Vous êtes sur la page 1sur 5

1

A Novel Web based Expert System Architecture for On-line and Off-line Fault Diagnosis and Control (FDC) of Transformers
M. Babita Jain, Member, IEEE, M.B.Srinivas, Member, IEEE, Amit Jain, Member, IEEE

Abstract-- The paper presents the structure and usage mode for
an integrated fault-diagnosis and control system for transformers. The fault diagnosis and control system allows online analysis on a desktop application, a web application and offline analysis for determining the transformer faults and remedies based on certain symptoms observed on the equipment and their comparison with the survey results of Cooling system condition, Bushing condition, Insulation system condition, Partial discharge evolution, Transformer switching in/switching off, exceeding of monitored parameter limits and estimation of remained life, the data of which is continuously updated into the database of the existing system. The proposed FDC system uses a Web based Expert System Architecture which has been proven as an effective platform for diagnostic and control applications of transformers. Conventional expert systems [1] are PC based and are not of much use where the expertise is required over a large network. The advent of the internet has strongly influenced modern power systems. A web based expert system [2] with knowledge base and database connectivity is a novel architecture and is an apt solution for both on-line and off-line applications. The purpose of this work is to develop a web based expert system shell and then use it for the fault diagnosis and control (FDC) [3] of transformers either off-line (on a stand alone system) or on-line through the internet with the facility of on-line up gradation of data related to periodic monitoring and maintenance of transformers. The proposed system goes beyond imitation of expertoperators' knowledge, being able to inference fault scenarios using both knowledge base and database of the required equipment with no limitation to the number of pieces of equipment being handled. It is easier, flexible, and more user friendly. It can be used for both stand-alone (off-line) and Webbased (on-line) applications.

Index Terms-- Control, Diagnosis, Database, Expert System, Expert System Shell, Fault, Knowledge Base, On-line diagnosis, Transformers, Web based expert system shell.

I. INTRODUCTION

n operation, the transformers are subjected to combined

stresses (electrical, thermal, mechanical and environmental), finally leading to the decrease of their life and appearance of some faults with irreversible evolution. Also taking into account the fact that most transformers are in operation since more than 20 years, it results the necessity of fitting out the transformers with systems for monitoring, diagnosing and control of faults. For carrying out the monitoring as source of a diagnosis as correct as possible, a mode for collecting data on the surveyed process, saving and storing them, filtering and verifying and finally transmitting them to the user is necessary. This calls for an efficient decision support system capable of handling large volumes of data, converting the data into meaningful information and supplying the appropriate personnel with the right information at the right time, supporting for on-line and off-line fault diagnosis and control system. This can be achieved through a web based expert system package with its unique architecture of connectivity to knowledge base as well as database. The present work is carried on in two phases: 1. Development of a web based expert system shell with a graphical user interface and a database connectivity 2. Development of an on-line and off-line fault diagnosis and control system (FDC) for Transformers using the web based expert system shell.

II. DEVELOPMENT OF THE WEB BASED EXPERT


M. Babita Jain is PhD student at Power Systems Research Center, International Institute of Information Technology, Hyderabad, Andhra Pradesh, India (e-mail: babita_j2000@yahoo.com). Dr M B Srinivas, Associate Professor, Head, VLSI & Embedded Systems at International Institute of Information Technology, Hyderabad, Andhra Pradesh, India (e-mail: srinivas@iiit.ac.in). Dr Amit Jain, Research Assistant Professor, Head, Power Systems Research Center at International Institute of Information Technology, Hyderabad, Andhra Pradesh, India (e-mail: amit@iiit.ac.in).

SYSTEM SHELL

An expert system [5] is a computer program that represents and reasons with knowledge of a specified subject with a view to solving problem directly or giving advice. The existing expert system architecture supports only Knowledge base and only on stand alone systems.

Fig 1 is the architecture of the developed web based expert system shell. The developed shell extends excellent support to knowledge bases as well as databases to multiple users round the globe in contrary to the existing expert system architecture with only single user and limited knowledge base compatibility. This module deals with the development of a web based expert system shell. In an environment where expert data and knowledge expertise is required but not available due to either lack of required personnel or with personnel not having required expertise in diverse fields due large size of data, a package developed using the web based expert system shell would be a fit of the tool to the problem. Experts Historical Data Books

application like fault diagnosis or condition monitoring or preventive maintenance or any other area where extensive data is available and assimilation of data or extracting useful information or knowledge from the data is difficult. It is a rule based web expert system shell developed using .NET technology coupled with the inherent property of storing data in text files and database tables. The architecture allows easy addition of new knowledge bases and databases of the application. From the beginning the web based expert system shell has been designed with extensibility and scalability as the major criterion. B. Design Details 1. Knowledge Base: The knowledge base of the expert system shell stores the extensive knowledge gathered from experts, historical data and books regarding the application in the form of rules. The knowledge can be either factual or heuristic. The knowledge is stored in the form of production rules, which contain if-else rules. The knowledge base is divided into three major sections namely: variable section, rules section and asks section. The variables section is used to declare all the variables used in the knowledge base, the rules section is used to define all the rules used to represent the knowledge and the asks section includes all the questions and options for the user to select and continue with the consultation for getting the required assistance from the expert system. The Fig 3 shows the typical format of the knowledge base required for the expert system.

Data Acquisition

Data Acquisition

Knowledge Base

Data Base

Inference E i Select Knowledge Base

DB Tools

Select Data Base

Main Menu

Graphical User Interface

user1 user2

user3 userx usery ADMIN usern

Fig. 2. Format of Knowledge Base of Web Based Expert System Shell

Fig. 1. Architecture of Web Based Expert System Shell

A. Functional Design The development of the web based expert system shell involves: 1. Knowledge Base Design 2. Database Design 3. Development of an inference engine 4. Development of a graphical user interface This expert system structure helps in refining the information from raw data to the one supporting a particular

2. Database Design: The information gathered is either factual, heuristic or query based. Of the data gathered, the data which can be put in the form of a question and answer are put in the database of the application. The database table of the expert system shell has three fields namely: the id, question and answer. The database can be accessed by entering the keyword related to the information seeked. This leads to displaying of all the questions containing the keyword and selection of any one of the questions would provide for the displaying of the required information.

3. Inference Engine: It provides the system control. It applies the expert domain knowledge to what is known about the present situation to determine new information about the domain. The inference engine is the mechanism that connects the user input in the form of answers to the questions to the rules of knowledge base and further continues the session to come to conclusions. This process leads to the solution of the problem. The inference engine also enables the expert system's interface to data sources and to the user. The inference engine also supports to identify the rules of the knowledge base used to get decision support from the system and also forms the decision tree. The inference engine is developed using VB.NET and it is a DLL (Dynamic Linking Library) file. 4. Graphical User Interface: The graphical user interface of the developed web based expert system shell comprises of: I. Input interface II. User interface and III. Working memory I. Input Interface: The input interface to the package is through the admin page of the GUI. This allows the expertise to be uploaded as the knowledgebase files in text format and also allows updating the database tables for given equipment by the expert or the knowledge engineer who would act as the administrator for the given package. The interface also allows uploading the images of the equipment into the database to help in presenting an easy user interface. The input interface has been developed using C#.NET and MS Access has been used to store information in database tables. The advantage of Access over SQL and other database servers is its ease of implementation while hosting the package on web. II. User Interface: The user interface to the package is through user page of the GUI. This allows for creation of new users and also allows the existing user to consult the expert system in a user friendly manner for decision support. The user can select between the database and knowledge base and can use the database with a very convenient search engine mechanism and the knowledge-base in the form of questions and answers. The user interface has been developed using the C#.NET technology. III. Working Memory: The working memory contains the information that the system has received about a problem at hand. In addition, any information the expert system derives about the problem is also stored in the working memory. III. DEVELOPMENT OF AN ON-LINE AND OFF-LINE FAULT DIAGNOSIS AND CONTROL (FDC) SYSTEM FOR TRANSFORMERS The developed web based expert system shell is now used for the development of an on-line and off-line FDC system. Expert system to be used for fault diagnosis and control calls

for a deep knowledge base. This is possible by extensive data collection by the knowledge engineer from the experts in the field, historical data of fault occurrences and from books related to the field. Lot of data is also obtained from the survey results of Cooling system condition, Bushing condition, Insulation system condition, Partial discharge evolution, Transformer switching in/switching off, exceeding of monitored parameter limits done at regular intervals. The data thus obtained is put in the form of either if-then rules in the knowledge base or in the form of question and answer in the database table, whichever format the data suits to. The expert in the particular domain should have extreme expertise in his field and preferably should be an "operational engineer". The knowledge base designed for this application of fault diagnosis and control of Transformers [4] deals with the symptoms, causes and remedies of various faults frequently occurring in different types of Transformers and the specifications and standard ratings of the equipment and also supports on-line data up gradation supporting for controlling faults. The data was collected from an electrical factory in Visakhapatnam, India. It is the data about equipment fault, its diagnosis and remedial measures and the database comprising of data from survey results. All the data collected was first put in tabular form for various types of AC Motors with equipment name, symptoms, causes and remedial measures as columns. Further they were transformed into decision trees and then as rules in the knowledge base of the expert system package. Let us consider a few fault clearance scenarios using the present FDC System. Based on the visual observations and information received from the sensors fitted to the transformer, the operator can easily detect the overheat on magnetic coil in the equipment i.e. more than the prescribed value (the prescribed value can be known from the application database); the operator can use the system and select for the transformer and symptom high temperature rise. This in turn leads the package seek information about the type of the transformer which can be known from the application database. The inference engine works on the input in relation with the knowledge base and finally displays the solution for clearing the fault or problem as reducing load and increasing power factor. In this way any fault related to various transformers can be very easily cleared with the assistance of the Fault Diagnosis and Control (FDC) system. Extensive data has been collected for power transformers, distribution transformers of all types from the experts, books, historical data and regular on-site survey and stored in the knowledge base and database of the package. The data is first transformed in the form of tables, decision trees and then as knowledge base rules and database queries for Transformer. The developed package can always be upgraded with new equipments very easily as extensibility is the main feature of our web based expert system package. The web based expert system package (FDC) developed can be hosted on web and

can be supported on both internet and intranet. Hence the user can use the package from anywhere round the globe. It is very user friendly. It starts with the login page asking for user id and password and also allowing for the registration of a new user. The login page connects the user to the main page displaying the equipment to be selected. This is a dynamic page which updates the images and list of equipments to the user as and when uploaded by the administrator. A click on the image of the required equipment redirects the user to the respective equipment page with options to select for either the knowledge base or the database of the selected equipment. On selecting the knowledge base the user enters into a question and answer session where the user selects for the symptoms based on the physical observations made for the fault. The session ends up finally providing the user with the remedy for the occurred fault. On selecting for the database the user is provided with a search engine mechanism where the user can get the details of the equipment specifications, ratings and other details by typing the keyword related to the requirement and the package places all the related data from the equipment database. The package gives complete freedom to the user to switch over from one equipment to other and to switch from the knowledge base and the database. The login with the administrator id and password links to the admin page where uploading of new knowledge bases and updating of existing databases can be easily done. The package once hosted on web for internet services or on local host for intranet services is completely available to the client (user) without any restrictions on the availability of the building technology on the client system. The user can always use the package in any web browsers like IE5 or Mozilla. etc. as the package is totally user friendly the user is not required to know any technology to work with the package and use it for fault diagnosis and control except for typing the URL to get to the login page. Snapshots of the developed on-line and off-line FDC System for Transformers are given below:

Fig. 4. Snapshot2 of the developed FDC System

IV. CONCLUSIONS The design and development of a web based expert system shell and its role in developing an on-line and off-line Fault Diagnosis and Control (FDC) system for all types of Transformers is presented in this paper. A brief description of expert system architecture [5] and issues involved in developing a web based expert system shell and the technology used is discussed. The concept of designing a web based expert system with a user friendly GUI is also discussed. The application of the shell to develop the package FDC system for fault diagnosis and control of Transformers which provides on-line and off-line help for diagnosing faults of various kinds of Transformers with various ratings and specifications and clearing them is discussed in detail in the paper. The package deals with data collected from an electrical factory in Visakhapatnam for power transformers, distribution transformers and various other type of transformers. The package can be easily extended to other equipments also. The package FDC system for Transformers is developed on the basis of expert's past experiences in fault clearance and the regular up gradation of data obtained from on-line survey. Thus it renders a cost-effective solution in clearing a fault that may occur again. The system can be used on a stand alone system and hence work off-line where the on-line connectivity is not possible and as it is a web based application anybody from anywhere can use it for fault diagnosis and control provided he is in the network. The package is very user friendly and is easily extensible.

V. REFERENCES Periodicals:
Fig. 3. Snapshot1 of the developed FDC System [1] A Moyes, J R Mcdonald, The Use of Intelligent Systems Techniques in Condition1 Monitomng of Electrical Plant Q3 1997 The Institution Of

5 Electrical Engineers, Printed and Published By The Iee, Savoy Place, London Wcpr Obl, Uk

[2]

Incorporating Expert Systems into web based environment from proceeding (385) Artificial Intelligence and Soft computing-2003

[3]

C. Han and S. Gerasenko (USA) AM. D. Judd, S. D. J. Mcarthur, Intelligent Condition Monitoring and Asset Management Partial Discharge Monitoring For Power Transformers, Power Engineering Journal December 2002

Papers from Conference Proceedings (Published):


[4] Miine,R, "Fault Diagnosis & Expert Systems, The 6lh International Workshop On Expert Systems And Their Applications. Avington, France K.S, Swarup and H. S. Chandrasekharaiah, "Diagnostic Reasoning Based On Structure and Behavior", Conference: Power System 2000

[5]

VI. BIOGRAPHIES
M Babita Jain is PhD student at the International Institute of Information Technology, Hyderabad, India since 2006. The areas of interest include IT Applications to Power Systems, Multi Agent Systems. She has actively involved in establishing IEEE student branch and Women in engineering affinity group activities.

M B. Srinivas is Associate Professor at the International Institute of Information Technology, Hyderabad, India where he is also heading center for VLSI and Embedded systems. He is a very active IEEE member and presently he is Chairman of IEEE Hyderabad section.

Amit Jain graduated from KNIT, India in Electrical Engineering. He completed his masters and Ph.D. from Indian Institute of Technology, New Delhi, India. He was working in Alstom on the power SCADA systems. He was working in Korea in 2002 as a Postdoctoral researcher in the Brain Korea 21 project team of Chungbuk National University. He was Post Doctoral Fellow of the Japan Society for the Promotion of Science (JSPS) at Tohoku University, Sendai, Japan. He also worked as a Post Doctoral Research Associate at Tohoku University, Sendai, Japan. Currently he is an Assistant Professor in IIIT, Hyderabad, India. His fields of research interest are power system real time monitoring and control, artificial intelligence applications, power system economics and electricity markets, renewable energy, reliability analysis, GIS applications, parallel processing and nanotechnology.

Vous aimerez peut-être aussi