基于递推子空间的机组数字孪生模型预测精度优化方法  

Recursive Subspace-based Model Refinement Method for Digital Twin of Thermal Power Unit

作　　者：赵彦博 蔡远利[1] 胡怀中[1] Zhao Yanbo;Cai Yuanli;Hu Huaizhong(School of Automation Science&Engineering,Xi'an Jiaotong University,Xi'an 710049,China)

机构地区：[1]西安交通大学自动化学院,陕西西安710049

出　　处：《系统仿真学报》2024年第1期195-202,共8页Journal of System Simulation

基　　金：国家重点研发计划(2018YFB1700100)。

摘　　要：由于机理分析的简化假设条件或设备实际运行中参数特性偏移等因素,导致机理建模不可避免存在模型误差。针对该问题,提出一种基于递推子空间的火电机组数字孪生模型预测精度优化方法。分析机组关键设备的运行机制,结合典型工况小样本数据,建立火电机组的全设备多工况非线性动态机理模型,确保数字孪生系统模型具有较好的可解释性与泛化性能;基于递推子空间辨识方法,建立预测精度优化模型并实时进行在线更新,补偿机理模型产生的误差,提高整体模型的预测精度,保证数字孪生模型的高保真性。仿真实验验证了所提方法的有效性。Due to factors such as simplified assumptions or equipment characteristic deviation,modeling errors are inevitable in the mechanism modeling of thermal power units.To deal with the problem,this paper proposes a novel model refinement method based on recursive subspace for the digital twin of thermal power units.Firstly,the digital twin models are built based on mechanism analysis and combined with small sample data of typical conditions,ensuring interpretability and generalization performance.Secondly,based on the recursive subspace identification method,the refinement model is built and updated online in real time to compensate for the modeling error,improving the prediction accuracy and ensuring the high fidelity of the overall digital twin model.Finally,simulation results validate the effectiveness of the proposed method.

关 键 词：数字孪生 火电机组 模型预测精度优化 子空间辨识 数据驱动 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TP11[自动化与计算机技术—计算机科学与技术]