Programming Model of Virtual Indicators for Delayed Cocking Unit

Topic: Programming model of Virtual indicators for Delayed Cocking Unit (DCU). Size of paper: 42 pages along with 11 pictures and 29 tables on them. 29 reference articles have been used during the compilation of the research. Key words: Virtual Analyzer, Machine Learning, Oil refinery. The object of the research was a delayed cocking unit of the refinery at Perm. The research subjects were models for virtual analyzers of light and heavy gasoil along with share of C5+ hydrocarbons in gas. Paper includes introduction, eight chapters as the main body and conclusion with results. Introduction contains a problem description, topicality of the topic and goals and tasks of the research. Chapter 2 states about theory of DCU. Section 3 includes literature overview with applications of machine learning in oil refinement. Chapter 4 presents current state of the virtual analyzers and indicate the problems to investigate. Theory of machine learning algorithms is presented in the section 5. Chapter 6 describes the implementation of the research systematically. Chapter 7 shows screenshots of a dashboard for virtual analyzers. Interim results are presented in the section 8. Conclusion highlights main aspects of the research made along with overall results reached and expected in the future.