Research On Optical Fiber Intelligent Sensing Technology And Safety Monitoring Model For Cracks In Hydraulic Concrete Structures

There are various types of defects in a large number of water conservancy proj ects in China,among which cracks are common defects in hydraulic concrete structures.Cracks not only affect the durability and beauty of the structure,but also seriously damage the integrity of the structure,leading to deterioration of the stress state of the structure and weakening the bearing capacity of the structure.Therefore,the monitoring of cracks in hydraulic concrete structures is of great significance to ensure the safe operation of structures.However,the traditional electrical sensor generally has some shortcomings,such as poor durability and vulnerability to electromagnetic interference,which make it difficult to meet the actual monitoring requirements of engineering.Therefore,it is urgent to study new sensor technology to adapt to the complex monitoring conditions of hydraulic concrete structures.At the same time,in the research of crack safety monitoring of hydraulic concrete structures,the traditional statistical model is vulnerable to multiple collinearity,and the neural network model is vulnerable to fall into local minimum,so the prediction effect is not good.Therefore,it is necessary to study the safety monitoring model for cracks in hydraulic concrete structures.In view of the shortcomings of traditional crack monitoring sensors and traditional statistical models in practical applications,a new type of fiber optic crack sensor based on fiber bending loss principle and a crack safety monitoring model of hydraulic concrete structures based on random forest algorithm are proposed.The main contents are as follows:(1)Based on the principle of optical fiber bending loss,a new type of optical fiber sensor for crack monitoring of hydraulic concrete structure is proposed.The structure and composition of the sensor are introduced,and the working principle of the sensor is analyzed.The theoretical relationship between the optical loss of the sensor and the crack width is deduced.The feasibility of the sensor used in structural crack monitoring is verified by experiments,and the technical specifications of the sensor are determined.The performance of the sensor is tested by experiments.(2)Aiming at the problem of laser source jitter,an improved fiber optic crack sensor based on splitter is proposed.The causes of laser source jitter and its adverse effects on the measurement results of the sensor are analyzed theoretically.The principle of the improved method is expounded.Experiments verify the effectiveness of the improved method in avoiding the adverse effects of light source jitter on the fiber optic sensor based on bending loss principle.Finally,the feasibility of applying the fiber optic crack sensor based on the principle of fiber optic bending loss to the crack monitoring of hydraulic concrete structures is further verified by the crack monitoring experiment of reinforced concrete beams(3)A crack safety monitoring model of hydraulic concrete structure based on Stochastic Forest algorithm is established.The measured crack data are selected,and the training model is built through training set,and the test set is predicted.The prediction results are compared with those of neural network model and least square method respectively,which proves that stochastic forest model has higher prediction accuracy.