## 45 Reputation

2 years, 293 days

## new approach I used, but recieved error...

Dear experts

I try to solve the partial integro-differential equation as following. but I recieved error. I would appreciate it if help me to solve this PIDE.

to have an idea about the equation and boundary conditions you can imagin a tank of water of length l and depth h. eta is considered as initial surface wave. then the PIDE shows the time evolution of the surface wave. (here eta and u(x,t) are the same). inface I try to use approximate solution eta for  initial condition

thank you in advance.

======================================================================================

restart:
with(plots):
with(PDEtools):

h:=1e-2:
H:=1e-3:
a:=H/2:
l:=1:
epsilon:= (a/l)^(2/3);

vel:= sqrt(9.81*h)*(1+epsilon/2-3/20*epsilon^2+3/56*epsilon^3);

X:=x-5:

eta:=h*( epsilon*(sech(X-vel*t))^2#-3/4*epsilon^2*(sech(X-vel*t))^2*(tanh(X-vel*t))^2 + epsilon^3*(5/8*(sech(X-vel*t))^2-101/80*(sech(X-vel*t))^4)*(tanh(X-vel*t))^2

eta:=h*( epsilon*(sech(X-vel*t))^2-3/4*epsilon^2*(sech(X-vel*t))^2*(tanh(X-vel*t))^2 + epsilon^3*(5/8*(sech(X-vel*t))^2-101/80*(sech(X-vel*t))^4)*(tanh(X-vel*t))^2

dx_eta:=diff(eta,x):

pdie:= 2*diff(u(x,t),t) + 3*diff(u(x,t),x) * u(x,t) +(1/3.0)*diff(u(x,t),x\$3)
-1/sqrt(Pi)* int(diff(u(x,tau),x)/sqrt(tau), tau= 0..t) = 0;
bounds:= {u(0,t)=0
, u(10,t)=0
, u(x,0)= eval(eta,t=0)
, D(u)(0,t)=eval(dx_eta,x=0)#0.2738612788e-1*sech(13.69306394+27.38612788*t)^2*tanh(13.69306394+27.38612788*t)
}:

#Generate IC for D(u)(x,0):
bounds:= bounds union{simplify@isolate}(
convert(value(eval(convert(eval(pdie, {t= 0}), Diff), bounds)), D),
D(u)(x,0)
);

#Convert original PDIE to a true PDE (i.e., without integrals):
pde:= diff(pdie, t);

#Solve PDE system:
psol:= pdsolve(pde, bounds, numeric, optimize=true):
psol:-plot3d(u(x,t), x= 0..10, t= 0..0.2, grid=[150,150],orientation=[-90,0,0],axes=boxed);

Error, (in pdsolve/numeric/plot3d) unable to compute solution for t>HFloat(0.0):
Newton iteration is not converging

=================================================================================

## range of K...

as you can see for K=1 the real and complex solutions exist. so K=1 is not interested. I look for (if it is possible) K=K1..K2 where all results (one or more) are imaginary for example. Or K=K3..K4 where all results (one or more) are complex, etc

## @acer  for example assume that Q a...

for example assume that Q and K are real. So Q:=alpha and K=k. and s:=7.44. I attached what I expect to see. I want to have plots like these.  Dear acer

Thnaks for your reply. I think 2D plots are enough to study the relationships. in the cae of (i) (p-K plot) we can fix s.

In this post ( https://www.mapleprimes.com/questions/228566-Study-Complex-Equation-And-Plot-It ) I explain how I get this equation by hand. I am also interested in doing this process with maple. could you please take a look at this and help me?

## is there any solution...

Dear users

is there any idea to solve the issue

## simple form of equation...

this equation can be rewrite like this p/(1+sK^2)=-F(K,Q)/G(K,Q)

maybe it is helpful

## Thanks...

Thanks a lot for comments on the code.

One more question. Do you have an idea to plot eq and see the zeros of the eq, beyond what I did? and what should I do I I want to study just the real Or just the imaginary section of the K or Q. In  other words, in the case of pure imaginary K, what knid of changes should be done in the code. Is it enought to set the K=0..14*I ?

## more details...

Dear , First of all, thank you for your code. It is very interesting for me.

Could you please explain me the last two lines?

```map(`@`(evalf, subs), sols, map(`@`(abs, lhs - rhs), EQ));
fnormal(%, Digits - 4);```

## --------------------...

Note that we aim to solve the eq=0. the term cosh(Q) in the denominator doesn't have a role. I use the denominator to have a clear plot.

could you please explain your post in detail?

 Page 1 of 1
﻿