DEVELOPMENT OF SOA BASED AUTO-MATED TESTING METHOD FOR SUBSTATION AUTOMATION
Ambaw, Tewodros Woldetensaie (2018)
2018
Kuvaus
Opinnäytetyö kokotekstinä PDF-muodossa.
Tiivistelmä
The thesis focuses on the development of an SOA based automated testing method for substation automation use in control systems. As an outcome of the research, a software called TestBase is developed. TestBase follows SOA model where individual service providers or services are implemented to deliver services to other services or components via a communication protocol. All these services communicate and exchanging messages using a C# RabbitMQ AMQP connector library, MML. MML allows services to ex-change messages independently by providing simple way to publish services to service bus from C# code. Google ProtoBuf is used for serialization in this SOA framework for its great advantage in terms of performance and message size for conversion of data structures into stream of bytes to store them in database.
TestBase can be used for system, functional and FATs in control systems to make sure that protection devices can protect against overcurrent, directional overcurrent, over-voltage etc. by following white box, black box and gray box testing methodologies.
This research also suggests to use BIST (mechanism in which a program tests itself) and machine learning algorithms in both SUT and test system. In BIST, services formed in SOA ought to contain a procedure or method that performs self-tests e.g. to test availa-bility of specific hardware or configuration. Machine learning can be applied with BIST to learn, predict and correct during unexpected test outcomes.
TestBase can be used for system, functional and FATs in control systems to make sure that protection devices can protect against overcurrent, directional overcurrent, over-voltage etc. by following white box, black box and gray box testing methodologies.
This research also suggests to use BIST (mechanism in which a program tests itself) and machine learning algorithms in both SUT and test system. In BIST, services formed in SOA ought to contain a procedure or method that performs self-tests e.g. to test availa-bility of specific hardware or configuration. Machine learning can be applied with BIST to learn, predict and correct during unexpected test outcomes.