Predictive state analytics of gas pumping units

Abstract

Diagnostics of the technical condition of equipment is an important factor in the efficiency of any production. This fully applies to the oil and gas industry with its huge fleet of pumping and compressor equipment, the downtime and repair of which, as a result of any accidents, causes significant damage to enterprises. The development of digital technologies, such as Big Data, IIoT, and machine learning makes it possible to move on to the so-called predictive or predictive analytics, which makes it possible, based on the analysis of trends in a large number of variables, in real time to predict the possible onset of an emergency. The article considers the possibility of building a local system of predictive analytics for a gas compressor unit based on the analysis of its behavior before the accident, during the accident and after it. The predictive analytics system (PSA) was built based on the real data of the unit, on which surge occurred due to fatigue failure of the high-pressure compressor elements. The evaluation of the resulting model showed its high accuracy even under conditions of a limited set of initial data. The use of such systems makes it possible to avoid accidents that occur as a result of a change in the technical condition of the equipment and are absent, in connection with this, in emergency protection scenarios.