Page 159 - Mathematical Models and Algorithms for Power System Optimization

150 Chapter 6

6.5.3 Algorithm based on Expert Rules for Discrete VAR Optimization 195
6.5.4 Implemetation 200
6.5.5 Summary 205
6.6 Discrete VAR Optimization based on GA 207

6.6.1 Overview 207
6.6.2 Necessity of Applying Artificial Intelligence Algorithms 208
6.6.3 GA-based Model for Discrete VAR Optimization 209
6.6.4 GA-based Algorithm for Discrete VAR Optimization 210
6.6.5 Implementation 214
6.6.6 Summary 216
6.7 Conclusion 218

6.1 Introduction

Reactive power optimization in power system operation is an important issue that
directly affects both the voltage quality of the system and the economic operation of the
power grid. Because system voltage is related to both the reactive and active power
distribution of the power grid, under the condition of meeting the normal operation of the
power system, the objective function of reactive power optimization can be the minimum
investment of the new reactive power equipment or it can be the minimum active power
losses of the system.

Generally, reactive compensation equipment, such as capacitors and other volt-ampere reactive
(VAR) sources, have been installed in substations of power systems. To meet the need of
present load and future system development, new capacitors should perhaps be installed.
Therefore, configuration planning of the new VAR devices is required. When planning a new
capacitor, it is necessary not only to consider the existing capacitors but also to consider the
discrete characteristics of the capacitor bank. With the expansion of the system, VAR
optimization issues involve not only massive integer variables but also nonlinear constraints of
a large-scale power system.

VAR optimization can be attributed to a nonlinear mathematical programming problem.
Considering the discrete characteristics of the number of transformer taps and capacitor banks,
it becomes a more complicated mixed-integer nonlinear programming (MINLP) problem.
Therefore, with regard to VAR optimization for a practical-scale power system, dealing with
dozens or even hundreds of integer variables is required.
This chapter studies the discrete VAR optimization of the power system and power distribution
system in detail, and proposes various models and algorithms for discrete reactive power
optimization.
This chapter deals with the problem of reactive power optimization as a mixed-integer,
nonlinear problem. The current algorithms for exact-integer programming optimization have