Component-Based Modeling of Turboprop Performance Loss Under Altitude and Thermal Variation

This is the authors' abstract. We don't add key points for this paper.

Accurate prediction of turboprop propulsion system performance under high-altitude conditions is critical for optimizing aircraft mission profiles and increasing operational efficiency. This study presents a dynamic and component-based digital twin model of a single-stage turboprop engine developed using the Siemens AMESIM environment. The developed model enables high-accuracy out-of-design steady-state performance analyses and offers a dynamic and low-cost approach for evaluating engine-aircraft integration. In this study, the effects of altitude and temperature variations on gas generator speed ([Formula: see text]) were investigated under four different operational scenarios: high-temperature–variable-altitude conditions, low-temperature–variable-altitude conditions, variable-altitude conditions under constant high temperature, and variable-temperature conditions under constant altitude. Through these analyses, performance losses due to altitude-dependent air density and temperature variations were quantitatively evaluated. In the model validation process, a comparative analysis was performed with open-source reference aircraft data, demonstrating the model’s strong predictive capabilities. Simulation results demonstrated high accuracy, with an error rate of 0.6%, particularly under medium-altitude conditions (14,000 ft, 253.15 K). Deviations observed at higher altitudes (reaching up to 9.82% at 18,000 ft) were attributed to unmodeled thermodynamic effects of auxiliary cooling systems. Thus, subsystem effects were isolated, and their impact on performance was analyzed. The results show that the AMESIM-based component-based digital twin approach is an effective tool for modeling turboprop engine performance losses. Furthermore, this study provides a significant foundation for the future integration of detailed thermal management and auxiliary system models.