知搜-全球专利检索

  1. 登录
  2. 注册
发明公开
CN107341581A 一种基于经验小波变换与高斯过程回归的新能源出力短期预测方法 无效 - 驳回 转让

一种基于经验小波变换与高斯过程回归的新能源出力短期预测方法
摘要：
本发明公开一种基于经验小波变换EWT与高斯过程回归GPR的新能源出力短期预测方法，包括以下步骤：1)采用EWT技术将原始风电功率序列分解为一系列特征互异的经验模式；2)对不同经验模式序列进行零均值化处理；3)对每个经验模式建立GPR模型，并求解模型超参数；4)将不同模型预测结果叠加，获得最终短期风电功率预测值。本发明结合EWT与GPR算法的优点，建立了EWT-GPR新能源出力短期预测方法，对风电功率进行预测。首先，采用EWT方法将原始风电功率序列分解为一系列特征互异的经验模式。然后，对每一经验模式建立高斯过程回归预测模型，并求解模型超参数，避免共轭梯度方法的不足。
摘要（英）：

The invention discloses a short-term new energy output prediction method based on empirical wavelet transform (EWT) and Gaussian process regression (GPR). The method comprises: (1), with an EWT technique, decomposing an original wind power sequence into a series of empirical modes with different characteristics; (2), carrying out zero equalization processing on different empirical mode sequences; (3), establishing a GPR model for each empirical mode and calculating a model hyperparameter; and (4), superposing different model prediction results to obtain a final short-term wind power prediction value. According to the invention, on the basis of combination of advantages of the EWT and the GPR algorithm, an EWT-GPR new energy output prediction method is established to predict the wind power. With the EWT method, the original wind power sequence is decomposed into a series of empirical modes with different characteristics; the GPR prediction model is established for each empirical mode; and the model hyperparameter is calculated. Therefore, defects of the conjugate gradient method are overcome.

0/0