Slamming Signal Detection Algorithm Based On HHT And SVM

As our country continues to promote the navigation strategy and the rapid development of the hull structural monitoring system,it is of great importance to have an effective analysis and processing for the data obtained from the monitoring system.When the ship sailing in the rough sea condition,due to the relative motion the hull will up the water and then immerse into the water quickly which will lead to the dramatic collision between the hull and the water surface,known as slamming.It has huge impact on the hull,on the one hand huge slamming forces have a great impact on the local structure of the hull,on the other hand,it will cause severe vibration throughout the hull,resulting in large vibration moment,at last the longitudinal strength will loss.Therefore,the study of detecting slamming signal in the hull structural monitoring system is very important,at the same time it is also one of the main content to build a sound structural monitoring system.At the beginning of the paper,it is a comparative study on the research status of slamming detection and recognition,and a detailed study of several typical time-frequency analysis methods to analyze their characteristics,applications and limitations,on the basis of this,a method based on Hilbert Huang Transform(HHT)and Support Vector Machine(SVM)classifier to detect slamming signal,this can reflect the strengths and advantages of the effectiveness of HHT in nonlinear and non-stationary signals analysis.Then by means of the improved Empirical Mode Decomposition(EMD)method Ensemble Empirical Mode Decomposition(EEMD)to decompose the data,according to the characteristics of the data to extract feature,we select representative feature variable to form feature vector,when we get enough samples and then randomly select samples to generate training samples and test samples to train the SVM,following the SVM will be used to identify slamming signal.At last,the above analysis process will be simulated in Matlab environment,by the results of analyzing and simulating validate the proposed method is rapid,simple and accuracy.It can provide reliable reference for the hull structural evaluation system.In the paper,the following tasks have been finished:1.Making a comparison of slamming signal research in the hull structure monitor,and analyzing several typical traditional time-frequency methods in signal processing applications,through matlab simulation to analyze the advantages and disadvantages of each method,finally to make a brief summary.2.A detailed analysis of HHT method,and with the HHT method in the detection of slamming signal,using EEMD to decompose signal for feature extraction,in order to validate the effectiveness of the selected method,we use EMD and EEMD to decompose the simulation signal respectively.3.Proposed a method based on HHT and SVM classifier to extract feature and classification,the implementation process will be described in detail.In the matlab environment the proposed method will be used for slamming signal detection,and the detection also use a sliding window technique to detect the real-time data,at last compare the threshold detection method with the method proposed in the paper.