Catalog Home Page

Integration of operational tasks in chemical plants

Nikraz, Magid (2007) Integration of operational tasks in chemical plants. PhD thesis, Murdoch University.

[img]
Preview
PDF - Front Pages
Download (100kB) | Preview
    [img]
    Preview
    PDF - Whole Thesis
    Download (6MB) | Preview

      Abstract

      The overall, coordinated management of different operational tasks in a chemical plant can improve operational efficiency. These operational tasks can be hierarchically categorised, from the lowest to highest level, as: data acquisition; regulatory control; monitoring; data reconciliation; fault detection and diagnosis; supervisory control; scheduling; and planning. Although each of these tasks is responsible for a particular function, they are dependent on each other, which is why an approach wherein all the different tasks can be integrated into a single unified framework is desirable. While integration has important benefits such as a significant reduction in operator workload and improved decision making, its realisation presents considerable challenges. Few previous works have addressed this topic and even fewer have investigated recent computing paradigms which may greatly assist in the development of a unifying framework.

      Multi-agent systems were introduced and investigated in this study as a possible means for achieving integration of operational tasks in chemical plants. Multi-agent systems are the subject of a sub-field of computing research known as agent-based computing. Agent-based computing represents a relatively recent and powerful high-level computing paradigm.

      Initially, a number of software applications were developed for the purposes of this study to assist realisation of the operational tasks. To simplify the process of system development and provide guidance for those unfamiliar with multi-agent systems wishing to adopt the proposed technique, an extensive methodology was devised. The operational tasks were then integrated using the proposed methodology to form an integrated multi-agent system, with the pilot plant at Murdoch University being used as a test base for the solution. The results were positive and demonstrated that the proposed agent-based solution was able to effectively account for the pilot plant setting. It was concluded that, in addition to presently available integration techniques and base technologies, the agent-based approach to integration of operational tasks in chemical plants presents a viable alternative solution.

      Publication Type: Thesis (PhD)
      Murdoch Affiliation: School of Engineering and Energy
      Supervisor: Bahri, Parisa
      URI: http://researchrepository.murdoch.edu.au/id/eprint/741
      Item Control Page

      Downloads

      Downloads per month over past year