CN105675962A - Voltage and current sampling method for on-line monitoring of transformer winding state - Google Patents

Abstract

The invention relates to a voltage and current sampling method for on-line monitoring of a transformer winding state. The method comprises: a transformer voltage or current signal collected by a transformer is obtained; and the voltage or current signal is processed to obtain a to-be-sampled waveform. In addition, the invention also includes: (1), a sampling frequency fs and a sampling point number N are set and the to-be-sampled waveform is sampled to obtain an N-point time domain sequence x(n); (2), time domain-frequency conversion is carried out on the time domain sequence x(n) to obtain a corresponding frequency domain signal waveform; (3), a target frequency range is set according to a first amplitude maximum point of the frequency domain signal waveform; and (4), refinement processing is carried out on a frequency within the target frequency range by using a frequency refinement algorithm to obtain a second amplitude maximum point, so that a final amplitude maximum point is obtained. Compared with the prior art, the provided method enables data collection and on-line monitoring accuracy to be improved effectively.

Description

A kind of electric current and voltage method of sampling for transformer winding state on-line monitoring

Technical field

The present invention relates to a kind of data sampling method, especially relate to a kind of electric current and voltage method of sampling for transformer winding state on-line monitoring.

Background technology

Transformer is one of electric installation the most important in power system, for guaranteeing that whole electric power netting safe running is significant. And the statistical information according to Accident of Transformer over the years, the power transformer of 35kV and above grade damages because of short trouble and accounts for the 50% of total number of accident. According to maintenance situation, this accident causes overwhelming majority transformer winding distortion in various degree and insulation breakdown. Along with China's rapid economic development, net capacity expands day by day, and system short-circuit capacity also increases thereupon, and the statistical number causing transformer to damage because of short trouble also rises year by year. Therefore, winding failure has become one of main reason of transformer fault and electric power system fault, how to monitor the slight deformation of winding in transformer operational process, follow the tracks of winding state, making early warning before it gross distortion occurs, the impact of electromagnetic transient enabling them better resist short-circuit current and generation thereof is the crucial matter of science and technology needing solution during transformer winding fault is monitored badly.

It is out of shape in the implementation process of on-line monitoring in transformer winding, the current transformer from operating transformer and voltage transformer is needed to obtain the electric current and voltage value of transformer actual motion, at signal after hall sensing device or electric parameters transducer, carrying out effective low pass filtered wave energy makes other harmonic components in signal obviously reduce, last it is sampled waveform to stablize, in this process, what should be noted is only that the phase differential that the filtering circuit adopted brings needs to compensate in computation program. Under the guaranteed prerequisite of hardware accuracy, the key obtaining accurate voltage current data is just the sampling accuracy rate of signal acquiring system. In actual moving process, the frequency of voltage and current signal constantly can change along with the change of electric network active, its frequency values is not " 50Hz " in broad sense cognition, and the current method of sampling to power system voltage electric current is much the resolving power based on 50Hz, this kind of sampling means can introduce certain error.

Summary of the invention

The object of the present invention is exactly provide the electric current and voltage method of sampling for transformer winding state on-line monitoring of a kind of effective raising data gathering and on-line monitoring accuracy to overcome defect that above-mentioned prior art exists.

The object of the present invention can be achieved through the following technical solutions:

A kind of electric current and voltage method of sampling for transformer winding state on-line monitoring, first the method obtains voltage or the electric current signal of the transformer gathered by mutual inductor, described voltage or electric current signal being processed, obtain waveform to be sampled, the method also comprises the following steps:

1) set sample frequency fs and sampling number N, described waveform to be sampled is sampled, obtain N point time domain sequences x (n);

2) described time domain sequences x (n) is carried out time domain frequency-domain transform, obtain corresponding frequency domain signal waveform;

3) according to the first amplitude maximum point target setting range of frequency of described frequency domain signal waveform;

4) adopt Frequence zooming algorithm that the frequency in described range of target frequencies is carried out thinning processing, obtain the 2nd amplitude maximum point, be final amplitude maximum point.

Described step 2) in, time domain sequences x (n) is carried out time domain frequency-domain transform and is specially:

The Fourier coefficient obtaining time domain sequences x (n) is

$\left\{\begin{array}{c}{a}_k=\frac{2}{N}\underset{n=0}{\overset{N-1}{\Σ}}x\left(n\right)\cos (2\mathrm{\π}nk/N)\\ b_k=\frac{2}{N}\underset{n=0}{\overset{N-1}{\Σ}}x\left(n\right)\sin (2\mathrm{\π}nk/N)\end{array}\right.,k=0,1...N/2$

Differentiating frequency k Δ f place amplitude spectrum vector expression for each is a_k-jb_k, i.e. a_kAnd b_kThe real part of the corresponding spectral magnitude of difference and imaginary part, Δ f=fs/N.

Described step 3) in, target setting range of frequency centered by the frequency corresponding to the first amplitude maximum point.

Described step 4) in, Frequence zooming algorithm is specially:

In range of target frequencies [f1, f2], by a_kAnd b_kExpression formula in k see the real variable f connecting change, then obtain:

$\left\{\begin{array}{c}a\left(f\right)=\frac{2}{N}\underset{n=0}{\overset{N-1}{\Σ}}x\left(n\right)\cos (2\mathrm{\π}nf/fs)\\ b\left(f\right)=\frac{2}{N}\underset{n=0}{\overset{N-1}{\Σ}}x\left(n\right)\sin (2\mathrm{\π}nf/fs)\end{array}\right.,f\∈\[f1,f2\]$

Wherein, f1 and f2 is respectively two borders of range of target frequencies, and f1 < f2, according to the expression formula of a (f) He b (f), get M number of sampling, obtain the continuous spectrum curve in range of target frequencies, obtain the 2nd amplitude maximum point according to described connection spectrum curve.

The frequency resolution ax f ' of described Frequence zooming algorithm meets Δ f '=(f2-f1)/(M-1).

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

1) the inventive method can accurately obtain system electric current and voltage, and then improves the accuracy of winding state monitoring.

2) the present invention adopts Frequence zooming algorithm in range of target frequencies, obtaining while accurate voltage current value, effectively reduce calculated amount, it is to increase data-handling efficiency.

3) range of target frequencies is carried out refinement by the present invention, can more accurately draw actual value and the phase place of signal, the calculating accuracy of short-circuit reactance is had greatly improved.

Accompanying drawing explanation

Fig. 1 is the schematic flow sheet of the present invention;

Fig. 2 is the frequency domain signal diagrams that existing method obtains;

Fig. 3 is the frequency domain signal diagrams that Frequence zooming algorithm of the present invention obtains.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is described in detail. The present embodiment is implemented premised on technical solution of the present invention, gives detailed enforcement mode and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

In the process of monitoring transformer winding state, first the voltage and current signal of transformer needs to convert through superpotential and current transformer, by hall sensing device and filtering circuit, it is processed the thresholding to reach digital sample further again, first sampling process is time-domain signal x (t) is sampled by sample frequency taking fs, sampling obtains N point time domain sequences x (n) afterwards is divided into two to wait long sequence with odd even, namely

$\left\{\begin{array}{c}{x}_1\left(r\right)=x\left(2r\right)\\ x_2\left(r\right)=x(2r+1)\end{array}\right.,r=0,1,...,N/2-1---\left(1\right)$

If the factorCalculate its frequency domain value

$\begin{array}{c}X\left(k\right)=\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)e^{-j\frac{2\mathrm{\&pi;}}{N}kn}=\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)W_N^{nk}\\ =\underset{r=0}{\overset{N/2-1}{\&Sigma;}}x\left(2r\right)W_{N/2}^{rk}+\underset{r=0}{\overset{N/2-1}{\&Sigma;}}x(2r+1)W_{N/2}^{(2r+1)k}\\ =\underset{r=0}{\overset{N/2-1}{\&Sigma;}}{x}_1\left(r\right)W_{N/2}^{rk}+W_N^k\underset{r=0}{\overset{N/2-1}{\&Sigma;}}{x}_2\left(r\right)W_{N/2}^{rk}\\ =X_1\left(k\right)+W_N^k{X}_2\left(k\right)\end{array},k=0,1,\mathrm{...}N/2-1---\left(2\right)$

For expression formula when k=N/2, N/2+1.....N-1, according toHave

$X_1(k+N/2)=\underset{0}{\overset{N/2-1}{\&Sigma;}}{x}_1\left(r\right)W_{N/2}^{r(k+N/2)}=\underset{0}{\overset{N/2-1}{\&Sigma;}}{x}_1\left(r\right)W_{N/2}^{rk}=X_1\left(k\right)---\left(3\right)$

With managing X₂(k+N/2)=X₂(k)

According to $W_N^{r(k+N/2)}=-W_N^k$ Then have

$\left\{\begin{array}{c}X\left(k\right)=X_1\left(k\right)+W_N^k{X}_2\left(k\right)\\ X(k+N/2)=X_1\left(k\right)-W_N^k{X}_2\left(k\right)\end{array}\right.,k=0,1,....N/2-1---\left(4\right)$

In said process, X (k) is the frequency domain value of respective signal, its time-domain value can be known by its real part, imaginary values size, the sampling resolution Δ f=fs/N that it can reach, it is subject to sample frequency fs and the restriction of sampling number N, under the prerequisite that sample frequency is constant, put forward high frequency resolution, just must increase sampling number, but increase the calculated amount that sampling number will increase said process, the present invention, in order to not increase calculated amount, adopts Frequence zooming algorithm in range of target frequencies.

As shown in Figure 1, a kind of electric current and voltage method of sampling for transformer winding state on-line monitoring that the present embodiment provides comprises the following steps:

S101, obtains voltage or the electric current signal of the transformer gathered by mutual inductor, described voltage or electric current signal is processed, obtain waveform to be sampled.

S102, setting sample frequency fs and sampling number N, samples to described waveform to be sampled, obtains N point time domain sequences x (n).

S103, carries out time domain frequency-domain transform to described time domain sequences x (n), and the Fourier coefficient obtaining time domain sequences x (n) is:

$\left\{\begin{array}{c}{a}_k=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)\cos (2\mathrm{\&pi;}nk/N)\\ b_k=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)\sin (2\mathrm{\&pi;}nk/N)\end{array}\right.,k=0,1...N/2---\left(5\right)$

Differentiating frequency k Δ f place amplitude spectrum vector expression for each is a_k-jb_k, i.e. a_kAnd b_kThe real part of the corresponding spectral magnitude of difference and imaginary part, Δ f=fs/N, thus obtain corresponding frequency domain signal waveform.

S104, according to the first amplitude maximum point target setting range of frequency of described frequency domain signal waveform. Range of target frequencies can set centered by the frequency corresponding to the first amplitude maximum point.

S105, adopts Frequence zooming algorithm that the frequency in described range of target frequencies is carried out thinning processing, obtains the 2nd amplitude maximum point, be final amplitude maximum point.

Know by sampling theorem, sequence x (n) contains the band information of 0～fs/2, if so calculating with continuous print Fourier pair frequency spectrum, spectrum curve is regarded as continuous print, namely the k in wushu (5) regards the real variable f (0≤f≤fs/2) of interval (0≤k≤N/2) interior consecutive variations as, then can formula (5) be become:

$\left\{\begin{array}{c}a\left(f\right)=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)\cos (2\mathrm{\&pi;}nf/fs)\\ b\left(f\right)=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)\sin (2\mathrm{\&pi;}nf/fs)\end{array}\right.---\left(6\right)$

Utilizing formula (6) just can obtain continuous print spectrum curve, thus make frequency resolving power no longer by the restriction of sampling number N, f is a continuous print frequency. At field of definition f ∈ [f1, f2] in, formula (6) is directly adopted to calculate, then the number of sampling M in frequency-selecting band is set according to required frequency resolution ax f ', frequency resolution ax f ' meets Δ f '=(f2-f1)/(M-1), obtain the continuous spectrum curve in range of target frequencies, obtain the 2nd amplitude maximum point according to described connection spectrum curve.

The calculating of concrete example is carried out according to aforesaid method:

AssumeWith sample frequency fs=1kHz, it is sampled by sampling number N=1000, calculate frequency domain signal such as accompanying drawing 2, the frequency of its amplitude maximum point is 50Hz, and amplitude size is 0.934, amplitude and original signal deviation 6.6%, create bigger deviation, this part deviation has a huge impact to the calculated value of short-circuit reactance, it can be seen that when frequency resolving power is 1Hz, the sampled value of amplitude can not reflect that it is truly worth.

According to Frequence zooming method, the frequency in this interval of 48.8Hz to 50.8Hz is carried out thinning processing, number of sampling M=21, then the frequency resolving power in refinement interval is 0.1Hz, and calculation result is such as accompanying drawing 3.

After Frequence zooming, amplitude maximum point is 49.8Hz, and amplitude size is 1.0014, differs 0.14% with true value, and frequency values is completely identical, compares with the result before refinement, and its accuracy improves greatly. Simultaneously value according to a (f), b (f) can calculate its phase place, its arc-tangent value is 1.743, corresponding radian is 1.0499,0.25% is differed with actual value, visible, range of target frequencies is carried out refinement, can more accurately draw actual value and the phase place of signal, the calculating accuracy of short-circuit reactance is had greatly improved.

Claims (5)

1. the electric current and voltage method of sampling for transformer winding state on-line monitoring, first the method obtains voltage or the electric current signal of the transformer gathered by mutual inductor, described voltage or electric current signal is processed, obtains waveform to be sampled, it is characterized in that, the method also comprises the following steps:

1) set sample frequency fs and sampling number N, described waveform to be sampled is sampled, obtain N point time domain sequences x (n);

2) described time domain sequences x (n) is carried out time domain frequency-domain transform, obtain corresponding frequency domain signal waveform;

3) according to the first amplitude maximum point target setting range of frequency of described frequency domain signal waveform;

4) adopt Frequence zooming algorithm that the frequency in described range of target frequencies is carried out thinning processing, obtain the 2nd amplitude maximum point, be final amplitude maximum point.

2. the electric current and voltage method of sampling for transformer winding state on-line monitoring according to claim 1, it is characterised in that, described step 2) in, time domain sequences x (n) is carried out time domain frequency-domain transform and is specially:

The Fourier coefficient obtaining time domain sequences x (n) is

$\left\{\begin{array}{c}{a}_k=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)cos(2\mathrm{\&pi;}nk/N)\\ b_k=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)\sin (2\mathrm{\&pi;}nk/N)\end{array}\right.,k=0,1...N/2$

Differentiating frequency k Δ f place amplitude spectrum vector expression for each is a_k-jb_k, i.e. a_kAnd b_kThe real part of the corresponding spectral magnitude of difference and imaginary part, Δ f=fs/N.

3. the electric current and voltage method of sampling for transformer winding state on-line monitoring according to claim 1, it is characterised in that, described step 3) in, target setting range of frequency centered by the frequency corresponding to the first amplitude maximum point.

4. the electric current and voltage method of sampling for transformer winding state on-line monitoring according to claim 2, it is characterised in that, described step 4) in, Frequence zooming algorithm is specially:

In range of target frequencies [f1, f2], by a_kAnd b_kExpression formula in k see the real variable f connecting change, then obtain:

$\left\{\begin{array}{c}a\left(f\right)=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)cos(2\mathrm{\&pi;}nf/fs)\\ b\left(f\right)=\frac{2}{N}\underset{n=0}{\overset{N-1}{\&Sigma;}}x\left(n\right)\sin (2\mathrm{\&pi;}nf/fs)\end{array}\right.,f\&Element;\&lsqb;f1,f2\&rsqb;$

Wherein, f1 and f2 is respectively two borders of range of target frequencies, and f1 < f2, according to the expression formula of a (f) He b (f), get M number of sampling, obtain the continuous spectrum curve in range of target frequencies, obtain the 2nd amplitude maximum point according to described connection spectrum curve.

5. the electric current and voltage method of sampling for transformer winding state on-line monitoring according to claim 4, it is characterised in that, the frequency resolution ax f ' of described Frequence zooming algorithm meets Δ f '=(f2-f1)/(M-1).