Soft-Sensing for the Feedwater Flowrate at PWRs Using a GMDH Algorithm

Currently, Venturi flow meters are being used to measure the feedwater flowrate in most pressurized water reactors (PWRs). These meters can induce measurement drift owing to corrosion product buildup near the meter orifice through long-term operation. Fouling of the Venturi meter decreases the accuracy of the existing hardware sensors. These requirements result in nuclear power plants operating at lower-than-planned power levels. Therefore, considerable research has been focused on resolving the inaccurate measurement issue of the feedwater flowrate. This study employed the Group Method of Data Handling (GMDH) which is one of the data driven models, such as Artificial Neural Networks (ANNs), to increase the thermal efficiency by accurately estimating online the feedwater flowrate. The GMDH method was trained using the informative data selected through a subtractive clustering (SC) scheme. In addition, the uncertainty of the GMDH algorithm was analyzed using a statistical uncertainty method