Simulink error: The data abscissae must be distinct and strictly monotonic.

Hi all,

I have constructed a simulink model, where within the model I have a MATLAB function block as you can see. Within that block, a very simple code loads data from a .mat file and interpolates using interp1.

However I get the following error when I run the model which I have managed to trace to the line in the MATLAB function block where the interpolation occurs:

An error occurred while running the simulation and the simulation was terminated

Caused by:

• The data abscissae must be distinct and strictly monotonic.

Component:Simulink | Category:Block error

This is simple enough, as one would assume that my x vector in the interp1 function includes non-monotonic data. But the problem is that it doesn't! I have tested that by running the code in the MATLAB function block by itself as a normal matlab file, and also used unique to tripple check and it works perfectly. I also checked where the current folder to ensure that the file is taken from the right place. So I really do not know why the interp1 function presents this problem when I am running it in Simulink. It must be something that I don't understand.

Any suggestions?

Kind Regards,

KMT.

P.S. Attached you can find the files that I am using. Just download and run if you need to.

Answer 1

[Daniel M]

None of the variables in the matfile are monotonic.

all(sign(diff(Tga))==1) % ans = 0
all(sign(diff(Tm1))==1) % ans = 0
all(sign(diff(Tm2))==1) % ans = 0
all(sign(diff(thx))==1) % ans = 0

x = -10:10;
all(sign(diff(x))==1) % ans = 1

Perhaps try unwrap, as the values of thx seem to be between -pi and pi.

[KostasK]

That is extremely weird, cause when I type those lines of code for thx I get 1. I suspect something might be happening with my directories in which case this is more of my problem rather than simulink's. Will keep you posted

[Daniel M]

Sorry, I edited my code to include the "==1", make sure you include that as well.

You can also check it visually by plotting the values:

figure
stem(thx)

[KostasK]

Hi Daniel,

So I did the checks that you did as well and you are right. The thx vector is a vetor that spans from -20 to -19 degrees, hence at 0 (or 180 degrees) the vector is non-monotonic. I managed to therefore make thx monotonic using sort.

However there is stil somethig weird happening, which was also happening before (hence why I assumed the vector was monotonic to begin with). When I use interp1 on the un-sorted thx vector it still works! I do not understand why... Does the same thing happen to you when you try to run interp1 with the un-sorted vector outside simulink?

e.g. if you try to run

interp1(thx, Tga, 1.2)

Edit: it worked in Simulink when I did the following:

[~,i] = sort(thx) ;
thxS = thx(i) ;
TgaS = Tga(i) ;

T_G = interp1(thxS, TgaS, input_variable) ;

[Daniel M]

I have solved it. If you unwrap thx such that it become monotonic, now there is a duplicate! It occurs because thx has entries for 3.1416 and -3.1416 (and unwrapped they become the same). So I removed the second entry.

Add this code to your matlab function after you load the data:

thx(201) = [];
Tga(201) = [];
thx = unwrap(thx);

And this function shouldn't have any more problems. And the simulink program generates an error further along the simulation (hey... it's still progress!).

[KostasK]

Ok, I understand. Thanks very much for your help!