Condition monitoring for maintenance support

BEERE William¹, BERG Øivind¹, WINGSTEDT Emil¹ SAVOLAINEN Samuli², and LAHTI Tero²

1. Institute for Energy Technology, OECD Halden Reactor Project, Norway(oivind.berg@hrp.no)

2. Fortum Power and Heat Oy, Loviisa Power Plant, Loviisa, Finland

Abstract: Considerable advancement has been made in computer and information technology that can benefit safety and economy in Operation and Maintenance (O&M). However, before implementing new technology in nuclear power plants there is a need for qualification of methods and related tools. The OECD Halden Reactor Project (HRP) has taken an active role in facilitating implementation of technology advances and in particular application of condition monitoring techniques for maintenance support.

TEMPO ^[1] is a system based on physical models for thermal performance monitoring and optimization developed at the HRP. The system aims at satisfying information needs associated with condition monitoring, on-line calibration monitoring of plant measurements, process fault detection and diagnosis.

The data-reconciliation ^[2] method used in TEMPO relies on fitting a simulation of the turbine cycle to the actual plant data. The difference between measurements and calculated values (residuals) are monitored to detect deviations. Each measurement point is assigned an uncertainty. How well the simulation fits to the measurements is compared to the given uncertainty. Traditionally this comparison is directly used to determine if there is a fault in the measurement.

By using a time series analysis of plant data, changes below single point statistical significance can be found. Variations in both individual residuals and the global object function, i.e. the sum of all residuals, are small and their values mostly static. Thus, trending the global object function value is important in order to identify possible faults. Comparing residuals with past behaviour enhances fault detection compared with a statistical analysis of each data point.

An example of fault detection is given from the analysis by TEMPO of data from the Loviisa 2 NPP in Finland.

Keyword: condition monitoring, maintenance support.