Page 526 - Decision Making Applications in Modern Power Systems

A reliable decision-making algorithm Chapter | 18 485

Stage-I Data preprocessing/Feature selection

Design of an input dataset for
Stage-II
training/ Testing of WPE_BDT modules

No-fault condition Detection of Fault condition
Stage-III System Condition
(WPE_BDT-1)

Power swing condition

Unstable
power swing Discrimination
Stage-IV OST of power swing
(WPE_BDT-2)

Stable power swing
No-fault
condition Detection of
Stage-V PSB shunt faults
(WPE_BDT-3)

Fault condition
Classification
Stage-VI of shunt faults
(WPE_BDT-4)
FIGURE 18.6 Flowchart of proposed reliable decision-making algorithm using WPE_BDT
modules. WPE_BDT, Wavelet packet energy and bagged decision tree.

been instigated as follows: (1) Stage I: data preprocessing and feature
selection by applying appropriate signal processing technique, (2) Stage II:
design of input dataset for training/testing of different WPE_BDT modules
considering wide range of variation in different parameters as per
Tables 18.2 18.5, (3) Stage III: WPE_BDT-1 for detection of power system
network condition such as no-fault, fault, PS conditions, (4) Stage IV:
WPE_BDT-2 for discrimination of stable and unstable PS, (5) Stage V:
WPE_BDT-3 for detection of faults during PS, and (6) Stage VI:
WPE_BDT-4 for classification of shunt faults during PS.
Four different WPE_BDT modules have been developed using distinct
and suitable features, such as S1 S5 are used for WPE_BDT-1 and -2 mod-
ule, S6 S15 are used for WPE_BDT-3 and -4 modules. The output of