Analysis Of The Modal Parameters Of Bridge On Various Environmental Conditions

As the booming of traffic, modern bridge always be overweight for a long time, what’s worse, in this period, it will be affected by the earthquake, windstorm and other natural disasters, which will have an adverse influence of the bearing capacity and the normal operation of structure. So the monitoring of the dynamic behavior which is the basement of damage detection make sense. However, a lot research shown that modal parameters will be various with different operational condition when the vibration-based methods was applied to large scale civil engineering structure. What’s important, these ambient variations of modal parameters can often mask subtle changes caused by damage. In general, the aspects which cause the change of structural modal parameters include the condition of environment--temperature, the state of foundation and humidity, the operational condition--vehicle load and other natural sources, the error resulting from test and post-process.This paper introduced the current studies in the change of modal parameters on various environmental condition firstly, then narrated the theory of Frequency Domain Decomposition method and Natural Excitation Technique with Eigen-system Realization Algorithms method, and applied these method to the analysis of accelerated signal. What’s more, vehicle load pattern was recognized by the data of strain, and the modal parameters and the characters of temperature and humility in light-vehicle pattern were extracted. At last, these data was used to build the model for the purpose of frequency prediction by the means of support vector machine, and this model was used to analysis how characters influence frequency.This paper combined the method of pattern recognition and machine learning and applied it to the analysis of modal parameters on various environmental condition at first time. Frequency prediction model was built by the use of monitoring data, and some new knowledge was found in the period of the analysis of how characters influence frequency by the built model, that will provide a new and efficient method for the study of modal parameters on various environmental condition.