Page 183 - Mathematical Models and Algorithms for Power System Optimization

174 Chapter 6

Referring to Tables 6.2 and 6.3, various operating conditions and candidate node sets have been
tested. The test cases show that the proposed algorithm can solve real-scale VAR planning
problems within a reasonable computation time.

6.3.5 Summary

An efficient algorithm for solving single-state discrete variable reactive power optimization
was proposed in this section, which could treat capacitors as discretely adjustable devices in
real-scale systems. The proposed approach is superior to other algorithms in the following
aspects:

(1) Reactive power optimization mathematical model is an MIP model, which takes capacitor
bank number as a discrete variable.
(2) Excellent approximation integer programming algorithm is used to solve large-scale
integer programming problems.
(3) Iteration method is used to solve LP, and the obtained solution is an AC power flow
solution rather than an approximation DC power flow solution.
(4) In accordance with actual situation of a power system, this chapter puts forward an
improved algorithm in allusion to integer programming, which can change two integers
each time. The algorithm can decrease fixed investment costs of capacitor bank and avoid
the unreasonable situation of setting a small amount of reactive power at adjacent nodes.

Calculation results based on an actual system with 135 nodes show that the algorithm is
reasonable and valid. The algorithm in this section provides basic mathematical tools and lays a
foundation for a discrete reactive power optimization algorithm.

6.4 Multistate Discrete VAR Optimization

6.4.1 Overview

The study in Section 6.3 comprehensively discussed discrete reactive power optimization under
single-state conditions. However, reactive power planning may also be implemented under
changing network (multistate) conditions. Multistate reactive power planning takes into
account normal operation, routine maintenance, and accidental failure, and is able to realize
reactive power configuration under multiple operating modes, so as to meet multistate
requirements of changing networks. There is some literature considering reactive power
planning under multiple power flows. Yet they did not take into account the discreteness of
reactive power compensation equipment. Based on the study in Section 6.3, this section
proposes a multistate discrete reactive power optimization model that can be applied to more
complicated actual situations.