IEEE Power Electronics Magazine - December 2020 - 20

What is AI? AI basically deals with computer emulation
of human thinking. A human brain with a biological nervous
system has natural intelligence and ability to think, that is,
to learn, reason and comprehend. The goal of AI is planting
human intelligence in a computer so that the computer can
think intelligently and solve problems like a human being. Can
a computer really think and take intelligent decision? Computational intelligence has been a debatable subject for a long
time. It is true that the intelligence of a computer is far inferior to the natural intelligence of a human being. But it is also
true that computer can help solving complex problems which
are difficult to solve by traditional methods. The AI technology has gone through evolution in the past several decades,
and has been extensively applied in information management,
space exploration, military systems, finance, industrial and
power systems, etc. which also include power electronics and
motor drives. AI techniques for applications in power electronics can be generally classified into four different areas.
These are: Expert systems (ES), fuzzy logic (FL), artificial
neural network (ANN or NNW), and genetic algorithm (GA)
or evolutionary computation (EC). However, the concept of
machine learning (ML) that involves the study of computer
algorithms that improve automatically through experience has
evolved over a long time. Lotfy Zadeh, the inventor of fuzzy
logic, defines ES as " hard " or precise computing, whereas FL,
NNW, and GA are defined as " soft " or approximate computing. The technology has gone through fast evolution in the
past several decades. In the recent years, AI has been almost
synonymous with the NNW applications. The potentiality of
this area appears to be tremendous. In this article, the principles of ES, FL and NNW will be briefly reviewed along with
discussion of an example application in each area.

Expert System
The ES, also known as classical AI, is the forerunner of all
the AI techniques. It is basically an intelligent computer pro-

gram that is designed to embed the expertise of a human
being in a certain domain, such as power electronics. A
power electronics engineer acquires the expertise by education and training over a prolonged period of time. The objective of an ES is to replace a power electronics expert by the
advanced computer-embedded program that is designed to
solve the problem [1]. Figure 1 shows the basic elements of
an ES with interconnection as follows:
■■Knowledge Base
■■Inference Engine
■■User Interface
■■Explanation Subsystem for user education.
The core of the ES is the knowledge base which consists of
expert knowledge that is supported by database as shown. The
expert knowledge basically consists of a matrix of IF .... THEN
rules using Boolean variables which are similar to logical thinking of a human being to solve a problem. An example rule for
fault-tolerant control of a converter-fed ac drive system may be
IF converter dc link voltage 1 200 V AND
ac line voltage = 0 AND
motor speed 2 1500 rpm
THEN reduce motor speed by 20% by regenerative
braking.
The expert knowledge is acquired from the domain expert
(i.e. a power electronics engineer) by a software expert
defined as knowledge engineer. The inference engine is basically the controller or executive software of the ES that tests
the rules of the knowledge base in sequence and tries to
draw inferences or conclusions that are supplied to the user.
The inference engine communicates with the user in a very
user-friendly dialog in a natural language, such as English.
Through the inference engine, the ES requests the parameter
values of the rules from the user to draw conclusions, or
solution of the problem. One very important function of the
ES is the user education through the explanation subsystem.

Explanation
Subsystem
(Help, Why, How)
(User Education)

User Dialog
(Semi Skilled
Engineer)

User Interface

Knowledge Base

Inference
Engine

Work Space

Expert
EXPERT
KNOWLEDGE
Knowledge

DATABASE
Database

Knowledge
Acquisition

Domain Expert
(Power Electronics
Engineer)
FIG 1 ES block diagram showing the different elements and their interconnection.

IEEE POWER ELECTRONICS MAGAZINE

z December 2020

IEEE Power Electronics Magazine - December 2020

Table of Contents for the Digital Edition of IEEE Power Electronics Magazine - December 2020