分析了工程抗风研究中对台风风观测数据可靠性和代表性判别的必要性以及可能对工程抗风参数的影响，给出了基础数据质量检验、处理的方法，初步提出了应用于工程抗台风研究的基础数据样本有效完整率标准以及台风强风数据代表性的判别依据和判别指标；通过“黑格比”台风观测数据的实验计算和对比分析，发现：1)对超声测风仪获取的高频采样测风数据以4倍截断方差法进行处理，能够较好地提高数据可靠性；2）风速序列中的野点数据对工程抗风特性参数的可靠性影响显著，即使只有2%～5%的野点数据，对风谱参数的影响也很显著；3) 台风强风数据的代表性判别指标应以风工程研究需要的参数在台风不同位置的特征变化为依据；4）台风强风数据代表性判别的基本指标是同时满足8级以上风速的风向连续转换120°以上方位角度、台风过程的风速时程曲线呈M型“双峰”型变化、“双峰”之间出现风速小于11m/s时可以判断为台风眼区。

The modern large-scale engineering of typhoon wind resistance mainly focus on the analysis of near surface wind

characteristics in typhoon boundary layer. Based on the near surface wind observation data of several typhoons in recent years and combined with the characteristic of non-homogeneous cyclone circulation of typhoon and the characteristic of highfrequency sampling anemometer, the necessity of reliability and representative assessment of typhoon wind observation data in the wind resistance study and its potential impact on wind resistance parameters are analyzed and the method of basic data quality test and control are given. Effective integrity rate standard of basic data and the data representative assessment criteria and index of typhoon strong wind are propounded to apply in the study of typhoon wind resistance engineering. By analyzing and comparing the values calculated by the observation data of typhoon Hagupit, the following results are found: 1) Highfrequency wind data observed by ultrasonic anemometer processed by 4 times truncation variance method have a better data representative. 2) The outlier in the wind speed time series has significant impact on the reliability of wind resistance parameters. Even 2%-5% outlier can have a large effect on the wind spectrum. 3) Typhoon strong wind representative assessment index for the observation data should be abided by the variation characteristics of parameters demanded by the study of wind engineering at different positions of typhoon. 4) The basic criteria of typhoon strong wind representative assessment are the wind direction alters over 120° in succession and the wind speed curve versus time shows an “M” shape bimodal distribution. If the lowest wind speed between two peaks is less than 11m/s, it can be judged as the typhoon eye area.