JP7304783B2 - Power system monitoring apparatus and method - Google Patents

Description

The present invention relates to a power system monitoring apparatus and method for locating a fault point in a power system.

A power system is a large-scale system constructed and operated by combining many devices and control methods for the stable supply of power. It is known that the state of an electric power system changes due to various factors, and is represented by physical quantities such as voltage, current, power, and frequency. The system state has a wide area created by the system configuration and is accompanied by temporal changes. It also varies greatly depending on the power generation and load connected to the grid. Since the energy supply should be stable, various control devices and control methods are utilized to maintain steady state.

When an accident that disrupts power supply occurs, it is essential to quickly identify the point of the accident (where the accident occurred) and start restoration work in order to increase the reliability of power supply. . In order to realize prompt recovery from the accident, it is necessary to quickly grasp the details of the accident. However, in the current situation, it is often the case that the accident point is specified by checking work by a person at the site where the accident occurred. Therefore, in addition to the time required to arrive at the site, there is work time for the person in charge to identify the accident point, and personal circumstances such as experience, knowledge, and physical condition are involved here.

On the other hand, in recent years, with the development of networks and digital signal processing, it has become possible to use signals measured by sensors installed in electric power systems. Many techniques for estimating the system state using sensor measurement signals have been proposed. A technique for estimating the fault point is called fault point location, and fault point location methods such as the surge method and the frequency method have been proposed.

Among these methods, the surge method locates the accident point using the arrival time of the surge propagating on the line. Surges are classified as traveling waves and are known to propagate at near-light velocities and to reflect at points of impedance mismatch. There is an accident point locating method for locating the accident point using digital data of the measured waveform obtained by sampling the waveform of this surge with an AD converter.

Patent Literature 1 provides a method for locating an accident point with high accuracy even when a surge generated at the time of an accident attenuates as it propagates through a line.

JP 2012-108011 A

The technique described in Patent Literature 1 is based on measuring the arrival times of surges that occur and propagate at a fault point in a power system at a plurality of points on a railroad track and calculate the difference in arrival times. The principle is to calculate the accident point by creating a relational expression using the distance from the accident point to the measurement point, the surge arrival time difference, and the surge propagation speed. However, there are the following problems for practical use.

Power system measurement signals may have distorted waveforms due to noise, harmonics, line-to-line crosstalk, etc., and it may be difficult to accurately detect the arrival time of a surge.

In order to measure surge arrival time differences at multiple locations, it is necessary to time-synchronize the measuring devices at each location. However, the time accuracy of, for example, GPS (Global Positioning System) is often on the order of microseconds, and the cost of equipment for improving the time accuracy increases.

Surge propagation speed is necessary to calculate the fault point using the arrival time difference, but it is known that the surge propagation speed changes depending on conditions such as line type. If a propagation velocity measured in advance experiments or the like is used, an error may occur in the measurement result due to changes in the propagation velocity. Alternatively, if a means for measuring the surge propagation speed of the line of the target system is used, the equipment cost will increase.

It is known that a surge propagating through a line is reflected at an impedance mismatch and travels in the opposite direction. Reflected waves are unavoidable because the points of reflection are branching points of railway lines and devices such as insulators. Therefore, the measured waveform includes a direct wave that propagates directly from the accident point and a reflected wave that propagates after being reflected at the reflection point. When detecting the arrival time of the waveform, the reflected wave may become noise and lower the detection accuracy.

In view of the above, it is an object of the present invention to provide an electric power system monitoring apparatus and method capable of achieving highly accurate location with a simple apparatus configuration.

In view of the above, the present invention provides a power system monitoring apparatus for locating fault points using surges in the power system, wherein first means for obtaining a measured waveform by measuring a surge waveform propagating in the power system; A second means for obtaining a simulated waveform by generating a simulated surge waveform propagating through the power system based on the system configuration information and the fault point information of the power system, determining the similarity between the measured waveform and the simulated waveform and locating the fault point. It is characterized by comprising a third means and a fourth means for outputting the location result of the located accident point.

Further, the present invention is a power system monitoring method for locating a fault point using a surge in the power system, wherein a surge waveform propagating in the power system is measured to obtain a measured waveform, and system configuration information of the power system and the fault point are obtained. Based on the information, a simulated surge waveform propagating through the power system is generated, a simulated waveform is obtained, the similarity between the measured waveform and the simulated waveform is determined, the fault point is located, and the location result of the determined fault point is output. Characterized by

According to the present invention, it is possible to achieve highly accurate orientation with a simple device configuration.

Specifically, there is no need to measure the arrival time of the surge waveform, no means for removing the reflected wave as noise, and no means for setting or measuring the propagation speed of the surge. All of these contribute to the simplification of the device configuration and the improvement of the positioning accuracy.

The figure which shows the phenomenon of surge propagation in case a surge generate|occur|produces in an accident point. The figure which shows the phenomenon of surge propagation when surge generate|occur|produces in an injection point. The figure which shows the processing procedure of accident point location of this invention. FIG. 4 is a diagram showing an example of simulated waveform generation by drawing; FIG. 4 is a diagram showing an example of simulated waveform generation by drawing; FIG. 4 is a diagram showing an example of simulated waveform generation by drawing; FIG. 4 is a diagram showing an example of simulated waveform generation by drawing; FIG. 4 is a diagram showing an example of simulated waveform generation by drawing; The figure which shows the notation example of the single-shot waveform arrange|positioned at surge arrival time in a drawing procedure. FIG. 4 is a diagram for explaining a method of calculating similarity between a measured waveform and a simulated waveform; The figure explaining the similarity of a measured waveform and a simulated waveform. The figure which shows the accident point when paying attention only to an accident point, and the relationship example of similarity. The figure explaining the processing procedure of accident point orientation. The figure explaining the apparatus structural example of accident point location. The figure explaining the structural example of the display screen of accident point location.

Embodiments of the present invention will be described below with reference to the drawings and the like. The following examples show specific examples of the content of the present invention, and the present invention is not limited to these examples. Various changes and modifications are possible.

Power system technology is sometimes divided into distribution and transmission systems. The following description focuses on distribution systems, but the technology can also be applied to power transmission systems.

In the description of the present invention, AC voltage and current are sometimes simply referred to as voltage and current.

The present invention is characterized by locating an accident point by performing signal processing on a measured waveform of a surge sampled using a measuring means.

The phenomenon of surge propagation in a power system will be described with reference to FIGS. 1a and 1b. Here, FIG. 1a shows a case where a surge occurs at an accident point, and FIG. 1b shows a case where a surge occurs at an injection point.

The horizontal axis of these figures indicates the line of the electric power system, and the fault point Pf (or the injection point Pi), the reflection point Pr, and the measurement point Ps are arranged. The vertical axis indicates the passage of time t after the occurrence of the accident. Since the surge propagates along the track with the passage of time, the relationship between position and time is drawn as oblique lines in the figure. In the following description, this oblique line will be called a propagation line.

A cause of surge generation is a surge (Fig. 1a) that is generated as a transient electrical phenomenon at the accident point Pf and propagates when an accident occurs. There is also a surge (FIG. 1b) generated by injecting a pulse into the line using some pulse generating means. In the latter case, a surge can be artificially generated at a location (injection point Pi) different from the accident point Pf at a timing different from the accident occurrence.

The surge generated in this way can be collected as a surge waveform by measuring the voltage or current waveform at a measuring point Ps prepared on the line. This measurement waveform is composed of a direct wave Wd that directly reaches the measurement point Ps from the fault point Pf (or the injection point Pi) and a reflected wave Wr that reaches the measurement point Ps after being reflected at the reflection point Pr. A reflection point Pr is a point where a surge is reflected due to an electrical impedance mismatch, etc., and includes facility equipment such as line branch points, insulators, line ends, and voltage regulators, and these are grasped in advance from the system configuration information. It is possible. Further, the accident point Pf itself becomes the reflection point Pr. Note that surges can occur in both positive and negative polar directions. In addition, the reflected wave Wr may have its amplitude reversed (positive/negative reversed) before and after reflection. If the amplitude reversal (positive/negative reversal) due to reflection can be correctly estimated, signal processing can be performed focusing on the magnitude, polarity (positive/negative), and position of the surge waveform. However, regarding these polarities, by taking the absolute value or square value of the waveform amplitude, it is possible to focus on the magnitude and position of the surge waveform regardless of the presence or absence of inversion.

In this way, the surge measurement waveform is related by the positional relationship between the accident point Pf (or the injection point Pi) and the reflection point Pr, the propagation speed of the surge, and the time of occurrence of the accident.

To summarize the above, the position of the fault point Pf (or the injection point Pi) and the surge waveform are related by the following known variables and unknown variables. First, known variables are the positions of the reflection point Pr and the measurement point Ps on the power system. On the other hand, the unknown variables are the position of the accident point Pf (or the injection point Pi), the accident occurrence time, and the surge propagation speed.

These variables can be expressed differently depending on the phenomenon of interest. For example, the accident occurrence time is the arrival time of the surge if we focus on the measured waveform. The surge propagation speed can be converted into the time width of the waveform by focusing on the measured waveform. In the following description, these terms may be used by appropriately replacing them.

In the present invention, the above variables are used to locate the accident point according to the processing procedure shown in FIG. 2 below.

In the process of locating an accident point, system configuration information is first inputted in process step S1. The positions of the reflection point Pr and the measurement point Ps can be determined based on the system configuration information. Also, the simulated waveform generating means is prepared. In this preparatory stage, simulated waveform generating means for generating a direct wave Wd and a reflected wave Wr of the surge is prepared using the accident point Pf (or injection point Pi), accident occurrence time, and surge propagation speed as variables. Also, set the initial values of the variables. A series of processes up to this point can be said to be a preparatory stage.

In processing step S2, the system waits until the occurrence of an accident is actually detected. When an accident occurs, a surge waveform is input in processing step S3. Here, a waveform in which the direct wave Wd and the reflected wave Wr of the surge are superimposed is input as the measured waveform Ws of the surge.

In processing step S4, a search for locating the accident point Pf is executed. Specifically, in processing steps S41, S42, and S43, the simulated waveform generation means prepared in advance in processing step S1 is used to provisionally set the fault point, set the delay time and the waveform width as variables, and generate the simulated waveform Wp. to generate Multiple sets of simulated waveforms Wp are generated by varying these variables. In processing step S44, the similarity between the measured waveform Ws and the simulated waveform Wp is calculated. In processing step S45, the relationship between the accident point and the similarity is stored.

In processing step S5, the search procedure is terminated if the termination condition is determined and satisfied. If not satisfied, the above processing steps S41, S42, and S43 are repeated as appropriate.

The search procedure of the present invention is to search for the solution (the variable with the highest similarity) using a numerical analysis technique. For this purpose, search techniques such as full search and optimization calculation are used. As the iteration termination condition, for example, the number of iterations, convergence of similarity, or the like is used.

Finally, in processing step S6, when the termination condition is satisfied, the fault point Pf at which the measured waveform Ws and the simulated waveform Wp are most similar is selected as the location result, and output in processing step S7. The orientation result is output using characters, figures, images, voice, etc. so as to be appropriately communicated to the person in charge of the removal work of the accident point Pf. In addition, progress is accumulated as log data.

As is clear from the above description, the present invention evaluates the similarity between the measured waveform Ws and the simulated waveform Wp using the similarity index. As will be described later, the similarity is calculated as a numerical value representing the similarity of waveforms independent of the waveform position (time) and waveform width (time width). Therefore, if appropriate conditions are added, it is possible to rephrase it into technical terms such as correlation calculation of waveforms, pattern matching of waveforms, and the like.

One of the features of the present invention is that a simulated waveform Wp is created by combining the direct wave Wd and the reflected wave Wr of the surge. Methods for creating the simulated waveform Wp include a drawing method and a formula method. Both are methods of representing the same surge propagation and are compatible.

In the following description of the embodiments, the principle of the drawing method will first be described with reference to FIGS. 3a to 3e, and then the mathematical method will be described.

Simulated waveform generation by drawing will be described with reference to FIGS. 3a to 3e. First, FIG. 3a illustrates a method of generating a simulated waveform Wp when the fault point Pf and the measurement point Ps are different. Here, it is assumed that there are a plurality of reflection points Pr in the front (right) direction and the rear (left) direction of the accident point Pf, respectively, and the measurement point Ps is adjacent to the rear of the accident point Pf. In the case of this positional relationship, the surge is measured at the measurement points Ps at times t1, t2, and t3. Conversely, therefore, a simulated waveform Wp in which a direct wave Wd and a reflected wave Wr appear as surge waveforms at times t1, t2, and t3 is created, and if this is similar to the measured waveform Ws, the fault point is specified as Pf. can be done. In this notation, the solid line represents the propagation path of the direct wave Wd, and the dotted line represents the propagation path of the reflected wave Wr.

FIG. 3b illustrates how the simulated waveform Wp is generated when there are two measurement points. Here, there are a plurality of reflection points Pr and a plurality of measurement points Ps in the forward (right) and rearward (left) directions of the accident point Pf. In the case of this positional relationship, the front measurement point Ps1 is subjected to surge measurement at times t2 and t4, and the rear measurement point Ps2 is subjected to surge measurement at times t1, t3 and t5. Conversely, therefore, simulated waveforms Wp1 and Wp2, in which direct waves Wd1 and Wd2 and reflected waves Wr appear as surge waveforms at times t1, t2, t3, t4, and t5, are created, and even if these waveforms are similar to the measured waveform Ws, , the fault point can be identified as Pf.

FIG. 3c illustrates a method of generating the simulated waveform Wp when the accident point Pf and the measurement point Ps are the same. In this example, the measurement point Ps is generally installed, for example, on a bus line in a substation, and it is assumed that an accident occurs on that bus line. Here, the accident point Pf and the measurement point Ps are located at the same place, and a plurality of reflection points Pr are present in the forward (right) direction and the rearward (left) direction thereof. In the case of this positional relationship, surge measurements are performed at the measurement points Ps at times t1, t2, t3, t4, and t5. Conversely, therefore, a simulated waveform Wp in which a direct wave Wd and a reflected wave Wr appear as surge waveforms at times t1, t2, t3, t4, and t5 is created, and if these are in a similar relationship with the measured waveform Ws, an accident point can be identified. Pf can be specified.

The above cases (FIGS. 3a to 3c) assume the case of FIG. 1a (when a surge occurs at the fault point) in which fault point location is performed when a fault occurs in the power system. On the other hand, the case of FIG. 1b (a surge occurs at the injection point) is the case of artificial pulse injection while the accident continues, and FIG. 3d illustrates the method of generating the simulated waveform.

In the example of FIG. 3d, the injection point Pi and the measurement point Ps are, for example, a work place near the accident site, and are set at substantially the same point. Reflected waves from are sequentially obtained. In this case, since the injection point Pi and the measurement point Ps are at the same place, all of them are reflected waves, but the first surge is treated as a direct wave Wd. In the case of this positional relationship, surge measurements are performed at the measurement points Ps at times t1, t2, t3, t4, and t5. Conversely, therefore, a simulated waveform Wp in which a direct wave Wd and a reflected wave Wr appear as surge waveforms at times t1, t2, t3, t4, and t5 is created, and if these are in a similar relationship with the measured waveform Ws, an accident point can be identified. Pf can be specified.

In the example of FIG. 3e, the system configuration is basically the same as that of FIG. 3a, and the positions of the points are the same. Therefore, the reflected wave Wr3 from the front at the measurement point Ps arrives later than in FIG. Even in this case, it is possible to locate the fault point by confirming the similarity similar to the case described above by creating a simulated waveform that takes into account the change in propagation velocity due to the difference in line type.

In each of the above cases, the horizontal axis indicates the positions of the fault point Pf, the reflection point Pr, and the measurement point Ps along the power system line, and the vertical axis indicates the passage of time t. The oblique lines in the figure are surge propagation lines that have a relationship between line position (horizontal axis) and time (vertical axis). The time axis t plots the passage of time from the time of occurrence of the accident as the origin.

According to the above display, the higher the speed of surge propagation, the more horizontal the slope of the propagation line becomes, and the earlier the arrival time becomes. Conversely, the slower the propagation speed, the steeper the gradient of the propagation line and the later the arrival time. The propagation speed can be converted into the time width of the waveform. The faster the surge propagation speed, the shorter the waveform width (time width), and the slower the surge propagation speed, the longer the waveform width (time width).

The surge includes a direct wave Wd that travels toward both sides of the system from the fault point Pf (or the injection point Pi) and a reflected wave Wr that travels in the opposite direction at the reflection point Pr. When these surges pass through the measurement point Ps, they are measured as a single surge waveform. Therefore, by arranging a single-shot waveform of the surge at the time when the direct wave Wd and the reflected wave Wr arrive at the measurement point Ps, a simulated waveform Wp is created by combining a plurality of single-shot waveforms. A surge can occur both positively and negatively, and the amplitude (positive/negative) of the reflected wave Wr may be reversed at the reflection point. Therefore, there is a method of taking the absolute value or square value of the surge or reflected wave and using it without depending on the positive or negative.

The present invention is characterized in that the combination of the direct wave Wd and the reflected wave Wr observed at the measurement point Ps is generated as the simulated waveform Wp.

As described above, in the present invention, the gradient of the propagation line in the drawing corresponds to the propagation speed of the surge. Therefore, when the difference in line type is known from the system configuration information, the change in propagation speed due to the difference in line type can be reflected as the change in gradient of the propagation line. At this time, it is sufficient that the ratio of the change in the propagation speed due to the line type can be plotted as the ratio of the change in the gradient of the propagation line, so it is not necessary to know the propagation speed quantitatively.

FIG. 4 shows a notation example of a single-shot waveform arranged at the surge arrival time in the drawing procedure described above. FIG. 4 exemplifies an impulse-like waveform, a sinusoidal waveform, a sinusoidal waveform with a partially reversed polarity, and a triangular waveform, and any of these waveforms may be used in the drawing of the present invention.

Although the actually observed surge waveform has a complicated shape, in the present invention, as will be described later, a combination of a plurality of direct waves Wd and reflected waves Wr is treated as a waveform instead of a single surge waveform. Therefore, the single waveform used to generate the simulated waveform Wp need not be similar to the measured waveform Ws. For example, instantaneous waveforms known as delta functions, square waves, triangular waves, etc. waveforms can be used. The amplitude and waveform width of these single-shot waveforms are not limited. The narrower the waveform width of the single-shot waveform, the higher the resolution in the direction of the time axis. Further, the amplitude of the single-shot waveform may be changed assuming that it is attenuated as the propagation distance increases.

The present invention compares the similarity of the temporal positional relationship between the direct wave Wd and the reflected wave Wr contained in the measured waveform Ws and the simulated waveform Wp, in other words, the similarity of the time interval between the direct wave Wd and the reflected wave Wr. is calculated as similarity. The waveform width (time width) of the measured waveform Ws and the simulated waveform Wp used for calculating the similarity is not limited. For example, the waveform width may be arbitrarily set so as to include an appropriate number of single-shot waveforms. Alternatively, the track range to be located and the surge propagation speed are provisionally set, the propagation time of the surge is calculated from the conditions, and the waveforms of the measured waveform Ws and the simulated waveform Wp are included so that the surge that reaches that time is included. A width (time width) may be set. At this time, the waveform width (time width) may be set in consideration of the number of reflection points included in the line range, the attenuation of the surge accompanying the propagation distance, and the like.

Next, as a way of making the simulated waveform Wp, a method using mathematical formulas will be described.

The simulated waveform Wp can be generated using a program or the like by expressing the procedure for generating the simulated waveform Wp by drawing as described above using a formula. For this reason, in the drawing described above, the propagation line indicating the surge propagation is expressed as a linear expression relating to the line direction and the time direction. The reflection point Pr and the measurement point Ps are set as track positions. The intersection of the measurement point Ps and the propagation line (linear expression) becomes the arrival time, and the simulated waveform Wp is generated by arranging the single-shot waveform of the surge (direct wave and reflected wave) at that time. At the intersection of the reflection point Pr and the propagation line (linear expression), the propagation line (linear expression) of the reflected wave traveling in the opposite direction is generated.

In these formulas, the fault point Pf is treated as an unknown variable. Then, using an index of similarity between the measured waveform Ws and the simulated waveform Wp, the fault point with the highest similarity is determined.

An example of the procedure for generating the simulated waveform Wp is shown below.

First, set the variables and conditions to be used for the orientation calculation. Here, the start point is appropriately set for the track position x. An initial value is appropriately set for the time t. The reflection point positions r(i), (i=1 to M) are set based on the system configuration information. The measurement point positions m(k), (k=1 to N) are set based on the system configuration information.

Next, prepare the variables (unknowns) to be used for the orientation calculation. Here, Fs is the position of the accident point, Ts is the arrival time of the surge, and Vs is the propagation speed of the surge. However, here, the surge arrival time Ts is the time when the first surge arrives, and is a converted value from the accident occurrence time.

An example of formulating the propagation line of surge propagation using the above symbols is shown in Equation (1). Note that there is a method of notating a plurality of surge propagation equations in combination in a matrix form, but this is omitted here.

Also, the following rules are prepared to generate a simulated waveform using the surge propagation formula (1). These rules are "propagate the surge waveform in both directions of the track from the accident point Pf", "generate a reflected wave Wr whose direction is reversed at the intersection of the surge waveform and the reflection point Pr", and "surge waveform". A single-shot waveform having a time width is placed at the intersection of the measurement point Ps and the measurement point Ps.

The intersection time Tr between the surge waveform and the reflection point Pr can be obtained by connecting the surge propagation formula S(t) and the position r(i) of the reflection point Pr with an equal sign and solving for the time t. From this intersection point time, a surge propagation formula for a new reflected wave Wr that travels in the opposite direction is created, and the calculation of the intersection point time with the reflection point Pr is repeated.

The intersection time Tm between the surge waveform and the measurement point Ps is obtained by connecting the surge propagation formula S(t) and the position m(k) of the measurement point Ps with an equal sign and solving for the time t. Using this result, a simulated waveform Wp is generated by arranging a single-shot waveform at the time Tm of the intersection.

Here, if there is no measurement error in the measured waveform Ws and there is no error in the system configuration information used for generating the simulated waveform, the fault point Pf (or injection point Pi) can be calculated. In practice, however, the measured waveform contains measurement errors, and the system configuration information used to generate the simulated waveform contains errors.

Therefore, in order to consider the error contained in the measured waveform Ws and the simulated waveform Wp, the present invention arranges single-shot waveforms having a time width as described above, and numerically calculates using a similarity index. Search for fault point Pf (or injection point Pi).

As will be described later, the present invention repeats the procedure for generating the simulated waveform Wp and calculating the similarity between the measured waveform Ws and the simulated waveform Wp by using full search or optimization calculation to obtain the most similarity. Determine the fault point at which .

Here, there is no particular restriction on the time width of the simulated waveform Wp, and it is set so that the direct wave Wd and the reflected wave Wr are appropriately combined. Therefore, a method of setting the time width fixedly, or a method of automatically setting the time width so as to include an appropriate number of direct waves Wd and reflected waves Wr, or the like may be used. If the calculation range is short, the calculation load will be reduced, but the orientation accuracy will decrease. Conversely, if the time width is long, the calculation load will increase, but the orientation accuracy will tend to improve.

In the calculation process of this search procedure, the similarity between the measured waveform Ws and the simulated waveform Wp is highest when the fault point Pf (or the injection point Pi) is correctly obtained, and the similarity is low otherwise. can get. For example, if the horizontal axis is the track position and the vertical axis is the similarity, a convex distribution characteristic with the highest similarity at the accident point position can be obtained. This distribution characteristic can be expressed as the basis for fault point determination or as the accuracy of fault point determination.

Next, calculation of similarity between the measured waveform Ws and the simulated waveform Wp will be described.

The present invention includes a procedure for calculating the similarity between the measured waveform Ws and the simulated waveform Wp. In the field of geometry, when considering the similarity of two figures, the term "similarity" is used as an index independent of the orientation and size of the figures. In the present invention, the similarity between the measured waveform Ws and the simulated waveform Wp is calculated as similarity in order to consider the similarity regardless of the start point and width (time width) of the waveform. Then, the present invention determines the accident point by utilizing the fact that the similarity between the measured waveform Ws and the simulated waveform Wp is highest when the accident point is correctly located.

Calculation of the similarity between the measured waveform Ws and the simulated waveform Wp will be described with reference to FIG. Here, the time-series information of the measured waveform Ws is represented by X(t), and the time-series information of the simulated waveform Wp is represented by Y(t). The measured waveform X(t) is measured as a wave, but the simulated waveform Y(t) is a simulated waveform based on an accident point, and is expressed as an impulse train including wave arrival time and propagation velocity (waveform width). .

As shown in FIG. 5, the measured waveform Ws and the simulated waveform Wp are represented by time series signals X(t) and Y(t) at time t. Here, time t indicates a discrete sampling time. Therefore, it may be replaced with a discrete sampling order. Examples of similarity calculation methods are shown in the following equations (2), (3), and (4). where the symbol Σ (sigma) denotes addition over time t.

The similarity 1 shown in the formula (2) is obtained by repeatedly multiplying X(t) and Y(t) at the same time in the set time width T (t=0 to T) in the time axis direction. By doing so, the similarity is calculated. This formula is equivalent to the correlation calculation formula, and the more similar the two are, the larger the value obtained. Therefore, the relationship between the fault point Pf and the similarity is calculated, and the fault point Pf can be calculated correctly when the similarity is the highest.

Similarity 2 shown in equation (3) is a modification of the similarity 1 calculation equation, and indicates a case where X(t) and Y(t) at the same time are converted to positive values and then similarity is calculated. . To get a positive value, there are a method of squaring, a method of taking an absolute value, and so on. Here we show how to take the absolute value. This formula can calculate the similarity focusing on the positional relationship even when the amplitude (positive/negative) of the reflected wave is reversed.

Similarity 3 shown in equation (4) is obtained by repeating the calculation of squaring the difference between X(t) and Y(t) at the same time within a set time span, using the time axis method. Calculate gender. The reason for squaring is to produce a positive value, and the absolute value may be taken instead. The denominator of this formula is equivalent to the formula for calculating the squared error, and the more similar the two, the smaller the value.

The present invention does not evaluate the similarity of a single surge waveform, but evaluates a waveform in which a direct wave Wd and a reflected wave Wr are combined as a simulated waveform Wp. Therefore, the single-shot waveform used to generate the simulated waveform Wp need not be similar to the single-shot waveform included in the measured waveform Ws. For example, a binary square wave of 0 and 1 may be used as a single-shot waveform for creating the simulated waveform Wp. At this time, for example, the similarity 2 in equation (3) can be obtained by simply adding the measured waveforms in the range where the simulated waveform is 1, as shown in equation (5).

The above can be extended to an equation for calculating similarity using measured waveforms Ws obtained from a plurality of measuring means. At this time, the measured waveform Xi(t) is sampled at the measurement point k (k=1 to N) to generate the simulated waveform Yk(t) corresponding to the measurement point k (k=1 to N). Then, the similarity is calculated using the measured waveform Xk(t) and the simulated waveform Yk(t). Examples of similarity calculation formulas when formulas (1), (2), and (3) are expanded are shown in formulas (6), (7), and (8).

In these formulas, the symbol Π (pi) represents the product for the measurement points k (k=1 to N), and the symbol Σ (sigma) represents the addition for the time t (t=0 to T). Calculating the similarity using a plurality of measured waveforms has the effect of canceling random noise components and improving the accuracy of accident point location.

Next, a similarity calculation method considering the starting point and width (time width) of the waveform will be described. In the present invention, as shown in FIG. 6, the similarity is calculated using the accident point, surge arrival time, and surge propagation speed as variables. Here, a measured waveform X(t) shown as a wave and a simulated waveform Y(t) shown as an impulse train are superimposed and displayed as simulated waveforms based on fault points. Here, the wave arrival times and propagation velocities (waveform widths) are superimposed and displayed. Therefore, the variables of the simulated waveform Y(t) are the distance to the fault point, the arrival time, and the propagation speed (waveform width).

Among these variables, the accident point is the place where the surge of the simulated waveform Y(t) occurs.

The time when the next variable, the surge, arrives at the measurement point is the start time of the simulated waveform Y(t). Assuming that the start point time of the waveform is Sa, the time-series signal of the simulated waveform Y(t) can be expressed as Y(t−Sa).

The last variable, the speed of propagation of the surge, reflects the width of the waveform. If the propagation speed is Sv, the time-series signal of the simulated waveform Y(t) can be expressed as Y(Sv·t). If the propagation speed Sv is large (if the propagation speed is fast), the waveform width will be shortened corresponding to advancing the time of the simulated waveform Y(t). ) The width of the simulated waveform Y(t) is lengthened correspondingly to delaying the time.

As described above, the present invention calculates the similarity for the entire waveforms of the measured waveform X(t) and the simulated waveform Y(t), so that the effects of noise, waveform distortion, and the like are reduced. In other words, the calculation procedure of the present invention itself has the effect of removing noise. For this reason, preprocessing such as noise removal and frequency filtering for the measured waveform is not required, or even if used, simple calculations are sufficient, which is effective in reducing the processing load.

Next, a search procedure based on numerical calculation will be described. In order to numerically search for an accident point where the measured waveform X(t) and the simulated waveform Y(t) are most similar, the accident point, surge arrival time, and surge propagation speed are prepared as unknown variables.

FIG. 7 illustrates the relationship between the accident point and the similarity when focusing only on the accident point among the variables. The relationship between the accident point and the similarity is that the similarity is highest when the accident point is set correctly, and the similarity decreases as the accident point is set, resulting in a distribution characteristic centered on the accident point. Similarly, regarding the relationship between the three unknown variables and the similarity, the value of the variable is determined by numerical calculation, utilizing the fact that the simulated waveform and the measured waveform have the highest similarity when they can be estimated correctly.

In the above description, system configuration information, which is known information, and fault point information, which is unknown information, are used to generate the simulated waveform Wp. Here, the system configuration information is system configuration information including reflection points, and the fault point information is the fault point, surge arrival time, and surge propagation speed, which are unknown variables. Then, in generating the simulated waveform Wp, a plurality of accident points are provisionally set, and a plurality of simulated waveforms Wp are obtained with the surge arrival time and the surge propagation speed as unknowns.

Each of the above processes can be realized in the flowchart of FIG. It should be noted that the present invention does not limit the method of search by numerical calculation, and a calculation method such as a full search or an optimization calculation for speeding up can be used. This search method is determined in consideration of calculation load, validity of search results, stability of search procedures, and the like.

For example, when searching by the exhaustive search method, create a triple loop that repeats calculations while changing the above-mentioned unknown variables such as accident point position, surge arrival time, and surge propagation speed, and finds similarity in each loop. Calculate and detect the variable when the similarity is the highest. In this way, the method of performing loop calculations covering the possible range of variables increases the calculation load, but the search is complete and the validity and stability of the results are excellent.

Also, the present invention does not limit the optimization calculation method and procedure. For example, when PSO (particle swarm optimization method) is applied to the present invention, a solution candidate is set on a space whose coordinate axes are the accident point position, the arrival time of the surge, and the propagation speed of the surge, and the most similarity A search is made for a solution candidate with a high . Then, by preparing some termination condition for the repeated calculation of the search, the end of the search is determined, the repeated calculation is terminated, and the result is output. By using the optimization calculation, it is possible to reduce the calculation load compared to the exhaustive search.

In Embodiment 1, the general description of the present invention and the case where a surge is generated mainly at the fault point of FIG. 1a are mainly described. On the other hand, in Example 2, the case where the pulse injection of FIG. 1b is performed will be described in detail.

The method using pulse injection is used in the field search scene to identify the detailed location of the accident point when workers arrive near the accident site and the accident continues after the accident has occurred and cannot be eliminated. often

As one embodiment of the present invention, an external device can be used to generate a surge by injecting a pulse into the railroad track after the occurrence of an accident, thereby locating the accident point.

In this method, the pulse injection point Pi becomes a surge generation point, and the accident point Pf becomes a surge reflection point Pr. Then, the simulated waveform Wp is generated in the same manner as described above for the surge generated and propagated at the pulse injection point Pi, and the accident point can be located by searching for similarity between the measured waveform Ws and the simulated waveform Wp. .

In this method, after an accident occurs, a surge is generated by using an appropriate point on the line as a pulse injection point. The present invention does not limit the principle of pulse injection, device configuration, working method, and the like. A surge can be generated by pulse injection using means such as electrode contact, electromagnetic induction, and the like. In addition, a surge can be generated by switching devices such as SVR, SVC, switches, etc. installed on the line. These devices may be used to detect changes in the state of the system by generating and measuring surges at appropriate timings when the system is healthy, not just after an accident.

The present invention uses the processing flow procedure shown in FIG. 8 to locate the accident point Pf.

First, in processing step S11, a pulse injection point Pi and a measurement point Ps are installed on a target railroad track. It should be noted that the two may be installed at the same location, but when this method is applied, they should be installed at the site.

Next, in processing step S12, a pulse is injected and a measured waveform Ws is sampled.

Next, in processing step S13, a simulated waveform Wp is generated based on the information on the installed pulse injection points and measurement points and the target system configuration. This processing can be realized by applying the processing of processing step S4 in FIG.

Finally, in processing step S14, the accident point is located using the similarity between the measured waveform Ws and the simulated waveform Wp. This processing can be realized by applying the processing of processing steps S6 and S7 in FIG.

According to the accident point locating method according to the second embodiment of the present invention, it is possible to locate the accident point after the occurrence of the accident, to set any location as the pulse injection point and measurement point, and to perform multiple locating operations. By repeating the work, it becomes possible to narrow down the accident point.”, “Existing equipment such as SVR, SVC, switches, etc. installed on the track can be used as a means of pulse injection (surge generation). ” effect can be obtained.

It takes a lot of manpower and time to patrol tracks on which measuring instruments are not installed and search for accident points. On the other hand, the accident point locating method by pulse injection according to the second embodiment of the present invention has the effect of reducing manpower and time required for searching for the accident point after the occurrence of the accident.

In the third embodiment, handling of the reflection point Pr will be described. This is the handling of the new occurrence or disappearance of the reflection point Pr.

As described above, there are branch points, insulators, and the like in the impedance unmatched portions that become the reflection points Pr of the surge. In addition, the reflection point Pr may appear and disappear for some reason. For example, due to changes in characteristics due to deterioration over time, new reflection points Pr may occur at locations where the insulation resistance is lowered. In addition, impedance mismatch may occur or disappear due to temporary contact with another object. The generation and disappearance of such reflection points Pr suggest some change in the state of the system, and can be used as an advance alarm as a concern that the supply reliability will be lowered even if it does not lead to an accident.

The present invention can detect the generation and disappearance of the reflection point Pr from the change in the surge measurement waveform Ws. The accident point locating technique of the present invention can be applied to ``locate a place where a reflection point is generated or disappeared'' and ``locate an accident point in a situation where a reflection point is generated or disappeared''.

An orientation method using the pulse injection method will be described below. Method 1 in this case is to output an alarm.

By comparing the currently measured waveform with the previously measured waveform, when a change in the reflection point is detected even though no accident has occurred, it is determined that a new reflection point has occurred. Then, the position of the new reflection point is calculated using the accident point locating method.

For example, a temporary contact with a tree may create a new reflection point, and repeated contact may cause an accident. The occurrence of a new reflection point can be output as an alarm in order to eliminate the cause of the accident.

Method 2 in this case is to narrow down the accident point after the accident. After the occurrence of the accident is detected, the measured waveform of the surge is sampled by the pulse injection method, and the position of the newly generated reflection point is located using the accident point location. Accident points can be narrowed down gradually by repeatedly injecting pulses and locating the accident points.

In the third embodiment of the present invention, the similarity of the entire simulated waveform is used to locate the accident point, so even if some reflection points are missing or added in the generation of the simulated waveform, the location is possible.

In a fourth embodiment, a description will be given of speeding up the arithmetic processing relating to the generation of the simulated waveform Wp and the location of the accident point.

The accident point location method of the present invention uses the accident point, accident occurrence time (surge arrival time), and surge propagation speed as variables to repeatedly calculate the similarity until the termination condition is satisfied. This iterative calculation can be realized using optimization calculation. Although the optimization calculation method is not limited, there is PSO (particle swarm optimization method), for example.

For the sake of simplicity, if the surge arrival time is set to 0, the fault point and the surge propagation speed are two unknown variables. Variables (fault point and surge propagation speed) are set in the repeated calculation, a simulated waveform is generated by the method described above, and the similarity with the measured waveform is calculated. Then, search for two variables (fault point and surge propagation velocity) that have the highest similarity. As end conditions for iterative calculations, judgments are made such as when the similarity exceeds a certain value, when the change in similarity becomes smaller than a certain value, or when the number of iterative calculations exceeds a certain value. .

Embodiment 5 describes the configuration of the power system monitoring apparatus of the present invention. In particular, a configuration for data transmission between slave stations and servers will be described.

FIG. 9 shows a device configuration example of a power system monitoring device that implements the present invention. The device configuration is determined based on criteria such as processing time, cost, amount of data to be transmitted, and the like.

In FIG. 9, the waveform is measured at the measurement point of the power system (in the example shown, the installation point of the switch with the sensor is shown), the measured waveform data is obtained on the server side, for example, via a network, and a predetermined Display the processing result on the screen.

When considering the configuration of the device that implements the present invention, the main elemental functions that realize this are input means for measured waveforms, means for generating simulated waveforms, means for calculating similarity, and accident points with the highest similarity. It can be said that it is a means to determine

These means can be distributed to the child station and the base server. Here, the base server indicates a centralized monitoring control device such as a power distribution automation system. The method of distributed arrangement is determined in consideration of equipment cost, processing time, and the like. In the case of distributed arrangement, data transmission means for exchanging data between them is prepared.

As one configuration example, there is a method of arranging all of the above means on the slave station side. With this configuration, only the location results need to be transmitted to the site server side, and the amount of data transmission can be reduced. Further, by generating the simulated waveform on the base server side and transmitting it to the slave station side, it is possible to reduce the processing load on the slave station side.

As another configuration example, there is a method of transmitting waveform data measured on the slave station side to the base server side and performing orientation using the measured waveform and the simulated waveform on the server side. This configuration enables signal processing utilizing the computing power of the server, and reduces the load on the slave station. In addition, by consolidating surge waveforms measured at multiple locations, it is possible to improve the location accuracy.

There is also a method of constructing a portable accident point locating device without using the slave station and base server. By providing a pulse injection means, a surge is generated at an arbitrary point after an accident to locate the accident point. In addition to being used for detailed searches in the vicinity of the accident, it is also used to confirm that the accident point has been removed after the accident restoration work. As a pulse injection method, there is a method of contacting an electrode with an electric wire. In order to collect the measured waveform, there is a method of bringing the electrode into contact with the electric wire. In order to improve workability, the pulse injection point and the measurement point can be set at the same point to collect the measured waveform. Then, a simulated waveform is generated using information on nearby reflection points, and accident point location can be performed using the similarity between the measured waveform and the simulated waveform. By using a portable device, it is possible to set an arbitrary timing and an arbitrary location after the occurrence of an accident and repeat orientation, thereby gradually narrowing down the accident point.

In the sixth embodiment, the output of the orientation result and how to use it will be explained.

FIG. 10 shows an example of screen configuration for displaying the result of accident point location according to the present invention. Here, an example of presenting the location result of the accident point, the grounds for the location result, and the accuracy of the location to the person in charge of restoration work is shown.

By temporarily storing and presenting the progress data of the search procedure for determining the accident point with the highest similarity, the grounds for the orientation can be shown. In many cases, the relationship between the position of the accident point and the similarity is a distribution characteristic centering on the accident point with the highest similarity. Presenting the relationship between the accident point and the similarity will show the grounds and accuracy of the orientation to the operator who uses the orientation result, and help the interpretation.

In addition, the present invention can use the history of the search procedure in the same manner as described above to prepare conditions for determining that accident point orientation is impossible for some reason, and present the determination results and grounds. For this reason, the judgment conditions that make it impossible to locate the accident point are that "measurement waveform distortion and noise are large, and location accuracy cannot be obtained", "there are no reflection points on the track, and the measurement "Cannot obtain a waveform or simulated waveform", "There are a large number of reflection points on the track, and the reflected waves overlap to form an averaged waveform", etc.

The present invention uses the above conditions to determine whether orientation is impossible. The threshold used for this determination may be fixed in advance, or may be set according to the characteristics of the target system. By outputting the determination result of whether or not the location can be determined together with the location result of the accident point, it can be used as a guideline for how to proceed with the restoration work.

In a seventh embodiment, determination of the presence/absence of an accident point will be described. Determining whether or not there is a fault point in the target system using fault point location according to the present invention will be described.

In the accident point location by pulse injection, the direct wave of the surge generated by the pulse injection, the reflected wave from the reflection point on the track, and the reflected wave reflected at the accident point are collected as measurement waveforms. On the other hand, the simulated waveform is generated by combining injection points for pulse injection, reflection points set from system configuration information, and fault points as variables. If a fault point is not set when generating a simulated waveform, a surge waveform that propagates through sound lines can be generated.

If the measured waveform has the highest similarity to the simulated waveform created without setting the fault point, the target line is determined to be healthy (no fault point). By outputting these determination results, it is possible to support determination of the necessity of restoration work.

In this way, the methods of judging the presence or absence of an accident point using fault point location include: "Confirmation of fault point removal by accident recovery work", "Presence or absence of trouble (fault point) in the It can be used for "confirmation of

For example, for a partial system (or microgrid) that was disconnected from the system using a switch or the like when an accident occurred, confirm that there is no fault point in the system or that the fault point has been removed before attempting to restore power. To do so, you can use the procedure above.

In this way, the present invention can be used as a method for confirming that there is no fault point in the target system in order to start power restoration after a fault.

Claims (11)

A power system monitoring device that locates a fault point using a surge in a power system,
A first means for obtaining a measured waveform by measuring a surge waveform propagating through an electric power system;
a second means for obtaining a simulated waveform by generating a simulated surge waveform propagating through the power system based on system configuration information and fault point information of the power system;
a third means for locating an accident point by judging the similarity between the measured waveform and the simulated waveform;
a fourth means for outputting a location result of the location of the accident point;
with
A power system monitoring apparatus, wherein the system configuration information of the power system includes information on reflection points, and means for determining the presence/absence of newly generated or disappeared reflection points. The power system monitoring device according to claim 1,
The power system monitoring apparatus according to the first means obtains a measured waveform by measuring a surge waveform generated at an accident point in the power system. The power system monitoring device according to claim 1,
A power system monitoring apparatus, characterized in that a surge is injected into a power system, and the first means measures a surge waveform due to the injected surge. The power system monitoring device according to any one of claims 1 to 3,
A power system monitoring apparatus, wherein the measured waveform and the simulated waveform are waveform trains in which a plurality of single-shot waveforms are combined in time series. The power system monitoring device according to any one of claims 1 to 4,
The power system monitoring apparatus, wherein the system configuration information is system configuration information including known reflection points, and the fault point information is unknown variables such as fault point, surge arrival time, and surge propagation speed. . The power system monitoring device according to claim 5,
A power system monitoring apparatus characterized in that, in generating the simulated waveform, a plurality of fault points are provisionally set to obtain a plurality of simulated waveforms with unknown values of surge arrival time and surge propagation speed. The power system monitoring device according to any one of claims 1 to 6 ,
Data transmission means is provided between the processing device comprising the second means and the third means, and the first means or the fourth means, and data is exchanged between them. and a power system monitoring device. The power system monitoring device according to any one of claims 1 to 7 ,
The power system monitoring apparatus, wherein the fourth means outputs both the location result of the accident point and the estimation accuracy of the location result. A power system monitoring method for locating a fault point using a surge in a power system,
Obtain a measured waveform by measuring the surge waveform propagating in the power system,
obtaining a simulated waveform by generating a simulated surge waveform propagating through the power system based on system configuration information and fault point information of the power system;
determining the similarity between the measured waveform and the simulated waveform to locate an accident point;
Outputting the orientation result of the oriented accident point ,
A power system monitoring method , wherein the system configuration information of the power system includes information on reflection points, and the presence or absence of newly generated or disappeared reflection points is determined. A power system monitoring method according to claim 9 ,
A power system monitoring method, wherein a surge propagation path is plotted on a position and time plane on the power system in order to obtain the simulated waveform. A power system monitoring method according to claim 9 ,
A method for monitoring a power system, wherein a numerical analysis method using formulas based on known system configuration information and unknown fault point information is executed in order to obtain the simulated waveform.

Images