2.5. Impedance of a closed tube using the inhomogeneous wave equation

Boundary conditions

We have already found the impedance of a closed tube by taking the solution to the following Helmholtz wave equation
(2x2+k2)p˜(x)=0
image (2.79)
and applying boundary conditions to the solution. It is known as a homogeneous wave equation because there are no sound sources explicit in the equation. These are included in the boundary conditions that are applied to the solution. Here we shall consider the inhomogeneous wave equation
(2x2+k2)p˜(x)=δ(xl)xp˜(x)|x=l,
image (2.80)
which includes the sound source at x   =   l on the right-hand side, where δ is the Dirac delta function. This useful function describes a singularity when its argument is zero (in this case when x   =   l) but returns a zero value for all other arguments. In solving the equation using this alternative method, we shall introduce some useful techniques for approaching acoustical problems in general, which make use of orthogonality and the properties of the Dirac delta function. The solution itself will provide useful identities for trigonometrical functions with numerical advantages over more conventional ones. In Fig. 2.6, the piston at x   =   l oscillates with velocity u˜0 image . Hence, using the relationship of Eq. (2.4a), gives
(2x2+k2)p˜(x)=jkρ0cδ(xl)u˜0.
image (2.81)
In other words, we are describing the piston as a point source at the end of the tube.

Solution of the inhomogeneous wave equation for a closed tube

Let the solution be in the form of an eigenfunction expansion
p˜(x)=n=0A˜ncos(nπx/l).
image (2.82)
When the piston is stationary, this satisfies the boundary conditions
xp˜(x)|x=0=0,xp˜(x)|x=l=0.
image (2.83)
Inserting Eq. (2.82) in Eq. (2.81) and multiplying both sides by cos(mπx/l) while integrating over the length of the tube gives
n=0(k2n2π2l2)A˜n0lcos(mπx/l)cos(nπx/l)dx=jkρ0cu˜00lcos(mπx/l)δ(xl)dx.
image (2.84)
The two integrals have the following identities
0lcos(mπx/l)cos(nπx/l)dx={0,mnl,m=n=0l/2,m=n0
image (2.85)
0lcos(mπx/l)δ(xl)dx=cos(mπ)=(1)m.
image (2.86)
The first is the property of orthogonality and the second is a property of the Dirac delta function
F(x)δ(xl)dx=F(l).
image (2.87)
Hence, cos(mπx/l) in this case is the orthogonal (or normalizing) function and we can eliminate the summation to yield
A˜n=jklρ0cu˜0(2δ0n)(1)nk2l2n2π2,
image (2.88)
so that the solution of Eqs. (2.80) and (2.81), given by Eq. (2.82), becomes
p˜(x)=jklρ0cu˜0n=0(2δ0n)(1)ncos(nπx/l)k2l2n2π2,
image (2.89)
where δ mn is the Kronecker delta function
δ0n={0,mn1,m=n.
image (2.90)
Eq. (2.89) is equivalent to the solution of Eq. (2.79), given by Eq. (2.69)
p˜(x)=jρ0cu˜0coskxsinkl.
image (2.91)

Impedance of the closed tube

We saw previously that the impedance of a closed tube is given by
Zs=p˜(l)u˜0=jρ0ccotkl,
image (2.92)
which is equivalent to the following expression obtained using Eq. (2.89)
Zs=p˜(l)u˜0=jρ0cn=0(2δ0n)klk2l2n2π2.
image (2.93)
Because modes occur when kl   =   , the eigenfrequencies are given by
f=nc2lorl=n2λ,n=1,2,3
image (2.94)

Expansions for cot and csc

We have thus obtained a useful expansion for cot kl, which converges when truncated to a finite number of terms and retains the singularities at kl   =   , unlike the more conventional formula cot kl   =   (cos kl)/(sin kl), where
sinkl=n=0(1)n(kl)2n+1(2n+1)!,
image (2.95)
coskl=n=0(1)n(kl)2n(2n)!.
image (2.96)
Hence
cotx=n=0(2δ0n)xx2n2π2.
image (2.97)
Similarly, by comparing the pressures given by Eqs. (2.89) and (2.91) at x   =   0, we obtain
cscx=n=0(2δ0n)(1)nxx2n2π2.
image (2.98)
It is also useful to have an expression for the difference between the above
cscxcotx=tanx2=n=04x(2n+1)2π2x2.
image (2.99)
Hence
tanx=n=02x(n+1/2)2π2x2.
image (2.100)
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset