Title: Modeling complex systems by simple mathematical models using selfsimilarity and fractionalorder syst
1Modeling complex systems by simple mathematical
models using self-similarity and fractional-order
system identification
- April 19, 2007
- Mechanical Systems GS
- Jason Mayes
2Outline
- Present Outline
- Talk
- Answer Questions
3Outline
- Mathematical modeling
- Philosophy
- Complex systems
- Self-similarity
- Mathematical
- Physical
- Fractional-order system identification
4Mathematical Modeling
xo
- Electro-mechanical model
- Equations of motion
- Add complexity
- Experiments
5Mathematical Modeling
- Finding a mathematical model is finding a
compromise between an intractable problem and
finding a model that sufficiently describes the
system - Complexity of our model depends on what we need
- Dont need great detail ? dont need detailed
model
6The philosophy of analysis
- Many ways to find a mathematical model
7Methodological Reductionism
- Decartes 1637 Discourse on Method
- The world is like a machine
- Everything can be reduced to many smaller,
simpler things - The best way to understand a system is to first
gain a clear understanding of its smallest
subsystems - Models are based on first principles
- Problems
- Size and complexity
8Holism
- Behavior must be studied on the level of the
system as a whole - Aristotles Metaphysics the whole is more than
the sum of its parts - Examples
- Neural nets
- On/off, PID, fuzzy logic
- Expert systems
- Are useful for describing or controlling systems,
but dont explain observed behavior
9A model-based compromise
- Mix of holistic and reductionist approaches
- Model-based system identification
- Can form a model containing a few free parameters
- Very common in heat transfer
- Nusselt number correlations
- Conductivity, convection
- Heat exchangers
- Trend in science Holistic?Reductionist analysis
- Example chemical reactions
10Complex systems
- What is a complex system?
- Physically any system composed of a large number
of components and interactions that creates
difficulties in both understanding and modeling - Mathematically a large system of coupled
equations which are either too complex or too
large to admit a sufficiently useful model of
solution - Properties
- Size/complexity
- Non-linear
- Emergent phenomena
- Memory
11New approaches
- Using self-similarity
- Can reduce complex mathematical models to simple
models - IF full solutions arent needed
- Fractional-order system identification
12Mathematical self-similarity
- Reducible mathematical models
- PDEs ? system of ODEs
- System of ODEs ? scalar ODE
- High-order scalar ODE ? lower order scalar ODE
- Mathematical similarity
- Equations of the same form
- Repeating pattern or coupling in equations
13Reducing infinite-order ODEs
- High-order ODEs can be reduced in the Laplace
domain
Geometric Series!!
n ? 8
14Reducing infinite-order ODEs
- Can reduce an infinite-order ODE to a simple,
finite-order ODE - Only useful in complex mechanical systems if
infinite-order ODEs occur in modeling
15Reducing infinite sets of ODEs
- High-order (infinite) ODEs result from large
(infinite) equations sets
16Reducing infinite sets of ODEs
- Consider a very large system of springs and masses
17Reducing PDEs
- Infinite sets of ODEs also result from the
reduction of PDEs - Finite-volume
- Spectral methods
- Finite difference
- Example heat equation
18Reducing PDEs
- Finite-volume formulation
19Reducing PDEs
- An infinite set of differential equations!
20Reducing PDEs
- Can now reduce the continued fraction
- Now take the inverse Laplace transform
21Reducing PDEs
- PDE has been reduced to a single fractional-order
ODE - Reduction process
- Alternative
- Solve PDE numerically
- Differentiate at the boundary
- Get a global solution
PDE ? System of ODEs ? Single high-order ODE ?
Single low-order ODE
22Applications
- Only need a local solution
- Change of boundary conditions
- Laser or cryogenic surgery
23Using physical self-similarity
- Self-similar equations sets also result from
physically self-similar systems
24Additional forms of similarity
- Similarity between operators can allow for
further reduction - Symmetric/Asymmetric
- Generation dependence
- Four cases
- Symmetric and generation independent
- Asymmetric and generation independent
- Symmetric and generation dependent
- Asymmetric and generation dependent
- All cases allow for reduction
25Fractional-order system identification
- Traditional system identification usually assumes
integer-order models - Fractional-order models can often provide a
better fit - For nearly-exponential systems
- Need models for control
- PID vs. PI?Dµ
- Parameter tuning for fractional-order controllers
26The end.
Questions?