Implementation of an External Code for the Control and Protection System of Atucha I Nuclear Power Plant

Abstract

One of the fields of study of paramount importance when modeling the behavior of a nuclear power plant is the control and protection systems. Depending on the scope of the implementation, the degree of accuracy of the control and protection actions may vary from a full spectrum coverage up to an extremely simplified model with respect to the real system installed in the plant. In the light of this consideration, certain tools and approaches might be more suitable for programming the control and protection logics depending on what the ultimate subject of the engineering study is (i. e. operational or safety transients). For instance, RELAP (which is a widely used code in the nuclear industry for performing transient calculations for nuclear installations) offers the possibility of modeling a control system through numerical cards that define generic control components such as arithmetic operations, integrators, lead-lag operators, mathematical functions, etc. Nevertheless, this tool is quite complex to be implemented if a high degree of accuracy is desired. On the other hand, this task could be far more simple to fulfill using a high-level programming language such as C or Fortran. If the protection and control actions are programed in an external code, both RELAP and the external program should be able to exchange information with each other (and eventually others) in order to perform coupled calculations successfully. Taking this last concept into account, an extension of RELAP, developed by TECNA S. A., capable of exchanging information through shared memory and synchronizing the different processes involved with semaphores has been used to perform a coupled calculation between the RELAP extension (modeling the plant) and an ad-hoc program written in Fortran (modeling the control and protection systems). In addition to this, a radiation model between the fuel rods and the coolant channel as well as the computation of the net reactivity were implemented. In this paper a general description of these models are presented as well as a brief description of the exchanged data flow. To conclude, the advantages of implementing this approach versus the one available with RELAP standalone are highlighted.