CN103293515B - Ship and warship line spectrum noise source longitudinal distribution characteristic measuring method - Google Patents

Abstract

The invention provides a ship and warship line spectrum noise source longitudinal distribution characteristic measuring method. First of all, fast Fourier transform is conducted on collected hydrophone receiving signals to analyze frequency spectrum, micro-region signals with line spectrums are separated out, and a Doppler analyzing signal is obtained in a pluralized mode; the Doppler signal to be analyzed is obtained and WVD is conducted on the Doppler signal: the right-horizontal moment estimated value when each low frequency line spectrum noise source on a ship or a warship passes the measurement hydrophone is obtained; time difference of passing the right-horizontal moment of passing the measurement hydrophone of each noise source and a beacon source is utilized and is combined with ship and warship motion speed to determine the position of each noise source distributed on the ship and warship in a longitudinal mode. The ship and warship line spectrum noise source longitudinal distribution characteristic measuring method is simple to implement, and higher in measuring accuracy than a passing characteristic method.

Description

A kind of naval vessel line spectrum noise source longitudinal distribution measuring method

Technical field

The present invention relates to signal transacting and underwater acoustic measurement field.

Background technology

Ship-radiated noise be mainly derived from the main frame propulsion system on naval vessel, screw propeller, subsidiary engine mechanical system and navigation time the hydrodynamic effect that produces.Wherein, the line spectrum noise caused by the to-and-fro movement of subsidiary engine machinery has that frequency is low, power is high and the feature such as stability is strong, and which carry the characteristic information that naval vessel is important, can, by long-range detection, be the key on underwater sound equipment detection naval vessel.Reduce ship noise and not only can reduce ship navigation to halobiontic impact, be also conducive to improving the performance of the various underwater sound equipments be equipped on naval vessel simultaneously.Therefore, naval vessel noise reduction becomes the urgent task of naval vessel acoustic design, and accurately identifies that the position of Main Noise Sources on naval vessel is that the acoustic design on naval vessel provides reference and foundation.

Mainly utilized it to be realized by characteristic to the measurement of ship-radiated noise in the past.When carrying out radiated noise and measuring, tested naval vessel does uniform rectilinear's navigation, draw near close to measurement hydrophone and from the close-by examples to those far off leave measurement hydrophone, the sound pressure level (or power spectrum level) of the tested ship-radiated noise of measurement hydrophone record and the corresponding relation of measurement hydrophone relative position in this process, what be generally referred to as ship-radiated noise passes through characteristic.Utilize undersea ranging information and the method for radiated noise synchronous recording can obtain position, naval vessel corresponding with radiated noise signals pass through family curve.Ship-radiated noise comprises overall level by characteristic by characteristic, and 1/3oct band level passes through characteristic by characteristic and line spectrum.Overall level corresponds to the maximum position of naval vessel overall level by characteristic maximal value, and 1/3oct sound control and line spectrum then reflect the noise level of different frequency and the relation of tested naval vessel diverse location by characteristic.The relation of reference positions some on measurement hydrophone and hull is determined, so on space (or time), measurement data and hull position just define one-to-one relationship, and when it reflects naval vessel by measurement hydrophone, acoustic pressure is along the distribution of hull.By to longitudinally by specificity analysis, the corresponding relation at different frequency bands noise level and hull position can be obtained, for determining that the noise source position on naval vessel provides a kind of method.Utilize naval vessel to carry out the method for ship noise source longitudinal distribution measurement by characteristic, test is simple but noise source positioning precision is undesirable.

Shi Jie (time clean, Yang Desen, Shi Shengguo. the moving acoustic sources based on the poorest performance optimization steadily and surely focus on positioning identifying method research [J]. Acta Physica Sinica, 2011; 60 (6): 1-11.), Wang Zhiwei (Wang Zhiwei, Xu Lingji, Yang Yixin etc. line array identification underwater movement objective noise source technique study [J]. vibration and impact, 2012; 31 (19): 118-122.) utilize the high resolution technique of sound focusing array to measure the noise source of underwater movement objective, positioning precision is higher.But for the low frequency spectrum lines sound source on naval vessel, the array aperture that said method needs is too large, and the shape need that pair array lays is strict, is difficult to be applied in actual ship-radiated noise test.

Summary of the invention

In order to overcome the deficiencies in the prior art, the invention provides a kind of naval vessel line spectrum noise source longitudinal distribution measuring method based on Wigner-Ville distribution, the Doppler shift characteristics utilizing tested naval vessel and measurement hydrophone relative motion to bring, only uses single nautical receiving set to realize the measurement to naval vessel line spectrum noise source longitudinal distribution.

The technical solution adopted for the present invention to solve the technical problems comprises the following steps:

The first step, any fixed position L on naval vessel _cplace lays a synchronous orientator of the underwater sound as beacon source, and naval vessel does uniform rectilinear's navigation with speed ν, and the velocity of sound is c under water, and the distance abeam of measurement hydrophone and Ship Motion track is R ₀, the acoustical signal that measurement hydrophone receiving ship beam is penetrated, and convert thereof into voltage signal s (t), 0 < t≤T, T represents the time span of data record;

Second step, first does fast fourier transform analysis spectrum nautical receiving set Received signal strength s (t) collected, finds the low frequency spectrum lines existed in signal, its frequency m represents the number of low frequency spectrum lines, isolates the frequency span that there is line spectrum to be by bandpass filter each zonule signal, obtains each low frequency spectrum lines noise source Doppler signal s _m(t), m=1,2 ..., M; By s _mt () plural numberization obtains Doppler analytic signal z _m(t)=s _m(t)+jH [s _m(t)], H [s _m(t)] represent signal s _mthe Hilbert transform of (t);

3rd step, the beam moment obtaining each low frequency spectrum lines noise source estimates evaluation t ' _m0after, by formula $\left\{\begin{array}{cc}0<t<{2t}_{m0}^{\&prime;}& t_{m0}^{\&prime;}\&le;T/2\\ {2t}_{m0}^{\&prime;}-T<t\&le;T& t_{m0}^{\&prime;}>T/2\end{array}\right.$ Intercept z _mt () obtains Doppler signal z ' to be analyzed _m(t); Then to the Doppler signal z ' after intercepting _mt () is WVD: $W_{z_m^{\&prime;}}(t,f)={\&Integral;}_{-\&infin;}^{\&infin;}{z}_m^{\&prime;}(t+\frac{\mathrm{\&tau;}}{2}){z_m^{\&prime;}}^{*}(t-\frac{\mathrm{\&tau;}}{2})e^{-j2\mathrm{\&pi;f\&tau;}}\mathrm{d\&tau;},$ Wherein t is the time, and f is frequency, and τ is time delay, and * represents conjugation, obtains the distribution of Doppler signal energy about temporal frequency; Utilize energy peak detection method to obtain on naval vessel each low frequency spectrum lines noise source by beam moment estimated value during measurement hydrophone

4th step, the synchronous orientator of the underwater sound can obtain known location L _cbeacon source is by beam moment t during measurement hydrophone _c0with Ship Motion speed ν, utilize each noise source and beacon source by the mistiming in measurement hydrophone beam moment, in conjunction with Ship Motion speed ν, each noise source position of genesis analysis on naval vessel can be determined

The invention has the beneficial effects as follows: measure on the platform of ship-radiated noise at existing single nautical receiving set, utilize the Doppler effect that when measuring, between naval vessel and measurement hydrophone, relative motion produces, in conjunction with time frequency analysis signal processing technology, realize the measurement of naval vessel low frequency spectrum lines noise source longitudinal distribution, the method is implemented simple, and measuring accuracy is than high by characteristic method.

Accompanying drawing explanation

Fig. 1 is ship-radiated noise measurement model schematic diagram;

Fig. 2 is the overall procedure block diagram of naval vessel line spectrum noise source genesis analysis localization method.

Embodiment

The present invention is fixed position L on naval vessel _cplace lays a synchronous orientator of the underwater sound as beacon source, and naval vessel does uniform rectilinear's navigation with speed ν, and draw near close to measurement hydrophone and from the close-by examples to those far off leave measurement hydrophone, the distance abeam of measurement hydrophone and Ship Motion track is R ₀, the measurement hydrophone receiving ship beam acoustical signal of penetrating in this process, and convert thereof into voltage signal s (t), 0 < t≤T, record by front putting rear data collecting instrument, T represents the time span of data record.

The receiving hydrophone signal demand pre-service of record, object is: one is the line spectrum noise source signal determining to exist; Two is out-of-band noises that Doppler signal is removed in filtering; Three is Doppler signals that plural numberization obtains analytical form.First nautical receiving set Received signal strength s (t) collected is done fast Fourier transform (FFT) analysis spectrum, find the low frequency spectrum lines existed in signal, its frequency m represents the number of low frequency spectrum lines, low frequency spectrum lines noise source number namely to be analyzed.Isolating by bandpass filter the frequency span that there is line spectrum is each zonule signal, this just obtains each low frequency spectrum lines noise source Doppler signal s _m(t), m=1,2 ..., M.By isolated line spectrum noise source Doppler signal s _mt () plural numberization obtains Doppler analytic signal z _m(t), wherein

Z _m(t)=s _m(t)+jH [s _m(t)], H [s _m(t)] represent signal s _mthe Hilbert transform of (t).

Owing to there is relative motion between tested naval vessel and measurement hydrophone, amplitude and the frequency of the line spectrum noise source signal that measurement hydrophone receives change, and the change of frequency is called Doppler shift.Doppler signal z _mt the amplitude of () is represent each line spectrum noise source radiation intensity A _m0with the decay of spherical wave form, t _m0for each line spectrum noise source by beam position nearest from measurement hydrophone time the beam moment; Signal z _mt the frequency of () is $f_m=\frac{f_{m0}{c}^2}{c^2-v^2}(1-\frac{v^2(t-t_{m0})}{\sqrt{{R_0}^2(c^2-v^2)+v^2{c}^2{(t-t_{m0})}^2}}),$ C is the velocity of sound in water, f _m0represent the frequency of each radiant rays spectral noise source signal on naval vessel.Doppler signal z _mt the frequency change of () is nonlinear, utilize secondary Time-Frequency Analysis Method Wigner-Ville distribution (WVD) to analyze Doppler signal z _mt () there will be the interference of Self-crossover item, this is disadvantageous to the instantaneous Frequency Estimation of signal, but the existence of Self-crossover item is to estimation beam moment t _m0useful.Because Doppler signal WVD Self-crossover item is in the convolution effect of frequency domain, cause except t=t _m0outside moment, (t, f on Doppler signal WVD time frequency plane should be gathered in _m) energy spread at place.Utilize this character of Doppler signal WVD Self-crossover item, directly carrying out WVD to Doppler signal can at time frequency plane (t _m0, f _m0) place's forming energy peak value.In addition because of Doppler signal z _mt () changes in amplitude causes the energy of signal at t _m0time also maximum, such Doppler signal WVD is at (t _m0, f _m0) energy peak at place is more obvious.Find the point of the energy peak on Doppler's time frequency plane position can to obtain on naval vessel each low frequency spectrum lines noise source by beam moment t during measurement hydrophone _m0.

In order to make full use of the energy spread effect of the Self-crossover item of Doppler signal WVD, reply z _mt () intercepts, the principle of intercepting is that the analysis time of Doppler signal is as far as possible long and with the beam moment of pre-estimation for intercepting the mid point of time period, the beam moment of pre-estimation can be adopted and be obtained by methods such as characteristic methods.Estimate evaluation t ' obtaining each noise beam moment in a steady stream _m0after, by formula $\left\{\begin{array}{cc}0<t<{2t}_{m0}^{\&prime;}& t_{m0}^{\&prime;}\&le;T/2\\ {2t}_{m0}^{\&prime;}-T<t\&le;T& t_{m0}^{\&prime;}>T/2\end{array}\right.$ Intercept z _mt () obtains Doppler signal z ' to be analyzed _m(t).Then to the Doppler signal z ' after intercepting _mt () is WVD: $W_{z_m^{\&prime;}}(t,f)={\&Integral;}_{-\&infin;}^{\&infin;}{z}_m^{\&prime;}(t+\frac{\mathrm{\&tau;}}{2}){z_m^{\&prime;}}^{*}(t-\frac{\mathrm{\&tau;}}{2})e^{-j2\mathrm{\&pi;f\&tau;}}\mathrm{d\&tau;}$ Obtain the distribution of Doppler signal energy about temporal frequency.WVD is the instantaneous autocorrelation function utilizing signal do Fourier transform, algorithm is generally realized by FFT.Because the movement velocity on tested naval vessel is slow, radiant rays spectral noise source frequency is positioned at low frequency, and Doppler frequency change is faint, so counting of FFT will obtain very greatly, so just can form one on the frequency axis trickle rule to observe small frequency change.Time frequency analysis for low frequency spectrum lines Doppler signal has that frequency band is narrow, refinement requires high feature, frequency band data not beyond Water demand Doppler shift, therefore use the FFT in the Chirp-Z method replacement implementation algorithm of local frequencies refinement, under the condition ensureing frequency resolution, improve counting yield.Finally utilize energy peak detection method to obtain on naval vessel each line spectrum noise source by beam moment estimated value during measurement hydrophone

The synchronous orientator of the underwater sound can obtain known location L _cbeacon source is by beam moment t during measurement hydrophone _c0with Ship Motion speed ν, utilize each noise source and beacon source by mistiming in measurement hydrophone beam moment, in conjunction with Ship Motion speed ν, according to formula each noise source position L of genesis analysis on naval vessel can be determined _m.

Below in conjunction with drawings and Examples, the present invention is further described.

Fig. 1 is the measurement model of ship-radiated noise, and measurement hydrophone is positioned at S place, tested naval vessel uniform motion, and speed is ν, and track is MN, beam position O be its running orbit from the nearest point of measurement hydrophone, fixed position L on naval vessel _cplace lays a synchronous orientator of the underwater sound as beacon, naval vessel is by the process of measurement hydrophone, the acoustical signal that measurement hydrophone receiving ship beam is penetrated, and convert thereof into voltage signal s (t), by using data collecting instrument with sampling rate F after enlarge leadingly _swriting time, length was the data of T, obtained the discrete-time series of signal s (t)

s(n),n＝1,…,N，N＝TF _S。

Fig. 2 is the overall procedure of naval vessel line spectrum noise source genesis analysis localization method, is specifically implemented as follows:

(1) to the pre-service of nautical receiving set Received signal strength, first signal s (n) collected is done fast fourier transform (FFT) analysis spectrum, determine the low frequency spectrum lines existed in signal, its frequency m represents the number of low frequency spectrum lines, the low frequency spectrum lines noise source number namely will analyzed.Isolating by FIR bandpass filter the frequency span that there is line spectrum is each zonule signal, obtains each line spectrum noise source Doppler signal

s _m(n),m＝1,2,…,M。By isolated line spectrum noise source Doppler signal s _mn () plural numberization obtains z _m(n), wherein z _m(n)=s _m(n)+jH [s _m(n)], H [s _m(n)] represent signal s _mthe Hilbert transform of (n).

(2) to the Doppler analytic signal z of previous step _mn () intercepts, the principle of intercepting is that the analysis time of Doppler signal is as far as possible long and with the beam moment of pre-estimation for intercepting the mid point of time period, utilizes can obtain the beam moment by characteristic method and estimate evaluation t ' _m0, by formula $\left\{\begin{array}{cc}0<n\&le;{2t}_{m0}^{\&prime;}{F}_S& t_{m0}^{\&prime;}\&le;T/2\\ ({2t}_{m0}^{\&prime;}-T)F_S<n\&le;T{F}_S& t_{m0}^{\&prime;}>T/2\end{array}\right.$ To sequence z _mn () intercepts and is always counted as the Doppler signal time series z ' to be analyzed of N ' _m(n), n=1,2 ..., N '.High-frequency resolving accuracy Doppler signal z ' is realized fast in conjunction with refinement Frequency Estimation Chirp-Z algorithm _mwigner-Ville distribution (WVD) of (n): first construct a N ' × N ' dimension matrix wherein 0 < n≤N ' ,-L≤l≤L, L=(N '-1)/2, N ' gets odd number; Each for matrix G (n, l) row is carried out ring shift, builds N ' × N ' reform matrix $X(n,l)=\left\{\begin{array}{cc}G(n,l)G^{*}(n,-l)& 0\&le;l\&le;L-1\\ G(n,l-2L)G^{*}(n,-1+2L)& L\&le;l\&le;2L\end{array}\right.,$ Wherein * represents conjugation, the row vector x of matrix X (n, l) _nl () represents that Doppler signal is at n/F _sthe instantaneous autocorrelation sequence in moment; With the numerical frequency low frequency of FIR bandpass filter in step (1) pre-service and high frequency thresholding f _b, f _eas original frequency and the termination frequency of zoom FFT, Δ f is refinement frequency interval, counting of refinement frequency structure frequency sampling factor z _k=AW ^-k, wherein $\left\{\begin{array}{c}A=e^{j4{\mathrm{\&pi;f}}_b}\\ W=e^{-j4\mathrm{\&pi;\&Delta;f}}\end{array}\right.;$ Structure length L ' unit response sequence $h\left(l\right)=\left\{\begin{array}{cc}{W}^{-l^2/2}& 0\&le;l\&le;K-1\\ 0& K\&le;l\&le;L^{\&prime;}-N^{\&prime;}\\ W^{-{(L-1)}^2/2}& L^{\&prime;}-N^{\&prime;}+1\&le;l\&le;L^{\&prime;}-1\end{array}\right.,$ L ' be closest to N '+K-1 2 power; To x _nl () is weighted process at y ' _nl zero padding in the middle of () data, obtains $y_n\left(l\right)=\left\{\begin{array}{cc}{y}_n^{\&prime;}\left(l\right)& 0\&le;l\&le;L-1\\ 0& L\&le;l\&le;L^{\&prime;}-L\\ y_n^{\&prime;}(l-L^{\&prime;}+N^{\&prime;})& L^{\&prime;}-L+1\&le;l\&le;L^{\&prime;}-1\end{array}\right.;$ Ask h (l), y respectively _nthe Fourier transform of (l) $\left\{\begin{array}{c}H\left(k\right)=\mathrm{FFT}\left[h\left(l\right)\right]\\ Y_n\left(k\right)=\mathrm{FFT}\left[y_n\left(l\right)\right]\end{array}\right.,$ Then be multiplied at frequency domain, and obtain convolution results q by inverse FFT computing _n(l)=IFFT [Y _n(k) H (k)]; Finally K point data before output sequence is exported and be weighted f _n(z _k) be n/F _sthe spectrum sequence that moment is corresponding, by the power spectrum sequence corresponding each moment composition matrix W (n, z _k), just can obtain signal z ' _mn the energy of () is about the distribution of time and frequency detected by energy peak $({\hat{f}}_{m0},{\hat{t}}_{m0})=\underset{t,f}{\mathrm{ArgMax}}\left\{W_{z_m^{\&prime;}}(t,f)\right\}$ Find the time that maximal value is corresponding, namely each noise is in a steady stream by the estimated value in beam moment during measurement hydrophone

(3) the synchronous orientator of the underwater sound can obtain known location L _cbeacon source is by beam moment t during measurement hydrophone _c0with Ship Motion speed ν, utilize each noise source and beacon source by mistiming in measurement hydrophone beam moment, in conjunction with Ship Motion speed ν, according to formula each line spectrum noise source position L of genesis analysis on naval vessel can be determined _m.

Claims (1)

1. a naval vessel line spectrum noise source longitudinal distribution measuring method, is characterized in that comprising the steps:

The first step, any fixed position L on naval vessel _cplace lays a synchronous orientator of the underwater sound as beacon source, and naval vessel does uniform rectilinear's navigation with speed v, and the velocity of sound is c under water, and the distance abeam of measurement hydrophone and Ship Motion track is R ₀, the acoustical signal that measurement hydrophone receiving ship beam is penetrated, and convert thereof into voltage signal s (t), 0 < t≤T, T represents the time span of data record;

Second step, first does fast fourier transform analysis spectrum nautical receiving set Received signal strength s (t) collected, finds the low frequency spectrum lines existed in signal, its frequency m=1,2 ..., M, M represent the number of low frequency spectrum lines, isolate the frequency span that there is line spectrum to be by bandpass filter each zonule signal, obtains each low frequency spectrum lines noise source Doppler signal s _m(t), m=1,2 ..., M; By s _mt () plural numberization obtains Doppler analytic signal z _m(t)=s _m(t)+jH [s _m(t)], H [s _m(t)] represent signal s _mthe Hilbert transform of (t);

3rd step, the beam moment obtaining each low frequency spectrum lines noise source estimates evaluation t ' _m0after, by formula $\left\{\begin{array}{cc}0<t<{2t}_{m0}^{\&prime;}& t_{m0}^{\&prime;}\&le;T/2\\ 2t_{m0}^{\&prime;}-T<t\&le;T& t_{m0}^{\&prime;}>T/2\end{array}\right.$ Intercept z _mt () obtains Doppler signal z ' to be analyzed _m(t); Then to the Doppler signal z ' after intercepting _mt () is WVD: wherein t is the time, and f is frequency, and τ is time delay, and * represents conjugation, obtains the distribution of Doppler signal energy about temporal frequency; Utilize energy peak detection method to obtain on naval vessel each low frequency spectrum lines noise source by beam moment estimated value during measurement hydrophone

4th step, the synchronous orientator of the underwater sound can obtain known location L _cbeacon source is by beam moment t during measurement hydrophone _c0with Ship Motion speed ν, utilize each noise source and beacon source by the mistiming in measurement hydrophone beam moment, in conjunction with Ship Motion speed ν, each noise source position of genesis analysis on naval vessel can be determined