## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

**Describing Function analysis
**

of

nonlinear systems

Prof. Elio USAI eusai@diee.unica.it

University of Leicester

Department of Engineering

**System definition and problem statement
**

d(t) r(t)

+ _

Plant P

process

e(t)

C

controller

u(t)

A

actuator

m(t) +

+

y(t)

z(t)

F

filter

ym(t)

+

+

T

trasmitter

n(t)

Controller

r: reference variable e: error signal u: control variable m: manipulated input y: controlled variable ym: measure of the output z: feedback signal d: disturbance n: measurement noise

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

**System definition and problem statement
**

In many cases the system presents a nonlinear phenomenon which is fully characterised by its static characteristics, i.e., its dynamics can be neglected

m

Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials

+M

u

-M

Dead zone in electromechanical systems

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

**System definition and problem statement
**

In many cases the system presents a nonlinear phenomenon which is fully characterised by its static characteristics, i.e., its dynamics can be neglected

e

**Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems
**

-E

+U

e +E

-U

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

e.System definition and problem statement In many cases the system presents a nonlinear phenomenon which is fully characterised by its static characteristics. i. its dynamics can be neglected m (driven gear) Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems -U +U u (driving gear) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 ..

e. its dynamics can be neglected m Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems -M0 +M0 u Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 ..System definition and problem statement In many cases the system presents a nonlinear phenomenon which is fully characterised by its static characteristics. i.

i.System definition and problem statement In many cases the system presents a nonlinear phenomenon which is fully characterised by its static characteristics. its dynamics can be neglected m Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems -U +U u Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .e..

they can also a have a “stabilising” effect r(t) + _ e(t) C u(t) A m(t) P y(t) z(t) H P(s ) = 1 s 1 + 1 .1s ) C (s ) = 10 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 H (s ) = (1 + 0.System definition and problem statement Nonlinearities are not always a drawback.4 s + s 2 ( ) A(s ) = N 1 O sat (± 2 ) 0.

System definition and problem statement Step response 20 NOT saturated actuator Saturated actuator 15 10 System output 5 0 -5 The saturated system oscillates but does not diverge -10 -15 -20 0 2 4 6 8 10 12 14 16 18 20 time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

47 deg Phase (deg) -180 -270 System: sys Frequency (rad/sec): 1.68 Phase (deg): -227 -360 10 -1 10 0 10 1 10 2 Frequency (rad/sec) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .941 Phase (deg): -180 The NOT saturated system is not stable since its stability margin are negatives • mg = -12 dB • mϕ = .System definition and problem statement Bode Diagram of the difference transfer function 50 System: sys Frequency (rad/sec): 1.0708 Magnitude (dB) 0 System: sys Frequency (rad/sec): 0.68 Magnitude (dB): -0.94 Magnitude (dB): 12 -50 -100 -150 -90 System: sys Frequency (rad/sec): 0.

System definition and problem statement Many systems can be reduced to a simplified form in which the all linear dynamics is concentrated in a unique block and the static non linear characteristics is represented by a separate block r(t) + _ e(t) C(jω) u(t) f(u) m(t) P(jω) y(t) z(t) H(jω) r(t) C(jω) r’(t) + _ u(t) f(u) m(t) P(jω)H (jω)C (jω) w(t) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

almost static.System definition and problem statement If constant (or very slowly varying) reference signals are considered. behaviour defined by a working nominal condition and a highfrequency behaviour due to small variations around such a working point r(t)=R C(jω) r’(t) + _ u(t) f(u) m(t) P(jω)H (jω)C (jω) w(t) u (t ) = u0 + Δu (t ) r’(t)=kCR + _ u(t) f(u) m(t) G (jω) w(t) m(t ) = m0 + Δm(t ) w(t ) = w0 + Δw(t ) u0 = kC R − kG m0 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . under some conditions it is possible to separate the low-frequency.

System definition and problem statement The non linear characteristics is translated so that the origin of the new reference Cartesian system is the point (u0. m0) in the original one Δm m kC R kG m0 kC R Δu u0 − 1 kG ⎧m0 = f (u0 ) ⎪ ⎨ ⎪ u0 = kC R − kG m0 ⎩ Δm0 = f ′(Δu ) f ′(Δu ) = f (u 0 + Δu ) − m0 u Δr’(t)=0 + _ Δu(t) f’(u) Δm(t) G (jω) Δw(t) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

i.System definition and problem statement The aim of the system analysis is to define which is the steady-state behaviour of the system defined by the variations around the nominal working point Δr’(t)=0 + _ Δu(t) f’(u) Δm(t) G (jω) Δw(t) If f’(u) is a passive sector function absolute stability tools allow for sufficient conditions for global asymptotic stability of the variation system. the steady state is characterised by constant values of the system variables If the origin of the variation system is not stable does the variables diverge to infinity or some periodic motion can appear? Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .e.

5 m0 Δu 0 u0 5 11 5 m0 = 11 u0 = -0.− 11 .5 m=− 1 1 u+ 10 2 1 m 0. 16 ) 11 2 Δm 1.5 -1 -6 -4 -2 0 2 4 6 8 u Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .4s + 1 ( ) 6 f (u ) = sat (1.Limit cycles Limit cycle: a periodic oscillation around a constant working point r’=5δ−1(τ) + _ u(t) m(t) w(t) f(u) G (jω) G (s ) = 10 (s + 1) s 2 + 1.

Limit cycles Limit cycle: a periodic oscillation around a constant working point Δr’(t)=0 + _ Δu(t) f’(u) Δm(t) G (jω) Δw(t) 8 7 6 output w(t) 5 w 4 0 5 11 5 m0 = 11 u0 = 3 2 1 w0 = 0 5 10 15 20 25 30 50 = 4.545 11 0 time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Limit cycles Limit cycle can be a drawback in control systems: Instability of the equilibrium point Wear and failure in mechanical systems Loss of accuracy in regulation Parameters of the limit cicle can be used to discriminate between acceptable and dangerous oscillations oscillation frequency oscillation magnitude Electronic oscillators can be based on limit cycles Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Describing Function .Assumptions The describing Function approach to the analysis of steady-state oscillations in non linear systems is an approximate tool to estimate the limit cycle parameters. It is based on the following assumptions There is only one single nonlinear component The nonlinear component is not dynamical and time invariant The linear component has low-pass filter properties The nonlinear characteristic is symmetric with respect to the origin Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

m(t) G (jω) w(t) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .Assumptions There is only one single nonlinear component The system can be represented by a lumped parameters system with two main blocks: •The linear part •The nonlinear part r’=const.L. + _ u(t) N.Describing Function .

Describing Function .Assumptions The nonlinear component is not dynamical and time invariant The system is autonomous. All the system dynamics is concentrated in the linear part so that classical analysis tools such as Nyquist and Bode plots can be applied. m +M u m=M u u +ε -M Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

K The more the low-pass filter assumption is verified the more the estimation error affecting the limit cycle parameters is small Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .Assumptions The linear component has low-pass filter properties This is the main assumption that allows for neglecting the higher frequency harmonics that can appear when a nonlinear system is driven by a harmonic signal G ( jω ) >> G ( jnω ) n = 2. 3.Describing Function .

Assumptions The nonlinear characteristic is symmetric with respect to the origin This guarantees that the static term in the Fourier expansion of the output of the nonlinearity. subjected to an harmonic signal. and it can be relaxed Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .Describing Function . can be neglected Δr’(t)=0 + _ Δu(t) Δm(t) Δw(t) f’(u) G (jω) Such an assumption is usually taken for the sake of simplicity.

K b0 = 0 2π T 2π T t )dt = π −π 2 ∫ f (t )sin (k t )d ( 2Tπ t ) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . T is a real constant a0 ∞ y (t ) = + ∑ (ak sin (k 2 k =1 2 2 a k = ∫ f (t ) cos(k T −T 2 2 2 bk = ∫ f (t )sin (k T −T 2 T T 2π T t ) + bk cos(k t )d ( 2Tπ t ) 2π T t )) 2π T t )dt = 1 π π 1 −π 2 π ∫ f (t )cos(k 2 2 2π T k = 0.2.Fourier expansion .1.Recall Consider a periodic function y (t ) = f (t ). y (t ) = y (t − T ).

Describing Function – Harmonic balance Δr’(t)=0 + _ Δu(t) f’(u) Δm(t) G (jω) Δw(t) Δu (t ) = U sin (ω t ) Δm(t ) = ∑ (a k cos(kω t ) + bk sin (kω t )) k =1 ∞ Δw(t ) = ∑ Gk (ak cos(kω t + ϕ k ) + bk sin (kω t + ϕ k )) Gk = G ( jkω ) k =1 ∞ ϕ k = ∠G ( jkω ) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Describing Function – Harmonic balance Consider the polar representation of a complex number associated with the exponential form of harmonic signals Δu (t ) = Ue jω t ∞ 2 M k = ak + bk2 a ϑk = arctan k bk m(t ) = ∑ M k e jϑk e jkω t k =1 Taking into account the low-pass property of the linear part of the system Δw(t ) = ∑ Gk e jϕ k M k e jϑk e jkω t ≅ G1e jϕ1 M 1e jϑ1 e jω t k =1 ∞ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Δm(t) G (jω) Δw(t) A permanent oscillation in the loop appears if Δu(t)=-Δw(t) jϕ1 Ue jω t = −G1e M 1e e jϕ1 jϑ1 jω t 1 + G1e M 1 jϑ1 e =0 U Harmonic balance equation 1 + G ( jω )N (U . ω ) = 0 N (U .L. ω ) = 1 (b1 + ja1 ) U is the Describing Function of the nonlinear term Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .Describing Function – Harmonic balance Δr’(t)=0 + _ Δu(t) N.

e. ω ) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . i.Describing Function – Harmonic balance The armonic balance equation is a necessary condition for the existence of limit cycles in the nonlinear system The approximate analysis gives good estimates if the low-pass filter hypothesis is strongly verified. It is a good tools for engineers The harmonic balance equation is similar to the characteristic polynomial function. it leads to the Nyquist condition for closed-loop stability The Describing Function is a linear approximation of the static nonlinearity limited to the first harmonic In most cases the Describing Function is not a function of the frequency and this simplifies the verification of the harmonic balance equation by means of the Nyquist plot of the transfer function G ( jω ) = − 1 N (U .

8 -0.ω 0 ) -\1/N(U) U=0 -6 -8 G(jω) ω=0 -10 -1.Describing Function – Harmonic balance Nyquist Diagram 2 ω=+ ∞ 0 -2 Imaginary Axis -4 +∞← U (U0 .6 -0.2 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .2 -1 + -0.4 -0.2 0 0.4 -1.

Describing Function – Harmonic balance Nyquist and DF Diagrams 1 0.021 Freq (rad/sec): 0.1 Imaginary Axis -1 U -1.01 Imag: -0.14 Imag: -2.997 -0.56234 -3.5 ω=17.5 -3 ω ω=0.5 System: sys Real: -1.02 Freq (rad/sec): 0.7828 ω=0.5 0 ω=100 System: sys Real: -2.651 -2 ω=3.5 -8 -7 -6 -5 -4 -3 -2 -1 0 1 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .1623 -2.

Describing Function – Computation The DF computation can be performed by means of its definition Usin(ωt) Δm(t) f’(u) N (U . ω ) = 1 (b1 + ja1 ) U 2 2 a1 = ∫ Δm(t ) cos(ωt )dt T −T 2 2 2 b1 = ∫ Δm(t )sin (ωt )dt T −T 2 T T The evaluation of coefficients a1 and b1 can be performed by means of both analytical calculation and numerical integration. depending on the type of nonlinearity involved Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

5 Δ u(t) 0 1 2 3 -2 0 0.Describing Function – Computation Ideal relay 3 2 Δu Δm +M 2 1.5 3 Time 2 2 a1 = ∫ Δm(t ) cos(ωt )dt = 0 T −T 2 T T Because of the odd symmetry of the Δm(t) signal π 2 2 4 2 4M b1 = ∫ Δm(t )sin (ωt )dt = ∫ M sin (ϑ )dϑ = T −T 2 π 0 π N (U .5 Δ m(t) 0 0 T -0. ω ) = 4M Uπ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 -1 -M -2 -1 -M -3 -3 -2 -1 -1.5 +M 1 1 0.5 2 2.5 1 1.

Describing Function – Computation Ideal relay 0.4M/π) c -1/N(U) -1 -1.2 0 0.5 -3 G(jω ) -3.6 -1.2 -1 -0.5 Nyquist and DF Diagrams U 0 The negative reciprocal of the DF is the negative real axis in backward direction Imaginary Axis -0.2 Real Axis The oscillation frequency is the critical frequency ωc of the linear system and the oscillation magnitude is proportional to the relay gain M Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 -4 -1.4 -1.5 -2 ω A limit cycle can exist if the relative degree of G(jω) is greater than two -2.8 -0.6 -0.4 -0.5 (ω .

ω ) = − j Uπ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 3 Time 2 2 b1 = ∫ Δm(t ) cos(ωt )dt = 0 T −T 2 T T Because of the even symmetry of the Δm(t) signal π 2 2 2 2 4M a1 = ∫ Δm(t ) cos(ωt )dt = − ∫ M cos(ϑ )dϑ = − T −T 2 π 0 π 4M N (U .Describing Function – Computation Pure hysteresis 2 2 1.5 -1 -1 -M -1.5 -1.5 2 2.5 -2 Δu 0 0.5 0 0.5 Δu Δm 1 1 0.5 2 2.5 -2 -2.5 Δm 0 0 -0.5 M 1.5 1 1.5 -1 -0.5 1 1.5 0.5 -2 -1.5 -0.

8 1 1.4 -0.Describing Function – Computation Pure hysteresis The negative reciprocal of the DF is the negative imaginary axis in backward direction A limit cycle can exist if the relative degree of G(jω) is greater than one and G(jω) is a type-0 system Nyquist and DF Diagrams 0.2 0.2 -0.6 -0.8 G(jω ) ω -1 -1.4 Real Axis The oscillation frequency is lower than the critical frequency ωc of the linear system and the oscillation’s magnitude is proportional to the relay gain M and to the modulus of the transfer function at phase -π/4 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .2 1.2 0 -0.4 U -1.4 Imaginary Axis -0.2 -1/N(U) -1.6 -0.6 0.4 0.2 0 0.

5 Δu 0 0.5 2 2.5 2 2.5 -2 -1.5 0 0.5 -2 -2.5 1 1 0.5 Δu Δm 1. ω ) = ⎨ 2 4M 4 Mβ ⎛β ⎞ ⎪ 1− ⎜ ⎟ − j ⎪ πU πU 2 ⎝U ⎠ ⎩ U ≤β U >β Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 1 1.5 Δm -β 0 β 0 t -0.5 +M 2 1.5 -1.5 -M -2 -2.5 γ -0.5 3 Time If U≤β the hysteretic relay behaves as pure hysteresis 8 8 4 4M ⎛β ⎞ b1 = − M ∫ sin (ωt )dt + M ∫ sin (ωt )dt = 1− ⎜ ⎟ T 0 T tγ π ⎝U ⎠ tγ T 2 8 8 a1 = − M ∫ cos(ωt )dt + M ∫ cos(ωt )dt = T 0 T tγ tγ T 4 The imaginary part of N(U) is proportional to the hysteresis area ⎧ 4M ⎪− j Uπ ⎪ N (U .5 1 1.Describing Function – Computation Hysteretic relay 2 2.5 0.5 -1 -1 -1.5 -1 -0.

8 -0.5 -3 G(jω ) -3.2 -4 Real Axis The oscillation frequency is lower than the critical frequency ωc of the linear system and the oscillation’s magnitude is proportional to the relay gain M Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .2 0 0.4 -0.5 -2 If β is larger than U the relay could not behave as a pure hysteresis -2.4 -1.5 ω -1.6 -0.Describing Function – Computation Hysteretic relay 0.2 -1 -0.5 Nyquist and DF Diagrams Imaginary Axis The negative reciprocal of the DF is the parallel to the negative real axis and with constant negative imaginary part 0 -0.5 U -1 -1/N(U) -1.

5 2 2.6 +M 0.8 -2 -1.4 -M -3 -0.5 3 Time If U≤Μ the saturation behaves as pure gain 2 2 a1 = ∫ Δm(t ) cos(ωt )dt = 0 T −T 2 T Because of the odd symmetry of the Δm(t) signal T tγ 2 M ⎞ ⎛M ⎞ 8 8 4 2kU ⎡ ⎛M ⎞ ⎤ 2 ⎢arcsin⎛ b1 = ∫ kU sin (ωt )dt + M ∫ sin (ωt )dt = ⎜ ⎟+⎜ ⎟ 1− ⎜ ⎟ ⎥ T 0 T tγ π ⎢ kU ⎠ ⎝ kU ⎠ kU ⎠ ⎥ ⎝ ⎝ ⎣ ⎦ M ⎧ U≤ ⎪k k ⎪ 2⎤ N (U ) = ⎨ ⎡ M 2k ⎛M ⎞ ⎛M ⎞ ⎛M ⎞ ⎥ ⎪ ⎢arcsin⎜ U> ⎟+⎜ ⎟ 1− ⎜ ⎟ k ⎝ kU ⎠ ⎝ kU ⎠ ⎝ kU ⎠ ⎥ ⎪π ⎢ ⎦ ⎩ ⎣ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 2 2.Describing Function – Computation Saturation 4 3 1 0.8 Δu Δm +M 0.6 -M -4 -2.5 1 1.4 2 1 k 0.5 1 1.5 -0.5 Δu -1 0 0.2 Δm 0 0 t γ -1 -0.2 -2 -0.5 0 0.5 -1 -0.

4 -1.5 G(jω ) -4 -1.4 -0.5 -3 ω -3.2 0 0.5 -2 A limit cycle can exist if the relative degree of G(jω) is greater than two and gain k is sufficiently high -2.8 -0.2 Real Axis The oscillation frequency is the critical frequency ωc of the linear system and the oscillation’s magnitude depends on the saturation parameters M and k Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .Describing Function – Computation Saturation 0.2 -1 -0.6 -1.6 -0.5 -1/k -1 -1.5 Nyquist and DF Diagrams -1/N(U) Imaginary Axis The negative reciprocal of the DF is part of the negative real axis in backward direction 0 U -0.

5 If U≤β the dead zone has no output 2 2 a1 = ∫ Δm(t ) cos(ωt )dt = 0 T −T 2 T 0 0.5 1 1.5 3 Time T Because of the odd symmetry of the Δm(t) signal 2 ⎡π ⎛β ⎞ ⎛β ⎞ ⎛β ⎞ ⎤ ⎢ − arcsin⎜ ⎟ − ⎜ ⎟ 1 − ⎜ ⎟ ⎥ ⎢2 ⎝U ⎠ ⎥ ⎝U ⎠ ⎝U ⎠ ⎣ ⎦ 8 4 2kU b1 = ∫ k (U sin (ωt ) − β )sin (ωt )dt = T tγ π β = U sin (ωtγ ) ⎧0 ⎪ 2k N (U ) = ⎨ k− ⎪ π ⎩ U ≤β ⎡ β ⎢arcsin⎛ ⎜ ⎢ ⎝U ⎣ 2 ⎛β ⎞ ⎤ ⎞ ⎛β ⎞ ⎟ + ⎜ ⎟ 1− ⎜ ⎟ ⎥ U > β ⎝U ⎠ ⎥ ⎠ ⎝U ⎠ ⎦ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 2 2.5 1 k 1 0.5 Dead zone 3 2 2 Δu Δm 1.Describing Function – Computation 2.5 -3 -2 -3 -2 -1 0 1 2 3 Δu -2.5 -1 t γ -1 -2 -1.5 Δm -β 0 0 +β -0.

5 -3 -2.5 -1/N(U) 0 -1/k Imaginary Axis U=+∞ -0.Describing Function – Computation Dead zone The negative reciprocal of the DF is part of the negative real axis in forward direction Nyquist and DF Diagrams 1 0.5 0 Real Axis The oscillation frequency is the critical frequency ωc of the linear system and the oscillation magnitude depends on the dead zone parameters β and k Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 -1 A limit cycle can exist if the relative degree of G(jω) is greater than two and gain k is sufficiently high G(jω ) -1.5 -3.5 -1 -0.5 -2 -1.5 ω -2 -2.

Describing Function – Computation 3 Dead zone 2 k The nonlinear characteristics of the Dead Zone can be computed by subtracting the Saturation characteristics from a linear one 1 saturation φ (Δ u) Δm 0 dead zone ψ (Δ u) -1 Ψ (Δu ) = k − Φ (Δu ) -2 -3 -3 -2 -1 0 1 2 3 Δu N Ψ (U ) = k − N Φ (U ) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

ω) N2(U.Describing Function – Computation The Describing function of a nonlinear characteristics can be computed as the combination of the Describing Functions of the elementary constituting nonlinear characteristics Nt(U.ω) Δm(t) G (jω) Δw(t) N t (U ) = N1 (U ) + N 2 (U ) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .ω) Δr’(t)=0 + Δu(t) _ N1(U.

Describing Function – Computation A number of Describing Function can be computed by particularisation of the function Φ (α ) = [arcsin(α ) + α π 2 1−α 2 ] in which the parameter α defines a peculiar point of the nonlinear characteristics m +M k -α +α u/U α ≥1 ⎧k ⎪ N (α ) = ⎨ ⎪k ⋅ Φ (α ) α < 1 ⎩ -M Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Describing Function – Computation +M k2 k1 u/U +α m -M -α α ≥1 ⎧k1 ⎪ N (α ) = ⎨ ⎪k 2 + (k1 − k 2 ) ⋅ Φ (α ) α < 1 ⎩ m k -α +α u/U α ≥1 ⎧0 ⎪ N (α ) = ⎨ ⎪k (1 − Φ (α )) α < 1 ⎩ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Describing Function – Computation m +M k2 -β -α +α k1 +β u/U α ≥1 ⎧0 ⎪ ⎪ N (α ) = ⎨k ⋅ (1 − Φ (α )) α <1≤ β ⎪ ⎪k ⋅ (Φ (β ) − Φ (α )) β < 1 ⎩ -M m +M k u/U -M N (α ) = k + 4M π Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Describing Function – Computation m +M u/U N (α ) = 4M π -M m +M −α +α -M u/U ⎧0 ⎪ N (α ) = ⎨ 4 M 1−α 2 ⎪ π ⎩ α ≥1 α <1 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

**Describing Function – Computation
**

m k -α +α u/U

α ≥1 ⎧0 ⎪ N (α ) = ⎨ k [1 − Φ(2α − 1)] − j 4kα (1 − α ) α < 1 ⎪2 π ⎩

m k -α +α u/U

α ≥1 ⎧0 ⎪ N (α ) = ⎨ k [1 − Φ(2α − 1)] + j 4kα (1 − α ) α < 1 ⎪2 π ⎩

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

**Describing Function – Computation
**

+M m

-α +α

u/U

⎧0 ⎪ N (U , ω ) = ⎨ 4 M 4 Mα 1−α 2 − j ⎪ πU πU ⎩

α ≥1 α <1

-M +M m

-α +α

u/U

⎧0 ⎪ N (U , ω ) = ⎨ 4 M 4 Mα 2 ⎪ πU 1 − α + j πU ⎩

α ≥1 α <1

-M

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

**Describing Function – Computation
**

Nt(U,ω)

Δr’=0 + Δu _ Δm 1/jω G (jω) Δw

**Negative inverse of the Describing Function
**

0.5

1 N t (U ) = N 1 (U ) + N 1 (U ) jω

N (U , ω ) = 4M 4M −j πU πUω

Imaginary axis

U

0 ω=100 -0.5 ω=17.7828

U

ω=0.1

-1

-1.5

U

-2

U

ω=3.1623

πUω −1 (ω + j ) =− 2 N (U , ω ) 4M 1 + ω

-2.5

(

)

-3 ω=0.56234 -3.5 -8 -7 -6 -5 -4 -3 -2 -1 0 1

Real axis

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

**Stability of limit cycles
**

Δr’(t)=0 + _ Δu(t) N.L. Δm(t) G (jω) Δw(t)

If the block N.L. is a pure constant k, the stability of the feedback system can be performed by means of the Nyquist criterion, which gives the number of roots with positive real roots of the Harmonic Balance Equation

1 1 + G ( jω ) = 0 k

The Nyquist criterion looks at the relative position of the transfer function G(jω) with respect to the point (-1/k, 0) in the complex plain. By extension the criterion can be applied to any point –α of the complex plain with reference to the Harmonic Balance Equation

1 + G ( jω )α = 0 , α ∈C

Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008

5 -4 -3 -2. i.5 -1 -0.5 G ( jω ) = ω=+∞ 0 1 2 jω ( jω ) + 0.5 -3 The closed loop is stable with respect to the point -α ω=0 + 0 -2 -1. -α -2 -2.5 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 Imaginary Axis -1 -1.5 -3. the reduced Nyquist criterion can be applied. Nyquist plot of G(jω ) 1 0.5 jω + 1 ( ) -1/k -0. the closed loop stability can be stated if the reference point –α in the complex plain remains on the left-side when running along the Nyquist plot of G(jω) from ω=0+ to ω =+∞.5 The closed loop is not stable with respect to the point (-1/k.0).Stability of limit cycles If the transfer function G(jω) represents a stable system.e.

Stability of limit cycles Can the generalization of the Nyquist criterion be used to analyse the stability of a limit cycle? YES. it can What does it mean that a limit cycle is stable? If a perturbation of the magnitude of the periodic oscillation occurs. Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . How can a limit cycle be considered from the Nyquist criterion point of view? It is a marginally stable condition. How can the magnitude of a limit cycle be represented? By a point in the Describing Function plot. it tends to the original value as time passes.

5 -1 -0.5 0 -0.Stability of limit cycles Nyquist and DF plots 2 Imaginary Axis Points B.5 ω -2.5 -2 -1. B”. B’. 1.5 U B B’ B” -1 A and B represent two possible limit cycles -1.5 0 -3 -3 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 -2 G(jω ) -2.5 1 A A’ 0. A and A’ represent possible magnitudes of the oscillation.

5 1 A A’ 0.5 -1 B’ B U B” G(j ω ) -1.5 ω -3 -2.5 0 -3 Real Axis A: Unstable limit cycle The linear system G(jω) is assumed to be stable in order to apply the reduced Nyquist criterion Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .Stability of limit cycles Applying the reduced Nyquist criterion with respect to: point B’: oscillations tend to decrease (stable system) point B”: oscillations tend to increase (unstable system) point A’: oscillations tend to decrease (stable system) B: Stable limit cycle Nyquist and DF plots 2 1.5 -2 -2.5 -2 -1.5 -1 -0.5 Imaginary Axis 0 -0.

The position of the motor shaft is measured by means of a rotational variable resistance. Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .c.c.DF Analysis – Example of application Consider a DC motor with permanent magnets +24 V M 0V -24 V d. The voltage on the rotational resistance drives the position of a relay that switches the motor supply voltage between +/.24 V d.

DF Analysis – Example of application The linear approximate model of a DC motor is the following di (t ) vr (t ) = Rr ir (t ) + Lr r + e(t ) dt Rr: rotor resistance dω (t ) Lr: rotor inductance = Cem (t ) + Bω (t ) J dt ke: voltage feedback constant dϑ (t ) kt: torque constant ω= dt Jm: motor inertia e(t ) = keω (t ) Jl: load inertia Cem (t ) = kt ir (t ) J = Jm + Jl B = Bm + Bl Bm: motor friction coefficient Bl: load friction coefficient vr: rotor supply voltage ir: rotor wound current Cem: electromagnetic torque ω: rotational speed θ: shatf angular position ⎧+ 24 ϑ < ϑd ⎪ vr (t ) = ⎨ ⎪− 24 ϑ ≥ ϑd ⎩ Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

DF Analysis – Example of application The motor transfer functions are the following kt Ω(s ) Wω (s ) = = Vs (s ) (sLa + Ra )(sJ + B ) + k t k e kt Θ ( jω ) = = ℜ( jω ) + jℑ( jω ) Wϑ ( jω ) = Vr ( jω ) jω (( jωLa + Ra )( jωJ + B ) + k t k e ) =- ω 2 (Ra J + La B )2 + (Ra B + k t k e − ω 2 La J ) −j k t ( R a J + La B ) 2 ω ω 2 (Ra J + La B )2 + (Ra B + k t k e − ω 2 La J ) [ k t R a B + k t k e − ω 2 La J ( ) 2 ] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

5 1 -20 -2 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . % rotor inductance ke=0. % voltage feedback constant kt=0.001.05. -18 ω -1. % motor inertia Jl=0. B=Bm+Bl. % motor friction coefficient Bl=0. % torque constant Imaginary Axis 0 U -2 -4 -6 Jm=0.3.09 % load inertia Bm=0.05.3.4.01.5 0 0.5 -1 -0.DF Analysis – Example of application Taking into account the parameters of the motor and of the load Nyquist and DF plots R=0. % rotor resistance L=0. %load friction coefficient -8 -10 -12 -14 -16 J=Jm+Jl.

05.73e-007 Freq (rad/sec): 36.3.05. % rotor inductance ke=0.8 0 U -1 -2 -3 -4 ω -5 -20 -15 -10 -5 0 x 10 -3 Real Axis ω ℑ(W = ω cr = p ( jω ))= 0 Ra B + k t k e = 36. % motor inertia Jl=0.056 rad/s La J U=- 4M π ⋅ ℜ(W p ( jω cr )) = 0. 1 System: Wp Real: -0. % voltage feedback constant Imaginary Axis 5 4 3 2 kt=0. %load friction coefficient J=Jm+Jl.4.001. % motor friction coefficient Bl=0.1759 rad Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . % torque constant Jm=0.00568 Imag: -8. B=Bm+Bl.DF Analysis – Example of application Taking into account the parameters of the motor and of the load x 10 -4 Nyquist and DF plots R=0.3.01.09 % load inertia Bm=0. % rotor resistance L=0.

056 rad/s La J U=- 4M π ⋅ ℜ(W p ( jω cr )) = 0. % motor friction coefficient Bl=0.73e-007 Freq (rad/sec): 36. %load friction coefficient J=Jm+Jl.1759 rad Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . 1 0 U -1 -2 -3 -4 ω -5 -20 -15 -10 -5 0 x 10 -3 Real Axis ω ℑ(W = ω cr = p ( jω ))= 0 Ra B + k t k e = 36.05.4.00568 Imag: -8. % torque constant Jm=0.8 2 kt=0. % rotor resistance L=0.09 % load inertia Bm=0.3.01.001.DF Analysis – Example of application Taking into account the parameters of the motor and of the load x 10 -4 Nyquist and DF plots R=0.05. % rotor inductance ke=0. % motor inertia Jl=0. % voltage feedback constant Imaginary Axis 5 4 3 STABLE limit cycle!!!! System: Wp Real: -0.3. B=Bm+Bl.

DF Analysis – Example of application 80 The system presents a periodic steady-state oscillation 60 angle speed current 40 20 0 -20 -40 -60 -80 0 1 2 3 4 5 6 7 8 9 10 Time Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

DF Analysis – Example of application 3 The system presents a periodic steady-state oscillation Angle 2.5 2 1.5 1 0.5 0 0 1 2 3 4 5 6 7 8 9 10 Time Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

65 1.75 8.45 8.DF Analysis – Example of application Tc=2π/ωc 1.6 8.65 8.55 2U=8M|G(jωc)|/π 1.5 8.45 1.4 8.75 Angle [rad] The system presents a periodic steady-state oscillation 1.7 1.4 8.6 1.8 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .55 8.5 1.7 8.

DF Analysis – Example of application What does it happen if the same control law is applied to the speed control problem? ang 24 Step Sign Gain V w Cr Scope i DC Motor kt Wω (s ) = (sLa + Ra )(sJ + B ) + k t k e The linear plant is characterised by an all-pole transfer function with relative degree two. therefore there is no contact point between the Nyquist plot and the real negative axis. but the origin at ω=+∞ (corresponding to U=0 in the DF plot) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

e.5 0 0.5 1 1.5 3 -1 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .2 U=0 ω =+∞ ω =0 -0. i..6 The Nyquist plot of the linear system is tangent to the negative reciprocal of the DF at the origin. U=0 and ω=+∞ Maybe a sliding mode behaviour is established asymptotically!? -0.DF Analysis – Example of application Nyquist and DF plots 0 U -0.5 2 2.8 -1 -1.2 ω -0.4 Imaginary Axis -0.

1 0.4 0.7 0.3 0.6 0.5 0.9 1 Time Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .8 0. or finite time stabilisation seams to appear 5 4 3 Can it be a sliding mode behaviour? 2 1 0 0 0.2 0.DF Analysis – Example of application 7 angle speed 6 Possibly a sort of dead-bit control.

08 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .04 0.01 0.DF Analysis – Example of application 7 Speed [rad/s] By reducing the integration step of the simulation it is apparent that the stabilisation is asymptotic 6 5 4 3 Can it be a sliding mode behaviour? 2 1 0 0 0.05 0.03 0.06 0.07 0.02 0.

08 0.01 0.07 0.1 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .09 0.05 0.DF Analysis – Example of application 60 Rotor current [A] By reducing the integration step of the simulation it is apparent that the stabilisation is asymptotic 50 40 30 20 10 The rotor current tends to a constant value. an asymptotic (2nd order) sliding sliding mode appears 0 -10 -20 -30 -40 0 0. i.06 0.e.02 0.03 0.04 0.

and if the system output is steared to zero in a finite time. than the system behaves as a first order system with time constant τ=0.5 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .DF Analysis – Example of application The DC motor is a second order system whose state variables are the rotor speed and current kt ⎤ k dω (t ) B ⎡ B 0 = ω (t ) + t ir (t ) ⎢ J ⎥ ⎡ω (t )⎤ ⎡ ⎤ J J dt J =⎢ k ⎥ + ⎢ 1 ⎥ v r (t ) ⎥⎢ k dir (t ) R R 1 ⎢ ⎥ = − e ω (t ) − r ir (t ) + v r (t ) ⎢− e − r ⎥ ⎢ir (t )⎥ ⎢ L ⎥ ⎣ ⎦ ⎣ r⎦ Lr ⎥ dt Lr Lr Lr ⎢ Lr ⎣ ⎦ kt B B k t ⎤ ⎡ω (t )⎤ ⎡ s (t ) = ω (t ) + ir (t ) + 2ω (t ) = ⎢2 + ⎢ ⎥ J J J J ⎥ ⎢ir (t )⎥ ⎣ ⎦⎣ ⎦ The system transfer function has a zero in –2.

the new closed loop gain (assuming the rotor velocity as the output) will be 1/2 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .DF Analysis – Example of application A MATLAB-Simulink scheme is the following Sign ang 24 V w r=2π Relay Manual Switch Gain Cr Scope i DC Motor du/dt Derivative 2 Gain1 Take care that because of the feedback.

5 3 2.5 1 0.5 0 0 0.DF Analysis – Example of application 3. the state trajectory tends to the steady-state values sliding on a linear manifold of the state space Rotor current [A] 1.5 1 1.5 0 0 0.5 2 2.5 2.5 1 0.5 3 3.5 3 3.5 1 1.5 Time [s] 2 Apart from a very short transient.5 s Speed [rad/s] 2 1.5 2 2.5 Speed [rad/s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 It is apparent that the shaft speed tends to the steady state value π as a first order system with τ = 0.

i.DF Analysis – Example of application Nyquist and DF plots 2 1 U 0 Imaginary Axis The Nyquist plot of the linear system cross the negative reciprocal of the DF at the origin. U=0 and ω=+∞ A sliding mode behaviour is established in a finite time ω -1 -2 -3 -4 -2 -1 0 1 2 3 4 5 6 7 8 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .e..

an ideal relay.e.DF Analysis & Sliding Modes Sliding modes are characterised by infinite frequency of proper ideal switching devices Most sliding mode controllers use a sign function in the controller. which can be approximately represented by its Describing Function By the example it is apparent that a sliding mode behaviour is established in a finite time if the Nyquist plot of cross the inverse negative of the describing function at the point (U=0. i.ω=+∞) TAKE CARE: the Describing function is an approximate tool Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 ..ω=+∞) It can also be derived that a sliding mode behaviour is established asymptotically if the Nyquist plot of is tangent to the inverse negative of the describing function at the point (U=0.

DF Analysis & Sliding Modes Substitute the ideal relay with a hysteretic one with β=1 Sign ang 24 r=0 Relay Manual Switch Gain V w Scope Cr i DC Motor du/dt Derivative 2 Gain1 It is apparent that an ideal sliding mode cannot appear because of the switching delay Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

with a period Tlc≅0.5 0 -0.4001 0.4002 0.4003 0.4008 0.4006 0.4007 0.6 ms The magnitude is very small because of the low-pass filter property of the motor transfer function 20 10 0 -10 -20 0.4001 0.401 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .4004 0.4006 0.4009 0.4009 0.5 -1 0.4005 0.4008 0.4005 0.4 0.4004 0.4 0.DF Analysis & Sliding Modes Supplu voltage [V] The limit cycle is apparent.4002 0.4007 0.4003 0.401 1 x 10 -5 Speed [rad/s] 0.

DF Analysis & Sliding Modes The Nyquist plot of the linear system has no common point with the negative inverse of the Describing Function of the hysteretic relay Describing function analysis can be useful in sliding mode systems when a common point is present Nyquist and DF plots 2 1 -1/N(U) U=1 0 Imaginary Axis ω =+∞ -1 ω =0 -2 G(jω) -3 -4 -2 -1 0 1 2 3 4 5 6 7 8 Real Axis Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

Analysis of the characteristics of a chattering behaviour due to unmodelled dynamics of sensors and/or actuators What is chattering? It appears as oscillations of the system variables.DF Analysis & Sliding Modes Most effective use of the Describing Function approach to sliding modes…. whose magnitude is related to the influence of the neglected dynamics on the system bandwidth It is very close to a limit cycle Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

1 s Scope2 Sign Scope3 24 r=2*pi Relay Manual Switch Gain 1 0.1s+1 Transfer Fcn ang V w Cr Scope i DC Motor du/dt Derivative 2 Gain1 Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .DF Analysis & Sliding Modes Consider the motor drive as a hysteretic switching device plus a time constant τa=0.

0328 Freq (rad/sec): 662 Real Axis Nyquist plot parameters ωlc=662 rad/s Ulm=1.0507 Imag: -0.1833 rad/s2 Uls represents the magnitude of the oscillation of the sliding variable G(j ω) Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .DF Analysis & Sliding Modes Nyquist and DF plots 2 1 -1/N(U) 0 Imaginary Axis -1 -2 -3 G(jω) -4 -2 -1 0 1 2 3 4 5 6 7 8 -1/N(U) System: untitled1 Real: -0.

92 2.99 3 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .8544 rad/s2 2 ] 0.9 Nyquist plot parameters ωlc=662 rad/s Ulm=1.95 2.DF Analysis & Sliding Modes Steady-state regime 2 1.93 2.7 ms ωlc=646 rad/s Ulm=1.96 2.91 2.97 2.5 Sliding variable [rad/s Simulation results Tls=9.94 2.5 0 -0.5 -2 2.5 -1 -1.1833 rad/s2 1 Differences are due to: • Numerical solution of the simulation • Approximation of the DF approach 2.98 2.

5s Rotor wound current presents large variations An approximate sliding mode is established 2.5 2 2.5 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .5 3 3.5 Velocity Current 2 1.5 0 0 0.5 1 0.5 1 1.5 3 The rotor velocity tends to its steady-state value as a first order system with time constant τ=0.DF Analysis & Sliding Modes 3.

DF Analysis & Sliding Modes kt ⎤ ⎡ B 0 & ⎡ω ⎤ ⎢ J ⎥ ⎡ω ⎤ ⎡ ⎤ J ⎢ ⎥=⎢ k ⎥ ⎢ ⎥ + ⎢ 1 ⎥vr R &⎦ ⎢ir ⎥ ⎢− e − r ⎥ ⎢ir ⎥ ⎢ L ⎥ ⎣ ⎦ ⎢ r⎥ ⎣ ⎣ ⎦ Lr Lr ⎥ ⎢ ⎦ ⎣ B k t ⎤ ⎡ω ⎤ ⎡ σ = ⎢2 + ⎥⎢ ⎥ J J ⎦ ⎢ir ⎥ ⎣ ⎣ ⎦ Full system dynamics kt B⎞ ⎛ ir = σ − ⎜ 2 + ⎟ω J J⎠ ⎝ Internal reduced-order dynamics & ω = −2ω + s & ⎜ σ = ⎜ 2 + − r ⎟σ + ⎜ ⎜ 2 + ⎟⎜ r − 2 ⎟ − e t ⎟ω + t v r ⎟ L J ⎟ ⎜⎝ J Lr ⎟ J ⎠⎜ Lr JLr r ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎛ B R ⎞ ⎛⎛ B ⎞⎛ R ⎞ k k ⎞ k Input-output dynamics Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .

99 3 Time [s] Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 .074 3.0029 rad/s 3.078 3.072 3.08 3.92 2.8544 rad/s2 Δωls=0.93 2.088 3.086 Speed [rad/s] Nyquist plot parameters ωlc=662 rad/s Uls=1.076 3.9 2.94 2.97 2.082 3.1833 rad/s2 Δωls=0.95 2.91 2.084 3.07 2.7 ms ωlc=646 rad/s Uls=1.0015 ⎛ ⎛ ⎞⎞ 1 ⎜ ω lc t + arg⎜ ⎟ sin ⎜ ω (t ) = ω 0 + U ⎜ jω + 2 ⎟ ⎟ ⎟ jω lc + 2 ⎝ lc ⎠⎠ ⎝ 3.DF Analysis & Sliding Modes Internal reduced-order dynamics At steady-state & ω = −2ω + σ 1 Ω(s ) = 1 Σ(s ) s+2 σ (t ) = σ 0 + U sin (ω lc t ) 1 jω ls + 2 ≅ 0.96 2.0018 rad/s Simulation results Tls=9.98 2.

DF Analysis & Sliding Modes The Describing Function approach to the analysis of the chattering phenomenon in sliding mode control systems can be used to have an estimate of the chattering parameters. i. therefore an equivalent gain of the nonlinearity has to be considered to estimate the constant mean steady-state value of the output Describing Function analysis of nonlinear systems – Prof Elio USAI – March 2008 . frequency and magnitude The estimates are affected by an error which depends on the low-pass properties of the linear part of the plant The sliding variable must be considered as the output of the nonlinear feedback system The actual system output behaviour can be estimated by considering the reduced order dynamics In the presence of a constant reference value.. the nonlinear function can be not symmetric.e.

- Exercises in Nonlinear Control Systems
- Describing Functions
- combustionj
- Describing Function Slides
- MCT Book - Chapter 7
- lec28
- MCT Module 2
- Engine Idle Speed Control Using Actuator Saturation
- phase plane analysis
- ELECTRIC HEATING AND WELDING
- 3
- EE704 Control Systems 2 (2006 Scheme)
- Index
- Exercises 09
- Total
- Decentralized Control
- Full Spectra
- HHT Time Integrator
- Nonlinear System
- Dynam Stab
- Data 10 Machine Pai
- Position Control of Arm Mechanism Using Pid Controller
- Process control lecture 1 & 2_Dr Pratik N Sheth_BITS.pdf
- CONTROL SYSTEMS.pdf
- 제어이론04장
- 10.1.1.38.7304.psa
- 05 Nonlinear Dynamics
- ijccms
- UJ Control SemesterProject July2014 (1)
- Extention of Routh’s Stability Criterion for the Analytic Conditions of Spontaneous Self Excitation in Induction Generators

Sign up to vote on this title

UsefulNot usefulClose Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Loading