Agent-Augmented Process Automation System

of The Master's Thesis The requirements for process automation systems have changed much since the adoption of automatic control. To survive in tightening global competition the developers of automation systems are introducing new tools to increase flexibility and fault-tolerance in process automation systems. This thesis proposes that agent technology as a new programming paradigm could be utilized in conjunction with the existing process automation system to meet the future requirements of automation. Agents are autonomous cooperative entities, having either purely software or physical appearance, that are executing predefined tasks. To study these issues the concept of agent-augmented process automation system has been introduced. The focus of this thesis is in the practical implementation of the concept to the laboratory test environment, which is used to test and demonstrate the technological feasibility of the agent approach. With agent technology it is possible to tackle the increasing complexity of automation systems, and agent negotiations can be utilized in a number of process automation system functionalities. This thesis begins with an introduction to agent technology and the FIPA agent standard. A short review of generic agent applications and modern-day automation solutions that uses agent technology is presented. The concept of agent-augmented process automation system, and potential exploitation of agent technology in a number of process automation system functions are analyzed. The integration of currently available agent technology implementations and conventional process automation systems require a number of technical problems solved, and these issues are discussed in this thesis. Furthermore, to validate the achievements of this thesis, an initial test run of agent-augmented fault handling application has been performed with the laboratory test environment. As a result of this research, knowledge of agent technology applied to process control domain has been gained and an operational agent-augmented process automation system was set up. In this thesis, the agent intelligence was based on reactive rules, which does not scale up easily to complex applications. Although the implementation work itself was relatively straightforward a number of open questions remains for future examination. Preface This thesis and work behind it have been conducted at the Automation Technology Laboratory of the Helsinki University of Technology during the year 2002 in a National Technology Agency (TEKES) funded project " Agent-based automation solution " (Agenttipohjainen automaatioratkaisu). I would like to thank all the project partners for the great opportunity to do interesting research. I wish to express my sincere gratitude to …