Benchmarking the NEM Real-Time Core Model for VVER-1000 Simulator Application: Asymmetric Core

Abstract

A real-time version of the Nodal Expansion Method (NEM) code is developed and implemented into Kozloduy 6 full-scope replica control room simulator. Combined with an enhanced thermal-hydraulics and I&C models the whole package is a high-fidelity simulation tool for operator training and various other applications. The fidelity and accuracy of simulation with emphasis on reactor core model is illustrated through comparison with plant-specific data. The transient of ‘Switching-off of One of the Four Operating Main Circulation Pumps at Nominal Reactor Power’ as described in OECD/NEA Kalinin 3 Coolant Transient Benchmark is an example of an asymmetric core scenario with a range of parameter changes. Simulation results concerning fuel assembly power and axial power distribution during the transient are compared with records from Kalinin 3 in-core monitoring system. Main operating parameters of nuclear steam supply system of a VVER-1000/V320 series units vary to a considerable degree. While Kalinin 3 benchmark specification contains very good description of the transient, as well as record of many parameters of the unit, the document provides only superfluous description of the reference unit. In such a case, an approach based on a ‘generic’ V320 model by default introduces deviations which are difficult to quantify. There are several examples which warrant discussion. One example is core coolant flow and pressure loss during the transient. Pump head and pressure loss across reactor vessel are measured and recorded and in-core monitoring system provides estimation of core coolant flow, which is quite high in comparison with some other V320 units (e.g. by about 5 % larger). Without more detailed pressure loss data across the main circulation loop and specific pump characteristics, however, one can only guess how much simulation is off the mark. Another detail of the same problem is coolant flow through a specific fuel assembly. The presence of a fuel assembly of different design (TVS-M type) surrounded by TVSA type fuel assemblies shall be thoroughly considered, because secondary sources indicate significant differences in fuel assembly pressure loss coefficients between the two types. Coolant flow affects coolant (and fuel) temperature profile and thus neutron cross-sections. Yet another example, even more strongly affecting our ability to interpret simulation results is core power reconstruction provided by the in-core monitoring system of the unit. The SPND (Self-Powered Neutron Detector) current readings are subject of conversion by an algorithm based upon simulated spatial neutron flux distribution across the reactor core. While error estimation of the parameters in stationary conditions is available from secondary sources, there is no reliable estimation of error magnitude during the transient.