A Functional Systems Framework and Blended Hazard Identification Methodology to Support Process Diagnosis

Benjamin Seligmann (2011). A Functional Systems Framework and Blended Hazard Identification Methodology to Support Process Diagnosis PhD Thesis, School of Chemical Engineering, The University of Queensland.

       
Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
s4055114_phd_abstract_finalthesis.pdf s4055114_phd_abstract_finalthesis.pdf Click to show the corresponding preview/stream application/pdf 11.21KB 4
s4055114_phd_finalthesis.pdf s4055114_phd_finalthesis.pdf Click to show the corresponding preview/stream application/pdf 3.73MB 12
Author Benjamin Seligmann
Thesis Title A Functional Systems Framework and Blended Hazard Identification Methodology to Support Process Diagnosis
School, Centre or Institute School of Chemical Engineering
Institution The University of Queensland
Publication date 2011-07
Thesis type PhD Thesis
Supervisor Professor Ian Cameron
Professor Katalin Hangos
Total pages 282
Total colour pages 27
Total black and white pages 255
Subjects 09 Engineering
Abstract/Summary Major industrial accidents continue to occur in process industries in spite of advances in risk management and abnormal condition management made over the last 40 years. A change in hazard identification (HAZID) theory and practice is required to better support process diagnosis and improve abnormal condition management, thus leading to a reduction in the number and severity of accidents. To improve HAZID for better supporting diagnosis the scope of HAZID analysis should be widened and knowledge generation and reuse associated with HAZID should be improved. To address these two major issues, a conceptual framework for systematically describing process system operation has been developed to guide the formation of a novel hazard identification methodology able to delivers outcomes with a high coverage of hazards and rich causal knowledge that can be used for supporting diagnosis. The conceptual framework was called the Functional Systems Framework. It allows qualitative knowledge associated with structure-function-goal relationships in process systems to be expressed in a clear and consistent manner. This knowledge describes how components, streams and connections form the structure of the system and how their capabilities describe system function. The capabilities explicitly describe how the actions of components and streams affect system properties to meet the goals of the system. The capability concept is crucial for describing the structure-function-goal relationships of process systems. The Functional Systems Framework was used to guide the formation of a novel HAZID method called the Blended Hazard Identification (BLHAZID) methodology. The primary purpose of the BLHAZID methodology is to generate knowledge to be used for supporting fault diagnosis. Therefore the outcomes of the method must contain a high coverage of hazards, richly describe failure causality and be able to be captured using computer-aided tools. This facilitates effective reuse of knowledge. The BLHAZID methodology takes advantage of the blending of two fundamental approaches to hazard identification: function-driven, such as Hazard and Operability Study (HAZOP), and component-driven, such as Failure Mode and Effects Analysis (FMEA). The BLHAZID methodology utilises the strengths of both approaches and minimizes their weaknesses by blending them together, forming a single approach with increased utility. Explicitly examining both functional and structural failures delivers outcomes with high coverage. The BLHAZID approach can be used to describe failure causality between process system equipment and streams in great detail and hence generate rich causal knowledge. The knowledge associated with the BLHAZID is captured and expressed in a structured language using computer based tools. This knowledge can then be reused for many applications, the most pertinent being diagnosis. Industrial case studies of increasing complexity show how the BLHAZID methodology is applied in practice. A BTX storage system and coke ovens gas bleeder, located at BlueScope Steel Ltd., are analysed to show the basic characteristics of the methodology. The BTX storage and transfer system is a combination of a batch and a continuous system, transferring BTX material periodically to a tanker. The bleeder is a safety relief system that flares coke ovens gas under high pressure conditions and therefore has two major operational modes: under normal pressure conditions and high pressure conditions when flaring occurs. A benzene saturation unit, at BP Refinery (Bulwer Island), is studied to show how the BLHAZID methodology can be applied to more intricate process systems. The benzene saturation unit contains a benzene hydrogenation reactor circuit combined with a control system and a large recycle of material and energy. These industrial case studies showed that the BLHAZID methodology generated outcomes that contained increased coverage and richer causality compared with other HAZID methods, with the extensibility and expressive power of the structured language allowing the effective capture of the BLHAZID outcomes. While the final quality of the outcomes is affected by the strengths of the analysis team, the BLHAZID methodology is indeed more effective at supporting fault diagnosis than other HAZID techniques because of the formal approach of relating capabilities to function and failure, the systematic nature of the workflow, the logical linking between subsystems that allows causality to be traced throughout a process system and the power of the structured language that supports effective reuse of BLHAZID knowledge. This work is a contribution towards a change in the field of HAZID, where detailed theory associated with hazard identification and failure propagation in process systems meets with an increased use of knowledge management tools and practices.
Keyword hazard identification
FAILURE
Function
Blending
Causality
Knowledge representation
System
Additional Notes Colour pages: 23, 61, 81, 100, 103, 123, 124, 126, 130, 145, 146, 149, 150, 156, 160, 161, 163, 168, 175, 192, 193, 194, 196, 200, 280, 281, 282. A3 pages: 196, 200 Note 1 - Pages 196 and 200 are BOTH colour and Landscape Note 2 - There are some pages that are NOT specified here that contain colour, but in terms of communicating the material in the thesis, these unspecified pages can remain black and white. Only the pages specified here should be in colour.

 
Citation counts: Google Scholar Search Google Scholar
Access Statistics: 212 Abstract Views, 16 File Downloads  -  Detailed Statistics
Created: Wed, 02 Nov 2011, 18:05:40 EST by Mr Benjamin Seligmann on behalf of Library - Information Access Service