RU2516371C1 - Method for determination of damaged feeder at earth fault in distributing mains - Google Patents

Abstract

FIELD: electricity.

SUBSTANCE: zero-sequence voltage samples at common buses and zero-phase sequence current samples in each feeder of the distributing mains are fixed with the preset sampling rate. The zero-sequence voltage samples are subject to analogue-to-digital conversion, then the converted voltage is delivered to the input of a zero-sequence model for each feeder at that models are made for the normal condition of the feeders. The zero-phase sequence current is subject to analogue-to-digital conversion for each feeder model with the preset sampling rate. Mismatch between the zero-phase sequence current for each actual feeder and current samples for each model is determined. The damaged feeder is determined against the mismatch value.

EFFECT: improving selectivity.

4 cl, 6 dwg

Description

The invention relates to the electric power industry, and in particular to relay protection of distribution networks, networks with low currents of an established earth fault, where there is a problem of determining a damaged feeder.

A known method for the recognition of emergency situations in the power line, based on the use of its model [1]. This or that emergency is modeled, and the damaged line model contains at least one variable parameter - the coordinate of the location of the alleged damage. The damage criterion helps to estimate the coordinate. In networks with a large earth fault current, the damage criterion is clearly formulated — the transition of the reactive power of the alleged damage through a zero value [2]. In distribution networks, where the closure process is intensive only for a limited time, such a criterion does not exist.

There is a more general method for recognizing emergencies in an electrical system using models of its power lines [3]. It allows you to distinguish between lines, identifying one damaged, but only under the condition of multilateral surveillance of the electrical system, which does not occur in distribution networks.

A known method for determining damaged phases or a damaged feeder based on the allocation of emergency components of electrical quantities [4]. The emergency ones include, among others, the most accessible for registration zero voltage components on the substation tires and zero current components in the feeders departing from it. In modern microprocessor relay protection, readings of electrical quantities are observed. The observed voltage can be included in the model of the electric network as a voltage source, but since the model is analog, then the fixed samples must be subjected to digital-to-analog conversion. Subsequent operations are determined by the accepted criterion for identifying a damaged feeder. In the discussed method [4], the criterion is the direction of energy transfer of emergency components of voltages and currents. This energy, according to theoretical concepts, should go out of the damaged feeder and go in intact. However, an experimental verification of this method in distribution networks revealed a significant drawback of the energy criterion. Energy is determined by the integration of instantaneous power over time, and errors inevitably accumulate due to the error of the measuring transformers.

The aim of the proposed technical solution is to increase the selectivity of the method for determining a damaged feeder during ground fault in the distribution network.

This goal is achieved by the fact that new, realizing the idea of rejecting the criteria of the damaged feeder and the transition to the criterion of an intact state are added to the known signs. The following features of the known technical solution are in demand: fixing (recording) zero-sequence voltage samples on the common buses of the distribution network, as well as fixing the zero-sequence current in each feeder, monitoring the network in discrete time with a given sampling rate; voltage of the zero sequence is subjected to digital-to-analog conversion, after which it is fed to the input of the model of each feeder of the controlled network in zero sequence. Known technical solutions are models of damaged feeders. Here other models are used, namely those that are characteristic of processes in each intact feeder. Each model responds to the effects of a zero-sequence voltage converted to an analog form by a zero-sequence current. Analog model currents undergo analog-to-digital conversion with the same sampling rate as the currents of the observed network. There is a discrepancy between the real and model currents of each feeder. It is determined and serves as a criterion for distinguishing between the damaged and undamaged condition of the feeder.

In additional claims, a mechanism for determining the discrepancy between the two discrete currents is disclosed, and operations specifying the procedure for identifying a damaged feeder are also indicated. It is proposed to form a discrepancy between the currents in the form of the standard deviation of the readings of these currents on the observation interval of the earth fault process, and it is proposed to identify a damaged feeder by comparing the discrepancy between the currents and the setpoint or by comparing the discrepancy between the currents of different feeders with each other.

Figure 1 shows a schematic diagram of a distribution network with two feeders and zero-sequence measuring transformers connected to the network, figure 2 shows a model of one of the feeders in a zero sequence, and figure 3 is an illustration of a procedure for comparing two discrete currents: one observed in real feeder, and the second one observed in the model of this feeder. Figure 4-6 illustrate the theoretical foundations of the proposed method, figure 4 shows a three-feeder network in the phase closure mode of one of the feeders to ground, figure 5 is a model of the same network in zero sequence, where the feeders are represented by single-line circuits, and in Fig. .6 the same model is depicted as a connection of one tri-terminal and three two-terminal.

The distribution network diagram shows common buses 1, feeders 2 and 3. One of the zero sequence current transformers 4, 5 is connected to each feeder, and the zero sequence voltage transformer is connected to common buses 6. The observed values of the zero sequence are fed to analog-to-digital converters 7 -9, issuing discrete currents i _{0, d} (k), i _{0, d + 1} (k) and discrete voltage u ₀ (k), where k is the discrete time (report number), d is the number of the feeder.

. На воздействие $$\stackrel{⌢}{u}{}_0(t)$$

модель откликается аналоговым током $$\stackrel{⌢}{i}{}_{\mathrm{0,}d}(t)$$

, который через трансформатор тока 12 подается на аналого-цифровой преобразователь 13. Последний работает с такой же частотой дискретизации, как и сетевые аналого-цифровые преобразователи 7-9, поэтому выходной сигнал $$\stackrel{⌢}{i}{}_{\mathrm{0,}d}(k)$$

модельного преобразователя 13 правомерно рассматривать в том же дискретном времени k, что и наблюдаемый ток d-го фидера i_0,d(k). Расхождение между током объекта и током его модели определяется компаратором 14, выделяющим среднеквадратическое отклонение двух выборок, каждая из которых состоит из n отсчетов своего тока:The zero sequence model of the undamaged d-th feeder is presented in the form of a chain structure 10 with longitudinal resistive inductive elements and transverse capacitors. The discrete voltage u ₀ (k) is supplied to the model through a digital-to-analog converter 11, which converts u ₀ (k) into an analog quantity $$\stackrel{⌢}{u}{}_0(t)$$

. On impact $$\stackrel{⌢}{u}{}_0(t)$$

the model responds with analog current $$\stackrel{⌢}{i}{}_{0d}(t)$$

, which is fed through a current transformer 12 to an analog-to-digital converter 13. The latter operates at the same sampling rate as the network analog-to-digital converters 7-9, so the output signal $$\stackrel{⌢}{i}{}_{0\mathrm{<figure-callout\; id="13"\; label="d\; model\; converter"\; filenames="00000012.png"\; state="\{\{state\}\}">d\; </figure-callout>}}{}^{}(k)$$

model converter 13 can rightly be considered in the same discrete time k as the observed current of the d-th feeder i _{0, d} (k). The discrepancy between the current of the object and the current of its model is determined by the comparator 14, which distinguishes the standard deviation of two samples, each of which consists of n samples of its current:

$$\Delta I{}_{0d}=\sqrt{\frac{\mathrm{one}}{n}{\displaystyle \sum _{k=\mathrm{one}}^n\left[i{}_{0d}(k)-\stackrel{⌢}{i}{}_{0d}(k)\right]}{}^2},$$

where ΔI _{0, d} is the output signal of the comparator 14.

The theoretical foundations of the proposed method are illustrated by the example of a three-feeder network 15-17, which in the earth fault mode has a single-line model 18-20 in zero sequence, and that in turn is implemented by a structure with one three-terminal 21 and three two-terminal 22-24.

In the normal mode of operation of the distribution network, the levels of all values of the zero sequence are assumed to be negligible, otherwise the process preceding the closure must be extrapolated for a time after the closure and removed from the current process of earth fault. After a circuit that occurred in one of the feeders, a zero-sequence voltage appears on the common buses 1 and zero-sequence currents in all feeders 2, 3, both in the damaged and intact ones. The values of the zero sequence are allocated by current transformers 4, 5 and voltage 6 and converted by analog-to-digital converters 7-9 into discrete signals i _{0, d} (k), i _{0, d + 1} (k), u ₀ (k)

подается на входы моделей всех фидеров. На фиг.2 показана только одна модель. Двойное преобразование напряжения нулевой последовательности необходимо и неизбежно, так как регистрация всех величин происходит в цифровой форме и именно так сохраняется в памяти микропроцессорного терминала, между тем как модель фидера функционирует в непрерывном времени. Выходное напряжение $$\stackrel{⌢}{u}{}_0(t)$$

преобразователя 11 подается на входы моделей всех фидеров распределительной сети, как это показано на фиг.2 применительно к модели 10 одного d-го фидера. Сигнал $$\stackrel{⌢}{u}{}_0(t)$$

в нормальном режиме работы сети находится на нулевом уровне. После замыкания на землю, случившегося в каком-нибудь, например в (d+1)-м фидере 3, уровень этого напряжения скачкообразно нарастает. В модели 10 неповрежденного d-го фидера напряжение $$\stackrel{⌢}{u}{}_0(t)$$

вызовет реакцию в виде аналогового тока $$\stackrel{⌢}{i}{}_{\mathrm{0,}d}(t)$$

. После его передачи трансформатором тока 12 и аналогово-цифровым преобразователем 13 создается дискретный ток $$\stackrel{⌢}{i}{}_{\mathrm{0,}d}(k)$$

, близкий к наблюдаемому i_0,d(k) неповрежденного фидера 2. Подобная ситуация сложится и в моделях всех прочих неповрежденных фидеров. Исключение составит лишь один поврежденный (d+1)-й фидер 3. Модель этого фидера, как и всех остальных, построена в предположении о его неповрежденном состоянии, вследствие чего ток модели $$\stackrel{⌢}{i}{}_{\mathrm{0,}d+1}(k)$$

будет далек от реального тока i_0,d+1(k). Соответственно, на выходе компаратора 14 неповрежденного d-го фидера появится сигнал низкого уровня ΔI_0,d а на выходе аналогичного компаратора поврежденного (d+1)-гo фидера сигнал ΔI_0,d+1 достигнет высокого уровня.The voltage u ₀ (k) is returned to analog form by a digital-to-analog converter 11, and its output signal $$\stackrel{⌢}{u}{}_0(t)$$

fed to the inputs of models of all feeders. Figure 2 shows only one model. Double conversion of the zero sequence voltage is necessary and inevitable, since the registration of all quantities occurs in digital form and this is how it is stored in the memory of the microprocessor terminal, while the feeder model operates in continuous time. Output voltage $$\stackrel{⌢}{u}{}_0(t)$$

the converter 11 is fed to the inputs of the models of all feeders of the distribution network, as shown in figure 2 with reference to model 10 of one d-th feeder. Signal $$\stackrel{⌢}{u}{}_0(t)$$

in normal mode, the network is at zero level. After an earth fault that happened in some kind of, for example, in the (d + 1) th feeder 3, the level of this voltage increases stepwise. In model 10, the intact d-th feeder voltage $$\stackrel{⌢}{u}{}_0(t)$$

will cause an analog current reaction $$\stackrel{⌢}{i}{}_{0d}(t)$$

. After its transfer by a current transformer 12 and an analog-to-digital converter 13, a discrete current is created $$\stackrel{⌢}{i}{}_{0d}(k)$$

, close to the observed i _{0, d} (k) of intact feeder 2. A similar situation will develop in the models of all other intact feeders. The exception is only one damaged (d + 1) th feeder 3. The model of this feeder, like all the others, is built on the assumption of its intact state, as a result of which the current of the model $$\stackrel{⌢}{i}{}_{0d+\mathrm{one}}(k)$$

will be far from the real current i _{0, d + 1} (k). Accordingly, at the output of the comparator 14 of the undamaged d-th feeder, a low level signal ΔI _{0, d} will appear and at the output of a similar comparator of the damaged (d + 1) -th feeder, the signal ΔI _{0, d + 1} will reach a high level.

The output signals of the comparators are compared with a setting that excludes the non-selective behavior of the observer due to some inadequacy of the model to the real object.

It is possible to protect all feeders leaving the substation at one terminal or at autonomous terminals for each feeder, but with the exchange of signals between them. In this case, the damaged feeder detects the operation of comparing the signals ΔI _{0, d of} different feeders with each other. The largest signal belongs to the damaged feeder.

The proposed method is based on the theoretical position explained in Figs. 4-6 and consisting in the fact that in the model of an intact feeder, the zero sequence current is completely determined by the fixed zero sequence voltage on the common buses u ₀ (t). Assume that a three-feeder circuit 15-17 has a ground fault in the third feeder 17, and the other two feeders 15.16 are not damaged. Feeders have distributed inductance and ground capacitance. These are zero sequence parameters. The mutual inductance between the feeders is assumed to be negligible. The network model in the zero sequence is a three-wire structure 18-20 (figure 5). Since the feeder loads are not connected to the ground, a single-wire model of the feeder in zero sequence does not bear the load. The observed zero sequence process is created by a 3i _0f current source operating at the fault location. With common tires it connects the three-terminal 21 - the left section of the feeder (Fig.6). The undamaged feeders 15, 16 represented by models 18, 19 with respect to the common buses are two-terminal 22, 23, as well as the right side 24 of the damaged feeder 20 with respect to the source 3i _0f . The structure of FIG. 6 fully explains the legitimacy of determining the currents 3i ₀₁ and 3i _{02 of} undamaged feeders 15 and 16 from the common voltage on the tires 3u ₀ for them.

The proposed method for detecting a feeder damaged by a ground fault has circumvented the problem of selecting damage criteria, which is especially acute in distribution networks where intermittent arcs occur and arc nonlinearity is sharply manifested. A similar approach, replacing the targeted search for the damaged part of the system by monitoring the health of its individual parts, became possible due to the fact that the voltage and current at the input of each undamaged feeder are connected by the laws inherent in the autonomous model of only one of it.

Information sources

1. RF patent No. 2033622, G01R31 / 11, H2N 3/28, 1989.

2. Lyamets Yu.Ya., Ilyin V.A., Podshivalin N.V. The software package for the analysis of emergency processes and determining the location of damage to the power line. -Electricity, 1996, No. 12, p.2-7.

3. RF patent No. 2033623, G01R 31/11, Н02НЗ / 28.1989.

4. RF patent №2050660, Н02Н 3/38, Н02Н 3/26, Н02Н 7/26, 1992.

Claims (4)

1. A method for determining a damaged feeder during an earth fault in a distribution network by fixing, at a given sampling rate, zero sequence voltage samples on common buses and zero sequence current samples in each distribution network feeder, digital-to-analog conversion of zero sequence voltage samples and supplying the converted voltage to the input of the model of each feeder in the zero sequence, characterized in that the said models make up for normal Feeder-being, perform analog-to-digital conversion of the residual current model each feeder at a predetermined sampling frequency determined difference between the zero sequence current samples each real current feeder and counts the model and largest discrepancies detected faulty feeder. 2. The method according to claim 1, characterized in that the divergence of currents is defined as the standard deviation of their readings in the interval of fixing the process of earth fault in the distribution network. 3. The method according to claim 1, characterized in that the damaged feeder is detected by comparing the magnitude of the divergence of currents with a given setting. 4. The method according to claim 1, characterized in that the damaged feeder is detected by comparing with each other the values of the divergence of currents in different feeders and determining the feeder with the greatest discrepancy.

Images