The established approach to safety management has failed to handle socio-technical systems that have become more complex. The main argument is this approach is based on assumptions that systems are protected against accidents by barriers (well-trained people, redundant mechanisms and safety devices, and procedures and safe systems of work). Complex systems, such as maintenance, are actually labour intensive; maintenance staff often works under pressure to finish tasks as rapidly as possible. They continuously adapt and make adjustments using available resources, time, knowledge, and competence to achieve success. Thus, they are accidents prone. Human factors inherent to maintenance accidents are most times difficult to identify. Research in this area in the oil and gas industry in maintenance management is limited in comparison to the aviation and nuclear sectors. Therefore, it has been suggested to overcome this lack by exploring the maintenance system and identifying appropriate methods and tools that lead a system to safety excellence. Resilience engineering (RE) approach has been found the suitable solution. Moreover, four system abilities (cornerstones of RE: ability to respond, to monitor, to anticipate, and to learn) have been identified to characterise the resilience of a system; if these abilities are known and increased, it will make the system As High Resilient As Possible (AHRAP). However, there is a need to bridge between RE theory and practice. Particularly, a tool that measures these abilities lacks in the oil and gas industry, specifically within the maintenance system. In doing so, a framework based on a Gap Analysis (GA) was outlined. A tool, the MAintenance System Resilience Assessment Tool- MASRAT, was developed to assess current system resilience and identify strategies for improvement to achieve safety excellence. The maintenance system of SONATRACH was explored by the analysis of the system documentation and processes, interviews with maintenance staff, questionnaires, field observations, storytelling, and functional analysis. MASRAT has been validated by means of congruency and principal components analysis, PCA (content validity), and Cronbach’s alpha (reliability). An expert panel testing was carried out to test its usability. The exploration of the system came up with a snapshot of daily activities as well as a better understanding of the maintenance system. The study identified the most significant human factors (resources, time pressure, and supervision/coordination) and their probable impact on plant safety. The elements of the system were found tightly coupled, hence the system complex. Stories describing the continuous adaptations of people to achieve assigned objectives were collected. On the other hand, MASRAT was validated. All items were rated above 0.75 in congruency test. The results of PCA for the three selected factors confirmed the items may be clustered after extraction into four components which interpretation represents the four cornerstones of RE. The analysis showed MASRAT is reproducible. Cronbach’s alpha results were found higher than what is required (0.7). MASRAT was found usable by maintenance expert panel. It was used to measure the maintenance department resilience. Strategies that may lead the system from current maturity level to excellence were identified. Eventually, recommendations were made to management to be implemented both at corporate and department levels. For the first time, the maintenance department resilience of petroleum assets was measured to fill in the gap between RE theory and practice. Besides, this can be of benefit to the petroleum industry by a better knowledge of the maintenance working environment and human factors impact on safety and by profiles determination and improvement strategies identification.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:692982 |
| Date | January 2016 |
| Creators | Ameziane, Said |
| Contributors | Kishk, Mohammed ; Steel, John A. |
| Publisher | Robert Gordon University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/10059/1565 |
Page generated in 0.002 seconds