Unable to obtain double precision from symbolic value

Hello everyone,

MATLAB returns a symbolic value for x(1,1). How do I go about converting it into a double precision?

Thank you!

clear;

clc;

close all;

%% Initial parameters %%

radian_to_deg = 180/pi;

degree_to_radian = pi/180;

prompt = "Enter a Mach Number ranging from 1.5 to 4: ";

Me = input(prompt);

prompt_1 = "Enter the number of divisions: ";

n_div = input(prompt_1);

prompt_2 = "Enter the throat radius: ";

TR = input(prompt_2);

g = 1.4;

mach_angle= [];

theta = [];

theta_a = [];

y=[];

x=[];

v=[];

v_a=[];

%% get mach angle

ma_f = @(x) asind(1/x) ;

%x = Me;

%ma_angle = ma_f(x);

%% get theta_max

A = sqrt((g+1)/(g-1));

B = (g-1)/(g+1);

pm_f = @(x) (A*atand(sqrt(B*(x^2-1))) - atand(sqrt(x^2-1)));

theta_max = pm_f(Me)/2;

%% extraction of fractional part of theta_max

%dT = theta_max/n_div;

%theta(:,1) = (dT:dT:theta_max);

theta_i = sign(theta_max)*(abs(theta_max) - floor(abs(theta_max)));

%{

for i = 2:n_div

theta(i,1) = theta(i-1,1) + 3.0;

end

%}

%% Storing different values of theta_a

%point a

for i = 1:n_div

theta_a(i,1) = theta_i ;

theta_i = theta_i + 3.0;

end

v_a = theta_a;

%% Getting properties of a and 1

ya=1;

xa=0;

%theta_a(1,1) = theta_i(1,1);

v_a(1,1) = theta_a(1,1);

%point 1

y(1,1) = 0;

Km_1 = theta_a(1,1) + v_a(1,1);

theta(1,1)= 0.5*(Km_1);

v(1,1) = 0.5*(Km_1);

%% get mach number for v(1,1)

syms x

f = @(x) (A*atand(sqrt(B*(x^2-1))) - atand(sqrt(x^2-1))-v(1,1));

f1 = matlabFunction(diff(f(x)));

%NR Method using a while loop

change = 10;

Mguess = 1.1;

while (change>1e-6)

Mnew = Mguess - f(Mguess)/f1(Mguess);

change = abs(Mguess - Mnew);

Mguess = Mnew;

end

mach_angle(1,1) = ma_f(Mnew);

x(1,1) = (-ya/tand(theta(1,1)-mach_angle(1,1)))

Answer 1

[Aquatris]

You need the vpa function

x = 2;
a = sym(1/3);
f = a*sin(2*pi*x)

f = 

fVpa = vpa(f,10)

fVpa = 

[Bamelari Jovani]

x(1,1) = vpa(-ya/tand(theta(1,1)-mach_angle(1,1)),4);

Thank you!. It does work now. However, x(1,1) does not get rounded off to 4 siginificant digits. Any reason behind this?

[Aquatris]

Because vpa uses "at least X digits" where X is 4 in your case. So it does not round of if it can do more than X digits. If you want to round off you should use round function with the output of the vpa function.

x = 0.95472222;
x_rounded = round(x,4)

x_rounded = 0.9547

x_rounded-x

ans = -2.2220e-05

[Bamelari Jovani]

Thank you!

Answer 2

[Ayush Modi]

Hi Bamelari,

You can convert the symbolic value into double precision value using double method. Here is a sample code to demonstrate this:

syms a b x

a = sym(3);
b = sym(2);

x = a / b;
disp(x); % This will display 3/2 (symbolic form)

% Convert the symbolic result to double precision
x_double = double(x);

disp(x_double);

    1.5000

For more information on double method, refer to the following MathWorks documentation:

https://www.mathworks.com/help/matlab/ref/double.html

[Bamelari Jovani]

x(1,1) = double(-ya/tand(theta(1,1)-mach_angle(1,1)));

I did try this. MATLAB still returns it in symbolic form.

Output of the above statement is:

1346076948282197/4503599627370496

[Ayush Modi]

You need to convert the symbolic variable to double. Right now, even though you are using double, you are still assigning the value to symbolic variable.

To get the final value in double precision, modify the code as below -

x(1,1) = -ya/tand(theta(1,1)-mach_angle(1,1));
double_x = double(x(1,1));
% Note that double_x is a new variable