9
C H A P T E R 2
Compatibility
We start our discussion by first solving the NPDE
xu
x
u
2
y
D 2u; (2.1)
subject to the boundary condition
u.x; x/ D 0: (2.2)
As with most introductory courses in partial differential equations (see, for example, [37]) we
use the method of characteristics. Here, we define F as
F D xp q
2
2u: (2.3)
e characteristic equations become
x
s
D F
p
D x; (2.4a)
y
s
D F
q
D 2q; (2.4b)
u
s
D pF
p
C qF
q
D xp 2q
2
; (2.4c)
p
s
D F
x
pF
u
D p; (2.4d)
q
s
D F
y
qF
u
D 2q: (2.4e)
In order to solve the PDE (2.1) we will need to solve the system (2.4). As (2.1) has a boundary
condition (BC), we will create BCs for the system (2.4). In the .x; y/ plane, the line y D x is
the boundary where u is defined. To this, we associate a boundary in the .r; s/ plane. Given the
flexibility, we can choose s D 0 and connect the two boundaries via x D r. erefore, we have
x D r; y D r; u D 0 when s D 0: (2.5)
To determine p and q on s D 0, it is necessary to consider the initial condition u.x; x/ D 0:
Differentiating with respect to x gives
u
x
.x; x/ C u
y
.x; x/ D 0: (2.6)
From the original PDE (2.1)
xu
x
.x; x/ u
2
y
.x; x/ D 0: (2.7)
If we denote p
0
D u
x
.x; x/ and q
0
D u
y
.x; x/, then (2.6) and (2.7) become
p
0
C q
0
D 0; rp
0
q
2
0
D 0: (2.8)