针对水电站厂房结构模态识别问题,建立了一种密集模态识别的组合方法。该方法通过将逆衰减指数窗和小波包分频结合,实现了降低振动信号的模态密集度,提高了模态分离的精度。振动信号模态分离后采用Random Decrement Technique(RDT)和Spare Time Domain(STD)方法识别水电站厂房结构的模态参数。以一大型地下水电站厂房结构为分析对象,对其原型振动测试数据开展了密集模态识别。结果表明,采用该组合方法识别结果与三维有限元计算结果一致,且识别出的阻尼比也处在合理范围之内。
This paper presents a combined method for modal identification of hydropower house structures with closely spaced modes, which can reduce mode intensity and improve the precision of modal separation by combining a reverse attenuation index window with wavelet packet frequency division. After modal separation of a signal, a random decrement technique and spare time domain method are used to identify its modal parameters. We have applied this combined method in a case study of the signals of in-situ vibrations in a large-size hydropower house structure with closely spaced modes, and compared with finite element analysis results. The calculations showed that the identification results were in good agreement with the finite element calculations and the damping ratios identified were in the normal range of reinforced concrete structures.