Scribd
Upload
1K views

Simulink Simulation

1. Simulate both models with constant, step, and sine inputs and compare the responses. 2. Compare the time-varying functions like step and sine inputs for both models when inputs are large and small deviations from steady-state. 3. Analyze any differences in the responses between the linear and nonlinear models to understand how well the linear model approximates the nonlinear model.

Uploaded by

poopoopho
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
1K views

Simulink Simulation

1. Simulate both models with constant, step, and sine inputs and compare the responses. 2. Compare the time-varying functions like step and sine inputs for both models when inputs are large and small deviations from steady-state. 3. Analyze any differences in the responses between the linear and nonlinear models to understand how well the linear model approximates the nonlinear model.

Uploaded by

poopoopho
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 15

Simulink:

ynamic Modeling

ChE 446
Simulink blocks
o Sources
• Constant (output a constant)
• Step (output a step change)
• Clock ( Outputs the time )
o Sinks
• Scope (simulation time plot of input)
• Display (real-time display of input value during
a simulation)
• To Workspace (saves the input to the matlab
workspace)
o Linear
• Sum (add or subtract inputs)
o Nonlinear
• S-Function (user defined function)
Simulink M-file S-Functions
o Primary purposes
• Simulating non-linear dynamics with MATLAB
o How they work
• Example M-file S-function script (Simulink/User-
Defined Functions/S-function/examples/M-
files/Level-1 M-files) explains the basics
• Each iteration, the S-function performs calculations
based on the value of a flag (initialize, find
derivatives, update actual values, etc.); it returns
the answer, then changes the flag for the next
iteration.
• The code is reasonably well-documented as to what
to enter where; we’ll help later.
Simulink M-file S-Functions
o Switch statements
o switch flag
case 0
Statements
case 1
statements
case 2
statements
otherwise
statements
end
Simulink M-file S-Functions
o Case 0: initialization
o First, Simulink sends flag=0, which:
o Declares things (x, t, y, u)
o Sets x to an initial value (x0)
o Returns ‘sys’, it’s standard return variable, set
equal to:
[#cont. states., #discrete states, #outputs,
#inputs, direct feedthrough?, #sample times]
o See
http://www.mathworks.com/access/helpdesk/help/too
for more information.
Simulink M-file S-Functions
o Case 1: calculate derivatives
o When Simulink sends flag=1, it expects the
function to return time derivatives
o You enter these derivatives like so:
 
dx 
= f ( x u ) ⇒ dxdt = [ f1 ( x(1) , x( 2 ) , ) f 2 ( x(1) , x( 2 ) , ) ]
dt

o Derivatives are returned (as ‘sys’) as a vector


o See
http://www.mathworks.com/access/helpdesk/help/too
for more information.
Simulink M-file S-Functions
o Case 3: calculate outputs
o When Simulink sends flag=2, it expects the
function to return outputs
o You enter these like so:
  
y = g ( x u ) ⇒ y = [ g1 ( x(1) , x( 2) , ) g 2 ( x(1) , x( 2 ) , ) ]

o Outputs are returned (as ‘sys’) as a vector or as


a scalar, depending on the number of inputs
o See
http://www.mathworks.com/access/helpdesk/help/too
for more information.
Simulink M-file S-Functions
o Cases 2,4,9: not used
o Case 2 updates discrete states, sample times
o Case 4 calculates the time that the next discrete
variable update occurs.
o Case 9 executes any statements you want to
perform at the END of the simulation. Plot the
output, for example.
o Cases 5-8 are not used; accordingly, there is an
“otherwise” statement to catch these cases
(which signal an error)
o See
http://www.mathworks.com/access/helpdesk/help/too
for more information.
Dynamic Simulation: Linear vs.
Nonlinear
o Solution to HW 2 will be posted on
website, including S-functions for
open loop and closed loop
simulation.
o In-class exercise: compare the
nonlinear model from the homework
to a linearized model
Nonlinear model

  
x1 x3
− ux +
dx
1 1+ x3 
= − ux +
2
2 x2 x3
5 + x3

dt  
 u (10 − x ) −2 x1 x3
− 8 x2 x3
 3 1+ x3 15+ 3 x3 
x1
y=
x1 + x2
x1′ = x1 − 52 x′2 = x2 − 32
x3′ = x3 − 3 y′ = y − 85
Linearized model

   5
x ′ − 5
2u
′  0 0  5
 − 52 
dx ′
32 3
   ′  3  ′
32

= ′
64 x3 − 2 u
′ = 0 0  ⋅ x +  − 2 u
15 3 15
64
dt  3  − 3 − 7
− x ′ −
 4 1 2 8 3 x ′ − 11 ′
x − 7 u ′   4 − 1 −  11
 
8

y′ = 16 x1′ − 32 x′2 = ( 16 − 32 0) ⋅ x ′
3 5 3 5

0 0 5
32 − 52 
A =  0 0 15
64

 B = − 32  C = [ 163 −5
32 0] D = [ 0]
− 34 − 1 − 
11
8
− 7 
If you build it, …
D

To Workspace

Scope 1
bioreact
1
S -Function r

0.76

Input x' = Ax + Bu
y = Cx + Du
Scope
State -Space
0.75

D_ss 0.625

r_ss
Bioreactor S-function
function [sys,x0] = bioreact(t,x,u,flag)

Kx=1.0; Ky=5.0; Yx=0.5; Yy=0.75;


muxmax=1.0; muymax=2.0; Si=10.0;
switch flag,

case 1,
D=u; X=x(1); Y=x(2); S=x(3);
mux=muxmax*S/(Kx+S);
muy=muymax*S/(Ky+S);
dxdt = [-D*X+mux*X, -D*Y+muy*Y, D*(Si-S)-mux*X/Yx-
muy*Y/Yy];
sys = dxdt;
Bioreactor S-function cont.
case 3,
X=x(1); Y=x(2); r=X/(X+Y); y = r;
sys = y;

case 0,
NumContStates = 3; NumOutputs = 1; NumInputs = 1;
sys = [NumContStates,0,NumOutputs,NumInputs,0,0];
x0 = [2.5 1.5 3.0];

case { 2, 4, 9 },
sys = [];

Otherwise
% error([’Unhandled flag = ’,num2str(flag)]);
end
Tasks
o Compare response of linear model to
non-linear model for:
o Constant input
o Step input
o Sine input
o Compare time-varying functions
(step, sine, etc.) for both large and
small deviations from steady-state

You might also like