CN105572650A - Broadband multiple correlation flow velocity measurement method - Google Patents

Abstract

The invention discloses a method for measuring broadband complex correlation flow velocity. The time-alternating sampling technique is used to sample the echo signal to obtain a discrete sampling sequence, and the discrete sampling sequence is respectively processed by a complex correlation algorithm and a cross-spectral autocorrelation algorithm. The complex correlation algorithm In the process of processing, the number of complex correlation points is increased, and the odd and even two-way parallel complex correlation processing is used to ensure the real-time performance of flow velocity measurement. The time-value roughly measured flow velocity value is used as the judgment standard of the flow velocity value interval, assists the flow velocity measurement of the complex correlation algorithm, and corrects the flow velocity measurement range; The sampling rate is high, combined with the cross-spectrum autocorrelation algorithm to roughly measure the flow velocity value to assist the complex correlation algorithm to measure the flow velocity, the measurement accuracy is high, and the measurement range is wide.

Description

A Broadband Complex Correlation Flow Velocity Measurement Method

technical field

The invention relates to a flow velocity measurement method, in particular to a broadband multiple correlation flow velocity measurement method.

Background technique

With the increasing scarcity of resources, the development of marine water resources has become an urgent need for economic development. The flow velocity of ocean waters is an important parameter of the environment of ocean waters, and it is also an important basis for activities in ocean waters. Existing flow velocity measurement methods are mainly based on broadband complex correlation technology. The document "Research on Velocity Measurement Method of Wideband Doppler Current Meter" proposes a flow velocity research based on broadband multiple correlation algorithm. In this document, the specific process of flow velocity measurement is as follows: firstly, the pulse signal generator is used to generate a single-frequency pulse signal, and the M-sequence code is used to binary code the single-frequency pulse signal to obtain a sub-pulse signal, and then multiple sub-pulse signals are sequentially spliced to obtain Modulate the pulse signal; receive the signal reflected by the modulated pulse signal through the ocean waters, and collect the echo signal of the signal; finally use the broadband complex correlation algorithm to process the echo signal to obtain the flow velocity.

In this flow velocity measurement method, the echo signal is a random signal containing noise. The strength of the echo signal, noise interference and the propagation loss of the echo signal all affect the accuracy of the flow velocity measurement. The complex correlation algorithm can filter out certain white noise. , has better anti-interference ability, so the flow velocity measurement method has higher accuracy, but in the process of processing the echo signal by the complex correlation algorithm, the phase information of the complex correlation function of the echo signal obtained is periodic, The measurement range of the flow rate is limited, so the method has a narrow range.

Contents of the invention

The technical problem to be solved by the present invention is to provide a wide-band complex correlation flow velocity measurement method with high measurement accuracy and wide measurement range.

The technical solution adopted by the present invention to solve the above-mentioned technical problems is: a kind of broadband complex correlation velocity measurement method, comprises the following steps:

① Obtain the echo signal

①-1 Use a pulse signal generator to generate a single-frequency pulse signal with a frequency of f ₀ , f ₀ =1.5MHz;

①-2 Use the M-sequence code with the number of symbols L and the symbol width Δτ to binary code the single-frequency pulse signal to obtain a sub-pulse signal, the number of symbols L=15, and the symbol width Δτ is filled with 10 single-frequency pulses signal period, The length of the sub-pulse signal is recorded as T, T=L×Δτ, and the symbol × is the multiplication symbol;

①-3 Continuously transmit multiple sub-pulse signals in the ocean waters where the flow velocity is to be measured, and form a modulated pulse signal by the multiple sub-pulse signals continuously transmitted in sequence, and record the modulated pulse signal as s(t), where t is the time domain time;

①-4 Obtain the signal reflected by the modulated pulse signal s(t) through the ocean waters, record the reflected signal as s'(t), and obtain the echo signal rr(t) to be processed according to s'(t) ), rr(t)=s'(t)+δ(t); wherein δ(t) is a white noise signal;

② Perform pre-amplification processing, band-pass filter processing and low-pass filter processing on the echo signal rr(t) in sequence to obtain a pre-sampling signal, which is recorded as r(t);

③Use the alternate sampling module to perform four-way time alternate sampling on the pre-sampling signal r(t) to obtain the discrete sampling sequence r(n), where n is the discrete sequence number, n=1, 2, 3,...; the alternate sampling module includes the first Analog-to-digital converter, second analog-to-digital converter, third analog-to-digital converter, fourth analog-to-digital converter and four-to-one selector, first analog-to-digital converter, second analog-to-digital converter, third analog-to-digital converter The converter and the fourth analog-to-digital converter are connected in one-to-one correspondence with the four selection ends of the four-choice selector, the first analog-to-digital converter, the second analog-to-digital converter, the third analog-to-digital converter and the fourth analog-to-digital converter The converter samples the pre-sampled signal r(t), and the one-of-four selector sequentially selects the first analog-to-digital converter, the second analog-to-digital converter, the third analog-to-digital converter and the fourth analog-to-digital converter, and the first The sampling signals of the analog-to-digital converter, the second analog-to-digital converter, the third analog-to-digital converter and the fourth analog-to-digital converter are sequentially output by a four-selector selector and sequentially spliced into a discrete sampling sequence r(n);

④ Perform complex correlation algorithm processing and cross-spectral autocorrelation algorithm processing on the discrete sampling sequence r(n) respectively;

The specific process of complex correlation algorithm processing is as follows:

a. adopt the DDS module to generate the quadrature modulation signal, adopt the quadrature modulation signal to digitally modulate the discrete sampling sequence r (n), and obtain two-way quadrature signal sequences, which are respectively denoted as I (n) and Q (n); Use I(n) and Q(n) to construct the first complex sequence X(n), take I(n) as the real part signal of the first complex sequence X(n), and Q(n) as the first complex sequence X( n) imaginary part signal, then the first complex number sequence represents X(n)=I(n)+jQ(n), wherein j represents the imaginary part unit of the complex number;

b. Divide the first complex number sequence X(n) into two complex number sequences according to the odd and even discrete serial numbers, and call the complex number sequence whose discrete serial number is an odd number the second complex number sequence, denoted as X(2m+1), The complex number sequence whose discrete serial number is an even number is called the third complex number sequence, which is denoted as X(2m+2), m=0, 1, 2, 3, ...;

c. The complex correlation function of the first complex number sequence X(n) is recorded as R ₁ (n), the complex correlation function of the second complex number sequence X(2m+1) is recorded as R ₂ (n), and the third complex number The complex correlation function of sequence X(2m+2) is denoted as R ₃ (n), $R_2\left(\mathrm{no}\right)=\underset{m=0}{\overset{N^{\′}}{\Σ}}x(2m+1+l)x^{*}(2m+1),R_3\left(\mathrm{no}\right)=\underset{m=0}{\overset{N^{\′\′}}{\Σ}}x(2m+2+l)x^{*}(2m+2),$ Wherein N' represents the complex correlation length of the second complex number sequence X(2m+1), N" represents the complex correlation length of the third complex number sequence X(2m+2), and l is the sampling delay value and its value is the encoding length of one M-sequence code, N=f _s ×(t ₁ -t ₀ ), f _s is the sampling frequency of the alternate sampling module, t ₀ is the Sampling start time, t ₁ sampling end time of alternate sampling module, X(2m+1+l)=I(2m+1+l)+jQ(2m+1+l), X(2m+2+l) =I(2m+2+l)+jQ(2m+2+l), I(2m+1+l) is the expression of I(2m+1) delay l, and I(2m+2+l) is The expression of I(2m+2) delay l, Q(2m+1+l) is the expression of Q(2m+1) delay l, Q(2m+2+l) is Q(2m+2) The expression of time delay l; the symbol * is the conjugate operator, X ^* (2m+1) is the conjugate complex sequence of X(2m+1), X ^* (2m+2) is the sequence of X(2m+2) sequence of conjugated complex numbers;

d. Substituting R ₂ (n) and R ₃ (n) into the formula R ₁ (n)=R ₂ (n)+R ₃ (n) to get R ₁ (n), the real part of R ₁ (n) Recorded as Re(R ₁ (n)), the imaginary part of R ₁ (n) will be recorded as Im(R ₁ (n));

e. Record the phase angle of the complex correlation function R ₁ (n) of the first complex number sequence X(n) as θ, and use the arctangent function to solve the phase angle θ, namely $\mathrm{\θ}=arcta\mathrm{no}\left(\frac{\mathrm{Im}\left(R_1\right(\mathrm{no}))}{\mathrm{Re}\left(R_1\right(\mathrm{no}))}\right);$

f. Record the Doppler frequency offset as f _d , and use the formula Where q is the coherent repetition period, that is, q=f _s ×T;

g. Record the flow velocity measured by complex correlation as v _c , and use the formula Calculate v _c , c is the speed of sound in water, c=1500m/s;

The specific process of the cross-spectral autocorrelation algorithm is as follows:

a. The spectrum of the discrete sampling sequence r(n) is denoted as r'(k), the autocorrelation function spectrum of the discrete sampling sequence r(n) is denoted as R(k), k represents the serial number in the frequency domain, and k=0 ,1,2,3,...;

b. The frequency spectrum r'(k) is obtained by FFT calculation, so that R(k)=r'(k)×r'(k);

c. Interpolate the autocorrelation function spectrum R(k) and then perform IFFT operation to obtain the autocorrelation function of the discrete sampling sequence r(n), and record the autocorrelation function of the discrete sampling sequence r(n) as R'( t);

d. Determine the moment when the first side lobe appears by the autocorrelation function R'(t) of the discrete sampling sequence r(n), and record the moment when the first side lobe appears as T ₁ ; in the ideal state where no frequency offset occurs The moment when the first side lobe appears is recorded as T ₂ , where T ₂ =T

e. Record the difference of Doppler time delay as ΔT, ΔT=T ₁ -T ₂ ;

f. Denote the flow velocity measured by the autocorrelation as v _z , using the formula Calculate the flow velocity v _z from the autocorrelation measurement;

⑤ Record the velocity correction parameter as α, α=[(v _z -v _c )/2v _max ], v _max is the fuzzy velocity, v _max =c/4f ₀ T, and the symbol “[]” is the rounding operator;

⑥Use the formula v=v _c +α2v _max to calculate the flow velocity v to be measured.

Compared with the prior art, the present invention has the advantage of adopting the time-alternating sampling technology to sample the echo signal to obtain a discrete sampling sequence, and in addition, the discrete sampling sequence is processed by a complex correlation algorithm and a cross-spectrum autocorrelation algorithm, and the time-alternating sampling On the premise of ensuring the sampling accuracy, the sampling rate is improved, the number of complex correlation points is correspondingly increased in the process of complex correlation algorithm processing, and the odd and even two-way parallel complex correlation operation is used for processing. On the basis of improving the current measurement accuracy of the complex correlation algorithm, the Calculate the speed to ensure the real-time performance of the flow velocity measurement. The cross-spectrum autocorrelation algorithm processing can improve the estimation accuracy of the time delay value by performing cross-spectrum interpolation delay estimation on the time-alternative sampling data, reduce the judgment error of the flow velocity range, and realize a wide range. For flow velocity measurement, the delay value is used to roughly measure the flow velocity value. There is no measurement ambiguity problem in the rough measurement flow velocity value, which can be used as a judgment standard for the size range of the flow velocity value, assisting the flow velocity measurement of the complex correlation algorithm, and overcoming the possible ambiguity in the flow velocity measurement of the complex correlation algorithm. In order to solve the limitations of the flow rate and flow rate measurement range, the flow rate measurement range was corrected, so that the complex correlation algorithm processing and the cross-spectral autocorrelation algorithm processing were combined, and the sampling rate of the signal was improved by using the time-alternating sampling technology. Combined with the cross-spectrum autocorrelation algorithm, the rough The measured flow velocity value assists the complex correlation algorithm to measure the flow velocity, with high measurement accuracy and wide measurement range.

detailed description

Below in conjunction with embodiment the present invention is described in further detail.

Embodiment: a kind of broadband complex correlation velocity measurement method, comprises the following steps:

① Obtain the echo signal

①-1 Use a pulse signal generator to generate a single-frequency pulse signal with a frequency of f ₀ , f ₀ =1.5MHz;

①-2 Use the M-sequence code with the number of symbols L and the symbol width Δτ to binary code the single-frequency pulse signal to obtain a sub-pulse signal, the number of symbols L=15, and the symbol width Δτ is filled with 10 single-frequency pulses signal period, The length of the sub-pulse signal is recorded as T, T=L×Δτ, and the symbol × is the multiplication symbol;

①-3 Multiple (i.e. at least two) sub-pulse signals are sequentially and continuously transmitted in the ocean waters where the flow velocity is to be measured, and the multiple sub-pulse signals continuously transmitted in sequence form a modulated pulse signal, and the modulated pulse signal is recorded as s( t), t is time domain time;

①-4 Obtain the signal reflected by the modulated pulse signal s(t) through the ocean waters, record the reflected signal as s'(t), and obtain the echo signal rr(t) to be processed according to s'(t) ), rr(t)=s'(t)+δ(t); wherein δ(t) is a white noise signal;

② Perform pre-amplification processing, band-pass filter processing and low-pass filter processing on the echo signal rr(t) in sequence to obtain a pre-sampling signal, which is recorded as r(t);

③Use the alternate sampling module to perform four-way time alternate sampling on the pre-sampling signal r(t) to obtain the discrete sampling sequence r(n), where n is the discrete sequence number, n=1, 2, 3,...; the alternate sampling module includes the first Analog-to-digital converter, second analog-to-digital converter, third analog-to-digital converter, fourth analog-to-digital converter and four-to-one selector, first analog-to-digital converter, second analog-to-digital converter, third analog-to-digital converter The converter and the fourth analog-to-digital converter are connected in one-to-one correspondence with the four selection ends of the four-choice selector, the first analog-to-digital converter, the second analog-to-digital converter, the third analog-to-digital converter and the fourth analog-to-digital converter The converter samples the pre-sampled signal r(t), and the one-of-four selector sequentially selects the first analog-to-digital converter, the second analog-to-digital converter, the third analog-to-digital converter and the fourth analog-to-digital converter, and the first The sampling signals of the analog-to-digital converter, the second analog-to-digital converter, the third analog-to-digital converter and the fourth analog-to-digital converter are sequentially output by a four-selector selector and sequentially spliced into a discrete sampling sequence r(n);

④ Perform complex correlation algorithm processing and cross-spectral autocorrelation algorithm processing on the discrete sampling sequence r(n) respectively;

The specific process of complex correlation algorithm processing is as follows:

a. Adopt DDS (DirectDigitalSynthesizer, Direct Digital Frequency Synthesizer) module to generate quadrature modulation signal, use the quadrature modulation signal to digitally modulate the discrete sampling sequence r(n), and obtain two quadrature signal sequences, respectively denoted as I(n) and Q(n); using I(n) and Q(n) to construct the first complex sequence X(n), using I(n) as the real part signal of the first complex sequence X(n), Q (n) as the imaginary part signal of the first complex number sequence X(n), then the first complex number sequence represents X(n)=I(n)+jQ(n), wherein j represents the imaginary part unit of the complex number;

b. Divide the first complex number sequence X(n) into two complex number sequences according to the odd and even discrete serial numbers, and call the complex number sequence whose discrete serial number is an odd number the second complex number sequence, denoted as X(2m+1), The complex number sequence whose discrete serial number is an even number is called the third complex number sequence, which is denoted as X(2m+2), m=0, 1, 2, 3, ...;

c. The complex correlation function of the first complex number sequence X(n) is recorded as R ₁ (n), the complex correlation function of the second complex number sequence X(2m+1) is recorded as R ₂ (n), and the third complex number The complex correlation function of sequence X(2m+2) is denoted as R ₃ (n), $R_2\left(\mathrm{no}\right)=\underset{m=0}{\overset{N^{\′}}{\Σ}}x(2m+1+l)x^{*}(2m+1),R_3\left(\mathrm{no}\right)=\underset{m=0}{\overset{N^{\′\′}}{\Σ}}x(2m+2+l)x^{*}(2m+2),$ Wherein N' represents the complex correlation length of the second complex number sequence X(2m+1), N" represents the complex correlation length of the third complex number sequence X(2m+2), and l is the sampling delay value and its value is the encoding length of one M-sequence code, N=f _s ×(t ₁ -t ₀ ), f _s is the sampling frequency of the alternate sampling module, t ₀ is the Sampling start time, t ₁ the sampling end time of the alternate sampling module, f _s , t ₀ and t ₁ are parameters that adopt the existing conventional setting method to set in the alternate sampling module, X(2m+1+l)=I (2m+1+l)+jQ(2m+1+l), X(2m+2+l)=I(2m+2+l)+jQ(2m+2+l), I(2m+1+ l) is the expression of I(2m+1) delay l, I(2m+2+l) is the expression of I(2m+2) delay l, Q(2m+1+l) is Q(2m +1) The expression of delay l, Q(2m+2+l) is the expression of Q(2m+2) delay l; the symbol * is a conjugate operator, X ^* (2m+1) is X( 2m+1) conjugate complex sequence, X ^* (2m+2) is the conjugate complex sequence of X(2m+2);

d. Substituting R ₂ (n) and R ₃ (n) into the formula R ₁ (n)=R ₂ (n)+R ₃ (n) to get R ₁ (n), the real part of R ₁ (n) Recorded as Re(R ₁ (n)), the imaginary part of R ₁ (n) will be recorded as Im(R ₁ (n));

e. Record the phase angle of the complex correlation function R ₁ (n) of the first complex number sequence X(n) as θ, and use the arctangent function to solve the phase angle θ, namely $\mathrm{\θ}=arcta\mathrm{no}\left(\frac{\mathrm{Im}\left(R_1\right(\mathrm{no}))}{\mathrm{Re}\left(R_1\right(\mathrm{no}))}\right);$

f. Record the Doppler frequency offset as f _d , and use the formula Where q is the coherent repetition period, that is, q=f _s ×T;

g. Record the flow velocity measured by complex correlation as v _c , and use the formula Calculate v _c , c is the speed of sound in water, c=1500m/s;

The specific process of the cross-spectral autocorrelation algorithm is as follows:

a. The spectrum of the discrete sampling sequence r(n) is denoted as r'(k), the autocorrelation function spectrum of the discrete sampling sequence r(n) is denoted as R(k), k represents the serial number in the frequency domain, and k=0 ,2,3,…;

b. The frequency spectrum r'(k) is obtained by FFT calculation, so that R(k)=r'(k)×r'(k);

c. Interpolate the autocorrelation function spectrum R(k) and then perform IFFT operation to obtain the autocorrelation function of the discrete sampling sequence r(n), and record the autocorrelation function of the discrete sampling sequence r(n) as R'( t);

d. Determine the moment when the first side lobe appears by the autocorrelation function R'(t) of the discrete sampling sequence r(n), and record the moment when the first side lobe appears as T ₁ ; in the ideal state where no frequency offset occurs The moment when the first side lobe appears is recorded as T ₂ , where T ₂ =T

e. Record the difference of Doppler time delay as ΔT, ΔT=T ₁ -T ₂ ;

f. Denote the flow velocity measured by the autocorrelation as v _z , using the formula Calculate the flow velocity v _z from the autocorrelation measurement;

⑤ Record the velocity correction parameter as α, α=[(v _z -v _c )/2v _max ], v _max is the fuzzy velocity, v _max =c/4f ₀ T, and the symbol “[]” is the rounding operator;

⑥Use the formula v=v _c +α2v _max to calculate the flow velocity v to be measured.

In this embodiment, both the alternate sampling module and the DDS module are mature products in their technical fields.

Claims (1)

1. a broadband multiple correlation flow-speed measurement method, is characterized in that comprising the following steps:

1. echoed signal is obtained

1.-1 utilize pulse signal generator generated frequency for f ₀pure-tone polse signal, f ₀=1.5MHz;

1.-2 adopt that code element numbers are L, symbol width is that the M sequence code of Δ τ carries out binary coding to pure-tone polse signal and obtains subpulse signal, code element number L=15, symbol width Δ τ is 10 pure-tone polse signal periods of filling, the length of subpulse signal is designated as T, T=L × Δ τ, symbol × be multiplication symbol;

1.-3 by multiple subpulse signal sequential filming successively in the maritime waters for the treatment of velocity measurement, and multiple subpulse signals of sequential filming form a modulated pulse signal successively, and this modulated pulse signal is designated as s (t), and t is the time domain time;

1.-4 the signal that is reflected back through maritime waters of these modulated pulse signals s (t) is obtained, the signal this be reflected back is designated as s'(t), according to s'(t) obtain pending echoed signal rr (t), rr (t)=s'(t)+δ (t); Wherein δ (t) is white noise signal;

2. successively enlarge leadingly process, bandpass filtering treatment and low-pass filtering treatment are carried out to echoed signal rr (t), obtain pre-sampling signal, pre-sampling signal is designated as r (t);

3. adopt alternating sampling module to carry out four road time-interleaveds to pre-sampling signal r (t) and obtain discrete sampling sequence r (n), n is discrete series number, n=1,2,3, alternating sampling module comprises the first analog to digital converter, second analog to digital converter, 3rd analog to digital converter, 4th analog to digital converter and four selects a selector switch, first analog to digital converter, second analog to digital converter, 3rd analog to digital converter and the 4th analog to digital converter and four select four of a selector switch selecting sides to connect one to one, first analog to digital converter, second analog to digital converter, 3rd analog to digital converter and the 4th analog to digital converter are sampled to pre-sampling signal r (t), four select a selector switch to select the first analog to digital converter successively, second analog to digital converter, 3rd analog to digital converter and the 4th analog to digital converter, first analog to digital converter, second analog to digital converter, the sampled signal of the 3rd analog to digital converter and the 4th analog to digital converter is selected a selector switch to export successively through four and is spliced into discrete sampling sequence r (n) according to the order of sequence,

4. multiple correlation algorithm process and the process of cross-spectrum auto-correlation algorithm are done respectively to discrete series r (n);

The detailed process of multiple correlation algorithm process is:

A. adopt DDS module to produce orthogonal demodulation signal, adopt this orthogonal demodulation signal to do digital modulation to discrete sampling sequence r (n), obtain two-way quadrature signal sequence, be designated as I (n) and Q (n) respectively; I (n) and Q (n) is adopted to construct the first sequence of complex numbers X (n), using the solid part signal of I (n) as the first sequence of complex numbers X (n), Q (n) is as the imaginary signals of the first sequence of complex numbers X (n), then the first sequence of complex numbers represents X (n)=I (n)+jQ (n), and wherein j represents the imaginary part unit of plural number;

B. the first sequence of complex numbers X (n) is divided into two sequence of complex numbers according to odd and even number discrete series number, the sequence of complex numbers being odd number by discrete series number is called the second sequence of complex numbers, be designated as X (2m+1), the sequence of complex numbers being even number by discrete series number is called the 3rd sequence of complex numbers, is designated as X (2m+2), m=0,1,2,3,

C. the compound correlative function of the first sequence of complex numbers X (n) is designated as R ₁n (), is designated as R by the compound correlative function of the second sequence of complex numbers X (2m+1) ₂n (), is designated as R by the compound correlative function of the 3rd sequence of complex numbers X (2m+2) ₃(n), $R_2\left(n\right)=\underset{m=0}{\overset{N^{\′}}{\Σ}}X(2m+1+l)X^{*}(2m+1),R_3\left(n\right)=\underset{m=0}{\overset{N^{\′\′}}{\Σ}}X(2m+2+l)X^{*}(2m+2),$ Wherein N' represents the multiple correlation length of the second sequence of complex numbers X (2m+1), N " represent the multiple correlation length of the 3rd sequence of complex numbers X (2m+2), and l is sample delay value and its value is the code length of 1 M sequence code, N=f _s× (t ₁-t ₀), f _sfor the sample frequency of alternating sampling module, t ₀for the sampling start time of alternating sampling module, t ₁the sampling finish time of alternating sampling module, X (2m+1+l)=I (2m+1+l)+jQ (2m+1+l), X (2m+2+l)=I (2m+2+l)+jQ (2m+2+l), I (2m+1+l) is the expression formula of I (2m+1) time delay l, I (2m+2+l) is the expression formula of I (2m+2) time delay l, Q (2m+1+l) is the expression formula of Q (2m+1) time delay l, and Q (2m+2+l) is the expression formula of Q (2m+2) time delay l; Symbol * is conjugate operation symbol, X ^*(2m+1) be the conjugate complex number sequence of X (2m+1), X ^*(2m+2) be the conjugate complex number sequence of X (2m+2);

D. by R ₂(n) and R ₃n () substitutes into formula R ₁(n)=R ₂(n)+R ₃r is obtained in (n) ₁n (), by R ₁n the real part of () is designated as Re (R ₁(n)), just R ₁n the imaginary part of () is designated as Im (R ₁(n));

E. by the compound correlative function R of the first sequence of complex numbers X (n) ₁n the phasing degree of () is designated as θ, adopt arctan function to solve phasing degree θ, namely $\mathrm{\θ}=\arctan \left(\frac{\mathrm{Im}\left(R_1\right(n\left)\right)}{\mathrm{Re}\left(R_1\right(n\left)\right)}\right);$

F. Doppler frequency deviation is designated as f _d, adopt formula wherein q is relevant repetition period, i.e. q=f _s× T;

G. the flow velocity that multiple correlation is measured is designated as v _c, adopt formula calculate v _c, c is the velocity of sound in water, c=1500m/s;

The detailed process of cross-spectrum auto-correlation algorithm process is:

A. the frequency spectrum of discrete sampling sequence r (n) is designated as r'(k), the autocorrelation function frequency spectrum of discrete sampling sequence r (n) is designated as R (k), and k represents the sequence number of frequency domain, k=0, and 1,2,3,

B. FFT computing is adopted to calculate frequency spectrum r'(k), make R (k)=r'(k) × r'(k);

C. carry out IFFT computing again after interpolation processing being done to autocorrelation function frequency spectrum R (k), obtain the autocorrelation function of discrete sampling sequence r (n), the autocorrelation function of discrete series r (n) is designated as R'(t);

D. by the autocorrelation function R'(t of discrete sampling sequence r (n)) determine the moment that the first secondary lobe occurs the moment that the first secondary lobe occurs to be designated as T ₁; The moment that ideally the first secondary lobe occurs that frequency deviation does not occur is designated as T ₂, wherein T ₂=T

E. the difference that Doppler's time delay occurs is designated as Δ T, Δ T=T ₁-T ₂;

F. the flow velocity of autocorrelation measurement is designated as v _z, adopt formula Δ T calculates the flow velocity v of autocorrelation measurement _z;

5. flow velocity corrected parameter is designated as α, α=[(v _z-v _c)/2v _max], v _maxfor fuzzy speed, v _max=c/4f ₀t, symbol " [] " is rounding operation symbol;

6. formula v=v is adopted _c+ α 2v _maxcalculate and treat velocity measurement v.