Computational intelligence based power tracing for non discriminatory losses charge allocation and voltage stability improvement / Zulkiffli Abdul Hamid

Abdul Hamid, Zulkiffli (2013) Computational intelligence based power tracing for non discriminatory losses charge allocation and voltage stability improvement / Zulkiffli Abdul Hamid. PhD thesis, Universiti Teknologi MARA.

[img] Text
TP_ZULKIFFLI ABDUL HAMID EE 13_5.pdf

Download (494kB)

Abstract

This thesis proposes a new power tracing technique using computational intelligence approach for non discriminatory losses charge allocation and voltage stability improvement. Contrary to conventional techniques which mainly rely on matrix operation, the proposed algorithm implements optimization technique as an alternative for performing the tracing process. At first, in producing a good optimization algorithm, a hybridization technique was proposed for adopting the finest features of two different algorithms; namely the Genetic Algorithm (GA) and continuous domain Ant Colony Optimization (ACOr). The hybrid algorithm is termed as the Blended Crossover Continuous Ant Colony Optimization (BX-CACO). It was found that performing power tracing via BX-CACO produced reliable tracing results as it is free from assumption like proportional sharing principle (PSP). Without treating the power system to be lossless, the tracing results are based on actual system condition; which means that they are consistent. Despite BX-CACO required computation time during optimization process, it is still within tolerable range. In addition, the proposed technique was able to promote fair losses charge allocation by involving imaginary consumers other than generation companies (GENCOs) and distribution companies (DISCOs); where, not all conventional tracing techniques include such consideration in their pricing scheme. Subsequently, the developed tracing algorithm was modified in the context of stability index tracing. At this stage, the limited application of power tracing in the field of power system economics was enhanced for the purpose of voltage stability improvement. By utilizing the Fast Voltage Stability Index (FVSI) as the index to be traced, the proposed stability index tracing is called FF57-Tracing (FVSI-T)', which consists of two schemes namely FFSZ-Generation Tracing (FVSI-GT) and FVSI-Load Tracing (FVSI-LT). From both schemes of FVSI-T, a ranking list indicating the priority of buses to be performed any countermeasures against voltage instability was derived. Contrary to conventional techniques such as sensitivity analysis and stability index approach, the derived ranking list consists of two types of locations ranked based on their priority; which are bus and line. In addition, the ranking list derived from BX-CACO-based-FF57-T is available at any system conditions and includes all possible sources and sinks that also participate in the system. This highlights its merit over conventional ranking techniques. After experiment and comparative studies, it was justified that the derived ranking list gave reliable signal for satisfactory and consistent voltage stability improvement in the problem of generation power dispatch, placement of compensating devices, and load shedding. Regardless of the change in system condition, the derived ranking list from FVSI-T resulted in steady trend of voltage magnitude and consistent losses reduction as compared to convetional methods. Eventually, in providing an effective load shedding scheme with reliable amount of load power to be shed, an intuitive Fuzzy Inference System (FIS) was designed using FVSI as the criterion when designing its membership functions and rules. Through experiment, the proposed method resulted in the most consistent voltage stability improvement over other methods as both FIS and the derived ranking list are based on FVSI.

Item Type: Thesis (PhD)
Creators:
CreatorsID Num.
Abdul Hamid, ZulkiffliUNSPECIFIED
Subjects: T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Electric apparatus and materials. Electric circuits. Electric networks
Divisions: Faculty of Electrical Engineering
Item ID: 25840
Uncontrolled Keywords: computational, intelligence, tracing
Last Modified: 27 Sep 2019 04:04
Depositing User: Admin Pendigitan 1 PTAR
URI: http://ir.uitm.edu.my/id/eprint/25840

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year