1.华北电力大学控制与计算机工程学院,北京市 102206;2.华北电力大学电气与电子工程学院,北京市 102206
对低频电力数据进行超分辨率重建有助于实现电力系统精准的态势感知与决策分析。现有的超分辨率重建算法存在重建效果不够精确、算法通用性不强等问题。文中提出了一种基于改进扩散模型的电力数据超分辨率重建技术,扩散模型能够捕捉到数据中细微且复杂的特征,可在低频数据引导下逐步生成高频电力数据。通过将长短期神经网络与扩散模型相结合,进一步增强了模型对时序数据的挖掘能力,提高模型超分辨率重建能力。文中使用负荷和谐波两种场景下的电力数据进行算例验证,实验表明所提方法能够精准重建高频数据。同时模型具备良好的泛化性和灵活性,可适用于未经训练的电气参数和建筑的数据,也可以重建不同精度的数据。
国家自然科学基金
Super resolution reconstruction of low-frequency power data is helpful to realize accurate situation awareness and decision analysis of power system. The existing super resolution reconstruction algorithms have some problems, such as not accurate reconstruction effect and not strong generality. In this paper, a super resolution power data reconstruction technique based on an improved diffusion model is proposed. The diffusion model can capture the subtle and complex features of the data, and can gradually generate high-frequency power data under the guidance of low-frequency data. By combining the long and short term neural network with the diffusion model, the model"s ability of mining time series data is further enhanced, and the model"s ability of super resolution reconstruction is improved. In this paper, the power data under two scenarios of load and harmonics are used for example verification. The experiments show that the proposed method can accurately reconstruct the high-frequency data. At the same time, the model has good generalization and flexibility, which can be applied to untrained electrical parameters and building data, and can also reconstruct data with different precision.
[1] | 薛彤丹,王红,齐林海,等.基于改进扩散模型的电力数据超分辨率重建[J/OL].电力系统自动化,http://doi. org/10.7500/AEPS20231017002. XUE Tongdan, WANG Hong, QI Linhai, et al. Power Data Super Resolution Technology Based on Improving the Diffusion Model[J/OL]. Automation of Electric Power Systems, http://doi. org/10.7500/AEPS20231017002. |