Improvement of the Method for Determining the Operating Mode of a Gas Turbine Engine in a Gas Pumping Unit

In this paper, we address the issue of improving the precision of determining the operational mode of a gas turbine engine (GTE) within a gas pumping unit (GPU). Traditional graphical characteristics exhibit limited accuracy and are time-consuming to apply, while analytical models require estimating the GPU load, which is a complex task. To overcome this challenge, we propose utilizing machine learning techniques to predict the rotational speeds of low- and high-pressure rotors. Parameters such as the centrifugal compressor characteristics and environmental conditions are used as features. This approach eliminates the reliance on manual analysis of graphical characteristics, thereby simplifying the determination of operating modes for gas turbine engines and enhancing accuracy and efficiency. The study focuses on identifying key parameters, including pressure and temperature at the intake and exhaust of the centrifugal compressor, the rotational speed of the power turbine, and environmental factors such as temperature and pressure. This integrated forecasting methodology enables more accurate predictions of the GTE's operational parameters.