Data set about Gas turbine engine monitoring system based on digital thermodynamical twin
Description
This study presents the design of an advanced Gas Turbine Engine Monitoring System intended to enhance engine reliability, operational safety, and predictive maintenance capabilities in modern aviation. The proposed system integrates thermodynamic modeling, structural analysis, real-time data acquisition, and machine learning algorithms to enable intelligent fault detection and remaining useful life estimation. A comprehensive thermodynamic model of a turbojet engine was developed using MATLAB/Simulink to simulate engine performance across a wide range of operating conditions and fault scenarios. In parallel, finite element analysis was conducted to assess thermal and mechanical stress distribution in critical components, such as combustion chamber, turbine blades and compressor discs, under realistic flight loads. Experimental validation was performed using a hardware-in-the-loop simulation environment, which emulated real-time sensor behavior, including vibration, pressure, temperature, and oil quality metrics. Data from fault injection tests were used to train supervised machine learning models with Scikit-learn and TensorFlow, achieving high classification accuracy for detecting degradation patterns. Additionally, a digital thermodynamic twin architecture and a wireless data transmission protocol were integrated to support real-time ground station diagnostics. This research contributes to the evolution of engine health management systems by combining physics-based modeling with data-driven intelligence for aerospace applications
Files
Institutions
- Nacional'nij tehnicnij universitet Ukraini Kiivs'kij politehnicnij institut imeni Igora Sikors'kogo