Active Noise Control

1. Field of the Invention

This invention relates to active control of noise in ducts.

2. Description of Related Art

The principles of active noise control were established by Paul Lueg in 1936 and basically consist of detecting by a microphone the noise which it is wished to control, and replaying the detected noise in anti-phase via a loudspeaker so that the regenerated noise destructively interferes with the source noise. Since that time there has been a great deal of research in the field of active noise control. However, the basic configuration for active noise control in a duct has been the provision of the microphone in the centre of the duct and of the loudspeaker in the duct wall. There are good reasons for this arrangement which will be gone into in greater detail later on in this specification.

However, it has been discovered that this known arrangement has disadvantages when there is a fluid medium flowing through the duct.

An object of the present invention is to provide an active noise control system which reduces these disadvantages.

According to the present invention there is provided on active noise control system comprising a duct through which noise to be controlled propagates; a microphone located in a wall of said duct, a source of anti-sound mounted substantially in the centre of said duct; and means responsive to the magnitude of the sound received by the microphone for driving the anti-sound source to reduce said magnitude.

The anti-noise source may comprise a loudspeaker mounted within the duct, or may comprise an outlet to which the output of a loudspeaker is piped, the loudspeaker itself being external of the duct.

An embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings, in which

FIG. 1 is a diagrammatic view of a known active noise control system according to the prior art,

FIG. 2 is a similar view of a system according to the present invention,

FIG. 3 is a perspective view of a duct,

FIG. 4 is a response graph, and

FIG. 5 is a block diagram of a control circuit.

Referring now to the drawings, FIG. 1 shows a known arrangement which is essentially that established by Paul Lueg. In this arrangement a sensing microphone 1 is positioned in the centre of a duct 2 and a loudspeaker 3 is located in the duct wall. This is the simplest form of an active attenuator for duct-borne sound. The system includes a controller 4 which includes an electrical signal delay to compensate for the acoustic propagation delay from the sensing microphone 1 to the antisource loudspeaker 3. If the microphone is placed directly in front of the loudspeaker piston, this acoustic delay is eliminated and the controller can be a simple inverting amplifier. Since the microphone senses the sound from the loudspeaker in addition to the primary noise, a closed loop configuration exists and there is a danger of instability. Substantial noise attenuation can be obtained provided the loop gain is high, but stability criteria limit this. In ducts, the positions of the sensor and the loudspeaker are important for stability and maximum obtainable attenuation, as will be shown in the following paragraph.

This can be appreciated by considering the duct shown in FIG. 3. A guided propagated wave in this rectangular duct can be described by equation (1):

where p is the acoustic pressure

.PSI..sub.n B is the pressure amplitude ##EQU1## with the constant D(n.sub.y,n.sub.z) determined from the identity for the orthogonality of eigenfunctions:

where A is the duct cross-sectional area.

A microphone placed in this duct will sense the pressure as described in equation (1) and measures both plane and transverse waves. The latter cannot be cancelled with a simple monopole antisource and the contribution of these waves to the total pressure, and consequently to the overall loop gain, does not contribute to the cancellation of plane waves. The phase shift caused by these transverse modes, especially at resonance frequencies, is also detrimental to the noise reduction which is obtainable. This is due to the reduction in the open loop gain necessary to maintain stability. It can be shown, however, that the ratio between total acoustic pressure and pressure due to plane waves is minimal when the microphone is placed in the duct centre. If the loudspeaker is mounted in the duct wall, as in FIG. 1, most transverse modes can be generated in addition to the plane wave mode, and the plane waves and even-numbered (n.sub.y and n.sub.z are even numbers) transverse modes are sensed by the centre mounted microphone. Positioning the microphone in the centre of the duct has, however, the disadvantage that airflow in the duct causes turbulence at the microphone resulting in a locally generated noise field. This gives rise to the electrical output of the microphone no longer being directly related to the acoustic field propagating down the duct. This severely restricts the obtainable attenuation and some form of microphone wind screening is essential.

Accordingly the present invention proposes that the microphone should be incorporated in and located flush with the surface of the duct wall, as is shown in FIG. 2 of the accompanying drawings. In this position the microphone no longer generates any flow noise because the air flow velocity at the duct surface is zero. However, there is limitation of attenuation because the microphone is no longer at a position where the contribution of transverse modes to the total acoustic pressure is minimal.

This problem can be alleviated by placing the antisource loudspeaker in the centre of a transverse cross-section of the duct. In this way a minimum number of transverse modes are generated. Thus if a point source (x.sub.o,Y.sub.o,z.sub.o) is placed in a duct the pressure amplitude can be written as ##EQU2## with S the monopole pressure amplitude. From equation (3) it can be shown that ##EQU3## is nonzero only if n.sub.y and n.sub.z are even integers. Hence, only one quarter of all transverse waves will be generated.

It has been found that an active noise control system with the configuration of a wall-mounted microphone and a centre-placed antisource yields satisfactory attenuation when there is airflow in the duct.

It will be appreciated that an antisource placed in the duct rather than in the duct wall will generally occupy a larger volume than a microphone and will therefore provide a larger obstruction to the airflow. In most practical applications, however, the active system will be integrated with a passive absorber, such as a splitter silencer. In such a case there would not be a significant increase in the overall air resistance.

Another important consideration is system stability. The active noise control system operates in a closed loop configuration due to the acoustic signal path from the loudspeaker back to the sensing microphone, and consequently the system could become unstable. To prevent this, stability criteria must be met and gain and phase need to be controlled. Since the amplitude-frequency response of a loudspeaker rolls off a low frequencies (i.e. a decreasing output with decreasing frequency), the open loop gain in this frequency region will decrease as well. The effect on the closed loop transfer function is that the loop phase goes through zero, which could lead to instability.

To meet this problem the system according to the invention incorporates an integration circuit. This is shown at 10 in FIG. 5 from which figure it can be seen that the control circuitry leading from microphone 1 to loudspeaker 3 comprises a microphone preamplifier 9, the integrator 10 and an inverting power amplifier 11. The inverting amplifier 11 provides the necessary phase shift to ensure that the output of the loudspeaker 3 interferes destructively with the noise detected by the microphone 1.

The integrator circuit 10 is intended not only to improve the loop gain at low frequencies thereby increasing the achievable attenuation, but also to modify the phase shift around the loop to secure operational stability. The integrator circuit 10 has therefore been given the amplitude-frequency response shown in the graph of FIG. 4. To produce this response the circuit 10 has a transfer function ##EQU4## where s=j..omega., j=.sqroot.-1, .omega. is the frequency in rads and .tau. the circuit time constant. High frequency stability can be ensured by reduction of gain by means of passive absorptive material placed on the walls of the duct.