关于风力发电系统短期风速预测仿真

Simulation of Short-term Wind Speed Forecasting for Wind Power Generation System

及时并准确地预测风力发电系统短期风速,对于电网调度具有重大意义。由于风速具有随机性、间歇性,且含有大量噪声,因此会对预测精度造成较大误差,为了对短期风速进行精确、稳定地预测,提出了一种基于变分模态分解和高斯过程回归的短期风速预测模型。首先,对原始风速序列进行变分模态分解,获得多个子序列;然后对各子序列分别建立高斯过程回归预测模型,并引入量子粒子群算法代替共轭梯度法,改进协方差函数的超参数寻优过程。最后将各序列的预测值进行叠加得到风速预测结果,并与高斯过程回归、经验模态分解-高斯过程回归模型进行对比。仿真结果表明,变分模态分解和高斯过程回归组合模型能够有效提高预测精度,并为类似工程提供借鉴。

Timely and accurate prediction of wind power system short-term wind speed grid scheduling is of great significance for the power. Because the wind is random and intermittent, and contains a lot of noises, it will lead to a significant error in the prediction accuracy of short-term wind speed. In order to predict the short-term wind speed accurately and steadily, a new model based on variational modal decomposition and Gaussian process regression is proposed. Firstly, the sub-sequence was obtained by the variational modal decomposition of the original wind speed sequence. Then, the Ganssian process regression prediction model was established for each sub-sequence, and the quantum particle swarm optimization algorithm was introduced instead of the conjugate gradient method, to improve the optimization process of super-parameter of variance function. Finally, the predicted values of each sequence were superimposed to obtain the wind speed prediction results, and compared with Gaussian process regression, the empiri- cal modal decomposition-Gaussian process regression model. The simulation results show that the variational modal decomposition and Gaussian process regression combination model can effectively improve the prediction accuracy and provide reference for similar projects.