Damage Identification Method Of Space Truss Structures Based On Strain Mode And RBF Neural Network

With the characteristics of large stiffness, light weight, rational force and well earthquake resistance, etc., space truss structures are widely used in industry and civil buildings. These structures inevitably emerge damage, aging and even seriously collapse with using time passing. In order to timely find the potential safety hazard and decreasing the fund of blind maintenance, it is important to detect the truss structures. Therefore, firstly damage location identification and damage degree judgment have been studied in this paper. Besides space truss structures are made up of multiple links, nodes and big degree of freedom, dynamics testing signal often appear repeated frequencies modes and closed frequencies modes. While testing signals directly related damage information, some research about single point optimal excitation position is considered to avoid this phenomenon.In short, some studies and conclusions in this paper are as follows:1) A method based on change of strain modal and RBF neural network has been obtained to identify damage location and judge damage degree of link elements. In this method, the link elemental axial difference of strain mode was considered as damage identification index. Using this index, the damage locations of link elements can be identified accurately. At the same time, thinking this index as training samples of RBF neural network, the damage degree of link elements can also be determined. At last, this theory was verified by a square pyramid truss structure stand triangular honeycomb cones truss structure. This process was realized by software Ansys and Matlab. The results show that this method can more accurately identify link element damage of truss structure.2) Single-point excitation was the common method to receive dynamics testing signals, and the excitation point directly affect the damage information. In this paper, the expression of structural free vibration response was gained under the action of single point excitation. Based on the expression, the optimal incentive position was obtained. Inciting this optimal incentive position, well dynamics testing signals will received without the phenomenon of repeated frequencies modes and closed frequencies modes. Finally, the effective of single point optimal incentive position was tested by two examples same as above.