Forecasting the reliability of components in thermal power plants using the VGB database KISSY

摘要

Power plant failure can be costly and any effort should be made to prevent such failures. Methods using Reliability Analysis Maintainability (RAM) tools are well known to answer questions on target values for reliability and availability of a plant, capacity expansion planning, optimum redundancy, maintenance optimisation, etc. However, such RAM tools need high-quality reliability data. These data are scarce for power plant components, therefore VGB PowerTech e.V. decided on a research project to arrive at such data from its KISSY database. While the yearly standard VGB reports sufficiently show which systems and components in thermal power plants are prone to unavailability, forecasting the reliability of components is not optimum as the number of components at risk is not taken into account. Therefore, one cannot determine the failure probability per unit time, the average repair time etc. for sub-systems and components from these reports. Also some components have HILP = High Impact Low Probability characteristics and, while the contribution to plant unavailability is in the standard report, for such components due to the low probability of occurrence, the frequency of occurrence is not in the reports.