In order to decrease the fuel consumption, a new flight mode is being considered for twin-engine helicopters, in which one engine is put into sleeping mode (a mode in which the gas generator is kept at a stabilized, sub-idle speed by means of an electric motor, with no combustion), while the remaining engine operates at nominal load. The restart of the engine in sleeping mode is therefore deemed critical for safety reasons. This efficient new flight mode has raised the interest in the modeling of the restart of a turboshaft engine.

In this context, the initial conditions of the simulations are better known relative to a ground start, in particular the air flow through the gas generator is constant, the fuel and oil system states are known and temperatures of the casings are equal to ambient. During the restart phase of the engine, the gas generator speed is kept at constant speed until the light-up is detected by a rise in inter-turbine temperature, then the starter torque increases, accelerating the engine towards idle speed. In this paper, the modeling of the acceleration of the gas generator from light-up to idle and above idle speeds is presented. Details on the light-up process are not addressed here.

The study is based on the high-fidelity aero-thermodynamic restart model that is currently being developed for a 2000 horse power, free turbine turboshaft. In this case, the term high-fidelity refers not only to the modeling of the flow path components but it also includes all the subsystems, secondary air flows and controls with a high level of detail.

The physical phenomena governing the acceleration of the turboshaft engine following a restart — mainly the transient evolution of the combustion efficiency and the power loss by heat soakage — are discussed in this paper and modeling solutions are presented. The results of the simulations are compared to engine test data, highlighting that the studied phenomena have an impact on the acceleration of the turboshaft engine and that the model is able to correctly predict acceleration trends.

This content is only available via PDF.
You do not currently have access to this content.