【摘要】 为提高天气突变时太阳能供电节点的能量预测准确率，提出基于改进Pro-Energy的太阳能短时动态预测方法。兼顾能量绝对误差和能量变化趋势，选取最相似天；引入缩放系数对最相似天预测时隙能量值进行修正，用能量修正值替代预测模型中最相似天预测时隙的能量值，建立能量预测模型；设计反映天气变化状况的动态权重因子来及时调整预测模型中各组成部分的比例；在每天预测结束后，根据相似度和时间新鲜度对历史能量模型进行更新；搭建实验平台收集太阳能能量数据，使用实验数据和橡树岭国家实验室能量数据进行预测验证。预测结果表明，所提方法的预测准确率在各种天气下均较对比方法提高7.16%以上。更多还原

【Abstract】 In order to improve the prediction accuracy of sensor nodes powered by solar energy when the weather changes suddenly, a dynamic solar energy prediction method for the short-term based on improved pro-energy is proposed. Firstly, according to the minimum weighted sum of the difference of standard deviation and the mean absolute error between current day’s energy and the stored profiles in the last k time slots, the most similar day is selected. Secondly, a scaling factor is introduced to modify the energy of the most similar day in the prediction time slot, and the corrected energy is used to replace the energy of the most similar day in the prediction time slot, and the prediction model is established. Finally, a dynamic weight factor is designed to adjust the proportion of each component in the prediction model when the weather changes. The stored profiles are updated after the end of the daily prediction. The experimental platform is built to harvest solar energy, and the performance of the prediction method is verified by the experimental data and the energy data of Oak Ridge National Laboratory. The prediction results show that the average prediction accuracy of the proposed method is at least 7.16% higher than all of the compared algorithms under various weather conditions.更多还原