Stochastic Kalman filtering for predicting the service life of a steam generator in a nuclear power plant

Determining the mechanism of damage accumulation in a material taking account of the statistical data remains an urgent problem when making service-life predictions for power equipment. The method of statistical prediction that is currently in use has substantial drawbacks: it is impossible to make predictions under varying operating conditions [1]. Therefore, the solution must be sought in the form of a procedure that takes account of the nature and randomness of the aging process. In addition, the data provided by monitoring give important additional real-time information. The operability of equipment such as the VVER steam generator is determined, first and foremost, by the integrity of the tube sheet. The metal of the heat-exchange tubes is monitored with a certain error. The statistics of plugged tubes include data on tubes not only with through defects but also with a metal shortage above 70%. This information can be used as a basis for making a probabilistic assessment of the service life of a tube sheet and determining the optimal plan for preventative maintenance. In the present paper, it is suggested that Kalman's method of stochastic filtering be used to solve this problem [2].