Course detail

Simulation and Optimalization of Electromechanical System

FEKT-MSIOAcad. year: 2012/2013

The survey of optimatisation methods, stochastic optimization methods and their utilization in electrical engineering. Parameters identification of elektromechanical systems. Artificial networks. Basic theory paradigms, learning algorithms. Artificial neural networks application in prediction and dynamic simulation.

Language of instruction

Czech

Number of ECTS credits

5

Mode of study

Not applicable.

Guarantor

doc. Ing. Čestmír Ondrůšek, CSc.

Department

Department of Power Electrical and Electronic Engineering (UVEE)

Learning outcomes of the course unit

The students become familiar with classical and stochastic optimisation methods. Genetic algorithms, simulated annealing and climbing algorithm. Parameter identification and optimisation of electrical mechines using artificial inteligence methods. Artificial neural networks. The basic paradigm and the the properties of artificial neural networks. Neural models of electrical machinesin steady state and in transients. Prediction of electrical machinesbehaviour. The use of artificial neural in electrical machines control. Short and long therm predictionof electrical load of power utility system using neural networks, radial base functions and fuzzy neural networks.

Prerequisites

The subject knowledge on the Bachelor´s degree level is requested.

Co-requisites

Not applicable.

Planned learning activities and teaching methods

Teaching methods depend on the type of course unit as specified in the article 7 of BUT Rules for Studies and Examinations.

Assesment methods and criteria linked to learning outcomes

Requirements for completion of a course are specified by a regulation issued by the lecturer responsible for the course and updated for every.

Course curriculum

Not applicable.

Work placements

Not applicable.

Aims

To by acquinted with up to date methods of simulation, optimisation and parameter determination of electromechanical systems.

Specification of controlled education, way of implementation and compensation for absences

The content and forms of instruction in the evaluated course are specified by a regulation issued by the lecturer responsible for the course and updated for every academic year.

Recommended optional programme components

Not applicable.

Prerequisites and corequisites

Not applicable.

Basic literature

Kosko,B.:Neural Networks for signal processing,Prentice Hall

Recommended reading

Not applicable.

Classification of course in study plans

  • Programme EEKR-M Master's

    branch M-SVE , 2 year of study, summer semester, elective specialised

  • Programme EEKR-M Master's

    branch M-SVE , 1 year of study, summer semester, elective specialised

Type of course unit

 

Lecture

26 hod., optionally

Teacher / Lecturer

doc. Ing. Čestmír Ondrůšek, CSc.

Syllabus

Survey of classical optimization methods.
Optimization methods based on artificial inteligence. Genetic algorithm.
Simulated annealing, climbing algorithm.
Principle of electrical machines parameters determination and optimisation.
Neural networks. The basic paradigma multilayer neural networks.
Learning algorithms. Back-propagation.
Recurent Neural Networks
Application of ANN to electrical mechine steady state and transient behaviour.
Radial basis function network (RBFN). Main features of RBFN. Aplication of RBFN to load forecasting.
Adaptively trained neural networks (ATNN). Problem description. Adaptive learning. Adaptive mechanism. Construction of ATNN architecture.
Hybrid fuzzy-neural networks (FNN). Motivation for using hybrid approach. Structure of FNN. Membership functions. Artificial neural networks.
Short therm load forecasting for large distribution system.
Optimization of electrical mechine design using artificial intelligence.

Exercise in computer lab

26 hod., optionally

Teacher / Lecturer

doc. Ing. Čestmír Ondrůšek, CSc.

Syllabus

Optimization and identification of electrical machines parameters by genetic algorithm.
Computer model of neural networks.
Learning algorithms. Back-propagation.