KR100918833B1 - Method for determining the location of impacts by acoustic imaging - Google Patents

Abstract

의 계산 단계 (c) 각각의 영역 k에 대한,

의 연산 단계 (d)

의 특성값은 k≠k_o인

의 해당 특성값보다 더 큰, k_o의 발견 단계를 포함한다. The method of determining the impact location on the surface 1 comprising the N acoustic sensors 2a, 2b, 2c, which transmits the sensed signal to the processing device, comprises: (a) P calculating the cross-correlation product (b) P Inverse Fourier Transform

(C) for each region k,

Calculation step of (d)

The characteristic value of is k ≠ k _o

The discovery step of k _o , which is greater than the corresponding characteristic value of.

Acoustic sensor, shock, Fourier transform

Description

Method for determining the location of impacts by acoustic imaging

The present invention relates to a method of determining an impact location on an object and to an apparatus using the method.

를 전송하도록 배열되어 있다.Known methods of determining the impact location on an object are based on acoustic image generation. According to this method, the surface of the object is subdivide in the M area, for example an array, and the N acoustic sensor is fixed on the surface, for example. Each acoustic sensor i receives an acoustic signal generated from the impact on the surface of the object and is detected by a processing unit.

Is arranged to send.

는 시간 지연 보상

으로 지연되고

를 형성하며 합산된다. 이 시간 지연 보상

는 영역 k의 개별 위치 및 센서 i의 개별 위치에 기초한 결정값이다. 도 1a에 나타난 바와 같이, 감지된 신호

가 정확하게 지연된다면, 즉 충격이 영역 k에서 발생된다면, 지연 신호들은 동위상이고

는 임펄스 형태를 가진다. 대조적으로 도 1b에 나타난 바와 같이, 감지된 신호

가 정확하게 지연되지 않는다면, 즉, 충격이 영역 k에서 발생되지 않는다면,

는 넓고 이것의 최대값은 낮다. 따라서, 어레이의 각각의 영역k에 대해,

최대값이 계산되고, 이에 의해 충격의 음향 이미지를 형성한다. 이후

의 비교는 표면상에 생성된 충격을 국부화(localize)하는 것을 허용한다. For each region k of the array, a sensed signal

Time delay compensation

Being delayed by

To form and add up. 2 time delay compensation

Is a determination based on the individual positions of the area k and the individual positions of the sensor i. As shown in FIG. 1A, the detected signal

If is correctly delayed, i.e., an impact occurs in area k, the delay signals are in phase and

Has an impulse form. In contrast, the detected signal, as shown in FIG.

If is not delayed correctly, i.e., no impact is generated in area k,

Is wide and its maximum is low. Thus, for each region k of the array,

The maximum value is calculated, thereby forming an acoustic image of the impact. after

The comparison of allows to localize the impact created on the surface.

In this method, however, the time origin is not calculated correctly, resulting in synchronization problems in the processing device. Thus, a failure occurs in determining the impact location. The accuracy of this method also depends on the shape, in particular the duration of the impact signal. At this time, the longer the shock signal lasts, the lower the accuracy.

It is an object of the present invention to provide a method for determining impact location, the accuracy of which does not depend on the impact signal waveform.

Another object of the present invention is to provide a robust method of determining the impact location.

을 전송하며, 상기 방법은 다음 단계를 포함한다:The present invention therefore proposes a method for determining an impact location on the surface of an object, the surface comprising N acoustic sensors and M crystal regions, where N is at least 3, the impact generating an acoustic signal and Each sensor i receives the acoustic signal and is detected by a processing device

And the method comprises the following steps:

의 연산(computing) 단계 (a) P intercorrelation product

Computing step

는 센서 i에 의해 감지된 감지 신호

의 푸리에 변환이고;here

Is the sensed signal detected by sensor i

Is a Fourier transform of;

는 센서 j에 의해 감지된 감지 신호

의 푸리에 변환이고;

Is the sensed signal detected by sensor j

Is a Fourier transform of;

* Is a complex conjugate operator;

의 P 역푸리에 변환

의 계산 단계;(b) above

P Inverse Fourier Transform

Calculating step of;

와

의 특성값(characterizing value)의 연산 단계, 여기서

는 영역 k의 개별 위치 및 센서 i 와 j의 개별 위치에 기초한 결정값이며;(c) for each region k,

Wow

Operation of characterizing value, where

Is a determination based on the individual positions of the area k and the individual positions of the sensors i and j;

의 특성값은 k≠k_o인

의 해당 특성값보다 더 크다. (d) determining k _o , where

The characteristic value of is k ≠ k _o

Is greater than the corresponding characteristic value of.

In various embodiments of the method according to the invention, at least one of the following characteristics can be used:

는-Inverse Fourier Transform

Is

와 동일하고

Same as

Where Ω is the wavenumber coefficient; And

Where m (Ω) is the frequency corresponding to the wave coefficient Ω according to the material dispersion relation;

는 영역 k와 센서 i 사이의 거리 및 영역 k와 센서 j 사이의 거리에 따른 길이이고; u is the distance

Is the length according to the distance between the area k and the sensor i and the distance between the area k and the sensor j;

와 동일하고, Material dispersion relations

Same as

Where α is a coefficient according to the object;
The mass dispersion relationship is substantially equal to Ω = αω, where α is a coefficient according to the object;

에 좌우되고, 여기서

는 영역 k의 개별 위치 및 센서 i의 개별 위치에 기초한 결정된 각도 값이고;The coefficient α is

Depends on, where

Is a determined angle value based on the individual position of area k and the individual position of sensor i;

에 따르며, T는 온도값이고;The coefficient α is the temperature of the object

T is a temperature value;

The method further comprises an initialization mode comprising determining a mass dispersion relationship of the object;

Determining the material dispersion relationship comprises the following sub-steps of generating an NSO impact at the determined location;

The NSO shock is generated at a position determined along the line connecting two of the N sensors and simultaneously spaced at the same time;

Determining the material dispersion relationship further comprises the following sub-steps:

의 연산 단계;With α test value

Operation of;

의 연산 단계, 및 각각의 α 테스트 값에 대해, 모든

의 합산 단계;For each α test value, for each impact m

Operation stage, and for each α test value, all

Summing step;

의 가장 큰 최대값을 제공하는

의 선택 단계.Sum

Which gives the largest maximum of

Step of choice.

The method further comprises an initialization mode comprising determining the position of the N sensors on the surface;

의 개수는

과 동일하고;-Cross-correlation product

The number of

Same as;

의 특성값은 다음 매개변수 중 하나이고:-

The property value of is one of the following parameters:

의 최대값;ㆍ

Maximum value of;

의 파워, 또는 최대 제곱 진폭;ㆍ

Power, or maximum square amplitude;

 Maximum peak-to-peak amplitude;

의 RMS(root mean square); 또는ㆍ

Root mean square of RMS; or

의 에너지, 이는

와 동일하고;ㆍ

Of energy, which

Same as;

의 펄스 폭 매개변수, 이는

와 동일하고, 여기서 REAL(x)는 복소수 x의 실수부이다;ㆍ

Of the pulse width parameter, which is

, Where REAL (x) is the real part of complex x;

의 특성값이 신뢰의 기설정된 임계보다 클 경우에만 충격이 영역 k_o에서 발생되었던 것으로 결론짓고;-

Conclude that the impact occurred only in the region k _o when the characteristic value of is greater than the predetermined threshold of confidence;

비에 의해 정의된 콘트라스트(contrast)값이 MEAN이 평균 연산자인 신뢰의 기설정된 임계보다 클 경우에만 충격이 영역 k_o에서 발생되었던 것으로 결론짓고;-

Conclude that the impact occurred in the region k _o only if the contrast value defined by the ratio is greater than the predetermined threshold of confidence that MEAN is the average operator;

는 정규화(normalize)된다. -

Is normalized.

Another object of the present invention is a device for determining an impact location on a surface of an object, the surface comprising M determined regions and the impact generating an acoustic signal,

The apparatus includes:

A processing device;

를 전송하며,N acoustic sensors adapted to be provided by the surface, where N is at least 3, and each sensor i receives the acoustic signals and processes the sensing signals autonomously

To send,

Wherein said processing device comprises:

을 연산하는 수단,(a) P cross-correlation product

Means for computing

는 센서 i에 의해 감지된 감지 신호

의 푸리에 변환이고;here

Is the sensed signal detected by sensor i

Is a Fourier transform of;

는 센서 j에 의해 감지된 감지 신호

의 푸리에 변환이고; 그리고

Is the sensed signal detected by sensor j

Is a Fourier transform of; And

* Is a complex conjugate operator;

의 P 역푸리에 변환

을 계산하는 수단;(b) above

P Inverse Fourier Transform

Means for calculating;

와

의 특성값(characterizing value)을 연산하는 수단, 여기서

는 영역 k의 개별 위치 및 센서 i 와 j의 개별 위치에 기초한 결정값이며;(c) for each region k,

Wow

Means for calculating the characteristic value of

Is a determination based on the individual positions of the area k and the individual positions of the sensors i and j;

의 특성값은 k≠k_o인

의 해당 특성값보다 더 크다. (d) means for determining k _o , where

The characteristic value of is k ≠ k _o

Is greater than the corresponding characteristic value of.

의 위상은 충격 파형에 좌우되지 않는다. 따라서, 충격의 파형은 본 발명의 방법의 정확성을 변경하지 않는다. Thus, by using the cross-correlation product of the sensed signals, it is not necessary to determine the time origin for processing the signal. Thus, there is no need for accurate synchronization of the acoustic sensors to implement the present invention. Also, as will be explained below, the cross-correlation product

The phase of is not dependent on the shock waveform. Thus, the waveform of the impact does not change the accuracy of the method of the present invention.

1A and 1B are chronograms that have already been described and represent acoustic image configurations according to technical levels;

2 is a schematic diagram of an apparatus according to the invention;

3 is a flow diagram illustrating the general operation of the apparatus of FIG. 2;

4 is a flowchart illustrating operation during the processing steps of the flowchart of FIG. 3;

5 is a flowchart illustrating operation during the initialization phase.

As illustrated in FIG. 2, the device 3 for determining the impact position according to the invention comprises N acoustic sensors, such as at least three sensors 2a, 2b, 2c shown. It may further comprise a processing device 4 connected to the acoustic sensor and a memory device, ie variable RAM and ROM for constant data and code programs, connected to the processing device. Also, for example, the acoustic sensor is fixed on the surface 1 of the object. The surface 1 may be subdivided into M sensitive areas 6.

가 단계 S102(샘플링) 동안 처리 장치(4)에서 샘플링(sampled)되고 저장된다. 이후, 처리 단계 S103(처리)가 실행된다. Before operation, as shown in FIG. 3, initialization mode S100 (initialization) is executed. This initialization mode will be described later. In step S101, a signal test (shock?) Is used to detect the shock. For example, if the signal sensed by the sensors exceeds the determined threshold, then the sensed signals transmitted by the N sensors

Is sampled and stored in the processing apparatus 4 during step S102 (sampling). Thereafter, processing step S103 (processing) is executed.

과 동일하고, 여기서 N은 센서의 개수이다. 각각의 상호상관 곱

는

와 동일하고, 여기서

는 센서 i에 의해 감지된 감지 신호

의 푸리에 변환이고,

는 센서 j에 의해 감지된 감지 신호

의 푸리에 변환이고, *는 복소수 공액 연산자이다. 푸리에 변환은 예를 들어 낮은 계산 전력을 필요로 하는 고속 푸리에 변환(FFT)이다. 실시예에 따라, 상호상관 곱은 예를 들어 에너지로 정규화된다. As shown in more detail in FIG. 4, the processing step starts cross-correlation product calculation (cross-correlation) in step S200. In practice, P independent cross-correlation products are performed. Where P is the sum of N-1 first integers, i.e.

, Where N is the number of sensors. Cross-correlation product of each

Is

Is the same as

Is the sensed signal detected by sensor i

Fourier transform of,

Is the sensed signal detected by sensor j

Is the Fourier transform of, and * is a complex conjugate operator. The Fourier transform is, for example, a fast Fourier transform (FFT) that requires low computational power. According to an embodiment, the cross-correlation product is normalized by energy, for example.

은 충격 파형과 시간 원점에 결정적으로 좌우되지 않는다. 사실,

을 고려하는 경우, 여기서 C_i는 센서 i의 주파수 복소수 응답이고, d_i는 충격 위치와 센서 i사이의 거리이고, E는 충격 파형 푸리에 변환이다. 이 예에서, 파수 k= ω/C이고, 여기서 C는 음향 전파 속도이다.이후, 상호상관 곱은This cross-correlation product

Is not critically dependent on the impact waveform and the time origin. Actually,

Is considered, where C _i is the frequency complex response of the sensor i, d _i is the distance between the impact location and the sensor i, and E is the shock waveform Fourier transform. In this example, the wave number k = ω / C, where C is the sound propagation velocity.

또는,

or,

와 동일하다.

Is the same as

이 실수임에 따라, 시간 상호상관 곱

은 주로 센서 응답에 좌우된다. 이후 큰 센서 응답 대역폭이, 충격 여진(impact excitation) 주파수 대역폭의 경로(course)의 한계에서, 실제로 짧은 임펄스를 얻도록 허용한다. 이 예에서, 센서 대역폭은 실질적으로 ［100 Hz, 7 kHz］이다. 또한, 상호상관 곱이 정규화된다면, 감지된 신호의, 그리고 이때 충격 강도의 진폭에 좌우되지 않는다.

Since this is a real number, time cross-correlation product

Depends mainly on the sensor response. The large sensor response bandwidth then allows to actually get a short impulse, at the limit of the course of the impact excitation frequency bandwidth. In this example, the sensor bandwidth is substantially [100 Hz, 7 kHz]. Also, if the cross-correlation product is normalized, it does not depend on the amplitude of the sensed signal and then the impact intensity.

의 역 푸리에 변환이 연산된다. 제 1 실시예에 따라, 대상물의 물질은 비 분산이다. 이는 음향 신호의 파 전파 속도가 이 신호의 주파수에 좌우되지 않는다는 것을 의미한다. 결과적으로, 이 물질에서 전파된 감지 된 신호의 파형은 근원(source)과 센서 사이의 거리 및 물질 속도의 함수인 지연의 지연된 원 신호와 실질적으로 동일하다. 이 경우에, 역 푸리에 변환은 일반적인 역 푸리에 변환 또는 역 고속 푸리에 변환일 수 있다. Then, in step S201,

The inverse Fourier transform of is computed. According to the first embodiment, the material of the object is non dispersion. This means that the wave propagation speed of the acoustic signal does not depend on the frequency of this signal. As a result, the waveform of the sensed signal propagated in this material is substantially the same as the delayed original signal of the delay, which is a function of the distance between the source and the sensor and the material velocity. In this case, the inverse Fourier transform may be a general inverse Fourier transform or an inverse fast Fourier transform.

(S203)의 연산을 그 후 수행하기 위해 1로 초기화되고, 여기서

는

의 역 푸리에 변환이다. 비 분산 물질에 대한 현재 경우에,

는 각각 영역 k에서 센서 i의 거리와 영역 k에서 센서 j의 거리 사이의 차에 기초한 지연이다. 이들 값은 장치(1)의 메모리 수단(5)에서 예를 들어 저장될 수 있다. 이들은 연산자에 의해 초기화 모드 S100에서 연산되거나 또는 이 장치의 설계 동안 결정될 수 있다. In step S202, k sums up for area k of each surface

Is initialized to 1 to perform the operation of (S203) thereafter, where

Is

Inverse Fourier transform. In the current case for non-dispersed materials,

Are respectively delays based on the difference between the distance of sensor i in area k and the distance of sensor j in area k. These values can for example be stored in the memory means 5 of the device 1. These can be calculated in the initialization mode S100 by the operator or determined during the design of the device.

는 동위상이고,

는 실질적으로 임펄스일 것이다. 대조적으로, 충격이 영역 k에서 생성되지 않는다면,

는 넓을 것이고 낮은 진폭을 가질 것이다. 이는 충격이 발생되었던 영역을 쉽게 결정하도록 한다. Thus, if an impact occurs in area k, then

Is in phase,

Will be substantially an impulse. In contrast, if an impact is not produced in region k,

Will be wide and have a low amplitude. This makes it easy to determine the area where the impact occurred.

의 특성 값이 계산되고 모든 특성 값 f(

)가 비교된다. 이 특성 값은 다음일 수 있다:This calculation is performed in loop steps S204 and S205 for all regions k. In step S206, each of

Is calculated, and all property values f (

) Is compared. This property value can be:

의 최대 값ㆍ

Maximum value of

의 파워, 또는 최대 제곱 진폭ㆍ

Power, or maximum square amplitude

Maximum peak-to-peak amplitude;

의 RMS(root mean square); 또는ㆍ

Root mean square of RMS; or

의 에너지, 이는

와 동일하고;ㆍ

Of energy, which

Same as;

의 펄스 폭 매개변수, 이는

와 동일하고, 여기서 REAL(x)는 복소수 x의 실수부이다. ㆍ

Of the pulse width parameter, which is

, Where REAL (x) is the real part of complex x.

)을 가지는 영역 k에 상응한다. 이때, 단계 S207에서, 가장 큰 값은, 검출된 충격이 영역 k₀에 있었다는 것을 보증하며, 신뢰의 임계와 비교될 수 있다. 실제로, 충격은 잡음 때문에, 또는 M 개의 결정된 영역에서보다 다른 영역에서 생성된 충격 때문에, 간섭(interference)일 수 있다. 이후, 이 마지막 테스트는 이 장치에 의해 제어된 바람직하지않은 행동을 피하도록 허용한다.In this case, the area k ₀ has the largest characteristic value f (

Corresponds to area k with At this time, in step S207, the largest value ensures that the detected shock was in the region k ₀ and can be compared with a threshold of confidence. In practice, the impact may be interference because of noise, or because of the impact generated in other regions than in the M determined regions. This last test then allows to avoid undesirable behaviors controlled by this device.

에 의해 정의된 콘트라스트의 테스트로 대체될 수 있거나 또는 결합될 수 있고, 여기서 MEAN은 k=k₀에 대한 것을 제외하고 f(

)의 모든 값의 평균 연산자이다. 콘트라스트 값이 기설정된 임계보다 더 큰 경우, 결과는 유효하다고 고려된다. 이후, 영역 k₀의 특성 값이 유효로서 고려된다면, 이 영역에서 충격이 일어났었다고 고려된다. 작동은, 예를 들어 이 장치가 인터페이스로서 사용된다면, 장치(3)에 의해 구동될 수 있다. 그렇지 않다면, 이 충격은 장치에 의해 무시될 수 있다. Also, this test is non

It can be substituted or combined with a test of contrast defined by MEAN, where MEAN is f (except for k = k ₀₎ .

Is the average operator of all values. If the contrast value is greater than the predetermined threshold, the result is considered valid. After that, the characteristic value of the area k ₀ If considered effective, it is considered that an impact has occurred in this area. The operation can be driven by the device 3, for example if the device is used as an interface. If not, this impact can be ignored by the device.

와 같은 보간(interpolation) 함수로 결과를 조정하는 것이 가능하다. It also allows to improve the resolution of this method,

It is possible to adjust the result with an interpolation function such as

This embodiment then provides a method of determining the location of the impact on the surface without precise synchronization, in particular for non-disperse materials.

However, in the case of a dispersing material, a second embodiment is provided by the present invention. The main difference in this embodiment is that the inverse Fourier transform of step S201 is a dispersion compensated Fourier transform. In practice, the dispersing material is a material whose speed of sound signal propagation depends on the frequency of the signal. Thus, the waveform of the sensed signal propagated in this material is different from the original signal waveform. In this case, as described below, the dispersion compensated Fourier transform must compensate for distortion of the signal due to propagation.

P.D. Following the progress of the Review of Progress in Quantitative Non-Destructive Evaluation entitled "A Signal processing technique to remove the effect of dispersion from guided wave signals" from Wilcox & al, It is possible to determine the distance:

, 여기서

는 감지된 음향 신호의 푸리에 변환이고, k(ω)은 관계 분산(relation dispersion)에 따른 파수이다. k(ω)= Ω와 ω =m(Ω)이라고 고려되는 경우, 상기 거리는 다음 공식으로부터 얻는다:

, here

Is the Fourier transform of the sensed acoustic signal, and k (ω) is the wave number according to the relation dispersion. Considering k (ω) = Ω and ω = m (Ω), the distance is obtained from the following formula:

, 이는

의 역 푸리에 변환이다. 얻어진 신호의 형태가 변경되지만, 그러나

가 일반적으로 완전히 실수이기 때문에 지연정보는 동일하다. 이후, 이 분산 보상 역 푸리에 변환이 제 1 실시예를 개선하기 위해 이용될 수 있고, 따라서 분산 물질 대상물의 표면 상에 충격을 결정하기 위한 방법을 제공한다.

, this is

Inverse Fourier transform. Although the shape of the obtained signal changes, but

The delay information is the same because is generally a full mistake. This dispersion compensation inverse Fourier transform can then be used to improve the first embodiment, thus providing a method for determining the impact on the surface of the dispersion material object.

과 동일하다는 것을 제외하고는, 실질적으로 동일하다. 이후, 이 푸리에 변환은 신호의 분산을 보상하도록 허용하고, 이에 따라 처리될 수 있는 신호를 제공한다. Thus, in the method according to the second embodiment, in step S201, the calculated inverse Fourier transform is performed.

Substantially the same, except that This Fourier transform then allows to compensate for signal variance and thus provides a signal that can be processed.

값은 영역 k에서 센서i의 거리와 영역 k에서 센서 j의 거리 사이의 차에 기초한 길이이다. 나머지 방법은 바뀌지 않는다.In this embodiment, in step S203

The value is a length based on the difference between the distance of sensor i in area k and the distance of sensor j in area k. The remaining methods do not change.

와 동일하고, α 는 물질에 따른 계수이다. In most of the examples, the wave generated by the impact on a substantially flat object can be considered a Lamb wave. As the waves are examined at low frequencies in this embodiment of the present invention, we can consider these waves to be mainly zero order lamb waves. After all, as the impact is created on the surface, the generated wave is an asymmetric zero order lamb wave. Thus, at low frequencies, the variance relationship is substantially

Is the same according to the material.

At this time, it is possible to use the method of acoustic imaging as a dispersing material for determining the location of the impact on the surface. When the following dispersion relationship is used, the method of this second embodiment can also be applied to non-dispersed materials: Ω = αω. Thus, this embodiment provides a general method applicable to any kind of dispersed or non-dispersed material. In practice, it is easy to interpolate the dispersion relationship of propagated waves over the full range of frequencies. The range of frequencies depends on the bandwidth of the material, the application, and the sensor. This embodiment is not limited to the current example of an asymmetric zero order lamb wave.

에 좌우되고,

는 영역 k와 센서 i의 각각의 위치에 기초하여 설정된 각도값이다. In addition, this coefficient α may also depend on the temperature of the material and this parameter may be considered. For example, a temperature sensor can be fixed to the object, and the coefficient α can be changed based on this coefficient. Thus, it is possible to provide a method which takes into account the temperature of the material. Likewise, if the material is anisotropic, ie the propagation speed depends on the direction of signal propagation, this can be compensated by the coefficient α which depends on the passage between the sensor i and the area k. For example, α

Depends on,

Is an angle value set based on each position of the area k and the sensor i.

According to the flow chart of FIG. 3, for example, the method starts with the initialization mode described by the flow chart of FIG. 5. During this initialization mode, the exact position of the sensor can be entered and stored in the device 3 (S300), but these positions can also be entered during the design of the device.

Thereafter, the dispersion relationship is determined. In general, the form of the dispersion relationship is known. This step is directed to provide accurate coefficients for interpolating this variance relationship. In the case of an asymmetric zero order lamb wave, we must determine the coefficient α. However, if the signal is examined at a different range of frequencies, it may require polynomial interpolation of the variance relationship. At this time, all polynomial coefficients must be determined.

이 연산되고 메모리에 저장된다. 실시예에서, 충격은 2 개의 센서에 연결되어 있는 선을 따라 생성되고, 균등하게 이격된다(spaced). 이는 최적값 α 계수를 쉽게 찾도록 한다. Also, although the material is considered as a non-dispersed material, this step can be used to determine the correct coefficient α corresponding to this non-dispersed material. In order to achieve the determination of the coefficient α, an NSO impact can be generated on the surface of the material (S301-S305) and a delay corresponding to each impact m

Is computed and stored in memory. In an embodiment, the impact is generated along a line connected to two sensors and evenly spaced. This makes it easy to find the optimal value α coefficient.

에서 검사될 것이다. 이 테스트값 α는 예를 들어 증분(increment) △α으로 균등하게 이격된다(S313). 각각의 테스트 값에 대해, 충격 m과 센서 i 및 j에 상응하는 상호상관 곱

는 이 테스트값 α로 연산되고(S309)

를 형성하기 위해 합해진다(S309). 이후, 모든 충격의

가 더해지고, 이 합의 최대 값이 단계 S312에서 저장된다. 따라서, 각각의 테스트 값α에 대해, 최대 F(α)은 저장된다. Then, the α test value is in the range set in step S306 to step S314.

Will be checked in. This test value α is equally spaced by increment Δα, for example (S313). For each test value, the cross-correlation product corresponding to impact m and sensors i and j

Is calculated with this test value α (S309)

It is combined to form a (S309). Since, all of the shock

Is added, and the maximum value of this sum is stored in step S312. Thus, for each test value α, the maximum F (α) is stored.

로 선택되고 장치의 동작동안 사용될 수 있다. 또한, 이분(dichotomy) 방법 단독으로 또는 마지막 방법과 결합하여 최적의 계수를 검색하는 데 사용될 수 있다. Then, in step S315, the largest maximum value, max (F (α)) is considered as the best value, so the corresponding α is the optimal coefficient

Can be selected and used during operation of the device. It can also be used to retrieve optimal coefficients, either alone or in combination with the last method.

Thus, it is possible to provide a method for determining the location of an impact on a surface with effective material dispersion compensation and based on acoustic imaging.

According to yet another embodiment, the initialization mode comprises determining a corresponding delay of the location and the M regions. For example, the operator may have to generate a shock on each sensing area, and the device stores a corresponding delay in the ROM 5 of the device.

Included in this specification.

Claims (18)

The surface comprises N acoustic sensors 2a, 2b, 2c, and M determined regions, where N is at least 3, an impact generates an acoustic signal, and each sensor i receives and processes the acoustic signal. Signal detected by device (4)

As a method of determining the impact location on the surface (1) of the object, (a)

Is the detected signal detected by sensor i

Fourier transform of,

Is the detected signal detected by sensor j

Fourier transform of, ＊ is a complex conjugate operator P cross-correlation product

Calculating a; (b) above

P Inverse Fourier Transform

Calculating; (c)

Is the determined value based on the individual positions of region k and the individual positions of sensors i and j, for each region k,

Wow

Calculating a characteristic value of; And (d)

The characteristic value of is k ≠ k _o

And determining k _o , which is greater than the corresponding characteristic value of. The method of claim 1, Inverse Fourier Transform

silver, When Ω is a wavenumber coefficient and m (Ω) is a frequency corresponding to the wavenumber coefficient Ω according to the material dispersion relation,

Impact location determination method on the surface of the same object as. The method of claim 2, U is the distance and

Is a length that depends on the distance between area k and sensor i and the distance between area k and sensor j. The method of claim 3, wherein The material dispersion relationship is, If α is a coefficient according to the object,

Impact location determination method on the surface of the same object as. The method of claim 3, wherein The material dispersion relationship is A method of determining the impact position on the surface of an object, where α is a coefficient according to the object, substantially equal to Ω = αω. The method according to claim 4 or 5, The coefficient α is

Depends on,

Is a determined angle value based on the respective position of region k and sensor i. The method according to claim 4 or 5, The coefficient α is the temperature of the object

The method of determining the impact position on the surface of an object, wherein T is a temperature value. The method of claim 2, And an initialization mode comprising determining a mass dispersion relationship of the object. The method of claim 8, Determining the material dispersion relationship includes the following sub-steps of generating an NSO impact at the determined location. The method of claim 9, And the NSO impact is generated at a position determined along a line connecting two of the N sensors and evenly spaced on the surface of the object. The method according to claim 9 or 10, Determining the material dispersion relationship With α test value

A sub-step of computing; For each α test value, for each impact m

To compute and for each α test value,

Sub-steps of summing; And Sum

Which gives the largest maximum of

And determining a location of impact on the surface of the object further comprising a sub-step of selecting. The method of claim 1, The method of determining the impact location on the surface of an object further comprising an initialization mode comprising determining the position of the N sensors (2a, 2b, 2c) on the surface (1). The method of claim 1, The cross-correlation product

The number of

A method of determining the impact location on the surface of the same object as. The method of claim 1,

The characteristic value of -

Maximum value of; -

Power, or maximum square amplitude;  Maximum peak-to-peak amplitude; -

Root mean square (RMS) of; or -

Of energy, which

Same as; REAL (x) is the real part of complex x,

Same as

A method of determining the impact location on the surface of an object, which is one of the pulse width parameters of. The method of claim 1,

A method of determining the location of an impact on a surface of an object that concludes that an impact occurred in region k _o only if the characteristic value of is greater than a predetermined threshold of confidence. The method of claim 1,

A method of determining the impact location on the surface of an object that concludes that an impact occurred in the region k _o only if the contrast value defined by the ratio is greater than the predetermined threshold of confidence, MEAN is the average operator. The method of claim 1, remind

A method of determining the impact location on the surface of an object where is normalized. A device for determining the impact location on the surface of an object, wherein the surface 1 of the object comprises M determined regions 6 and the impact on the surface generates an acoustic signal, The determination device A processing device 4; A signal in which N is at least 3 and each sensor i receives the acoustic signal and is sensed by a processing device

In case of transmitting, N acoustic sensors (2a, 2b, 2c) installed on the surface (1), The processing device 4 (a)

Detected signal detected by sensor i

Fourier transform of,

Detected signal detected by sensor j

Fourier transform of, If ＊ is a complex conjugate operator, P cross-correlation product

Means for calculating; (b) above

P Inverse Fourier Transform

Means for calculating; (c)

Is a determined value based on the individual positions of the regions k and the individual positions of the sensors i and j, for each region k,

Wow

Means for calculating a characteristic value of; And (d)

Has a characteristic value of k ≠ k _o

And a means for determining k _o that is greater than the corresponding characteristic value of.

Images