Estimation of the coolant flow through a natural circulation BWR fuel channel applying and equivalent electrical model

This work presents the design and implementation of an advanced controller for a reduced order model of a BWR reactor core cooled by natural circulating water, which allows real time estimates of coolant flows through fuel assemblies about standard neutron flux strings. Nuclear power plants with boiling water reactors control individual fuel assembly coolant flows by forced circulation using external or internal water pumps and different core support plate orifices. These two elements reduce flow dependency on local channel pressure drops. In BWR reactors using only natural circulation coolant flows, these two elements are not available and therefore individual channel coolant flows are highly dependent in local conditions, such as power distributions and local pressure drops. Therefore it is expected that grater uncertainties in these variables be used during safety, fuel management and other analysis, which in turns may lead to increased operation penalties, such as tighter operating limits. The objective of this work is to asses by computer simulations means to reduce uncertainties in the measurement of fuel assembly coolant flows and eventually the associated penalties. During coolant phase transitions, pressure drops and local power may alter local natural circulation through fuel assemblies and flow estimates can be helped or not by control rod moves. This work presents the construction of an optimal controller for a core flow estimator based on a reduced order model of the coolant going though the reactor vessel components and nuclear core. This model is to be driven by plant signals from standard BWR instrumentation in order to estimate the coolant flows in selected fuel assemblies about a LPRM string. For this purpose an equivalent electrical model has been mathematically developed and numerically tested. The power-flow maps of typical BRW are used as steady state references for this equivalent model. Once these were fully reproduced for steady state conditions, the model was scaled and compared to nominal values of a 4500 MWt ESBWR. The design of the reduced order model controller was mathematically based on the theory of an extended Kalman filter, whose algorithm allows to carry out an advanced control of the system. The estimator uses the equivalent electrical model, which was developed from the system. The estimator uses the equivalent electrical model, which was developed from the analogies between electrical current and voltage with coolant flows and pressure drops. (Author)