Damage Detection Of Shell Structures Based On GIS And Artificial Neural Network

The GIS has been wildly applied in structural health monitoring system GISSHM (Geographic Information System for Structure Health Monitoring) and wildly applied on the health monitoring of large bridges and arch dams。In this paper, following the works of"the GIS-based structural health monitoring system"in the task team, one double-layer spherical lattice shell as a example, a program of health monitoring and damage detection coupled with the functions of GIS specially is designed, The program has powerful function of visualization and can analyze the monitoring data of multi- sensor by visual graph。Firstly, the method, the tools and the functions which is based on the GIS in designing Health Monitoring System (DDS) is analysised and the structure of datas in this system is brought forward in this paper。with one double-layer spherical lattice shell as a example, the programme modules is designed, which is to link the geography information and property information, thus to make all types of information visualized, and to label the datas on the map in the style of statistics chart realized by the function of theme particular to GIS, thus to make analysis more directly and vivid.Then the technique of damage detection is coupled with the GIS-based health monitoring system, the technique of damage detection is mostly based on mode parameters of structures and the system of damage identification is formed including mode identification, damage detection and damage displaying. The GIS-based structural health monitoring is applied on the health monitoring of double-layer spherical shell。Based on the finite element model and damage dectection analysis of a 120m span double-layer spherical lattice shell, the artificial neural network is trained by the mode parameters. the fequence and the mode parameters at some keypoints before and after the damage are used to make the damage indexs to train the artificial neural network, there are satisfying precisions in identifying the damage position and the damage dagree,and the net has a ability of self-study when being used in identifying unkown damage types。...