Analysis Of Parametric Transmission Path Based On Combined Noise Reduction Of EEMD And Wavelet Packet

Under the working conditions of the car,the parameterized transmission path method can identify the influence of the vibration generated by the engine and other vibration sources on a certain point in the car,but due to some external disturbances and noise signals during the actual vehicle test,the measured working condition data It is inevitable that noise will be mixed,which will affect the construction of the parameterized model and hide the specific problem path.Therefore,the combined noise reduction method of EEMD and wavelet packet is used to reduce the noise of large and complex working condition data,improve the prediction accuracy of the parameterized transfer path model,and identify hidden paths.The sample entropy is introduced as the quantization criterion of the noise signal component in the signal,and the sample entropy of the noise signal that has been eliminated is used as the information cost function,and the particle swarm optimization algorithm is used to find the optimal threshold to solve the effect of the effective signal component on the noise variance estimation.The problem of threshold estimation accuracy.Combining the respective normalized sample entropy of the noisy IMF components and their decomposed normalized sample entropy,the tuning parameters are designed,and the tuning parameters are applied to the threshold function proposed in this paper to achieve the combined noise reduction of EEMD and wavelet packet Adaptive improvement.Simultaneously,a simulation signal is built,and the existing wavelet packet threshold estimation methods and threshold functions are listed.This article and the existing threshold estimation method are combined with the threshold function to improve the EEMD and wavelet packet combined noise reduction algorithm and reduce the noise of the simulated signal.Among them,the pure signal and the reconstructed signal after noise reduction in the method in this paper have SNR and The RMSE is 23.669 and 0.2566 respectively,the noise reduction effect is the best.Using the existing threshold function combined with the threshold estimation method of this article,using the existing threshold estimation method combined with the threshold function of this article to improve the OPAX model of EEMD and wavelet packet noise reduction,the image results show the improved 2nd order synthetic response and 2nd order actual measurement The response consistency becomes better.Among them,the SNR and RMSE of the target point synthetic response and the measured response obtained by the combination of the threshold estimation method and the threshold function in this paper are 17.8731 and 0.0097,which proves that they have the best consistency,the effect of noise is basically eliminated,and the improvement effect is the best.And using the improved method of this article to analyze the contribution amount,it is found that the problem path hidden by the noise is displayed,which can accurately guide the vibration optimization.