RU2568680C1 - Method for determining location of feeder short-circuiting with two-way observation - Google Patents

Abstract

FIELD: electricity.

SUBSTANCE: according to the method data components of monitored current and voltage at the feeder ends are recorded and used as input voltage and the first input current values of the feeder model. At that respective voltage values are supplied to inputs of the undamaged feeder model, second input currents are defined as the model reaction to applied voltage, the third currents are defined as difference of the respective first and second currents, levels of the third currents and degree of their identity are controlled at opposite ends of the model and in case of zero level for the third current at one of inputs the conclusion is made on short-circuiting at the other feeder input. In case of the third currents identity the conclusion is made on short-circuiting in the feeder middle, and if the third current at one input exceeds the level of the third current at the other input, then conclusion is made on short-circuiting in a half of the feeder with bigger current value. Both inputs of the model are shunted, the model is divided into submodels of damaged and undamaged halves of the feeder, the third current of the respective shunted input of the model is accepted as the first current of the damaged feeder submodel and the first current and voltage of the other input are generated in the undamaged feeder submodel out of the third current of its shunted input. In the damaged half feeder submodel with one shunted input the same operations are repeated to determine the second and third currents, to control level of third currents and degree of their identity, to determine damaged half of the model, which have been performed earlier for the initial feeder model; the above operations are repeated till identity in the third input currents is detected for the feeder submodel, which points to short-circuiting in the middle of the modelled feeder section, zero level is detected at one of the third input currents, which points to short-circuiting at the other input.

EFFECT: improved accuracy in detecting location of short-circuiting.

2 cl, 4 dwg

Description

The invention relates to the electric power industry and electrical engineering, namely to relay protection and automation of electrical systems, especially distribution electric networks with nominal voltages of 6, 10, 20, 35 kV. In these networks operating in the mode with isolated or compensated neutral, the problem of location of a single-phase circuit is acute. Damage of this kind is accompanied by an intense, but very short-lived, transient process. In the steady state, information about the place of damage is carried by the main harmonic of the zero sequence current, but its level is insignificant, and the zero sequence current is strongly distorted by noise.

The invention is aimed at solving the problem of determining the location of a single-phase circuit in the feeder - a power line in the distribution network. It is assumed that the network is equipped with modern means of control, and the feeder is observed synchronously on both sides. Observation consists in recording digital waveforms of currents and voltages. Observations are synchronized on opposite sides of the feeder via satellite communications.

In the classification of methods for determining the location of a power line circuit, a special role is played by the criteria by which damage is searched. Known methods based on the criterion of resistivity of the damage model [1, 2], which is relatively simple to implement in the basis of sinusoidal currents and voltages. These methods are widely used in high and ultra-high voltage networks operating with dead-grounded neutral [3]. Recently, however, in connection with a noticeable increase in the speed of circuit breakers, the problem of isolating the sinusoidal components of the observed values has arisen due to the fact that the oscillograms of currents and short-circuit voltages are recorded over a very limited time interval.

Two-sided observation of a power line has a clear advantage over one-way [4]. Firstly, the accuracy of determining the location of the circuit is increased, since the uncertainty in setting the parameters of the system connected to the unobservable side of the transmission is eliminated. Secondly, it becomes possible to turn to other damage criteria, simpler than the resistance criterion. The simplicity criterion is distinguished by the coincidence criterion at the point of fault closure determined by the results of observation of each of the two sides separately [5].

In the basis of sinusoidal values, the voltage at the end of an undamaged section of the line is determined by elementary transformations of complex signals - complexes of the observed current and voltage. Another thing is the transformation in the time basis of the instantaneous values of the values observed during the registration of a short-term intensive transient process that occurs during a single-phase circuit in the feeder. All known methods for determining the location of power transmission damage, including the most common of them, presented in [6], use the same operation, which is difficult to implement and leads to an error in determining the coordinates of the fault location. This is the operation of converting the current and voltage readings recorded on one side of the feeder to the voltage of the location of the proposed circuit. Closing can be assumed anywhere, i.e. this is an arbitrary place.

The purpose of the present invention is to improve the accuracy of determining the coordinates of the location of the circuit feeder and at the same time simplify the process of its search. This goal is achieved through the discovery of a new set of technical operations, which can be described as a whole as a way to localize the circuit. It is based on a new criterion, incomparably more general than the known ones, and reduced to the following position. In an object model that is symmetrical about the fault location and activated only by the fault current, the currents at the opposite inputs are the same. This criterion has no restrictions; it is valid for any objects, regardless of the nature of the process caused by closure. Of course, a natural question arises about the symmetry of the model relative to an unknown point of closure. To answer it, it is necessary to solve the problem of balancing the model. The present invention is precisely dedicated to solving this problem. An integral part of the symmetrization problem is the separation of the model of the damaged part of the object from its other parts. In the proposed method, localization is achieved by transferring the damaged feeder model to the mode with shunted inputs. A model with balanced inputs remains non-symmetrical with respect to the unknown circuit location. However, the levels of currents observed at the inputs of this model of the currents make it possible to judge which side of the input is shorted to the side of the circuit. Balancing is done by dividing the length of the simulated area in half; intact half at each division is excluded from the model. The degree of non-identity of the observed currents is proposed to be judged by the value of the parameter determined in such a way that it was little affected by noise. A single value of the non-identity parameter indicates that a closure occurred in the middle of a uniform portion of the feeder. A smaller unit indicates that a closure takes place in the first half of the section, a larger unit in the second half.

In FIG. 1 shows a model of a distribution network with a controlled feeder in which an earth fault occurred, in FIG. 2 - illustrations of operations performed in the process of finding the location of the feeder circuit. FIG. 3 illustrates a key operation of locating a fault, and FIG. 4 - search criteria.

Three-phase feeder 1 was selected from the distribution network model. The rest of the network is an active six-terminal 2, the outputs of which are connected to the observed inputs of the feeder A _s , B _s , C _{s of the} left side s and A _r , B _r , C _{r of the} right side r. The arbitrary phase of the feeder is marked by the index ν = A, B, C. Currents i _sv , i _rv and voltages u _sv , u _rv are observed. A single-phase closure occurred at a location with the x _f coordinate to be determined.

The search for the place of damage is carried out in the two-wire model of the feeder, which in FIG. 2a is shown in the form of a circuit 3 with distributed parameters, observed in the mode preceding the closure. The input quantities of the two-wire model are the components of a three-phase system of quantities. These can be zero sequence components.

or non-zero components

In FIG. 2 these components are indicated generically as i _s , u _s on the left input of the model and i _r , u _r - on the right. In the previous mode - with the index "PD".

In FIG. 2b shows a two-wire model of feeder 4 in the current single-phase circuit at an unknown location x _f . The observed values are indicated with the index "tk". In FIG. Figure 2c shows Model 5, which is in a purely emergency mode, derived from the two observed modes - the previous (model 3) and the current (model 4). In model 5, emergency components of the observed quantities operate - the difference between the values of the current mode and the corresponding values of the previous mode extrapolated to the time after closure

, u_ав=u⁽⁰⁾, говорящие о том, что определением аварийных составляющих наблюдаемых величин завершается процедура подготовки информационно важных сигналов. Фиг. 2г - 2л иллюстрируют операции, совершаемые с этими сигналами. Заметим, что модели 4 и 5 идентичны по своей структуре, но различаются токами и напряжениями на входах. В модели 4 величины i_тк, u_тк создаются как неизвестным током замыкания i_f, так и внешними по отношению к фидеру источниками распределительной сети. В модели 5 величины i_ав, u_ав инициируются одним лишь источником тока i_f, а влияние внешних источников на эту модель исключено.where t≥0, t = 0 is the moment of closure, the upper symbol means extrapolation. Usually, the values of the previous regime are periodic, and extrapolation is reduced to their continuation according to the same law. In FIG. 2c, for signals (4), the notation $$i_{\mathrm{but}\mathrm{at}}=i_{\mathrm{one}}^{\left(0\right)}$$

, u _av = u ⁽⁰⁾ , indicating that the determination of emergency components of the observed values completes the procedure for preparing informationally important signals. FIG. 2g - 2l illustrate the operations performed with these signals. Note that models 4 and 5 are identical in structure, but differ in currents and voltages at the inputs. In model 4, the values of i _tk , u _tk are created by both an unknown fault current i _f and external distribution network sources with respect to the feeder. In model 5, the values of _iav , _{uav are} initiated by a single current source i _f , and the influence of external sources on this model is excluded.

и правым $$u_r^{\left(0\right)}$$

. Как следует из фиг. 2в, эти источники представляют собой зарегистрированные по результатам наблюдения фидера аварийные составляющие его напряжений u_s,ав и u_r,ав. Модель 3 откликается на приложенные напряжения вторыми токами $$i_{s2}^{\left(0\right)}$$

и $$i_{r2}^{\left(0\right)}$$

.In FIG. 2d illustrates the operation of activating the passive model 3 of an intact feeder by two voltage sources 6 and 7 - left $$u_s^{\left(0\right)}$$

and right $$u_r^{\left(0\right)}$$

. As follows from FIG. 2c, these sources represent the emergency components of its voltages u _{s, av} and u _{r, av} registered as a result of observation of the feeder. Model 3 responds to applied voltages with second currents $$i_{s2}^{\left(0\right)}$$

and $$i_{r2}^{\left(0\right)}$$

.

In FIG. 2e shows the active model 8 with an unknown current source i _f , which differs from models 4 and 5 in that its inputs are shorted. Short circuits 9 and 10 appear instead of previously operating voltage sources 6 and 7. The action of a single model source i _f is manifested by currents at the inputs

и $$i_{r1}^{\left(0\right)}=i_{s,ав}$$

являются аварийными составляющими наблюдаемых токов i_s и i_r. В свою очередь токи $$i_{s2}^{\left(0\right)}$$

, $$i_{s3}^{\left(0\right)}$$

; $$i_{r2}^{\left(0\right)}$$

, $$i_{r3}^{\left(0\right)}$$

моделей на фиг. 2г и 2д представляют собой разные компоненты аварийных составляющих $$i_{s2}^{\left(0\right)}$$

и $$i_{r1}^{\left(0\right)}$$

.where the superscript is the sequence number of the cycle for converting the input quantities. Toki $$i_{s\mathrm{one}}^{\left(0\right)}=i_{s,\mathrm{but}\mathrm{at}}$$

and $$i_{r\mathrm{one}}^{\left(0\right)}=i_{s,\mathrm{but}\mathrm{at}}$$

are emergency components of the observed currents i _s and i _r . In turn, currents $$i_{s2}^{\left(0\right)}$$

, $$i_{s3}^{\left(0\right)}$$

; $$i_{r2}^{\left(0\right)}$$

, $$i_{r3}^{\left(0\right)}$$

models in FIG. 2d and 2d are different components of emergency components $$i_{s2}^{\left(0\right)}$$

and $$i_{r\mathrm{one}}^{\left(0\right)}$$

.

при нулевом входном напряжении в выходные величины $$i_{s1}^{\left(1\right)}$$

, $$u_s^{\left(1\right)}$$

.FIG. 2e illustrates the operation of dividing the model into two submodels 11 and 12. The undamaged in this case submodel 11 functions as an input current converter $$i_{s3}^{\left(0\right)}$$

at zero input voltage to output quantities $$i_{s\mathrm{one}}^{\left(\mathrm{one}\right)}$$

, $$u_s^{\left(\mathrm{one}\right)}$$

.

.The latter serve as input values for the damaged submodel 12, in which the short circuit 10 with its current is stored $$i_{r\mathrm{one}}^{\left(\mathrm{one}\right)}=i_{r3}^{\left(0\right)}$$

.

14. Реакциями на воздействие 14 являются токи $$i_{s2}^{\left(1\right)}$$

, $$i_{r2}^{\left(1\right)}$$

. На фиг. 2з изображена активная подмодель 15, отличающаяся от подмодели 12 появлением еще одной закоротки 16. Наблюдаемые здесь токи определяются по аналогии с операцией (6)Submodel 12 differs from the original model 5 only in its length and the presence of a short-circuit 10. The illustrations in FIG. 2g, h show the same operations with submodel 12, which were previously performed with model 5 (Fig. 2d, e). In FIG. 2g, submodel 13 is obtained from submodel 12 when the fault current i _{f is} excluded and the voltage source is turned on $$u_s^{\left(\mathrm{one}\right)}$$

14. The responses to exposure 14 are currents $$i_{s2}^{\left(\mathrm{one}\right)}$$

, $$i_{r2}^{\left(\mathrm{one}\right)}$$

. In FIG. 2c shows the active submodel 15, which differs from submodel 12 by the appearance of another short circuit 16. The currents observed here are determined by analogy with operation (6)

при нулевом входном напряжении в выходные сигналы $$i_{r1}^{\left(2\right)}$$

, $$u_r^{\left(2\right)}$$

, которые в свою очередь становятся входными сигналами для подмодели 18 поврежденной части фидера. Именно эта подмодель на фиг. 2к и 2л принимает вид сначала подмодели 19, свободной от повреждения и активизируемой источником напряжения 20, а затем подмодели 21 с закороткой 22 на месте этого источника.FIG. 2i - 2l demonstrate the continuation of the process of localizing the circuit by further crushing the feeder model. FIG. 2i illustrates the next, second division of the feeder in half into the undamaged part 17 and the damaged 18. The short circuits 10 and 16 remain in place, and the division of the submodels 17, 18 takes place at the coordinate 3l / 4, where l is the length of the feeder. Submodel 17 converts input current $$i_{r3}^{\left(\mathrm{one}\right)}$$

at zero input voltage to the output signals $$i_{r\mathrm{one}}^{\left(2\right)}$$

, $$u_r^{\left(2\right)}$$

, which in turn become input signals for submodel 18 of the damaged part of the feeder. It is this submodel in FIG. 2k and 2l takes the form of first a submodel 19, free of damage and activated by a voltage source 20, and then a submodel 21 with a short circuit 22 in place of this source.

The proposed method is based on the regularity explained by the example of the model of FIG. 3a. Suppose that the closure occurred in a homogeneous feeder, the model of which was able to be brought into the mode with shunted inputs at zero initial conditions, while maintaining the unknown current i _f at the closure point x _f unchanged. Let x _s and x _r be the coordinates of the feeder inputs, x _s -x _r = l. Known currents at inputs i _s3 and i _r3 . Regardless of the type of current i _f they carry information about the place of damage, in three cases indisputable. If the current i _{r3 is} zero, then x _f = x _s . If the current level is i _s3 , then x _f = x _r. Finally, if the currents i _s3 and i _{r3 are} identical, then the circuit is located in the middle of the feeder, i.e.

In the event that none of the three conditions listed is fulfilled, the ratio between the currents i _s3 and i _r3 tells you in which half of the feeder the circuit is located.

The controlled object is observed by microprocessor relay protection in discrete time k = 0, 1, ...; the countdown is from the moment of registration of the circuit k = 0. To compare the current levels i _s3 and i _r3, it is necessary to introduce a criterion that can testify to their identity. We use the least squares criterion in two versions

where N is the number of the last observed reference, λ _s and λ _r are the identity parameters determined from (9) and (10) as

If the closure is close to the start of the model of FIG. 3a (x _f → x _s ), then i _r3 → 0, and then

while the parameter λ _r becomes undefined. In the opposite situation, when the closure is close to the end of the model (x _f → x _r ), we see that i _s3 → 0, as a consequence

and the parameter λ _{s is} indefinite. Finally, in the last characteristic case (8), the equality

If none of the conditions (13) - (15) is satisfied, then in the general case it is not possible to determine the coordinate x _f from the ratio of the parameters λ _s and λ _r , however, it can be judged in which of the two halves of the feeder there is a closure . For λ _r >1> λ _s , half with the s-th input is damaged, and for λ _s >1> λ _r - with the r-th input.

In the particular case when, as a result of the transformations, the initial feeder model was reduced to the size of a submodel of a section of length Δl (Fig. 3b), the shunted inputs of the submodel so weaken the influence of the distributed capacitance that the current distribution is determined mainly by the specific inductance L ⁰ and the specific resistance R ^{0 of the} two-wire system. If the coordinate x _{f is} measured from the middle of the section with a length Δl, i.e. the circuit is located at a distance Δl / 2 + x _f from the beginning of the section and ΔΖ / 2-x _f from the end, then the value of x _f is determined by the ratio of the third currents

In the steady-state sinusoidal mode, when the model can be represented in the basis of complex quantities (Fig. 3c), relation (16) takes the form

where I is the current complex, and in the circuit of FIG. 3c Z ⁰ = R ⁰ + jωL ⁰ , ω is the network frequency. The applicability of dependence (17) is controlled by the condition for the phase-matching of the sinusoidal components of the currents i _s3 and i _r3 , in which

and the coordinate x _f determined by formula (17) is a real number.

The criteria for recognizing a fault location or a damaged part of the feeder are presented on a plane with coordinates λ _s , λ _r (Fig. 4), where conditions (13) - (15) are taken into account, as well as those that distinguish faults in different halves of the feeder.

The procedure for determining the location of the feeder by the proposed method begins with the registration of phase currents and voltages i _sv , u _sv ; i _rv , u _rv in two network modes - before and after closure. Informational components are distinguished from phase quantities, for example, components of the zero sequence (1), (2). In the previous mode, the zero sequence is usually absent, and then the zero sequence in the values of the current mode belongs to their emergency components. If, in the previous mode, the presence of distinguished information components is detected, then the operation of separating the emergency components (4), (5) from the composition of phase quantities or from their previously determined components is performed. Model 3 of an intact feeder reproduces the previous mode, which continues after a circuit that has occurred at the facility. Model 4 of the damaged feeder is passive. Sources are located outside it in model 2 of the main part of the network to which the controlled feeder 1 is connected. Finally, model 5 for the information components of currents and voltages is active. Its sole source i _{f is} unknown, as is its location x _f .

, $$u_s^{\left(0\right)}$$

; $$i_{r1}^{\left(0\right)}$$

, $$u_r^{\left(0\right)}$$

служат исходными сигналами для всех последующих преобразований. В данном способе присутствуют три типа входных токов. Первые токи создаются источниками сети 2. Вторые токи являются реакциями модели 3 неповрежденного фидера на действие двух источников напряжения, приложенных к входам этой модели. Третьи токи - разности первых и вторых. На начальном этапе преобразований выделенные из фазных величин напряжения $$u_s^{\left(0\right)}$$

и $$u_r^{\left(0\right)}$$

представляются в виде ЭДС 6 и 7, которые подключают к входам модели 3 неповрежденного фидера (фиг. 2г). Фиксируют токи $$i_{s2}^{\left(0\right)}$$

и $$i_{r2}^{\left(0\right)}$$

на входах этой модели как ее реакции на воздействие источников 6 и 7. Далее определяют третьи токи $$i_{s3}^{\left(0\right)}$$

и $$i_{r3}^{\left(0\right)}$$

как разности (6). В информационном плане это наиболее ценные сигналы. Их подвергают информационному анализу как на данном начальном этапе, так и на последующих. Если m - номер этапа, $$i_{s3}^{\left(m\right)}$$

и $$i_{r3}^{\left(m\right)}$$

- третьи токи на этом этапе, то дискретные сигналы $$i_{s3}^{\left(m\right)}\left(t_k\right)$$

и $$i_{r3}^{\left(m\right)}\left(t_k\right)$$

подвергают операциям свертки и квадратирования, определяя промежуточные величиныInput Model 5 Values $$i_{s\mathrm{one}}^{\left(0\right)}$$

, $$u_s^{\left(0\right)}$$

; $$i_{r\mathrm{one}}^{\left(0\right)}$$

, $$u_r^{\left(0\right)}$$

serve as initial signals for all subsequent transformations. In this method, there are three types of input currents. The first currents are created by network sources 2. The second currents are the responses of model 3 of an intact feeder to the action of two voltage sources applied to the inputs of this model. The third currents are the differences of the first and second. At the initial stage of the transformations, the voltages isolated from the phase quantities $$u_s^{\left(0\right)}$$

and $$u_r^{\left(0\right)}$$

are presented in the form of EMFs 6 and 7, which are connected to the inputs of model 3 of an intact feeder (Fig. 2d). Fix currents $$i_{s2}^{\left(0\right)}$$

and $$i_{r2}^{\left(0\right)}$$

at the inputs of this model as its reaction to the effects of sources 6 and 7. Next, determine the third currents $$i_{s3}^{\left(0\right)}$$

and $$i_{r3}^{\left(0\right)}$$

as a difference (6). In terms of information, these are the most valuable signals. They are subjected to information analysis both at this initial stage and at subsequent ones. If m is the stage number, $$i_{s3}^{\left(m\right)}$$

and $$i_{r3}^{\left(m\right)}$$

- third currents at this stage, then discrete signals $$i_{s3}^{\left(m\right)}\left(t_k\right)$$

and $$i_{r3}^{\left(m\right)}\left(t_k\right)$$

subjected to convolution and squaring operations, determining intermediate values

и $${\lambda }_r^{\left(m\right)}$$

как отношения (11), (12), т.е.and further parameters $${\lambda }_s^{\left(m\right)}$$

and $${\lambda }_r^{\left(m\right)}$$

as relations (11), (12), i.e.

By values (22), (23) in accordance with the criteria presented in FIG. 4, fix the circuit at the ends of the feeder or its middle, and if neither one or the other takes place, then establish which of the half of the feeder it is in.

и $${\lambda }_r^{\left(0\right)}$$

показал, что замыкание следует искать во второй половине фидера (фиг. 2д). Тогда выполняют операции перехода к следующему этапу (m=1): разделяют модель 8 на две подмодели 11 и 12 (фиг. 2е). В подмодели И неповрежденной половины фидера известны сигналы на закороченном входе - ток $$i_{s3}^{\left(0\right)}$$

и нулевое напряжение. На выходе этой подмодели они преобразуются в сигналы $$i_{s1}^{\left(1\right)}$$

, $$u_s^{\left(1\right)}$$

, которые позволяют приступить к следующему этапу преобразований, на этот раз в подмодели 12 поврежденной половины фидера, на выходе которой сохраняется шунт 10 с известным током

. Этот этап с первым номером начинается с тестирования подмодели 12, находящейся в неповрежденном состоянии, когда она принимает вид 13, свободный от тока повреждения i_f (фиг. 2ж). Тестирование заключается в подаче на подмодель 13 ЭДС 14, равной зафиксированному в подмодели 11 напряжению $$u_s^{\left(1\right)}$$

. Реакцией подмодели 13 на воздействие ЭДС являются токи на ее зажимах $$i_{s2}^{\left(1\right)}$$

и $$i_{r2}^{\left(1\right)}$$

. Далее на текущем этапе определяют разностные третьи токи

. Найденные третьи токи представляют собой результат распределения неизвестного тока i_f, действующего в подмодели 15 половины фидера, которая отличается от подмодели 13 состоянием левого входа. У подмодели 15 оба входа зашунтированы, как ранее у модели фидера 8. Произошла, следовательно, локализация места замыкания относительно входов моделируемого участка. Как и на начальном этапе преобразований входных величин, на нынешнем этапе третьи токи $$i_{s3}^{\left(1\right)}$$

, $$i_{r3}^{\left(1\right)}$$

используют для формирования оценок $${\lambda }_s^{\left(1\right)}$$

и $${\lambda }_r^{\left(1\right)}$$

и применения к ним критериев распознавания места замыкания. Если бы оказалось, что координата x_f близка к середине моделируемого участка (четверть длины линии от ее конца), то сигналы $${\lambda }_s^{\left(1\right)}$$

и $${\lambda }_r^{\left(1\right)}$$

отвечали бы условию (15) (фиг. 4). Предположим, что условие (15) на данном этапе не выполняется, а имеют место менее информативные условия

,

, говорящие о том, что замыкание находится в первой части моделируемого участка фидера. Эта часть протяженностью в четверть длины фидера начинается в середине фидера и завершается за четверть длины до его конца (фиг. 2и). Предстоит выполнить с подмоделью 18 поврежденной части фидера следующий этап преобразований (m=2). Ток на левом зашунтированном входе подмодели 18 сохраняется таким, каким он был на предыдущем этапе:

, а для определения входных сигналов на правом входе $$i_{r1}^{\left(2\right)}$$

, $$u_r^{\left(2\right)}$$

используют подмодель 17 неповрежденной четверти фидера 17. Ток предшествующего этапа $$i_{r3}^{\left(1\right)}$$

на зашунтированном входе 10 преобразуется подмоделью 17 в выходные сигналы $$i_{r1}^{\left(2\right)}$$

, $$u_r^{\left(2\right)}$$

. Теперь состояние обоих входов подмодели 18 определено, и могут быть выполнены операции данного этапа. Подмодель 19 неповрежденного отрезка фидера в четверть его длины тестируют путем включения ЭДС 20, равной напряжению $$u_r^{\left(2\right)}$$

(фиг. 2к). Шунт на первом входе подмодели 18 сохраняют и при испытании подмодели 19. Определяют токи $$i_{s2}^{\left(2\right)}$$

и $$i_{r2}^{\left(2\right)}$$

на ее входах. Тем самым обнаруживают и третьи токи $$i_{s3}^{\left(2\right)}=i_{s1}^{\left(2\right)}-i_{s2}^{\left(2\right)}$$

, $$i_{r3}^{\left(2\right)}=i_{r1}^{\left(2\right)}-i_{r2}^{\left(2\right)}$$

на данном этапе, когда замыкание локализовано на отрезке фидера между координатами l/2 и 3l/4 (фиг. 2л).Suppose that at the initial stage of transformations (m = 0), the analysis of parameters $${\lambda }_s^{\left(0\right)}$$

and $${\lambda }_r^{\left(0\right)}$$

showed that the closure should be sought in the second half of the feeder (Fig. 2e). Then, operations of transition to the next stage are performed (m = 1): model 8 is divided into two submodels 11 and 12 (Fig. 2e). In the submodel And the undamaged half of the feeder, signals at a shorted input are known - current $$i_{s3}^{\left(0\right)}$$

and zero voltage. At the output of this submodel, they are converted into signals $$i_{s\mathrm{one}}^{\left(\mathrm{one}\right)}$$

, $$u_s^{\left(\mathrm{one}\right)}$$

, which allow us to proceed to the next stage of transformations, this time in submodel 12 of the damaged half of the feeder, at the output of which a shunt 10 with a known current is stored

. This stage with the first number begins with testing the submodel 12, which is in an intact state, when it takes the form 13, free from the damage current i _f (Fig. 2g). Testing consists in supplying to the submodel 13 EMF 14 equal to the voltage recorded in the submodel 11 $$u_s^{\left(\mathrm{one}\right)}$$

. The response of submodel 13 to the influence of EMF are currents at its terminals $$i_{s2}^{\left(\mathrm{one}\right)}$$

and $$i_{r2}^{\left(\mathrm{one}\right)}$$

. Next, at the current stage, the difference third currents are determined

. The found third currents are the result of the distribution of the unknown current i _f acting in the submodel 15 of the half of the feeder, which differs from submodel 13 in the state of the left input. In submodel 15, both inputs are bridged, as previously in the model of feeder 8. Therefore, localization of the circuit location relative to the inputs of the simulated section occurred. As at the initial stage of the transformation of input quantities, at the current stage the third currents $$i_{s3}^{\left(\mathrm{one}\right)}$$

, $$i_{r3}^{\left(\mathrm{one}\right)}$$

used to form ratings $${\lambda }_s^{\left(\mathrm{one}\right)}$$

and $${\lambda }_r^{\left(\mathrm{one}\right)}$$

and applying to them the criteria for recognizing a fault location. If it turned out that the coordinate x _f is close to the middle of the simulated section (a quarter of the length of the line from its end), then the signals $${\lambda }_s^{\left(\mathrm{one}\right)}$$

and $${\lambda }_r^{\left(\mathrm{one}\right)}$$

would satisfy condition (15) (Fig. 4). Suppose that condition (15) is not fulfilled at this stage, but less informative conditions

,

, indicating that the closure is in the first part of the simulated section of the feeder. This quarter-length portion of the feeder begins in the middle of the feeder and ends a quarter of the length before its end (Fig. 2i). The next stage of transformations (m = 2) is to be performed with submodel 18 of the damaged part of the feeder. The current at the left shunted input of the submodel 18 is maintained as it was at the previous stage:

, and to determine the input signals on the right input $$i_{r\mathrm{one}}^{\left(2\right)}$$

, $$u_r^{\left(2\right)}$$

use submodel 17 of the undamaged quarter of feeder 17. Current of the previous stage $$i_{r3}^{\left(\mathrm{one}\right)}$$

at the shunted input 10 is converted by the submodel 17 into output signals $$i_{r\mathrm{one}}^{\left(2\right)}$$

, $$u_r^{\left(2\right)}$$

. Now the state of both inputs of the submodel 18 is determined, and the operations of this stage can be performed. Submodel 19 of the intact feeder segment in a quarter of its length is tested by turning on the EMF 20 equal to the voltage $$u_r^{\left(2\right)}$$

(Fig. 2k). The shunt at the first input of the submodel 18 is also saved when testing the submodel 19. The currents are determined $$i_{s2}^{\left(2\right)}$$

and $$i_{r2}^{\left(2\right)}$$

at her entrances. Thus, third currents are also detected. $$i_{s3}^{\left(2\right)}=i_{s\mathrm{one}}^{\left(2\right)}-i_{s2}^{\left(2\right)}$$

, $$i_{r3}^{\left(2\right)}=i_{r\mathrm{one}}^{\left(2\right)}-i_{r2}^{\left(2\right)}$$

at this stage, when the circuit is localized on the feeder segment between the coordinates l / 2 and 3l / 4 (Fig. 2l).

и $$i_{r3}^{\left(2\right)}$$

в параметры (19)-(23) при m=2 и по критериям, представленным на фиг. 4, проверяют, не произошло ли замыкание посередине моделируемого участка длиной l/4. Если оказывается, что оно не там, то устанавливают, в какой из половин этого участка оно находится. Преобразования моделей участка с их делением пополам выполняют до тех пор, пока замыкание не будет локализовано на малом участке фидера, с достаточной точностью указывающем место повреждения по соотношению третьих входных токов (16).Next, the third currents are converted $$i_{s3}^{\left(2\right)}$$

and $$i_{r3}^{\left(2\right)}$$

to parameters (19) - (23) with m = 2 and according to the criteria presented in FIG. 4, check whether a short circuit has occurred in the middle of the simulated section of length l / 4. If it turns out that it is not there, then establish which of the halves of this section it is in. Transformations of the site models with their division in half are performed until the closure is localized on a small feeder site, which indicates the location of damage with sufficient accuracy by the ratio of the third input currents (16).

As you can see, the proposed method is not associated with any assumptions that reduce the accuracy of determining the location of the circuit. In addition, it consists of a strictly defined set of operations performed with models of feeder sections and consisting in the conversion of currents and voltages observed on two sides of the feeder.

Information sources

1. RF patent No. 2033622, G01R 31/11, H2N 3/28, 1989.

2. RF patent No. 2033623, G01R 31/11, Н02Н 3/28, 1989.

3. 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, S. 2-7.

4. Lyamets Yu.Ya., Voronov P.I. Location of damage to a multi-wire network during two-way observation. - Proceedings of the RAS. Energy, 2013, No. 3, S. 96-107.

5. RF patent No. 2492493, G01R 31/08, 2011.

6. RF patent No. 2492565, Н02Н 3/28, 2012.

As in the prototype, in the proposed method, the information components of the currents and voltages observed on both sides of the feeder are isolated, which makes it possible to use two-wire two-input models of the feeder in the basis of these components. The observed processes are recorded, i.e. save in the memory of the microprocessor protection system and automation. The components of currents and voltages at the ends of the feeder are used in different ways, although both are input values of the model of the feeder. Input voltages are used once, and they cannot be converted. In contrast, input currents are divided into components. And here lies the fundamental difference from the prototype. If we accept that the observation results give the first input currents, then the second currents are determined using the feeder model. To do this, known voltages are fed to the model inputs and determine its response. Third currents are found as the differences between the first and second, therefore, the second and third currents are components of the first current. Third currents carry important information. Their levels, as well as the degree of their identity at the opposite inputs of the model, make it possible to judge the location of the feeder damage. If the third current of one of the inputs has a zero level, then this is enough to state a short circuit at the other input. If the third currents of both inputs are identical, a short circuit is detected in the middle of the feeder. Finally, if there is neither one nor the other, then it is not the exact value of the circuit location that is judged, but where in the feeder it occurred. It is noted that a circuit takes place in that half of the feeder where the level of the third current is higher. This is a general case, and it requires additional operations to determine the location of the damage with the necessary accuracy. The feeder model is transformed: its inputs are shunted and subdivided into submodels of the damaged and undamaged halves. The submodel of the undamaged part is used only to convert the third current of its shunted input into the current and voltage of its other input, which are taken as the first current and voltage of the corresponding input of the submodel of the damaged part. The other input of this submodel is shunted, and the known current in the shunt is taken as the first current of the submodel. Thus, the submodel of the damaged half of the feeder is prepared for the next conversion with the same set of operations as previously with respect to the original model. The converted submodel saves the shunted input, and in this sense is simpler than the original model, since one of its input voltages is known to be zero.

In the submodel of the damaged half of the feeder, the components of the first currents are again determined - the second and third currents, they control the levels of the third currents and check the degree of their identity. Now, just as in operations with the original model, it is checked whether the place of damage is located in the middle of the damaged half of the feeder, and if this is not so, the damaged part is again determined, this time not only of the feeder, but only its half. This sequence of model conversion operations is repeated until they are convinced of the identity of the third input currents or in the zero level of one of them, which indicates the location of the feeder circuit - the middle of its damaged section or one of its ends.

It is proposed to judge the degree of identity of the third currents by the values of the parameters determined by the readings of each current:

Here s and r are the indices of the beginning and end of the damaged section of the feeder, t _k are the discrete moments of determining the samples of currents i _s3 (t _k ), i _r3 (t _k ), N + 1 is the number of samples on the window of the feeder observation. The length of the damaged area is l / 2 ^m , where l is the length of the feeder, m is the number of divisions of the damaged area in half. If λ _s → 0, and the value of λ _{r is} not determined, a short circuit is detected at the s-th input of the section. If λ _r → 0, and the value of λ _{s is} not determined, then a circuit is detected at the rth input. If λ _s → 1 and λ _r → 1, a short circuit is detected in the middle of the section. For λ _s <1, λ _r > 1, a short circuit is detected on the s-th half of the section to which its s-th input belongs, and for λ _s > 1, λ _r <1 - on the r-th half.

Claims (2)

1. A method for determining the location of the feeder circuit during the two-sided observation of currents and voltages by extracting information components of the observed values and using a two-wire two-input model of the feeder in the basis of these components by registering the information components of the observed currents and voltages at the ends of the feeder and using these components as input voltages and first input currents of the feeder model, characterized in that the inputs of the intact feeder model are supplied with the corresponding voltage, determine the second input currents as the reaction of the model to the applied voltages, determine the third currents as the differences of the corresponding first and second currents, control the levels of the third currents and the degree of their identity at the opposite inputs of the model, and in the case of a zero level of the third current of one of the inputs, detect a short circuit at the other input feeder, in the case of the identity of the third currents, a circuit is detected in the middle of the feeder, and if the third current exceeds one of the inputs of the third current level of the other input, circuit in the feeder half with a large current, both inputs of the model are shunted, the model is divided into submodels of the damaged and undamaged half of the feeder, the third current of the corresponding shunted input of the model is taken as the first current of the model of the damaged half of the feeder, and the first current and voltage of the other input of this submodel are formed into submodels of the intact half of the feeder from the third current of its shunted input, repeat the same operations in the submodels of the damaged half of the feeder with one shunted input the second and third currents, monitoring the level of third currents and the degree of their identity, determining the damaged half of the model that were previously performed in the original model of the feeder, and repeat the above operations until they find the identity of the third input currents of the submodel of the feeder, indicating the closure in the middle of the simulated section of the feeder, or the zero level of one of the third input currents, indicating a circuit at the other input. 2. The method according to p. 1, characterized in that the degree of identity of the third currents is judged by the values of the parameters

Where

s and r are the indices of the ends of the model of the feeder or the ends of the model of the damaged section of the feeder l / 2 ^m long, where l is the length of the feeder, m is the number of divisions of the damaged section in half, t _k are the discrete time instants, i _s3 (t _k ), i _r3 (t _k ) are the readings of the third currents, N + 1 is the number of samples on the observation window, for a zero value of λ _s and an indefinite value of λ _{r, a} short circuit at the s-th input is detected, at zero λ _r and an indefinite λ _{s, a} short circuit at r- at the input level, at a single level, both λ _s and λ _r , a short circuit is detected in the middle of the simulated section, with λ _s <1, λ _r > 1 Tattoo circuit in the s-th half of the simulated area, and for λ _r <1, λ _s > 1 - in the r-th half.

Images