Fault Diagnosis of Automaton Based on EMD and Close Degree

Targeting the non-stationary characteristics of short and transient impulse vibration signals of high speed automaton, a new method based on empirical mode decomposition (EMD) and close degree to diagnose fault for high speed automaton is proposed in this paper. Firstly, original acceleration vibration signals are decomposed into a finite number of stationary intrinsic mode functions (IMFs), and then the EMD energy of each IMF is calculated. The correlation analysis is applied and the results show that the first three IMFs contained the most dominant fault information; therefore, the energy feature extracted from the first three IMFs that could serve as a feature vector for fault patterns recognition of automaton. Finally close degree evaluation method was used to diagnose the automaton fault. The experimental results indicate that the proposed approach put forward in this paper can effectively identify automaton fault patterns and it has a great application potential in condition monitoring and fault diagnosis of automaton.