CN105071362B - A kind of distributed feeder automation new protective method applied to FTU - Google Patents

Abstract

The invention discloses a new method for automatic protection of distributed feeders applied to FTU, which accurately judges the fault moment through the zero-sequence current of the feeder; calculates the phase current and zero-sequence current transient fault components; and compares the phase current transient fault components Perform four-level wavelet decomposition to obtain the fault phase; perform four-level wavelet decomposition on the zero-sequence current transient fault component, and judge whether to enter the grounding protection through the energy ratio of the free oscillation component of the zero-sequence current transient fault component; determine the fault phase transient state The ratio of the total energy of the capacitive current to the total energy of the average transient capacitive current of the non-fault phase determines whether to enter the ground protection; through the cross-correlation coefficient of the autocorrelation estimation sequence of the fault phase and the other two non-fault phases, it is judged whether to enter the fault phase ground protection. The invention introduces the shape gradient technology to process the zero-sequence current, extracts the current mutation information and accurately captures the time when the fault occurs, and directly and accurately identifies whether the feeder is faulty based on the three-phase current information collected by the FTU.

Description

A New Method of Distributed Feeder Automatic Protection Applied to FTU

technical field

The invention relates to the field of relay protection of electric power systems, in particular to a new method for automatic protection of distributed feeders applied to FTU. It is suitable for single-phase grounding protection of neutral point non-effective grounding system.

Background technique

Most distribution networks with a voltage level of 66kV and below in my country adopt the non-effective neutral point grounding method, and this system is often called a small current grounding system. In this mode, when a single-phase grounding fault occurs in the power grid, the fault current is small (especially the grounding system through the arc suppressing coil), although the regulations allow continuous operation for 1 to 2 hours, the voltage of the non-grounded phase after the fault becomes the line voltage, which is harmful to the system. Insulation brings hidden dangers and threats. It is necessary to cut off the faulty line as soon as possible to avoid two-point grounding and short-circuit accidents. Ferromagnetic resonance overvoltage may also cause voltage transformer burnout accidents and voltage transformer circuit fuses are frequently blown, which poses a serious threat The safety and reliability of the distribution network. Therefore, when a single-phase grounding fault occurs in a small current grounding system, the faulty line can be detected correctly and in time, and the faulty line can be directly and automatically cut off or the fault can be removed by manual processing by sending a signal. Power supply quality and operation level have important practical significance.

Aiming at the difficulty of single-phase grounding detection in small current grounding systems, a large number of research and development personnel from universities and power companies at home and abroad invested in research, and proposed a variety of protection schemes: zero-sequence current overcurrent protection, insertion resistance method, reactive power directional protection, Harmonic current protection, first half-wave principle, zero-sequence admittance grounding protection, etc. Recently, transient zero-sequence current protection based on signal processing methods such as wavelet transform has been proposed. At the same time, companies at home and abroad have developed line selection devices based on various protection schemes, all of which are based on the centralized comparison of the fault information of each outgoing line (zero-sequence current magnitude and direction, reactive power direction, etc.), and the use of line selection devices in substations Centralized judgment or adding a centralized line selection module in the master station system.

The existing single-phase grounding protection devices for small current systems are still mainly based on line selection devices for steady-state power frequency signals. According to the statistical results of authoritative departments, the overall line selection accuracy of line selection devices in actual operation is less than 70%. Misselection, The probability of rejection is very high, and the line selection of steady-state power frequency signals is affected by the following aspects: (1) the influence of the compensation current of the arc suppression coil; (2) the influence of the transition resistance; (3) the influence of the unbalanced current of the current transformer (4) The influence of the ground fault location.

Because a single-phase grounding will produce a transient fault component that is several to ten times larger than the steady-state component, the small current grounding protection method based on the transient signal will have a good fault identification ability, but the transient signal appears for a short time, The attenuation is fast, and accompanied by high-frequency interference, the detection is relatively difficult, and it is necessary to use appropriate modern signal processing methods for analysis and processing. Currently, wavelet analysis is generally used to process fault transient signals, which is easily affected by impulse noise. Small current ground fault protection based on transient signals is still in the stage of experimentation and promotion.

At present, the detection of single-phase ground faults on feeder lines generally adopts the centralized line selection method, and the faulty line needs to be judged by collectively comparing the fault information of each outgoing line, and further fault location and fault isolation are required. With the development of distribution network automation, the data processing and storage capabilities of feeder intelligent terminals such as FTU have been very powerful, which can meet the requirements of intelligent algorithms; at the same time, some load switches and section switches have been replaced with circuit breakers, which have the ability to break short-circuit current The ability to realize the distributed grounding protection of the small current grounding system can complete the line selection, positioning and isolation at one time, realize the distributed intelligent control of the feeder automation system, realize the rapid isolation of faults, and ensure the safety of the distribution network and the safety of power consumption of users , to better realize distribution network automation.

Contents of the invention

The purpose of the present invention is to overcome the deficiencies in the prior art and provide an on-site distributed single-phase grounding protection method that only needs feeder three-phase current information and can be installed in the FTU.

The object of the invention is achieved through the following technical solutions:

A new method for automatic protection of distributed feeders applied to FTU, comprising the following steps:

Step 1. When the zero-sequence current I0 of the feeder is greater than the set current Iset, record the current time as the fault time, and record the phase current and zero-sequence current of the two cycles before the fault time and the four cycles after the fault time on the feeder;

Step 2. Select a period of zero-sequence current signal for each half period before and after the fault time, and filter and process the gray-scale gradient filter. The time when the first pulse of the filtered zero-sequence current appears is the precise time when the fault occurs;

Step 3. Based on the precise fault occurrence time obtained in step 2, obtain the phase current fault component Δi _p by calculating the fault phase current after the fault occurrence time minus the load phase current before the fault occurrence time; Subtract the unbalanced zero-sequence current before the fault occurrence time from the fault zero-sequence current after the time instant to obtain the zero-sequence current fault component Δi ₀ ; , get phase current transient fault component Δi _p.tr and zero-sequence current transient fault component Δi _0.tr ;

Step 4: Select the phase current transient fault component Δi _p.tr and the zero-sequence current transient fault component Δi _0.tr of the two cycle data after the fault occurrence time, and use the adaptive morphological filter for filtering and denoising preprocessing;

Step 5. Use cubic spline wavelet to decompose the phase current transient fault component Δi _p.tr after filtering and denoising pretreatment with four-layer wavelet, and obtain A, B, and C three-phase current transient fault components in 16 sub-frequency bands respectively energy on beg with choose with The phase corresponding to the maximum value is regarded as the fault phase;

Step 6: Use cubic spline wavelet to decompose the zero-sequence current transient fault component Δi _0.tr after filtering and denoising preprocessing into four layers of wavelet decomposition, and calculate the energy on 16 sub-frequency bands, among which the energy on the (4,0) frequency band The energy is the energy E _ld of the DC attenuation component of the zero-sequence current transient fault component Δi _0.tr , and the sum of the energies in the other frequency bands is the energy of the free oscillation component of the zero-sequence current transient fault component Δi _0.tr E _c.os ;

Find the ratio of E _ld and E _c.os When η>1, it is judged that a single-phase ground fault has occurred in the feeder, and the fault phase-to-ground protection action; when η≤1, go to step 7 for further judgment;

Step 7. According to the 16 frequency band energies of A, B, and C three-phase current transient fault components obtained in step 5 beg When i takes A, B, and C, it corresponds to the total energy of the transient capacitive current of phases A, B, and C According to the fault phase, determine the total energy of the fault phase transient capacitive current as _Ef and the average transient capacitive current total energy of the non-fault phase When η _p >K _e , it is determined that a single-phase ground fault has occurred in the feeder, and the fault phase ground protection action; when η _p ≤ K _e , go to step 8 for further judgment, K _e is the safety factor;

Step 8: Carry out correlation analysis on the waveform-wavelet coefficients of the feeder three-phase current transient fault component Δi _p.tr on the other sub-frequency bands except (4,0) frequency band, first make an automatic analysis of the transient fault components of each phase current Correlation analysis obtains the autocorrelation estimation sequence, and then calculates the cross-correlation coefficient of the autocorrelation estimation sequence of the fault phase and the other two non-fault phase current transient fault components respectively, and obtains the fault phase and the other two non-fault phase current transient fault components the mean of the cross-correlation coefficients of the estimated series of autocorrelations of the components ,when When it is determined that a single-phase ground fault has occurred on the feeder, the fault phase-to-ground protection will act; It is judged that there is no single-phase ground fault on the feeder.

As mentioned above, the adaptive morphological filter in step 4 is:

y(x)=α ₁ (x)Oc(f(x))+α ₂ (x)Co(f(x))

In the formula, α ₁ (x), α ₂ (x) are weight coefficients, f(x) is the original signal to be processed, and the structural elements adopted are incremental structural elements: g=[0.050.10.6], Oc(f(x) ) indicates that the morphological opening and closing operation is performed on f(x) by using the structural element g, and Co(f(x) indicates that the morphological closing and opening operation is performed on f(x) by using the structural element g.

The grayscale morphological gradient filter as described above is:

Among them, g(x) is the structural element, f(x) is the original signal to be processed, means to use the structural element g(x) to perform the open operation on f(x), and (fΘg)(x) means to use the structural element g(x) to perform the closed operation on f(x). Using incremental structuring elements: g = [0.050.10.6].

As mentioned above, K _e takes 1-2; K _M takes 0.5-1.

The advantages of the present invention are:

In the existing non-effectively grounded system single-phase ground fault protection schemes, centralized line selection devices are used, which require fault information of all outgoing lines, which does not meet the requirements of feeder automation. The present invention analyzes the characteristics of fault transient current when single-phase grounding occurs, introduces shape gradient technology to process zero-sequence current, extracts current mutation information, accurately captures the moment of fault occurrence, and introduces self-adaptive shape filtering technology to effectively eliminate white noise and pulse noise , Based on the three-phase current information collected by the FTU, it can directly and accurately identify whether the feeder is faulty, without other feeder fault information, and truly realize the distributed single-phase grounding protection technology that can be applied to the FTU, greatly improving the automation level of the feeder.

Description of drawings

The present invention will be further described by using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention.

Figure 1 Equivalent wiring diagram of the compensation network when a single-phase ground fault occurs;

Figure 2 implementation flow chart;

Figure 3 uses the morphological gradient MMG to identify the moment of fault occurrence; the dotted line in the figure is the zero-sequence current waveform, and the dotted solid line is the waveform after morphological gradient filtering.

Figure 4. Waveform and original waveform after adaptive morphological filtering; (a) is the zero-sequence current waveform of the feeder polluted by white noise and impulse noise, and (b) is the waveform after adaptive morphological filtering.

Figure 5. Fault transient current of feeders 1 and 2 (feeder 1A phase ground fault); (a) is the ABC phase transient current waveform of feeder 1, and (b) is the ABC phase transient current waveform of feeder 2.

Fig.6 The autocorrelation curve of the transient capacitive current of the three phases of feeder 1; the ground fault of feeder 1A phase, (a) is the autocorrelation curve of the A phase transient capacitive current of feeder 1, (b) is the autocorrelation curve of feeder 1 The autocorrelation curve of B-phase transient capacitive current, (c) is the autocorrelation curve of C-phase transient capacitive current of feeder 1.

Fig. 7 The autocorrelation curve of the transient capacitive current of the three-phase feeder 2. (a) is the autocorrelation curve of A-phase transient capacitive current of feeder 1, (b) is the autocorrelation curve of B-phase transient capacitive current of feeder 1, (c) is the C-phase transient capacitance of feeder 1 Autocorrelation curve of sex current.

detailed description

The present invention will be described in detail below by means of embodiments in conjunction with the accompanying drawings.

The technical solution of the present invention introduces a morphological filter, pre-identification of system operating conditions, and a fault location method based on the fault information of the feeder itself, thereby solving the limitations and shortcomings of the existing solutions, and realizing that it can be directly applied to FTU intelligent terminals Distributed feeder single-phase ground protection.

In order to realize the distributed feeder single-phase grounding protection, the method needs to adopt the following processing flow:

Step 1. Use the zero-sequence current as the starting element. When the zero-sequence current I0 of the feeder is greater than the set current Iset, record the current time as the fault time, start the protection algorithm and record the two cycles before the fault time on the feeder and the four cycles after the fault time. Phase current (including A, B, C three phases) and zero sequence current of a cycle.

Step 2: Select a zero-sequence current signal of half a cycle before and after the fault time, and filter it with a gray-scale gradient filter. The time when the first pulse of the filtered zero-sequence current is the precise time when the fault occurs. The grayscale morphological gradient filter G _grad (f) used is:

In the above formula, g(x) is the structural element, f(x) is the original signal to be processed, means to use the structural element g(x) to perform the open operation on f(x), and (fΘg)(x) means to use the structural element g(x) to perform the closed operation on f(x). Use increasing structural elements: g = [0.05 0.1 0.6].

Step 3. Based on the precise fault occurrence time obtained in step 2, calculate the phase current fault component Δi _p and zero-sequence current fault component Δi ₀ according to the following formula; that is, by calculating the fault phase current after the fault occurrence time minus The phase current fault component Δi _p is obtained from the load phase current before the fault occurrence time; the zero-sequence current fault component Δi ₀ is obtained by subtracting the unbalanced zero-sequence current before the fault occurrence time from the fault zero-sequence current after the fault occurrence time;

Δi=i _back -i _front

In the formula, _after i is the fault current after the fault, and before i is the load current _before the fault. Then a band-stop filter is used to filter out the steady-state components of the fault components of the phase current and zero-sequence current, and the phase current transient fault component Δi _p.tr and the zero-sequence current transient fault component Δi _0.tr are obtained.

Step 4: Select two cycle data after the fault occurrence time in the phase current transient fault component Δi _p.tr and the zero-sequence current transient fault component Δi _0.tr , and use an adaptive morphological filter for filtering and denoising preprocessing. The adaptive morphological filter y(x) used is:

y(x)=α ₁ (x)Oc(f(x))+α ₂ (x)Co(f(x))

In the formula, α ₁ (x), α ₂ (x) are weight coefficients, which are obtained by RLS (least square method) adaptive iteration, f(x) is the original signal to be processed, and the structural elements used are incremental structural elements : g=[0.05 0.1 0.6], Oc(f(x)) means to use structural element g to perform morphological opening and closing operations on f(x), Co(f(x) means to use structural element g to perform morphological operations on f(x) Close and open operations.

Step 5. Use cubic spline wavelet to decompose the phase current transient fault component Δi _p.tr after filtering and denoising pretreatment with four-layer wavelet, and obtain A, B, and C three-phase current transient fault components in 16 sub-frequency bands respectively energy on beg with choose with The phase corresponding to the maximum value is regarded as the fault phase.

Step 6: Use cubic spline wavelet to decompose the zero-sequence current transient fault component Δi _0.tr after filtering and denoising preprocessing into a four-layer wavelet decomposition, the essence of which is to let the signal pass through a set of high- and low-pass conjugated orthogonal filters The device group continuously divides the signal into different frequency bands. Then use the following formula to calculate the energy of the zero-sequence current transient fault component Δi _0.tr in the 16 sub-bands obtained by four-layer wavelet decomposition. The energy in the (4,0) frequency band is the energy (E _ld ) of the DC attenuation component of Δi _0.tr , and the sum of the energies in the other frequency bands is the energy of the free oscillation component of Δi _0.tr (E _{c. os} )

In the formula Coefficients under the (4,k)th sub-band are decomposed by wavelet, k=0,1,...,15.

n is the sequence number of the wavelet coefficient obtained under the (4,k) sub-band.

Find the ratio of E _ld and E _c.os When a single-phase-to-earth fault occurs on the feeder, the zero-sequence current transient fault component Δi _0.tr includes transient inductive current (ie, DC attenuation component E _ld ) and transient capacitive current (ie, free attenuation component E _c.os ) , and when the fault angle is around 0 degrees, E _ld >E _c.os , so when the criterion η>1, it is judged that a single-phase-to-ground fault occurs in the feeder, and the fault phase-to-ground protection operates. If η≤1, go to step 7 for further judgment.

Step 7. According to the 16 frequency band energies of A, B, and C three-phase current transient fault components obtained in step 5 beg When i takes A, B, and C, it corresponds to the total energy of the transient capacitive current of phases A, B, and C Let us assume that the fault phase is phase A, then the total energy of the transient capacitive current of the fault phase is The total energy of the average transient capacitive current of the non-fault phase Find the energy ratio η _p of the faulty phase and the non-faulty phase

When η _p >K _e , it is determined that a single-phase ground fault has occurred in the feeder, and the fault phase ground protection action; when η _p ≤K _e , go to step 8 for further judgment. K _e is the safety factor, which is 1.5 here.

Step 8: On the other sub-bands except (4,0) frequency band, conduct correlation analysis on the wavelet coefficient of the three-phase current transient fault component Δi _p.tr of the feeder, and first make the transient fault component of each phase current The autocorrelation analysis obtains the autocorrelation estimation sequence, and then calculates the cross-correlation coefficient of the autocorrelation estimation sequence of the transient fault component of the fault phase and the non-fault phase current, as in step 7, assuming that the fault phase is A phase, then it can be obtained in 15 sub-phases Cross-correlation coefficients over frequency bands j=1,...15, find the mean value of the cross-correlation coefficient with Find the mean again when When it is determined that a single-phase ground fault has occurred on the feeder, the fault phase-to-ground protection will act; It is judged that there is no single-phase ground fault on the feeder. After a large number of simulation experiments to verify and analyze, K _M is recommended to be 0.7.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than limit the protection scope of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that Modifications or equivalent replacements are made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims (4)

1. A new method for distributed feeder automatic protection applied to FTU, is characterized in that, comprises the following steps: Step 1. When the zero-sequence current I0 of the feeder is greater than the set current Iset, record the current time as the fault time, and record the phase current and zero-sequence current of the two cycles before the fault time and the four cycles after the fault time on the feeder; Step 2. Select a period of zero-sequence current signal for each half period before and after the fault time, and filter and process the gray-scale gradient filter. The time when the first pulse of the filtered zero-sequence current appears is the precise time when the fault occurs; Step 3. Based on the precise fault occurrence time obtained in step 2, obtain the phase current fault component △i _p by calculating the fault phase current after the fault occurrence time minus the load phase current before the fault occurrence time; The zero-sequence current fault component △i ₀ is obtained by subtracting the unbalanced zero-sequence current before the fault occurrence time from the fault zero-sequence current after the fault occurrence time; The steady-state component in the fault component, the phase current transient fault component △i _p.tr and the zero-sequence current transient fault component △i _{0.tr are obtained} ; Step 4: Select the phase current transient fault component △i _p.tr and the zero-sequence current transient fault component △i _0.tr of the two cycle data after the fault occurrence time, and use the adaptive morphological filter for filtering and denoising pre-processing deal with; Step 5: Use cubic spline wavelet to decompose the phase current transient fault component △i _p.tr after filtering and denoising pretreatment by four-layer wavelet, and obtain A, B, C three-phase current transient fault components in 16 sub-components respectively. energy in frequency band beg with choose with The phase corresponding to the maximum value is regarded as the fault phase; Step 6: Use cubic spline wavelet to decompose the zero-sequence current transient fault component △i _0.tr after filtering and denoising preprocessing into four layers of wavelet decomposition, and calculate the energy on 16 sub-frequency bands, among which the (4,0) frequency band The energy of the zero-sequence current transient fault component △i _0.tr is the energy E _ld of the DC attenuation component of the zero-sequence current transient fault component △i 0.tr, and the sum of the energy in the other frequency bands is the free oscillation of the zero-sequence current transient fault component △i _0.tr The energy E _c.os of the component; Find the ratio of E _ld and E _c.os When η>1, it is judged that a single-phase ground fault has occurred in the feeder, and the fault phase-to-ground protection action; when η≤1, go to step 7 for further judgment; Step seven, according to the energy on each of the 16 sub-frequency bands of the A, B, and C three-phase current transient fault components obtained in step five beg When i takes A, B, and C, it corresponds to the total energy of the transient capacitive current of phases A, B, and C According to the fault phase, determine the total energy of the fault phase transient capacitive current as _Ef and the average transient capacitive current total energy of the non-fault phase When η _p >K _e , it is determined that a single-phase ground fault occurs in the feeder, and the fault phase ground protection action; when η _p ≤ K _e , go to step 8 for further judgment, and K _e is the safety factor; Step 8. On the other sub-bands except (4,0) frequency band, conduct correlation analysis on the waveform-wavelet coefficient of the three-phase current transient fault component △i _p.tr of the feeder, and first make the transient fault component of each phase current Autocorrelation analysis obtains the autocorrelation estimation sequence, and then calculates the cross-correlation coefficient of the autocorrelation estimation sequence of the fault phase and the other two non-fault phase current transient fault components respectively, and obtains the fault phase and the other two non-fault phase current transients The mean value of the cross-correlation coefficient of the autocorrelation estimation sequence of the failure component when When it is determined that a single-phase ground fault has occurred on the feeder, the fault phase-to-ground protection will act; It is judged that there is no single-phase ground fault on the feeder. 2. a kind of distributed feeder automatic protection new method that is applied to FTU according to claim 1, is characterized in that, in described step 4, adaptive shape filter is: y(x)=α ₁ (x)Oc(f(x))+α ₂ (x)Co(f(x)) In the formula, α ₁ (x), α ₂ (x) are weight coefficients, f(x) is the original signal to be processed, and the structural elements adopted are incremental structural elements: g=[0.05 0.1 0.6], Oc(f(x) ) indicates that the morphological opening and closing operation is performed on f(x) by using the structural element g, and Co(f(x)) indicates that the morphological closing and opening operation is performed on f(x) by using the structural element g. 3. a kind of distributed feeder automatic protection new method that is applied to FTU according to claim 1, it is characterized in that, described grayscale shape gradient filter is: G _grad (f)=(f⊕g)(x)-(fΘg)(x) Among them, g(x) is the structural element, f(x) is the original signal to be processed, (f⊕g)(x) means to use the structural element g(x) to open f(x), (fΘg)(x ) indicates that f(x) is closed by structural element g(x), and the incremental structural element is adopted: g=[0.05 0.1 0.6]. 4. A new method for distributed feeder automatic protection applied to FTU according to claim 1, characterized in that, said K _e is 1-2; K _M is 0.5-1.