Optimierung der zustandsorientierten Instandhaltung von Kraftwerken mit datengestützten Online-Modellen (English shop)

Preview

Extract, PDF (250 KB)
Table of Contents, PDF (80 KB)

As a result of high cost pressure, it is becoming increasingly important for power plant operators to run their plants optimally at every required load point and to detect deviations from the target condition at an early stage. This work presents a unified and transferable methodology for the design and implementation of performance indicators for the online condition monitoring of different components in power plants on the basis of artificial neural networks. By extrapolating the performance indicator or cost trend over time as a result of condition deterioration, it is possible to determine the optimum maintenance time for a component.

To minimise costs, it is important to operate power plants optimally at every load requested. For this reason, an early detection of deviations is necessary. In this work, a unified and transferable methodology is introduced for the development and implementation of performance indicators. Based on artificial neural networks, these indicators are used for online condition monitoring. By extrapolating the performance indicator or cost trend due to degradation, it is possible to determine the optimum time for maintenance for the associated component.