Converting xlsx files to frequency domains using the fft algorithm

The above graph is a graph using Var1_Time and Hammer (input) data Var2_1 through Var2_5, the time data of the referenced file 120mm.xlsx.

>> plot(Var1_Time,Var2_1)

>> hold on

>> plot(Var1_Time,Var2_2)

>> plot(Var1_Time,Var2_3)

>> plot(Var1_Time,Var2_4)

>> plot(Var1_Time,Var2_5)

But I want to make a graph about Hammer (input) data by changing Var1_Time, which is time data, to frequency domain using the fft algorithm, but I don't know how. I'd really appreciate your help.

Answer 1

[Star Strider]

Try this —

T1 = readtable('120mm.xlsx')

T1 = 10240×13 table
    Var1_Time      Dev3_ai0      Dev3_ai1       Var2_1        Var2_2        Var2_3        Var2_4        Var2_5        Var3_1        Var3_2        Var3_3       Var3_4        Var3_5  
    __________    __________    __________    __________    __________    __________    __________    __________    __________    ___________    _________    _________    __________

             0     -0.078707      0.052439     -0.043113     -0.015541      0.033087      0.015541     -0.078707     0.0090577       0.022406     0.036231     0.016685      0.052439
    0.00097656     -0.074696     -0.038138    -0.0035092     -0.012032       0.01905      0.023061     -0.074696      0.068171    -0.00095344     0.016685     0.031464     -0.038138
     0.0019531      0.069182      0.050056     -0.038602     -0.087229    -0.0095251      -0.01905      0.069182      0.030033       0.028126     0.024789    -0.041951      0.050056
     0.0029297      0.066174       0.14254      0.037599     -0.076702     -0.054644      0.011029      0.066174        0.0553     -0.0042905    0.0066741    -0.012395       0.14254
     0.0039062     -0.036095      0.036231     -0.030079     -0.014538      0.037098     -0.052137     -0.036095      0.082473       0.018115      0.03194     -0.03051      0.036231
     0.0048828     -0.030079      0.024313     0.0035092      0.027573     -0.013034    0.00050132     -0.030079      0.036231       0.038138     0.054823    -0.023836      0.024313
     0.0058594      0.014538      0.030033     -0.030079     -0.051135    -0.0070185     -0.066174      0.014538      0.054346       0.018115    -0.013348    -0.017162      0.030033
     0.0068359     0.0095251      0.017162      -0.01153      0.015541      -0.04562     -0.051135     0.0095251     -0.022883       0.040044     -0.02765    -0.054823      0.017162
     0.0078125      0.016544      0.023836      -0.01153      0.012032    -0.0085224      0.037098      0.016544      0.057206       0.057206    -0.011441     0.048149      0.023836
     0.0087891     -0.052137     -0.008581     -0.048628     -0.041108      -0.09124      0.015541     -0.052137      0.035277       0.013348     -0.06245     0.014778     -0.008581
     0.0097656     0.0020053    -0.0057206     0.0035092     -0.048628     -0.023061     0.0080211     0.0020053      0.080566     -0.0023836     -0.03194     0.078659    -0.0057206
      0.010742    -0.0035092      0.051009             0      0.089736     -0.026069     0.0030079    -0.0035092      0.064834      0.0028603    -0.065311     0.055776      0.051009
      0.011719      0.014538      0.044335        0.0381     0.0050132      0.041609      0.055145      0.014538      0.067218       -0.02765     0.051486     0.043382      0.044335
      0.012695     0.0090237      0.031464      -0.05314     0.0025066    -0.0040105      0.028074     0.0090237    -0.0014302    -0.00095344    -0.016685     0.051009      0.031464
      0.013672     -0.012032             0     -0.046623    -0.0040105     -0.025567     -0.050132     -0.012032      0.010488      -0.018115    -0.022883     0.053869             0
      0.014648     -0.001504     -0.046719     -0.015541      0.016544      0.021055     -0.047124     -0.001504     0.0052439       0.029557     0.023359     0.074845     -0.046719

VN = T1.Properties.VariableNames([1 4:end]);
t = T1{:,1};
s = T1{:,4:end};

Ts = mean(diff(t))

Ts = 9.7656e-04

Tsd = std(diff(t))

Tsd = 4.0216e-05

Fs = 1/Ts

Fs = 1.0240e+03

[sr, tr] = resample(s, t, Fs);

[FTs1,Fv] = FFT1(sr,tr);

figure
plot(Fv, mag2db(abs(FTs1)*2))
grid
xlim([0 max(Fv)])
xlabel('Frequency')
ylabel('Magnitude (dB)')
legend(strrep(VN(2:end),'_','\_'), 'Location','NE', 'NumColumns',2)

idx = reshape([1:5; 6:10], [],1);
figure
tiledlayout(5,2)
for k = 1:size(FTs1,2)
    nexttile
    plot(Fv,  mag2db(abs(FTs1(:,idx(k)))*2))
    title(strrep(VN{idx(k)+1},'_','\_'))
    grid
    axis([0 max(Fv)  -100 10])
end

function [FTs1,Fv] = FFT1(s,t)
    t = t(:);
    L = numel(t);
    if size(s,2) == L
        s = s.';
    end
    Fs = 1/mean(diff(t));
    Fn = Fs/2;
    NFFT = 2^nextpow2(L);
    FTs = fft((s - mean(s)) .* hann(L).*ones(1,size(s,2)), NFFT)/sum(hann(L));
    Fv = linspace(0, 1, NFFT/2+1)*Fn;
    Iv = 1:numel(Fv);
    FTs1 = FTs(Iv,:);
end

Your data are not regularly sampled, so it is necessary to use the resample function here to produce reliable results, since fft and nearly every other signal processing function require regularly-sampled data.

I created the ‘FFT1’ function for my convenience. Feel free to use it.

.

[호태 강]

Wow, thank you so much. I'll bring you as a big brother. However, I need to divide the Var2_1~Var2_5 and Var3_1~Var3_5 graphs in order to organize the experimental data, so can I get one more help?

[Star Strider]

As always, my pleasure!

I will be happy to help you, however I am not certain what you want. What do you intend by ‘... I need to divide the Var2_1~Var2_5 and Var3_1~Var3_5 graphs in order to organize the experimental data ...’? How do you want to divide them?

Perhaps this —

T1 = readtable('120mm.xlsx');
VN = T1.Properties.VariableNames([1 4:end]);
t = T1{:,1};
s = T1{:,4:end};

Ts = mean(diff(t));
Tsd = std(diff(t));
Fs = fix(1/Ts)

Fs = 1024

[sr, tr] = resample(s, t, Fs);

[FTs1,Fv] = FFT1(sr,tr);

figure
tiledlayout(5,1)
for k = 1:5
    nexttile
    plot(Fv,  mag2db(abs(FTs1(:,k))*2))
    title(strrep(VN{k+1},'_','\_'))
    grid
    axis([0 max(Fv)  -100 10])
end

figure
tiledlayout(5,1)
for k = 1:5
    nexttile
    plot(Fv,  mag2db(abs(FTs1(:,k+5))*2))
    title(strrep(VN{k+6},'_','\_'))
    grid
    axis([0 max(Fv)  -100 10])
end

function [FTs1,Fv] = FFT1(s,t)
    t = t(:);
    L = numel(t);
    if size(s,2) == L
        s = s.';
    end
    Fs = 1/mean(diff(t));
    Fn = Fs/2;
    NFFT = 2^nextpow2(L);
    FTs = fft((s - mean(s)) .* hann(L).*ones(1,size(s,2)), NFFT)/sum(hann(L));
    Fv = linspace(0, 1, NFFT/2+1)*Fn;
    Iv = 1:numel(Fv);
    FTs1 = FTs(Iv,:);
end

.

[호태 강]

Thanks for help, But what i want is your 1st graph(Frequency-Magnitude graph including Var2_1~Var3_5) i need 2 Frequency-Magnitude graph. One is Var2_1~2_5. Other is 3_1~3_5. Again, thanks for help

[Star Strider]

As always, my pleasure!

O.K.

Try this —

T1 = readtable('120mm.xlsx');
VN = T1.Properties.VariableNames([1 4:end]);
t = T1{:,1};
s = T1{:,4:end};

Ts = mean(diff(t))

Ts = 9.7656e-04

Tsd = std(diff(t))

Tsd = 4.0216e-05

Fs = 1/Ts

Fs = 1.0240e+03

[sr, tr] = resample(s, t, Fs);

[FTs1,Fv] = FFT1(sr,tr);

figure
plot(Fv, mag2db(abs(FTs1(:,1:5))*2))
grid
xlim([0 max(Fv)])
xlabel('Frequency')
ylabel('Magnitude (dB)')
legend(strrep(VN(2:6),'_','\_'), 'Location','best')

figure
plot(Fv, mag2db(abs(FTs1(:,6:end))*2))
grid
xlim([0 max(Fv)])
xlabel('Frequency')
ylabel('Magnitude (dB)')
legend(strrep(VN(7:end),'_','\_'), 'Location','best')

function [FTs1,Fv] = FFT1(s,t)
    t = t(:);
    L = numel(t);
    if size(s,2) == L
        s = s.';
    end
    Fs = 1/mean(diff(t));
    Fn = Fs/2;
    NFFT = 2^nextpow2(L);
    FTs = fft((s - mean(s)) .* hann(L).*ones(1,size(s,2)), NFFT)/sum(hann(L));
    Fv = linspace(0, 1, NFFT/2+1)*Fn;
    Iv = 1:numel(Fv);
    FTs1 = FTs(Iv,:);
end

.

[호태 강]

Thanks!!! I really appreciate

[Star Strider]

As always, my pleasure!