利用扰动激励的机组轴系扭振响应辨识次同步振荡模态及安全预警的研究

Along with the wide use of series-compensated capacitor transmission and AC/DC hybrid transmission in point to network based long distance power transmition of large-scale energy bases in China, generator-grid subsynchronous oscillation(SSO) and generator unit torsional vibration have become a key technical problem threatening the safety of the generator unit and the regional power grid. Currently, great progress has been made in the research on prevention against torsional vibration of the generator unit and the suppression of SSO of the generator-grid after the SSO occured, but the research into real-time SSO modal monitoring and safety early warning in normal operation is still very weak. If a reliable measurement of SSO modes and damping under the current mode of the generator set and grid operation can be realized, the preventive monitoring and security early warning of the risk of SSO for the power plant and its sending out system can also be realized and thus it provides a new defensive line against SSO and torsional vibration of the generator set. This project starts with the analysis of the electromechanical dynamic process of the power plant and the regional power grid which have the risk of SSO and uses the shaft of the generator unit as the sensor. Through real-time measurement of the rotor rotation vibration response excited by the grid fault disturbance, manual operation disturbance and ambient noise, combined with the probability modal analysis of linear systems under the disturbances, the online identification of SSO and torsional vibration modal parameters(modal frequency and damping) excited by the running condition of the generator unit and grid are studied. And the further research is carried out into the wide-area security early warning against SSO based on the modal damping and the operation mode.

随着我国大型能源基地点对网远距离经串补输电、交直流混合输电技术的广泛应用，机网次同步振荡及机组扭振成为威胁发电机组及区域电网安全的关键技术难题。目前在次同步振荡发生后的机组扭振保护及机网次同步振荡抑制技术的研究有很大进展，但在运行中的次同步振荡模态实时监测和安全预警方面的研究还很薄弱。如果在机网运行中对当前运行方式下的次同步振荡模态及阻尼实现可靠测量，即可实现对发电厂及其送出系统的次同步振荡风险进行预防性监控及安全预警，从而为次同步振荡及机组扭振提供一道新的防线。课题从分析具有次同步振荡风险的发电厂及区域电网的机电动态过程入手，以机组轴系做为传感器，通过实时测量转子受电网故障扰动、操作扰动及环境噪声激励下的旋转振动响应，结合线性系统受扰下的概率模态分析，研究机网运行环境激励下的次同步振荡及扭振模态参数（模态频率及阻尼）的在线辨识，进而研究基于模态阻尼及运行方式的广域次同步振荡安全预警方法。

当前机网次同步振荡（SSO）及机组轴系扭振成为威胁发电机组及区域电网安全的关键技术难题，在运行中的SSO模态实时监测和安全预警方面的研究还很薄弱。课题从分析具有SSO风险的发电厂及区域电网的机电动态过程入手，通过测量转子的旋转振动响应，研究机网运行环境激励下的SSO及扭振模态参数的在线辨识，进而研究基于模态阻尼和多参数的SSO安全预警方法。.主要研究内容和成果如下： .1. 研究了机网互作用机理及模型。依据发电机组和区域电网实际参数，建立机组轴系扭振的有限元模型和集中质量计算模型，以及机网SSO模型。计算出轴系扭振的固有模态和运行模态，获得扭振传感和辨识所需的频率和振型特征。用Holf分岔等理论分析了机网模型稳定性及敏感参数稳定域；仿真分析了多种机械和电气扰动激励的SSO动态过程，发现当存在持续的在轴系固有频率上的扰动，且阻尼不足时，将引发发散的SSO；发现转子转速是SSO的最佳观测点。.2. 研究高可靠扭振传感技术。通过实测数据分析得出满足工程应用的扭振测量精度需求；研究转速脉冲定时测振法的误差，提出多传感器矢量传感等提高灵敏度和准确度的改进措施，研究并实现一种高分辨率扭振信号模拟方法，研制出一套新型扭振仿真测试仪器；提出一种保偏光子晶体光纤扭振测量新方法，研究了其扭转应力敏感特性，提出并验证了扭振测量方案。.3. 深入研究环境激励下微弱SSO的模态阻尼辨识算法。提出“模态预分离+自由响应分量提取+模态参数辨识算法”的方案。详细研究了多种算法的缺陷，提出满足时频特性的改进方案；利用随机减量技术实现自由响应分量提取，研究了关键参数的选取原则，实现微弱SSO模态阻尼有效辨识，可用于低频振荡和网侧次/超同步间谐波检测，能连续获得电力系统的动态特征。 .4. 研究了SSO风险评估策略。提出SSO的安全预警判据；利用神经网络实现对SSO的预测；研究了基于录波大数据异常检测的次同步振荡风险评估算法；提出在线监测保护系统实现方案和系统架构。.项目实现在机网运行中对SSO模态及阻尼的连续测量算法，对SSO风险评估及安全预警提出有效方法和措施，为SSO及机组扭振安全提供一道新的防线。