Upload
sai0319
View
217
Download
0
Embed Size (px)
Citation preview
8/7/2019 orginal ppt2003
1/13
BY
B SAI HARSHA07F91A0255
REALTIME POWER SYSTEMSECURITY ASSESSMENT USINGARTIFICAL NEURAL NETWORKS
8/7/2019 orginal ppt2003
2/13
ABSTRACTContingency analysis of a power system is a major activity
in power system planning and operation.
The traditional approach of security analysis involving thesimulation of all conceivable contingencies by full AC loadflows, becomes prohibitively costly in terms of time and
computing resourcesA new approach using Artificial Neural Network s has been
proposed in this paper for real-time network securityassessment.
The proposed paradigms are tested on IEEE 14 bus and
30 bus systems.
8/7/2019 orginal ppt2003
3/13
INTRODUCTIONSecurity refers to the ability of the system to withstand the
impact of disturbance
The process of investigating whether the system secure orinsecure in a set of proposed contingencies is called SecurityAnalysis.
Security assessment has two functions the
first is violation detection in the actual system operatingstate.
The second, much more demanding, function of securityassessment is contingency analysis
In this paper, for the determination of voltage contingency
ranking, a method has been suggested, which eliminatesmisranking and masking effects and security assessment hasbeen determined using Radial Basis Function (RBF) neuralnetwork for the real time control of power system
8/7/2019 orginal ppt2003
4/13
Artificial Neural NetworksArtificial neural networks (ANN) are massively parallel
inter connected networks of simple elements known as
artificial neurons and their connectivity is intended tointeract with the objects of real world, in a similarmanner as the biological nerves systems do.
The simple neuron model is shown in fig(a). unitmultiplies each input x by a weight w and sums the
weighted inputs. The output of the figure is
NET = x1w1 + x2w2 + .+xnwn :
OUT = f(NET)
input ouput
hidden layer
F
8/7/2019 orginal ppt2003
5/13
Basic features of ANNs are:
High computational rates due to the massive parallelism.
Fault tolerance.
Training the network adopts itself, based on the informationreceived from the environment
Programmed rules are not necessary.
Primitive computational elements.
Radial basis function networks:
The Radial Basis Function is similar to the Gaussianfunction, which is defined by a center and a widthparameter.
The Gaussian function gives the highest output when theincoming variables are closest to the center appositionand decreases monotonically as the rate of decrease.
8/7/2019 orginal ppt2003
6/13
Training of the RBF networka)Computation of RBF parameters:
Initialize the center of each cluster to a randomlyselected training pattern.
Assign each training pattern to the nearestcluster. This can be accomplished by calculating
the Euclidean distances between the trainingpatterns and the cluster centers.
When all the training patterns are assigned,calculate the average position for each cluster
center. Then they become new cluster centers.Repeat steps (2)and (3) until the cluster centers
do not change during the subsequent iterations.
8/7/2019 orginal ppt2003
7/13
8/7/2019 orginal ppt2003
8/13
To achieve an accurate picture of systems vulnerability tooutage event several issues used to be considered. They are :
a) system model
b) Contingency definition
c) Contingency list
d) performance
e) Modeling details
When finding critical contingencies and giving ranking tothem, we should consider the following.
a)Magnitude of voltage violation
b)Number of violations occurringc)Relative importance to each voltage violationd)Nearness of voltages to the security limitse)Load level of the system when evaluated
8/7/2019 orginal ppt2003
9/13
Real power performance index:
Real power based performance indices (PIp) to calculate PI valuesof each line under outage conditions, is defined as
On-line security analysis:
There are three basic elements of on-line security analysis andcontrol, namely, monitoring assessment and control. They aretied together in the following framework.
Step-1) Security monitoring :
Using real-time systems measurements. Identify
whether the system is in the normal state or not. If the system isin an emergency state, go to step-(3). If load has been lost, go
to step-(4).
=
=
L
kii i
ipip
P
PWPI
1
max2
8/7/2019 orginal ppt2003
10/13
Step-2) Security Assessment:
If the system is in the normal state, determinewhether the system is secure or insecure with
respect to a set of next contingencies.
Step-3) Emergency control:
Execute proper corrective action to bringthe system back to the normal back to the normal
state following a contingency which causes thesystem to enter an emergency state. This issometimes called remedial action.
Step-4) Restorative Control:
Restore service to system loads
8/7/2019 orginal ppt2003
11/13
Simulation results:
The training patterns were generated forbase case and different network outages using aNewton-Raphson load flow program by varyingthe loads at each bus randomly covering the
whole range of operating condition up to 110%of the base case loading value. For each power system total 100 patternswere generated. Out of these, 75 patterns were
used to train the RBF network and remaining 25patterns were used to test the accuracy androbustness of the trained RBF network.
8/7/2019 orginal ppt2003
12/13
CONCLUSION
In this paper power system static security assessmenthas been investigated. The test results presented onIEEE 14-bus system and IEEE 30-bus system providesthe following observations.
A new method has been reported for calculating voltageperformance index for contingency ranking. Whicheliminates misranking and masking problems? Rankingof all contingencies is same irrespective of values ofweights supplied.
The RBF neural network model provides more accurateresults for both the security and insecurity cases.
Training is very fast as the RBF network has thecapability of handling large dateTesting time is less than 0.2 micro sec.
8/7/2019 orginal ppt2003
13/13