Maintenance Strategies for Major Accident Prevention

Abstract

The overall objective of this PhD thesis has been to develop new strategies for the prevention of maintenance-related major accidents in the process industries. By virtue of the new knowledge developed in this PhD project, the decision-makers are expected to gain a better insight into the pros and cons of maintenance, how maintenance influences major accidents, the maintenance-related major accident trend, the degree and distribution of the causes of maintenance-related major accidents as well as strategies for the prevention of such accidents.

This PhD thesis has been a mix of qualitative and quantitative analysis and synthesis of concepts and theories in relation to major accident causation (e.g. organizational accident perspectives and previous accidents investigation/analysis reports) and preventive initiatives in relation to maintenance-related organizational robustness and resilience as well as maintenance optimization (aimed at minimizing the major accident risk).

This PhD thesis investigates all the aforementioned aspects of maintenance-related major accidents in the hydrocarbon and chemical process industries, and significant results have been achieved as regards a better understanding of the characteristics of maintenance-related major accidents and the most reasonable preventive efforts. The main contributions of this PhD project to the body of knowledge are documented in the form of six articles, five published and one under review in international peer-review publication channels. Maintenance-related major accidents, like other major accidents, can be caused by technical, human and organizational factors. However, some individuals and organizations may tend to consider the technical and human factors only, while neglecting the rest. Since accidents are not only products of technical/human failures, it is important to consider all the possible types of influences in order to develop a comprehensive prevention strategy. In this PhD thesis, a better insight is being given into how maintenance influences major accidents with respect to the aforementioned types of influences.

Maintenance-related major accidents may be classified based on existing schemes encompassing maintenance work/management process, threats and error management (TEM) framework and MTO (Man, Technology and Organization) etc., but these fail to explicitly address the accident process itself. In this PhD thesis, how maintenance influences the major accident risk has been analyzed and investigated. The work and accident processes combined into a new scheme called Work and Accident Process (WAP) scheme is being proposed for maintenance-related major accidents classification. This scheme has been tested with maintenance-related major accidents cases from the U.S. Chemical Safety Board (CSB) and the validity has been demonstrated. The PhD thesis also presents a classification approach based on both the accident process and the maintenance management process. Statistics on maintenance-related major accidents are important for decision-making, most especially, in relation to the prioritization of preventive efforts. However, most of the statistics identified on maintenance-related accidents are about 25 years old. The validity of such statistics is constrained by the uncertainty associated with the assumption that future failure will continue at the same rate as immediate previous experience. Besides, some of the statistics were a mixture of major accidents, occupational accidents and serious incidents and this is also adds to uncertainties in relation to maintenance-related major accidents. This PhD thesis provides a current update on the trend of maintenance-related major accidents in the 21st century and the degree and distribution of the causes. The thesis also proposes WAP-FMEA (Work and Accident Process Failure Modes and Effects Analysis), i.e. a FMEA which integrates the maintenance work process with the accident process for the purpose of prevention of maintenance related major accidents. As regards the applicability of the statistical findings, the frequencies can be used to determine probabilities which in turn will be useful in maintenance-related, major-accident risk modeling and in the probability column of the suggested WAP-FMEA. The purpose of maintenance should not be limited to the traditional idea of retaining systems in or restoring them to a functioning state. Maintenance can also contribute to improved system knowledge and inter-discipline coordination that may benefit the entire organization. The organizational value-adding potential of maintenance has never been investigated. Maintenance can be investigated as a contributor to robustness and resilience of organizations such that the ability to prevent or limit unexpected events is improved. Hence, maintenance-related major accidents (i.e. maintenance-related organizational accidents) can also be viewed from the six perspectives of organizational accidents, i.e. Energy-Barrier model, Normal Accident Theory (NAT), High Reliability Organizations (HRO), the Man-made Disaster (MMD) theory, Conflicting Objectives and Resilience Engineering. These perspectives have been studied by several people, but have not been investigated thoroughly by any in relation to maintenance. This PhD thesis systematically investigates the relationship between maintenance and the six perspectives of organizational accidents. This encompasses the significance of the perspectives to maintenance. Finally, recommendations on how to improve the robustness and resilience of maintenance within the process industries evolve from the aforementioned investigation. This is an added knowledge for a better insight into the link between maintenance and major accident Increasing or decreasing frequencies of maintenance increase the major accident risk when they are off the optimal level. Too much maintenance visits increase the exposure of personnel to risk, the probability of introducing new hazard and failures and the wear-out potential of safety barriers, whereas too little maintenance visits provide an opportunity for failure mechanisms to degrade the major accident barriers. Hence, there is the need to optimize maintenance. However, none of the existing optimization methods has accounted for the exposure of personnel to the major accident risk. This PhD thesis improves maintenance optimization in terms of risk such that the exposure of humans to risk is being adequately accounted for.