1. Introduction
Based on information from the International Air Transportation Association [
1], 3.8 billion passengers traveled by air in 2016, which is 8% more than the previous year. The Organization for Economic Cooperation and Development forecasts that air transport CO
emissions will grow by 23 % by 2050, if no measures for their abatement are taken [
2]. Considering this, stringent environmental regulations are already in place with the ultimate goal to cut net emissions to half of the 2005 level by 2050. It is for this reason that engine manufacturers focus on possible ways to increase engine efficiency. At the same time, stationary gas turbines are the only thermal power plant technology capable of delivering both secondary and tertiary control reserve from idle [
3]. The rapid expansion of renewable generation in Europe is expected to double the demand for both reserves in the coming decade [
4]. If one considers that gas turbines are very likely to be able to convert hydrogen into electricity at a large scale, an increase in their efficiency can prove very valuable on the road towards carbon free power generation.
Pressure Gain Combustion (PGC) has the potential to increase the propulsion efficiency of aero-engines and the thermal efficiency of stationary gas turbines. Up to date, detonative combustion processes have been the primary method to realize pressure gain combustion, such as pulsed [
5] and rotating detonation combustion [
6], with the latter gaining more attention. Two alternative approaches are the shockless explosion combustion [
7] and pulsed resonant combustion [
8]. Both use resonant pressure waves in a combustor to realize quasi constant volume combustion. The ideal thermodynamic cycles that model gas turbines with pressure gain combustion are the Humphrey and the ZND cycle, presented in
Figure 1 along with the Joule cycle. The Humphrey cycle models gas turbines with ideal constant volume combustion and is best suited for the cases of shockless explosion combustion and resonant pulsed combustion. The ZND cycle models the application of detonative combustion in gas turbines.
Heiser and Pratt [
9] were the first to theoretically demonstrate the potential of pressure gain combustion to raise the efficiency of gas turbines. Their analysis focused on the ideal Humphrey and ZND cycles and concluded that the main reason for their higher efficiency is the lower entropy increase during combustion. By extending their analysis to include the turbomachinery isentropic efficiency, they showed the importance of the expander efficiency for the cycle efficiency. However, the T-s diagrams and the respective ideal cycle calculations do not model the actual physical phenomena in pressure gain combustion systems in a satisfactory way. In fact, the processes are periodic and time-dependent in the combustor, while they can be easily represented by quasi steady values in time at the outlet of the compressor. Another very important aspect of the analysis in [
9] was the assumption that expansion, and thus work extraction from the working medium, starts at the highest temperature point of the cycle (point 3 in
Figure 1). Nalim [
10] indicated both shortcomings. He thus proposed a simplified model that accounted for the internal expansion in a pressure gain combustor. This model delivers an equivalent steady thermodynamic state at the outlet of constant volume combustors that can be then used for an analysis similar to that in [
9]. This model does not account for entropy generation due to shock in detonations, but it is a good approximation of the physical phenomena taking place in the two PGC technologies, which are best modeled by the Humphrey cycle (see
Section 2.1). Paxson et al. [
11] proposed a more detailed way to account for the time variation at the outlet of pulsed detonation combustors. They used a typical operational map of a turbine expander and computed the work output with a quasi steady-state model. This approach has been adopted in the work of Stathopoulos [
12] and Rähse [
13,
14] to compute the thermal efficiency of the pulsed detonation and the shockless explosion cycles. In this case, the processes in the combustor are resolved in time by solving the 1-D time dependent Euler equations with source terms for the chemical reaction. The time-resolved combustor outlet conditions were then fed to a turbine expander model that computed the generated work in a similar way as in [
11]. Nordeen applied a similar method to resolve the outlet conditions in a rotating detonation engine, also with the aim to compute the thermodynamic efficiency of the cycle [
15]. Irrespective of the type and approach of the aforementioned models, effects such as detonation-to-deflagration transition, quasi constant volume combustion and the pressure drop at the combustor inlet and outlet have not been accounted for in a holistic manner.
The exhaust flow of pressure gain combustors is characterized by strong pressure, temperature and velocity fluctuations [
16,
17]. The main challenge in the practical implementation of PGC into gas turbines is the lack of turbomachinery that can efficiently harvest work from the PGC exhaust gas. Although still a topic of active research, it is generally accepted that conventional turbine expanders have a lower isentropic efficiency when they interact directly with pressure gain combustors [
18,
19]. To address this challenge, one can follow two extreme methods. According to the first, a plenum or combustor outlet geometry could be designed to adapt the exhaust stream of a PGC to an extent that it could be fed to a conventional turbine. In this case, the latter would operate at its design efficiency. The other approach focuses on a dedicated turbine design that could directly expand the outlet flow of a PGC. A much more rational approach would be to optimize the combination of a PGC outlet geometry and an adapted turbine design to achieve the maximum possible work extraction. The current work aims at benchmarking the latter approach for the cases of shockless explosion combustion and pulsed resonant combustion. In this way, insights on the allowable limits for the losses in exhaust gas conditioning devices and the maximum allowable reduction in turbine efficiency can be gained.
Another aspect of the cycles that has been neglected in all previous thermodynamic evaluations is turbine cooling. This topic has two implications. On the one hand, the combustor is expected to deliver an average pressure increase over a limit cycle. This implies that the cooling air for the first turbine stage has to be compressed by an additional compressor. On the other hand, turbine cooling reduces the cycle efficiency for the same turbine inlet temperature and its effect on PGC gas turbine cycles has not been analyzed yet. Furthermore, it has been shown by numerous studies on turbine integration that the pressure, velocity and temperature fluctuation stemming from PGC combustors are largely attenuated through the first turbine stage [
18,
20]. This means that the remaining turbine stages will most probably work at their nominal isentropic efficiency. Up to date, the expansion efficiency has been lumped in one equivalent efficiency of the whole turbine. The current work aims at resolving this issue and its impact on cycle efficiency.
In summary, the present work aims at resolving several open questions on the Humphrey gas turbine cycle. More specifically, the effect of excursions from ideal constant volume combustion on the cycle and its thermal efficiency are explored. The current work is also the first that accounts only for reductions in the the first turbine stage efficiency and thus clarifies the demand for further research in the field of turbine design. In the same scope, the sensitivity of the cycle efficiency on the installation of exhaust gas conditioning devices at the turbine inlet is studied. The current work also aims at clarifying the importance of turbine cooling for the efficiency of the Humphrey cycle, as it is compared to an equivalent Joule cycle with turbine cooling. Moreover, the impact of an additional compressor that delivers cooling air to the first turbine stage is analyzed.
To answer these questions, a new steady state model of the Humphrey cycle was developed in Aspen plus, the details of which are presented in
Section 2.
Section 3 presents the results of the analysis, and the current work concludes with some recommendations for further work on the attempted cycle analysis.