Advanced signal processing tools for the vibratory surveillance of assembly faults in diesel engine cold tests

This paper addresses the use of several signal processing tools for monitoring and diagnosis of assembly faults in diesel engines through the cold test technology. One specific fault is considered here as an example: connecting rod with incorrectly tightened screws. Firstly, the experimental apparatus concerning the vibration tests is introduced. Subsequently, the dynamic analysis of the engine has been carried out in order to calculate the connecting rod forces against the crankpin for predicting the position where mechanical impacts are expected. Then, a vibration signal model for this type of faults is introduced. It deals with the cyclostationary model in which the signal is subdivided into two main parts: deterministic and non-deterministic. Finally, the acceleration signals acquired from the engine block during a cold test cycle at the end of the assembly line are analyzed. For quality control purposes in order to obtain reliable thresholds for the pass/fail decision, a method based on the image correlation of Symmetrized Dot Patterns (SDP) is proposed. This method visualizes vibration signals in a diagrammatic representation in order to quickly detect the faulty engines in cold tests. Moreover the fault identification is discussed on the basis of the cyclostationary model of the signals. The first-order cyclostationarity is exploited by the analysis of the Time Synchronous Average (TSA). In addition the residual signal is evaluated by subtracting the TSA from the raw synchronized signal, and thus the second-order cyclostationarity analysis is developed by means of the Wigner-Ville Distribution (WVD), Wigner-Ville Spectrum (WVS) and Mean Instantaneous Power (MIP). Moreover, Continuous Wavelet Transform (CWT) is presented and compared with the WVD and WVS.