WPA.zip_wpa_信号降噪_小波包 去噪_小波包降噪_降噪处理

%===================================================================================== % Program functions: % For the given structural vibration response data(time sequence signal), % the program can decompose these response signal into multiple orders % wavelet sub-signals, then calculate the energy spectum of the sub-signals. % % Data file preparation: % Simulation data: save as the format "save 3mm_crack RS3mm" including file name % and variable name % Experimental data: (1)sampling data uing "Signal analyzer B&K 3557" are % saved as "CVP" format flie % (2) CVP file is edited using "Excel" and save as "TXT" file % (3) TXT file is called using MATLAB software and save as % " M" file % (4) In matlab program, using special sentence call M data file, % i.e., for example "x0=load('ad800t.m')" %===================================================================================== % to load the data files of structural vibration response clear all clc format long e load trans_nodamg_var2to5.txt load res_dmg10perc.txt %---------------------- time=trans_nodamg_var2to5(:,1); l_time=size(time,1); node_46_nodmg=trans_nodamg_var2to5(:,2); node_46_dmg10perc=res_dmg10perc(:,2); node_64_nodmg=trans_nodamg_var2to5(:,3); node_64_dmg10perc=res_dmg10perc(:,3); node_82_nodmg=trans_nodamg_var2to5(:,4); node_82_dmg10perc=res_dmg10perc(:,4); node_100_nodmg=trans_nodamg_var2to5(:,5); node_100_dmg10perc=res_dmg10perc(:,5); %------------------------- %================================== nn=length(time); %measure data length of structural response (all data files % should have the same data length %nn=1200; % One can select part data from a response signal to make wavelet analysis %Put one group of damaged structual response data and one group of intact %structural response data into two vectors "xx0 and xx1", so that one can compare % differences between them using wavelet signal shape and energy spectrum for i=1:nn xx0(i)=node_100_nodmg(i); xx1(i)=node_100_dmg10perc(i); end %================================== %Using "plot(x,y)" function to draw the waveshapes of the response signals of %intact and damage structures. i=1:nn; plot(i,xx0) title(('Case 0: no damage')) xlabel('time-') ylabel('Response signal of the intact structure (mV)') grid on pause close i=1:nn; plot(i,xx1) title(('Case 1: damage 10%')) xlabel('time-') ylabel('Response signal of the damaged structure ') grid on pause close %============================================= % Decompose two kinds of signals upto the LN layer of wavelet. % First, one can get all wavelet coefficients using "wpdec";then we can % recompose every sub-wavelet signals using these wavelet coefficients and "wprcoef" % When a signal is decompsed to LN layer, one can get % s1=xx0; % response signal of non damage plate s2=xx1; %response signal of damage plate % To draw raw waveshape of the response signals of intact and damaged structures subplot(11,2,1); plot(s1); title('Intact structure') axis off ylabel('raw signal') subplot(11,2,2); plot(s2); title('Damaged structure 10%') axis off %=============================================================== % Decompose intact structural response signal to the LN layer of wavelet using wavelet package % for the LN layer wavelet, we can get the 2**LN sub-wavelet signals % LN=5; SUB_W_LN=2.^LN t=wpdec(s1,LN,'db1','shannon'); %one-dimension wavelet decomposition %--------------------------------- for i=1:SUB_W_LN gsi=wprcoef(t,[LN,i-1]); % wavelet reconstruction for j=1:nn gs(i,j)=gsi(j);% Matrix "gs" containing the SUB_W_LN sub-wavelet signals of % the intact structure. end end %=============================================================== % Decompose damaged structural response signal to the LN layer of wavelet using wavelet package % for the LN layer wavelet, we can get the 2**LN sub-wavelet signals % t=wpdec(s2,LN,'db1','shannon');%wavelet decomposition %---------------------------- for i=1:SUB_W_LN dsi=wprcoef(t,[LN,i-1]);% wavelet reconstruction for j=1:nn ds(i,j)=dsi(j);% Matrix "ds" containing the SUB_W_LN sub-wavelet signals of % the damaged structure. end end %=============================================================== % To draw the first 10 orders sub-wavelet signals of the intact structure for i=1:10 subplot(11,2,2*i+1); plot(gs(i,:)); axis off end %------------------------------------------------------ %To draw the first 10 orders sub-wavelet signals of the damaged structure for i=1:10 subplot(11,2,2*i+2); plot(ds(i,:)); axis off end pause close %=================================================================== % To calculate the energy spectrum of the sub-wavelet signals %-------------------------------------------------------- for i=1:SUB_W_LN for j=1:nn si(j)=gs(i,j); end goodnorm(i)=norm(si); %the energy spectrum of the response signal of the intact structure end %---------------------- for i=1:SUB_W_LN for j=1:nn di(j)=ds(i,j); end badnorm(i)=norm(di);%the energy spectrum of the response signal of the damaged structure end %====================================================== % To calculate the variation percent of every wavelet-order energy spectrum between % the intact and damaged structures for i=1:SUB_W_LN dr(i)=100*(goodnorm(i)-badnorm(i))/goodnorm(i);%the variation percent end %====================================================== % to display the energy spectrum valuse for i=1:SUB_W_LN result(i,1)=i; result(i,2)=goodnorm(i); result(i,3)=badnorm(i); result(i,4)=dr(i); end format short e result; MAXESV=max(abs(dr)); bar(dr,'group'); title(('Structural damage index')) xlabel('wavelet orders') ylabel('variation percent of wavelet energy spectrum (%)') axis([0 SUB_W_LN -1.5*MAXESV 1.5*MAXESV]) grid on pause close