A methodological approach to identifying destructive information actions on a distributed process control system

Abstract

The aim of this study is to improve the methodological support for evaluating a possibility of implementing a sequence of information actions on a distributed control system leading to emergency situations. The problem of evaluating a possibility of implementing a sequence of information actions on a distributed control system leading to emergency situations is solved. A methodological approach is proposed to identify information actions sequences leading to an emergency situation at industrial facilities, based on the use of the method of time differences of machine learning with reinforcement. The effectiveness of the considered methodological approach in comparison with the method of equal probability search is estimated. An example of the implementation of the methodological approach in relation to the technological process used at industrial facilities is given.

Keywords

• industrial facility
• an emergency situation
• information actions sequence
• distributed control system.