基于DNN的电热综合能源系统快速状态估计

United Fast State Estimation Combined Heat and Power Based on Deep Neural Network

综合能源系统引入大量多源异构的运行数据,并且设备采集的运行数据不可避免地存在误差,因此难以准确获取系统运行状态。提出一种基于DNN的电-热综合能源系统快速状态估计模型。上述方法首先利用蒙特卡洛采样和相关性分析,挖掘子网络和耦合设备的相关性。接着根据相关性分析结果提取状态量特征,即量测量。最后对数据集添加高斯白噪声,搭建DNN对状态量进行快速状态估计。通过算例验证,并与相同条件下的传统方法对比。结果表明,利用文中方法对电-热IES进行联合状态估计,可以充分挖掘子网络参数之间的相关性,和基于模型的传统方法相比,DNN状态估计平均误差更低,实时性更强,鲁棒性更好。

The integrated energy system(IES)introduces multi-source heterogeneous operating data,and the oper-ating data collected by the equipment inevitably has errors,so it is difficult to accurately obtain the operating status of the system.In this paper,a fast state estimation model of electric-heat integrated energy system based on deep neural network(DNN)is proposed.The article uses Monte Carlo sampling and correlation analysis to mine the correlation be-tween the parameters of the sub-network.Combining the results of correlation analysis,a DNN fast state estimation model is established.It is verified by calculation examples and compared with traditional methods under the same conditions.The results show that the joint state estimation of electric-thermal IES using the method in this paper can fully tap the correlation between the sub-network parameters,and the state estimation average error is lower,the real-time performance is stronger,and the robustness is better.