Research On Modal Parameter Identification Of Super High-rise Buildings Under Strong Win

With the continuous development of economy and technology,supertall buildings have gradually become the landmark buildings in various cities in China.Field measurement is one of the reliable methods to obtain the wind-induced response of such structures under the action of strong wind(typhoon).The use of measured wind response to identify and study the dynamic characteristics of the structure at work is of great significance and value to the development and improvement of wind resistance theory and design of supertall buildings.In addition,although supertall buildings are designed for a long service time,their final failure form is still difficult to be determined,so the structural damage needs to be monitored regularly,and the structural damage is often reflected by the changes of structural dynamic characteristics,so the tracking study of the dynamic characteristics of such buildings can help to ensure their safe use.Moreover,the measured structural dynamic properties can be used for model correction.Up to now,most of the time domain modal parameter identification methods are based on certain ideal assumptions,and there may be various factors that conflict with the assumptions in practice,which may affect the applicability of the methods.Therefore,investigating the robustness of such methods and optimizing their identification process can help the study of modal parameters identification of supertall buildings under strong wind.Considering the possible interference factors in the actual measurement process,this paper first investigates the effects of noise interference,measurement direction,and parameter changes on the recognition results of the mainstream time-domain modal parameter identification methods such as NEx T-ERA,SSI-DATA,SSI-COV,and BMID by using numerical simulation methods.The improvement is also carried out to address the noise modal interference introduced by the high-order calculation of the stability diagram algorithm used to assist in the screening of recognition results.Subsequently,the acceleration response of two supertall buildings during typhoon transit is studied,and the modal parameters are identified using the BMID algorithm,SSI-DATA,and SSI-COV in combination with the improved stability diagram algorithm.The details are developed as follows:(1)First,the principles of the methods covered in this paper are briefly introduced,including the four types of NEx T-ERA,SSI-COV,SSI-DATA,and BMID.The auxiliary tools of the first three methods are also introduced,and the basic principles of the stable diagram algorithm based on potential-based hierarchical agglomerative(PHA)are presented.(2)Numerical simulation experiments were carried out to consider the effects of noise interference,the existence of parameter variations in the structure itself,and response measurement angle deviations on the recognition effects of each method.The results show that the addition of noise makes the average error of the recognition results of NEx T-ERA fluctuate greatly and the stability of the algorithm is poor.In contrast,the remaining three methods,SSICOV,SSI-DATA,and BMID,have better noise immunity and can restore the trend of the modal parameters over time.In addition,the change of the measurement direction of the response has a major impact on the recognition results of these three algorithms in the direction of the lower amplitude in the modal vibration pattern,and the impact on the overall vibration pattern is not obvious.BMID,however,is limited by the number of measurement points and cannot completely separate all the modes,thus affecting the results to a certain extent.(3)Improvement methods are proposed for the problems of stable noise modes and missing stable axes caused by the high-order calculation in the stability diagram method.This paper proposes to downsample the collected response data by low-pass filtering to obtain multiple downsampled data of the same time period,and use its identification results to replace the highorder calculation results,so as to reduce the calculation order requirement.In addition,the noise and physical mode differentiation and order discrimination are achieved by parameter settings.Finally,the optimized algorithm is validated by numerical simulations,and the results show that the required computational order is reduced after the improvement,thus avoiding the negative effects and improving the stability of the algorithm.(4)Hitachi Elevator Test Tower and Leatop Plaza were taken as the research objects respectively,and the time-varying phenomenon of their structural dynamic properties during typhoon impact and its relationship with amplitude were analyzed based on three methods: SSIDATA,SSI-COV and BMID,and the results were analyzed probabilistically.The results show that the modal frequency of the super high-rise building decreases with the increase of amplitude,while the damping ratio only has a discrete distribution within a certain range at low amplitude and does not show an obvious trend of change,while it increases with the increase of amplitude when the amplitude is larger.By comparing the amplitude of the two buildings at large,medium and small amplitudes,it is found that the modal vibration pattern identification results do not change significantly.In addition,the modal frequencies of each order at low amplitude basically obeyed normal distribution,while the type of modal frequency distribution at high amplitude was still open to question.Compared with the log-normal distribution,the results of fitting the damping ratio with the Gamma distribution may be more generalizable.