I have an integral that cannot be evaluated symbolically. When I try to do it numerically I want to keep some symbols in the result (to have an output like 10*x1+0.5*x3 where x1 and x3 are symbols) because in a further step I want to build a system of two equations that will solve for x1 and x3. How should I do it? Thanks
for i=1:length(A(1,:))
x3=sym('x3(t)'); x1=sym('x1(t)'); x1_diff=diff(diff(x1,t),t); r=D_o-thickness; fI1=symfun((x3*sqrt(r^2+(z-zG)^2))/(- (3*x3^5)/40 - x3^3/6 - x3 + pi/2+atan(r/(z-zG))),[t z]); fI1_diff=diff(diff(fI1,t),t); fI1_diff_z=@(z)(fI1_diff); % balast maybe need to revisit if integration doesn't work --> add x1
% and x3 as variables besides z
fI1_t=@(z) (density2.*A(2,i)*(x1_diff-fI1_diff)); % steel cyl
fI1_t2=@(z) (ro_steel.*A(1,i)*(x1_diff-fI1_diff)); % deck
fI1_t3=@(z) (ro_steel.*(A(1,i)+A(2,i))*(x1_diff-fI1_diff)); % added mass
fI1_t4=@(z)(density.*(A(1,i)+A(2,i))*(x1_diff-fI1_diff)); % keep x1 and x3 intact
FI1(1,i)=integral(fI1_t,-hw,hb-hw)+integral(fI1_t2,-hw,hc-hw)+integral(fI1_t3,hc-hw-hd/2,hc-hw+hd/2)+integral(fI1_t4,-hw,0);
Best Answer