إزالة الضوضاء عالية التردد من إشارات أجهزة استشعار الاهتزاز أثناء تشخيص اهتزاز المحامل

import scaleogram as scg import pandas as pd from pylab import * import pywt filename_n = ['Test_set/Bearing1_3/acc_00001.csv', 'Test_set/Bearing2_3/acc_00001.csv', 'Test_set/Bearing3_3/acc_00001.csv'] for filename in filename_n: df = pd.read_csv(filename, header=None) signal = df[4].values wavelet = 'cmor1-0.5' ax = scg.cws(signal, scales=arange(1, 40), wavelet=wavelet, figsize=(8, 4), cmap="jet", cbar=None, ylabel=' ', xlabel=" ", yscale="log", title='-  %s \n( )'%filename) show() 

 import pywt from scipy import * import pandas as pd from pylab import * filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) signal = df[4].values (cA1, cD1) = pywt .dwt (signal, 'db2', 'smooth') r_signal = pywt.idwt (cA1, cD1, 'db2', 'smooth') fig, ax =subplots(figsize=(8,4)) ax.plot(signal, 'b',label=' ') ax.plot(r_signal, 'r', label='  ', linestyle='--') ax.legend(loc='upper left') ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) ax.set_title('   ( pywt.dwt()) \n   ( pywt.idwt()) ') show() 

 import pywt import pandas as pd from pylab import * filename = 'Test_set/Bearing3_3/acc_00026.csv' df = pd.read_csv(filename, header=None) signal = df[4].values coeffs = pywt.wavedec(signal, 'db2', level=8) r_signal = pywt.waverec(coeffs, 'db2') fig, ax = plt.subplots(figsize=(8,4)) ax.plot(signal, 'b',label=' ') ax.plot(r_signal, 'r ',label= ' ', linestyle='--') ax.legend(loc='upper left') ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) ax.set_title('    - level.\n   ( pywt.wavedec()) ') show() 

 from scipy import* from pylab import* import pywt s = linspace(-4, 4, 1000) s_soft = pywt.threshold(s, value=0.5, mode='soft') s_hard = pywt.threshold(s, value=0.5, mode='hard') s_garrote = pywt.threshold(s, value=0.5, mode='garrote') figsize=(10, 4) plot(s, s_soft) plot(s, s_hard) plot(s, s_garrote) legend(['soft (0.5)', 'hard (0.5)', 'non-neg. garrote (0.5)']) xlabel('   ') ylabel('  ') show() 

 import pandas as pd from pylab import * import pywt filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) signal = df[4].values v='bior4.4' thres=['soft' ,'garrote'] for w in thres: def lowpassfilter(signal, thresh, wavelet=v): thresh = thresh*nanmax(signal) coeff = pywt.wavedec(signal, wavelet, level=8,mode="per" ) coeff[1:] = (pywt.threshold(i, value=thresh, mode=w ) for i in coeff[1:]) reconstructed_signal = pywt.waverec(coeff, wavelet, mode="per" ) return reconstructed_signal fig, ax = subplots(figsize=(8,4)) ax.plot(signal, color="b", alpha=0.5, label=' ') rec = lowpassfilter(signal, 0.4) ax.plot(rec, 'r', label='DWT  ', linewidth=2) ax.legend() ax.set_title('      :%s\n   :%s'%(v,w), fontsize=12) ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) show() 

 import pandas as pd from pylab import * import pywt filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) signal = df[4].values v='bior4.4' thres=[0.1,0.4,0.6] for w in thres: def lowpassfilter(signal, thresh, wavelet=v): thresh = thresh*nanmax(signal) coeff = pywt.wavedec(signal, wavelet, level=8,mode="per" ) coeff[1:] = (pywt.threshold(i, value=thresh, mode='soft' ) for i in coeff[1:]) reconstructed_signal = pywt.waverec(coeff, wavelet, mode="per" ) return reconstructed_signal fig, ax = subplots(figsize=(8,4)) ax.plot(signal, color="b", alpha=0.5, label=' ') rec = lowpassfilter(signal,w) ax.plot(rec, 'r', label='DWT  ', linewidth=2) ax.legend() ax.set_title('      :%s\n    %s'%(v,w), fontsize=12) ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) show() 

 import pywt from pylab import* db_wavelets = ['db1', 'db38'] fig, axarr = subplots(ncols=2, nrows=5, figsize=(14,8)) fig.suptitle('  : db1,db38', fontsize=14) for col_no, waveletname in enumerate(db_wavelets): wavelet = pywt.Wavelet(waveletname) no_moments = wavelet.vanishing_moments_psi family_name = wavelet.family_name for row_no, level in enumerate(range(1,6)): wavelet_function, scaling_function, x_values = wavelet.wavefun(level = level) axarr[row_no, col_no].set_title("{} : {}.   : {}. :: {} ".format( waveletname, level, no_moments, len(x_values)), loc='left') axarr[row_no, col_no].plot(x_values, wavelet_function, 'b--') axarr[row_no, col_no].set_yticks([]) axarr[row_no, col_no].set_yticklabels([]) tight_layout() subplots_adjust(top=0.9) show() 

 import pandas as pd import pywt filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) data = df[4].values w=['db1', 'db38'] for v in w: n_level=pywt.dwt_max_level(len(data),v) print('  %s   : %s ' %(v,n_level)) 

 import pandas as pd import scaleogram as scg from pylab import * import pywt filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) signal = df[4].values discrete_wavelets =[('db38', 5),('db1',11)] for v in discrete_wavelets: def lowpassfilter(signal, thresh = 0.63, wavelet=v[0]): thresh = thresh*nanmax(signal) coeff = pywt.wavedec(signal, wavelet, mode="per" ) coeff[1:] = (pywt.threshold(i, value=thresh, mode='soft' ) for i in coeff[1:]) reconstructed_signal = pywt.waverec(coeff, wavelet, mode="per" ) return reconstructed_signal wavelet = pywt.DiscreteContinuousWavelet(v[0]) phi, psi, x = wavelet.wavefun(level=v[1]) fig, ax = subplots(figsize=(8,4)) ax.set_title("   : %s,level=%s"%(v[0],v[1]), fontsize=12) ax.plot(x,phi,linewidth=2) fig, ax = subplots(figsize=(8,4)) ax.plot(signal, color="b", alpha=0.5, label='   ') rec = lowpassfilter(signal, 0.4) ax.plot(rec, 'r', label='DWT  ', linewidth=2) ax.legend() ax.set_title('   \n   :%s,level=%s'%(v[0],v[1]),fontsize=12) ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) wavelet = 'cmor1-0.5' ax = ax = scg.cws(rec, scales=arange(1,128), wavelet=wavelet,figsize=(8, 4), cmap="jet", ylabel=' ', xlabel=" ", yscale="log", title='CWT  \n(   DWT  )') show() 

 from scipy import* from pylab import* import pandas as pd def kalman_filter( z, F = eye(2), #     (transitionMatrix) Q = eye(2)*3e-3, #   (processNoiseCov) H = eye(2), #   (measurement) R = eye(2)*3e-1 #    (measurementNoiseCov) ): n = z.shape[0] #   #m = z.shape[1] #   x = z[0,:].transpose() #      P = Q #       r = [ x ] #  for i in range(1,n): # : x = F.dot( x ) #  ()   P = F.dot(P).dot( F.transpose() ) + Q #       # : y = z[i] - H.dot(x) #  ,      ()  S = H.dot(P).dot(H.transpose()) + R #      ( ) K = P.dot(H.transpose()).dot(linalg.inv(S)) #     x = x + K.dot(y) #     P = ( eye(z.shape[1]) - K.dot(H) ).dot(P) #       r.append( x.transpose()) return array(r) filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) signal = df[4].values N= signal .shape[0] #   # 2 dimensions y0 =linspace(0.0,2500.0,N) y1 =signal d = array(list(zip(y0,y1))) dn = dr = kalman_filter(dn) fig, ax = subplots(figsize=(8,4)) ax.plot( d[:,0], d[:,1],'b', label=' ' ) ax.plot( r[:,0], r[:,1],'r', label='kalman' ) ax.legend(loc='upper left') ax.set_title('       \n (  )', fontsize=12) ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) show() 

 from pykalman import KalmanFilter import pandas as pd from pylab import * import scaleogram as scg filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) signal = df[4].values measurements =signal kf = KalmanFilter(transition_matrices=[1] ,# A      t  t+1 observation_matrices=[1],# C   initial_state_mean=measurements[0],#   initial_state_covariance=[1],#    observation_covariance=[1],# R    transition_covariance= 0.01) # Q     state_means, state_covariances = kf.filter(measurements)# ,   state_means=[w[0] for w in state_means] state_std = sqrt(state_covariances[:,0]) fig, ax = subplots(figsize=(8,4)) ax.plot(measurements ,'b', label=' ') ax.plot(state_means ,'r', label='kalman- ') ax.legend() ax.set_title('      \n ( pyKalman)', fontsize=14) ax.set_ylabel(' ', fontsize=12) ax.set_xlabel('', fontsize=12) wavelet = 'cmor1-0.5' ax = ax = scg.cws(signal, scales=arange(1,40), wavelet=wavelet, figsize=(8, 4),cmap="jet", ylabel=' ', xlabel=" ", yscale="log", title='CWT    \n( )') ax = ax = scg.cws(state_means, scales=arange(1,40), wavelet=wavelet, figsize=(8, 4), cmap="jet", ylabel=' ', xlabel=" ", yscale="log", title='CWT     \n(  )') show() 

 <source lang="python">from scipy import * import pandas as pd from pylab import * import pywt import scaleogram as scg def get_ave_values(xvalues, yvalues, n = 6): signal_length = len(xvalues) if signal_length % n == 0: padding_length = 0 else: padding_length = n - signal_length//n % n xarr = array(xvalues) yarr = array(yvalues) xarr.resize(signal_length//n, n) yarr.resize(signal_length//n, n) xarr_reshaped = xarr.reshape((-1,n)) yarr_reshaped = yarr.reshape((-1,n)) x_ave = xarr_reshaped[:,0] y_ave = nanmean(yarr_reshaped, axis=1) return x_ave, y_ave def plot_signal_plus_average(time, signal, average_over = 5): fig, ax = subplots(figsize=(8, 4)) time_ave, signal_ave = get_ave_values(time, signal, average_over) ax.plot(time_ave, signal_ave,"b", label = '   (n={})'.format(5)) ax.set_xlim([time[0], time[-1]]) ax.set_ylabel(' ', fontsize=12) ax.set_title('     SMA', fontsize=14) ax.set_xlabel('', fontsize=12) ax.legend() return signal_ave filename = 'Test_set/Bearing2_3/acc_00001.csv' df = pd.read_csv(filename, header=None) df_nino = df[4].values N = df_nino.shape[0] time = arange(0, N) signal = df_nino signal_ave=plot_signal_plus_average(time, signal) wavelet = 'cmor1-0.5' ax = ax = scg.cws(signal, scales=arange(1,40), wavelet=wavelet, figsize=(8, 4),cmap="jet", ylabel=' ', xlabel=" ", yscale="log", title='CWT  %s \n(     )'%filename) ax = ax = scg.cws(signal_ave, scales=arange(1,40), wavelet=wavelet, figsize=(8, 4), cmap="jet", ylabel=' ', xlabel=" ", yscale="log", title='CWT  %s \n(    )'%filename) show()