Research On Reverse Identification Of Wind Loads On High-rise Buildings Based On Dynamic Response

In general,the wind load is difficult to achieve through field measurement.Although it is able to obtain the structural wind loads by the wind tunnel test and numerical simulation,the accuracy of wind load assessment results need to be further verified due to the flow field simulation techniques and calculation methods.Considering the fact that the measurement technology of wind response is relatively mature,the measurement accuracy of structural acceleration and displacement response is much higher than that of load,so it is a practical way to identify the wind load based on wind response.In this paper,two methods of wind load identification for tall buildings are proposed.Wind tunnel tests of Hengqin international financial center building and typical square section building model were carried out.The engineering overview,test equipment and flow field simulation of high-rise building were introduced in detail,which laid a foundation for the follow-up method verification work.Based on the pressure-type wind tunnel test,the test data are transformed corresponding to the actual structure,and the wind pressure integral of the measuring point is obtained.The dynamic wind load of each floor is calculated by linear interpolation method,and the wind-induced response is calculated.Finally,the limited floor responses are used to conduct identification research.The accuracy of the DKF method is verified by wind tunnel test of Hengqin International Financial Center Building.The algorithm uses only six floors of acceleration responses or displacement responses to identify the wind loads of each floor and wind-induced response time-history,and the results are satisfactory.The recognition accuracy of the method based on Kalman filter and least squares method(KFLS)was evaluated by wind tunnel test of a typical square section building.The algorithm identified the results from the displacement responses of six floors,and the results showed that the wind load and wind induced response of the structure based on the displacement response are in good agreement with the true results.And the characteristics of the two recognition methods are compared.The optimized layout of the response points corresponding to the floors is studied.The optimal arrangement of measuring points at the first few modes is given,and the comparison of the measured points on different floors is given to identify the wind loads.It is proved that the optimized measuring floor placement is important for the identification of wind load.The effects of parameter estimation errors on wind load identification results were investigated by increasing or decreasing magnitude of the structural modal parameters(including natural frequency and damping ratio).The results show that the influence of modal parameter error on inversion accuracy is negligible,which means that the two methods are insensitive to the estimation error of modal parameter.The results of wind load and response estimation of the state vector with large initial value are given.The results of the recognition are in good agreement with the accurate results.It is proved that the tolerance of the initial value of the state vector is better than that of the two methods small.The L-curve method is used to calculate the magnitude of the covariance of the measured noise,and the calculated value of _vR is used to identify the wind noise accurately under the measurement noise.And this method overcomes the problem that the statistical characteristics of noise must be known in advance by using the Kalman filter theory,and the above results show that the method has some anti-noise ability,and the recognition results can meet the needs of practical engineering.