An approach for the development of the methodology for automated determination of the state of an emergency NPP unit with VVER-1000 based on the analysis of numerical values of a limited list of technological equipment parameters is proposed.
The advantage of automated determination is that it is possible to obtain accident developments forecast before arrival of experts. This will significantly facilitate experts’ work and allow to assess the current state of the emergency power unit and to obtain preliminary forecast results.
Possible ways to solve this problem as well as their advantages and disadvantages are presented. Based on the analysis of these methods, a methodology on the basis of the finite state machines is proposed.
The methodology is based on dividing the task into separate independent subtasks. Each subtask is an assessment of particular equipment state (pump, diesel generator, steam dump valve, etc.). The logic of equipment state assessing is developed on the basis of finite state machines application, considering expert experience.
The methodology of a finite state machine diagram developing for determining the diameter of a small leakage from the primary circuit is presented as an example. The primary circuit leakage identification is based on the simultaneous presence of two phenomena – primary circuit pressure decreasing and pressure and temperature increasing in the containment. Prevention of false leakage identification is carried out by the pressure and temperature insensitivity range in the containment. The leakage diameter is estimated by interpolating the primary pressure reduction rate table, obtained by calculation. The developed finite state machine allows to determine the diameter of a small leakage in the range of 20–90 mm with an error of up to 15 %.
Keywords: VVER-1000, leakage, methodology, neural network, finite-state machine, emergency response.
Article language: Russian. Pp. 40–52. DOI: 10.26277/SECNRS.2021.100.2.004.
