Re: Electrical signal analisis

[Jaime Alberto Silva]

Thanks a lot for your help but those libraries are for signals that

change frecuency in time. I need simpler tools for signals that keep the 
same frecuency all the time. For the moment I'll use the functions 
attached to find the delay between signals and the fft to obtain the 
fundamental frecuency (with max()).

Joe Koski wrote:

> on 8/21/04 11:10 AM, robert Macy at address@hidden wrote:
>
>  
>
> > Group (and Joe)
> >
> > http://iut-saint-nazaire.univ-nantes.fr/~auger/tftb.html
> > The URL at France didn't seem active.  Any new location for
> > same information?
> >
> >     - Robert -
> >
> >    
> Robert,
>
> I just tried and downloaded the full UNIX tftb.tar.gz version from the
> mirror site without any problems. I did not try the .zip version or the main
> site.
>
> Joe Koski 
>
>
>  
>
> > On Wed, 18 Aug 2004 10:30:45 -0600
> > Joe Koski <address@hidden> wrote:
> >    
> >
> > > There are two more signal processing libraries (besides
> > > those with octave
> > > and octave-forge) of which you should also be aware, and
> > > possibly install.
> > > The URLs for them are
> > >
> > > http://www-stat.stanford.edu/~wavelab/
> > > http://wiki.octave.org/wiki.pl?WavelabOnOctave
> > >
> > > And 
> > >
> > > http://iut-saint-nazaire.univ-nantes.fr/~auger/tftb.html
> > > http://perso.ens-lyon.fr/patrick.flandrin/emd.html
> > >
> > > The latter has a routine extr.m that finds all local
> > > maxima and minima of a
> > > time (or frequency) series.
> > >
> > > For help with octave and octave-forge signal processing
> > > routines, check the
> > > information at 
> > >
> > > http://octave.sourceforge.net/index/index.html
> > >
> > > There is a wealth of good signal processing software
> > > available. Good luck.
> > >
> > > Joe Koski
> > >
> > > on 8/17/04 7:10 PM, Jaime Alberto Silva at
> > > address@hidden wrote:
> > >
> > >      
> > >
> > > > I'm new in the use of this tools (octave, mathlab,
> > > >        
> > > etc.) and since I'm a
> > >      
> > >
> > > > full time Linux user I am using octave.
> > > >
> > > > Currently, I'm trying to do some electrical signal
> > > >        
> > > analysis and I will
> > >      
> > >
> > > > like to know if there are predefined functions for some
> > > >        
> > > calculations or
> > >      
> > >
> > > > how can I do them with octave:
> > > >
> > > > 1. Frecuency: I have founded in the "Signal Processing"
> > > >        
> > > section the fft
> > >      
> > >
> > > > function, OK it is great but how do I find the
> > > >        
> > > frecuencies from the
> > >      
> > >
> > > > returned vector. I know that using the max function I
> > > >        
> > > can find the
> > >      
> > >
> > > > fundamental frecuency like this:
> > > >
> > > > # Create the signal:
> > > > octave:190> t= [0: 1/2000: 5 - 1/2000];
> > > > octave:191> sint= sin(2*pi*60*t);
> > > > # Transform the signal:
> > > > octave:192> ft= abs(fft(sint));
> > > > # Find the position of the max value of the
> > > >        
> > > transformation:
> > >      
> > >
> > > > octave:193> [val, idx]= max(ft);
> > > > # Use the sampling rate to obtain the fundamental
> > > >        
> > > frecuency:
> > >      
> > >
> > > > octave:194> fund_frec= 2000 * idx/ length(ft)
> > > > fund_frec = 60.200
> > > > octave:195>
> > > >
> > > > But how can I find the other frecuencies in case it has
> > > >        
> > > more components
> > >      
> > >
> > > > like in
> > > > f= sin x + 0.5 sin 3x + 2 sin 10x ?
> > > >
> > > > 2. Phasors: If I have 2 signals like v1 an v2 how do I
> > > >        
> > > find the angle of
> > >      
> > >
> > > > v2 with respect of v1. I mean if v1= sin (t) and v2=
> > > >        
> > > sin (t + 15°) and I
> > >      
> > >
> > > > have sampled those signals in time, how do I find that
> > > >        
> > > v2 is delayed 15
> > >      
> > >
> > > > degrees with respect to v1? taking in account that v2
> > > >        
> > > samples where
> > >      
> > >
> > > > taken in different times to v1 samples like this: v1(1)
> > > >        
> > > was taken at T0,
> > >      
> > >
> > > > v2(1) was taken at T0 + dT. And, of course, you know
> > > >        
> > > the value of dT.
> > >      
> > >
> > > > 3. Electric Power: If I have sampled the signals of
> > > >        
> > > voltage (V) and
> > >      
> > >
> > > > current (I) are there any tools that l can use to
> > > >        
> > > obtain the active (P)
> > >      
> > >
> > > > and reactive (Q) power, the S vector angle, etc. ? I
> > > >        
> > > know that I can
> > >      
> > >
> > > > multiply the signals but, you know if there are tools
> > > >        
> > > for that why do it
> > >      
> > >
> > > > again.
> > > >
> > > > Thanks in advance for any help you can bring me.



--
Jaime Alberto Silva Colorado

function d = signal_delay (signal1, signal2, t)

                                % d = SIGNAL_DELAY (sig1, sig2, t) finds
                                % the delaying of sig2 with respect to
                                % sig1.
%
                                % Inputs: 
                                % - sig1: reference sinusoidal signal.
                                % - sig2: delayed sinusoidal signal.
%
                                % Outputs:
                                % - d: delay of sig2 with respect to sig1.

  if (nargin < 2)
                                % error
  end
  if (nargin == 2)
    t= 1:length(signal1);
  end

  sig1= [];
  sig2= [];
  s= size(signal1);
  if (s(1,1) != 1)
    sig1= signal1';
  else
    sig1= signal1;
  end
  s= size(signal2);
  if (s(1,1) != 1)
    sig2= signal2';
  else
    sig2= signal2;
  end
  

  iz1= zero_crossings (sig1,t);
  iz2= zero_crossings (sig2,t);
  dif1= diff (sig1);
  dif2= diff (sig2);
  iz2r= [];
  if (dif1(iz1(1)) < 0)
    for i = iz2
      if dif2(i) < 0
        iz2r= i;
        break;
      end
    end
  else
    for i = iz2
      if dif2(i) >= 0
        iz2r= i;
        break;
      end
    end
  end

    
  lx= iz1(1); # x coord at the left of the cross
  ly= sig1(lx); # y coord at the left of the cross
  rx= lx + 1; # x coord at the rigth of the cross
  ry= sig1(rx); # y coord at the rigth of the cross
  m= (ry - ly)/(rx - lx);
  b= ly - m * lx; # y = mx + b => b= y - mx
  t1= -1 * b / m; # y= mx + b, y= 0 => x = -b/m

#   disp ("x1:"), disp (lx);
#   disp ("y1:"), disp(ly);
#   disp ("y'1:"), disp(dif1(lx));

  lx= iz2r; # x coord at the left of the cross
  ly= sig2(lx); # y coord at the left of the cross
  rx= lx + 1; # x coord at the rigth of the cross
  ry= sig2(rx); # y coord at the rigth of the cross
  m= (ry - ly)/(rx - lx);
  b= ly - m * lx; # y = mx + b => b= y - mx
  t2= -1 * b / m; # y= mx + b, y= 0 => x = -b/m

#   disp ("x2:"), disp (lx);
#   disp ("y2:"), disp(ly);
#   disp ("y'2:"), disp(dif2(lx));

  d= t2 - t1;
end

  
  

function indzer = zero_crossings (data,t);

% indzer = ZERO_CROSSINGS (x,t) finds zero-crossings
%
% inputs : - x : analyzed signal
%          - t (optional) : sampling times, default 1:length(x)
%
% outputs : - indzer = indices of zero-crossings

  if(nargin==1)
    t=1:length(x);
  end
  m = length(t);
  x= data(t(1):t(m));

  x1= x(1:m-1);
  x2= x(2:m);
  indzer = find(x1.*x2<0);

  if any(x == 0)
    iz = find( x==0 );
    indz = [];
    if any(diff(iz)==1)
      zer = x == 0;
      dz = diff([0 zer 0]);
      debz = find(dz == 1);
      finz = find(dz == -1)-1;
      indz = round((debz+finz)/2);
    else
      indz = iz;
    end
    indzer = sort([indzer indz]);
  end
end