Reduce the number of peaks from findpeaks command?

My code does a few things: it averages data, smooths it, finds the peaks in the data, and then calculates the damping ratio of the data.

I have wayyyy too many peaks. I only want/need about 5 peaks--circled in blue. Right now, my plot looks like this:

How can I do this?

Thanks in advance!

clear

close

clc

load data

v_data = data;

v_mean = mean(v_data,2)'; %getting mean

v_std = std(v_data, 0, 2)'; %getting standard deviation

v_smooth = smooth(v_mean); %smoothing out noisy data

t = v_data(:,1); %picking off a time vector

ind_1 = 1:length(t);

t = t(ind_1);

v = v_smooth(ind_1);

ind_2 = 1:length(t);

[ pks, locs] = findpeaks(v(ind_2));

n_peaks = numel(locs);

%---Calculate Logarithmic Decrement and damping ratio

log_dec = zeros(length(n_peaks));

for nn = 1:n_peaks-1

log_dec(nn)= log(pks(nn)/pks(nn+1)); % computed with n = 1 periods apart

end

mean_dec = mean(log_dec); %average decrement

%---damping ratio---

damp_ratio = 1/sqrt(1+((2*pi/(mean_dec))^2))

pf = polyfit(t(locs), log(pks), 3);

curve_fit = exp(polyval(pf,t));

figure()

axes

plot(t, v_data(:,2), 'Displayname', 'Data 1'); hold on;

plot(t, v_data(:,3), 'Displayname', 'Data 2');

plot(t, v_data(:,4), 'Displayname', 'Data 3');

plot(t, v_data(:,5), 'Displayname', 'Data 4');

plot(t, v_data(:,6), 'Displayname', 'Data 5');

plot(t(locs), pks, 'rd', 'markersize', 10, 'displayname', 'Peaks'); hold on;

plot(t, v_smooth, 'k', 'linewidth', 1.5, 'Displayname', 'Smoothed Average Data');

plot( t, curve_fit,'g-.','LineWidth', 1.5, 'Displayname', 'Peak Envelope');

grid minor

title('Data: 1 - 5');

legend('location', 'ne');

xlabel('Time (s)'); ylabel('Velocity (m/s)');

Answer 1

[Star Strider]

The findpeaks function has a number of name-value pair arguments that you can use. The two I use most often are 'MinPeakProminence' and 'MinPeakDistance'. The first selects the peaks based on their relation to the signal levels around them, and the second does not return peaks within a specified distance of identified peaks. One or botth of those (and possibly others) should do what you want.

load('data.mat')
% whos

v_data = data;
v_mean = mean(v_data,2)'; %getting mean
v_std = std(v_data, 0, 2)'; %getting standard deviation
v_smooth = smooth(v_mean); %smoothing out noisy data
t = v_data(:,1); %picking off a time vector
ind_1 = 1:length(t);
t = t(ind_1);  
v = v_smooth(ind_1);  
ind_2 = 1:length(t);
[ pks, locs] =  findpeaks(v(ind_2), 'MinPeakProminence',7.5, 'MinPeakDistance',280);
n_peaks = numel(locs);  
%---Calculate Logarithmic Decrement and damping ratio
log_dec = zeros(length(n_peaks));  
for nn = 1:n_peaks-1 
    log_dec(nn)= log(pks(nn)/pks(nn+1));  % computed with n = 1 periods apart
end  
mean_dec = mean(log_dec); %average decrement
%---damping ratio--- 
damp_ratio = 1/sqrt(1+((2*pi/(mean_dec))^2))

damp_ratio = 0.0624 + 0.1839i

pf = polyfit(t(locs), log(pks), 3);
curve_fit = exp(polyval(pf,t));
figure()
axes
plot(t, v_data(:,2), 'Displayname', 'Data 1'); hold on;
plot(t, v_data(:,3), 'Displayname', 'Data 2'); 
plot(t, v_data(:,4), 'Displayname', 'Data 3'); 
plot(t, v_data(:,5), 'Displayname', 'Data 4'); 
plot(t, v_data(:,6), 'Displayname', 'Data 5'); 
plot(t(locs), pks,  'rd', 'markersize', 10, 'displayname', 'Peaks'); hold on;
plot(t, v_smooth, 'k', 'linewidth', 1.5, 'Displayname', 'Smoothed Average Data'); 
plot( t, curve_fit,'g-.','LineWidth', 1.5, 'Displayname', 'Peak Envelope');

Warning: Imaginary parts of complex X and/or Y arguments ignored.

grid minor
title('Data: 1 - 5');
legend('location', 'ne');
xlabel('Time (s)'); ylabel('Velocity (m/s)');

I do not understand your data, and apparently the smoothing steps interact with the returned peaks. This is an improvement, however you will need to continue experimenting yourself to get the results you want.

Meanwhile see: How to filter noise from time-frequency data and find natural frequency of a cantilever? for one approach to fitting this sort of problem.

.