CN104569581A - Multi-level set and single-cycle estimation method of power grid frequency measuring - Google Patents

Abstract

The invention provides a multi-level set and single-cycle estimation method of power grid frequency measuring. The method includes: performing low-pass filter on an input signal, and then sampling the signal; setting more than two thresholds horizontally intersecting with a sinusoidal periodic signal, determining the adjacent intersection points, whose slope is identical with the slope of two tangent lines of the sinusoid, of each threshold and signal sampling points close to the two intersection points, directly calculating the fundamental frequency of the threshold if the two intersection points of the threshold coincide with the signal sampling points, or else respectively substituting the signal sampling points close to the intersection points into a Lagrange interpolation formula, and calculating the fundamental frequency estimation value corresponding to the threshold; determining the weight of the multi-level set fundamental frequency estimation value acquired by each threshold according to the absolute value of the tangent line slope of the intersection points of each threshold and the sinusoidal periodic signal, and using a weighted average method to obtain the final fundamental frequency estimation value. The method has the advantages that multi-level set and single-cycle estimation frequency is achieved by multiple thresholds, and the method is good in instantaneity and high in precision.

Description

A kind of multilevel collection monocycle method of estimation of grid frequency measurement

Technical field

The invention belongs to power domain, be specifically related to a kind of multilevel collection monocycle method of estimation of grid frequency measurement.

Background technology

Along with improving constantly of national economy and people's living standard, people grow with each passing day to the demand of electric energy and the requirement of the quality of power supply.Wherein, the frequency of electrical network is a very important power quality index, and now when measuring fundamental frequency, frequent employing integer-period sampled some counting method, this method is when the sampling period is less, its the non-constant of precision, when the sampling period is long time, its real-time is very poor, and this method is when survey frequency, filtering process is not carried out to high frequency harmonic signals, when harmonic wave interferes with to a certain degree, can, on the raw certain impact of fixed output quota really of whole primitive period, measured frequency finally can be caused to have very large error.

Summary of the invention

The object of the invention is the deficiency in order to overcome above-mentioned power grid frequency measurement method, propose a kind of multilevel collection monocycle method of estimation of grid frequency measurement, the method effectively can solve the existing frequency measurement method problem that precision is poor and real-time is poor when measuring mains frequency.

The present invention proposes a kind of multilevel collection monocycle method of estimation of grid frequency measurement, described method comprises the steps:

Step one: carry out low-pass filtering treatment to input signal low-pass filter, eliminates the interference of high-frequency harmonic;

Step 2: select suitable sample frequency f _swith sampling length N, equal interval sampling is carried out to signal, sample frequency f _s2 times of the frequency of most higher harmonic component contained by signal should be not less than, obtain signal discrete sample sequence x (t _n), n=0,1,2 ..., N-1;

Step 3: find out discrete sampling sequence x (t _n) in maximal value MAX and minimum value MIN, then m threshold value Y be set _i, m be more than or equal to 2 integer, i=1,2 ..., m;

Step 4: determine adjoining nodes that each threshold value is identical with sinusoidal 2 tangent slopes and close on the signal sampling point of these 2 intersection points, the time interval of the adjoining nodes that 2 tangent slopes are identical is 1 primitive period of signal; If 2 of certain threshold value intersection points all overlap with signal sampling point, then directly calculate fundamental frequency estimated value f _i; If 2 of certain threshold value intersection points do not overlap with signal sampling point, then the signal sampling point of this threshold value and the above-mentioned near intersections of sinusoidal signal is substituted into Lagrange interpolation formula, calculate this threshold value Y _icorresponding fundamental frequency estimated value; After all threshold calculations, m fundamental frequency estimated value of acquisition is multilevel collection fundamental frequency estimated value f _i;

Step 5: to obtaining multilevel collection fundamental frequency estimated value f in step 4 _ibe weighted average, calculate final fundamental frequency estimated value f, wherein weights q _iaccording to the sinusoidal curve tangent slope absolute value of each threshold value and sinusoidal intersection point | K _i| determine, slope absolute value is larger, and weights are less, final fundamental frequency estimated value f=q ₁f ₁+ q ₂f ₂+ ... + q _mf _m.

Described method, in step 3, described threshold value Y _ithe maximal value in discrete sampling sequence x (n) should be less than and be greater than the minimum value in discrete sampling sequence x (n), and for ensureing that set threshold value is in rational scope, eliminate the interference that other uncertain factor causes, make it meet: 0.9MIN+0.1MAX<Y _i<0.9MAX+0.1MIN.

Described method, in step 4, if 2 of certain threshold value intersection points do not overlap with signal sampling point, then substitutes into Lagrange interpolation formula by the signal sampling point of this threshold value and the above-mentioned near intersections of sinusoidal signal, calculates this threshold value Y _icorresponding fundamental frequency estimated value is simplified operation, the sampled signal discrete data [t, x (t)] of this threshold value and sinusoidal signal near intersections can be substituted into 4 Lagrange interpolation formulas, through calculating threshold value Y _icorresponding fundamental frequency estimated value f _i.

Described method, in step 5, weights q _iaccording to the sinusoidal curve tangent slope absolute value of each threshold value and sinusoidal intersection point | K _i| determine, f _icorresponding weights are:

$q_i=\frac{1/\left|K_i\right|}{1/\left|K_1\right|+1/\left|K_2\right|+...+/\left|K_m\right|}.$

Beneficial effect: the present invention adopts multilevel collection to realize grid frequency measurement, overcomes in the past that single threshold method is easily by the defect of noise, and the present invention can realize frequency measurement within the monocycle, has the advantages that precision is high, real-time is good.

Accompanying drawing explanation

Fig. 1 is the theory diagram for the treatment of scheme of the present invention;

Fig. 2 is single threshold Frequency Estimation schematic diagram of the present invention;

Fig. 3 is the Frequency Estimation schematic diagram that intersection point of the present invention overlaps with sampled point;

Fig. 4 is the Frequency Estimation schematic diagram that intersection point of the present invention does not overlap with sampled point;

Fig. 5 is multi thresholds Frequency Estimation schematic diagram of the present invention.

Embodiment

The present invention proposes a kind of multilevel collection monocycle method of estimation of grid frequency measurement.The present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

1. the signal of the embodiment of the present invention to input first carries out low-pass filtering treatment with low-pass filter, then with sample frequency f _s=6400Hz carries out periodic sampling to signal, finds out maximal value MAX and minimum value MIN;

The model of first-harmonic sinusoidal periodic signal is:

x(t)＝A sin(2πft+θ)+B (1)

In formula, A represents the amplitude of sinusoidal periodic signal, and f represents the frequency of sinusoidal periodic signal, and θ represents the initial phase of sinusoidal periodic signal, and B represents the DC component of sinusoidal periodic signal.As shown in Figure 2, this sinusoidal periodic signal x (t) intersects with certain threshold value b and the time interval of identical 2 the adjacent intersection points of sinusoidal curve tangent slope is 1 primitive period of signal.Therefore, by these 2 intersection points, a fundamental frequency estimated value of signal just can be obtained.

2. with sample frequency f _scarry out equidistant discrete sampling to the cycle simulating signal shown in formula (1), generally, in real process, threshold value b is generally difficult to overlap with signal sampling point with the intersection point of sinusoidal signal, therefore in two kinds of situation:

1. as shown in Figure 3, the above-mentioned intersection point of threshold value b overlaps with signal sampling point, directly should calculate the corresponding fundamental frequency estimated value of this threshold value:

$f_g=\frac{1}{t_b^{\text{'}}-t_b}=\frac{1}{t_i-t_k}---\left(2\right)$

F in formula _gthe fundamental frequency estimated value corresponding to threshold value b;

2. as shown in Figure 4, the intersection point of threshold value b does not overlap with signal sampling point, obtains intersection point (t by sampling _b, the sampled point [t b) _k-2, x (t _k-2)], [t _k-1, x (t _k-1)], [t _k, x (t _k)], [t _k+1, x (t _k+1)] and [t _k+2, x (t _k+2)].Adopt Lagrange method of interpolation, 4 interpolation polynomial L can be obtained ₄(t) be:

$L_4\left(t\right)=\underset{i=k-2}{\overset{k+2}{\mathrm{\&Sigma;}}}\left(\underset{j\&NotEqual;i}{\overset{k+2}{\underset{j=k-2}{\mathrm{\&Pi;}}}}\frac{t-t_j}{t_i-t_j}\right)x\left(t_i\right)---\left(3\right)$

By (t _b, b) substitute into (3) and obtain:

L ₄(t _b)＝b (4)

T is solved by formula (4) _b, in like manner can obtain t ' _b, have:

$f_g=\frac{1}{t_b^{\text{'}}-t_b}---\left(5\right)$

F in formula _gthe fundamental frequency estimated value corresponding to threshold value b.So far, the estimation to fundamental frequency value corresponding to threshold value b is completed.

3. for sake of convenience, only the situation that 5 threshold values are crossing with sinusoidal periodic signal is explained below:

5 threshold value Ys crossing with sinusoidal cycles sampled signal are set _i, and make it in the reasonable scope, there is 0.9MIN+0.1MAX<Y _i<0.9MAX+0.1MIN, wherein i=1,2,3,4,5.By the fundamental frequency estimated value that above-mentioned principle can be corresponding with each threshold value, can with threshold value Y ₁corresponding fundamental frequency estimated value is f ₁; With threshold value Y ₂corresponding fundamental frequency estimated value is f ₂; With threshold value Y ₃corresponding fundamental frequency estimated value is f ₃; With threshold value Y ₄corresponding fundamental frequency estimated value is f ₄; With threshold value Y ₅corresponding fundamental frequency estimated value is f ₅.According to each threshold value Y _iwith sinusoidal curve place, sinusoidal cycles sampled signal intersection point place tangent slope absolute value | K _i| size, wherein K _ivalue by Y _iwith 4 interpolation polynomial L of the above-mentioned intersection point of sinusoidal curve ₄t the derivative of () is obtained, that is:

$K_i=L_4^{\text{'}}\left(t_{Y_i}\right)---\left(6\right)$

To the fundamental frequency estimated value f corresponding to each threshold value _iweighting, absolute value is larger, and weights are less, f _icorresponding weights are i=1,2,3,4,5; Recycling calculated with weighted average method obtains final frequency estimation:

$\begin{array}{c}f=\frac{1/\left|K_1\right|}{1/\left|K_1\right|+1/\left|K_2\right|+...+1/\left|K_5\right|}{f}_1+\frac{1/\left|K_2\right|}{1/\left|K_1\right|+1/\left|K_2\right|+...+1/\left|K_5\right|}{f}_2\\ +\frac{1/\left|K_3\right|}{1/\left|K_1\right|+1/\left|K_2\right|+...+1/\left|K_5\right|}{f}_3+\frac{1/\left|K_4\right|}{1/\left|K_1\right|+1/\left|K_2\right|+...+1/\left|K_5\right|}{f}_4\\ +\frac{1/\left|K_5\right|}{1/\left|K_1\right|+1/\left|K_2\right|+...+1/\left|K_5\right|}{f}_5\end{array}---\left(7\right)$

In formula, f is final fundamental frequency estimated value.

So far, the multilevel collection monocycle completing grid frequency measurement is estimated.The method calculates simple, and the adjoining nodes identical by the sinusoidal tangent slope of threshold value and sampled signal just can record electrical network fundamental frequency, and have employed the method for 4 Lagrange method of interpolation and weighting, has extraordinary real-time and accuracy.

Claims (4)

1. a multilevel collection monocycle method of estimation for grid frequency measurement, it is characterized in that, described method comprises the steps:

Step one: carry out low-pass filtering treatment to input signal low-pass filter, eliminates the interference of high-frequency harmonic;

Step 2: select suitable sample frequency f _swith sampling length N, equal interval sampling is carried out to signal, sample frequency f _s2 times of the frequency of most higher harmonic component contained by signal should be not less than, obtain signal discrete sample sequence x (t _n), n=0,1,2 ..., N-1;

Step 3: find out discrete sampling sequence x (t _n) in maximal value MAX and minimum value MIN, then m threshold value Y be set _i, m be more than or equal to 2 integer, i=1,2 ..., m;

Step 4: determine adjoining nodes that each threshold value is identical with sinusoidal 2 tangent slopes and close on the signal sampling point of these 2 intersection points, the time interval of the adjoining nodes that 2 tangent slopes are identical is 1 primitive period of signal; If 2 of certain threshold value intersection points all overlap with signal sampling point, then directly calculate fundamental frequency estimated value f _i; If 2 of certain threshold value intersection points do not overlap with signal sampling point, then the signal sampling point of this threshold value and the above-mentioned near intersections of sinusoidal signal is substituted into Lagrange interpolation formula, calculate this threshold value Y _icorresponding fundamental frequency estimated value; After all threshold calculations, m fundamental frequency estimated value of acquisition is multilevel collection fundamental frequency estimated value f _i;

Step 5: to obtaining multilevel collection fundamental frequency estimated value f in step 4 _ibe weighted average, calculate final fundamental frequency estimated value f, wherein weights q _iaccording to the sinusoidal curve tangent slope absolute value of each threshold value and sinusoidal intersection point | K _i| determine, slope absolute value is larger, and weights are less, final fundamental frequency estimated value f=q ₁f ₁+ q ₂f ₂+ ... + q _mf _m.

2. method according to claim 1, is characterized in that, in step 3, and described threshold value Y _ithe maximal value in discrete sampling sequence x (n) should be less than and be greater than the minimum value in discrete sampling sequence x (n), and for ensureing that set threshold value is in rational scope, eliminate the interference that other uncertain factor causes, make it meet: 0.9MIN+0.1MAX<Y _i<0.9MAX+0.1MIN.

3. method according to claim 1, it is characterized in that, in step 4, if 2 of certain threshold value intersection points do not overlap with signal sampling point, then the signal sampling point of this threshold value and the above-mentioned near intersections of sinusoidal signal is substituted into Lagrange interpolation formula, calculate this threshold value Y _icorresponding fundamental frequency estimated value is simplified operation, the sampled signal discrete data [t, x (t)] of this threshold value and sinusoidal signal near intersections can be substituted into 4 Lagrange interpolation formulas, through calculating threshold value Y _icorresponding fundamental frequency estimated value f _i.

4. method according to claim 1, is characterized in that, in step 5, and weights q _iaccording to the sinusoidal curve tangent slope absolute value of each threshold value and sinusoidal intersection point | K _i| determine, f _icorresponding weights are: $q_i=\frac{1/\left|K_i\right|}{1/\left|K_1\right|+1/\left|K_2\right|+\&CenterDot;\&CenterDot;\&CenterDot;+1/\left|K_m\right|}.$