基于最大谱的非平稳随机振动数据分析处理方法研究

摘要

ObjectiveTo solve the problem that the environmental measurement of non-stationary random vibration data can’t deal with power spectral density.Methods Two methods of maximum spectral analysis were summed up, namely FFT and digital filter which were based on minimum mean-square error criterion. The advantages, disadvantages and error of two me-thods with different bandwidth were discussed, and then examples were given to testify the better method.Results The theoretic analysis and simulation results showed that: the error was minimum with conventional FFT computations (proportional resolu-tion bandwidth). The conventional FFT computations were applied to 9 sorties of non-stationary vibration data processing and environmental condition determination for a certain missile, and the result was reasonable and effective.Conclusion The con-ventional FFT is the best method to deal with non-stationary random vibration data, which provides guidance for formulating environmental condition based on maximum spectral.%目的：解决非平稳随机振动环境实测数据无法采用功率谱密度分析方法的问题。方法基于最小均方误差准则，归纳总结 FFT 和数字滤波两种最大谱分析方法。通过对比两种方法在不同分辨率带宽时计算的优缺点，并分析其误差。对优选出来的最大谱分析方法进行举例验证。结果理论分析和仿真结果均表明基于比例分辨率带宽的FFT最大谱分析方法的误差最小。将FFT最大谱分析方法应用于某型炸弹9个架次自由飞振动数据的处理和环境条件确定中，结果合理有效。结论可用比例分辨率带宽的 FFT 法分析非平稳随机振动环境实测数据，为基于最大谱制定型号振动环境试验条件提供指导。