摘要
根据非平稳过程的进化谱理论,导出基于TARMA模型的非平稳脉动风速模拟式。基于模拟解析式,得到一些空间点非平稳下击暴流风速的模拟时间序列;运用经验模式分解(EMD)和基于粒子群优化(PSO)的最小二乘支持向量机(LSSVM)(简称为PSO-LSSVM)算法,经MATLAB平台编制程序,根据上下空间点风速样本预测出中间高度处的非平稳下击暴流风速时程。通过功率谱、自相关和互相关函数预测值与模拟值的比较及平均误差(AE)、均方根误差(MSE)和相关系数(R)的评价,验证了基于时变ARMA模型和EMD-PSO-LSSVM算法的下击暴流风速模拟与预测的可行性。
Using the theory of evolutionary power spectral density for nonstationary processes,the formula of the time-varying auto regressive moving average（ TARMA） model was derived to simulate nonstationary downburst wind velocity. The simulation of nonstationary downburst wind velocity time histories at some space points was conducted using TARMA. Adopting the empirical mode decomposition（ EMD） and the particle swarm optimization（ PSO） based least squares support vector machines（ LSSVM） making a programming with MATLAB, the prediction of nonstationary downburst wind velocity time histories at middle space points was accomplished using the nonstationary downburst wind velocity samples at upper and lower two space points. Through comparing the simulated values and the predicted ones of the power spectrum,autocorrelation and cross-correlation functions,respectively,it was shown that the TARMA and EMD-PSO-LSSVM algorithm-based simulation and prediction are feasible for nonstationary downburst wind velocity.
出处
《振动与冲击》
EI
CSCD
北大核心
2016年第17期33-38,51,共7页
Journal of Vibration and Shock
基金
国家自然科学基金(51378304)
关键词
下击暴流
预测
时变ARMA
经验模式分解
最小二乘支持向量机
downbursts
prediction
time-varying ARMA
empirical mode decomposition（EMD）
least squares support vector machines（LSSVM）
