed, including the selection of DFOS based technique, the process of installation for the optical fiber sensors, the multi-field data characterization、 processing and analysis.(5) A fiber optic sensing based multi-fields monitoring system was set up for Majiagou slope that near Three Gorges reservoir. Two years of multi-fields data of this bank slope was successfully obtained. The sliding planes and the sliding boundary of the slope were accurately positioned, and the value of displacement was calculated as well.(6) A data mining method that based on Apriori algorithm was used to extract the association rules of multi-fields information. It is found that the reservoir water level is the main factor that affected the deformation of Majiagou slope. Overall, the higher of water level is the smaller of the deformation rate will be. And the deformation rate increases when the water level is low.(7) A numerical model was established to analyze the effect of seepage field on the safety of Majiagou slope. Safety factor under different reservoir water level and changing rate was obtained. It shows that the slope safety factor is large when the water level is high and small when the water level is low. There is an obvious zoning phenomenon that the lowest safety factor appeared when the water level was around 150m. The safety factor would reduce rapidly when the water level fell into the scope of [152 m,155 m] and the change rate into the scope of [-0.45m/day,0.26m/day]. The monitoring results of multi-fields confirm this conclusion as well.(8) Nonlinear dynamic equations of Majiagou slope were established with both displacement and strain data and they were proved equivalent Particle swarm optimization (PSO) algorithm was used to improve the accuracy of the model. Based on the catastrophe theory, it is concluded that Majiagou bank slope is not in the stage of mutation development and it creaps steadily. |
PDF Full Text Request