Are the variables passing through ode45?

Hi, I have two arrays - k1 and k2. I am trying to solve the ode45 in the for loop by varying the k1 and k2. Since k1 and k2 are global, I did not include it as an input for the function. My question is - why are my outputs of the ode45 the same for all of the k1 and k2 I pass in?

keff = 18000;
k2RangeHop = [linspace(25,55,6).*1000 35000000];
k1RangeHop = (k2RangeHop.*keff)./(k2RangeHop-keff);
initialStates(1) = L0+l0;          % initial y
initialStates(2) = -1.2830;        % initial yDot

y1_2a = zeros(25,7);
yDot_2a = zeros(25,7);
yDDot_2a = zeros(24,7);
groundReactionForce = zeros(24,7);
y2_2a = zeros(25,7);
deltaL = zeros(25,7);

for i = 1:length(k1RangeHop)
    
    k1 = k1RangeHop(i);
    k2 = k2RangeHop(i);
    options = odeset('events', @endStance);
    [t,states] = ode45(@(t,states) springMass(t,states),timespan, initialStates,options);
    disp(k1)
    disp(states)
    y1_2a(:,i) = states(:,1);
    yDot_2a(:,i) = states(:,2);
    yDDot_2a(:,i) = diff(yDot_2a(:,i),1);
   
    groundReactionForce(:,i) = yDDot_2a(:,i);
    y2_2a(:,i) = ((-k1.*L0)+(k1.*y1_2a(:,i))+(k2.*l0))./(k2-k1);
    deltaL(:,i) = -L0 + (y1_2a(:,i)-y2_2a(:,i));
    
end

function stateDot = springMass(t, states)

global k1 k2 m g L0 l0


% Equations of motion
y_pos = states(1);
yDot = states(2);
yDDot = -g + ((1/m)*((k1*k2)/(k1+k2))*((L0+l0)-y_pos)); % YOUR CODE HERE

% Collect all states
stateDot = [yDot;yDDot]; % YOUR CODE HERE

end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%% End stance
function [value, isterminal, direction] = endStance(t, states) % end stance
% returns event function for passive walking simulation

global k1 k2 L0 l0

% The event occurs when value = 0. For this problem, the simulation should
% stop when the leg spring is extended beyond its resting length

y1 = states(1);
y2 = ((-k1*L0)+(k1*y1)+(k2*l0))/(k2+k1);
cll = y1 - y2; % current leg length
value = L0 - cll; % YOUR CODE HERE

isterminal = 1;
direction =  -1;

end

[Torsten]

My question is - why are my outputs of the ode45 the same for all of the k1 and k2 I pass in?

Did you include the global line also in the script part of your code where you define k1 and k2 ? In the above code this is not the case.

[Vanessa Chung]

Yes, I have this on the top of my code: global k1 k2 m g L0 l0

[Torsten]

Do you see the reason for identical results ?

keff = 18000;
k2RangeHop = [linspace(25,55,6).*1000 35000000];
k1RangeHop = (k2RangeHop.*keff)./(k2RangeHop-keff);
(k1RangeHop.*k2RangeHop)./(k1RangeHop+k2RangeHop)

ans = 1×7

18000 18000 18000 18000 18000 18000 18000

[Stephen23]

"Since k1 and k2 are global..."

Avoid using global variables.

"...I did not include it as an input for the function."

You should pass them as function inputs:

https://www.mathworks.com/help/matlab/math/parameterizing-functions.html

Answer 1

[Steven Lord]

I recommend you not use global variables to share data between the ODE function and the function where you call ode45 with the ODE function as input. Instead use one of the techniques shown on this documentation page. Since you're already using anonymous functions, that technique would probably be easiest to adapt your code to use.