Academia.eduAcademia.edu

Figure 13 – uploaded by Dieter Peitsch

See full PDFdownloadDownload figure
Figure 14 Loss coefficient under steady and unsteady conditions at 0.5 % sealing flow

Figure 14 Loss coefficient under steady and unsteady conditions at 0.5 % sealing flow

Related Figures (15)

Table 1 Hot-Acoustic-Testrig parameters HAT 1s a modular test facility for the investigation of acoustic and aero-thermal phenomena on liner and other thermally stressed surfaces in aero engines. This novel test rig is a joint facility of DLR and TU Berlin. The HAT provides the possibility of steady or unsteady flow conditions, with controlled high pressure and / or high temperature environments. In addition, a secondary cooling air supply is installed, which can be used to simulate a secondary air flow. An overview of the main characteristics is shown in Table 1.
Table 1 Hot-Acoustic-Testrig parameters HAT 1s a modular test facility for the investigation of acoustic and aero-thermal phenomena on liner and other thermally stressed surfaces in aero engines. This novel test rig is a joint facility of DLR and TU Berlin. The HAT provides the possibility of steady or unsteady flow conditions, with controlled high pressure and / or high temperature environments. In addition, a secondary cooling air supply is installed, which can be used to simulate a secondary air flow. An overview of the main characteristics is shown in Table 1.
In the cavities, there are highly unsteady mixing processes due to the in and out coming flow. One driving force is the inhomogeneous circumferential pressure distribution around the annulus flow, described by [Phadke and Owen]. This leads to a high frequency interaction of the flow with different temperature and pressure levels in a very small geometry. In the last years, there have been investigations into the rim sealing.
In the cavities, there are highly unsteady mixing processes due to the in and out coming flow. One driving force is the inhomogeneous circumferential pressure distribution around the annulus flow, described by [Phadke and Owen]. This leads to a high frequency interaction of the flow with different temperature and pressure levels in a very small geometry. In the last years, there have been investigations into the rim sealing.
Two different rim-seal designs are examied for this work, one beeing a straight slot and one beeing an axial gap design (Fig. 4).
Two different rim-seal designs are examied for this work, one beeing a straight slot and one beeing an axial gap design (Fig. 4).
Following the air heater is a modular duct system with a 70 mm diameter. While the measurement section is located at the center, microphone probes and loud speakers are arranged up- and downstream. The thermo- acoustic setup is described by [Knobloch and Lahri]. The liner bias flow, e.g. cooling flow and/or secondary The rim-seal designs While the slot design represents a simple reference case, the axial gap is a known and investigated rim-seal design [Schuler]. Both contours agree in dimensions.
Following the air heater is a modular duct system with a 70 mm diameter. While the measurement section is located at the center, microphone probes and loud speakers are arranged up- and downstream. The thermo- acoustic setup is described by [Knobloch and Lahri]. The liner bias flow, e.g. cooling flow and/or secondary The rim-seal designs While the slot design represents a simple reference case, the axial gap is a known and investigated rim-seal design [Schuler]. Both contours agree in dimensions.
Figure 5 Setup for the pressure actuation
Figure 5 Setup for the pressure actuation
To investigate the pressure loss downstream the cascade, a combination of pressure tabs and a calibrated five-hole-probe is used. The general setup is shown in Figure 7 as an overview. The five-hole-probe has a conical head with a tip-diameter of 1.8 mm. It is located 1/3 of the blade span behind the vane. This is the area were the rotor-stator-cavity is located [Owen]. The probe covers a straight line on the y-axis at heights of 9 mm, 11.5 mm, 16.5 mm and 21.5 mm bottom up, with the last value corresponding to half of the sections total height. The five-hole-probe is connected to the HDO- Pressure-sensors, which have a measuring range up to 1000 mbar and a relative error of max. +0.5% Full Scale Output (FSO). All 8. signals were obtained simultaneously at a sampling rate of 1 kHz, using the 64-channel signal amplifier as well. The record time was 60 s for each set point. Since the test rig is under pressure, there are two sealed window configurations, with a total of four probe holes shared.
To investigate the pressure loss downstream the cascade, a combination of pressure tabs and a calibrated five-hole-probe is used. The general setup is shown in Figure 7 as an overview. The five-hole-probe has a conical head with a tip-diameter of 1.8 mm. It is located 1/3 of the blade span behind the vane. This is the area were the rotor-stator-cavity is located [Owen]. The probe covers a straight line on the y-axis at heights of 9 mm, 11.5 mm, 16.5 mm and 21.5 mm bottom up, with the last value corresponding to half of the sections total height. The five-hole-probe is connected to the HDO- Pressure-sensors, which have a measuring range up to 1000 mbar and a relative error of max. +0.5% Full Scale Output (FSO). All 8. signals were obtained simultaneously at a sampling rate of 1 kHz, using the 64-channel signal amplifier as well. The record time was 60 s for each set point. Since the test rig is under pressure, there are two sealed window configurations, with a total of four probe holes shared.
The measurements were carried out for two Mach numbers (Ma = 0.1 & Ma = 0.35) and temperatures around 288 K, while the supply pressure of the valves was varied from p, = 7 bar (abs) up to p, = 12 bar (abs). In Fig. 8 the time-resolved behavior of the pressure probe signal is depicted. the peak-to-plateau values for a recording time ¢ = 1 s at a frequency f= 40 Hz for both locations of the pressure probes. No phase deviation between the two measurement-/probe-locations is detected.
The measurements were carried out for two Mach numbers (Ma = 0.1 & Ma = 0.35) and temperatures around 288 K, while the supply pressure of the valves was varied from p, = 7 bar (abs) up to p, = 12 bar (abs). In Fig. 8 the time-resolved behavior of the pressure probe signal is depicted. the peak-to-plateau values for a recording time ¢ = 1 s at a frequency f= 40 Hz for both locations of the pressure probes. No phase deviation between the two measurement-/probe-locations is detected.
The influence of the sealing air upon the main annulus flow can be validated by the loss coefficient (Eq. 1).
The influence of the sealing air upon the main annulus flow can be validated by the loss coefficient (Eq. 1).
Trequency generator. Figure 9 shows the averaged peak-to-plateau value of the pressure fluctuation over valve frequency at Ma = 0.1. It is recognizable that the pressure amplitude increases with rising supply pressure. The highest pressure amplitude is p,q. = 130 mbar at f = 40 Hz, which results in a relative pressure gradient of 8% compared to the absolute pressure of the main flow. A frequency of f= 40 Hz has been chosen for the pressure fluctuations, because at this frequency the highest pressure amplitudes are reached. In literature [Roy] states a pressure raise from 10 - 20%, but a pressure increase up to 30% is targeted within the Collaborative Research Center. So it is essential to gain a higher pressure amplitude which will be incorporated in the future. This can be achieved by doubling the number of valves from 6 up to 12 via a Y-Mounting. By doing that it is possible to increase the mass flow and therefore the pressure amplitude at parallel operation, or the increase the frequency at serial operation. Another obvious method would be a shock tube. This would provide a higher pressure amplitude, but the downside of this technique is it's singularity; with a shock tube setup it is not possible to investigate the periodic effects of the flow.
Trequency generator. Figure 9 shows the averaged peak-to-plateau value of the pressure fluctuation over valve frequency at Ma = 0.1. It is recognizable that the pressure amplitude increases with rising supply pressure. The highest pressure amplitude is p,q. = 130 mbar at f = 40 Hz, which results in a relative pressure gradient of 8% compared to the absolute pressure of the main flow. A frequency of f= 40 Hz has been chosen for the pressure fluctuations, because at this frequency the highest pressure amplitudes are reached. In literature [Roy] states a pressure raise from 10 - 20%, but a pressure increase up to 30% is targeted within the Collaborative Research Center. So it is essential to gain a higher pressure amplitude which will be incorporated in the future. This can be achieved by doubling the number of valves from 6 up to 12 via a Y-Mounting. By doing that it is possible to increase the mass flow and therefore the pressure amplitude at parallel operation, or the increase the frequency at serial operation. Another obvious method would be a shock tube. This would provide a higher pressure amplitude, but the downside of this technique is it's singularity; with a shock tube setup it is not possible to investigate the periodic effects of the flow.
Conditions of the campaigne are summarized in Tab 2. In Figure 11 the normalized total pressure downstream the cascade is shown. There is a obvious pressure gain caused by the sealing flow, while th difference between the two geometries is very little (Fig. 11). Just a slightly higher pressure loss for the slot gap design, because the sealing air goes straight into the main flow. For the axial gap design there are more losses, because of swirls
Conditions of the campaigne are summarized in Tab 2. In Figure 11 the normalized total pressure downstream the cascade is shown. There is a obvious pressure gain caused by the sealing flow, while th difference between the two geometries is very little (Fig. 11). Just a slightly higher pressure loss for the slot gap design, because the sealing air goes straight into the main flow. For the axial gap design there are more losses, because of swirls
caused by the contour. The normalized pressure loss is at max. 4-5% of the upstream flow. In [Bunker et al.], a pressure loss of around 10% is measured in a rotating setup with realistic vane profiles. Considering the geometrical limitations of the HAT-measurement- section, a 5% pressure loss is acceptable. The use of bigger vane profiles would lead to a larger wake / pressure loss, but also to a blocking of the flow.
caused by the contour. The normalized pressure loss is at max. 4-5% of the upstream flow. In [Bunker et al.], a pressure loss of around 10% is measured in a rotating setup with realistic vane profiles. Considering the geometrical limitations of the HAT-measurement- section, a 5% pressure loss is acceptable. The use of bigger vane profiles would lead to a larger wake / pressure loss, but also to a blocking of the flow.
Figure 12 Total pressure at 1% sealing flow Figure 11 Total pressure at 1% sealing flow The loss coefficient for the two geometries at two different sealing air rates is depicted in Figure 13. It shows the highest losses for the slot design at 1% sealing flow rate. The pressure losses are decreasing with the sealing air ratio. With the slot contour having the highest value at 1 % sealing air and the lowest value at 0.5 %, while the axial gap design is not that much affected by the sealing mass flow rate, but still shows a notable effect. It is obvious that doubling the sealing flow rate will increase the speed of the jet blowing out of the gap. This will turns into a bigger impact to the main annulus flow and potential downstream located blades.
Figure 12 Total pressure at 1% sealing flow Figure 11 Total pressure at 1% sealing flow The loss coefficient for the two geometries at two different sealing air rates is depicted in Figure 13. It shows the highest losses for the slot design at 1% sealing flow rate. The pressure losses are decreasing with the sealing air ratio. With the slot contour having the highest value at 1 % sealing air and the lowest value at 0.5 %, while the axial gap design is not that much affected by the sealing mass flow rate, but still shows a notable effect. It is obvious that doubling the sealing flow rate will increase the speed of the jet blowing out of the gap. This will turns into a bigger impact to the main annulus flow and potential downstream located blades.
Figure 13 Loss coefficient axial-gap and slot design at 0.5 % and 1 % sealing air
Figure 13 Loss coefficient axial-gap and slot design at 0.5 % and 1 % sealing air
Related topics:
Mechanics
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved