DAQ - Measurement - Extraction of DataSet - PCA Analysis

[inst_soleil]

Hi Henrik,

Amazing!

I can see the different colors that correspond to the different data set in the whole data. Thank you very much for sharing your screen. I am aware of the noise and wanted use of signal filtering. I will also try to implement Savitzky-Golay and compare the results.

Could you please share the code with me? Then I will be able to play around with it and, more significantly, gain an understanding of how it works. Due to the fact that I perceived it to be highly complicated, I was unable to avoid it. Next, I will share the final result in the community if it is ok with you too.

All the Best

M

Thanks very much in advance for your time, and effort. 

Best,m

[Henrik_Volkers]

Oh sorry, missed to attach the vi ..

will post it tomorrow, when I am back at my PC 

[inst_soleil]

Wonderful.

looking forward 

best, M

[Henrik_Volkers]

here we go

[inst_soleil]

Hi Henrik,

I could jsut resolve the Labview in my PC. I had to ask IT to reseolve the issue with Lavbiew.

Thanks for the code. I am going to explore it now. I will update you asap.

Best

m

[inst_soleil]

Hi Henrik,

I have looked at your code. The question I have is that you are seperating each column (4 of them) from the data I gave, classing them to plot and draw only one of them?

Column1 and column2

Column1 and column3

Column1 and column4

Column1 and column5

[Henrik_Volkers]

are we talking about this part?

this was just a quik and dirty way to select the summ of all three data columns or one selected column.

the get array element with unwired index input will 'auto index' the elemnts/data columns (0 to 4) . the 5th get array element  has an wired index input 😉

I assume it's a good idea to choose the same time slice for all three data sets , to do so,  just use the added columns data, extract the index and length information (output of second last loop) , the output arrays are index and length. the last loop cut the data, loop over all data columns and you can feed your fit. 

You need to define good filter settings and min peak step.

the min peak step need some individual data based setting. I would try some percentage of the min max difference ...

BTW: Kudos wellcome 😉

[inst_soleil]

Ok. I understand what do you mean, but I need a dynamic data which means, I am collecting datta in 1D array of 4 elements. I nthe diagram you sent there is 2D array and that is splitted. My labview knowledge is limited so I cannot comprehand

[Henrik_Volkers]

if you look into the other part of the hard wired case structure, you migth find how i read the file 😉

try to avoid dynamic data  , have a look at the shipped exsamples and use DAQ vis (help.find exsamples)

you get waveforms (wfrms) or arrays of wfrms ..

[inst_soleil]

 

Hi again,

In this case (see screenshot), I cannot upload or read the *.csv file that I want, the one I uploaded previously that contains two columns of data where I plot column2 vs. column1. I do not know why!

Column2 contains the measured data, and column1 contains the time that corresponds to the measured data.

I choose T (boolen) on this case structure, and it asks me to select the file. If I choose a *.csv file that contains only one column (column2), it reads it and plots the curves vs. time, which is the system time, which I don't want. But when I want to select data with two columns (time vs. measured data), it doesn't read it. I need to plot column2 vs. column1 to be able to define the max and min values (peaks and Valleys) that correspond to the values in column1. That is, I also need the column1 values.

In other examples in Labview, such as Peak Finder VI, the same thing happens. It asks for only one column of data and processes it to find the peaks. But what I need is to plot column2 vs. column1 and find the peaks&valleys and corresponding peaks&valley locations. Peak&valley locations should be from column1, not from system time.

Any idea?

Thanks for helping.