"break" in for loop seems to not be working

I'm having an issue with a current bit of code I am working on (see below). It seems that the "break" in my for loop is not working. I must have an error somewhere but I cannot see it. Here is a rundown of the section of code:

1. counter and relevant vectors are initialized
2. start of for loop
3. if statement that is used to revert back to previous loop data if a condition fails
4. calculations
5. end

My output is in the image below, note that all of the preload_sf and some of the sep_sf values should not be there, as the break will cause the for loop to convert current values to previous values that satisfy the conditions (user inputs), along with ending the loop. Instead, it seems that the break does nothing, and the loop cycles through all available preload values. So does anyone have an idea as to why my break command is not working for me in this context?

Hopefully this makes sense with only this small section of code. The script right now is around 400 lines, and I don't think most of that is necessary to understanding this problem.

        preload_vector=0.50:0.01:0.90; %creating vector of preload values
        length1=length(preload_vector); %determining the length of the vector for later use
        
        if analysis == "Static" %implying static analysis           
            count2=0; %initializing second counter
            %initializing safety factors
            Ni=1:length1; %initializing
            Nsep_calc=1000000:10000000; %arbitrary vector for initialization
            Ny_calc=1000000:10000000; %arbitrary vector for initialization
            
            for L=1:length1 %cycling through all available preload values until safety factors are met
                
                %stopping secondary loop if safety factors are too low, one
                if Ni(L)<1.0 || Nsep_calc(L)<Nsep || Ny_calc(L)<Ny %if one of these fails - then the loop must revert
                    Ni(L)=Ni(count2); %sets to value in last loop
                    Nsep_calc(L)=Nsep_calc(count2); %sets to value in last loop
                    Ny_calc(L)=Ny_calc(count2); %sets to value of previous loop
                    break
                end
                    
                count2=count2+1; %used to recall values from of previous loop
                
                preload=preload_vector(count2);
                %Fastener Calculations
                Fi=preload*ps*At; %calculating force on fastener due to preload for each grade
                max_stress=(Pb+Fi)/At; %calculating stress due to preload and external loads for each grade
                Ti=u*Fi*major_dia; %calculating torque to acheive preload from above
                tau=(16*Ti)/(pi*((2*sqrt(Am/pi))^3)); %calculating shear stress from preload for each grade
                eff=sqrt((Fi/At)^2+(3*tau^2)); %calculating effective stress from preload for each grade
                Psep=Fi/(1-C); %load at which sep. occurs
                
                %storing last value of Ni for each secondary loop iteration
                Ni(L)=ts/eff; %calculating preload application safety factor
                
                %calculating yield safety factor
                Ny_calc(L)=ys/max_stress;
                
                %calculating separation safety factor
                Nsep_calc(L)=Psep/max_load;
            end
                        %storing static analysis data
            Ti_stored(i)=Ti;
            preload_stored(i)=preload; %percent preload
            Ny_calc_stored(i)=Ny_calc(L);
            Nsep_calc_stored(i)=Nsep_calc(L);
            Ni_stored(i)=Ni(L);
        end

Answer 1

[JAKE SEIFERT]

So, after some sleep and more troubleshooting today I figured out a few things:

1. The "break" command does not work if the criteria used to break the loop is a function of the loop index itself. I'm not sure why that is. What lead me to this conclusion is in the code below. When an arbitrary number was used in the conditional statement, the for loop broke as expected. Very interesting, as at some point in the above code, L at an iteration of the loop was equal to 6.
2. For the second item I got lucky, since all of my calculations below my previous break statement were functions of the preload variable, which is a function of the counter, I was able to accomplish what I needed by storing only the counter variable for the entire loop sequence, and then when a condition was met for the current loop, reducing that counter by an increment of one and recalculating. Not the most elegant solution in my opinion (I wish my first bit of code would have worked), but the code still runs incredibly fast.

Thank you all for your help!

%1*****troubleshooting statement*****
if Ni(6)<1.0
    break
end
%2*****updated code*****
for L=1:length1 %cycling through all available preload values until safety factors are met
                
                %stopping secondary loop if preload safety factor is too low
                %***bolts will break before Ny, Nf3, or Nsep is reached,
                %those are filtered out at the end of the code
                if Ni<1.0 && count2>2 %***if the count is too low there will be no integers for the next line
                    preload=preload_vector(count2-1); %reverses to the previous preload value that satisfied the above condition
                    %performing calculations with the acceptable preload
                    Fi=preload*ps*At; %calculating force on fastener due to preload for each grade
                    max_stress=(Pb+Fi)/At; %calculating stress due to preload and external loads for each grade
                    Ti=u*Fi*major_dia; %calculating torque to acheive preload from above
                    tau=(16*Ti)/(pi*((2*sqrt(Am/pi))^3)); %calculating shear stress from preload for each grade
                    eff=sqrt((Fi/At)^2+(3*tau^2)); %calculating effective stress from preload for each grade
                    Psep=Fi/(1-C); %load at which sep. occurs

                    %storing last value of Ni for each secondary loop iteration
                    Ni=ts/eff; %calculating preload application safety factor

                    %calculating yield safety factor
                    Ny_calc=ys/max_stress;

                    %calculating separation safety factor
                    Nsep_calc=Psep/max_load; 
                     
                    break
                end
%****************************************************************************************************************************************                 
                count2=count2+1; 
                
                preload=preload_vector(count2);
                %Fastener Calculations
                Fi=preload*ps*At; %calculating force on fastener due to preload for each grade
                max_stress=(Pb+Fi)/At; %calculating stress due to preload and external loads for each grade
                Ti=u*Fi*major_dia; %calculating torque to acheive preload from above
                tau=(16*Ti)/(pi*((2*sqrt(Am/pi))^3)); %calculating shear stress from preload for each grade
                eff=sqrt((Fi/At)^2+(3*tau^2)); %calculating effective stress from preload for each grade
                Psep=Fi/(1-C); %load at which sep. occurs
                
                %storing last value of Ni for each secondary loop iteration
                Ni=ts/eff; %calculating preload application safety factor
                
                %calculating yield safety factor
                Ny_calc=ys/max_stress;
                
                %calculating separation safety factor
                Nsep_calc=Psep/max_load;
            end
            %storing static analysis data
            Ti_stored(i)=Ti;
            preload_stored(i)=preload; %percent preload
            Ny_calc_stored(i)=Ny_calc;
            Nsep_calc_stored(i)=Nsep_calc;
            Ni_stored(i)=Ni;
        end

[Walter Roberson]

It would help if we had the complete program to test with.

>> for i = 1 : 100; if i^2 - 2*i + 5 > 17; break; end; end
>> i

i =

     5

Looks like functions work.

[JAKE SEIFERT]

I have the program working flawlessly now with the corrections that I mention above.