2. Objectives of Power System
Protection
Selectivity
Speed
Reliability
Stability
Adequateness
Sensitivity
Adaptiveness
3. .
.
Development in Power System Relaying
.
Performance
1900 years 1960 1975 2000
Electromechanical Relays
Microprocesso
r-
Based Relays
(Digital)
Static
Relays
Electroni
c Circuits
Digital ICs
(mP,DSP,ADC,)
Digital Proc.
Algorithms Digital ICs
(mP,DSP,ADC,
neuro-IC
fuzzy-IC)
AI-based
Methods
Communication
Facility
AI-Based Relays
(Intelligent)
4. Scope of the Study
AI Applications to Digital Protection like:
Transmission Line Fault Classification
Distance relaying
Machine Winding Protection
Transformer Differential Protection
Transformer Fault Diagnosis
5. .
.
XX---Relay setting& coordination
---XXHIF detection
---XXTransformer fault diagnosis
--XXXTransformer differ. relaying
-X-XX
Machine Winding Relaying
XXXXXDistance Relaying
-XXXXTL fault classification
SelectivitySpeedSecurityDependabilityProtection Area
Shortcomings of Conventional Protection
Systems
Key: “-” no problem, “X” some problems, “XX” big problems
6. .
.
Characteristics of Digital Relaying
Self-diagnosis: improving reliability.
Programmability: multi-function, multi-
characteristic, complex algorithms.
Communication capability: enabling
integration of protection & control.
Low cost: expecting lower prices.
Concept: no significant change (smart copy of
conventional relays).
7. .
.
Motivation for AI-Based Protection
Enabling the introduction of new relaying
concepts capable to design smarter, faster, and
more reliable digital relays.
Examples of new concepts: integrated protection
schemes, adaptive protection & predictive
protection.
9. Expert System
Definition: Expert System is a computer
program that uses knowledge and inference
procedures to solve problems that are
ordinarily solved through human expertise
10. .
.
Structure of Rule-Based Expert System
Knowledge
Acquisition
Facility
Explanation
Facility
User Interface
Knowledge
Base (Rules) Inference Engine
Data Base
(facts)
11. ANN Models
Feedback
Constructed Trained Nonlinear
Adaptive
Resonance
Hopfield
(recurrent)
Linear
Kohonen
(Self-
Organizing
Map)
Unsupervised Supervised
MLP
(Back-
Propagation
Feed
Forward
Classification of ANN Models
12. Fuzzy If-Then Rules
If X1 is BIG and X2 is SMALL
Then Y is ON,
If X1 is BIG and X2 is BIG Then
Y is OFF.
..
DefuzzificationFuzzy
Inference
Inference methods:
Max-Min
composition,
Max-Average comp.,
..
Fuzzification
Membershi
p functions
Input
variables
Defuzzification
methods:
Center of area
Center of sums
Mean of Maxima,..
Output
Decision
X1 is 20% BIG&
80% MEDIUM
Main Components of Fuzzy Logic Reasoning
13. Samples of 3-ph
Voltages &
Currents
Filtered
Samples
Simulation
Environment
“EMTP”
Fault type,
location &
duration
System
model,
parameters &
operating
conditions Pattern
Classifier
Performance Evaluation
Anti-
aliasing
& other
Filters
Feature
Extraction
Training Set
Testing Set
Classifier
output
(training)
Pattern
Classifier
Training target
Classifier
parameters
Training error
Testing target
Testing error
Classifier
output
(testing)
Steps of Designing an AI-Based Protective Scheme
14. Modules of Intelligent Transmission Line Relaying
Fault
Detection
Trip Signal
Data
Processing
Transmission Line
Fault Identification
Direction
Discrimination
Fault
Location
Arcing
Detection
Faulted Phase
selection
Fault Type
Classification
Decision Making
Features
V
I
15. Application 1
Transmission Line Fault Classification
Conventional schemes: cannot adapt to changing
operating conditions, affected by noise& depend on
DSP methods (at least 1-cycle).
Single-pole tripping/autorecloser SPAR requires the
knowledge of faulted phase (on detecting SLG
Single-pole tripping is initiated, on detecting arcing
fault recloser is initiated).
Motivation
16. ANN4
20-15-10-1
ANN1
30-20-15-11 Control Logic
Arcing
fault
phase-T
1/4 cycle
each
(5
samples)
VR,VS,VT
IR,IS,IT
ANN3
20-15-10-1
Decision
K
N
O
W
LE
D
G
E
B
AS
E
One
cycle
each
(20
samples)
VS
VT
VR
Arcing
fault
phase-S
Arcing
fault
phase-R
ANN2
20-15-10-1
Enabling Signals
Fault Type
RST
RG
Transmission Line Relaying Scheme
45000 training
patterns
5-7 ms
25 ms
18. Other AI Applications
Fuzzy & fuzzy-neuro classifiers used for fault
type classification (1-cycle).
Pre-processing: 1- Changes in V&I,
2- FFT to obtain fundamental V&I,
3- Energy contained in 6 high freq. bands
obtained from FFT of 3-ph voltage.
Measures from two line ends.
Implementation of a prototype for ANN-based
adaptive SPAR
19. Application 2:
Distance Relaying
Motivation
Changing the fault condition, particularly in the
presence of DC offset in current waveform, as well
as network changes lead to problems of underreach
or overreach.
Conventional schemes suffer from their slow
response.
20. AI Applications in Distance Relaying
Using ANN schemes with samples of V&I measured
locally, while training ANN with faults inside and
outside the protection zone.
Same approach but after pre-processing to get
fundamental of V&I through half cycle DFT filter.
Combining conventional with AI: using ANN to
estimate line impedance based on V&I samples so as
to improve the speed of differential equation based
algorithm.
21. AI Applications in Distance Relaying
Pattern Recognition is used to establish the
operating characteristics of zone-I. The impedance
plane is partitioned into 2 parts: normal and fault.
Pre-classified records are used for training.
Application of adaptive distance relay using
ANN,where the tripping impedance is adapted
under varying operating conditions. Local
measurements of V&I are used to estimate the
power system condition.
22. Application 3:
Machine Winding Protection
Motivation
If the generator is grounded by high
impedance, detection of ground faults is not
easy (fault current < relay setting).
Conventional algorithms suffer from poor
reliability and low speed (1-cycle).
23. DFT Filtering
In5 In6In3 In4In1 In2
Ia2 Ib2Ib1Ia1
Ra
Ic1 Ic2
A
C
B
L-L
ANN2
L-L-L
ANN3
L-G
ANN1
OutputOutputOutput
Iad(n) = Ia1(n)- Ia2(n)
Iaa(n) = ( Ia2(n) + Ia1(n) )/2
Current Manipulator
Icd(n) Ica(n)Ibd(n) Iba(n)Iad(n) Iaa(n)
Sampling
Ib2(n) Ic2(n)Ic1(n) Ia2(n)Ia1(n) Ib1(n)
ANN-Based Generator Winding Fault Detection
24. Application 4:
Transformer Differential Relaying
Motivation
Conventional differential relays may fail in
discriminating between internal faults and other
conditions (inrush current, over-excitation of core, CT
saturation, CT ratio mismatch, external faults,..).
Detection of 2nd and 5th harmonics is not sufficient
(harmonics may be generated during internal faults).
25. Multi-Criteria Differential Relay based on
Self-Organizing Fuzzy Logic
One differential relay per phase.
12 criteria are used and integrated by FL.
Examples of criteria: (ID=differential current)
q1
q3
q4
q6
q1> highest expected inrush current
q3 < 10-15%
q4 > current for over-excitation
q6 < 30%
ID1
ID2/ID1
ID1
ID5/ID1
Definition Criterion StatementSign
26. APPLICATION 5:
Transformer Fault Diagnosis
Motivation
Conventional methods, e.g., Dissolved Gas Analysis
(DGA), suffers from imprecision & incompleteness.
IEC/IEEE code for DGA relates the fault type to the ratios
of gases; e.g.,
IF (C2H2/C2H4 =0.1-3) AND (CH4/H2 < 0.1) AND
(C2H4/C2H6 < 1) THEN (the fault is High energy partial
discharges)
28. Each subspace is described by a fuzzy if-then rule based on the
patterns of training set.
C2H4/C2H6
C2H2/C2H4
S M L
S
M
L
S
M
CH4/H2
L
29. CONCLUSION
The applications of Artificial Intelligence in the arena
of Relaying employs the methods of ANN,ES and FL.
Adaptiveness and smartness get highly improved by
inculcating the AI methods into Conventional
Relaying.
There is a great scope of exceptional developments in
this arena ,hence imparting a smart outlook for the
entire power system.
30. REFERENCES
Artificial Intelligence Techniques in Power Systems by
K. Warwick, Arthur Ekwue, Raj Aggarwal, Institution of
Electrical Engineers.
http://web.stanford.edu/class/cs227/Lectures/lec01.pdf
Computational Intelligence Systems and Applications:
Neuro-Fuzzy and Fuzzy logic By Marian B. Gorzalczany