Extreme Response Analysis Of Transmission Tower Under Non-Gaussian Wind

The damages of transmission towers are mainly caused by winds. The turbulence winds in current design standards are usually assumed as stationary Gaussian stochastic processes. The Gaussian assumption of wind inflows is appropriate for boundary layer wind fields associated with flat and homogenous terrains. However measured data have shown that turbulence wind in complex terrain exhibits non-Gaussian characteristics.The assessment of extreme values of stochastic structural response with given mean recurrence intervals by adequate modeling and characterization of loads is essential for a reliability- and performance-based structural design. But this is not reflected in current transmission tower design standards. Transmission tower extreme responses under both Gaussian and non-Gaussian wind are investigated in this paper, the main contents includes:The Gaussian wind inflows are simulated by spectrum representation method. the non-Gaussian wind inflows with given power spectrum density function, skewness and kurtosis are simulated by translation theory, the translation function of this method is modeled by Hermite function. The turbulence wind time history samples of multiple mean wind speeds are simulated with 50 samples under each mean wind speed. The wind loads are calculated according to the wind speed time history samples,and transient analysis of a finite element model of a 500 KV transmission tower established by ANSYS is investigated.The power spectrum density function,mean value, variance, skewness, kurtosis of the responses at each mean wind speed under both Gaussian and non-Gaussian wind inflows are calculated, their variations with mean wind speeds are investigated. The short-term 10-min extreme response distributions conditional on various wind speeds are determined using global maxima method and ACER method. This extreme response distributions conditional on various wind speeds are then combined with the distribution of mean wind speed for the evaluation of overall short-term extreme response, a Rayleigh distribution is used in this paper. Under the assumption that the short-term extremes are mutually independent, the extreme values with various given mean recurrence intervals are determined, for instance,the 50-year extreme response corresponds to a probability of exceedance of 3.8×10-7 in terms of 10m min extreme. The results shows that long-term extreme values under non-Gaussian wind inflows are larger than the results under Gaussian wind inflows, so adequate modeling of the wind inflow is important for the safety of the transmission tower.