For looping/if statement issue (simple hopefully)

I'm trying to obtain a temperature based on the given polynomial profiles over the length z. There are two conditions, if z is less than or equal to 0.32 use Ts1 and Tg1, and if z is greater than 0.32 use Ts2 and Tg2. I know what the final values should be because I generated these equations from a plot. So, I've tried the loop using only a range that suits Ts1 and Tg1 conditon and it worked. Then a tried a range that suits Ts2 and Tg2 condition and it worked. But when I try the whole range, it only uses Ts2 ad Tg2 equations, it is skiping the first if statement condition.

z=linspace(0,0.407);

i = 100;

for c = i

%explicit equations

Ts1 = ((-3508440.83).*(z.^5))+((3221859.34).*(z.^4))-((1089760.82).*(z.^3))+((165361.13).*(z.^2))-((12015.96).*z)+(1179.21);

Ts2 = ((-1906981.2).*(z.^3))+((2073664.15).*(z.^2))-((754331.65).*(z))+(92096.47);

Tg1 = ((2796881.8).*(z.^5))-((2553240.18).*(z.^4))+((852324.51).*(z.^3))-((124780.34).*(z.^2))+((6092.92).*(z))+(889.33);

Tg2 = ((-1567143.11).*(z.^3))+((1751141.7).*(z.^2))-((653704.46).*(z))+(81812.04);

%Temperature Loop

if z(:,c)>=0 && z(:,c)<=0.32

Ts = Ts1;

Tg = Tg1;

elseif z(:,c)>0.32 && z(:,c)<=0.414

Ts = Ts2;

Tg = Tg2;

end

Tg

Ts

end

APPROXIMATE FINAL VALUES FOR REFERENCE

Length [m] Temperature Solid [C] Temperature Gas [C]

0 1200 890

0.007 1085.714 920

0.017 1000 965

0.02857 950 985.71

0.05 890.5714 970

0.1 828.857 870

0.15 772 800

0.2 714 742.8

0.25 642 685.7

0.3 600 620

0.32 572 600

0.335 400 400

0.35 357.14 360

0.3893 200 257.14

0.4 85.714 200

0.407 28 185

Answer 1

[Voss]

There are a few problems.

Problem 1:

for c = i

tells the loop to iterate once, with the value of c being i. I imagine you meant:

for c = 1:i

which tells the loop to iterate i times, with the c taking values starting with 1 and going to i.

Problem 2:

if ...
    Ts = Ts1;
    Tg = Tg1;
elseif ...
    Ts = Ts2;
    Tg = Tg2;
end

This overwrites Ts and Tg on each iteration, so after the loop, you only ever have the value from the last iteration (which are the entire Ts1/Tg1 or Ts2/Tg2 vectors).

Instead, take only the appropriate single element each iteration:

if ...
    Ts(c) = Ts1(c);
    Tg(c) = Tg1(c);
elseif ...
    Ts(c) = Ts2(c);
    Tg(c) = Tg2(c);
end

Problem 3:

if z(:,c)>=0 && z(:,c)<=0.32
    % ...
elseif z(:,c)>0.32 && z(:,c)<=0.414
    % ...
end

What happens if z(:,c)<0 or z(:,c)>0.414? Then neither of those conditions is true, so Ts and Tg do not get assigned to, so they have their value left over from the previous iteration, or they are undefined if this happens on the first iteration. As you said, there are two conditions ("z is less than or equal to 0.32 ... z is greater than 0.32"), so there is only one check necessary:

if z(:,c)<=0.32
    % ...
else
    % ...
end

Inefficiency (but not a problem in that it doesn't impact the result): You can calculate Ts1, Ts2, Tg1, Tg2 one time before the loop, rather than on each iteration of the loop, because these vectors contain the values for all elements of z, so they don't depend on the loop iterator variable c.

However, all that being said, you can replace your loop approach with a simple and efficient logical indexing approach:

z=linspace(0,0.407);

%explicit equations
Ts1 = ((-3508440.83).*(z.^5))+((3221859.34).*(z.^4))-((1089760.82).*(z.^3))+((165361.13).*(z.^2))-((12015.96).*z)+(1179.21);
Ts2 = ((-1906981.2).*(z.^3))+((2073664.15).*(z.^2))-((754331.65).*(z))+(92096.47);
Tg1 = ((2796881.8).*(z.^5))-((2553240.18).*(z.^4))+((852324.51).*(z.^3))-((124780.34).*(z.^2))+((6092.92).*(z))+(889.33);
Tg2 = ((-1567143.11).*(z.^3))+((1751141.7).*(z.^2))-((653704.46).*(z))+(81812.04);

Tg = NaN(size(Tg1)); % initialize Tg and Ts
Ts = NaN(size(Ts1));

idx = z <= 0.32; % idx is a vector of logicals (true/false) saying whether each element of z is <= 0.32

Tg(idx) = Tg1(idx);   % take Tg1 when z <= 0.32
Ts(idx) = Ts1(idx);   % take Ts1 when z <= 0.32
Tg(~idx) = Tg2(~idx); % take Tg2 when z > 0.32
Ts(~idx) = Ts2(~idx); % take Ts2 when z > 0.32

plot(z,Tg)
hold on
plot(z,Ts)
legend({'Gas' 'Solid'})
ylabel('T')
xlabel('z')

[Jenna Broders]

Thanks so much! This was very helpful.

A further question, if each Tg and Ts values are used to find another component, lets say density for example which depends on temperature. Would the changing size of T impact the other variables that depend on T, or would they those other component just grow in size as well as the iterations proceed? In reality, determining the temperature is only a portion of the desired code, in the end i'm hoping to find the change in composition of gases along the length.

[Voss]

You're welcome!

Once you have Tg and Ts (which are each the same size as z) then you can calculate density or whetever else based on those Tg and Ts values. They have the dependency on z "built in" because they were calculated using the equations appropriate for the condition z <= 0.32.

In general, if calculating something in a loop one element at a time (which, to reiterate, is not necessary for Tg and Ts), then the array you're calculating grows as the loop iterates, if you don't pre-allocate the array to its final size before the loop:

% no pre-allocation of x
for ii = 1:5
    x(ii) = ii
end

x = 1

x = 1×2

1 2

x = 1×3

1 2 3

x = 1×4

1 2 3 4

x = 1×5

1 2 3 4 5

y = zeros(1,5); % pre-allocate y to contain 5 elements
for ii = 1:5
    y(ii) = ii
end

y = 1×5

1 0 0 0 0

y = 1×5

1 2 0 0 0

y = 1×5

1 2 3 0 0

y = 1×5

1 2 3 4 0

y = 1×5

1 2 3 4 5

[Jenna Broders]

Right makes sense! Thanks again!! This was very helpful

[Voss]

You're welcome! Any other questions, let me know. Otherwise, please "Accept This Answer". Thanks!