Improved empirical mode decomposition method for fault diagnosis of ship electric propulsion system

The invention discloses an improved empirical mode decomposition method for fault diagnosis of ship electric propulsion system, which comprises the following steps: obtaining fault data of ship electric propulsion system; improved empirical mode decomposition (EMD) being used to obtain the intrinsic mode function (IMF) data from the fault data; RBF neural network being used to analyze the intrinsic mode function data of different parts and the fault reason is obtained. The improved Empirical Mode Decomposition (EMD) process is as follows: signal input; determining an initialization parameter Tby using the cosine window definition; the original data are extended by genetic algorithm. Windowing the data to improve the endpoint effect; Empirical Mode Decomposition for Eliminating Modal Aliasing of Data; The intrinsic mode function data are intercepted. An improved empirical mode decomposition (EMD) method for fault diagnosis of ship electric propulsion system is disclosed, which is suitable for non-stationary, n