@janhardo  i have to find for all B[ij] becuase i have other function too, this is uncompleted we have to find B[12],B[13], and B[23] one by one and in other function sometime we have B[15] and so on   hence we need generalized this for finding all of them

restart;

# Kortere versie
e_ij := ((3*c1*ki*(ki - kj)*lj - c3)*li^2 - 3*(c1*kj*(ki - kj)*lj - (2/3)*c3)*lj*li - c3*lj^2) /
        ((3*c1*ki*(ki + kj)*lj - c3)*li^2 + 3*(c1*kj*(ki + kj)*lj + (2/3)*c3)*lj*li - c3*lj^2);

# Bereken direct B13 met k1 -> 0 limiet
B13_direct := limit((subs(ki=k1, kj=k3, li=l1, lj=l3, e_ij) - 1)/(k1*k3), k1=0);

# Vereenvoudig
B13_result := simplify(B13_direct);

"B_{13} = ";
B13_result;

in this part  if you can use limit for finding B12 and B23 and B13 it will be great can you i think in B12  K[1] and K[2]-->0 and in B13 K[1]-->0 and in B23 K[2]-->0

@dharr  the problem for thus ways is you can do a lot thing which not done before the idea is remains and open you can do anything and use any series combined with another ode which have a lot of solution and constract the ode and pde solution, i  combined the G'/G method with solution of herota bilinear method  and it give me answer of pdes so there is a lot idea is still open  and really if intrested in such topic i can put my 100 on this work with you in future and i am sure with your idea we can construct a lot new methods .  i will put a lot question which really are intresting after new year and really i have to construct a new method  and the graph of solution must be different, i hope  or i wish have a meeting with you and negotation about that topic

@sand15 thanks for your help and explaining that points, thanks a lot.

@dharr  your idea about hundling thus equation are emazing i liked a lot, as i mention before is a little bit long for explanation but i will try, the procces of solving pdes is like that i have transform the pde for two part real and imaginary then i take real part of pdes as my ode and i will use a methods for solving the ode then this solution of my ode will be solution of my pdes automatically becuase of that i use thus analyticall methods and there is a lot of them which i can't do it yet and becuase construct paper i don't have much time, there is some other way for solvng pdes which i don't have idea but the visualization is great i don't know where to start , like lax pair, lie symmetry  and darboux transformation thus are my future work and keep continuos on analytical methods  when something new come up i will try to solve, and really sorry for that if the questions are a little bit similar, in my mind still i am looking for long wave limit too, thank you a lot and any time have any question i will response and is my honor 

@janhardo don't have link yet

@dharr the whole idea is a little bit long becuase of that i didn't mention all, just the place i stuck, @sand15  my questions are different a lot each time i stuck in a place and i figure out i need your help , and i am really true about that without Dr.David and your help i was not success at all, beside this i am work on unpublished work now becuase of that i need to figure out and construct the idea and i am not sure i am stuck at coding a lot but in idea of undrestanding math i don't have issue at all,  i will upload the picture of one of my article but not published yet and i need to find the solution of all of them and each time use a different ode for getting my solution but the solution are different as clear in picture,

about the parameter at least one of the (F) function must not be zero, and the parameter in orginal equation must not be zero , in here again  i used your code for sech&csch is work but when i want to used for sinh and cosh is not worked and there is one of derivative remain  

Dr.D-thanks.mw

@acer  This version works and is also faster. However, I have another problem the code is too long. I wish there were a much shorter version so I could easily replace my function, and the plotting would be even better.

I also used a 3D plot for that function, but the result is not very visually appealing. If possible, I would like to generate a better-designed 3D plot, as it would be very helpful for my future article. Right now, I am using the following approach.

@sand15  when i run it it take a lot of time visualization is perfect more than matlab even, How about the 3D we can have something like that , but be a little fast work too is will be better

@janhardo  i did this but result is different

@janhardo  i need just one step no more

@janhardo  i need evaluate |u|^2 but by code what code we have to use for get the same result?

@sand15 something wrong in my changing pdes which is related to complex number and i don't have idea about that kind of substitution which work automatically 

@dharr  i try another method but you did the same as paper did , i am looking for a methods for finding this kind of pdes i will try more and more  thanks a lot 

@janhardo i need solving not writing