Sensitivity analysis and estimation method of natural frequency for large cooling tower based on field measurement

Abstract

The natural frequency is a key factor for estimation of dynamic deformation and mechanical performance. As simple and effective estimation equation for natural frequency of cooling tower is absent, current investigations of natural frequency of cooling towers are basically based on finite element analyses. In this study, 38 models of a 179 m cooling tower were established by tuning key structural parameters (e.g., tower height, throat height, throat diameter, inlet height, and pillar sectional area) and dynamic characteristics of these models were analyzed. Also, effects of structural parameters on fundamental frequency and overturning frequency were investigated. The sensitivity analysis of structural natural frequency for cooling towers was executed using the perturbation method and the Latin hypercube sampling method and sensitivity factors of different parameters corresponding to different orders were obtained. Based on that, multi-parameter empirical estimation equations for fundamental frequency and overturning frequency considering weighted sensitivity factor were proposed. Then, the estimation equation of natural frequency is verified by field tests of eight cooling towers with typical tower heights and configurations. Specifically, the measured acceleration signals were pre-treated using random decrement method (RDT) and natural excitation technique (NExT) and the first 10 order natural frequencies of the cooling tower using three time-domain modal identification methods (ARMA, ITD, and STD). Finally, structural parameters of cooling tower obtained by field tests were fitted. The results indicated that the fundamental frequency of cooling towers decreases as tower height and throat height increase and increases as throat diameter, inlet height, and pillar sectional area increase. Although sensitivity factors obtained by the two methods are similar, the LHS method shows higher accuracy. The effect of tower height on natural frequency is most significant among all parameters, which means the sensitivity factor of tower height is higher than those of other parameters. The measured fundamental frequencies of cooling towers were between 0.6 Hz and 1.9 Hz and heights and configurations of cooling towers have significant effects on their dynamic characteristics. The field test results of eight cooling towers demonstrated reasonable effectiveness of the proposed empirical estimation equations for the fundamental frequency and overturning frequency of cooling towers (the maximum goodness of fit of fundamental frequency and overturning frequency were 0.996 and 0.975, respectively). Error analysis indicated that the proposed estimation equation for natural frequency is highly accurate and reliable. This study provides references for determination of structural natural frequency of large cooling towers and future studies on structural natural vibration characteristics.