摘要

Impact excitations, time-varying transfer properties and non-stationary random response are typical characteristics of the reciprocating engine vibration. These characteristics of the engine vibration make its dynamic analysis and signature extraction much more difficult than that of the rotational machinery. Focused on this problem, this paper summarizes a thorough investigation into it from theoretical analysis to an applicable diagnosing approach design, as well as experimental verification. Firstly, the impact excitations inside the engine are reasonably analyzed and classified. Based upon this, an analytic model of the non-stationary engine vibration considering its time-varying transfer properties is developed and discussed in details on its time domain and time-frequency domain characteristics. Directed by these discussions, an applicable signature extracting and diagnosing approach is designed. This includes the software-based synchronization for pretreatment of the non-stationary vibration signal, wavelet packet based multi-band filtering techniques for signature extraction of the impact excitations, as well as the auto-regressive model based pseudo-Wigner-Ville distribution for an integrated time-frequency signature extraction of the engine vibration. These methodologies are successfully used in the practical diagnosis of the 6190Z(L)C diesel engine.