Research On The Bridge Damage Identification Based On Curvature Mode And Extreme Learning Machine

In view of the importance and particularity of the bridge structure,it is easily affected by the external environment and the aging of its own materials during use.Safety issues cannot be ignored.It is particularly important to accurately detect and prevent major damage to the bridge structure in a timely manner.In recent years,machine learning technology has been widely used in all walks of life,but in the field of bridge structure damage recognition,it still needs to be further improved and developed.As an application of machine learning technology,extreme learning machine is superior to classification and regression problems.Traditional learning algorithm.Therefore,based on the bridge dynamic parameters,combined with the intelligent damage recognition technology of the extreme learning machine,the bridge damage is studied.The main work contents are as follows:First,the research background and significance of bridge damage identification are described,and the development of various parameter labels and damage identification methods that affect bridge damage identification at home and abroad and the existing problems are summarized.Machine learning methods and The combination of pattern recognition methods has advantages for bridge damage identification,and the extreme learning machine is selected as the reason for bridge damage identification analysis.Then use ANSYS to establish the finite element model of the simply supported beam bridge,select the three indexes of natural frequency,displacement mode,and curvature mode,and compare and analyze its adaptability,accuracy,and sensitivity to find that the curvature mode is the best Damage identification index.Taking the actual three-span girder bridge as an example,the curvature mode is used as the input parameter of the extreme learning machine,the classification model of the extreme learning machine is selected to identify the damage of the bridge’s position(single position and multi-position),and then the regression is selected by the extreme learning machine.The model identifies the degree of damage to the bridge and determines the best activation function by comparing different activation functions.Finally,the recognition rate is used to verify that the extreme learning machine model is good at recognizing the location and degree of damage.In summary,the curvature mode can be used as a parameter index for bridge damage identification.The selected extreme learning machine classification model can also be used for bridge damage location.The extreme learning machine regression model can also predict the degree of bridge damage.It shows that the extreme learning machine has a very broad application prospect in the field of bridge damage identification,but the specific theoretical methods still need to be further improved.