Thermal nuclear reactor power is typically evaluated with secondary system calorimetric calculations based on feedwater flow rate measurements. The feedwater flow rate should therefore be measured accurately.
Most pressurized water reactors(PWRs) use Venturi flow meters to measure the feedwater flow rate. However, fouling phenomena, which make corrosion products to accumulate and increase the differential pressure across the Venturi flow meter, can result in an overestimation of the flow rate.
Therefore, in this thesis, a support vector regression (SVR), a fuzzy support vector regression (FSVR) and a fuzzy neural network (FNN) model has been developed in order to accurately estimate online the feedwater flow rate, and also to monitor the status of the existing hardware sensors. Also, the data for training the data-based methods are selected by using a subtractive clustering scheme to select informative data from among all acquired data.
The proposed inferential sensing and monitoring algorithm is verified by using the acquired real plant data of Yonggwang Nuclear Power Plant Unit 3. In the simulations, since the root mean squared error and the relative maximum error are so small and the proposed smart software sensor early detects the degradation of an existing hardware sensor, it can be applied successfully to validate and monitor the existing hardware feedwater flow meters.