Analysis of VVER-1000 main circulation pump condition in operation

The paper presents a method and algorithms for detecting abnormal conditions of the main circulation pumps (MCP) of NPPs with VVER-1000 reactors based on their in-process testing results. The methodological basis for the algorithms is the presentation of the nuclear power plant equipment as a complicated system described by the N-dimensional vector in the space of its conditions. A large number of process parameters describing the equipment condition using the Karhunen-Loeve transform is reduced to a much smaller number of informative criteria and presented in the form convenient for an analysis. The effectiveness of the method has been demonstrated in detecting the MCP abnormal behavior at the power units of Kalinin and Novovoronezh NPPs. The method and algorithms developed for monitoring the VVER-1000 MCP condition make it possible to detect an abnormality based on the pump operating data long before it is detected by the regular control systems.