Мир Знаний

Tupolev 154M noise asesment (стр. 3 из 3)

 

Corresponding drag coefficient is determined from Fig. 6.6.

Some corrections must be made to calculated values of drag and lift coefficients. It is necessary to take into account the influence of the landing gear which creates additional drag and decreases lift. The influence of flaps and slats is little and can be neglected.

 

 

Necessary thrust is calculated using the following formula

, where

 

is drag and q is approach path angle which is equal to 3 degrees.

 

Calculated results for five different landing weights are shown in the table 6.3.

Table 6.3 Calculation results for Tupolev 154M at approach configuration.

Weight, % MLW MLW 95% 90% 85% 80%
Weight, kg 80000 76000 72000 68000 68000
Vapp, m/s 74,8 72,91 70,964 68,965 66,91
Thrust, kg 8445,63 8024,67 7601,88 7179,66 6758,58
LA , dBA 96,74 96,05 95,35 94,66 93,97
EPNL, EPNdB 112,17 111,32 110,48 109,64 108,79
∆LA, dBA 0 0,69 0,7 0,69 0,69
∆EPNL, EPNdB 0 0,85 0,84 0,84 0,85
SQRT (Wing Load) 21,082 20,548 20 19,437 18,856
Thrust To Weight rt. 0,10557 0,105588 0,105582 0,105583 0,105603

Tupolev 154M has the same aerodynamics as Tupolev 154, thus the necessary thrust for both of them during approach is almost the same. Tupolev 154M has more powerful engines and it can carry more payload. Its maximum landing weight is 2 tons greater than that one of 154. Noise parameters are different for these aircraft (table 6.2), and the calculated noise levels slightly differ as well.

7 Noise Suppression

7.1 Suppression of Jet Noise

Methods for suppressing jet noise have exploited the characteristics of the jet itself and those of the human observer. For a given total noise power, the human impact is less if the frequency is very high, as the ear is less sensitive at high frequencies. A shift to high frequency can be achieved by replacing one large nozzle with many small ones. This was one basis for the early turbojet engine suppressors. Reduction of the jet velocity can have a powerful effect since P is proportional to the jet velocity raised to a power varying from 8 to 3, depending on the magnitude of uc . The multiple small nozzles reduced the mean jet velocity somewhat by promoting entrainment of the surrounding air into the jet. Some attempts have been made to augment this effect by enclosing the multinozzle in a shroud, so that the ambient air is drawn into the shroud.

Certainly the most effective of jet noise suppressors has been the turbofan engine, which in effect distributes the power of the exhaust jet over a larger airflow, thus reducing the mean jet velocity.

In judging the overall usefulness of any jet noise reduction system, several factors must be considered in addition to the amount of noise reduction. Among these factors are loss of thrust, addition of weight, and increased fuel consumption.

A number of noise-suppression schemes have been studied, mainly for turbofan engines of one sort or another. These include inverted-temperature-profile nozzles, in which a hot outer flow surrounds a cooler core flow, and mixer-ejector nozzles. In the first of these, the effect is to reduce the overall noise level from that which would be generated if the hot outer jets are subsonic with respect to the outer hot gas. This idea can be implemented either with a duct burner on a conventional turbofan or with a nozzle that interchanges the core and duct flows, carrying the latter to the inside and the former to the outside. In the mixer-ejector nozzle, the idea is to reduce the mean jet velocity by ingesting additional airflow through a combination of the ejector nozzles and the chute-type mixer. Fairly high mass flow ratios can be attained with such arrangements, at the expense of considerable weight.

The most promising solution, however, is some form of “variable cycle” engine that operates with a higher bypass ratio on take-off and in subsonic flight than at the supersonic cruise condition. This can be achieved to some degree with multi-spool engines by varying the speed of some of the spools to change their mass flow, and at the same time manipulating throttle areas. Another approach is to use a tandem-parallel compressor arrangement, where two compressors operate in parallel at take-off and subsonically, and in series at a supersonic conditions.

7.1.1 Duct Linings

It is self evident that the most desirable way to reduce engine noise would be to eliminate noise generation by changing the engine design. The current state of the art, however, will not provide levels low enough to satisfy expected requirements; thus, it is necessary to attenuate the noise that is generated.

Fan noise radiated from the engine inlet and fan discharge (Fig. 7.1) of current fan jet airplanes during landing makes the largest contribution to perceived noise.

Figure 7.1 Schematic illustration of noise sources from turbofan engines

Figure 7.2. shows a typical farfield SPL noise spectrum generated by a turbofan engine at a landing-approach power setting. Below 800 Hz, the spectrum is controlled by noise from the primary jet exhaust. The spectrum between 800 and 10000 Hz contains several discrete frequency components in particular that need to be attenuated by the linings in the inlet and the fan duct before they are radiated to the farfield.

 

Figure 7.2 Engine-noise spectrum

The objective in applying acoustic treatment is to reduce the SPL at the characteristic discrete frequencies associated with the fan blade passage frequency and its associated harmonics. Noise reductions at these frequencies would alleviate the undesirable fan whine and would reduce the perceived noise levels.

A promising approach to the problem has been the development of a tuned-absorber noise-suppression system that can be incorporated into the inlet and exhaust ducts of turbofan engines. An acoustical system of this type requires that the internal aerodynamic surfaces of the ducts be replaced by sheets of porous materials, which are backed by acoustical cavities. Simply, these systems function as a series of dead-end labyrinths, which are designed to trap sound waves of a specific wavelength. The frequencies for which these absorbers are tuned is a function of the porosity of flow resistance of the porous facing sheets and of the depth or volume of the acoustical cavities. The cavity is divided into compartments by means of an open cellular structure, such as honeycomb cells, to provide an essentially locally reacting impedance (Fig. 7.3). This is done to provide an acoustic impedance almost independent of the angle of incidence of the sound waves impinging on the lining.

The perforated-plate-and-honeycomb combination is similar to an array of Helmholtz resonators; the pressure in the cavity acts as a spring upon which the flow through the orifice oscillates in response to pressure fluctuations outside the orifice.

Figure 7.2 Schematic of acoustic damping cavities in an angine duct. The size of the resonators

is exaggerated relative to the duct diameter.

The attenuation spectrum of this lining is that of a sharply tuned resonator effective over a narrow frequency range when used in an environment with low airflow velocity or low SPL. This concept, however, can also provide a broader bandwidth of attenuation in a very high noise-level environment where the particle velocity through the perforations is high, or by the addition of a fine wire screen that provides the acoustic resistance needed to dissipate acoustic energy in low particle-velocity or sound-pressure environments. The addition of the wire screen does, however, complicate manufacture and adds weight to such an extent that other concepts are usually more attractive.

Figure 7.3 Acoustical lining structure.

Although the resistive-resonator lining is a frequency-tuned device absorbing sound in a selected frequency range, a suitable combination of material characteristics and lining geometry will yield substantial attenuation over a frequency range wide enough to encompass the discrete components and the major harmonics of most fan noise.

7.1.2 Duct Lining Calculation

First we have to determine the blade passage frequency:

, where z is number of blades, n is RPM.

Blade passage frequencies for different engine modes are given in table 7.1

Next we determine the second fan blade passage harmonic frequency, which is two times greater than the first one:

 

.

Table 7.1 Fan blade passage frequencies for different engine modes.

Take-off Nominal 88%Nom 70%Nom 60%Nom 53%Nom Idle
RPM 10425 10055 9878 9513 9315 8837 4000
1st harmonic freq., Hz
 
 
5195,083
 
5103,633
 
4915,05
 
4812,75
 
4565,783
 
2066,667

2nd harmonic freq., Hz

 
10772,5
 
10390,17
 
10207,27
 
9830,1
 
9625,5
 
9131,567
 
4133,333

Using experimental data, we determine lining and cell geometry:

For the first harmonic, parameters will be:

· Distance between linings 28.5 cm;

· Lining length 45 cm;

· Lining depth 2.5 cm;

· Cell length 2 cm..

For the second harmonic, parameters will be the following:

· Distance between linings 4.5 cm;

· Lining length 5 cm;

· Lining depth 2.5 cm;

· Cell length 0.4 cm.

Figure 7.4 shows the placement of the lining in engine nacelle.

Figure 7.4 Lining placement in the nacelle.