Machinery Fault Diagnosis Using Signal Analysis

Signal analysis methods are important for machinery fault diagnosis. Signals provided by faulty manufacturing systems are usually arbitrary and nonlinear. These detected changes carry information about system functioning. Compared to conventional signal analysis methods, adaptive mode decomposition based procedures have good versatility and high adaptability in nonlinear signal analysis. This paper presents an instantaneous frequency estimation based on energy separation approach which has effectiveness in estimating both instantaneous frequency and instantaneous amplitude of arbitrary time-varying signals. The proposed algorithm is tested on synthetic signals created for this purpose. This paper presents the principles of this fault diagnosis method, results indicate that the proposed procedure can handle fault detection issues even in noisy environment.