Scribd
Upload
105 views17 pages

SYSID Lecture 04

This document discusses discrete identification models for system identification. It begins by introducing discrete state space models, which are needed because sensor outputs and force inputs are recorded as discrete numbers in modern measurement. A simple discrete approximation is presented using zero-order hold input. The discrete forms of the internal variable, state space model, output sequence, and Markov parameters are then defined. The document concludes by introducing a state observer model that can be used to add artificial damping and deal with noise when identifying systems with little damping.

Uploaded by

sam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
105 views17 pages

SYSID Lecture 04

This document discusses discrete identification models for system identification. It begins by introducing discrete state space models, which are needed because sensor outputs and force inputs are recorded as discrete numbers in modern measurement. A simple discrete approximation is presented using zero-order hold input. The discrete forms of the internal variable, state space model, output sequence, and Markov parameters are then defined. The document concludes by introducing a state observer model that can be used to add artificial damping and deal with noise when identifying systems with little damping.

Uploaded by

sam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

OSE801

Engineering System Identification


Spring 2010

Lecture 4: Discrete Identification Models


and Markov Parameters

Instructors: K. C. Park (Division of Ocean Systems Engineering) and


Y. J. Park (Division of Mechanical Engineering)
DISCRETE STATE SPACE MODELS FOR SYSTEM IDENTIFICATION

So far we have studied several model equations for structural


system identification, which are continuous in time. In modern
measurement practice, discrete sampling has become the
standard methodology. In other words, the sensor output as well
as the force input are recorded in discrete numbers. Therefore, an
appropriate model for system identification must be cast into a
discrete form that reflects the measurement environment. The
derivation of a discrete state space model begins with the solution
of the continuous state space model which is recalled here for
convenience:
DISCRETE STATE SPACE MODELS FOR SYSTEM IDENTIFICATION
- cont’d

A simple discrete approximation of the above continuous


solution is to employ sampling with a zero-order hold. By zero-
order hold it is meant to hold the excitation input u constant
during each sampling step:
DISCRETE STATE SPACE MODELS FOR SYSTEM IDENTIFICATION
- cont’d

Discrete solution form of the internal variable (x):


DISCRETE STATE SPACE MODELS FOR SYSTEM IDENTIFICATION
- cont’d

Discrete form of the state space model:

It is noted that C and D have not been affected by the discrete form.
Why?
Output Sequence of Discrete State Space Model

Consider an initially-relaxed system, namely, the system with the


trivial initial condition (x(0) = 0). For this case the output sequence
is given by
Markov Parameters
(Discrete Impulse Response Functions(DIRFs) )

When u(0) = 1, and


u(1) = u(2) = … = u(k) =0, the excitation (input) is called
the discrete impulse excitation. The corresponding discrete
Solution sequence is thus the discrete impulse response functions,
or is referred to Markov parameters defined by:

For a general train of excitations, the response is given in terms of


the Markov parameters as
STATE OBSERVER MODEL

The identification models discussed so far in the preceding sections


are based on the physical models that are derivable from analytical
deterministic modeling process. It will be seen in the subsequent
chapters that the length of the input data used for system
identification assumes that the impulse response h(t) decays to a
sufficiently small amplitude within that time duration.

There are many instances where the system to be identified exhibits


very little damping such as composite structures and many light-
weight space structures, often contaminated with noises. For such
systems, one may inject an artificial damping by augmenting the
output into the state space model by modifying the model equation or
embed a filter to deal with noises.
STATE OBSERVER MODEL - cont’d

The state observer model equation begins with

stated as
STATE OBSERVER MODEL - cont’d
STATE OBSERVER MODEL - cont’d
TRANSFER FUNCTION FOR DISCRETE STATEE SPACE MODEL

Continuous transfer function:

Discrete transfer function:


Relation between the two-sided Laplace Transform
and z-Transform: http://en.wikipedia.org/wiki/Z-transform
Matlab Toolbox Functions
Luckily, Matlab has the conversion routine!

>> num=[1/2 1 1/2]


>> den =[1 -2 1];
>> [A, B, C, D] = tf2ss(num,den)
A =[
2 -1
1 0];
B =[
1
0];
C = [2 0]
D = 0.5000

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy