基于优化本征正交分解法的复杂屋盖风压场预测

Prediction of wind pressure on the complex roof based on optimized orthogonal decomposition technique

为了提高屋盖表面风压数据密度,通过有限测点的风压特征信息准确预测未布置测点的风压信息。风洞试验中,复杂的屋盖结构可以同时引入测点面积和脉动风压自根方差参量,采用优化本征正交分解法(POD法),提高识别精度。利用优化的POD法,在此基础上插值预测节点处的荷载本征模态值,从而进一步获得预测节点处的风压时间序列。对预测节点处实测风压与预测风压的均值、根方差误差进行比较分析,预测的风压均值误差最大达到8%,根方误差最大达到15.8%,同时分别在时域和频域范围内对实测风压与预测风压的时程曲线和功率谱进行比较,可知该方法能较好地对未布置测点的节点风荷载进行有效预测。

In order to improve the data density of wind pressure on the roof surface,the characteristic of wind pressure of the node is predicted accurately by wind pressure under limited measuring points. In the wind tunnel test,area of the measuring point and fluctuation wind pressure variance parameter was introduced into the complex roof structure to optimize orthogonal decomposition technique and improved recognition accuracy. By using of optimized eigen orthogonal decomposition technique,the intrinsic modal values at the predicted nodes are interpolated to obtain the wind pressure time series at the nodes. Based on the analysis of the measured and predicted wind pressure of the mean wind pressure and root mean square error,the prediction of wind pressure mean maximum error reaches 8%,root mean square error maximum reached 15. 8%,respectively. Meanwhile,in the time domain and frequency domain,the curve and the power spectrum comparision of measured and predicted of wind pressure indicates that the method can forecast the wind load node without the layout of measuring points effectively.