Variable-load piston type engine impact feature extraction method based on EMD and Teager operators

Abstract

The invention relates to a variable-load piston type engine impact feature extraction method based on EMD and Teager operators, and the method comprises the steps: carrying out the processing of an engine cylinder vibration impact signal through employing the EMD and the Teager operators, and obtaining a reconstructed vibration waveform; then, aiming at the variable load working condition of the engine, the signal data change rate is used for extracting an impact judgment threshold value by the signal data change rate under the normal working condition, and solving the problems that impact identification is greatly influenced by loads and deviation exists due to a single fixed threshold value in the past are solved; and finally, by utilizing the characteristic that the vibration signals of the engine are distributed according to a fixed phase, segmenting the vibration signals are segmented, comparing the impact peak value of each segment is compared with an impact judgment threshold, effectively identifing the impact frequency of the whole period is effectively identified, and t. The method can be used for judgment and alarm of impact faults such as engine valves, cylinder collision and connecting rod breakage.

Description

technical field [0001] The invention belongs to the technical field of engine fault monitoring and diagnosis, and relates to a method for extracting shock features of a variable-load piston engine based on EMD and a Teager operator. Background technique [0002] Piston engines are widely used as power equipment in the fields of ships, naval vessels, petroleum and petrochemical and nuclear power, and are of great significance to the monitoring and diagnosis of engines. Compared with rotating machinery, the most essential feature of the engine vibration signal is the impact feature, the system structure is complex, and there are many vibration excitation sources; the signal is time-varying, which belongs to a typical time-varying system. Due to the large number of shocks, the signal as a whole shows strong non-Gaussian characteristics; in the vibration signals collected at different measuring points of the same unit, the vibration shocks are mixed with each other, and the exci...

AI Technical Summary

Problems solved by technology

Due to the large number of shocks, the signal as a whole shows strong non-Gaussian characteristics; in the vibration signals collected at different measuring points of the same unit, the vibration shocks are mixed with each other, and the excitation is complex, so it is difficult to extract and identify the vibration shocks, determine the shock source, and diagnose faults happened

[0003] With the development of engine fault diagnosis technology at home and abroad, the research on vibration and shock identification is relatively lacking; most units currently focus on thermal parameter monitoring, which cannot meet the needs of on-site unit inspection and maintenance

However, the method of vibration signal monitoring and diagnosis still has problems such as variable signals, multiple sources of impact, complex components, and difficulty in extracting fault symptoms.

Images