Internal Flow Characteristics of Novel Turbine Performance Enhancement Method through Specific Fluid Injection

Featured Application

In this paper, the authors discuss the internal flow characteristics of a novel method for turbine performance enhancement at partial loads using fluid injection into specific sections of turbine vanes. The results will increase the understanding of the flow through the injection channels and contribute to the development of the injection method as a viable solution for existing engines.

Abstract

Gas turbine engines are an essential component for many industries, such as aerospace, marine propulsion, energy generation, etc., with modern engines capable of achieving high powers, efficiencies and reliability. However, these high performances are achieved on a narrow interval of working regimes; when operating at partial loads, a drastic decrease in performance is expected. In order to mitigate this drawback, a novel injection method has been proposed to improve axial turbine and gas turbine engine performance at these regimes. The method consists of the injection of a fluid in specific sections of turbine vanes to accelerate the flow, modify the velocity triangles and increase the generated power at partial loads. In this paper, the authors discuss the internal flow characteristics of the injection channels using numerical studies to determine the flow fields for different working regimes. The results show that the power generated by the rotor can be improved by 10% to 21% for different operating regimes without considering the internal geometries. The introduction of internal flow configurations led to smaller improvements in power generation, obtaining injection system pressure losses of between 10% and 20%. The paper concludes that the flow through channels is not uniform, with smaller dimensions of the internal geometries leading to higher pressures and an increased influence of the injection system.

1. Introduction

Axial turbines are a main component of turbomachinery used in different industries, such as aviation, power generation, marine propulsion, etc., due to their high power output, small dimensions and high efficiencies. These systems have been in a continuous process of evolution, achieving high performances despite harsh operating conditions (very high temperatures, pressures, and mechanical loads). This evolution was possible due to advances in different fields, such as fluid mechanics, materials and heat transfer.

The use of numerical design, Computer Aided Design (CAD) methods and Computational Fluid Dynamics (CFD) led to the introduction of evermore complex geometries and systems (such as cooling systems). A less studied drawback of axial turbines is the rapid decrease in performance at partial loads. Since turbines are designed for a set of nominal inlet parameters and power requirements, when operating at partial regimes (i.e., different inlet temperatures, pressures and mass flows according to the working line of the respective engine) the performance drops significantly. This is caused by a change in flow characteristics in turbine channels, which are not suited for the fixed geometries of the vanes and blades and are designed for nominal regimes. A method to address this decrease in performance is the use of adaptive turbine geometry. Through changing the turbine geometry, depending on the inlet conditions, the flow through the turbine can be controlled to better extract the available energy of the hot gases at the respective regime. The use of adaptive geometry is not a new idea, with a patent obtained as early as 1966 by Clarence Edgar Le Bell and Alvin Taub [1] for a variable stator turbine. However, due to the extreme environment at which the turbine operates, high temperatures and pressures, a variable geometry turbine could not be used for modern engines without affecting engine reliability.

With the aim of controlling the flow through the turbine at partial regimes, in order to improve the performance without affecting engine reliability, an active flow control system has been proposed and characterized in previous works [2,3]. Instead of using variable geometries, the method consists of the injection of a fluid in specific sections of vanes in order to modify the flow sections, accelerate the flow through the channels and increase the power output. Previous studies showed that power generation can be increased by as much as 30%, depending on the injection system configuration, without increasing the total mass flow through the turbine [2]. The studies have focused on determining the influence of different parameters on the influence of the injection system. It was also showed that the system can be applied to different turbine sizes and for a large operating interval [3].

A similar injection method was researched by numerous authors with a similar aim, flow control in turbine channels. The main difference with the aforementioned studies is that the aim is to decrease certain losses not to control the flow to directly influence power generation. In [4,5,6,7,8,9,10], the authors introduced an injection method in order to accelerate the flow on the suction side of the blade in order to reduce or eliminate the flow detachment at different Reynolds numbers. The results showed that through injecting a fluid in these regions the flow detachment process can be delayed. The authors also studied the use of unsteady jets as a method of controlling flow detachments. The results showed that through using transient flows similar results can be obtained if the correct frequencies are achieved. Similar studies were conducted in [11,12,13,14,15,16,17].

A different approach but with the same goal, that of reducing the losses caused by the appearance of detachments on the suction side of the vanes, was undertaken by Rohr and Yang [18]. The authors performed a numerical study in which they used an injection system in the trailing edge area of the vane for a turbine operating at partial regimes. Through injecting the fluid into the working channel, a reduction of the section is obtained, resulting in lower pressure gradients, thus reducing the area of detachments.

The process of fluid injection in turbine channels and the flow through internal geometries of the vanes is not a new concept, similar systems are currently used in modern gas turbine engines for cooling purposes. Through using bleed air from the compressor and injecting this fluid in the vanes, after passing through the internal geometries, a fluid film around the vane walls is achieved, thus protecting the materials from the high temperature gases. The topic has been intensively studied by numerous authors to determine the cooling efficiency, the flow around injection orifices, pressure losses and fluid interactions [19,20,21,22,23,24].

The paper is composed of two parts; first, the influence of the injection system on the turbine flow field and performance is determined for several partial regimes through comparing the cases with and without injection. For this study the injection channels are not considered, the boundary conditions being set at the exit of the injection orifices. A second numerical study is conducted to determine the flow field inside the injection channels with boundary conditions set further upstream on the injection path. Four configurations of the internal flow channels are considered with different dimensions of the chamber diameter.

2. Materials and Methods

In this study, the injection system is applied to a small size axial flow turbine, part of the gas generator assembly for a microjet engine. At a nominal regime, the turbine generates approximately 274 kW for a mass flow of 1.4 kg/s. The 3D models for the turbine parts and assembly are shown in Figure 1 with the main parameters presented in

Table 1. Axial turbine main parameters.

Parameter	Measurement Unit	Value
Vane shroud diameter	mm	89.8
Rotor shroud diameter	mm	92.7
Vane maximum height	mm	19.8
Rotor maximum height	mm	25.7
Rotor tip clearance	mm	0.5
Number of vanes	-	26
Number of rotor blades	-	37
Nominal rotational speed	rpm	45,000
Nominal mass flow	Kg/s	1.427
Nominal power	KW	274
Pressure ratio	-	2.1
Isentropic efficiency	%	84.5

.

To determine the influence of the injection system on the axial turbine flow field, a numerical study was conducted using CFD tools (Computational Fluid Dynamics). For the studies presented in this paper, ANSYS CFX was used to perform numerical simulations and to characterize the flow and resulting performance of the turbine.

The numerical model of the turbine, which incorporates the injection system, was constructed based on previously formulated methodology [1], starting with the identification of the critical section of the vanes and its intersection with the suction side, followed by the translation of the intersection curve upstream with predefined values. The injection orifices, with predefined dimensions and similar size and shape, were placed equidistant from the hub and shroud on this translated curve. The numerical model for this study is presented in Figure 2.

For the second part of the study, the internal flow was considered, thus injection channels were introduced to the numerical model. The injection inlet is moved upstream on the injection path and a cylindrical chamber inside the vane is created, which is connected by a number of small channels to the turbine main flow. This numerical model is presented in Figure 3.

For this model, the numerical grid is unstructured and was generated using ANSYS Mesh. In order to determine the grid influence on the numerical result, a grid analysis was performed using three configurations with the number of elements increased by a total factor of 3.4. The results from the grid analysis are presented in

Table 2. Numerical grid analysis results.

Configuration	Number of Elements	Generated Power [kW]	Isentropic Efficiency
1	699,864	176.803	0.844
2	1,216,848	175.027	0.845
3	2,380,918	174.812	0.845

. The differences between the results from the grid configurations are negligible. For this study, the second configuration was used as the values of y+ for this grid were in the recommended interval for the respective turbulence model and the computational demands were acceptable. The general dimension of the cell is 0.5 mm with smaller values near the walls. For the injection channels, the dimension for the cell is 0.05 mm and 0.1 mm for the internal cylindrical chamber. Numerical grids for both the vane and rotor are presented in Figure 4.

The boundary conditions and numerical case are typical for axial turbine simulations, with the total pressure and temperature set at the vane inlet and mass flow set at the rotor outlet. A single vane channel and respective single rotor channel were employed using rotational periodicity conditions in order to lower the computational demands. At the vane–rotor interface stage, mixing conditions were used. The walls were considered adiabatical walls with a no-slip condition. For the injection system, total temperature and 2.5% of working fluid were set for the injection orifices, respective to the cylindrical chamber inlet in the second part of the study, using the inlet boundary condition. Air ideal gas was used for these studies for both working flow and injection fluid.

The k–ε model turbulence model was used for these studies as fully turbulent flows are expected as well as strong turbulence at the interaction between injection flow and the flow through the turbine. This model is used in many industrial processes as it is known for its robustness, reasonable accuracy and lower computational demands.

For the convergence criteria, three parameters were monitored: residuals, mass flow imbalance and rotor blade torque variation. The residual target was set at 10⁻⁶, reaching at least 10⁻⁴ for each case. In terms of mass flow imbalance, both vane and rotor volumes were monitored, the imbalance achieving values smaller than 0.1% for each simulation. The last parameter monitored was the rotor blade torque. The case was considered converse if the variation of rotor blade torque was less than 0.2% over 300 iterations.

Based on previous studies [2,3] a configuration of the injection system was selected, with the system main parameters presented in

Table 3. Injection system parameters.

Parameter	Measurement Unit	Value
Injection diameter ${\mathrm{Ø}}_{\mathrm{i}\mathrm{n}\mathrm{j}}$	mm	0.5
Number of orifices	-	22
Coverage degree	%	55.5
Axial distance $Za$	-	0.2
Injection angle	°	90
Injection mass flow ${\dot{M}}_{inj}$	kg/s	$2.5\% \dot{M}$
Injection total temperature $T_{inj}^{\ast }$	K	$T_{inj}^{\ast }=T_{in}^{\ast }$

. Perpendicular injection was selected because in previous studies [2] it was determined that the injection system has, for this angle, a greater influence. The injection orifices were limited at 0.5 mm and the number of injection sections at 22 due to manufacturing considerations. With these values and the height of the vane, a coverage degree of approximately 55.5% resulted [2]. The injection mass flow was limited at 2.5% of the working fluid mass flow to limit the influence of the bleed air on the engine cycle. The injection temperature was considered equal to the turbine inlet temperature for each of the studied regimes.

In the first part of this study, the influence of the injection system has been determined at five partial regimes between 71% and 85% nominal speed. The input parameters for each partial regime have been determined using turbine maps and the engine working line. The main parameters for the studied regimes are presented in

Table 4. Partial regimes’ parameters.

Case	Regime [%]	$\dot{\mathbf{M}} [\mathbf{k}\mathbf{g}/\mathbf{s}]$	${\mathbf{P}}_{\mathbf{i}\mathbf{n}}^{\mathbf{\ast }} \left[\mathbf{b}\mathbf{a}\mathbf{r}\right]$	${\mathbf{T}}_{\mathbf{i}\mathbf{n}}^{\mathbf{\ast }} \left[\mathbf{K}\right]$	Speed [rpm]	Power [KW]
Nominal	100%	1.427	4.55	1180	45,000	274.3
Partial 1	71%	0.852	2.27	774	32,400	78.4
Partial 2	75%	0.936	2.5	798	34,200	96.6
Partial 3	80%	1.053	2.9	852	36,900	123.1
Partial 4	83%	1.093	3.04	876	37,800	136.8
Partial 5	85%	1.17	3.36	932	39,600	153.4

.

3. Results

The results of this paper are divided in two parts. First, the influence of the injection system on the flow field in turbine channels is determined as well as the performance improvements in terms of power generation. For this part, the internal flow of the injection system is not represented, only the turbine channels flow is studied. In the second part of the study, several geometries for the internal cavities of the injection system are investigated.

3.1. Performance Improvement

As seen in previous studies [2,3], the main effect of the injection is the deviation of the working flow from the suction side of the vane, which leads to a smaller flow section and increased velocities at the vane outlet. This effect can be determined from Figure 5, which presents the velocity contour at the mean radius of the vane before and after injection. The injection fluid interacts with the main flow creating a low-pressure zone downstream of the injection orifice, which leads to a decrease in the vane flow section and an acceleration of the working fluid. The decrease in flow section is visible when plotting the absolute pressure in the vane channels, as presented in Figure 6.

Using this injection system for the previously presented partial regimes, it was determined that the generated power can be improved for a wide range of regimes. The results of the injection process are presented in

Table 5. Partial regimes’ analysis results.

Name	Regime [%]	Power without Injection [kW]	Power after Injection [kW]	Power Increase [%]	Equivalent Regime [%]
Partial 1	71	78.4	86.6	10.5	73
Partial 2	75	96.6	110.5	14.4	78
Partial 3	80	123.1	146.1	18.7	83
Partial 4	83	136.8	166.0	21.4	87
Partial 5	85	153.4	186.3	21.5	90

. Graphically representing the regimes calculated before and after the injection, as in Figure 7, it can be observed, as in previous research, that there is an increase in the slope of the power generated by the turbine at different regimes. Thus, higher powers can be achieved without changing the turbine input parameters. In this way, the engine speed can be increased without adding fuel.

3.2. Internal Flow

The flow through the internal channels of the injection system represents an important topic. As the injection fluid is envisioned to be bleed air form the engine compressor, the total pressure loss across the channels from the bleed point to the injection orifices will represent an important parameter in determining the available mass flows and pressures for the injection system. Thus, in this paper, a preliminary study is conducted to determine the influence of the internal geometry of the channel. Four geometries are considered with different diameters of the vane interior injection chamber and characterized based on their influence on the pressure loss inside the injection channels and the overall performance of the turbine.

Through representing velocity streamlines from the injection chamber inlet, it can be determined that smaller velocities are achieved in the lower part of the chamber, as presented in Figure 8b, because of constant mass flow decrease through injection in the turbine channels. After injection, the fluid interacts with the working fluid creating a deviation followed by the rapid change of direction and attachment to the suction side of the vane, as presented in Figure 8a.

As the injection chamber dimensions decrease, the velocity through these parts of the injection system increase, as can be determined from Figure 9, leading to higher pressure losses. The evolution of these losses with the increase in chamber diameter is presented in Figure 10, with pressure losses computed as the ratio between the total pressure at the chamber inlet and the mean value of total pressure at the injection points. The small dimensions of the chamber led to high pressure losses, over 20%. For higher dimensions, the influence of the chamber diameter decreases, the pressure losses being determined by the injection channels.

In terms of overall performance, the introduction of the inside flow geometry to the numerical model resulted in lower influence of the injection system. This effect can be determined from

Table 6. Turbine power generation for different internal flow configurations.

Configuration	Internal Injection Chamber Diameter	Injection Mass Flow [%]	Regime [%]	Power without Injection [kW]	Power after Injection [kW]	Power Increase [%]	Pressure Losses [%]
Without internal geometry	-	2.5	85	153.4	186.3	21.5	-
1	3.5	2.5			174.8	14.0	9.7
2	3	2.5			176.8	15.2	10.7
3	2.5	2.5			178.3	16.2	13.3
4	2	2.5			178.5	16.4	21.4

where the four internal flow configuration results are compared with the ideal case (without considering the internal losses). As the numerical cases are constructed with the injection mass flow set at the chamber inlet, the pressure losses of the internal geometry led to higher pressures needed for injection in order to achieve a 2.5% injection mass flow.

Comparing the four configurations studied, it was determined that decreasing the dimensions of the injection chamber, which led to an increase in injection pressure (in order to achieve the 2.5% injection mass flow), led to an increased pressure in the injection chamber and a better distribution of the injection fluid across the injection orifices from shroud to hub. This in turn leads to a slightly better deviation of the working fluid and increased influence of the injection system. The higher chamber pressure is shown in Figure 11 which presents a comparison of total pressure distribution for the configuration with the highest and the lowest diameter.

4. Conclusions

In this paper, a two-part study was conducted in order to characterize the performance enhancement method for axial turbines. The system, which implies the injection of a fluid in specific parts of turbine vanes, was studied in terms of overall performance variation, considering in the initial part of the study an ideal injection geometry, and in terms of internal flow characteristics, considering in the second part four configurations of the internal geometry.

Considering the influence of the system in turbine channels, the injection process leads to a deviation of the working fluid form the suction side of the vane downstream of the injection points and a decrease in the flow section. Through placing the injection orifices in the vicinity of the minimal section of the vanes, a small overall flow path is obtained, resulting in increased velocities through the channels and higher powers generated by the rotor. For the turbine studied in this paper, a power increase of more than 21% was obtained when compared with the case without injection, with higher values obtained for higher partial regimes. The power increase is achieved without adding mass in the channel as the mass flow is fixed at the rotor outlet, the added mechanical power being a result of a modified flow geometry better adapted to the respective partial regime. The results are in line with previous works conducted on the axial turbine performance enhancement method.

The injection system internal flow was discussed in terms of flow characteristics for four configurations of the internal geometry. The decrease in injection chamber diameter leads to higher velocities, as the aim of the system is to achieve a 2.5% injection mass flow, which led to higher pressure losses (over 20% for the geometries studied in this paper). Higher dimensions of the chamber led to lower pressure losses, approximately 10%, with values being determined by the injection channels. In terms of overall performance, the introduction of the internal geometry resulted in a decrease in added power with the best results obtained for smaller dimensions of the injection chamber as the increased pressure in the chamber led to better mass flow distribution across the vane height.