KR20180086655A - Method for Estimating Model Parameters of Generator Based on PMU(Phasor Measurement Unit) Measurement - Google Patents

Abstract

The present invention relates to a method for estimating a generator constant based on a PMU measurement, comprising the following steps of: (A) receiving a measurement value including instantaneous voltage and current values for each generator online through a network by using a measurement device installed on a power system including one or more generators connected to a power network through a transformer; (B) generating simulation result data for an output of a simulation regeneration generator by inputting simulation input data including a plurality of estimation target constants determined based on the measurement value to the simulation regeneration generator and performing simulation, with respect to a simulation equivalent system for simulation of the output of the simulation regeneration generator on the power network using a predetermined program; and (C) correcting at least one of the estimation target constants based on sensitivity analysis for the estimation target constants, and determining model constants for the generator according to an error by performing the simulation.

Description

[0001] The present invention relates to a PMU measurement based generator parameter estimation method,

More particularly, the present invention relates to a method of estimating a generator constant of a system based on actual measurement data obtained when a system fault occurs, using a PMU (Phasor Measurement Unit) apparatus connected to a real system .

As the scale of the power system grows and the number of new new facilities increases, a variety of technologies are required to operate the power system more sophisticatedly. Among them, the basic technology of stable system operation and planning is system stability analysis. For accurate and precise stability analysis, the parameters of the stability analysis model should be tuned similar to the actual generator. Tuning of existing generator parameters has been accomplished through installation of new generators or control systems or periodic performance tests. However, existing generator performance tests have limited the ability to track generator parameters that reflect system characteristics because the generator uses measured data based on test scenarios under specific conditions.

For example, a typical conventional technique uses a method of deriving a model constant using a first-order equivalent model using data acquired through a generator characteristic test. The generator characteristic test that is currently performed is to apply a specific test condition to the generator if there is no generator model constant, and derive the model constant using the acquired data. However, since the existing method is an integer obtained by using the characteristics of the generator response to the specific test condition, the generator is operating in connection with the system, so that the response to the fault occurring during operation in connection with the actual system may be significantly different from the test condition have. Accordingly, when the existing model is used, the output may be more or less than the actual generator output.

SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide a PMU installed in a power system, It is possible to reflect the fluctuation characteristics due to the system failure of the generator connected to the actual system to the model constants to derive more accurate and practical model constants and to obtain accurate model constants, This method is applied to the establishment and operation of the operation and planning method of the system, and to provide a method of estimating the generator water constant which can be utilized for the establishment of measures for system failure.

In order to achieve the above object, a method for estimating a generator model constant according to an aspect of the present invention includes: (A) Receiving a measurement value including an instantaneous voltage and a current for each generator using a measurement device; (B) generating simulation input data including a plurality of estimated object constants determined based on the measured values for a simulated reduction system for simulating the output of the simulated regenerative power generator on the power grid using a predetermined program, Generating simulated result data for an output of the simulated regeneration generator by performing simulation on the simulation output data; And (C) re-executing the simulation after correcting for at least one of the estimated constants based on the sensitivity analysis for the estimated constants, and determining model constants for the generator according to the error .

If the error between the measured value and the simulation result data is smaller than the threshold value, the estimation target constants are determined as model constants for the generator.

In the step of generating the simulation result data, steps (A) and (B) are performed for a predetermined start time to end time, and the simulation is restarted according to the error, (A) and (B) are performed for each sample interval by dividing the sample interval into a plurality of sample intervals, and the simulation is re-executed for the entirety of the plurality of sample intervals according to the error after the simulation for the entire plurality of sample intervals One of which is used.

The measured value is data for the primary or secondary of the transformer connected to the generator. The measured value includes an effective value voltage, frequency, active power and reactive power converted from the instantaneous value voltage and current for the generator.

A central server on the network receives the measurements online from the measurement device to perform the simulation, modify the estimated constants, and determine model constants for the generator. The model constants for the plurality of generators can be simultaneously determined by performing the time synchronization with respect to the plurality of generators.

Calculating a Jacobian matrix J for the sensitivity on the basis of the difference between the measured value y and the simulation result data according to the correction amount delta of the estimation target constants beta , Determining the estimated constants ? That are variables of the objective function f such that the difference between the measured value y and the simulation result data is minimized and the following equation is satisfied,

Here, I is an identity matrix, and μ is an adjustment coefficient for preventing divergence and improving convergence speed in the sensitivity calculation process.

In the step of determining the model constants, a method of estimating a parameter, which is the estimation object constants, from a plurality of measured values while adding a new measured value except for a part of a previously used measurement value in the next estimation, A method of selecting representative values from a plurality of parameters estimated after separately estimating parameters from measured values while adding the estimated parameters from the new measured values to the extent of the previously used parameters except for some of the previously used parameters is used.

In the step of determining the model constants, in order to select any one of the estimated values for the estimated constants as a representative value, the distribution of the sum of the gravitational fields with respect to the ordered pair of the estimated constants is calculated, The corresponding value of the strong point can be selected as the representative value of the ordered pair of the estimated object integers.

According to another aspect of the present invention, there is provided a recording medium embodied on a computer-readable code, comprising: (A) a plurality of generators connected to a power grid through a transformer, A function for receiving a measured value including an instantaneous value voltage and current; (B) generating simulation input data including a plurality of estimated object constants determined based on the measured values for a simulated reduction system for simulating the output of the simulated regenerative power generator on the power grid using a predetermined program, To generate simulated result data for the output of the simulated regeneration generator; And (C) after correcting at least one of the estimated constants based on the sensitivity analysis on the estimated constants, performing the simulation again to determine model constants for the generator according to the error , The model constants of the generators in the grid system can be estimated.

According to the method of estimating the generator constant according to the present invention, the generator model constant can be estimated and adjusted so that the dynamic characteristics of the generator when the system failure occurs in the actual system operation conditions and the dynamic characteristics of the generator appear in the simulation. The simulation results are compared with the results obtained by extracting PMU data related to the system from the results of the simulation. By adjusting the generator model constants to be similar to the PMU data, simulation results closer to the real system can be obtained. The system stability simulation is the most basic simulation in planning and operating the power system. If the reliability of the simulation is raised to reflect the characteristics of the actual system in the simulation, it is possible to obtain more accurate and reliable stability analysis results As a result, it becomes the base technology of stable system planning and operation, so that it can be used accurately and easily for establishment of measures for system failure.

1 is a view for explaining a method of measuring measured data at a generator terminal or a system connection point according to an embodiment of the present invention.
2 is a diagram for explaining a measurement method of measured data that minimizes a cost according to an embodiment of the present invention.
3 is a diagram for explaining a method of generating virtual data for dynamic simulation according to an embodiment of the present invention.
4A is a diagram for explaining an example of a generator model constant estimation method of the present invention.
FIG. 4B is a diagram for explaining another example of a generator model constant estimation method of the present invention.
5 is a diagram for explaining an example of a model constant estimation and representative integer selection method in the present invention.
FIG. 6 is a diagram for explaining a method of using statistical techniques for model constant estimation in the present invention.
7 is an example of a three-dimensional distribution of the sum of the gravitational fields generated by each point when there are arbitrary 10 ordered pairs in the two-dimensional plane.
FIG. 8 is a diagram showing the distribution of the sum of the gravitational fields of FIG. 7 by contour lines. FIG.
Fig. 9 is a contour map showing the distribution of the sum of gravity fields for a large number of ordered pairs composed of two model integer pairs.
FIGS. 10, 11, and 12 show an ordered pair of distributions, contour lines, and reactive power-related constants of the estimated model constants K and T5 based on the measured data based on any generator.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same components are denoted by the same reference symbols as possible. In addition, detailed descriptions of known functions and / or configurations are omitted. The following description will focus on the parts necessary for understanding the operation according to various embodiments, and a description of elements that may obscure the gist of the description will be omitted. Also, some of the elements of the drawings may be exaggerated, omitted, or schematically illustrated. The size of each component does not entirely reflect the actual size, and therefore the contents described herein are not limited by the relative sizes or spacings of the components drawn in the respective drawings.

First, a generator constant estimation method according to an embodiment of the present invention may be implemented using an electronic device such as an EMS (Energy Management System) or a central server as described below, , Software such as an application program, or a combination thereof. The electronic device may be a computing system (e.g., a personal computer (PC), a large-capacity computer, etc.) and may include at least one processor, memory, user interface input device, user interface output device, storage, and network interface. A processor may be a central processing unit (CPU) or a semiconductor device that performs processing on instructions stored in memory or storage.

<1. Measurement>

1 is a view for explaining a method of measuring measured data at a generator terminal or a system connection point according to an embodiment of the present invention.

Referring to FIG. 1, one or more constants to be estimated are selected from among generators G ₁ ,..., G _n that supply power to a main grid through a bus, (M1s, ..., Mnns) of the secondary side of the transformer connected to the output of each generator which is the generator terminal or the instantaneous value voltage / current (M1s, ..., Mnp) Is measured by using a measuring device such as PT (Transformer) / CT (Potential Transformer / Current Transformer), MOF (Metering Out Fit) or PMU (Phasor Measurement Unit) Can be measured in units of time (t). The above EMS (control device linked with the central server) can store and manage measured values converted into actual value voltage, frequency, and valid / reactive power from the measured data of the measured target generator to a memory or the like. Thus, the estimation accuracy can be improved by measuring both the grid connection point and the generator terminal as required. Also, in order to accurately reflect the state, an automatic generation control (AGC) signal transmitted from the EMS to the generator at the time of measurement may be acquired together and used for synchronization or the like to improve estimation accuracy.

The same point in time for the individual measured data (M1s, M1p, ..., Mns, Mnp) each also possible to estimate the constant generator to separately used, but in the estimation target generator _{(G 1, ..., G n} ) The individual measurements for all the individual generators must be made at the same time so that the characteristics of each of the generators in the generator are accurately reflected. For this, each individual measurement is made at the same time based on the start of measurement and time synchronization using the AGC signal, and the time synchronization between the measurement devices is performed through local communication between measurement devices or GPS (Global Positioning System) . Especially, wide - area time synchronization such as the nationwide range can be achieved mainly through GPS visual reception.

Using such a measurement method, wide-area actual data on a nationwide basis can be obtained at the same time. The measured data is transmitted online, and the central server receiving the data can simultaneously determine the correct model constants of all generators at the critical point. This minimizes field testing and greatly shortens the delay time to reflect the generator model, contributing to the accuracy of the system analysis. Incidentally, it is possible to minimize the cost and manpower of the field test.

Since it is important to minimize the cost for nationwide measurement, a measurement method that minimizes the cost among the measurement methods at various positions as shown in Fig. 1 can be shown in Fig.

2 is a diagram for explaining a measurement method of measured data that minimizes a cost according to an embodiment of the present invention.

Referring to FIG. 2, there is shown a method of eliminating redundancy of measurement and minimizing the cost of measurement. That is, CT is installed at the grid connection point of the primary side of the transformer connected to each generator output to measure the instantaneous value current at the grid connection point of each generator and PT is installed to measure the instantaneous value voltage at grid connection point of each generator . In accordance with the measured instantaneous value voltage and current, the EMS (control device linked to the central server) converts the measured value converted into the effective value voltage, frequency, effective and reactive power inputted from the individual generators into the system, , A wireless communication network, etc., to a central server through a network.

<2. Virtual data generation and verification>

As shown in FIG. 1, when data is acquired at all the points on the primary and secondary sides of each generator, dynamic simulations can be performed without separate virtual data. However, based on the acquired data, virtual data at another point may be generated and compared with the measurement data at the corresponding point, thereby verifying the accuracy of the measurement to remove the defective data.

The virtual data can be generated based on the impedance or line impedance of the transformer depending on the actual system configuration, and it is possible to generate the secondary data using the primary data or generate the data in the reverse direction. For example, when the voltage, frequency, effective, and reactive power data of the transformer secondary side are measured at the G terminal of the generator as shown in Fig. 3, the impedance of the transformer between the generator G and the PB generator (simulated regenerative generator for estimating the system change) It is possible to generate PB (play back) virtual data at the primary side of the opposite transformer by reflecting variables affecting the data such as line impedance, and it is also possible to generate virtual data in the reverse direction. That is, for dynamic simulation, the simulator of the central server directly measures the data at the PB generator position when inputting the simulated input data including the estimation object constant determined based on the measurement data to the PB generator , It is possible to generate dynamic simulations by generating virtual data (data on the transformer primary side corresponding to the transformer secondary data) and reflecting the simulated input data.

<3. Generator model constant estimation>

The reduction system for estimating the generator model constants in the simulator of the central server is composed of a transformer primary side connected to the power grid (main grid) for estimating the model constants for the estimation target generator G on the secondary side of the transformer, And a PB generator (a simulated regenerative generator for estimating the system change). This simulated contraction system contributes to minimizing the operator circle for estimating the model constants.

The simulated input data including the estimated object constants determined based on the measurement data is input to the PB generator to simulate the output of the PB generator in the system of such a grid network and the generator G ), The simulator of the central server can simulate using a predetermined application program for simulating the operation of the circuit, and the output voltage, frequency, effective and reactive power of the PB generator And so on. Using the simulation result data of the PB generator, the model constant of the estimation subject generator G that can be satisfied to approximate the value measured in the estimation subject generator G according to the principle of FIG. 3 can be determined.

4A is a diagram for explaining an example of a generator model constant estimation method of the present invention. Here, a method of determining the model constants of the generator G by performing simulations from a predetermined start time (t ₀ ) to an end time (t _f ) is shown.

Referring to FIG. 4A, in an example of a method of estimating the generator model constants performed in the simulator, first, a method of estimating the generator model constants is performed on the grid on the main grid by using the measured values such as the effective value voltage, frequency, Simulation results (simulated results) of the output of the PB generator in the power network from the simulated start time (t ₀ ) to the end time (t _f ) together with the simulated input data of the PB generator including the estimated target constants determined on the basis (E.g., voltage, frequency, effective, and reactive power, etc.) in order to simulate a part of the estimation target constant after modifying the estimation target constant The sensitivity of the model to the target integer is analyzed (S11). The number of simulations for sensitivity analysis increases linearly with increasing number of estimated constants.

For example, one or more model constants that are highly sensitive to parameter value changes among the estimation target constants (e.g., values associated with input / output voltage / current, effective / reactive power, gain, impedance, etc.) . Thereafter, the value of the corresponding estimation target integer selected on the basis of the sensitivity calculation is modified (S12), and the simulation results (e.g., voltage, frequency, validity and reactive power, etc.) Is calculated (S13). If the error due to the estimated integer is sufficiently small (smaller than the threshold value) (S14), the estimation target constants are determined as the model constants of the generator G and reflected in the generator model of the predetermined systematic analysis program (S15) If the error is sufficiently small, it is repeated from the above sensitivity analysis to determine the estimated constants including the finally corrected model constants as the model constants of the generator (G). This is the case for a typical local optimization technique, and the method of FIG. 4A would be superior to the method of FIG. 4B, so it would be reasonable to select it.

FIG. 4B is a diagram for explaining another example of a generator model constant estimation method of the present invention. Here, a method of determining the model constants of the generator (G) by performing a system simulation by dividing one sample of measured values is shown. This is equivalent to the Kalman filter tracking technique. Estimation of the generator model constants is very difficult due to the strong nonlinearity. The ensemble Kalman filter, which has a very good nonlinear analytical ability due to its probabilistic approach, is applicable to the case.

Referring to FIG. 4B, when a measured value such as the effective value voltage, frequency, effective, and reactive power is received for the system on the main grid, a simulation of the PB generator including the estimated target constants determined based thereon but with the input data to the output of the PB generator in the grid generate simulation results (for example, voltage, frequency, active and reactive power, etc.), given the up from a predetermined starting time (t _0), end time (t _f) the n samples of the interval (in hours) (△ t) when divided by _{(t n = t 0 ~ t} n = t f), and starts to perform the simulation in the first place the sample interval (S21). A statistical analysis such as a sensitivity analysis for determining the estimation target constants as the model constants of the generator G is performed (S22) using the ensemble Kalman filter method or the like with respect to the simulation result (S22) (S23). This process is repeatedly performed for all the sample intervals (t _n = t ₀ to t _n = t _f ) (S 24, S 25). That is, after systematic simulation is performed for one sample period, a plurality of virtual samples of probable target constants are generated, and a system simulation is performed for each virtual sample. In this process, hundreds of virtual samples can be generated and many (eg, hundreds) of systematic interpretations are involved per analysis of the original data-based system received.

After the system simulation for each virtual sample is performed, the most likely new virtual sample is statistically obtained, and the virtual samples of the estimated target constants are modified based on the statistical result. More. If the estimated result is sufficiently small compared with the measured value, the estimated estimated constants to date are determined to be the model constants of the generator G and updated (S27). However, like other Kalman filters for nonlinearity, Since the ensemble Kalman filter does not guarantee that the final convergence value is the optimal value, if it is determined that the error is not a sufficiently small error through comparison of the final results (S26), all the above processes must be repeated (S27).

The integer estimates of the generator (G) are very nonlinear, and the various constants of the generator (G) exhibit various characteristics in combination. That is, there may be a plurality of combinations of integers that can accurately represent measurement data. Finding a global optimal value for a problem involving multiple local optimal points is defined as a global optimization problem and can be solved by global optimization techniques such as Genetic Algorithm, Particle Swarm Optimization, and Gravity Algorithm. The above global optimization technique requires much more operator resources than local optimization, and the possibility that the derived values become the global optimal values increases as the operator input amount increases. Therefore, it is not suitable for estimating the number of generators of a nationwide scale, and it is difficult to reflect the program analysis program of the actual generator characteristics.

 Since the ultimate goal of estimating the generator constant is to accurately reflect the characteristics of the actual generator, it is not important whether the combination of the constants of the generator is the same as the actual one. Therefore, it is reasonable to apply the local optimization technique to the generator constant estimation. Also, in general, the constant of a generator does not change significantly in a short period of time, so that data measured within a certain period of time shows a consistent result. That is, as the measurement data accumulates, the characteristics of the generator must satisfy various conditions, and each of the estimated generator constants to satisfy all the data approaches the actual generator constant.

The Levenberg-marquardt algorithm (LM algorithm), which is one of the most powerful local optimization methods, was applied first to estimate generator constants. The algorithm operates according to [Equation 1]. That is, the Jacobian matrix J for the sensitivity is calculated based on the difference between the measured value y according to the correction amount ? Of the estimation target constants ? And the simulation result data, and the Jacobian matrix J for the sensitivity is calculated. And the estimation target constants ? That are variables of the objective function f are determined so that the difference between the simulation result data is minimized and the following equation is satisfied.

[Equation 1]

Where the matrices, J and I, are a Jacobian matrix and an identity matrix for sensitivity, respectively. J ^T is the transpose matrix of J. The vector β (β _P , β _Q ) is a group of parameters that determine the objective function (f ()) (eg, values associated with input / output voltage / current, valid / reactive power, gain, And a vector delta consisting of the amount of change in parameters (correction amount), and is updated every time the iteration is performed to optimize the objective function. That is, the vector β is the measured value (y) and the simulation result data (e. G., Voltage, frequency, real power (P) and reactive power (Q), etc.) in calculating the sensitivity (J) as the difference between the measured value (y) The simulation results are determined to minimize the difference between the data. The relatively small or large adjustment coefficient μ reduces the error by the Gauss-Newton or gradient descent method of the LM algorithm and improves the convergence speed and prevention of divergence in the sensitivity calculation process. Each parameter of the LM algorithm is determined by [Equation 2] - [Equation 5]. That is, the variation amount of the root mean square error (RMSE) with respect to the effective power P and the reactive power _Q with respect to the respective parameters β _P and β _Q in the active power P and the reactive power Q is calculated , The parameters β _P and β _Q related to the generator (G) and its associated governor, PSS (Power System Stabilizer), exciter and the like can be estimated.

&Quot; (2) &quot;

&Quot; (3) &quot;

&Quot; (4) &quot;

&Quot; (5) &quot;

In this way, it is possible to target all the values to the integer estimation, but in this case, the computation amount is very high and it is not possible to confirm the success of the estimation. Therefore, it is important to select the main parameters appropriately. For example, for IEEEG1 in the generator (G) model, the main constants of the governor are K and T5, and the major constants of the exciter (eg ESST4B) are Kpr, Kir, and Kp. The main integer of PSS2A (power stabilizer) is Ks1. Based on this, Equations (2) to (5) can be rewritten as Equations (2) to (5). Such a generator (G) model constant is an example, and an appropriate constant can be selected according to a simulation purpose or a development environment.

&Quot; (6) &quot;

&Quot; (7) &quot;

&Quot; (8) &quot;

&Quot; (9) &quot;

A Gauss-Newton algorithm can be applied in a similar manner, which operates on Equation (10).

&Quot; (10) &quot;

<4. Estimation of model constants based on multiple measurement data>

When the measurement is continued, a plurality of data acquired at various points of time exist. Since the generator (G) constants estimated at a specific point in time can not be regarded as representative of characteristics at all points in time, Should be.

5 is a diagram for explaining an example of a model constant estimation and representative integer selection method in the present invention.

In the simplest way, there is a way to estimate the integer by setting the RMSE for all the data as an objective function. (Estimation target constants / model constants) from a plurality of measured values Data (for example, four in the drawing) as in 510 of FIG. 5, but in the next estimation, a part of the previously used measurement values ), The method of repeating the estimation while adding the newly measured value may be simple. However, as the measurement data increases, there is a disadvantage that the number of operators required for estimation increases. Also, when new data are measured, all the data used in the existing estimation must be used again with the new data, so that it takes a lot of space to store the measurement data and many operators must be mobilized in each estimation. Even when the estimation of the old data is not performed due to the storage space problem or the accuracy of the generator constant, the calculation resources are very much used as the old data must be deleted and the remaining data must be estimated again. In addition, this method has a fatal problem in that, when the generator constants satisfying the respective measured values are placed in directions opposite to each other, all the opposite measured values are not reflected and an incorrect integer is derived.

As a method of drastically reducing the computational resources and the storage space, there is a method of estimating parameters (estimation target constants / model constants) separately for each measured value as shown in 520 of FIG. 5, In the next estimation, there is a method of repeating the estimation while adding the estimated parameter from the newly measured value except for a part of the parameter used previously (for example, one). The effect of this method depends on how to select effective data selection and representative values, and statistical techniques can be used as the simplest method.

FIG. 6 is a diagram for explaining a method of using statistical techniques for model constant estimation in the present invention. FIG. 6 shows an example in which ten ordered pairs each consisting of two model integer pairs represent estimated values in a relative positional relationship.

If the ordered pair representing the average of the individual integers is the same as the point 610, this can lead to a problem that results in an erroneous result that does not represent any integer pair of ten. The ordered pair having the mode of each integer is represented by point 620, which also does not represent any integer pair. 6, only two integer pairs of 10 integer pairs are concentrated in a specific area, and the remaining 8 integer pairs are sporadic, so that 8 integer pairs are removed and representative points 630 of two concentrated integer pairs are selected It would be more reasonable to do so. The rationality of this choice will be increased if there are a large number of estimated integer pairs.

When the estimated integers are sporadic, it is very difficult to establish a criterion for selecting the integer pairs and the integer pairs to be used in the representative value selection. Therefore, it is practically impossible for a person to directly classify the data and extract representative values. In addition, when the concept of FIG. 6 is introduced, since the ordered pairs are listed in the total six-dimensional space, the above-mentioned estimation target constants are six (K, T5, Kpr, Kir, Kp and Ks1) Work.

In order to select any one of the estimated values for the estimation target constants as the representative value, a method of solving it by the gravitational field theory will be described as follows.

7 is an example of a three-dimensional distribution of the sum of the gravitational fields generated by each point when there are arbitrary 10 ordered pairs in the two-dimensional plane. As a way to solve very difficult data selection problems, we can introduce the gravitational field theory that exists in nature. When there are arbitrary ten ordered pairs (corresponding to the estimation target integers) in the two-dimensional plane, the sum of gravity fields generated by each point can be distributed as shown in FIG.

FIG. 8 is a diagram showing the distribution of the sum of the gravitational fields of FIG. 7 by contour lines. FIG. In Fig. 8, the average values of arbitrary points (a green point 810), a global optimal point (black asterisk), a global optimum point (black cross 830) Blue X 850), and the actual optimum point (red cross 840).

Since a strong gravitational field is formed in a place where many points are gathered intensively, a value corresponding to the strongest point of the gravitational field (a value of the estimation target constant) can be selected as a representative value. That is, although there are six points on the lower side with respect to the vertical axis 0 in FIG. 8, they exist sporadically and do not produce a strong gravitational field. On the other hand, the upper four points have few absolute numbers, but they exist intensely in one place, creating a strong gravitational field. Therefore, this is suitable as a method for solving the problem mentioned in Fig.

Fig. 9 is a contour map showing the distribution of the sum of gravity fields for a large number of ordered pairs composed of two model integer pairs. In the case of Fig. 8, it is also possible to directly classify by a person, but it is very difficult for a person to directly classify if there are many ordered pairs as shown in Fig. However, the representative value can be selected without difficulty by the proposed technique.

Since there are several pole values in the curved surface shown in Fig. 7, there are several local optimum points. Therefore, global optimization algorithms such as genetic algorithm, particle swarm optimization and gravity algorithm should generally be applied to find the global optimal point. However, as mentioned above, the global optimization algorithm has a problem in that it requires a lot of operator resources and an increase in operator input to increase the reliability of the result.

The data classification and representative value selection method using this gravity field theory solves the problem of applying this global optimization algorithm. In other words, the gravitational field appears strongly at each point, so there are many points around the optimal point. Thus, starting from each of the points producing the gravitational field and performing a local optimization, each point converges to a local optimal point existing around it. Therefore, one or more points are converged on the individual local optimum points, and the largest number of points converge on the global optimum point (representative value). Therefore, by comparing the intensity of the gravity field possessed by the local optimum point and selecting the global optimum point as the highest point, the global optimum point can be selected and used as a representative value.

Local optimization can be applied to all local optimization algorithms such as Levenberg-Marquardt (L-M), Gauss-Newton (G-N) and Gradient descent, and L-M or G-N uses [11] and [12].

&Quot; (11) &quot;

&Quot; (12) &quot;

At this time, the variables exist in the six-dimensional space, and when integers for the effective and reactive power are separated as shown in [Equation 13] and [Equation 14], they exist in the two-dimensional plane and the four-dimensional space, respectively.

&Quot; (13) &quot;

&Quot; (14) &quot;

FIGS. 10, 11, and 12 show an ordered pair of distributions, contour lines, and reactive power-related constants of the estimated model constants K and T5 based on the measured data based on any generator.

For example, the distribution of model constants K and T5 related to the active power among the constants estimated on the basis of the measured data in a predetermined generator using a PMU or the like makes the gravity field as shown in FIG. 10, do. In Fig. 11, the average value (blue x 920) of an arbitrary point (green dot), the local optimal point (black star), the global optimum point (black cross 910) 930).

The green point is an integer ordered pair estimated from the effective power of each measurement data, and the representative value 910 is located at the most concentrated point. Whereas the mean value 920 refers to the wrong spot where there are no nearby points. Order pairs of reactive power-related constants appear as shown in FIG. 12, and representative values indicated by red crosses indicate places where ordered pairs are dense.

It is confirmed that the generator model constant estimation method based on the actual data of the present invention as described above is effectively applied to the simulation results of other power plants as well.

In the method of estimating generator model constants in an electronic device such as a computer according to an embodiment of the present invention, the functions used for input / output data processing are implemented as computer-readable codes on a recording medium readable by an apparatus such as a computer A combination of such a recording medium and a computer can be used to input or output data and information necessary for performing a function. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, a hard disk, a removable storage device, And the like. In addition, the computer-readable recording medium may be distributed in a computer system connected to a network (e.g., the Internet, a mobile communication network, or the like) and may store computer readable codes in a distributed manner and may be executed through a network.

As described above, according to the method for estimating the generator model constant according to the present invention, it is possible to estimate and adjust the generator model constant so that the dynamic characteristics of the generator when the system failure occurs in the actual system operation conditions and the dynamic characteristics of the generator, . The simulation results are compared with the results obtained by extracting PMU data related to the system from the results of the simulation. By adjusting the generator model constants to be similar to the PMU data, simulation results closer to the real system can be obtained. The system stability simulation is the most basic simulation in planning and operating the power system. If the reliability of the simulation is raised to reflect the characteristics of the actual system in the simulation, it is possible to obtain more accurate and reliable stability analysis results As a result, it becomes the base technology of stable system planning and operation, so that it can be used accurately and easily for establishment of measures for system failure.

As described above, the present invention has been described with reference to particular embodiments, such as specific elements, and specific embodiments and drawings. However, it should be understood that the present invention is not limited to the above- Those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the essential characteristics of the invention. Therefore, the spirit of the present invention should not be construed as being limited to the embodiments described, and all technical ideas which are equivalent to or equivalent to the claims of the present invention are included in the scope of the present invention .

Claims (11)

(A) receiving a measurement including an instantaneous voltage and current for each generator using a measurement device installed on a power system including at least one generator connected to the power grid via a transformer;
(B) generating simulation input data including a plurality of estimated object constants determined based on the measured values for a simulated reduction system for simulating the output of the simulated regenerative power generator on the power grid using a predetermined program, Generating simulated result data for an output of the simulated regeneration generator by performing simulation on the simulation output data; And
(C) re-executing the simulation after correcting at least one of the estimated constants based on the sensitivity analysis on the estimated constants, and determining model constants for the generator according to the error
And estimating a generator model constant. The method according to claim 1,
If the error between the measured value and the simulation result data is smaller than a threshold value, estimating target constants are determined as model constants for the generator. The method according to claim 1,
A step of performing the steps (A) and (B) for the predetermined start time to ending time in the step of generating the simulation result data and re-executing the simulation according to the error, or
(A) and (B) are performed for each sample interval by dividing the start time to the end time by a plurality of sample intervals, wherein the plurality of sample intervals A method of re-executing the simulation for the entire section
The method estimating the generator model constant. The method according to claim 1,
Wherein the measured value is data for a primary side or a secondary side of the transformer connected to the generator. The method according to claim 1,
Wherein the measured value includes an effective value voltage, a frequency, an effective power, and an ineffective power converted from the instantaneous value voltage and current for the generator. The method according to claim 1,
Wherein the central server on the network receives the measured values from the measurement device on-line to perform the simulation, modify the estimated constants, and determine model constants for the generator. Way. The method according to claim 1,
The model constants for the plurality of generators are simultaneously determined for the plurality of generators through the time synchronization, and the model constants for the plurality of generators are simultaneously determined. The method according to claim 1,
Calculating a Jacobian matrix J for the sensitivity on the basis of the difference between the measured value y and the simulation result data according to the correction amount delta of the estimation target constants beta , Determining the estimated constants ? That are variables of the objective function f such that the difference between the measured value y and the simulation result data is minimized and the following equation is satisfied,

Wherein I is an identity matrix, and μ is an adjustment coefficient for preventing divergence and improving convergence speed in the sensitivity calculation process. The method according to claim 1,
In determining the model constants,
In the next estimation, a method of estimating a parameter, which is the estimation object constants, from a plurality of measurement values while adding a new measurement value as much as a part of the previously used measurement values, or
In the following estimation, except for a part of the parameters used previously, a parameter is selected from a plurality of parameters estimated after separately estimating the parameter from the measured values while adding the estimated parameter from the new measured value,
Is used for estimating the generator model constants. The method according to claim 1,
In determining the model constants,
Calculating a distribution of a sum of the gravitational fields with respect to a pair of ordered pairs of the estimated object constants to select one of the estimated values for the estimated constants as a representative value, And a representative value of an ordered pair of integers. (A) receiving a measurement comprising an instantaneous voltage and current for each generator using a measurement device installed on a power system comprising at least one generator connected to the grid via a transformer;
(B) generating simulation input data including a plurality of estimated object constants determined based on the measured values for a simulated reduction system for simulating the output of the simulated regenerative power generator on the power grid using a predetermined program, To generate simulated result data for the output of the simulated regeneration generator; And
(C) a function of re-executing the simulation after correcting at least one of the estimated constants based on the sensitivity analysis on the estimated constants, and determining model constants for the generator according to the error
And estimating the model constants of the generators on the power grid system.

Images