DIAGNOSTIC DIAGRAM OF STATE OF WELL MECHANICAL CLEANING UNIT

Relevance Approximately one fifth of oil reserves of the Russian Federation belong to oil with high content of asphalt-resin-paraffin deposits (ARPD). When thermobaric conditions in a producing oil well change (temperature and pressure drop), ARPD crystallize from oil and precipitate on tubing walls. This leads to a decrease in their working cross-section and oil flow rate, an increase in the amount of consumed electricity, unproductive downtime of wells and, in general, reduces the efficiency of oil field operation. That is why wells are cleaned from ARPD at certain intervals, most often mechanically with the help of various scrapers. Aim of research The main aim is to develop a diagnostic diagram of the state of the mechanical well cleaning unit from ARPD in real time and to choose a type of mathematical model for the digital twin of the unit. Research methods For realization of the set goal in the article the working conditions and possible states of the mechanical well cleaning unit from ARPD are considered. A comparative analysis of several methods of mathematical models synthesis most often used in the oil and gas industry is carried out. Results The diagnostic diagram of the unit state in real time is developed, which allows estimating the probability of occurrence of specific abnormal or emergency situations. It is established that the most expedient methods for modeling of the state diagnostics for mechanical cleaning unit from ARPD are methods based on fuzzy logic using Mamdani algorithm and neural networks of multilayer perseptron class.