论文部分内容阅读
为实现水轮机转轮状态及时有效地识别,运用基于模型的状态识别方法建立水轮机转轮模型,从而提取转轮不同状态下的特征,达到状态识别的目的。鉴于水轮机转轮模型的输入量非常复杂且不易测得的特点,假定水轮机转轮受到的不均匀力作为一个随机输入,利用水导轴承摆度作为输出,提出了基于Volterra模型盲辨识的水轮机转轮状态识别方法。利用在线监测系统中已有的水导轴承摆度数据作为输出,辨识出水轮机转轮Volterra时域模型。结合多维傅里叶变化,建立转轮的广义频率响应模型(generalized frequency response function,GFRF)。通过模型频域上的变化,分析转轮状态的变化。仿真实验和实例研究表明,Volterra模型盲辨识方法能够很好地反映原系统的频域特性,实现了水电机组转轮频域特征的良好提取,进而可采用该方法对水轮机转轮进行状态识别。
In order to realize the timely and effective identification of turbine runner status, a model-based state identification method is used to establish a turbine runner model so as to extract features under different states of the runner so as to achieve state recognition. Considering that the input of the turbine runner model is very complicated and difficult to measure, assuming that the unevenness of the turbine runner is a random input and the turbine bearing swing is used as the output, the turbine rotation based on Volterra model blind identification is proposed Wheel status identification method. Using the existing data of the slewing bearing slew in the on-line monitoring system as the output, the Volterra time-domain model of the turbine runner is identified. Combining the changes of multidimensional Fourier transform, a generalized frequency response function (GFRF) was established. Through the changes in the frequency domain of the model, the change of wheel state is analyzed. The simulation experiments and case studies show that the Volterra model blind identification method can well reflect the frequency domain characteristics of the original system and realize the good extraction of the frequency domain features of the turbine runner. The method can be used to identify the turbine runner status.