基于失真数据降噪的数据预处理方法及其在风电功率预测中的应用

提高风电数据精度对于建设泛在电力物联网具有重要意义。风电功率预测对历史风电数据集的要求较高，现研究多集中于通过建立不同预测模型或提出不同预测算法以提高风电功率预测准确性，对于风电功率历史数据集本身的噪声数据的处理关注并不多。为此，提出一种针对风电历史数据降噪的方法，该方法主要作用于数据集本身，通过清除历史风电数据中的失真数据，降低历史数据中无用数据的数量，在提高风电功率预测准确性的同时，尽可能缩短数据建模、预测的时间。

Data Pre-processing Method Based on Distorted Data Noise Reduction and Its Application in Wind Power Prediction

Improving the accuracy of wind power data is of great significance for building ubiquitous power internet of things (UPIoT). Wind power prediction has a high demand for historical data sets. Most of research was focused on improving the prediction accuracy by establishing different prediction models or proposing different prediction algorithms. There is not much attention on noise data elimination. Thus, a noise reduction method for the historical wind power data was proposed, which was mainly applied to the data set, by eliminating the distorted data in the historical wind power data, the amount of useless data was reduced, the accuracy of wind power prediction was improved, and the data modeling and prediction time was shortened.