基于数据动态降阶的发电设备物理场级数字孪生模型构建方法

Construction Method of Physical Field Digital Twin Model of Power Generation Equipment Based on Dynamic Degradation of Data

摘要

本文提出一种基于数据动态降阶的物理场级数字孪生模型构建方法。首先阐述数字孪生范式下模型降阶的核心价值，分析传统仿真技术在跨领域、跨生命周期复用中的瓶颈；其次，从三维仿真与系统仿真的跨平台交互、设计阶段与运行阶段的模型迁移两个维度，揭示模型降阶技术在打破仿真壁垒、实现知识复用中的关键作用；进而详细介绍模型降阶的技术原理，包括离线-在线两阶段计算架构、主流降维方法及非线性场景的研究现状；最后总结模型降阶在交互性、可靠性、连续性、可访问性与可分布性方面的核心优势。

Abstract

This paper presents a data-driven dynamic dimensionality reduction method for constructing physical field-level digital twin models. First, it elucidates the core value of model dimensionality reduction under the digital twin paradigm and analyzes the limitations of traditional simulation technologies in cross-domain and lifecycle reuse. Second, it highlights the pivotal role of model dimensionality reduction in breaking down simulation barriers and enabling knowledge reuse, focusing on two dimensions: cross-platform interaction between 3D simulation and system simulation, and model migration between design and operational phases. The paper then details the technical principles of model dimensionality reduction, including its two-phase (offline-online) computational architecture, mainstream dimensionality reduction methods, and research progress in nonlinear scenarios. Finally, it summarizes the core advantages of model dimensionality reduction in enhancing interactivity, reliability, continuity, accessibility, and distributability.

关键词

数字孪生 / 模型降阶 / 跨生命周期复用 / 发电设备 / 实时仿真

Key words

digital twin / model reduction / lifecycle reuse / power generation equipment / real-time simulation

参考文献

[1] 王青山, 严波, 陈岩, 等. 基于降阶模型和数据驱动的动态结构数字孪生方法[J]. 应用数学和力学, 2023, 44(7):757-768.

[2] 仝小冬, 鲍劲松, 陶飞. 协同进化的数字孪生——数字工程的多维动态方法[J]. Advanced Engineering Informatics, 2024,

[3] 佚名. 大型发电机端部绕组电磁力物理数字孪生建模[J]. 计算机集成制造系统, 2023, https://doi.org/10.13196/j.cims.2023.0633.

[4] 李琦, 张霖, 陶飞. 数字孪生模型的复用与演化机制研究[J]. 计算机集成制造系统, 2022, 28(5):1421-1433.

[5] 赵星海, 刘畅, 王鹏. 发电设备多物理场耦合仿真的模型降阶技术应用[J]. 中国电机工程学报, 2021, 41(18):6198-6208.