SE513: Systems Identification

Lecture 1
Introduction to Systems
Identification
Dr. Samir Al-Amer

(c) 2007 Dr. Samir Al-Amer

Lecture Outlines
O What is Identification
O Modeling vs. Identification
O Types of models
O Flow chart for systems identification
O Problems in identification

(c) 2007 Dr. Samir Al-Amer

Systems Identification
O Identification is a process in which
experimental data is used to obtain a
mathematical model describing the system
Measured and unmeasured disturbances

u System ? y

O Identification is a process that uses u(t) and

y(t) to determine the dynamics of the system

(c) 2007 Dr. Samir Al-Amer

 Why do we need models?
O Models can be used to improve our understanding
of the process.
O Models are needed in designing better controllers

O A Model can be a part of some controllers like feed

forward controllers
O A model can be very valuable in optimizing the
operating conditions to get the best performance
of control systems

(c) 2007 Dr. Samir Al-Amer

The Four Problems of Identification

Representation Measurement

Estimation Validation
(c) 2007 Dr. Samir Al-Amer
Representation
What types of models are used to
represent the system?
Types of Models:
O Mental, intuitive, verbal
O Graphs and tables: step response, Bode plot
O Mathematical models
static/dynamic, linear/nonlinear,
lumped/distributed, continuous/discrete,
time-domain/frequency domain

(c) 2007 Dr. Samir Al-Amer

Representation
issues

Some of the issues

O How to select the model structure
O Flexibility of the model

O Complexity of the model

(c) 2007 Dr. Samir Al-Amer

Measurement

O Which physical quantities should be

measured?
O How should we measure them?

O Some of the issues:

Some variables of interest are difficult or
impossible to measure
Presence of noise in the measured data

(c) 2007 Dr. Samir Al-Amer

Estimation

OHow do we estimate the model

parameters from the measured
data?
How do we estimate the
nonparametric model from the
measured data?

(c) 2007 Dr. Samir Al-Amer

Validation

Can the model explain the

measured data?
Are the confidence limits on
the model acceptable?

(c) 2007 Dr. Samir Al-Amer

The Four Problems of Identification

Representation Measurement

Estimation Validation
(c) 2007 Dr. Samir Al-Amer
How do we obtain mathematical models?
Modeling Identification
(Theoretical) (Experimental)
O Construct a simplified version O Conduct an experiment
using idealized elements O Collect data
O Write element laws O Fit data to a model
O Write interaction laws O Verify the model
O Combine element laws and
interaction laws to obtain the
model

(c) 2007 Dr. Samir Al-Amer

 Natural Laws in Modeling
In processing systems,
Conservation of energy
Conservation of mass
Conservation of individual components
Conservation of momentum
are useful in Obtaining mathematical models

(c) 2007 Dr. Samir Al-Amer

 Conservation Principle
over any time period,
the rate of accumulation of S within the system
= flow rate of S in the system
flow rate of S out of the system
+ rate of the amount generated within the system
rate of the amount consumed within the system

(c) 2007 Dr. Samir Al-Amer

 Heated Stirred Tank

Ti, wi Mixing V: volume of liquid

T: temperature inside tank

V
temperature at outlet
Q
T, w Ti : temperature at inlet
wi : inlet flow rate

Heater w: outlet flow rate

(c) 2007 Dr. Samir Al-Amer

 Heated Stirred Tank
Conservation of energy

rate of rate of rate of rate of energy

energy = energy of energy of + provided
accumulation inlet material outlet material by heater

= {wi C (Ti Tref )} {w C (T Tref )}+ {Q}

dT
V C
dt
dT
V C = w C (Ti T ) + Q
dt
V : volume C : specific heat w : outlet flow rate Q : heat flow rate
Ti : inlet temp. T : outlet temp Tref : Reference temp

(c) 2007 Dr. Samir Al-Amer

Heated Stirred Tank
Conservation of mass and Energy

Mass Conservation
rate of mass rate of mass rate of mass
=
accumulation of inlet material of outlet material
d (V )
= wi w
dt

Energy Balance

C
(
d V (T Tref ) ) = w C (T T ) w C (T Tref ) + Q
i i ref
dt

(c) 2007 Dr. Samir Al-Amer

 Heated Stirred Tank Model
Mass Conservation
d (V )
= wi w
dt
Energy Balance

C
(
d V (T Tref ))= wi C (Ti Tref ) w C (T Tref ) + Q
dt
Combining
dT
C (T Tref )( wi w) + V C = wi C (Ti Tref ) w C (T Tref ) + Q
dt

(c) 2007 Dr. Samir Al-Amer

 Heated Stirred Tank Model

Alternative Model
dV 1
= ( wi w)
dt
dT wi Q
= (Ti T ) +
dt V V C
further simplifications are possible : Example V is constant,

(c) 2007 Dr. Samir Al-Amer

Electrically Heated Stirred Tank
Model
Assumptions :
wi = w (inlet flow rate = outlet flow rate)
(volume = constant)
(Heater has significant heat capacitance)
Electrical energy heats the coils
then the coil heats the liquid.

(c) 2007 Dr. Samir Al-Amer

Modeling verses Identification
O Modeling:
The have some physical insight
In many cases modeling from basic laws may be too
complex to be practical
In some cases the model derived from the basic laws may
contain unknown parameters
O Identification:
Limited validity models (depends operating point, input,)
Little physical insight (some parameters have no physical meaning)
They are easy to construct and use

(c) 2007 Dr. Samir Al-Amer

System Identification
O Perform an experiment
O Collect data
O Assume a model
O Use data to estimate unknown
parameters
O Validate model

(c) 2007 Dr. Samir Al-Amer

Flow Chart for System Identification
Start
Apriori
Design of Experiment Knowledge

Perform Experiment and Collect data

Choose Model Structure

Choose method& Estimate parameters

Model Validation
No Yes
Model Accepted? STOP New data

(c) 2007 Dr. Samir Al-Amer

Problems in Identification
O Model Selection
O Presence of noise in the measured data
O Some variables of interest are difficult or
impossible to measure
O If the system is time-varying system,
problems may occur if a time-invariant
model is used.

(c) 2007 Dr. Samir Al-Amer

Course Objectives
O To be able to design an identification experiment
O To be able to perform identification experiment
and collect data
O To be able to solve parameter estimation
problems
O To be able to perform model validation
O To be exposed to current research on
identification
O To be able to use existing software / write own
programs to solve identification problems

(c) 2007 Dr. Samir Al-Amer

Reading Material
O Chapter 1 (Soderstrom & Stoica)

(c) 2007 Dr. Samir Al-Amer

 _O

(c) 2007 Dr. Samir Al-Amer

Write
O One idea that you completely
understand from the material of this
lecture.

O One Question that you want to ask

(c) 2007 Dr. Samir Al-Amer

Models
Step Response

Step
Step system
response

(c) 2007 Dr. Samir Al-Amer

(c) 2007 Dr. Samir Al-Amer