基于改进HHT的泵站管道工作模态辨识

为提高管道工作模态辨识精度,在希尔伯特-黄变换(Hilbert-Huang Transform,HHT)辨识基础上,发展奇异值分解(singular value decomposition,SVD)和经验模态分解(empirical mode decomposition,EMD)联合滤波技术与HHT时频域辨识结合的模态辨识方法。SVD-EMD联合滤除结构振动信号中的强噪声,凸显结构振动特性,有效避免了后期HHT参数辨识过程中虚假模态干扰,提高辨识精度和准确性。将该方法应用于景泰二期工程3泵站2管道的模态参数辨识问题中,建立该管道有限元模型并计算其结构动力特性。对比该文方法辨识结果、随机子空间辨识结果与有限元计算结果,该文方法辨识结果稍小于随机子空间辨识结果,与有限元计算结果更接近,其最大辨识误差为3.6%。研究表明,该方法能准确辨识管道频率,且有效降低管道结构背景强噪声。该研究可为管道安全运行和在线健康监测提供参考。

Modal parameter identification for pipeline of pumping station based on improved Hilbert-Huang transform

Abstract: For large pipeline structure, high-frequency white noise and low-frequency noise are mixed into vibration information, which belongs to one kind of non-stationary and nonlinear signal in low signal-to-noise ratio (SNR). In order to improve the precision of modal parameter identification for pipeline, on the basis of the Hilbert-Huang Transform (HHT) modal parameter identification theory, an improved HHT modal parameter identification method was proposed, which combined the united filtering technique of singular value decomposition (SVD) and empirical mode decomposition (EMD) as pretreatment. The basic of SVD is to process the online data or discrete data with the theory of matrix SVD to obtain the feature information of pipeline structure. The core of EMD is to decompose self-adaptively the signal into a series of intrinsic mode functions (IMFs) from high frequency to low frequency based on its time scale characteristics. Firstly, the pipeline structure vibration signal was processed with SVD, and high-frequency white noise was filtered out. Then the further EMD was conducted on de-noised signal processed by SVD, and through analyzing the spectrum diagram of every IMF component, low-frequency noise was filtered out. So the combined SVD-EMD filtering method was used to process vibration signal to achieve a higher precision de-noised signal. When strong noise was filtered out by the combined SVD-EMD filtering technique, the useful dominant dynamic characteristics of structure were highlighted, which decreased the noise interference to a large extent and avoided the false modal interference effectively during the later HHT processing. Structure system order was determined by singular entropy increment. Finally the de-noised signal was conducted by the improved HHT method, and the structure modal parameter was obtained. Taking the No.2 pipeline of Pumping Station 3 in Jintai River pumping irrigation as the research object, this proposed method was used to identify vibration response data to achieve modal parameter identification. Three-dimensional finite element model (FEM) of No.2 pipeline was constructed according to fluid-solid interaction theory, through which the structure modal parameter was obtained. Stochastic subspace identification (SSI) is one kind of high precision modal parameter identification method, which is used in many fields in recent years. By comparing the frequency results of improved HHT method, SSI and FEM analysis, the results showed that the frequency identified by improved HHT method was slightly smaller than that by SSI, and closed to that by FEM with the maximum error of 3.6%. This improved method can accurately identify the frequency of pipeline, which reduces the strong noise of pipeline and improves the modal parameter identification precision, and thus can be extended to lager pipeline structure. This proposed method provides a new idea for safe operation and online monitoring of the pipeline, and can be used effectively to solve the problem of structure modal parameter identification under ambient excitation, especially under the background with strong noise. This method has a broad prospect in engineering application.