CN102889926B - Ultrasonic wave anti-noise detection method and system - Google Patents

Abstract

The invention discloses an ultrasonic wave anti-noise detection method and an ultrasonic wave anti-noise detection system. The method comprises the following steps of: firstly, performing generalized time-frequency domain average analysis on an ultrasonic wave signal which is obtained by repeated measurement, and thus obtaining a time-frequency analysis chart; secondly, identifying time and frequency corresponding to an energy center in the time-frequency analysis chart; and finally, slicing at the frequency corresponding to the energy center along a time shaft, and thus obtaining the detected ultrasonic wave signal. The invention has the advantages that by using a generalized time-frequency domain average technology, the defects of propagation path deflection caused by an installation effect of an ultrasonic wave detection device in a speed field, a pipeline manufacturing process, energy attenuation of ultrasonic wave in gas and low ratio of sound speed to flow speed, channel gain variation caused by gas pressure fluctuation, acoustic interference caused by a gas pressure regulating device, and the like are overcome, so the detection accuracy of the ultrasonic wave is high.

Description

A kind of ultrasound wave anti-noise detection method and system

Technical field

The invention belongs to the ultrasonic signal detection field, be specifically related to a kind of ultrasound wave anti-noise detection method and system.

Background technology

Ultrasound wave has that frequency is high, wavelength is short, the diffraction phenomenon is little, particularly good directionality, can become ray and the characteristics such as direction propagation.Ultrasound wave can be propagated in various different media, and moderate with the interaction of propagation medium, the medium status information of the relevant ultrasonic propagation that is easy to carry about with one.As information carrier and form of energy, the combinations such as ultrasonic technology and other electronic technology, optical technology extensively are applied, and develop rapidly.In recent years, based on hyperacoustic various application, because making it, its security, robustness and satisfied precision be used widely in fields such as ultrasonic flow rate measurement, distance/wall thickness measuring and Ultrasonic NDT.

At present, most ultrasonic applications all are based on and estimate transmission time (Time of Arrival, TOA) or the flight time (Time of Flight, TOF) of ultrasound wave in medium.Yet the ultrasound echo signal in above-mentioned application, owing to disturbed by noise or other signals, causes amplitude or waveform distortion etc. problem.Ultrasonic signal itself shows very strong non-stationary, non-linear characteristics.Above factor often detects and brings difficulty to TOA/TOF, finally causes ultrasonic applications to realize.Up to the present, existing multi-signal treatment technology is applied in the middle of ultrasonic applications, such as cross-correlation method, match tracing, Kalman filtering, method of estimation, wavelet transformation and empirical mode decomposition etc. based on model.Yet the noise in signal and other interference components are also to cause above-mentioned these technology can not the hyperacoustic principal element of fine detection.

The general temporal evolution of the instantaneous frequency of non-stationary signal and energy, the fundamental purpose of time frequency analysis be exactly the research instantaneous frequency of signal content and amplitude/energy thereof the time become situation.Various time-frequency analysis technologies can provide the time domain of signal, the information of frequency domain simultaneously, but different Time-Frequency Analysis Method has different time frequency resolution and concentration of energy.Time-frequency analysis technology is various features in detection signal well, but can't distinguish useful and useless information.

Summary of the invention

Technical matters to be solved by this invention is to provide a kind of ultrasound wave anti-noise detection method and system, by the ultrasonic signal to repeatedly measuring, carry out broad sense time-frequency domain average treatment, the impact of the various interference such as noise in erasure signal, obtain effectively, ultrasonic signal reliably.

For addressing the above problem, the present invention is achieved by the following scheme:

A kind of ultrasound wave anti-noise detection method, comprise the steps:

(1) set sample frequency and sampling number, record the ultrasound data of repeatedly measuring;

(2) respectively the ultrasound data of repeatedly measuring is carried out to time-frequency conversion;

(3) the time-frequency conversion result of the ultrasound data that will repeatedly measure is carried out superposed average, obtains the Time-Frequency Information matrix, and generates broad sense time-frequency domain mean chart according to the Time-Frequency Information matrix;

(4) utilize the extreme value finding method, corresponding time of center of energy and the frequency information of search broad sense time-frequency domain mean chart;

(5) at the maximum frequency point of center of energy, along the time shaft section, obtain the maximum frequency place and be detected ultrasonic signal.

In above-mentioned steps (1), the ultrasound data of sampling need, successively through after analog to digital conversion step and pre-filtering step, just go on record.

The described time-frequency conversion of above-mentioned steps (2) is selected linearity, secondary and/or self-adaptation time-frequency conversion method.

A kind of ultrasound wave anti-noise detection system comprises as lower unit:

Record cell, according to sample frequency and the sampling number set, record the ultrasound data that ultrasonic receiver is repeatedly measured;

The time-frequency conversion unit, carry out time-frequency conversion to the ultrasound data of repeatedly measuring respectively;

Broad sense time-frequency domain averaging unit, carry out superposed average by the time-frequency conversion result of the ultrasound data repeatedly measured, obtains the Time-Frequency Information matrix, and generate broad sense time-frequency domain mean chart according to the Time-Frequency Information matrix;

The extrema-finding unit, utilize the extreme value finding method, corresponding time of center of energy and the frequency information of search broad sense time-frequency domain mean chart;

The signal determining unit, at the maximum frequency point of center of energy, along the time shaft section, obtain the maximum frequency place and be detected ultrasonic signal.

Above-mentioned ultrasound wave anti-noise detection system also further comprises AD conversion unit and pre-filtering unit, wherein AD conversion unit is carried out analog to digital conversion to the ultrasound data of sampling, and the data of pre-filtering unit after to analog to digital conversion are carried out after pre-filtering sending into record cell and carried out record.

In such scheme, described time-frequency conversion unit selects linearity, secondary and/or self-adaptation time-frequency conversion method to carry out time-frequency conversion to the ultrasound data of repeatedly measuring.

The present invention is based on broad sense time-frequency domain averaging and realize that ultrasonic signal detects.The first step is advanced with the average analysis of broad sense time-frequency domain repeatedly measuring ultrasonic signal, obtains its time frequency analysis figure.Corresponding time of center of energy and frequency in second step identification time-frequency figure.When cutting into slices along time shaft at this frequency place, obtain detected ultrasonic signal.

Compared with prior art, the present invention has following features:

1, the present invention can overcome ultrasonic detection device in less the brought travel path deflection of installation effect, pipeline manufacturing process, ultrasonic energy attenuation, the velocity of sound/flow velocity ratio in gas of velocity field distribution, the channel gain change that the gaseous tension fluctuation causes, the acoustic interference that gas pressure regulating apparatus causes etc.;

2, the ultrasound wave time precision that the present invention detects is high;

3, the present invention can be different types of ultrasonic applications device use such as ultrasonic flowmeter, supersonic diagnostic set, and therefore, the wide popularization and application had is worth.

The accompanying drawing explanation

Fig. 1 is the scheme of installation of the present invention in the ultrasonic gas flow velocity measuring system.

Fig. 2 is that broad sense time-frequency domain averaging and the ultrasound wave peak value while adopting SPWV, CWT and HHT detects schematic diagram (this schematic diagram only means the averaging of time frequency analysis of the same race).

Fig. 3 (a)-(d) for ultrasonic signal, carry out the time domain average analysis result under different operating modes.

Fig. 4 (a1)-(c1) ultrasonic signal when P=0psig carries out broad sense time-frequency domain average analysis result, i.e. broad sense time-frequency domain mean chart.

Fig. 4 (a2)-(c2) is ultrasonic signal peak value slice map when P=0psig.

Embodiment

Content below in conjunction with accompanying drawing embodiment in ultrasonic flowmeter to tool of the present invention is described in further detail:

The ultrasonic flow detection technique is one of technology with fastest developing speed in instrument monitoring, control and fields of measurement in recent years.Ultrasonic flowmeter is divided into gas and liquid two large field types, both ultimate principles are identical, all utilize ultrasound wave respectively when mobile and static Propagation, with respect to fixed coordinate system (as instrument housing), its velocity of propagation is discrepant, and the variable quantity of velocity of propagation is relevant with velocity of medium.Can obtain velocity of medium according to the variation of ultrasonic propagation velocity thus, under the prerequisite of known cross-sectional flow area and then obtain flow.Yet, various unfavorable factors due to outwardness in the measurement of gas flow, comprise less the brought travel path deflection of installation effect, pipeline manufacturing process, ultrasonic energy attenuation, the velocity of sound/flow velocity ratio in gas of velocity field distribution, the channel gain change that the gaseous tension fluctuation causes, the acoustic interference that gas pressure regulating apparatus causes etc., make the ultrasonic gas flow detection compare liquid extremely difficult.

The present embodiment adopts the broad sense time-frequency domain on average the ultrasound wave in ultrasonic flowmeter to be detected.

The collection of the present embodiment external ultrasonic signal adopts clamp-mounting type (clamp-on) ultrasonic gas flow quantity detecting system, and its mounting structure as shown in Figure 2.A pair of ultrasonic generator and ultrasonic receiver are installed on the pipeline that diameter is 6 inches (about 152.4mm) face-to-face.The amplitude that can inspire certain frequency when ultrasound transmitter device A is triggered by monopulse is the vibration ultrasound wave of first increases and then decreases gradually, and this ultrasound wave is received by ultrasonic receiver B after penetrating pipeline.Although, experimental results demonstrate interference and the loss of signal of 500kHz resonant frequency energy noise decrease.But, due to the intrinsic influence factor that is clamp-mounting type and gas flow measurement, make the signal received still have the problems such as signal to noise ratio (S/N ratio) is low, waveform distortion.Particularly lower (such as 0psig at gaseous tension, 1psi=6.89kPa psig is gauge pressure here, 0psig means that external and internal pressure equates, be also 1atm or 0.1MPa) and flow velocity when very fast signal interference problem even more serious, this brings great challenge to traditional ultrasonic detection method.

Gas velocity and air in pipeline pressure P are controlled by servomotor frequency (Motor Frequency, MF) and flow control valve respectively.In the larger corresponding system of MF numerical value, gas flow rate is faster, and in the present embodiment, the servomotor frequency is that the corresponding flow velocity of 60Hz(is about 121.6ft/s or 37.06m/s).In addition, the temperature variation of gas can be brought impact to measuring-signal, so proving installation outfit cooling system can make temperature remain on 80 ℉ (approximately 26.7 ℃) left and right.

The signal that ultrasound wave anti-noise detection system in computing machine collects ultrasonic receiver according to following steps is processed, and this implements described ultrasound wave anti-noise detection method as shown in Figure 2.

(1) set sample frequency and sampling number, record the ultrasound data of repeatedly measuring.

In the present embodiment, the setting sample frequency is 20MHz, and sampling number is 2048 points.

(2) respectively the ultrasound data of repeatedly measuring is carried out to time-frequency conversion.In the present invention, ultrasound data is carried out to time-frequency conversion and can select linearity, secondary and/or self-adaptation time-frequency conversion method.

For signal s (t), can adopt linear time-frequency conversion method to have: Short Time Fourier Transform (STFT), continuous wavelet transform (CWT) etc.

If the employing Short Time Fourier Transform, the time-frequency conversion of signal s (t) is

$W_{\mathrm{STFT}}^{s,g}(t,f)=\underset{-\&infin;}{\overset{\&infin;}{\&Integral;}}s\left(\mathrm{\&tau;}\right)g^{*}(\mathrm{\&tau;}-t)e^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="HDA00002314123200031.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}\&CenterDot;\mathrm{\&tau;}}\mathrm{d\&tau;}$

Wherein, g (t) is institute's added-time window function.

If the employing wavelet transformation, the time-frequency conversion of signal s (t) is

$W_{\mathrm{WT}}^{s,\mathrm{\&psi;}}(t,a)\stackrel{\mathrm{\&Delta;}}{=}\frac{1}{\sqrt{a}}\underset{-\&infin;}{\overset{\&infin;}{\&Integral;}}s\left(\mathrm{\&tau;}\right){\mathrm{\&psi;}}^{*}\left(\frac{t-\mathrm{\&tau;}}{a}\right)\mathrm{d\&tau;}$

Wherein, ψ is mother wavelet function.ψ is a function that is zero in the time shaft upper integral, to ψ, after flexible and translation, can obtain small echo atom ψ _{τ, a}(t)=a ^1/2ψ ((t-τ)/a), mother wavelet function is chosen one dimension Gabor small echo, and its expression formula is formula

$\mathrm{\&psi;}\left(t\right)=\frac{1}{\sqrt{2\mathrm{\&pi;}{\mathrm{\&sigma;}}_t^2}}{e}^{-{\frac{{\mathrm{<figure-callout\; id="2"\; label="t"\; filenames="HDA00002314123200031.png"\; state="\{\{state\}\}">t</figure-callout>}}^2}{2{\mathrm{\&sigma;}}_t^2}}_{{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="HDA00002314123200031.png"\; state="\{\{state\}\}">e</figure-callout>}}^j}}$

Wherein, σ _tthe frequency span that means small echo, f ₀centre frequency for small echo.

For signal s (t), can adopt secondary time-frequency conversion method to have: Cohen class time-frequency distributions, smooth Wigner-Ville time-frequency distributions (SPWV) etc.

If adopt Cohen class time-frequency distributions, the time-frequency conversion of signal s (t) is

$W_K^s(t,f)={\&Integral;}_{-\&infin;}^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}K(t,f;\mathrm{\&upsi;},\mathrm{\&tau;})s(t+\frac{\mathrm{\&tau;}}{2})s^{*}(t-\frac{\mathrm{\&tau;}}{2})\mathrm{d\&upsi;d\&tau;}$

Wherein, K (t, f; υ, τ) be called the kernel function that meets certain condition.

If adopt smooth pseudo-Wigner-Ville time-frequency distributions, the time-frequency conversion of signal s (t) is

$W_{\mathrm{SPWV}}^{s,g,h}(t,f)\stackrel{\mathrm{\&Delta;}}{=}\underset{-\&infin;}{\overset{\&infin;}{\&Integral;}}g(u-t)\underset{-\&infin;}{\overset{\&infin;}{\&Integral;}}h\left(\mathrm{\&tau;}\right)s(u+\frac{\mathrm{\&tau;}}{2})s^{*}(u-\frac{\mathrm{\&tau;}}{2})e^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="HDA00002314123200031.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f\&tau;}}\mathrm{d\&tau;du}$

G in above formula (t) and h (t) are windowed function.Smooth pseudo-Wigner-Ville distributes and can effectively suppress cross term.

For signal s (t), can adopt self-adaptation time-frequency conversion method to have: Hilbert-Huang spectrum (HHS) etc.

If adopt the Hilbert-Huang spectrum, the time-frequency conversion of signal s (t) is

Hilbert-Huang that instantaneous amplitude and instantaneous frequency are constructed spectrum (Hilbert-Huang Spectrum, HHS) can the concentrated expression signal characteristic information, and the Time-Frequency Information matrix can be obtained by following formula.

$W_{\mathrm{HHT}}^s(f,t)\stackrel{\mathrm{\&Delta;}}{=}{(f_i\left(t\right),a_i\left(t\right))}_{i=1,...,n,j=1,...,n}$

(3) the time-frequency conversion result of the ultrasound data that will repeatedly measure is carried out superposed average, obtains the Time-Frequency Information matrix, and generates broad sense time-frequency domain mean chart according to the Time-Frequency Information matrix.Above-mentioned time-frequency conversion result of carrying out superposed average can be the time-frequency conversion result that adopts same time-frequency transformation analysis method to obtain, and can be the time-frequency conversion result that adopts time-frequency transformation analysis method not of the same race to obtain.

By broad sense time-frequency average treatment, obtain the time-frequency figure after noise is eliminated.?

If s _i(t) be one group of time-domain signal in N group measuring-signal,

s for this reason _i(t) the time-frequency conversion result of signal, T means adopted a certain time-frequency conversion, it can be linearity, secondary and adaptive Time Frequency Analysis method.For N group signal, the Time-Frequency Information matrix obtained after time-frequency domain is average

for

${\stackrel{\&OverBar;}{W}}_T^s(t,f)=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{W}_T^{s_i}(t,f)$

If time-frequency conversion adopts, be smooth pseudo-Wigner-Ville while distributing, the average result of broad sense time-frequency domain is available

mean.If time-frequency conversion adopts while being a certain time-frequency computing in SPWV, Scalogram or HHS, correspondingly the average result of broad sense time-frequency domain can be noted by abridging respectively and is

with

if time-frequency conversion adopts wavelet transformation and two kinds of Time-frequency method analytic signal s of Bilinear TFD (t), now will obtain the broad sense time-frequency domain averaging similar to formula and be

${\stackrel{\&OverBar;}{W}}_{\mathrm{\&Xi;}\&cup;\mathrm{\&Omega;}}^s(t,f)=\frac{1}{J}\underset{j=1}{\overset{J}{\mathrm{\&Sigma;}}}{W}_{\mathrm{\&Xi;}}^{{\mathrm{\&Omega;}}_s(t,f^j)}(t,f)$

Wherein: Ξ ∪ Ω means the joint time-frequency operation.

(4) utilize the extreme value finding method, corresponding time of center of energy and the frequency information of search broad sense time-frequency domain mean chart;

(5) at the maximum frequency point of center of energy, along the time shaft section, obtain the maximum frequency place and be detected ultrasonic signal.

Below by the validity of concrete data verification said method:

At MF=60Hz and P=0psig(0.1Mpa), 5psig(0.134Mpa), 15psig(0.203Mpa) or the about 0.445Mpa of 50psig() in situation, following current (being ultrasonic generator A → B direction in Fig. 2) gathers respectively 80 groups of data with adverse current (B → A direction).All data store in computing machine after A/D conversion and near pre-filtering (500kHz).

Along with the increase of gaseous tension, the signal to noise ratio (S/N ratio) of following current and adverse current collection signal all is enhanced, signal when during following current, collection signal is subject to the impact of external interference to be less than adverse current.Therefore, at low pressure (as P=0psig and 5psig), countercurrent direction signal, be subject to extraneous interference larger, signal to noise ratio (S/N ratio) is lower.To time domain average technology (TDA) analysis result of 80 groups of data gathering under above-mentioned every kind of operating mode as shown in Figure 3, can find out when P=0psig and 5psig, TDA detects effect neither be desirable especially, and reason is that the TDA technology is limited in one's ability when processing waveform distortion or exist than the signal of strong jamming composition.Therefore, the present embodiment, mainly for above-mentioned P=0psig and two kinds of operating modes of 5psig, adopts broad sense time-frequency domain averaging (ETFDA) technology to be analyzed, and carrys out the validity of verification method.

When p=0psig, AVSPWV, AVSG and AVHHS analysis result be respectively as shown in Fig. 4 (a1), (b1), (c1), therefrom can obviously find out that three kinds of ETFDA technology locate successfully to detect the ultrasonic signal (arrow indication in figure) of following current and countercurrent direction in 500kHz frequency of operation (or yardstick 50).Fig. 4 (a2), (b2), (c2) show the 500kHz frequency place section of corresponding three time-frequency figure, and following current (solid line) all detects effectively with adverse current (dotted line) direction ultrasound wave.In such cases, the ultrasound wave moment point (in Fig. 4 (a2), (b2), (c2), dotted line means) that corresponding A VSPWV, AVSG and AVHHS technology for detection go out following current and countercurrent direction is respectively 611 and 868,611 and 869,609 and 862.Therefore, even when there is strong interference component in signal, three kinds of ETFDA technology still successfully detect the ultrasound wave in signal, and the time is basically identical, and the mistiming is about 12.85us(257/20M).

Designed a kind of ultrasound wave anti-noise detection system according to said method comprises as lower unit:

AD conversion unit, carry out analog to digital conversion to the ultrasound data of sampling.

The pre-filtering unit, the data of pre-filtering unit after to analog to digital conversion are carried out after pre-filtering sending into record cell and are carried out record.

Record cell, according to sample frequency and the sampling number set, record the ultrasound data that ultrasonic receiver is repeatedly measured.

The time-frequency conversion unit, carry out time-frequency conversion to the ultrasound data of repeatedly measuring respectively.Described time-frequency conversion unit selects linearity, secondary and/or self-adaptation time-frequency conversion method to carry out time-frequency conversion to the ultrasonic signal of repeatedly measuring.

Broad sense time-frequency domain averaging unit, carry out superposed average by the time-frequency conversion result of the ultrasound data repeatedly measured, obtains the Time-Frequency Information matrix, and generate broad sense time-frequency domain mean chart according to the Time-Frequency Information matrix.

The extrema-finding unit, utilize the extreme value finding method, corresponding time of center of energy and the frequency information of search broad sense time-frequency domain mean chart.

The signal determining unit, at the maximum frequency point of center of energy, along the time shaft section, obtain the maximum frequency place and be detected ultrasonic signal.

Above-described embodiment, be only the specific case that purpose of the present invention, technical scheme and beneficial effect are further described, and the present invention not is defined in this.All any modifications of making, be equal to replacement, improvement etc., within all being included in protection scope of the present invention within scope of disclosure of the present invention.

Claims (6)

1. a ultrasound wave anti-noise detection method, is characterized in that comprising the steps:

(1) set sample frequency and sampling number, record the ultrasound data of repeatedly measuring;

(2) respectively the ultrasound data of repeatedly measuring is carried out to time-frequency conversion;

(3) the time-frequency conversion result of the ultrasound data that will repeatedly measure is carried out superposed average, obtains the Time-Frequency Information matrix, and generates broad sense time-frequency domain mean chart according to the Time-Frequency Information matrix;

(4) utilize the extreme value finding method, corresponding time of center of energy and the frequency information of search broad sense time-frequency domain mean chart;

(5) at the maximum frequency point of center of energy, along the time shaft section, obtain the maximum frequency place and be detected ultrasonic signal.

2. a kind of ultrasound wave anti-noise detection method according to claim 1 is characterized in that: in step (1), the ultrasound data of sampling need, successively through after analog to digital conversion step and pre-filtering step, just go on record.

3. a kind of ultrasound wave anti-noise detection method according to claim 1 and 2, it is characterized in that: the described time-frequency conversion of step (2) is selected linearity, secondary and/or self-adaptation time-frequency conversion method.

4. a ultrasound wave anti-noise detection system is characterized in that comprising as lower unit:

Record cell, according to sample frequency and the sampling number set, record the ultrasound data that ultrasonic receiver is repeatedly measured;

The time-frequency conversion unit, carry out time-frequency conversion to the ultrasound data of repeatedly measuring respectively;

Broad sense time-frequency domain averaging unit, carry out superposed average by the time-frequency conversion result of the ultrasound data repeatedly measured, obtains the Time-Frequency Information matrix, and generate broad sense time-frequency domain mean chart according to the Time-Frequency Information matrix;

The extrema-finding unit, utilize the extreme value finding method, corresponding time of center of energy and the frequency information of search broad sense time-frequency domain mean chart;

The signal determining unit, at the maximum frequency point of center of energy, along the time shaft section, obtain the maximum frequency place and be detected ultrasonic signal.

5. a kind of ultrasound wave anti-noise detection system according to claim 4, it is characterized in that: also further comprise AD conversion unit and pre-filtering unit, wherein AD conversion unit is carried out analog to digital conversion to the ultrasound data of sampling, and the data of pre-filtering unit after to analog to digital conversion are carried out after pre-filtering sending into record cell and carried out record.

6. according to the described a kind of ultrasound wave anti-noise detection system of claim 4 or 5, it is characterized in that: described time-frequency conversion unit selects linearity, secondary and/or self-adaptation time-frequency conversion method to carry out time-frequency conversion to the ultrasound data of repeatedly measuring.

Images