QuarterlyIssue 3 (211)
CONTENTS
Foreword
- Henryk KOCOT, Roman KORAB, Kurt ŻMUDA: Unit commitment and dispatch on energy market – a survey of applied methods
- Bernard BARON, Artur PASIERBEK, Paweł SOWA: A power system’s AC model optimized for the power flow and the optimal power flow tasks
- Bernard BARON, Artur PASIERBEK: Comparison of conjugate gradient and quasi-Newton BFGS algorithms in the optimal power flow problem
- Radosław SOKÓŁ: Base mathematical classes library
- Radosław SOKÓŁ: Base matrix classes library. Part I
- Radosław SOKÓŁ: Base matrix classes library. Part II
- Radosław SOKÓŁ: Object-oriented power system model representation
- Marcin POŁOMSKI: Optimal power flow by non-interior-point method
- Tomasz CIEŚLA, Bogusław GRZESIK: Analysis proporties of high frequency flyback-topology SMPS
- Bogusław GRZESIK, Radosław JEŻ: Fem analysis of the irregularity of the magnetic field distribution inside the static magnetizer coil
Henryk KOCOT, Roman KORAB, Kurt ŻMUDA
Instytut Elektroenergetyki i Sterowania Układów, Politechnika Śląska w Gliwicach
UNIT COMMITMENT AND DISPATCH ON ENERGY MARKET – A SURVEY OF APPLIED METHODS
Summary. In the paper the security constrained unit commitment method is dis-cussed. Two stages of the method are distinguished. In the first stage, for the assumed planning horizon, a set of generating units is determined without taking into account net-work constraints. By contrast, in the second stage, the optimum operating policy for the fixed set of generating units is found, while considering network constraints for both normal and abnormal (n - k) states of the power system. In the second stage, there is pos-sible to apply the economic dispatch method that works with the classical power flow al-gorithm or the optimal power flow method as well. The application of the optimal power flow method allows determining the locational marginal prices, on which a model of the energy market may be based.
Bernard BARON, Artur PASIERBEK
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
Paweł SOWA
Instytut Elektroenergetyki i Sterowania Układów, Politechnika Śląska w Gliwicach
A POWER SYSTEM’S AC MODEL OPTIMIZED FOR THE POWER FLOW AND THE OPTIMAL POWER FLOW TASKS
Summary. Solving the optimal power flow (OPF) problem requires defining a mathematical model including all the power system components which influence a solution. Historically, OPF problem solutions used to rely on DC models of a power system. However, algorithmic and performance advances have caused that a more precise AC model can be applied. Because of a size of a typical power system (a large number of nodes and lines), calculating its optimal state is possible only by means of a numerical optimisation. This task requires a model suitable both for numerical application and programming needs. The paper contains a description of the power system mathematical model, created in the framework of the N511 001 32/0852 research project.
Bernard BARON, Artur PASIERBEK
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
COMPARISON OF CONJUGATE GRADIENT AND QUASI-NEWTON BFGS ALGORITHMS IN THE OPTIMAL POWER FLOW PROBLEM
Summary. Solving the optimal power flow (OPF) problem requires defining an objective function and selecting an efficient optimization algorithm. Because of non-linearity of the problem, there is no single perfect solution to it. Each nonlinear problem has to be treated independently. The paper presents results of a comparison of the conjugate gradient and the quasi-Newton BFGS algorithms in the optimal power flow problem. These two methods are primarily used for non-constrained optimisation. However, by modification of the objective function (by introducing a penalty function), one can make these algorithms only search for a solution in the constrained space. The authors propose using an external or hybrid penalty function. The methods presented were analysed and tested in the framework of the N511 001 32/0852 research project.
Radosław SOKÓŁ
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
BASE MATHEMATICAL CLASSES LIBRARY
Summary. The C++ language is not suitable for performing algebraic operations on vectors. While its performance makes it a perfect tool for implementation of complex computational algorithms, coding these algorithms may be a daunting task. The paper presents an object-oriented library implementing vector abstraction. The work was funded from the science budget as the research project No N511 001 32/0852.
Radosław SOKÓŁ
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
BASE MATRIX CLASSES LIBRARY. PART I
Summary. The C++ language does not directly support matrix algebra (implementing every operation requires open-coding a loop); it also does not support sparse matrix techniques. The end effect are code duplication and the necessity of optimising each independent code block separately. One also cannot avoid code readability degradation, caused by replacing a clean algebraic form with C++ code snippets. To solve this problem and improve code creation time (especially in case of the OPF-problem solving software, being developed as a part of a research project), a C++ class library was created. The library consists of classes implementing basic algebraic operations on simple and sparse matrices. The work was funded from the science budget as the research project No N511 001 32/0852.
Radosław SOKÓŁ
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
BASE MATRIX CLASSES LIBRARY. PART II
Summary. The C++ language does not directly support matrix algebra (implementing every operation requires open-coding a loop); it also does not support sparse matrix techniques. The end effect are code duplication and the necessity of optimising each independent code block separately. One cannot also avoid code readability degradation, caused by replacing a clean algebraic form with C++ code snippets. To solve this problem and improve code creation time (especially in case of the OPF-problem solving software, being developed as a part of a research project), a C++ class library was created. The library consists of classes implementing basic algebraic operations on simple and sparse matrices. The work was funded from the science budget as the research project No N511 001 32/0852.
Radosław SOKÓŁ
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
OBJECT-ORIENTED POWER SYSTEM MODEL REPRESENTATION
Summary. An implementation of an optimization algorithm working on data of a power system state requires defining an in-memory representation of the data. A high level of usage friendliness and flexibility can be achieved by means of object-oriented programming techniques. The paper presents an object-oriented power system model representation, created as a part of the research project No N511 001 32/0852.
Marcin POŁOMSKI
Instytut Elektrotechniki i Informatyki, Politechnika Śląska w Gliwicach
OPTIMAL POWER FLOW BY NON-INTERIOR-POINT METHOD
Summary. In the paper a variant of non-interior point method algorithm for solving the nonlinear optimal power flow problem (OPF) is proposed. The OPF problem optimality conditions can be regarded as a particular case of the nonlinear complementarity problem (NCP), in which the complementarity conditions are handled by Chen-Harker-Kanzow-Smale smoothing functions (NCP-functions). The presented method was experimentally verified by applying it to various test power systems including the Polish power system. The obtained results confirm that the new variant is appropriate for optimising large power systems.
Tomasz CIEŚLA, Bogusław GRZESIK
Katedra Energoelektroniki, Napędu Elektrycznego i Robotyki, Politechnika Śląska w Gliwicach
ANALYSIS PROPORTIES OF HIGH FREQUENCY FLYBACK-TOPOLOGY SMPS
Summary. The paper presents properties of high frequency DC-DC flyback-topology SMPS with input: 230 V AC, and output: 5 V, 5 W. The aim of the work is to determine the limit frequency. The SMPS is based on a cascode switch that allows operating with frequency of 3 MHz. A laboratory model of the SMPS was realised and measured. A computer model was developed in Simplorer software and its features were compared with those obtained from measurements.
Bogusław GRZESIK, Radosław JEŻ
Katedra Energoelektroniki, Napędu Elektrycznego i Robotyki
FEM ANALYSIS OF THE IRREGULARITY OF THE MAGNETIC FIELD DISTRIBUTION INSIDE THE STATIC MAGNETIZER COIL
Summary. The paper deals with the analysis of the magnetic field distribution irregularity inside a static magnetizer coil used for magnetization of polymeric membranes made on a basis of neodymium (Nd). The FEM calculations of the magnetic field distribution are presented.