Harmonic Balance Method
Harmonic Balance Method
Module 3 Lecture 6
Harmonic balance method is the most commonly used method to study the nonlinear vibration
problems. Here, the response of the system is assumed in terms of a Fourier series and using this
expression in the governing differential equation and separating the coefficients of the harmonic
terms one can obtain the unknown coefficients and frequency amplitude relation of the nonlinear
system. One may assume the response in the following form.
M M
x= ∑ Am cos(mωt ) + Bm sin(mωt )= ∑ Am cos(mωt + mβ 0 ) (3.6.1)
=m 0=m 0
Then substituting (3.6.1) in the governing equation and equating the coefficient of each of the
lowest M + 1 harmonics to zero, one obtains a system of M + 1 algebraic equation relating ω
and the Am . Usually these equations are solved for A0 , A2 , A3 ,....., Am and ω in terms of A1 . The
accuracy of the resulting periodic solution depends on the value of A1 and the number of
harmonics in the assumed solution. The method is illustrated using the following examples.
Example 3.6.1:
Find the expression for frequency amplitude relation for the single degree of freedom system
with both quadratic and cubic nonlinearities using harmonic balance method by taking one , two
and three terms in the expansion of the Fourier series.
x + ω02 x + α 2 x 2 + α 3 x3 =
0 (3.6.2)
Solution :
Taking only one term expansion, from equation (3.6.1) one has
=x A1 cos(ωt +=
β 0 ) A1 cos φ (3.6.3)
3
ω=
2
ω02 + α 3 A12 (3.6.5)
4
which for small A1 becomes
(3.6.6)
Comparing (3.6.6) with (3.2.14) we conclude that only part of the nonlinear correction to the
frequency has been obtained.
Now taking two terms and following Nayfeh and Mook (1979) by putting
=x A0 + A1 cos φ (3.6.7 )
in (3.6.2) one obtains
2 2 3
ω0 A0 + α 2 A0 + 2 α 2 A1 + α 3 A0 + 2 α 3 A0 A1 + − (ω − ω0 ) A1 + 2α 2 A0 A1 + 3α 3 A0 A1 + 4 α 3 A1 cos φ
2 1 2 3 3 2 2 2 3
1 3 1
+ α 2 A12 + α 3 A0 A12 cos 2φ + α 3 A13 cos 3φ = 0 (3.6.8)
2 2 4
Equating the constant term (terms with magenta colour) and the coefficient of cos φ (terms with
blue colour) to zero, one obtains the following equations.
1 3
ω02 A0 + α 2 A02 + α 2 A12 + α 3 A03 + α 3 A0 A12 = 0 (3.6.9)
2 2
− (ω 2 − ω02 ) + 2α 2 A0 + 3α 3 A02 + α 3 A12 =
3
0 (3.6.10)
4
When A1 is small, neglecting terms containing A02 , A12 , A03 , from Eqs. (3.6.9) and (3.6.10) one
can write
1 α2 2
A0 = − A + O( A14 ) (3.6.11)
2 ω0
2 1
3 α 22 2
ω2 =
ω02 + α 3 − A (3.6.12)
4 ω02 1
Hence,
3α 3ω02 − 4α 22 2
= ω ω0 1 + A1 (3.6.13)
8ω04
It may be noted that this expression for frequency is not same as that we obtained by using
method of multiple scales or L-P method. Hence to obtain a consistent solution by using the
method of harmonic balance, one need either to know about the solution a priori or one has to
take many terms in the Fourier series and make a convergence analysis. Otherwise one might
obtain an inaccurate approximation.
Using two harmonic terms
x= A0 + A1 cos φ + A2 cos 2φ (3.6.14)
where φ= ωt + β 0 and . Substituting Eq. (3.6.14) in Eq. (3.6.2) and equating the
coefficient of the constant part, coefficient of cos φ and cos 2φ equal to zero, one obtains the
following equations.
Constant terms
1 1 3 3 3
ω02 A0 + α 2 A02 + A12 + A22 + α 3 A03 + A0 A12 + A0 A22 + A12 A2 =
0 (3.6.15)
2 2 2 2 4
Coefficient of cos φ
(ω − ω 2 ) + 2α 2 A0 A1 + α 2 A1 A2 + 3α 3 A02 A1 + α 3 A13 + 3α 3 A0 A1 A2 + α 3 A1 A22 =
2
0
3
4
3
2
0 (3.6.16)
Coefficient of cos 2φ
Assuming A1 to be small, one can observe from Eqs. (3.6.15-3.6.17) that A0 and A2 are of the
order of A12 . So neglecting the terms of O ( A14 ) and higher order terms one can write A0 in terms
of A1 form Eq. (3.6.15) as follows.
1 α2 2
A0 = − A1 + O ( A14 ) (3.6.18)
2 ω02
Now multiplying 4A2 in Eq.(3.6.16) and subtracting it from Eq.(3.6.17) one obtains the
following equation.
4 (ω02 − ω 2 ) A2 + 8α 2 A0 A2 + 4α 2 A22 A2 + 12α 3 A02 A2 + 3α 3 A12 A2 + 12α 3 A0 A22 + 6α 3 A23 + ( −ω02 + 4ω 2 ) A2
1 3 3
−α 2 A12 + 2 A0 A2 − α 3 A12 ( A0 + A2 ) + 3 A02 A2 + A23 =
0
2 2 4
(3.6.19)
1
or , 3ω02 A2 − α 2 A12 = 0
2
α 2 A12 + O ( A14 )
1
=
or , A2
6ω02
Substituting the expressions for A0 and A2 in Eq. (3.6.16) one obtains
3 3
ω2 =ω02 + 2α 2 A0 + α 2 A2 + 3α 3 A02 + α 3 A12 + 3α 3 A0 A2 + α 3 A22
4 2
1 α2 2 1 α2 2 3
Or, ω 2 = ω02 + 2α 2 − A + α2
2 1
A + α 3 A12 + O ( A14 )
2 1
2 ω 0 6 ω 0 4
5 α2 3
Or, ω 2 = ω02 + − + α 3 A12 + O ( A14 )
6 α1 = ω0 4
2
(3.6.20)
18α 3ω02 − 20α 2 2
Or, ω 2 =
ω02 + A1 + O ( A1 )
4
24 ω 2
0
1/2
18α 3ω02 − 20α 2 2
Or, ω = ω0 1 + A1
24ω04
9α ω 2 − 10α
=
Or, ω ω0 1 + 3 0 4 2 A12
24ω0
By substituting A1 = ε a , this expression is same as that obtained by applying method of multiple
scales and Lindstedt Poincare’ technique.
Now substituting the expression of A0 and A2 in Eq. (3.6.14) one obtains
1 α2 2 1
x=A1 cos φ − A 1 − cos 2φ
2 1
(3.6.21)
2 ω0 3
Though the harmonic balance method is the most commonly used method for analyzing the
nonlinear structural vibration, it has several disadvantages. First the formulation is very tedious
not only for a multi degree of freedom nonlinear system but also with higher harmonic terms
taken into account. Second, to obtain a consistent solution, one needs to know a priori which
harmonic terms to be included in the analysis. Third a separate analysis is required to study the
stability of the system.
Exercise problems:
Determine the frequency response of a 3-degree of freedom system given by the following
equation
x1 + ω02 x1 + c12 ( x1 − x2 ) + α ( x1 − x2 ) + 2ς ( x1 − x2 )= P cos Ωt
3