Why am i getting an unknown variable while taking the inverse Laplace of a 4th order system?

I am calculating the inverse Laplace of a 4th order transfer function with the following code:

syms s t
B=((0.8236*s^2+0.4877*s+1.315)/(s^4+1.385*s^3+1.978*s^2+1.317*s));
y=ilaplace(B,s,t)

Here is the ilaplace i get:

y =
1315/1317 - (3682969*sum((r3*exp(r3*t))/(2*(1500*r3^2 + 1385*r3 + 989)), r3 in RootOf(s3^3 + (277*s3^2)/200 + (989*s3)/500 + 1317/1000, s3)))/6585 - (1315000*sum((r3^2*exp(r3*t))/(2*(1500*r3^2 + 1385*r3 + 989)), r3 in RootOf(s3^3 + (277*s3^2)/200 + (989*s3)/500 + 1317/1000, s3)))/1317 - (19587691*sum(exp(r3*t)/(2*(1500*r3^2 + 1385*r3 + 989)), r3 in RootOf(s3^3 + (277*s3^2)/200 + (989*s3)/500 + 1317/1000, s3)))/13170

After that i am calculating integral squared error with the following code:

h=(1/s);
Y=ilaplace(h,s,t);
Ise=int((Y-y)^2,t,0,100)

and getting :

Ise =
((5260*exp(100*r3))/1734489 + (1729225*exp(200*r3))/3468978 - 193305/385442)/r3 + 400/1734489

Now, the problem is that i am getting this unknown variable 'r3' due to which i am unable to calculate a numerical value of integral squared error when i am using :ISE=double(Ise)

Can anyone please let me know how to get a numerical value of integral squared error. Thank you in advance.

[Walter Roberson]

experiment with vpa(y)

[Preeti Prajapati]

I tried vpa(y) too, but it didn't work.

[Star Strider]

Only a partial fraction decomposition will work. Since you rejected my Answer with that result and the code that generated it, I deleted it.

[Preeti Prajapati]

No, Sir, I didn't reject your answer. Else i tried what you suggested and commented on it. But now when I checked again then i didn't find your answer and my comment on it

[Preeti Prajapati]

Sir, i tried the partfrac command, but it showed an error 'Undefined function 'partfrac' for input arguments of type 'sym''. I am working in 2013a. Does it only work in 2015a version?

[Star Strider]

Yes.

I also posted the inverse Laplace transform of the partial fraction decomposition.