GB2310039A - Turbulence-noise cancelling means - Google Patents

Abstract

The turbulence noise cancelling means comprises an acoustic sensor 2 eg in the form of a microphone having a pressure sensing diaphragm 4. An anemometer 6 is mounted on the microphone. The anemometer senses the speed of the surrounding fluid eg air speed, and the output signal from the anemometer is applied to a circuit 8 for calculating the air speed, the output of which is applied to a squaring circuit 10. The output of the squaring circuit 10 represents dynamic pressure and is applied to an input of an adaptive filter 12. The dynamic pressure signal from the adaptive filter 12 is subtracted at combiner 16 from the pressure signal derived from the microphone 2. The microphone (figure 5, not shown) may include a transducer (20) in a form of a loud speaker which receives the output signal from the adaptive filter 12 directly thereby avoiding the need for a combiner 16.

Description

TURBULENCE-NOISE CANCELLING MEANS The present invention relates to turbulence noise cancelling means and in particular to the reduction of turbulence induced noise in acoustic sensing applications.

The applications include, but are not limited to, sonar sensors underneath ships, towed arrays, or on submarines, acoustic sensors mounted on vehicles, acoustic sensors mounted on aircraft, acoustic sensors mounted out of doors and subjected to wind, for example, in making recordings out of doors, and acoustic sensors in pipes or ducts.

Acoustic sensors, for example, microphones, and sonar transducers, are intended to measure the pressure variations associated with a propagating acoustic wave. However, turbulent fluid flow past an acoustic sensor gives rise to unwanted pressure variations that interfere with those due to the wanted signal. The traditional way of tackling this is to place a large porous windshield structure around the sensor to reduce airflow in the vicinity of the sensor and damp out turbulence. This approach can be effective, but does considerably increase the size and conspicuousness of the sensor, and results in increased disturbance of the fluid flow pattern in the vicinity of the sensor.

An aim of the present invention is to provide turbulence noise cancelling means which does not require the use of a large porous windshield structure.

According to the present invention there is provided turbulence noise cancelling means comprising a pressure sensor including an anemometer, said pressure sensor being arranged to detect a pressure wave and generate a signal indicative of the detected pressure, said anemometer being arranged to detect the speed of air in the vicinity of the anemomoter, filter means for generating a dynamic pressure signal derived from said air speed signal, and combining means for combining the pressure signal and the filtered dynamic pressure signal to produce an output signal devoid of noise generated by turbulence.

The pressure sensor is an acoustic sensor in the form of a microphone.

The pressure sensor is an acoustic sensor in the form of a hydrophone.

The anemometer is a hot wire or a hot thermistor or a cross venturi or vortex-shedding type anemometer.

An ultrasonic or any other device capable of measuring the speed of fluid flow in the volume adjacent to the microphone may be used.

Pressure variations can be generated by two distinct mechanisms, compressive and inertial. Compressive variations are associated with sound waves, and propagate with a speed which is a characteristic of the medium. Inertial variations are associated with turbulence, which consists of circulating 'eddies' that are convected along with the general drift of the fluid. In general acoustic sensing applications, only the compressive pressure variations are of interest, those due to turbulence being indicative of noise.

Acoustic waves are described by the equations:

whereas, fluid flow is described by the Navier Stokes equation + U. = #u = vp + vV2u + F V u = O (2) p p where the symbols have the following meanings p pressure difference from ambient p density u particle velocity c speed of sound t time v kinematic viscosity F external force.

The version of the Navier Stokes equation is that specific to incompressible flow, which is valid for speeds much less than the speed of sound.

The two equations model entirely distinct mechanisms by which pressure variations can be generated. In the case of sound waves, the pressure variations are compressive, and are associated with fluctuations in density in the local fluid. In particular, the term V u plays an integral role in the acoustic equations, but, in the case of fluid flow, the pressure variations are principally associated with fluctuations in the particle flow speed, and V U is assumed to be zero.

These two different mechanisms for pressure generation each result in very different relationships between pressure and particle velocity. In the case of a plane sound wave p = pcuX (3) and, in the case of turbulent pressure variations, the relationship is p = -'2p1U12 (4) Equation 4 is derived from the Navier Stokes equation by assuming U does not vary with time, that there is no external force, and that viscous stresses are unimportant. The equation is then used to model a rotating body of fluid, and an expression derived to establish the pressure on the periphery, where the tangential velocity is u.

The table shown in Figure 1 gives some examples of the relationship between particle velocity and pressure for the two processes. The most obvious conclusion to be drawn from Figure 1 is that particle velocities associated with turbulent pressure variations are generally very much greater than those associated with acoustic pressure variations, particularly when the absolute level of the pressure variations is small. Thus, wherever turbulence contributes significantly to the pressure at a point, the particle velocities will be almost entirely due to the turbulence, with only a very small component attributable to acoustic propagation. This observation is the basis of the present invention.

An embodiment of the present invention will now be described with reference to the accompanying drawings wherein: FIGURE 2 shows a turbulence cancelling microphone; FIGURE 3 shows a turbulence cancelling signal processing circuit; FIGURE 4 shows a microphone including a windshield and, FIGURE 5 shows a circuit for subtracting turbulent pressure acoustically.

Referring to Figure 2 and 3 in accordance with the present invention an acoustic sensor 2 is provided which consists of a pressure sensor 4, an air-flow/air speed sensor 6, and a means as shown in Figure 3, of combining the signals from the two to produce a resultant signal which corresponds only to the acoustic component of the pressure variation.

The airspeed sensor 6 must be able to respond quickly, and to be equally sensitive to air flowing across the diaphragm in any direction. Therefore, a hot-wire or hot-thermistor type might be the most suitable, but other types, such as crossed venturi, acoustic or vortex-shedding types could be used also. Another important property of the airspeed sensor is that it should be small enough that airflow past it does not cause significant additional turbulence.

Referring to Figure 3 there is shown one example of a means for cancelling noise generated from turbulence, and the microphone bears the same numerical references as in Figure 2.

The microphone 2 comprises a pressure sensing diaphragm 4 and an anemometer 6 for measuring the speed of air. The output of the anemometer 6 is connected to an input of a circuit 8 which is used to calculate the air speed. The output signal indicative of air speed is passed to a squaring circuit 10, the output of which represents the dynamic pressure. The signal from the squaring circuit 10 is applied to an adaptive filter 12 and to a filter controller 14. The output of the adaptive filter 12 is connected to a negative input of a subtractor 16. A positive input to the subtractor 16 receives the pressure signal generated from the output of the microphone 2. The output of the subtractor 16 represents the acoustic pressure signal which is the desired signal, and this signal is applied to a further input of the filter controller 14, and used for controlling the filter 12.

The signal processing apparatus shown in Figure 3 is only an example of the type of arrangement that might be used. It would be possible to remove the filter controller and replace the adaptive filter with an adjustable attenuator which is set up manually only once. Such a simplified arrangement might still prove capable of making a significant reduction in noise, but would not be able to take account of relative variations in the sensitivity and frequency response of the pressure and airspeed sensors due to temperature and density variations, and due to ageing.

Figure 3 does not explicitly show any analogue to digital conversion. In principal, the whole system could be implemented using analogue circuitry but if an adaptive filter and filter controller are used, they would be more effective as digital circuits. Therefore, analogue digital converters would be used for the pressure and anemometer signals, and all the subsequent processing would be carried out digitally.

The operation of the circuit of Figure 3 will now be described. The signal generated by the anemometer 6 represents air speed and is applied to a calculation circuit 8. The circuit 8 generates an output signal indicative of the air speed and this signal is squared by the circuit 10 to generate a signal indicative of dynamic pressure. This signal represents the unwanted noise and is applied to the adaptive filter 12 which is controlled by the filter controller 14 in a manner dependant upon the dynamic pressure signal and the acoustic pressure signal. The output signal from the adaptive filter 12, indicative of the filtered dynamic pressure is applied to a subtractor 16. The subtractor 16 receives the pressure signal from the microphone and the two signals are subtracted in order to remove the noise generated through turbulence as sensed by the microphone and the anemometer. The output of the subtractor 16 therefore represents the acoustic pressure signal which is the wanted signal.

One potential difficulty with the technique is that the dynamic range of the pressure sensor and the circuitry associated with it, have to be able to accommodate the full amplitude of the turbulent pressure variations. There are two possible ways of reducing this requirement. The first is illustrated in Figure 4, which shows a pressure sensor 2 with an airspeed sensor 6 combined with a small windshield 18. The airspeed sensor 6 might be inside the windshield 18, combined with the windshield 18, or mounted on the external surface of the windshield 18. This arrangement would be useful in cases where there is enough space to accommodate a small windshield. The windshield 18 is capable of attenuating some of the higher-frequency airflow noise, but is unable to deal with the very low frequency noise associated with scales of turbulence much larger than the windshield. This leaves the adaptive turbulence noise canceller to deal only with the lower frequencies. This could be made more effective by including a controlled degree of pre-emphasis in the pressure sensor to roll-off the sensitivity at low-frequencies, with a corresponding de-emphasis filter after the turbulence noise has been cancelled to restore a flat frequency response.

Another way of dealing with dynamic range is shown in Figure 5. In this arrangement, the subtractor shown in Figure 3 is effectively moved into the microphone, and the operation is performed acoustically by means of a transducer 20 in the form of a loudspeaker 20 driven from the filtered dynamic pressure signal derived from the airspeed sensor 6. This removes the need for the microphone electronics to be able to accommodate the full dynamic range of the turbulent pressure variations. Of course, it would be possible to combine this technique with that shown in Figure 4 by integrating the pressure sensor, the anemometer, and the windshield.

The invention therefore reduces the level of turbulence induced noise in an acoustic sensor system without the need for a bulky windshield assembly. This will enable the acoustic sensor system to achieve greater sensitivity and range. In addition, the reduction in size, or complete absence, of a windshield assembly will minimise any disturbance to the airflow caused by the acoustic sensor system, as well as rendering the acoustic sensor less bulky and less conspicuous.

Claims (10)

1. Turbulence noise cancelling means comprising a pressure sensor including an anemometer, said pressure sensor being arranged to detect a pressure wave and generate a signal indicative of the detected pressure, said anemometer being arranged to detect the speed of air in the vicinity of the anemoter, filter means for generating a dynamic pressure signal derived from said air speed signal, and combining means for combining the pressure signal and the filtered dynamic pressure signal to produce an output signal devoid of noise generated by turbulence.

2. Turbulence noise cancelling means as claimed in Claim 1, wherein the pressure sensor is an acoustic sensor in the form of a microphone.

3. Turbulence noise cancelling means as claimed in Claim 1, wherein the pressure sensor is an acoustic sensor in the form of a hydrophone.

4. Turbulence noise cancelling means as claimed in Claim 1, Claim 2 or Claim 3, wherein the anemometer is a hot wire or a hot thermistor anemometer.

5. Turbulence noise cancellation means as claimed in Claim 1, Claim 2 or Claim 3, wherein the anemometer is a cross venturi or vortex- shedding type anemometer.

5. Turbulence noise cancelling means as claimed in any preceding claim, wherein the output signal generated from the anemometer is applied to a circuit for calculating air speed, the output of which is applied to a squaring circuit, the output of said squaring circuit is applied to said filter means controlled in a manner to generate an output signal which represents the noise components generated by turbulence, said combining means receives a signal from the pressure sensor and from the filter means for generating an output signal which is devoid of noise generated by turbulence.

6. Turbulence noise cancelling means as claimed in Claim 5, wherein the output signal from the combining means is applied to the filter controller.

7. Turbulence noise cancelling means as claimed in any preceding claim, wherein the pressure sensor may include a windshield upon which is mounted on an external surface thereof, the anemometer.

8. Turbulence noise cancelling means as claimed in Claims 1 to 4 and Claim 7 wherein the pressure sensor includes a transducer which is controlled by the output of the filter means and is arranged to control the operation of a pressure sensing diaphragm in a manner to remove the components generated by turbulence, from which the output acoustic pressure signal is generated.

9. Turbulence noise cancelling means as claimed in any preceding claim, wherein the filter means is an adaptive filter.

10. Turbulence noise cancelling means substantially as hereinbefore described with reference to Figures 2, 3, 4 and 5, of the accompanying drawings.