电力系统运行与规划中需重点关注到楼宇空调负荷的不确定性,可将楼宇空调负荷变化的不确性场景转化为多个确定性场景的场景生成问题。提出了楼宇空调负荷场景生成问题的基本分析框架,深入分析了楼宇空调负荷的用能特征,挖掘了楼宇空调...电力系统运行与规划中需重点关注到楼宇空调负荷的不确定性,可将楼宇空调负荷变化的不确性场景转化为多个确定性场景的场景生成问题。提出了楼宇空调负荷场景生成问题的基本分析框架,深入分析了楼宇空调负荷的用能特征,挖掘了楼宇空调负荷用能时序序列数据所蕴含的动静态特征。将楼宇空调负荷数据的动静态特征作为条件监督项,将无监督对抗训练与监督训练相结合,设计了联合训练损失函数与全局优化损失函数,并在此基础上提出了一种基于条件时序生成对抗网络(time series generative adversarial nets,TimeGAN)的楼宇空调负荷场景生成方法。最后,通过算例验证了所提方法的可行性与有效性。研究成果对提高楼宇空调负荷主动参与电力系统的运行规划有积极的意义。展开更多
This study proposes a hybrid network model based on data enhancement to address the problem of low accuracy in photovoltaic(PV)power prediction that arises due to insuffi cient data samples for new PV plants.First,a t...This study proposes a hybrid network model based on data enhancement to address the problem of low accuracy in photovoltaic(PV)power prediction that arises due to insuffi cient data samples for new PV plants.First,a time-series gener ative adversarial network(TimeGAN)is used to learn the distri bution law of the original PV data samples and the temporal correlations between their features,and these are then used to generate new samples to enhance the training set.Subsequently,a hybrid network model that fuses bi-directional long-short term memory(BiLSTM)network with attention mechanism(AM)in the framework of deep&cross network(DCN)is con structed to effectively extract deep information from the origi nal features while enhancing the impact of important informa tion on the prediction results.Finally,the hyperparameters in the hybrid network model are optimized using the whale optimi zation algorithm(WOA),which prevents the network model from falling into a local optimum and gives the best prediction results.The simulation results show that after data enhance ment by TimeGAN,the hybrid prediction model proposed in this paper can effectively improve the accuracy of short-term PV power prediction and has wide applicability.展开更多
Load modeling is one of the crucial tasks for improving smart grids’ energy efficiency. Among manyalternatives, machine learning-based load models have become popular in applications and have shownoutstanding perform...Load modeling is one of the crucial tasks for improving smart grids’ energy efficiency. Among manyalternatives, machine learning-based load models have become popular in applications and have shownoutstanding performance in recent years. The performance of these models highly relies on data quality andquantity available for training. However, gathering a sufficient amount of high-quality data is time-consumingand extremely expensive. In the last decade, Generative Adversarial Networks (GANs) have demonstrated theirpotential to solve the data shortage problem by generating synthetic data by learning from recorded/empiricaldata. Educated synthetic datasets can reduce prediction error of electricity consumption when combined withempirical data. Further, they can be used to enhance risk management calculations. Therefore, we proposeRCGAN, TimeGAN, CWGAN, and RCWGAN which take individual electricity consumption data as input toprovide synthetic data in this study. Our work focuses on one dimensional times series, and numericalexperiments on an empirical dataset show that GANs are indeed able to generate synthetic data with realisticappearance.展开更多
文摘电力系统运行与规划中需重点关注到楼宇空调负荷的不确定性,可将楼宇空调负荷变化的不确性场景转化为多个确定性场景的场景生成问题。提出了楼宇空调负荷场景生成问题的基本分析框架,深入分析了楼宇空调负荷的用能特征,挖掘了楼宇空调负荷用能时序序列数据所蕴含的动静态特征。将楼宇空调负荷数据的动静态特征作为条件监督项,将无监督对抗训练与监督训练相结合,设计了联合训练损失函数与全局优化损失函数,并在此基础上提出了一种基于条件时序生成对抗网络(time series generative adversarial nets,TimeGAN)的楼宇空调负荷场景生成方法。最后,通过算例验证了所提方法的可行性与有效性。研究成果对提高楼宇空调负荷主动参与电力系统的运行规划有积极的意义。
基金supported by the Regional Innovation and Development Joint Fund of National Natural Science Foundation of China(No.U19A20106)the Science and Technology Major Projects of Anhui Province(No.202203f07020003)the Science and Technology Project of State Grid Corporation of China(No.52120522000F).
文摘This study proposes a hybrid network model based on data enhancement to address the problem of low accuracy in photovoltaic(PV)power prediction that arises due to insuffi cient data samples for new PV plants.First,a time-series gener ative adversarial network(TimeGAN)is used to learn the distri bution law of the original PV data samples and the temporal correlations between their features,and these are then used to generate new samples to enhance the training set.Subsequently,a hybrid network model that fuses bi-directional long-short term memory(BiLSTM)network with attention mechanism(AM)in the framework of deep&cross network(DCN)is con structed to effectively extract deep information from the origi nal features while enhancing the impact of important informa tion on the prediction results.Finally,the hyperparameters in the hybrid network model are optimized using the whale optimi zation algorithm(WOA),which prevents the network model from falling into a local optimum and gives the best prediction results.The simulation results show that after data enhance ment by TimeGAN,the hybrid prediction model proposed in this paper can effectively improve the accuracy of short-term PV power prediction and has wide applicability.
文摘Load modeling is one of the crucial tasks for improving smart grids’ energy efficiency. Among manyalternatives, machine learning-based load models have become popular in applications and have shownoutstanding performance in recent years. The performance of these models highly relies on data quality andquantity available for training. However, gathering a sufficient amount of high-quality data is time-consumingand extremely expensive. In the last decade, Generative Adversarial Networks (GANs) have demonstrated theirpotential to solve the data shortage problem by generating synthetic data by learning from recorded/empiricaldata. Educated synthetic datasets can reduce prediction error of electricity consumption when combined withempirical data. Further, they can be used to enhance risk management calculations. Therefore, we proposeRCGAN, TimeGAN, CWGAN, and RCWGAN which take individual electricity consumption data as input toprovide synthetic data in this study. Our work focuses on one dimensional times series, and numericalexperiments on an empirical dataset show that GANs are indeed able to generate synthetic data with realisticappearance.
