基于级联结构的航空发动机振动信号盲分离

摘要

级联型的盲抽取紧缩方法具有算法快速、冗余性小、单元结构简单和易于实现等优点,在盲信号处理领域有着广泛的应用.针对该结构在航空发动机振动信号盲分离中的应用进行了研究,首先介绍了这种级联结构的可行性和优越性以及抽取紧缩单元的优化算法,然后对某型航空发动机实测的故障振动信号利用级联盲抽取结构进行振源分离,并通过对分离出的振源信号进行功率谱分析诊断出导致发动机振动过大产生故障的原因.仿真结果验证了该方法的有效性,可以应用于航空发动机振动信号盲分离领域.%The cascade type blind separation method has the advantages such as fast algorithm, little redundancy, simple unit-structure and easy to realize, so it is widely applied in blind signal processing field. The use of this structure in aviation engine vibration signal blind separation was researched. At first, the feasibility and superiority of the cascade extraction structure, and the optimization algorithm of extraction and deflation units were introduced. After, a real aviation engine vibration signal was separated by using the cascade type blind extraction structure, then power spectrum analysis for the separated vibration signal was done, at last the cause of excessive vibration fault was diagnosed. Simulation results demonstrate the effectiveness of the proposed method, and it can be used in aviation engine vibration signal blind separation.