Reliability analysis and cost modeling of degrading systems

Degradation is an on-going process in systems, equipments and components subjected to various stresses and adverse operating conditions. The factors influencing the degradation process may adversely affect the system or component performance. A study of these factors will provide a basis for making correct decisions concerning corrective and preventive measures. Such a study is also useful for reliability analysis of the degrading systems/components for making maintenance decisions or for initiating measures for changes in the design. Many times it is not possible to implement design changes due to complexities and cost considerations, as in the case of railway infrastructures, etc. In such situations operational reliability is assured through effective maintenance actions. Knowledge of the technical condition of components is important to achieve the optimal maintenance policy in order to minimize the total system risk. A methodology for rail defect prioritization and risk assessment is developed to support the decision- making process during the effective scheduling of the inspection frequency based on the type of defect and its risk of developing into a rail break. The present research work also demonstrates an application of reliability analysis to improve system reliability based on design changes. This has been illustrated with an example from a manufacturing industry. The objective is achieved using cost-benefit analysis in combination with failure data and root cause analysis. The analysis of failure data with the different cost elements involved in the operation and maintenance of the complex systems is presented as a basis for choosing between alternative designs. Furthermore, an optimization model has been developed to estimate the optimum inspection frequency required at the minimum maintenance cost based on the technical condition of the component. The model has been validated by a case study of an offshore oil and gas platform. The consequences of not choosing the right distribution have also been discussed in the thesis. The concept of the virtual failure state has been introduced to estimate the failure distribution of highly critical components. The factors influencing the degradation process have been identified and studied in detail. A framework for classification of rail failure data has also been developed for rail life estimation using reliability analysis. In short, the thesis discusses the application of reliability analysis and cost modeling techniques to support the decision-making process in operation and maintenance activities and demonstrates its usefulness in real life.Godkänd; 2008; 20080328 (ysko)