Code No: D52A2

BVR AJU IN S TITU TE O F TE CHNO LO GY , N AR S APUR

(UGC - AUTONOMOUS)
I M.Tech II Semester Regular Examinations, OCT 2022
Artificial Intelligence Techniques in Power Systems
(Electrical Power Systems)
Time: 3 Hours Max Marks: 70
1 a. Discuss the key developments and its other significant contributions of artificial Intelligence?
5M
Key Developments of Artificial Intelligence

 AI Workforce Augmentation. In the future, we'll see even more of our jobs being outsourced to AI.
 Better Language Modeling Capability. ...
 Cybersecurity. ...
 The Metaverse. ...
 Low-Code or No-Code AI. ...
 Data-Centric AI. ...
 Autonomous Cars. ...
 Creative AI.
Significant contributions of AI in the society.
Letting AI and automation take over tasks and data analysis can reduce or even eliminate human error.
This is one of the major pros of artificial intelligence, as it doesn't make mistakes like we do. AI is also
immune to factors that affect performance in humans such as stress and fatigue .

b. Illustrate the different methods of Artificial Intelligence 5M

Among the artificial intelligence (AI) techniques, the main algorithms applied in power systems are: 
 artificial neural networks,
 fuzzy logic systems, genetic algorithm,
 particle swarm optimization, colony optimization,
 Simulated annealing and evolutionary computing .

c. Discuss about the growth of AI in industry 4M

According to Precedence Research, the artificial intelligence (AI) market size is projected to surpass
around US$ 1,597.1 billion by 2030 and is expanding growth at a CAGR of 38.1% from 2022 to
2030.

 Companies Covered: Intel Corporation, Micros...

 CAGR: 38.1% from 2022 to 2030
 Largest Market: North America
 Report Scope: Details
2 a. Discuss various types of neural network architectures. 7M
What are the various neural network architectures?

 Feed-Forward Neural Networks. 

 Recurrent Neural Networks (RNN) 
 Convolutional Neural Networks (CNN) 
 Generative Adversarial Networks (GANs)

b. Compare the main differences among the three models of artificial neuron, namely, McCulloch-
Pitts, Perceptron and Adaline? 7M

The McCulloch–Pitt Perceptron Adaline

The McCulloch–Pitt neural network A perceptron model, in Machine ADALINE (Adaptive Linear
is considered to be the first neural Learning, is a supervised Neuron or later Adaptive
network. The neurons are learning algorithm of binary Linear Element) is an early
connected by directed weighted classifiers. A single neuron, the single-layer artificial
paths. McCulloch–Pitt neuron allows perceptron model detects whether neural network and the
binary activation (1 ON or 0 OFF), any function is an input or not and name of the physical device
i.e., it either fires with an activation 1 classifies them in either of the that implemented this
or does not fire with an activation of classes. network.
0
3 a. Explain in detail “Error back propagation training algorithm”. 8M

It is an error reducing algorithm used in artificial neural networks. Artificial neural networks are
networks based on the human’s nerve system. These networks contain well defined set of inputs and
outputs. The network is used to describe the complex relationship between the inputs and outputs of the
network. The name comes from the complexity of the network because the human’s nerves system is so
much complex which is known by all of us.
       We can categorize the error propagation algorithms into two types as follows,

                                           1. Forward propagation algorithms

                                           2. Back propagation algorithms

Our concentration now is on back propagation algorithms.

In this method, the traversal is from output node to various input nodes and hence called back
propagation algorithm. At each node the errors are analyzed and the error with minimum gradient weight
is chosen and is referred to as local minima. So at the end of the traversal we will have a set of local
minimas. Again an analysis is made on the local minimas and an error is chosen which have minimum
gradient weight. This final minima is called global minima. After finding the global minima the
propagation process comes to end.

b. Identify and list the limitations of back propagation algorithm? 6M

The biggest disadvantages of back propagation are: 

 Back propagation could be rather sensitive to noisy data and irregularity.
 The performance of back propagation relies very heavily on the training data.
 Back propagation needs a very large amount of time for training.

Most prominent advantages of Back propagation are:

 Backpropagation is fast, simple and easy to program.

 It has no parameters to tune apart from the numbers of input.
 It is a flexible method as it does not require prior knowledge about the network.
 It is a standard method that generally works well.
4 a. Explain the following fuzzy logic controller components.
i) Fuzzy inference engine

What is inference engine in fuzzy logic?

The inference engine is responsible for applying the inference rules to the fuzzy input in order to
generate the fuzzy output. In particular, the inference rules are used to evaluate the linguistic values and
map them to a fuzzy set which requires the defuzzification process to transform it into a crisp value.

ii) Rule base

A rule-based system is used to store and manipulate knowledge to interpret information in a useful way.
It is often used in artificial intelligence applications and research.
Normally, the term rule-based system is applied to systems involving human-crafted or curated rule sets.
Rule-based systems constructed using automatic rule inference, such as rule-based machine learning, are
normally excluded from this system type.

iii) Data base

iv) Fuzzification

Fuzzification is the process of converting a crisp input value to a fuzzy value that is performed by
the use of the information in the knowledge base. Although various types of curves can be seen in
literature, Gaussian, triangular, and trapezoidal MFs are the most commonly used in the fuzzification
process.
b. Explain the following membership value assignment 7M
i) Intuition
Intuition method is based upon the common intelligence of human. It is the capacity of the
human to develop membership functions on the basis of their own intelligence and
understanding capacity. There should be an in-depth knowledge of the application to which
membership value assignment has to be made. The Figure below shows various shapes of
weights of people measured in kilogram in the universe. Each curve is a membership
function corresponding to various fuzzy (linguistic) variables; such as very light, light,
normal, heavy and very heavy. The curves are based on context functions and the human
developing them. For example, if the weights are referred to range of thin persons we get one
set of curves, and if they are referred to range of normal weighing persons we get another set
and so on. The main characteristics of these curves for their usage in fuzzy operations are
based on their overlapping capacity.

ii) Inference
The inference method uses knowledge to perform deductive reasoning. Deduction
achieves conclusion by means of forward inference. There are various methods for
performing deductive reasoning. Here the knowledge of geometrical shapes and geometry
is used for defining membership values. The membership functions may be defined by
various shapes: triangular, trapezoidal, bell-shaped, Gaussian and so on. The inference
method here is discussed via triangular shape.

iii) Rank ordering

The formation of government is based on the polling concept; to identify a best student,
ranking may be performed; to buy a car, one can ask for several opinions and so on. All
the above mentioned activities are carried out on the basis of the preferences made by an
individual, a committee, a poll and other opinion methods. This methodology can be
adapted to assign membership values to a fuzzy variable. Pairwise comparisons enable
 us to determine preferences and these results in determining the order of the
membership.
iv) Inductive reasoning :
An automatic generation of membership functions can also be accommodated by using the
essential characteristic of the inductive reasoning, which derives a general consensus from
the particular (derives the generic from specific). The induction is performed by the entropy
minimisation principle, which clusters most optimally the parameter corresponding to the
output classes. This method is based on an ideal scheme that describes the input and
output relationship for a well established database, i.e. the method generates memberships
functions based solely on data provided. The method can be quite useful for complex
systems where the data are abundant and static. In situations where the data are dynamic,
the method may not be useful, since the membership functions will continually change with
time. The intent of induction is to discover a law having objective validity and universal
application. The laws of induction are summarized here : Basics of Artificial Intelligence
Tools
• Given a set of irreducible outcomes of an experiment, the induced probabilities are
those probabilities consistent with all available information that aximize the entropy
of the set.
• The induced probability of a set of independent observations is proportional to the
probability density of the induced probability of a single observation.
 The induced rule is that rule consistent with all available information of which the
entropy is minimum

5 a. Two fuzzy sets define by A={(X1,0.3) (X2,0.7) (X3,1)} and B= {(Y1,0.4) (Y2,0.9)} Solve R=A × B 7M

Ans: Problem solution

b. Define the following with example i) Membership ii) power set iii) super set iv) cardinality 7M

i) Membership

We already know that fuzzy logic is not logic that is fuzzy but logic that is used to describe
fuzziness. This fuzziness is best characterized by its membership function. In other words,
we can say that membership function represents the degree of truth in fuzzy logic.
Power Set Definition

A power set is defined as the set or group of all subsets for any given set, including the
empty set, which is denoted by {}, or, ϕ. A set that has 'n' elements has 2n subsets in all. For
example, let Set A = {1,2,3}, therefore, the total number of elements in the set is 3. Therefore,
there are 23 elements in the power set. Let us find the power set of set A.
Set A = {1,2,3}
Subsets of set A = {}, {1}, {2}, {3}, {1,2}, {2,3}, {1,3}, {1,2,3}
Power set P(A) = { {}, {1}, {2}, {3}, {1,2}, {2,3}, {1,3}, {1,2,3} }

Superset 
In set theory, set A is considered as the superset of B, if all the elements of set B are the
elements of set A.
For example, if set A = {1, 2, 3, 4} and set B = {1, 3, 4}, we can say that set A is the superset
of B. As the elements of B [(i.e.,)1, 3, 4] are in set A. We can also say that B is not a superset
of A.
Cardinality
The cardinality of a set is the total number of elements in the set. A power set contains the
list of all the subsets of a set. The total number of subsets for a set of 'n' elements is given
by 2n. Since the subsets of a set are the elements of a power set, the cardinality of a power
set is given by |P(A)| = 2n. Here, n = the total number of elements in the given set.
Example: Set A = {1,2}; n = 2
|P(A)| = 2n = 22 = 4.

Subsets of A = {}, {1},{2},{1,2}

Therefore, |P(A)| = 4.

6 a. Define GA search space? Explain the most important aspects of using genetic algorithm. 7M
The set of all possible solutions or values which the inputs can take make up the
search space. In this search space, lies a point or a set of points which gives the optimal
solution. The aim of optimization is to find that point or set of points in the search space.
Five phases are considered in a genetic algorithm.

• Initial population.
• Fitness function.
• Selection.
• Crossover.
• Mutation.

b. Write the chronological order of historical development of genetic algorithm. 7M

Evolutionary computing is a family of stochastic search techniques that mimic the natural
evolution proposed by Charles Darwin in 1858. In the realm of search techniques the
following classification indicates the position of evolutionary algorithms:

7 a. Define the objective function. Discuss the creation of offspring. 7M

The objective function is a means to maximize (or minimize) something. This something is

a numeric value. In the real world it could be the cost of a project, a production quantity,
profit value, or even materials saved from a streamlined process. With the objective function,
you are trying to arrive at a target for output, profit, resource use, etc.We need to look at the
relationships between constraints and any limitations within the business itself. These can
include production capacity limits, resource availability, or even technology.From a
mathematical perspective, the technical representation of the objective function is what
you're seeing on your screen right now:

It looks like a scary function, but let's break it down and take a look at its individual
components, giving an example of maximizing profit:
 ci is the coefficient that matches the ith variable.
 Xi is the ith decision variable.

b. Illustrate the working principle of GA. Explain i) Permutation encoding ii) tree
encoding. 7M
Genetic Algorithm (GA) is a search-based optimization technique based on the principles
of Genetics and Natural Selection. It is frequently used to find optimal or near-optimal
solutions to difficult problems which otherwise would take a lifetime to solve. It is frequently
used to solve optimization problems, in research, and in machine learning.

Introduction to Optimization: Optimization is the process of making something better. In

any process, we have a set of inputs and a set of outputs as shown in the following figure.

i) Permutation encoding: In this method of encoding, each chromosome is a string of

numbers, representing numbers in a sequence. Each chromosome is essentially an
arrangement, or permutation, of a set of genes. The search space is the total number of
possible permutations, and the population is the current set of permutations. Values must
not be repeated in a single chromosome.

For example, given a set of genes {A, B, C, D, E}, then some possible chromosomes are:

Chromosome 1: ABDCE
Chromosome 2: DBCAE
Chromosome 3: CBADE
This type of encoding is only really useful for problems dealing with ordering, such as the
classic traveling salesman problem. Crossover and mutation operations on this kind of
encoding must take care not to alter the chromosomes in an inconsistent fashion.
ii) Tree Encoding

Tree encoding is typically used to encode programs and expressions

for genetic programming. This encoding is typically a tree of objects
(such as functions or commands in a programming language) that
are interconnecting in specific orders. Frequently, the programming
language LISP is used as the representation, as programs in LISP
are easily represented as trees, and hence easily parsed, and
crossover/mutation operations are relatively easy to perform. The
example on the left is the formula (a-b)+4 encoded as a tree.

8 a. How can an Artificial Neural Network be applied for reactive power control? Explain in detail
Ans: Give the application of any reactive power control

b.Explain the applications of AI techniques in the Speed Control of dc and ac motors 7M

Ans: Give the application of any DC motor speed control

****End ****