The Dimension Reduction Simulation Of Non-Stationary Turbulence Stochastic Field For Wind-Resistance Dynamic Reliability Analysis Of Communication Towers

On the basic of the 14 th Five-Year Plan,the construction of data center infrastructures has been considered as the crucial element for the development of 5G communication technology in China,thus it is necessary to ensure the security and reliability of this kind of structures for the breakthrough of communication industry.Furthermore,iron communication towers with much flexibility are extremely sensitive to wind load,especially the heavy turbulence load,therefore,it is urgent to research the dynamic reponse and wind-resistant reliability of communication towers under this special load.The simulation of nonstationary fluctuating wind velocity might be a keypoint before the evaluation of wind-resistant reliability for communication towers.Generally,the double proper orthogonal decomposition（DPOD）model could be used to obtain wind velocity sample,however the extreme amount of comsumed time,unaccepatable restriction of simualated points and the complicated mathematical derivation prevent itself from being applied widely.Thus,an efficient model for nonstationary turbulence fluctuating wind velocity is presented in this paper through the connection with the spectral decomposition theory and the wavenumber spectral theory.Firstly,the DPOD expression for nonstationary descrete multi-variables random processes is given in the form of orthogonal random variables.The WSRM equation is developed for the nonstationary two-dimension and three-variable（2D-3V）continuous random fields through the decomposition of evolutionary wavenumber spectral matrixes,which solves the drawback of the DPOD model.Moreover,the wavenumber spectral matrixes is factorized as the product of a series of matrixs including coherence matrixs,leading to the orthogonal random variable presensentation of WSRM expression based on the decomposition of coherence matrixs,which could decreases the calculated time.Then,the random function is expressed by merely 3 elementary random varibles,and the number-theoretic method is recommended to generate representative point sets for elementary random varibles,leading to the dimension-reduction expression for the DPOD model and the WSRM model.Meanwhile,the POD-Fast Fourier Transform（POD-FFT）method is introduced for the aim to ensure the acceptable simulation efficiency.Thirdly,the simulation of nonstationary turbulence fluctuating wind velocity field along the span of long span bridges and the height of communication towers are implemented by the means of the dimension-reduction DPOD（DR-DPOD）model and the dimension-reduction WSRM（DR-WSRM）model,and the effectiveness of the two models is demonstrated convincingly on the basic of the simulation accuracy.Furthermore,the simulation accuracy and efficiency of the DR-POD and the DR-WSRM over the case of different number of wind velocity point are compared to recognize the suitable engineering cases for the two methods.The wind-induced dynamic response of the communication tower is calculated under the case of extreme longitudinal wind load and the extreme turbulence wind load on the basic of the probability density evolutionary theory;The comparion of the wind-resistance reliability between the two load cases is implemented with the aid of quantitative evaluation index which is inter-storey displacement angle,demomstrating the vadility of dimension-reduction menthod in real engineering.At last,a brief recap is shown in the last chapter of this paper,together with some improvement and expectation about this research.