3.4 Linearization of Nonlinear State Space Models: 1 F X Op 1 F X Op 2 F U Op 1 F U Op 2

F. Haugen: Kompendium for Kyb.
2 ved Høgskolen i Oslo 21
3.4 Linearization of nonlinear state space models
The formulas for linearizing nonlinear discrete-time state space models are
presented without derivation below. They can be derived in the same way
as for linearizing nonlinear continuous-time models [1]. In the formulas
below it assumed a second order system. I guess it is clear how the
formulas can be generalized to higher orders.
Given the following discrete-time nonlinear state space model
x1 (k + 1) = f1 [x1 (k), x2 (k), u1 (k), u2 (k)]

(3.9)
x2 (k + 1) = f2 [x1 (k), x2 (k), u1 (k), u2 (k)]
y1 (k) = g1 [x1 (k), x2 (k), u1 (k), u2 (k)]

(3.10)
y2 (k) = g2 [x1 (k), x2 (k), u1 (k), u2 (k)]
where f1 and f2 are nonlinear functions. The corresponding linear model,
which defines the system’s dynamic behaviour about a specific operating
point, is
¯ ¯ ¯ ¯
∂f1 ¯ ∂f1 ¯ ∂f1 ¯ ∂f1 ¯
∆x1 (k + 1) = ∂x1 ¯ ∆x1 (k) + ∂x2 ¯ ∆x2 (k) + ∂u1 ¯ ∆u1 (k) + ∂u2 ¯ ∆u2 (k)
op op op op
¯ ¯ ¯ ¯
∂f2 ¯ ∂f2 ¯ ∂f2 ¯ ∂f2 ¯
∆x2 (k + 1) = ∂x1 ¯ ∆x1 (k) + ∂x2 ¯ ∆x2 (k) + ∂u1 ¯ ∆u1 (k) + ∂u2 ¯ ∆u2 (k)
op op op op
(3.11)
¯ ¯ ¯ ¯
∂g1 ¯ ∂g1 ¯ ∂g1 ¯ ∂g1 ¯
∆y1 (k) = ∂x1 ¯ ∆x1 (k) + ∂x2 ¯ ∆x2 (k) + ∂u1 ¯ ∆u1 (k) + ∂u2 ¯ ∆u2 (k)
op op op op
¯ ¯ ¯ ¯
∂g2 ¯ ∂g2 ¯ ∂g2 ¯ ∂g2 ¯
∆y2 (k) = ∂x1 ¯ ∆x1 (k) + ∂x2 ¯ ∆x2 (k) + ∂u1 ¯ ∆u1 (k) + ∂u2 ¯ ∆u2 (k)
op op op op
(3.12)
or
∆x(k + 1) = A∆x(k) + B∆u(k) (3.13)
∆y(k) = C∆x(k) + D∆u(k) (3.14)
where  
∆x1 (k)
 .. 
∆x(k) =  .  (3.15)
∆x2 (k)
F. Haugen: Kompendium for Kyb. 2 ved Høgskolen i Oslo 22
and similarly for ∆u(k) and ∆y(k). The system matrices are1
 ¯
∂f1 ∂f1 ¯ ¯
¯ ∂f ¯¯
 ∂x1 ∂x2
¯
A=  ¯ = (3.16)
∂f2 ∂f2 ¯ ∂xT ¯op
∂x ∂x
¯
1 2 op
 ¯
∂f1 ∂f1 ¯ ¯
∂u1 ∂u2 ¯ ∂f ¯¯
 ¯
B= ¯ = (3.17)
∂f2 ∂f2 ¯ ∂uT ¯op
∂u1 ∂u2
¯
op
 ¯
∂g1 ∂g1 ¯ ¯
∂x1 ∂x2 ¯ ∂g ¯¯
 ¯
C= ¯ = (3.18)
∂g2 ∂g2 ¯ ∂xT ¯op
∂x1 ∂x2
¯
op
 ¯
∂g1 ∂g1 ¯ ¯
∂u1 ∂u2 ¯ ∂g ¯¯
 ¯
D= ¯ = (3.19)
∂g2 ∂g2 ¯ ∂uT ¯op
∂u1 ∂u2
¯
op
In the formulas above the subindex op is for operating point, which is a

particular set of values of the variables. Often, the operating point is
assumed to be an equilibrium (or static) operating point, which means
that all variables have constant values there.
Example 1 Linearization
Given the following non-linear state-space model:
x1 (k + 1) = ax1 (k) + bx1 (k)x2 (k) + cx1 (k)u1 (k) (3.20)

| {z }
f1
x2 (k + 1) = dx2 (k) (3.21)
| {z }
f2
y1 (k) = x1 (k) (3.22)
| {z }
g1
The system matrices of the corresponding linear model are

 ¯  ¯
∂f1 ∂f1 ¯
∂x1 ∂x2 ¯ a + bx2 (k) + cu1 (k) bx1 (k) ¯¯
 ¯ ¯
A=  ¯ =  ¯ (3.23)
∂f2 ∂f2 ¯ ¯
∂x ∂x
¯ 0 d op
1 2 op
1
Partial derivative matrices are denoted Jacobians.
F. Haugen: Kompendium for Kyb. 2 ved Høgskolen i Oslo 23
 ¯  ¯
∂f1 ¯ ¯
∂u1 ¯ cx1 (k) ¯
 ¯ ¯
B= ¯ = ¯ (3.24)
∂f2 ¯ ¯
∂u1
¯ 0 op
op
h i¯ £ ¤
∂g1 ∂g1 ¯
C = ∂x 1 ∂x2 ¯ = 1 0 (3.25)
op
h i¯ £ ¤
∂g1 ¯
D = ∂u 1
¯ = 0 (3.26)
op
[End of Example 1]
3.5 Calculating responses in discrete-time state

space models
3.5.1 Calculating dynamic responses
Calculating responses in discrete-time state space models is quite easy.

The reason is that the model is the algorithm! For example, assume that
Euler’s forward method has been used to get the following discrete-time
state space model:
x(k) = x(k − 1) + hf (k − 1) (3.27)
This model constitutes the algorithm for calculating the response x(k).
3.5.2 Calculating static responses
The static response is the response when all input variables have constant
values and all output variables have converged to constant values. Assume
the following possibly nonlinear state space model:
x(k + 1) = f1 [x(k), u(k)] (3.28)
where f1 is a possibly nonlinear function of x and u. Let us write xs and

us for static values. Under static conditions (3.28) becomes
xs = f1 [xs , us ] (3.29)
which is an algebraic equation from which we can try to solve for unknown
variables.
If the model is linear:
x(k + 1) = Ax(k) + Bu(k) (3.30)

3.4 Linearization of Nonlinear State Space Models: 1 F X Op 1 F X Op 2 F U Op 1 F U Op 2

Uploaded by

Copyright:

Available Formats

3.4 Linearization of Nonlinear State Space Models: 1 F X Op 1 F X Op 2 F U Op 1 F U Op 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

3.4 Linearization of Nonlinear State Space Models: 1 F X Op 1 F X Op 2 F U Op 1 F U Op 2

Uploaded by

Copyright:

Available Formats

F. Haugen: Kompendium for Kyb.

2 ved Høgskolen i Oslo 21

3.4 Linearization of nonlinear state space models

Given the following discrete-time nonlinear state space model

x1 (k + 1) = f1 [x1 (k), x2 (k), u1 (k), u2 (k)]

y1 (k) = g1 [x1 (k), x2 (k), u1 (k), u2 (k)]

In the formulas above the subindex op is for operating point, which is a

Given the following non-linear state-space model:

x1 (k + 1) = ax1 (k) + bx1 (k)x2 (k) + cx1 (k)u1 (k) (3.20)

The system matrices of the corresponding linear model are

3.5 Calculating responses in discrete-time state

3.5.1 Calculating dynamic responses

Calculating responses in discrete-time state space models is quite easy.

3.5.2 Calculating static responses

x(k + 1) = f1 [x(k), u(k)] (3.28)

where f1 is a possibly nonlinear function of x and u. Let us write xs and

If the model is linear:

x(k + 1) = Ax(k) + Bu(k) (3.30)

You might also like