CN103050942B - Current transformer (CT) saturation detection method based on Hilbert-Huang transformation (HHT) - Google Patents

Abstract

The invention discloses a current transformer (CT) saturation detection method based on Hilbert-Huang transformation (HHT), and the method comprises the followings steps that instantaneous frequency spectrum, Hilbert spectrum and Hilbert marginal spectrum are obtained through EMD decomposition and Hilbert diversion by collecting differential current signals on both sides of a transformer; the detection of saturation faults and CT saturation outside a transformer area is completed respectively arranged to the three types of spectrum so as to realize the quick action and the reliable action of differential protection under CT saturation; the method serves as a preferential improvement to a traditional CT saturation detection method and combines instantaneous frequency criteria, Hilbert spectrum criteria and Hilbert marginal spectrum criteria in the forms of 'and' and 'or'; when a judgment result has errors, another method can correctly judge, so that the reliability of differential protection is improved; and in addition, fault current outside the transformer area, fault current in the area, conversion fault current and the like under the two circumstances of CT saturation and unsaturation are fully considered, and the CT saturation detection method based on HHT has the advantages of stronger functions, higher efficiency, higher reliability and the like.

Description

A kind of CT saturation detection method based on HHT

Technical field

The invention belongs to field of relay protection in power, be specifically related to a kind of CT saturation detection method based on HHT.

Background technology

In China be in the intelligent grid process of construction of bulk transmission grid with extra-high voltage grid, the expansion of electric power system scale and the raising of electric pressure, objectively require to configure the power equipment of more Large Copacity and more voltage levels.Putting into operation of high-power transformer, proposes higher requirement to relaying protection.Recent statistics data show, the average motion accuracy of 2001 ~ 2011 years 220kV and above transformer is 79.794%, lower than generator and route protection about 20 percentage points.At present, differential protection is as one of the main protection of power transformer interior fault, and its reliability is by the saturated impact of CT.Under CT saturated conditions, when external area error occurs, the differential current of transformer both sides current transformer is very large, can cause the misoperation of differential protection.

At present, in engineering reality, the general ratio braking method that adopts is differentiated, but this kind of method can not ensure the reliability of differential protection under CT extremely saturated conditions, must be aided with other and differentiate the measure that CT is saturated.In addition domestic and international proposed recognition methods also has time difference method, harmonic restraint method, Wavelet Singularity detection method, second order or three order derivative methods etc.; these methods can ensure the fail safe of transformer differential protection when experiencing fault substantially; but also there are some problems on the swiftness of time or accuracy, some method above-mentioned is too complicated and be not suitable for practical application.

Hilbert-Huang transform (Hilbert-Huang Transform), i.e. HHT conversion, as the emerging intelligent algorithm that a kind of 1998 propose, application in field of power just starts, analyze read lot of documents basis on by this algorithm application in CT saturation detection, a kind of simple, efficient new approaches can not only be provided for CT is saturated, widened the application of this algorithm simultaneously to a certain extent, there is very high theory value and Research Significance.

Summary of the invention

The object of the embodiment of the present invention is to propose a kind of CT saturation detection method based on HHT; can under CT saturated conditions by HHT; easily by transformer external area error, troubles inside the sample space and and district outside turn to the evolved fault in district correctly to differentiate, be beneficial to and solve the misoperation problem of differential protection.

To achieve these goals, the present invention adopts following technical scheme:

Based on a CT saturation detection method of HHT, comprise the steps:

A, collection transformer both sides differential current signal, obtain the intrinsic mode function that several meet IMF condition, i.e. IMF component after EMD decomposes;

B, carry out Hilbert conversion to above-mentioned IMF component, obtain instantaneous frequency, obtain Hilbert spectrum and Hilbert marginal spectrum, detailed process is as follows:

IMF component is designated as X (t), and carrying out Hilbert conversion Y (t) to it is:

$Y\left(t\right)=\frac{1}{\mathrm{\π}}{\∫}_{-\∞}^{+\∞}\frac{X\left(\mathrm{\τ}\right)}{t-\mathrm{\τ}}\mathrm{d\τ}---\left(1\right)$

Then the analytic signal Z (t) of X (t) is

Z(t)=X(t)+jY(t)=a(t)e ^jθ(t) （2）

In formula for instantaneous amplitude, θ (t)=arctan (Y (t)/X (t)) is phase place, according to phase value θ (t)

Can instantaneous frequency be obtained further:

ω(t)=d(θ(t))/dt (3）

Amplitude and the phase place of instantaneous frequency can be expressed by formula (3) clearly, and then reflect the instantaneity of data;

Can draw further according to formula (3):

$f\left(t\right)=\frac{1}{2\mathrm{\π}}\mathrm{\ω}\left(t\right)=\frac{1}{2\mathrm{\π}}\×\frac{d\left(\mathrm{\θ}\left(t\right)\right)}{\mathrm{dt}}---\left(4\right)$

Can obtain like this:

$X\left(t\right)=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j{\mathrm{\θ}}_i\left(t\right)}=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j{\∫\mathrm{\ω}}_i\mathrm{dt}}---\left(5\right)$

Expansion is called that Hilbert composes above, is denoted as:

$H(\mathrm{\ω},t)=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j\∫{\mathrm{\ω}}_i\mathrm{dt}}---\left(6\right)$

Define marginal spectrum again:

$h\left(\mathrm{\ω}\right)={\∫}_0^{T}H(\mathrm{\ω},t)\mathrm{dt}---\left(7\right)$

In formula (7), T is the length of signal, as can be seen here, what marginal spectrum characterized is [0, T] Cumulate Sum of each frequency content energy in the time, the amplitude that H (ω, t) accurately can describe signal on whole frequency band in time with the rule of conversion of frequency, h (ω) then reflects the amplitude of signal situation of change with frequency on whole frequency band;

Wherein,

Utilize instantaneous frequency time-frequency spectrum, differentiate the CT saturated generation moment, and then identify saturated fault outside transformer district, when there is external area error and causing CT saturated, Output rusults " 0 ", when generating region internal fault, Output rusults " 1 ", and start differential protection;

Compose the instantaneous frequency fluctuation situation reflected according to Hilbert, whether saturated fault outside cog region, detect CT and occur saturated, when detecting that CT occurs saturated, and Output rusults " 0 ", when not detecting that CT occurs saturated, Output rusults " 1 ";

By the Hilbert marginal spectrum that Hilbert obtains time integral, whether saturated fault outside cog region, detect CT and occur saturated, when detecting that CT occurs saturated, and Output rusults " 0 ", when not detecting that CT occurs saturated, Output rusults " 1 ".

Preferably, after described step b, also step c is comprised:

The result that c, the result first Hilbert spectrum judged and Hilbert marginal spectrum judge is carried out or computing; again by through or computing after result and the result that judges of instantaneous frequency time-frequency spectrum carry out and computing, then according to computing after result judge whether to start differential protection:

When through with computing after result be " 0 ", Blocking Differential Protection, when through with computing after result be " 1 ", startup differential protection.

Preferably, realize the described hardware configuration based on the CT saturation detection method of HHT and comprise CPU module, analogue collection module, control output module, communication module and MMI module.

Preferably, described CPU module adopts 32 high-performance digital signal processors.

Advantage of the present invention is:

A kind of CT saturation detection method based on HHT that the present invention addresses; by gathering transformer both sides differential current signal; and after EMD decomposition, Hilbert conversion, obtain instantaneous frequency time-frequency spectrum ω (t), Hilbert spectrum H (ω successively; t), Hilbert marginal spectrum h (ω); complete the detection of saturated fault and CT saturation outside transformer district respectively according to above three kinds of frequency spectrums, and then realize quick-action and the reliability action of differential protection under CT saturated conditions.

As a kind of advantageous embodiment of the present invention; by instantaneous frequency criterion, Hilbert composes criterion and Hilbert marginal spectrum criterion combines with "AND", "or" form; when a kind of judged result is made a fault; another kind can correctly judge; improve the reliability of differential protection; and the multiple situation such as transformer external area error electric current, troubles inside the sample space electric current and evolved fault electric current under taking into full account the saturated and unsaturated two kinds of situations of CT, the method have with better function, efficiency is higher and reliability more advantages of higher.

Accompanying drawing explanation

Fig. 1 is the FB(flow block) of a kind of CT saturation detection method based on HHT in the embodiment of the present invention;

Fig. 2 is the method flow diagram that in Fig. 1, EMD decomposes;

Fig. 3 a is the oscillogram of the embodiment of the present invention at the outer fault current of CT saturated conditions inferior segment;

Fig. 3 b is the instantaneous frequency time-frequency spectrum that in Fig. 3 a, external area error electric current is corresponding;

Fig. 3 c is the oscillogram of embodiment of the present invention troubles inside the sample space electric current under CT saturated conditions;

Fig. 3 d is the instantaneous frequency time-frequency spectrum that in Fig. 3 c, troubles inside the sample space electric current is corresponding;

Fig. 4 a is the Hilbert spectrum that external area error electric current that the embodiment of the present invention utilizes Hilbert to compose to detect under CT saturated conditions is corresponding;

Fig. 4 b is the frequency change figure that in Fig. 4 a, Hilbert spectrum is corresponding;

Fig. 4 c is the Hilbert spectrum that troubles inside the sample space electric current that the embodiment of the present invention utilizes Hilbert to compose to detect under CT saturated conditions is corresponding;

Fig. 4 d is the frequency change figure that in Fig. 4 c, Hilbert spectrum is corresponding;

Fig. 5 a is the embodiment of the present invention Hilbert marginal spectrum that fault current is corresponding outside CT saturated conditions inferior segment;

Fig. 5 b is the Hilbert marginal spectrum that embodiment of the present invention troubles inside the sample space electric current under CT saturated conditions is corresponding;

Fig. 6 is a kind of further decision flow chart of CT saturation detection method on Fig. 1 FB(flow block) basis based on HHT in the embodiment of the present invention;

Fig. 7 is the hardware block diagram adopted based on the CT saturation detection method of HHT a kind of in the embodiment of the present invention.

Embodiment

Shown in composition graphs 1, Fig. 2, HHT algorithm comprises empirical mode decomposition (being called for short EMD) and Hilbert analysis of spectrum (being called for short HAS) two parts.Gather transformer both sides differential current signal and can obtain several intrinsic mode functions meeting certain condition (being called for short IMF component) and a surplus after EMD decomposes.Subsequently Hilbert conversion is carried out to these IMF components, obtain instantaneous frequency, obtain Hilbert spectrum and Hilbert marginal spectrum.

Detailed process is as follows:

IMF component is designated as X (t), and carrying out Hilbert conversion Y (t) to it is:

$Y\left(t\right)=\frac{1}{\mathrm{\π}}{\∫}_{-\∞}^{+\∞}\frac{X\left(\mathrm{\τ}\right)}{t-\mathrm{\τ}}\mathrm{d\τ}---\left(1\right)$

Then the analytic signal Z (t) of X (t) is

Z(t)=X(t)+jY(t)=a(t)e ^jθ(t) （2）

In formula for instantaneous amplitude, θ (t)=arctan (Y (t)/X (t)) is phase place, according to phase value θ (t)

Can instantaneous frequency be obtained further:

ω(t)=d(θ(t))/dt (3）

Amplitude and the phase place of instantaneous frequency can be expressed by formula (3) clearly, and then reflect the instantaneity of data;

Can draw further according to formula (3):

$f\left(t\right)=\frac{1}{2\mathrm{\π}}\mathrm{\ω}\left(t\right)=\frac{1}{2\mathrm{\π}}\×\frac{d\left(\mathrm{\θ}\left(t\right)\right)}{\mathrm{dt}}---\left(4\right)$

Can obtain like this:

$X\left(t\right)=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j{\mathrm{\θ}}_i\left(t\right)}=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j{\∫\mathrm{\ω}}_i\mathrm{dt}}---\left(5\right)$

Expansion is called that Hilbert composes above, is denoted as:

$H(\mathrm{\ω},t)=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j\∫{\mathrm{\ω}}_i\mathrm{dt}}---\left(6\right)$

Define marginal spectrum again:

$h\left(\mathrm{\ω}\right)={\∫}_0^{T}H(\mathrm{\ω},t)\mathrm{dt}---\left(7\right)$

In formula (7), T is the length of signal, as can be seen here, what marginal spectrum characterized is [0, T] Cumulate Sum of each frequency content energy in the time, the amplitude that H (ω, t) accurately can describe signal on whole frequency band in time with the rule of conversion of frequency, h (ω) then reflects the amplitude of signal situation of change with frequency on whole frequency band.

In the once-through operation process of above-mentioned HHT, three judgements can be completed:

(1) owing to having obvious catastrophe point in fault current saturated outside district, and significantly do not suddenly change in troubles inside the sample space electric current, thus utilize instantaneous frequency time-frequency spectrum, the CT saturated generation moment can be differentiated, and then identify saturated fault outside transformer district, when there is external area error and causing CT saturated, Output rusults " 0 ", Blocking Differential Protection, when generating region internal fault, Output rusults " 1 ", and start differential protection;

(2) because fault current harmonic component decay saturated outside transformer district is very fast, instantaneous frequency fluctuates obviously in time, and troubles inside the sample space electric current keeps fundamental frequency constant substantially, thus compose the instantaneous frequency fluctuation situation reflected according to Hilbert, saturated fault outside district can be identified, detect CT and whether occur saturated, when detecting that CT occurs saturated, Output rusults " 0 ", when not detecting that CT occurs saturated, Output rusults " 1 ";

(3) owing to containing a large amount of high order harmonic component in fault current saturated outside transformer district, and troubles inside the sample space electric current is close to sine wave, thus by Hilbert marginal spectrum that Hilbert obtains time integral, can saturated fault outside cog region, whether detect CT occurs saturated, when detecting that CT occurs saturated, and Output rusults " 0 ", when not detecting that CT occurs saturated, Output rusults " 1 ".

Fig. 3 a to Fig. 3 d shows and utilize instantaneous frequency time-frequency spectrum identification transformer fault under CT saturated conditions, from Fig. 3 a, Fig. 3 b, when CT is saturated, when there is external area error, corresponding 4 the instantaneous frequency catastrophe points (minimum point) of differential current in each cycle; And from Fig. 3 c, Fig. 3 d, corresponding 1 the instantaneous frequency catastrophe point of troubles inside the sample space electric current in each cycle, can realize the reliable recognition of external area error and troubles inside the sample space thus.

Fig. 4 a to Fig. 4 d shows and utilize Hilbert spectrum discrimination transformer fault under CT saturated conditions, from Fig. 4 a, Fig. 4 b, under CT saturated conditions, when there is external area error, due to containing a large amount of higher harmonic components in differential current, be reflected in Hilbert spectrum and occur larger frequency fluctuation, frequency change rate is larger; And from Fig. 4 c, Fig. 4 d, when generating region internal fault, differential current is approximately a sine wave, frequency content is single, and therefore Hilbert composes medium frequency and is basically stable at 50Hz, and frequency change rate is less, can realize the identification of external area error and troubles inside the sample space thus.

Fig. 5 a, Fig. 5 b show and utilize Hilbert marginal spectrum identification transformer fault under CT saturated conditions, from Fig. 5 a, under CT saturated conditions, when there is external area error, due in differential current waveform containing a large amount of higher harmonic content, therefore to account for the percentage of current signal lower for fundamental frequency; And from Fig. 5 b, troubles inside the sample space electric current is only containing a small amount of harmonic wave, the percentage that fundamental frequency accounts for current signal is higher, can realize the identification of troubles inside the sample space and external area error thus.

Shown in composition graphs 6; as a kind of advantageous embodiment of the present invention; by instantaneous frequency criterion, Hilbert composes criterion and Hilbert marginal spectrum criterion combines with "AND", "or" form; be exactly specifically; the result that the result first Hilbert spectrum judged and Hilbert marginal spectrum judge is carried out or computing; again by through or computing after result and the result that judges of instantaneous frequency time-frequency spectrum carry out and computing, then according to computing after result judge whether to start differential protection.

When there is external area error and causing CT saturated, instantaneous frequency criterion exports " 0 ", and through Hilbert spectrum and Hilbert marginal spectrum, all can detect that CT is saturated, Output rusults is also " 0 ", pass through or export " 0 " behind the door, differentiate that result is passed through and exported behind the door as " 0 " for twice, therefore differential protection does not perform an action; Otherwise; (the saturated two kind situations unsaturated with CT of CT are comprised) when generating region internal fault; instantaneous frequency correctly can detect troubles inside the sample space; Output rusults " 1 "; and Hilbert spectrum and Hilbert marginal spectrum also correctly can identify troubles inside the sample space; Output rusults " 1 ", twice differentiation is passed through and is exported " 1 " behind the door, starts differential protection.

If due to the impact of other factors such as harmonic wave; there is mistake in the differentiation of Hilbert spectrum or Hilbert marginal spectrum; namely external area error is identified as troubles inside the sample space and exports " 1 " by mistake; then through or door differentiate after Output rusults " 1 "; and instantaneous frequency criterion correctly can detect external area error; Output rusults " 0 ", exports " 0 " after process and door, Blocking Differential Protection.

Therefore; three kinds of criterions can form a complementary relationship; when one differentiates that result is made a fault; another kind can correctly judge; the Reliability of Microprocessor of differential protection can not be affected; and the multiple situation such as transformer external area error electric current, troubles inside the sample space electric current and evolved fault electric current under taking into full account the saturated and unsaturated two kinds of situations of CT, this determination methods have with better function, efficiency is higher and reliability more advantages of higher.

Shown in composition graphs 7; the present invention's application HHT algorithm carries out CT saturation detection and evolved fault identification; amount of calculation is larger; especially when high yardstick, and various theory during embody rule, needs the various situation more complicated considered in engineering; logic judgment unit is more; in order to the requirement of satisfied protection quick-action, accuracy, coordinate the renewal of hardware, this hardware core adopts 32 high-performance digital signal processor TMS320F2812 of TI to come simultaneously.

Whole hardware configuration comprises CPU module, analogue collection module, control output module, communication module and MMI module five part composition, and wherein MMI module comprises input through keyboard submodule and liquid crystal display submodule.

Claims (1)

1., based on a CT saturation detection method of HHT, it is characterized in that comprising the steps:

A, collection transformer both sides differential current signal, obtain the intrinsic mode function that several meet IMF condition, i.e. IMF component after EMD decomposes;

B, carry out Hilbert conversion to above-mentioned IMF component, obtain instantaneous frequency, obtain Hilbert spectrum and Hilbert marginal spectrum, detailed process is as follows:

IMF component is designated as X (t), and carrying out Hilbert conversion Y (t) to it is:

$Y\left(t\right)=\frac{1}{\mathrm{\π}}{\∫}_{-\∞}^{+\∞}\frac{X\left(\mathrm{\τ}\right)}{t-\mathrm{\τ}}\mathrm{d\τ}---\left(1\right)$

Then the analytic signal Z (t) of X (t) is

Z(t)＝X(t)+jY(t)＝a(t)e ^jθ(t) (2)

In formula for instantaneous amplitude, θ (t)=arctan (Y (t)/X (t)) is phase place, can obtain instantaneous frequency further according to phase value θ (t):

ω(t)＝d(θ(t))/dt (3)

Amplitude and the phase place of instantaneous frequency can be expressed by formula (3) clearly, and then reflect the instantaneity of data;

Can draw further according to formula (3):

$f\left(t\right)=\frac{1}{2\mathrm{\π}}\mathrm{\ω}\left(t\right)=\frac{1}{2\mathrm{\π}}\×\frac{d\left(\mathrm{\θ}\left(t\right)\right)}{\mathrm{dt}}---\left(4\right)$

Can obtain like this:

$X\left(t\right)=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j{\mathrm{\θ}}_i\left(t\right)}=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j\∫{\mathrm{\ω}}_i\mathrm{dt}}---\left(5\right)$

Expansion is called that Hilbert composes above, is denoted as:

$H(\mathrm{\ω},t)=\mathrm{RP}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i\left(t\right)e^{j\∫{\mathrm{\ω}}_i\mathrm{dt}}---\left(6\right)$

Define marginal spectrum again:

$h\left(\mathrm{\ω}\right)={\∫}_0^{T}H(\mathrm{\ω},t)\mathrm{dt}---\left(7\right)$

In formula (7), T is the length of signal, as can be seen here, what marginal spectrum characterized is [0, T] Cumulate Sum of each frequency content energy in the time, the amplitude that H (ω, t) accurately can describe signal on whole frequency band in time with the rule of conversion of frequency, h (ω) then reflects the amplitude of signal situation of change with frequency on whole frequency band;

Wherein,

Utilize instantaneous frequency time-frequency spectrum, differentiate the CT saturated generation moment, and then identify saturated fault outside transformer district, when there is external area error and causing CT saturated, Output rusults " 0 ", when generating region internal fault, Output rusults " 1 ", and start differential protection;

Compose the instantaneous frequency fluctuation situation reflected according to Hilbert, whether saturated fault outside cog region, detect CT and occur saturated, when detecting that CT occurs saturated, and Output rusults " 0 ", when not detecting that CT occurs saturated, Output rusults " 1 ";

By the Hilbert marginal spectrum that Hilbert obtains time integral, whether saturated fault outside cog region, detect CT and occur saturated, when detecting that CT occurs saturated, and Output rusults " 0 ", when not detecting that CT occurs saturated, Output rusults " 1 ";

The result that c, the result first Hilbert spectrum judged and Hilbert marginal spectrum judge is carried out or computing; again by through or computing after result and the result that judges of instantaneous frequency time-frequency spectrum carry out and computing, then according to computing after result judge whether to start differential protection:

When through with computing after result be " 0 ", Blocking Differential Protection, when through with computing after result be " 1 ", startup differential protection.