基于对抗式迁移学习的含柔性高压直流输电的风电系统次同步振荡源定位

Location Method of Subsynchronous Oscillation Source in Wind Power System with VSC-HVDC Based on Adversarial Transfer Learning

风电场经柔性高压直流输电(VSC-HVDC)接入交流系统会产生次同步振荡(SSO),定位风电场SSO源并及时采取针对性抑制措施是迫切需要解决的问题。该文通过建立风电场经VSC-HVDC并网电力系统线性化模型,分析风电场因发生交互而诱发SSO的机理,提出基于对抗式迁移学习的风电场SSO源定位方法。该方法通过对仿真系统与实际系统的振荡特征进行对抗学习,缩小了仿真系统与实际系统的域差异,实现仿真系统离线建立的定位模型能够迁移到实际系统中进而对风电场次同步振荡源进行在线定位。通过设计多风电场经VSC-HVDC并网电力系统应用案例,验证分析了所提方法在不同系统中均具有较高的定位精度。这对电网调度运行基于广域测量系统识别振荡源或提供振荡抑制策略具有重要参考价值。

Wind farms connected to AC systems via voltage source converter based high voltage direct current transmission(VSC-HVDC),which will generate subsynchronous oscillations(SSO).It is an urgent problem to locate the source of SSO in wind farms and take targeted suppression measures in time.This paper establishes the linearization model of the power system of wind farms connected to the grid via VSC-HVDC,analyzes the mechanism of wind farms inducing SSO due to interaction,and proposes a wind farm SSO source location method based on adversarial transfer learning.This method reduces the domain difference between the simulation system and the actual system by learning from the oscillation features of the two systems,and realizes that the location model established by the simulation system offline can be transferred to the actual system,thus to locate the SSO source in the wind farms.Through designing a power system case where multiple wind farms are connected to the grid via VSC-HVDC,it is verified and analyzed that the proposed method has high locating accuracy in different systems.This has important reference value for the power grid dispatching operation based on the wide-area measurement system to identify the source of oscillation or to provide an oscillation suppression strategy.