Research Of Diagnosis Method Of Damages Upon Shalloe-Covered Structures Dased On Neural Network

As the construction scale of civil air defense engineering expands, it is in urgent need of solving the important problems of damage diagnosis and safety evaluation on the civil air defense engineering. This thesis analyzes how to diagnose the damages of shallow-covered frame structures by combining the analysis approach of dynamic properties of underground structures with the theory of diagnosing damages on the buildings above ground, and by combining the structure vibration characteristics with neural network. The research is mainly consisting of the following:(1) A finite element model under the interaction of dynamics from underground structures was set up with the common finite element analysis software named Marc, by using the method of dispersing dynamic from the system consisting of structures and foundations, so as to analyze the dynamic properties of underground structures. The characteristic quantity of damage diagnosis DEΦ reflecting the sensitivity on the damage of shallow-covered frame structures has been proposed after computation and analysis.(2) The damage degree was characterized as per the decrease of elastic modulus on local damages, and a standard spectrum with dynamic finite element algorithm on the diagnosis of damages upon shallow-covered structures was set up, i.e. to set up a coincidence relation between the characteristic quantity of damage diagnosis DEΦ and the state of damages (damage positions and degrees). The neural network is trained by the standard spectrum and the damage is therefore diagnosed.(3) The neutral network method on diagnosing the damages upon shallow-covered frame structures was proposed, and an ameliorated...