CN103926463A - Phase difference measurement method related to data continuation - Google Patents

Abstract

The invention relates to the field of signal processing, in particular to a phase difference measurement method related to data continuation. The suitable object of the method is phase difference measurement of any same frequency sinusoidal signals. The phase difference measurement method comprises the following steps that first, cycle pre-estimation is conducted on sampling signals; then whether data lengths are in full cycle or not is judged, full-cycle data continuation processing is conducted on non-full cycle data lengths, the length of related calculation is just or approximately the full cycle, and therefore the purpose of improving correlation method phase difference measurement precision is achieved; at last, a correlation method phase difference computational formula is used for calculating to obtain the phase difference of signals. According to the phase difference measurement method related to data continuation, implementation is simple, the anti-noise performance is higher, the precision is higher, the influence of non-full cycle sampling on the correlation method phase difference measurement precision is removed, and the correlation method phase difference measurement precision can be improved together by combining other methods.

Description

The relevant method for measuring phase difference of a kind of Data extension formula

Technical field

The present invention relates to signal process field, particularly the relevant method for measuring phase difference of a kind of Data extension formula.

Background technology

The measurement of phase differential all has important and application widely in fields such as fault diagnosis, navigator fix, test control, power electronics, communication, radars.For example, in electrical network electric energy metrical, power factor (PF) is definite, relates to the Measurement accuracy of phase differential between relevant voltage and electric current; In high-precision radar precision positioning, depend on the high-acruracy survey to same-frequency signal phase difference; The Coriolis flowmeter of being used widely in high-precision flow fields of measurement, the measuring accuracy of phase differential directly affects fluid mass measuring accuracy.Therefore, realize the high-acruracy survey of phase differential is seemed and is even more important.

Cross-correlation function value when correlation method utilizes the time delay of the same frequency of two-way sinusoidal signal to be zero obtains phase differential to the principle that the cosine value of its phase differential is directly proportional, and in signal is processed, is widely applied.Because noise signal is common and useful signal correlativity is less, correlation method has good noise inhibiting ability, can suppress all and incoherent various forms noise of reference signal, but for the harmonious wave interference of the stronger undesired signal of correlativity, measuring error is larger under Low SNR.In addition, correlation method requires signal to carry out strictly integer-period sampled, can realize the Measurement accuracy of phase differential, but measuring accuracy need to improve under non-integer-period sampled condition under integer-period sampled condition.For improving the measuring accuracy of phase differential, in succession there is multiple phase difference measurement new algorithm or improvement algorithm based on relative theory, mainly contain following several:

(1) list of references [1]: improve oneself in a left side, Xu Yang, Cao Xiaolong, Deng. calculate the improvement [J] of the related function method of capacitance type equipment dielectric dissipation factor. electric power network technique, 2004,28 (18): the basic thought of 53-57. document [1] method is to adopt interpolation to round the method in cycle integrating range is revised, with trapezoidal integration, replace rectangular integration logarithm value integration method to improve, and choose suitable sample frequency and sampling number improves computational accuracy.The method computational accuracy is subject to the impact of integrating range, numerical integration method and numerical integration step-length, exists more than influence factor and the problem of calculation of complex.

(2) list of references [2]: Li Jianmin, Zhao Peng, Hou Wen, Deng. the error research [J] of the phase differential algorithm based on correlation theory. Northcentral University's journal (natural science edition), 2009,30 (6): 616-619. document [2] adopts two correlation methods, need to produce two with reference signal frequently, that there are 90 ° of phase shifts, input and output signal is carried out to the phase differential that related operation is tried to achieve signal with two reference signals respectively.The method has reduced the phase error that randomness causes, but phase difference error in non-integer-period sampled situation is still larger.

(3) list of references [3]: Zheng Shengfeng, building petrel, model essay is brilliant, Deng. a kind of method for measuring phase difference [J] based on multiple correlation method. aerospace instrumentation technology, 2008,28 (5): 57-60. list of references [4]: Zheng Shengfeng, Chen Suming, Di Jinhai, Deng. a kind of phase difference measurement new method [J] based on multiple simple crosscorrelation. aerospace instrumentation technology, 2012,32 (1): 34-40. document [3] and document [4] utilize the special nature of sinusoidal signal, by signal being carried out to auto-correlation and computing cross-correlation repeatedly, recycling relative theory solves the phase differential of signal.The method can improve the Phase Difference Measuring Precision in low signal-to-noise ratio situation, the whether impact of complete cycle but measuring accuracy is also sampled.

In sum, the method for measuring phase difference of correlation method has important research meaning and using value, but there are problems in existing method, these methods or have the problem that calculated amount is larger, cannot eliminate the non-integer-period sampled impact that measurement result is brought to error completely, a kind of calculating need to be proposed simple, the non-integer-period sampled method for measuring phase difference that measurement result is brought to error effect can be eliminated completely.

Summary of the invention

The object of the invention is to propose a kind of calculating simple, can eliminate the non-integer-period sampled method for measuring phase difference that measurement result is brought to error effect, be applicable to the same phase difference measurement of sinusoidal signal frequently arbitrarily, solve the subject matter that existing correlation method phase difference measurement exists, expand its range of application.

The present invention proposes the relevant method for measuring phase difference of a kind of Data extension formula.

The present invention includes following steps:

The first step: the frequency of signal or cycle are carried out to pre-estimation.Suppose that the sampled data that obtains two-way short signal is respectively x=[x ₁, x ₂..., x _n-1, x _n], y=[y ₁, y ₂..., y _n-1, y _n], P ∈ Z ⁺for the sampling number of one-period, cycle estimated result is made as P.

Second step: whether judge complete cycle to data.Suppose that N is k divided by the business of P, remainder is m.If m=0, matches the complete cycle of persistence length and signal, become integral multiple relation, now data are without processing.If m > 0, illustrates that persistence length surpasses k cycle but do not reach k+1 cycle.

The 3rd step: carry out data processing.Can carry out in two kinds of situation data processing:

1) carry out the rejecting of data, according to the periodicity of signal, from sampled data x, in y, reject m data [x outside data complete cycle _kP+1..., x _kP+m], [y _kP+1..., y _kP+m], making persistence length is complete cycle, the burst of rejecting after data is:

x _e=[x ₁，x ₂，…，x _kP]，y _e=[y ₁，y ₂，…，y _kP]

2) carry out the continuation of data, according to the periodicity of signal, from sampled data x, in y, search P-m data [x _{(k-1) P+m+1}..., x _kP], [y _{(k-1) P+m+1}..., y _kP], being merged to former data sequence x, after y, making persistence length is complete cycle, the burst after Data extension is:

x _e=[x ₁，x ₂，…，x _N，x _(k-1)P+m+1，…，x _kP]，y _e=[y ₁，y ₂，…，y _N，y _(k-1)P+m+1，···，y _kP]

It is to be noted, for sinusoidal signal in short-term, due to the sampled data of picked up signal, count lessly, and in signal, be also with noisy interference, by data, reject and adjust the way that burst is complete cycle and can improve algorithm measurement precision, but can reduce data itself with quantity of information; The way that adopts Data extension, on the basis of counting at legacy data, has increased the quantity of information of data, when improving arithmetic accuracy, also can improve the noise immunity of algorithm, and its practicality is stronger.

The 4th step: utilize correlation method to calculate phase differential.The signal x after Data extension is processed _e, y _emake respectively auto-correlation and computing cross-correlation, recycling cross correlation function calculates phase differential.Suppose that Δ θ is the phase differential of two paths of signals, R _xy(0), R _xx(0), R _yy(0) be respectively x _e, y _esimple crosscorrelation and autocorrelation function,

$\mathrm{\Δ\θ}=\arccos \left(\frac{R_{\mathrm{xy}}\left(0\right)}{\sqrt{R_{\mathrm{xx}}\left(0\right)\sqrt{R_{\mathrm{yy}}\left(0\right)}}}\right)$

Accompanying drawing explanation

The present invention is further elaborated with concrete enforcement with reference to the accompanying drawings below.Take and describe as example containing white noise sinusoidal signal.

Fig. 1 is method realization flow figure.

In figure: 1 represents known sinusoidal sequence x (n) and y (n); 2 represent frequency or the cycle of x (n) and y (n) to carry out pre-estimation; 3 represent whether judge complete cycle to data, use N divided by P, and business is k, and remainder is m, and N represents that signal sampling counts, P ∈ Z ⁺sampling number for one-period; 4 represent signal x (n) and y (n) to carry out Data extension processing complete cycle; 5 represent to utilize correlation method phase difference calculating formula to try to achieve the phase differential of signal; 6 represent when m ≠ 0; 7 represent when m=0.

Fig. 2 is x (n) and the y (n) before continuation.

In figure: signal frequency is 1Hz, sample frequency is 15Hz, and signal to noise ratio (S/N ratio) is 20dB, sampling number N=20.

Fig. 3 is x (n) and the y (n) after continuation.

Embodiment

Method realization flow of the present invention as shown in Figure 1.

The specific embodiment of the present invention is as follows:

The two-way of take describes as example with sinusoidal signal (shown in Fig. 2) frequently.

1) offset of sinusoidal signal x, the frequency of y or cycle are carried out pre-estimation.Cycle estimated result is made as P ∈ Z ⁺, the sampling number that P is one-period.

2) to data, whether judge complete cycle.N is k divided by the business of P, and remainder is m.If m=0, matches the complete cycle of persistence length and signal, become integral multiple relation, now data are without processing.If m > 0, illustrates that persistence length surpasses k cycle but do not reach k+1 cycle.

3) carry out Data extension processing.According to the periodicity of signal, from sampled data x, in y, search P-m data [x _{(k-1) P+m+1}..., x _kP], [y _{(k-1) P+m+1}..., y _kP], being merged to former data sequence x, after y, making persistence length is complete cycle, the burst after Data extension is: x _e=[x ₁, x ₂..., x _n, x _{(k-1) P+m+1},, x _kP], y _e=[y ₁, y ₂..., y _n, y _{(k-1) P+m+1}..., y _kP] (as shown in Figure 3).

4) utilize correlation method phase difference calculating formula to calculate the phase differential of signal.

$\mathrm{\Δ\θ}=\arccos \left(\frac{R_{\mathrm{xy}}\left(0\right)}{\sqrt{R_{\mathrm{xx}}\left(0\right)}\sqrt{R_{\mathrm{yy}}\left(0\right)}}\right)$

Claims (1)

1. the relevant method for measuring phase difference of Data extension formula, is characterized in that: applicable object is the same phase difference measurement of sinusoidal signal frequently arbitrarily;

The method comprises the following steps:

The first step: the frequency of signal or cycle are carried out to pre-estimation.Suppose that the sampled data that obtains two-way short signal is respectively x=[x ₁, x ₂,, x _n-1, x _n], y=[y ₁, y ₂..., y _n-1, y _n], P ∈ Z ⁺for the sampling number of one-period, cycle estimated result is made as P.

Second step: whether judge complete cycle to data.Suppose that N is k divided by the business of P, remainder is M.If m=0, matches the complete cycle of persistence length and signal, become integral multiple relation, now data are without processing.If m > 0, illustrates that persistence length surpasses k cycle but do not reach k+1 cycle.

The 3rd step: carry out data processing.Can carry out in two kinds of situation data processing:

1) carry out the rejecting of data, according to the periodicity of signal, from sampled data x, in y, reject m data [x outside data complete cycle _kP+1..., x _kP+m], [y _kP+1..., y _kP+m], making persistence length is complete cycle, the burst of rejecting after data is:

x _e=[x ₁，x ₂，…，x _kP]，y _e=Ly ₁，y ₂，…，y _kP]

2) carry out the continuation of data, according to the periodicity of signal, from sampled data x, in y, search P-m data [x _{(k-1) P+m+1}..., x _kP], [y _{(k-1) P+m+1}..., y _kP], being merged to former data sequence x, after y, making persistence length is complete cycle, the burst after Data extension is:

x _e=[x ₁，x ₂，…，x _N，x _(k-1)P+m+1，…，x _kP]，y _e=[y ₁，y ₂，…，y _N，y _(k-1)P+m+1，…，y _kP]

It is to be noted, for sinusoidal signal in short-term, due to the sampled data of picked up signal, count lessly, and in signal, be also with noisy interference, by data, reject and adjust the way that burst is complete cycle and can improve algorithm measurement precision, but can reduce data itself with quantity of information; The way that adopts Data extension, on the basis of counting at legacy data, has increased the quantity of information of data, when improving arithmetic accuracy, also can improve the noise immunity of algorithm, and its practicality is stronger.

The 4th step: utilize correlation method to calculate phase differential.The signal x after Data extension is processed _e, y _emake respectively auto-correlation and computing cross-correlation, recycling cross correlation function calculates phase differential.Suppose that Δ θ is the phase differential of two paths of signals, R _xy(0), R _xx(0), R _yy(0) be respectively x _e, y _esimple crosscorrelation and autocorrelation function,

$\mathrm{\Δ\θ}=\arccos \left(\frac{R_{\mathrm{xy}}\left(0\right)}{\sqrt{R_{\mathrm{xx}}\left(0\right)}\sqrt{R_{\mathrm{yy}}\left(0\right)}}\right)$