Abstract: To better combine the knowledge of physics-informed model and data-driven model,and transfer the test model to the real unit-model in the "Smart Turbine",the numerical reliable simulation data was further realized,and the real operation data were mutually verified and the mixed modeling was driven.By taking 600MW large turbo-generator as the research object,the body model of turbine-generator stator winding with integrated evolution factors were established,the research framework of integrated digital-twin smart winding was proposed,and the formal definition of smart winding was put forward.The mathematical and physics-informed model of electromagnetic force for the stator-end winding was established based on the mirror method,the air-gap rotating current applied superposition,the Bio-Savart law and the Ampere law.Finite element electromagnetic force and digital electromagnetic force were compared to verify the rationality of the established digital physics-informed model of electromagnetic force.The vibration response was analyzed to verify the digital physics-informed modeling of electromagnetic force and electromagnetic force in finite element.In digital twin,the research would support the fault diagnosis and prediction of the vibration dynamics of "Smart Turbine" from the maintenance based on the mixed modeling of physical mechanism and data-driven model.

Key words: stator end winding, electromagnetic force, physics-informed modelling, digital twin, smart turbine, vibration response

摘要： 为了在“智慧汽轮机”中更好地结合物理信息模型先验知识与数据驱动模型,并向真实机组模型迁移试验模型,进一步实现可靠的实验室数值仿真数据和真实运行数据互相印证与混和建模驱动,以600 MW大型汽轮发电机为研究对象,建立融合演进因素的汽轮发电机定子绕组本体模型,提出集成数字孪生智能绕组研究框架,并提出智能绕组内涵的形式化定义。基于镜像法、气隙回转电流应用叠加法、毕奥-萨伐尔定律和安培力定律建立定子端部绕组电磁力的数学机理物理信息模型。对有限元电磁力和数字化电磁力进行了分析对比,验证了所建立电磁力数字化物理信息模型的合理性。对比分析振动响应结果,进一步验证了数字化电磁力物理信息建模的正确性以及有限元电磁力对比验证的正确性。研究将对“智慧汽轮机”运维的振动动态基于物理机理与数据驱动混和建模的数字孪生故障诊断与预测提供有力的支撑。

关键词: 定子端部绕组, 电磁力, 物理信息建模, 数字孪生, 智慧汽轮机, 振动响应