CN112881855A - High-voltage direct-current transmission line lightning stroke interference identification method based on generalized S transformation - Google Patents

Abstract

The invention discloses a method for identifying lightning strike interference of high-voltage direct current transmission lines based on generalized S-transformation. First, phase-mode transformation is performed on the positive and negative voltage and current signal fault components of the rectifier side to obtain a 1-mode component; secondly, 3ms after the protection is started The 1-modulus component of the voltage inside is subjected to generalized S-transformation to calculate the low-frequency energy and high-frequency energy; then the ratio of the voltage low-frequency energy to the high-frequency energy is calculated, and according to the result, it is judged whether lightning strike interference or line fault has occurred; finally, if lightning strike interference occurs, protection Return, if a fault occurs, perform generalized S transformation on the 1-mode component of the current 1ms after the protection starts, calculate the ratio of the low-frequency energy to the high-frequency energy of the current, and judge whether a lightning strike fault or an ordinary short-circuit fault has occurred according to the result, and protect the outlet. The invention can accurately identify the lightning strike interference, improve the reliability of the DC line transient protection, further distinguish the lightning strike fault and the common short-circuit fault, and has guiding significance for the lightning protection, operation and maintenance of the line.

Description

Lightning disturbance identification method for HVDC transmission lines based on generalized S transform

technical field

The invention belongs to the field of power systems, relates to the field of relay protection of high-voltage direct current transmission lines, and in particular relates to a method for identifying lightning strike interference of high-voltage direct current transmission lines based on generalized S transform.

Background technique

As an effective way to solve the power transmission, HVDC transmission has the advantages of low line cost, simple structure, large transmission capacity, small loss, and long transmission distance. It is also widely used in grid connection and urban cable power supply. The protection principle based on transients only needs single-ended fault information, and is often used as the main protection of the line, but the high-frequency signal injected when the line is struck by lightning easily affects the protection based on transients. In view of the above problems, fast and reliable anti-lightning interference technology is required to accurately distinguish lightning interference from ordinary short-circuit faults or faults caused by lightning strikes, so that when the line is disturbed by lightning strikes, the protection will not malfunction, and when it suffers from lightning strikes or ordinary faults In the event of a short-circuit fault, the protection does not refuse to act. This technology plays a decisive role in the practical application of ultra-high-speed transient protection of HVDC transmission lines. At the same time, further distinguishing the lightning strike fault and ordinary short-circuit fault of the line can provide important data support for the line lightning protection, and has guiding significance for the operation and maintenance of the high-voltage DC line.

At present, the existing achievements on the identification of lightning strikes are studied from the perspectives of numerical and waveform characteristics in the time domain, energy distribution characteristics in the frequency domain, and artificial intelligence methods. Among them, the extraction of numerical features in the time domain is simple, but its threshold is easily affected by lightning parameters and has the disadvantage of being difficult to set. The extraction of waveform features needs to be based on a long time window, which is difficult to cooperate with ultra-high-speed transient protection. In the frequency domain, methods such as wavelet transform and HHT are usually used, which can effectively extract the difference in the spectral energy distribution of interference and faults. However, the selection of the wavelet basis and the determination of parameters for the former are highly empirical, and the latter has unavoidable modal aliasing. Phenomenon. The artificial intelligence method avoids complex mathematical operations. The purpose of distinguishing interference and faults can be achieved through offline learning of the training set. However, there are few actual examples that can be used as training sets, and it needs to be based on a large number of simulation results. May not be applicable in actual power system.

SUMMARY OF THE INVENTION

In order to solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a method for identifying lightning strike interference of high-voltage direct current transmission lines based on generalized S transform, which can quickly and reliably identify lightning strike disturbances, and can reduce the types of faults and disturbances, lightning strikes, and lightning strikes. The influence of factors such as current parameters, refraction and reflection traveling waves, etc., realizes rapid identification of lightning strike interference within a short data window, and further distinguishes lightning strike faults and ordinary short-circuit faults on this basis.

To achieve the above object, the technical scheme adopted in the present invention is:

A method for identifying lightning strike disturbance of HVDC transmission lines based on generalized S transform, based on short data windows, has good adaptability to different types of faults and disturbances, and is not affected by transition resistance and lightning current parameters, including the following steps:

Step 1: collect the positive and negative voltage signals and current signals of the rectifier side at a preset sampling frequency, and perform phase-mode transformation on the voltage signal and the current signal respectively to obtain the voltage 1-mode component and the current 1-mode component;

Step 2: Set the protection start time as t, and the lightning strike interference identification system automatically records the voltage 1-mode component and the current 1-mode component within [t ms, (t+3) ms] after the protection starts;

Step 3: Calculate the time-frequency matrix E _U1 of the voltage 1-modulus component in the 3ms time window based on the generalized S transform:

Among them, T is the sampling time interval, the unit is ms, N=3/T is the number of sampling points within 3ms, the value range of n, m, and v is [0, N-1], and U ₁ (v) is the voltage 1 modulo Discrete Fourier transform of component u ₁ (nT), where k is the window width adjustment coefficient.

Step 4: Sum the voltage square values of each frequency within a 3ms time window, and calculate the voltage frequency marginal spectrum E _U1 (n):

Step 5: Taking f _th =1kHz as the threshold value of high frequency and low frequency, calculate the low frequency energy E _{LF_U1} and the high frequency energy E _{HF_U1} of the voltage 1 mode component, and calculate the ratio ρ _U1 :

Among them, f _s is the signal sampling frequency.

Step 6: determine whether the ratio ρ _U1 exceeds the threshold ρ _set1 , if it exceeds, it is judged as a fault, and the process goes to step 7;

Step 7: Calculate the time-frequency matrix E _I1 of the current 1-mode component in the 1ms time window based on the generalized S transform:

Among them, T is the sampling time interval, the unit is ms, N=1/T is the number of sampling points within 1ms, the value range of n, m, and v is [0, N-1], and I ₁ (v) is the voltage 1 modulo Discrete Fourier transform of component i ₁ (nT), where k is the window width adjustment coefficient.

Step 8: Sum the current square values of each frequency within a 1ms time window, and calculate the current frequency marginal spectrum E _I1 (f):

Step 9: take f _th =1kHz as the high frequency and low frequency thresholds, calculate the low frequency energy E _{LF_I1} and the high frequency energy E _{HF_I1} of the 1-mode component of the current, and calculate the ratio ρ _I1 :

Among them, f _s is the signal sampling frequency.

Step 10: Determine whether the ratio ρ _I1 exceeds the threshold ρ _set2 , if it exceeds, it is judged as a common short-circuit fault, if not, it is judged as a lightning strike fault, the algorithm ends, and the normal exit is protected.

Preferably, the preset sampling frequency in step 1 is set to 1 MHz.

Preferably, the threshold ρ _set1 described in step 6 is set to 3.4133.

Preferably, the threshold ρ _set2 described in step 10 is set to 10.6361.

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

The identification scheme of the present invention based on the ratio of the low-frequency and high-frequency energy of the voltage and current moduli after phase-to-mode transformation as the criterion has obvious advantages, avoiding the influence of the absolute magnitude of the electrical quantity, so compared with the previous method of extracting numerical features According to the research, the scheme has good adaptability in terms of lightning current parameters and transition resistance. In addition, for interference and various faults, the ratio has an order of magnitude difference, and the requirements for the threshold value are more tolerant, avoiding the problem of difficult tuning. This solution only needs data in a 3ms time window, which satisfies the quickness of protection and has lower requirements on computing speed and storage capacity.

Description of drawings

FIG. 1 is a model diagram of a primary system of HVDC power transmission suitable for the method of the present invention.

Figure 2 is a flow chart for implementing the method of the present invention.

Figure 3(a), Figure 3(b), Figure 3(c), Figure 3(d) are the voltage 1-mode component waveform, the current 1-mode component waveform when a lightning strike fault occurs, the judgment result of whether a fault has occurred, and whether the fault has occurred or not. Discrimination results are caused by lightning strikes.

Fig. 4(a), Fig. 4(b), Fig. 4(c), Fig. 4(d) are the voltage 1-mode component waveform, current 1-mode component waveform when lightning disturbance occurs, the judgment result of whether a fault has occurred, and whether the fault has occurred or not. Discrimination results are caused by lightning strikes.

Fig. 5(a), Fig. 5(b), Fig. 5(c), Fig. 5(d) are respectively the waveform of voltage 1-mode component, current 1-mode component waveform, fault judgment result, failure Whether the judgment result is caused by lightning strike.

Detailed ways

The present invention will be described in further detail below with reference to the accompanying drawings and embodiments.

As shown in Figure 1, the primary system of DC transmission includes models such as DC transmission line, converter transformer, converter valve, AC filter bank and reactive power compensation, DC filter bank and smoothing reactor. This model adopts various parameters of a DC transmission project of ±500kV. This project is an important channel for sending electricity from coal power bases and hydropower in the upper reaches of the Yellow River. Usually, the sending-end converter station is used as the rectifier station, and the receiving-end converter station is used as the reverse power station. Change station operation, the line is 1043 kilometers long.

When the DC transmission line is 450km away from the rectifier side protection installation place and the top of the tower is struck by lightning, the method provided by the invention can eliminate the interference of the lightning strike, and further distinguish the lightning strike fault and the common short-circuit fault. The lightning current waveform is set to 2.6/50μs, the amplitude is -120kA, the insulator flashover, and the line is faulty. The identification scheme includes the following steps, as shown in Figure 2:

Step 1: Collect the positive and negative voltage signals and current signals of the rectifier side at a certain sampling frequency, and perform phase-mode transformation on the voltage signals and current signals respectively to obtain the voltage 1-mode component and the current 1-mode component;

Step 2: Set the protection start time as t, and the lightning strike interference identification system automatically records the voltage 1-mode component and the current 1-mode component within [t ms, (t+3) ms] after the protection starts;

In this example, the sampling frequency is 1MHz, and the phase-to-analog transformation is performed using the following matrix:

It can be derived from:

The fault components of the 1-mode voltage and current waveforms after phase-to-mode conversion are shown in Figure 3(a) and (b), respectively.

Step 3: Calculate the time-frequency matrix E _U1 of the voltage 1-modulus component in the 3ms time window based on the generalized S transform:

Among them, T is the sampling time interval, the unit is ms, N=3/T is the number of sampling points within 3ms, the value range of n, m, and v is [0, N-1], and U ₁ (v) is the voltage 1 modulo Discrete Fourier transform of component u ₁ (nT), where k is the window width adjustment coefficient.

Step 4: Sum the voltage square values of each frequency within a 3ms time window, and calculate the voltage frequency marginal spectrum E _U1 (n):

Step 5: Taking f _th =1kHz as the threshold value of high frequency and low frequency, calculate the low frequency energy E _{LF_U1} and the high frequency energy E _{HF_U1} of the voltage 1 mode component, and calculate the ratio ρ _U1 :

Among them, f _s is the signal sampling frequency.

In this example, after the frequency marginal spectrum is integrated, the low frequency energy E _{LF_U1} =0.594, the high frequency energy E _{HF_U1} =0.0429, and the ratio ρ _U1 =13.8569 of the voltage 1 modulo component can be obtained.

Step 6: judge whether the ratio ρ _U1 exceeds the threshold ρ _set1 , if it exceeds, it is judged to be a fault, and proceed to step 7; if it does not exceed, it is judged to be lightning strike interference, and the protection is reset;

In this example, the threshold ρ _set1 = 3.4133 for judging lightning interference and faults is set. If the ratio ρ _U1 exceeds the threshold, it is judged that a fault has occurred. The judgment result signal is shown in Figure 3(c), where 0 means lightning interference has occurred, and 1 means A fault has occurred, and the next step is to determine whether a lightning strike fault or an ordinary short-circuit fault has occurred.

Step 7: Calculate the time-frequency matrix E _I1 of the current 1-mode component in the 1ms time window based on the generalized S transform:

Among them, T is the sampling time interval, the unit is ms, N=1/T is the number of sampling points within 1ms, the value range of n, m, and v is [0, N-1], and I ₁ (v) is the voltage 1 modulo Discrete Fourier transform of component i ₁ (nT), where k is the window width adjustment coefficient.

Step 8: Sum the current square values of each frequency within a 1ms time window, and calculate the current frequency marginal spectrum E _I1 (f):

Step 9: take f _th =1kHz as the high frequency and low frequency thresholds, calculate the low frequency energy E _{LF_I1} and the high frequency energy E _{HF_I1} of the 1-mode component of the current, and calculate the ratio ρ _I1 :

Among them, f _s is the signal sampling frequency.

In this example, after the frequency marginal spectrum is integrated, the low frequency energy E _{LF_I1} =0.1875, the high frequency energy E _{HF_I1} =0.1369, and the ratio ρ _I1 =1.3692 of the current 1-modulus component can be obtained.

Step 10: Determine whether the ratio ρ _I1 exceeds the threshold ρ _set2 , if it exceeds, it is judged as a common short-circuit fault, if not, it is judged as a lightning strike fault, the algorithm ends, and the normal exit is protected.

In this example, the threshold ρ _set1 = 10.6361 for judging lightning interference and faults is set. If the ratio ρ _I1 does not exceed the threshold, it is judged that a lightning fault has occurred. The judgment result signal is shown in Figure 3(d), where 0 indicates that a common short-circuit fault has occurred. , 1 indicates that a lightning strike fault occurs, the algorithm ends, and the normal exit is protected.

In addition, when the DC line is 450km away from the rectifier side protection installation site, the lightning strikes the top of the tower and does not cause a fault. The waveform of the voltage 1-mode component, the current 1-mode component waveform, the judgment results of whether a fault has occurred, and whether the fault is caused by lightning strikes are shown in Figure 4 ( a), as shown in Figure 4(b), Figure 4(c), and Figure 4(d), it is judged as lightning strike interference, and the protection is reset. When the DC line is 450km away from the rectifier side protection installation, a positive metal grounding short-circuit fault occurs, the voltage 1-mode component waveform, the current 1-mode component waveform, the judgment results of whether the fault occurs, and whether the fault is caused by lightning strikes are shown in Figure 5(a) , Figure 5 (b), Figure 5 (c), Figure 5 (d), it is judged that an ordinary short-circuit fault occurs, and the normal outlet is protected.

Claims (4)

1. a kind of high-voltage direct current transmission line lightning disturbance identification method based on generalized S transformation, it is characterized in that: based on short data window, to different types of faults, good interference adaptability, not affected by transition resistance and lightning current parameters, including the following step: Step 1: collect the positive and negative voltage signals and current signals of the rectifier side at a preset sampling frequency, and perform phase-mode transformation on the voltage signal and the current signal respectively to obtain the voltage 1-mode component and the current 1-mode component; Step 2: Set the protection start time as t, and the lightning strike interference identification system automatically records the voltage 1-mode component and the current 1-mode component within [t ms, (t+3) ms] after the protection starts; Step 3: Calculate the time-frequency matrix E _U1 of the voltage 1-modulus component in the 3ms time window based on the generalized S transform:

Among them, T is the sampling time interval, the unit is ms, N=3/T is the number of sampling points within 3ms, the value range of n, m, and v is [0, N-1], and U ₁ (v) is the voltage 1 modulo Discrete Fourier transform of component u ₁ (nT), k is the window width adjustment coefficient; Step 4: Sum the voltage square values of each frequency within a 3ms time window, and calculate the voltage frequency marginal spectrum E _U1 (n):

Step 5: Taking f _th =1kHz as the threshold value of high frequency and low frequency, calculate the low frequency energy E _{LF_U1} and the high frequency energy E _{HF_U1} of the voltage 1 mode component, and calculate the ratio ρ _U1 :

Among them, f _s is the signal sampling frequency; Step 6: determine whether the ratio ρ _U1 exceeds the threshold ρ _set1 , if it exceeds, it is judged as a fault, and the process goes to step 7; Step 7: Calculate the time-frequency matrix E _I1 of the current 1-mode component in the 1ms time window based on the generalized S transform:

Among them, T is the sampling time interval, the unit is ms, N=1/T is the number of sampling points within 1ms, the value range of n, m, and v is [0, N-1], and I ₁ (v) is the voltage 1 modulo Discrete Fourier transform of component i ₁ (nT), k is the window width adjustment coefficient; Step 8: Sum the current square values of each frequency within a 1ms time window, and calculate the current frequency marginal spectrum E _I1 (f):

Step 9: take f _th =1kHz as the high frequency and low frequency thresholds, calculate the low frequency energy E _{LF_I1} and the high frequency energy E _{HF_I1} of the 1-mode component of the current, and calculate the ratio ρ _I1 :

Among them, f _s is the signal sampling frequency; Step 10: Determine whether the ratio ρ _I1 exceeds the threshold ρ _set2 , if it exceeds, it is judged as a common short-circuit fault, if not, it is judged as a lightning strike fault, the algorithm ends, and the normal exit is protected. 2 . The method for identifying lightning strike interference of HVDC transmission lines based on generalized S transform according to claim 1 , wherein the preset sampling frequency in step 1 is set to 1 MHz. 3 . 3. A method for identifying lightning strike disturbance of HVDC transmission lines based on generalized S transform according to claim 1, wherein the threshold ρ _set1 described in step 6 is set to 3.4133. 4 . The method for identifying lightning strike interference of HVDC transmission lines according to claim 1 , wherein the threshold ρ _set2 described in step 10 is set to 10.6361. 5 .

Images