Evolution Rule And Abnormality Diagnosis Of Crack Of Important Hydraulic Concrete Structure

Aiming at crack problems in important hydraulic concrete structures, the cause, evolution rule and abnormality diagnosis of the crack are studied in this dissertation by using advanced theories and mathematics and mechanics, such as rough set theory, data mining, neural network, wavelet analysis, chaos theory, and so on, combining with prototype observation data. The main contents are as follows.(1) Mining of cause of crack is proposed by using logic diagnosis, rough set theory, and data mining technique. Firstly, The possible cause of crack is analyzed by using fault tree method. Then, The data mining approach to the cause of the crack is presented using rough set theory, the rule of the cause of the crack is extracted, and the main cause of the crack is analyzed.(2) The simulation of temperature field is discussed. The thin layer element and smeared crack model is studied. The effect of load on the crack is discussed using numerical simulation approach. Then, the data mining approach to the unfavorable load combination of the crack is proposed using artificial neural network. The rule of the unfavorable load combination of the crack is generated and three dimensional load combination is given.(3) The contact model is used to simulate the crack. On the basis of project, the evolution of crack opening and crack depth related to loads such as hydrostatic pressure and temperature are studied using numerical simulation approach, and the load effect on the crack evolution is discussed.(4) According to observation data, the approach to the time varying rule is proposed using wavelet theory and phase space reconstruction, the evolution of the crack opening varying with time is studied and proved by a project example.(5) On the basis of extracting the time-dependent component using wavelet analysis. Phase space reconstruction technique is used to reconstruct the phase plane. The approach to detect the abnormal point of the tune-dependent of the crack by phase space is developed. Combined with gray system and catastrophe theory, according to observation data of dam cracks, a gray cusp catastrophe model is established to diagnose the abnormality of the crack.(6) The abnormality criteria of fracture mechanics are employed to diagnose the abnormality of the crack, and the inversion of the concrete fracture roughness is studied.The space prediction model of the anomalous load is presented using neural network. The load surface of stress intensity factor and its corresponding plane is given and employed to diagnose and predict the abnormal load of the crack.