Вы находитесь на странице: 1из 9

Frequency-Domain Analysis Methods for Modelica Models

Frequency-Domain Analysis Methods for Modelica Models


Andreas Abel Tobias Nähring

ITI GmbH
Webergasse 1
01067 Dresden, Germany

{andreas.abel,tobias.naehring}@iti.de

Abstract However, all these tasks would typically be assigned


to different software tools, which is not really neces-
In addition to time-domain simulation methods, en- sary.
gineers from different application fields require fur- Modelica forms the ideal base also for frequency-
ther types of analysis to be performed on their sys- domain analyses, since it provides complete system
tems. In particular results from frequency domain descriptions in an analytic form. However, so far
analysis play an important role – this includes the Modelica is used almost exclusively for transient
calculation of natural frequencies and vibration time-domain simulation.
modes, but also the computation of transfer functions In this paper we will show, how Modelica models
or the simulation of steady-state behaviour. are used in order to compute frequency-domain re-
If the system equations are formulated using the sults and how these processes are integrated into the
Modelica language, there is the potential to use one Modelica simulation environment SimulationX.
and the same model for time-domain as well as fre- The paper will treat the following topics:
quency-domain computations.
• Nonlinear periodic steady-state simulation
In this paper we will show, how the different meth- and generation of spectral results based on
ods can be applied to a Modelica model, what kind harmonic balance
of prerequisites and adjustments are required in order
to perform the different types of analysis and how • Natural frequencies, vibration modes and
these methods can be seamlessly integrated into a energy distributions based on models lin-
Modelica simulation environment. earized in an operating point
• Computation of transfer functions based on
models linearized in an operating point
Keywords: Modelica, Steady State Simulation,
Transfer Function Analysis, Natural Frequency We focus on the periodic steady-state simulation
Analysis since this is the most recent innovation in
SimulationX.

1 Introduction
2 Periodic Steady-State Simulation
In many engineering disciplines frequency-domain
methods play an important role. Powertrain engi- 2.1 Application to Modelica Models
neers for instance not only exploit transient simula-
tions, but to a large extend assess the behaviour of The main area of application for the nonlinear peri-
their systems based on the natural frequencies, the odic steady-state simulation in SimulationX is the
resulting vibration models, and also in terms of vibration analysis of powertrains.
steady state results, which show vibrations under The example Modelica model in Fig. 1 is an adaption
stationary conditions resulting from the uneven and from [4] p. 246 with some added damping and cylin-
multi-order excitation of the driveline by the engine. ders including oscillating masses and driven by some
Other engineering domains and tasks also require typical combustion engine cylinder pressure.
frequency-domain approaches.

The Modelica Association 383 Modelica 2008, March 3rd − 4th , 2008
A. Abel, T. Nähring

The steady-state for a range of mean rotational the calculated harmonic torque components in the
speeds of the engine has to be computed. The oscilla- mass-damper spring are shown in Fig. 2.
tion time period is determined by the engine speed In SimulationX the initial conditions corresponding
and the periodicity of the excitation over the crank to the results of the periodic steady-state simulation
angle. can be calculated and used to initialize a successive
Engine Damper transient simulation.

4-Cylinder Engine

Fly Wheel

Gear Box and Shaft

Differential Gear, Side Shaft

Tyre, and Car Load Torque


tau

Fig. 1: Example of a Modelica model of a powertrain


analyzed with the periodic steady-state simulation Fig. 2: Spectral results for the torque of the spring in
But the method is also applicable to driven systems the engine damper; the sum curve and the amplitudes
in other physical domains. For non-linear electronic of the first harmonic components are shown, the larg-
amplifiers and filters most often the frequency or est amplitudes are labelled with the oscillation orders
amplitude of the driving generator is swept and the In this way the periodic solution can be recalculated
period is measured at its phase. Therefore, a general with a transient simulation and the steady-state re-
approach is needed. In SimulationX the following sults can easily be checked. Fig. 3 shows a very good
procedure has been implemented: The user chooses match of the steady-state simulation result with the
the varying reference quantity (e.g. mean engine transient simulation result.
speed or generator frequency) and the period vari- 200
able (e.g. crank angle or generator phase) from Mod-
elica model trees containing all variables and pa- 150
Torque of Engine Damper/Nm

rameters. For powertrains (or more general whenever 100


the reference quantity is not a parameter but the
50
mean value of a variable) the user also distinguishes
some model parameter as compensation parameter - 0
such as the load torque of the powertrain. The algo- −50
rithm adjusts the compensation parameter for the
steady-state, i.e. the mean engine torque and the load −100

torque are balanced out by the algorithm. No special −150


preparation of the Modelica model is needed to en-
−200
able the steady-state simulation. The same model 0 120 240 360 480 600 720
may be used for a simulation in time-domain too. Crank Shaft Angle/degree
During the simulation the computed spectra of the Fig. 3: Torque in the spring of the engine damper (full
Modelica variables are written to special steady-state line: steady-state simulation, dashed line: transient
protocols. Those results can be visualized in several simulation); the results are almost identical
different representations (amplitudes, phases, fluc-
tuations, spectral powers and so on). For the power-
train example from Fig. 1 some of the amplitudes of

The Modelica Association 384 Modelica 2008, March 3rd − 4th , 2008
Frequency-Domain Analysis Methods for Modelica Models

2.2 Computational Background • computation of the periodic function ~


x
which will be described in the following two sec-
In this subsection we give some insight in the com- tions.
putational background specific to the periodic
steady-state simulation. If the reader is only inter- 2.2.2 Period Vector Computation
ested in applications he may safely skip to subsec-
tion 2.3. The user selects one model variable as the period
variable (cf. section 2.1). We denote the index of that
For the periodic steady-state simulation the harmonic variable as iP . For this variable the user specifies
balance method is employed. This method gives a
the period length p . The model equations (2) are
high spectral precision of the results and prepares the
numerical base for behavioural modelling in the fre- then solved for the static case (i.e. ω = 0 ) once with
quency domain. ϕ = 0 and once with ϕ = 2π . Because of the 2π -
periodicity of ~ x the difference of these two solu-
2.2.1 System Equations tions just gives the period vector
The symbolic analysis compiles from the Modelica xP = x (2π ) − x (0 ) . (4)
model a system of equations for the stationary simu-
lation. If the simulation time appears explicitly in the At ϕ = 0 the displacement of the system (e.g. the
model equations (for instance in a driven system) it rotational position of a powertrain) is determined by
is replaced by a state x time with dx time / dt = 1 which the additional condition xi P (0) = 0 . This together
leaves us with an autonomous algebraic differential with (2) and (3) results in the overall system
f ( x (0),0, xC ) = 0; xi P (0) = 0
equation system
(5)
f ( x (t ), x& (t ), xC ) = 0 (1)
for the case ϕ = 0 which consists of n + 1 equations
where x is the R n -valued state vector with corre- for the n + 1 unknowns composed of the n
sponding time-derivative x& , and xC ∈ R is the com- states x (0) and the compensation quantity xC (e.g.
pensation parameter (see section 2.1). It is conven- the load torque of a powertrain).
ient to represent oscillations not over time but over
the phase angle ϕ := ω t for which the period length For ϕ = 2π we use the user-defined periodicity of
keeps constant at 2π independent of the period du- the state vector component iP and solve
ration ( ω is the phase velocity of the oscillation). f ( x (2π ),0, xC ) = 0; xi P (2π ) = p . (6)
Substituting the derivative w.r.t. time through the
derivative w.r.t. phase x& (t ) = ω x ' (ϕ ) in eq. (1) The condition that xC is the same in (5) and (6) of-
gives fers a possibility to check the computed solutions.
f ( x (ϕ ), ω x ' (ϕ ), xC ) = 0 .
For driven systems the equations in (5), (6) may not
(2)
be simultaneously solvable. In that case in each of
Throughout the remainder of this section we repre- these systems the static equation
sent x in dependence of the phase angle. f ( x,0, xC ) = 0;
The system is assumed to be freely displaceable in
one direction of the state space. Therefore, we chose is replaced by the condition
a combination of a 2π -periodic function ~ x and a f ( x, v, xC ) = 0; v 2
→ min
component linearly dependent on the phase angle as
a solution ansatz where v 2
denotes the Euclidian norm of v .
xPϕ ~ In practice it has proven sufficient to solve the result-
x (ϕ ) = + x (ϕ ) (3)
ing restricted minimization problems by a modified

Gauss-Newton algorithm.
for the system equation (2) with a constant vector
xP ∈ R n , called period vector in the sequel. 2.2.3 Harmonic Balance
This setup is rather general. It includes freely rotat- For the computation of the periodical part ~x in the
ing powertrains and periodically driven systems. ansatz (3) equation (2) is reformulated as the varia-
Solving (2) can now be divided into the two tasks tional equation
• computation of the period vector xP

The Modelica Association 385 Modelica 2008, March 3rd − 4th , 2008
A. Abel, T. Nähring

1 2π 2 N + 1 left-hand sides of (7) become the


∫ ψ (ϕ ) ⋅ y (ϕ ) dϕ = 0 (7)
2π 0 first 2 N + 1 Fourier coefficients fˆ ( xˆ , ω , xc )[k ] of
with ψ varying over all continuous R n -valued the left-hand side of (2), i.e. Fourier coefficients of
the time-domain residuals. Equations (7,8,9,10) to-
functions fulfilling the condition ψ (0 ) = ψ (2π ) and gether then give the harmonic balance equation sys-
with tem
⎛x ϕ
y (ϕ ) := f ⎜⎜ P + ~
⎛x
x (ϕ ), ω ⎜ P + ~
⎞ ⎞
x ' (ϕ )⎟, x C ⎟⎟ . (8) fˆ ( xˆ , ω, xC ) = 0
⎝ 2π ⎝ 2π ⎠ ⎠ xˆ i P [0] = 0 (13)
~
For a fixed phase velocity ω the solution x is only xˆ i R [0] = r
determined up to a multiple of xP and a correspond-
of n (2 N + 1) + 2 scalar equations for the
ing phase shift (e.g. for a powertrain the arbitrary
initial angular position). To formally fix the initial n (2 N + 1) unknowns in x̂ and the additional two
disposition, additionally the mean value of the period unknowns ω , xC . The fast Fourier transformation
variable is balanced to zero: (FFT) is used to approximate the Fourier-coeffi-
1 2π cients of y . Because of the nonlinearities in f the
∫ xi P (ϕ ) dϕ = 0 .
~ (9)
2π 0 spectrum of y is wider than that one of x and some
In some cases the user does not want to prescribe the oversampling is needed for the FFT to keep the alias-
phase velocity ω directly (e.g. for powertrains it is ing error low.
usual to prescribe the mean rotational speed of the For solving system (13) Newton's algorithm is ap-
engine instead). For that reason the user-chosen ref- plied. Deriving the Newton corrector equation in
erence quantity was introduced in section 2.1. Let time-domain and then transforming it into frequency-
iR be the index of the reference quantity and r the domain gives good insight into the structure of the
wanted mean value for that variable. Then instead of resulting system of equations. A first order Taylor
a direct assignment to ω the equation approximation of (2) in the current numerical ap-
1 2π proximation of (~x , ω, xC ) yields the equation
∫ xi R (ϕ ) dϕ = r
~ (10)
2π 0
f + ∂ 1 f ⋅ δ x + ∂ 2 f ⋅ (ωδ x '+ x ' δω ) +
(14)
is added to the variational system. + ∂ 3 f ⋅ δ xC = 0
Following Galerkin for the numerical treatment which determines with (3) the Newton correc-
of (7,8) the function space for ψ and ~
x is restricted tion (δ ~
x , δω, δ xC ) (note: (i) here ∂ k f stands for
to the finite-dimensional space spanned by the har-
the derivative of f w.r.t. the k th argument, and
monic orthogonal system of base functions
(ii) for clarity we have omitted the arguments
ψ [k ] := exp( j kϕ ) with k = − N , K , N . (11)
(x, ωx' , xC ) of f , (iii) x, δ x, δ ~x are functions
In the following we keep using lower indexes for the of ϕ ). The time-domain products in (14) correspond
state vector components but we use Modelica index
to frequency-domain convolutions. E.g., the FFT
notation to organize the frequency components (as
transforms ∂ 1 f ⋅ δ x into
we have already done so by defining ψ [k ] above).
Using the base (11) for the periodical part ~
((∂ )
N
x in (3)
1 f ) *δ x ∑ (∂ f ) [k − l ] δ xˆ[l ] .
^ ^
ˆ [k ] = 1 (15)
the ansatz becomes l =− N

With I : (I xˆ )[k ] := kxˆ[k ] the spectrum of the de-


N
x (ϕ ) =
~
∑ exp( j kϕ ) xˆ[k ] (12)
rivative x& can be written as ( x ' ) = j I xˆ . So, after
k =− N ^

where xˆ[k ] is the k -th frequency component of the shifting f to the right-hand side (14) is transformed
state space vector (we use a hat x̂ or ( x ) to denote
^
by the FFT into the equation
complex amplitudes). Since ~ x is real xˆ[k ] is the ((∂ 1 ) (
f ) * δ xˆ + j ω (∂ 2 f ) * (Iδ xˆ ) +
^ ^
) (16)
complex conjugate of xˆ[ − k ] . Thus, the values of x̂ + (∂ 2 f ⋅x ' ) δω + (∂ 3 f ) δ xC = − fˆ
^ ^

are determined by n (2 N + 1) real numbers. With ψ


replaced by ψ [k ] for k = − N , K , N the resulting

The Modelica Association 386 Modelica 2008, March 3rd − 4th , 2008
Frequency-Domain Analysis Methods for Modelica Models

for the unknown Newton correction (δ xˆ , δω, δ xC ) size is implemented in SimulationX. A short outline
in the frequency domain. Together with (9) and (10) of this algorithm shall conclude this subsection.
written as Only at the starting value rStart and the end
δ xˆi P [0] = 0; δ xˆi R [0] = 0 (17) value rStop of the interval for the reference quan-
this system formally determines the Newton correc- tity xiR the full system (13) is solved. At intermedi-
tion in the frequency domain completely. ate points for xiR the last equation determining the
With the number of n (2 N + 1) + 2 real unknowns value of the reference quantity is removed resulting
the system is rather large and the convolution opera- in
tor in (16) causes large fill-in of the system matrix
making direct solving infeasible in real-world appli- X F(X ) = 0
cations. Therefore, the iterative GMRES algorithm is
used instead (see e.g. [5]). This method only requires F ( X ) = F (X P ( k ) )
δX ⊥(k )
the evaluation of the left-hand side of (16) for X P( k )
known (δ xˆ , δω, δ xC ) . This also makes it possible X (k )
to replace the frequency-domain convolutions in (16) δX ||(k −1)
by the cheaper corresponding time-domain products X ( k −1) xiR
in (14) (together with the therefore needed FFT- rStart rStop
operations). GMRES only works well with an appro-
priate pre-conditioner. Thus, one must be able to Fig. 4: Curve tracing algorithm (see text for details)
roughly solve systems with the left-hand side of (16) ⎛ fˆ ( xˆ , ω , xC )⎞
fast. For this end the block-diagonal preconditioner F ( X ) = 0 with F ( X ) := ⎜⎜ ⎟
⎟ (19)
is used (see e.g. [6]). This approximates the convolu- ⎝ xˆ iP [ 0 ] ⎠
tions by only retaining the mean value component and with the unknowns collected in X := ( xˆ , ω , xC ) .
of (∂ k f )^ : Since (19) has one scalar equation less than un-
((∂1 f )^ * δ xˆ )[k ] ≈ (∂1 f )^ [0] ⋅ δ xˆ[k ] knowns it formally defines a solution curve (see also
((∂ 2 f )^ * ( Iδ xˆ ))[k ] ≈ (∂ 2 f )^ [0] ⋅ kδ xˆ[k ] (18) upper branch in Fig. 4) instead of a single point.
Given the last solution point X (k −1) on the solution
The so approximated system (16) can be solved fre-
quency-component wise. curve and the tangent direction δ X ||(k −1) of the solu-
If the dynamical system is linear then the Jacobi- tion curve in that point a prediction
ans ∂ 1 f , ∂ 2 f are constant in time and the corre- X P (k ) = X (k ) + s δ X ||(k −1)
sponding higher spectral components in the convolu- for the new solution point is computed. Thereby, the
tions (e.g. (∂ 1 f ) [k − l ] with k − l ≠ 0 in (15)) are
^
step size s is chosen in dependence of the estimated
zero. In this case ` ≈ ´ in (18) can be replaced by `=´ curvature of the solution path, the estimated distance
and the approximations are exact. For increasing of X P (k ) to the solution path, and the local conver-
nonlinearities the higher spectral components of gence behaviour of Newton's algorithm (for details
∂ 1 f , ∂ 2 f omitted in the preconditioner gain influ- see [2]). In the predicted point a new estima-
ence, the approximations become more coarse. In tion δ X ⊥(k ) for the tangent vector is computed as
general one can say that with stronger nonlinearities the solution of the system
D F (X P (k ) )δ X ⊥( k ) = 0,
the number of GMRES iterations per Newton step
and the number of Newton-iterations increase.
If the local Newton method does not converge fast (δ X ( ) )
|| k −1 T
⋅ δ X ⊥(k ) = 1.
enough then the Newton-algorithm with backward- This is not the tangent direction to the solution curve
error minimization via backtracking (see [1] and [7])
is applied. For a better numerical condition the states
but to the curve defined by F ( X ) = F X P (k ) (see ( )
Fig. 4). Nevertheless, these curves and their tangents
are automatically scaled during the computation.
are supposed to be close to each other. The Newton
In section 2.3 we will give an example of a nonlinear correction for the computation of the next solu-
system with a turning point in its frequency re-
tion X (k ) of (19) is then carried out in the affine
sponse. To make the computation of such points pos-
sible a curve tracing algorithm with variable step- plane with X P (k ) as origin and δ X ⊥(k ) as normal

The Modelica Association 387 Modelica 2008, March 3rd − 4th , 2008
A. Abel, T. Nähring

direction. The point X (k ,0 ) := X P (k ) is used as an


initial guess and the Newton corrections δ X (k ,i ) as
well as the iterated solution approximations X (k ,i )
(i = 0,1, K) are defined by the system
DF (X (k ,i ) )⋅ δ X (k ,i ) = − F (X (k ,i ) ),
(δ X ( ) )⊥k T
⋅ δ X (k ,i ) = 0, (20)
X (k ,i +1) = X (k ,i ) + δ X (k ,i ) .
As Fig. 4 suggests X ⊥(k ) is a better approximation
of the tangent to the solution curve at the new solu- Fig. 6: Frequency response with turning-point for the
tion point X (k ) than X ||(k −1) . Using X ⊥(k ) lets the angular speed of inertia1 in the nonlinear spring-mass-
Newton iterations run on nearly the shortest path to system; the sum curve and the first three harmonic
the solution curve, gives (20) a better numerical con- components are distinguishable in this diagram
dition, and avoids jumping between different solu- As an example in Fig. 7 the limit cycles from the two
tion branches at sharp turning points of the solution stable branches (lowest and highest) of the frequency
path. characteristic at 0.405 Hz are shown.

2.3 Example: Nonlinear Spring-Mass-System


with Turning-Point in Frequency Response 10

Unlike linear systems nonlinear systems may exhibit


Speed of Inertia rad/s

5
turning points in the frequency characteristic. The
curve tracing algorithm implemented in SimulationX
makes the computation of such kind of frequency 0

characteristics possible.
−5
The simple mechanical system of Fig. 5 is a torque
excited spring-mass-oscillator. The frequency of the
sinusoidal torque source is chosen as the reference −10

quantity and swept between 0.2 Hz and 0.7 Hz . 0 60 120 180 240 300 360
Since this reference quantity is a parameter and not a Phase of Oscillation / degree
variable SimulationX chooses it automatically as Fig. 7: Angular speed curves for the two possible stable
compensation parameter as well. The phase of the limit cycles of the nonlinear spring-mass-system at
sine oscillator is the period variable with period 2π . excitation frequency 0.405 Hz represented over
The quadratic term added to the spring characteristic phase.
makes the system nonlinear in such a way that it We kept this example simple to demonstrate that
shows a turning point in the frequency characteristic even very basic nonlinear systems may have fre-
(see Fig. 6). quency responses with turning-points. More compli-
sqr
cated examples can be found in [8], and [9].

2.4 Example: Active Electronic Filter


angleSensor1
torque2
sineOsc
torque1 inertia1 p SD1=8 The periodic steady state simulation is not restricted
fixed1= 0

tau
to mechanical systems. As an example the periodic
amplitude=1
J=1
d=0.1 steady state simulation is applied to a Modelica
Fig. 5: Nonlinear Spring-Mass-system model for an active electronic pass-band filter (see
Fig. 8). The reference and compensation quantity in
In the interval from 0.397 Hz to 0.426 Hz the fre-
this example is the frequency of the sinusoidal
quency characteristic is multi-valued. That corre-
source vin and its phase is the phase variable.
sponds to multiple periodic limit cycles at those exci-
tation frequencies.

The Modelica Association 388 Modelica 2008, March 3rd − 4th , 2008
Frequency-Domain Analysis Methods for Modelica Models

rc1 6
vc1

C=6.4e-9

R=250e3

rb1
c2
V R=20e3 5
q2

Collector Voltage of q1/V


rin c1 q1
4

power…

+
R=2.5e3 C=6.4e-9

-
R=1e3
3

re2
vin

ground1 ground3 ground2 ground4 1


Fig. 8: Modelica model of the active electronical filter
At resonance frequency the transistor amplifier of 0
0 60 120 180 240 300 360
the pass-band filter is overdriven which causes Phase of Oscillation / degree
nonlinear harmonic distortions. The nonlinear fre- Fig. 10: Collector voltage of q1 in the active electronic
quency response of the collector voltage of transis- filter at excitation frequency 1.15 kHz represented over
tor q1 is shown in Fig. 9. phase; full line: periodic steady state simulation,
dashed line: transient simulation;
The results are in good accordance. Nevertheless, a
slight difference of the results from the periodic
steady state simulation and the transient simulation is
visible at about 75° . The steep slopes of the spike
are somewhat smoothened by the limited number of
equidistant sample-points for the steady state simula-
tion (256 sample points per period were used).

3 Transfer Function Analysis and


Natural Frequencies
3.1 Linear System Analysis

Beside the nonlinear algorithm for the steady-state


simulation also linear frequency-domain analysis
methods are applicable to Modelica models and are
implemented in SimulationX. Those are based on the
linear system which results from the linearization of
the nonlinear system equations for the Modelica-
model in the current operating point. The operating
point may be determined by a previous transient
simulation or an equilibrium computation (in elec-
tronics also called DC-analysis). Some of the algo-
Fig. 9: Frequency response of the collector voltage of rithms may be applied to any Modelica model with-
q1 in the active electronic filter; top: sum signal and out changes by the user. This includes the computa-
first harmonic, bottom: zoomed view of the other har- tion of the eigensystems, the Campell diagram, and
monics in the resonance region where the amplifier is methods for the animation of the eigenmodes.
overdriven; the harmonics are decreasing with order,
only the 2nd and 3rd harmonic are labelled Other frequency-domain results such as the devia-
tions in mechanical quantities (vibration modes) and
In Fig. 10 the periodic steady state result and the
the distribution of vibration energies and losses re-
time domain result of this voltage over phase angle
quire special internal blocks that can be included into
for an excitation frequency of 1.15 kHz are com-
the Modelica-model. The following Modelica source
pared. At about 75° the base-emitter diode of q2 code shows how the inertia from the standard Mode-
blocks and the voltage amplification of q1 grows lica library can be supplemented with an internal
which causes the spike in the collector voltage of q1.

The Modelica Association 389 Modelica 2008, March 3rd − 4th , 2008
A. Abel, T. Nähring

energy calculation block which SimulationX uses in earized system. Any result variable of the model may
order to compute the energy distribution. be used as the system output. SimulationX has a spe-
cial class of signal inputs that may be open even for
model RotInertiaEnergyBlock the top-level model. Those inputs may be used for
the input-output-analysis. In Fig. 11 a cut-out of the
import M=Modelica.Mechanics;
powertrain from Fig. 1 is shown where a torque
extends M.Rotational.Inertia; source with such an input has been added. The input-
Mechanics.Rotation.CalcEnergyBlock eb; output behaviour is described by the frequency re-
equation sponse function and the pole-zero diagram of the
eb.dom = w; system.
eb.T = J*a;
end RotInertiaEnergyBlock;
The modification of the Type SpringDamper is
similar. For a demonstration the (rotational and
translational) masses and spring-dampers in the
powertrain from Fig. 1 have been substituted by the
modified types. The distribution of energy calculated Fig. 12: Element linSysAnaOpenInput in the example
by SimulationX for the eigenmode at 1.6664 Hz is from Fig. 1 with open input for the input-output-
shown in Fig. 12. In practical applications such rep- analysis
resentations show the engineer which masses, Fig. 13 and Fig. 14 show the pole-zero plot and the
springs, and dampers dominate the behaviour in cer- frequency characteristic, resp., for the powertrain
tain eigenmodes of the system, so he can take sys- from Fig. 1 with the torque at the first cylinder as
tematic countermeasures to avoid unwanted oscilla- input (Fig. 12) and the torque in the engine damper
tions. as output.
Up to now these blocks are not documented and only
used for the internal element libraries of Simula-
tionX. But this may change in future.

Fig. 13: Pole-zero plot of the system in Fig. 1; crosses:


poles, circles: zeros
For further analysis in external tools the linearized
system matrices may be exported in Modelica or
MATLAB syntax.

4 Conclusions and Outlook


Fig. 11: Distribution of energy for the powertrain ex-
ample from Fig. 1
Periodic steady state simulation proves useful for the
vibration analysis of nonlinear systems. SimulationX
3.2 Input-Output Analysis allows its application to Modelica models, in particu-
lar to powertrains, without the decomposition into
For the analysis of the input-output-behaviour the nonlinear exciter and linear drivetrain. Furthermore,
user must select the input and the output of the lin- the method is applicable to driven systems of other

The Modelica Association 390 Modelica 2008, March 3rd − 4th , 2008
Frequency-Domain Analysis Methods for Modelica Models

physical domains since it is purely equation-based. • Autonomous Systems: The ansatz used for the
Only very little knowledge of the system is required harmonic balance also bears the potential for the
from the user. Two mechanical examples and one simulation of autonomous systems. The required
from electronics were given in the paper. randomization of the start values for the harmonic
balance could be implemented.
• Detection of Stable/Unstable Limit Cycles: Up to
now there is no automatic discrimination of the sta-
ble and unstable branches in the nonlinear frequency
response computed via harmonic balance. This can
be implemented by an eigenvalue analysis of the
monodromy matrix of the computed limit cycles.

References
[1] J. E. Jr. Dennis and Robert B. Schnabel:
Numerical Methods for Unconstrained Opti-
mization and Nonlinear Equations. SIAM
1996.
Fig. 14: Frequency response of the system in Fig. 1; [2] E. L. Allgower and K. Georg: Numerical
top: amplitude, bottom: phase Continuation Methods: An Introduction.
Furthermore, we discussed methods for the small- Springer-Verlag, 1990.
signal analysis in the current operating point (result- [3] http://www.simulationx.com
ing from a transient or equilibrium computation). [4] H. Dresig and F. Holzweißig: Maschinendy-
Beside pole-zero plots and frequency response func- namik. 5th ed., Springer-Verlag Berlin, 2004.
tions also some remarks about the deviation- and
[5] A. Meister: Numerik linearer Gleichungssys-
energy distribution analysis for oscillation modes
teme. Vieweg-Verlag, Wiesbaden, 2005.
were given. They are especially useful for the me-
chanical engineer to detect the powertrain elements [6] Ognen J. Nastov: Methods for Circuit Analy-
which participate in selected oscillation modes. sis. PHD-theses, Massachusetts Institute of
Technology, 1999.
• Behaviour Description In Frequency Domain:
In future it is planned to include a behavioural de- [7] U. Feldmann, U. A. Wever, Q. Zheng, R.
scription in frequency domain (e.g., for modeling of Schultz, and H. Wriedt: Algorithms for Mod-
dynamic stiffness) for the periodic steady state simu- ern Circuit Simulation. AEÜ, Vol. 46 (1992),
lation as well as for the frequency response computa- No. 4.
tion, which was one main argument for the harmonic [8] A. Al-shyyab and A. Kahraman: Non-linear
balance method to be preferred over the shooting dynamic analysis of a multi-mesh gear train
method (see e.g. [11] for a short introduction and using multi-term harmonic balance method:
further references). One major reason for the fre- period-one motions. Journal of Sound and
quency domain description not yet being imple- Vibration, 284 (2005) 151-172.
mented in SimulationX is that Modelica currently [9] Wen-I Liao, Tsung-Jen Teng, and Chau-
still lacks a standardized way for computations with Shioung Yeh: A method for the response of
complex numbers (even if some steps in this direc- an elastic half-space to moving sub-Rayleigh
tion have already been taken, see e.g. [10]). point loads. Journal of Sound and Vibration
• Event Iterations: Event iterations are already em- 284 (2005) 173-188.
bedded into the harmonic balance algorithm. But [10] Peter Aronsson at al.: Meta Programming
there remains still some work for the treatment of and Function Overloading in OpenModelica.
time-discrete variables in special cases. Modelica 2003, November 3-4, 2003.
• Improved Convergence for Strongly Nonlinear [11] Kenneth S. Kundert: Introduction to RF
Systems: As long-term objective the convergence Simulation and Its Application. IEEE Journal
speed of the harmonic balance for strongly nonlinear of Solid-State Circuits, Vol. 34, No. 9, Sep-
systems can be improved by time domain precondi- tember 1999.
tioners (see [6]).

The Modelica Association 391 Modelica 2008, March 3rd − 4th , 2008

Вам также может понравиться