JP3472697B2 - Power system frequency stabilizer - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a power system frequency stabilizer for a power system, and particularly to a frequency in the future after the observation time by observing the frequency at multiple points in the case of system separation. The present invention relates to a power system frequency stabilizing device that performs control for frequency maintenance by predicting changes.

[0002]

2. Description of the Related Art In the operation of a power system, when a power line route disconnection occurs due to a transmission line accident or the like, a frequency deviation occurs in a separated system due to supply and demand imbalance. In order to eliminate this and stabilize the frequency fluctuations in the isolated system, it is necessary to correct the supply and demand imbalance in the isolated system, and therefore the power supply or load is forcibly cut off. For this purpose, a frequency stabilizer has been conventionally installed.

However, since it is difficult to directly grasp the amount of supply and demand imbalance by observation, the current system estimates the supply and demand imbalance based on the observation of frequency.

The elimination of the supply / demand imbalance amount estimated from the observed frequency is considered as a control target, and control having a corresponding control amount is executed to maintain the supply / demand balance and stabilize the frequency. As a general stabilizing device which is a conventional technique, Japanese Patent Laid-Open No. 2-266831 is known.

[0005]

In the conventional stabilizing device, a method of calculating a control amount using a settling value from a frequency calculated from a bus voltage or a rate of change thereof, or a fluctuation equation and an operation of a generator control system In addition to the frequency response formula theoretically created based on the formula, methods that predict the convergence destination of the frequency response are adopted.In these methods, the settling value is prepared in advance corresponding to the system state. Therefore, it was necessary to use this set value.

However, in these methods, the error between the settling value and the true value causes an error in the prediction of the frequency response convergence destination, so that the method using such a setting value has a drawback. Further, when the stabilizing device controls, the system state changes due to the control, so the frequency response changes in the middle and causes a prediction error. In particular, the smaller the isolated system, the greater the influence of the control on the frequency characteristics of the system. When adding control in multiple stages, parameters that depend on system characteristics must be set in advance and enormous processing is required.

Further, in the above-mentioned known example, it was assumed that there was no frequency deviation at the time starting point of the frequency response. Therefore, if the frequency has a deviation from the rated frequency at the start of frequency response, a prediction error will occur. Especially, when the system state changes in the middle of frequency response due to the control performed by the device, it is a remarkable example of the above case, and when the correction control becomes necessary after the device execution control, it is erroneous. There was a problem that control was carried out.

[0008]

In order to achieve the above object, the present invention relates to a power system frequency stabilizer for stabilizing the frequency of a power system, and when an accident occurs in the power system, it is separated by the accident. Based on the frequency time series data by the storage means for storing the observation frequency observed by the observation means as time series data To
The frequency response of the electric power system separated by the accident is formulated, and the arbitrary frequency prediction after the observation time is performed.

According to the present invention, a means for predicting the frequency response after the observation point is formed by formulating the estimated frequency time response that minimizes the evaluation of the error between the observed value and the estimated value for the time series data of the observation frequency. Is provided.

Further, the present invention is such that every time the frequency in the power system is observed in time series, the frequency response of the power system is mathematically formulated using the frequency time series data observed in the past. is there.

According to the power system frequency stabilizing device of the present invention, it is possible to perform the process of estimating the frequency response based on a large number of observed time series data of frequency, without using a prior parameter dependent on the system. Become. This is because the time-series data of the observation frequency can be used to predict the frequency at any point after the observation point by formulating the temporal frequency response. By this processing, it becomes possible to predict the frequency without preliminarily stabilizing the parameters depending on the system state.

Further, even when the frequency fluctuation does not start from the steady state, that is, when the frequency which is the starting point of the frequency response has a deviation from the rated frequency, the frequency after the observation can be predicted.

As a result, the impact of the accident that occurred, or
The effects of the controls that have already been implemented can be foreseen at their mid-course. Therefore, at a point in the middle of the process in which the frequency response continues, control of the frequency after the response convergence within an appropriate range is realized.

Further, since the control can be executed without waiting for the response to converge, a system for surely performing the necessary control in multiple stages becomes possible, and excessive control can be prevented, so that the separated system can be maintained. In terms of, the reliability of control is improved. Further, it is possible to expect economical loss control by surely reducing the control amount.

Since there is little dependence on the operating / operating states of the generators in the system and the parameters for pre-settling showing the characteristics of the load condition, the necessity of adjustment or repair work becomes low.

[0016]

DETAILED DESCRIPTION OF THE INVENTION An embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 1 shows an embodiment of a power system frequency stabilizing device for stabilizing the frequency of a separated system according to the present invention.

First, in the power system, the interconnection flow tides 102 at a plurality of assumed grid separation accident points 101 and points included in the separated grid 103 caused by the separation accident at the assumed grid points, for example, the lower side of the assumed grid point 101. Busbar 1
Observe 04 frequency. As a specific method of frequency observation, a method of measuring voltage is used.

When system separation occurs, a large supply-demand imbalance occurs in the separated system.
Generally, it corresponds to the tidal flow rate of the interconnection power flow 102 immediately before separation. Therefore, immediately after the occurrence of the separation, the tidal component correcting means 105 outputs the monitoring cutoff command 106 of the generator or the load having the content corresponding to the tidal flow of the interconnection line tidal flow 102, and generally maintains the supply and demand balance in the separated system. To do.

Further, the frequency multipoint observing means 107 determines the frequency of the system, which changes as a result of this control, the result of the autonomous control system of the generator, the influence of load drop and recovery, at multiple points in time. Observe the system frequency. The result is stored in the frequency time series data storage device 108.

Next, an embodiment of the frequency multipoint observation means is shown in FIG.
Will be explained.

The detailed means of the frequency multipoint observing means 107 is performed, for example, by the procedure shown in FIG. When the frequency is calculated from the voltage observation value as in the embodiment, the process of acquiring the frequency when necessary is performed from the voltage observing means 21 which regularly repeats the observation.

First, start observation of the start time of time series data
It is set by the time setting 22. Then, the observation time determination 23
It was decided by the
If the frequency is measured, the frequency observing means 24 observes the voltage.
Frequency acquisition from the means 21 to store frequency time series data
It is stored in the device 108. Then, in the next observation time measurement 25
Therefore, the next scheduled observation time is set. And
Whether the required amount of time series data has been acquired Observation end judgment 2
Judgment according to 6, and as a result, if insufficient, observe
From the time judgment 23, the frequency observation process is continued and necessary
When the observation is performed, the frequency multipoint observation means 107
finish.In addition, in the above-mentioned one embodiment,
The frequency of the lower bus 104 is detected as one location
Detects frequencies from more than one location in the grid.
Therefore, it is also possible to perform frequency observation.

Based on the frequency time-series data stored in the storage device 108 acquired by the frequency multi-point observation means 107 as described above, the frequency response prediction means 109 approximates the temporal change of the frequency. Express with a mathematical formula.
The major procedure of the frequency response predicting means 109 is to prepare a template of a frequency response equation and to perform a calculation based on this to minimize the error between the estimation equation and each observed value. An example of the calculation will be described below with reference to FIG.

First, the estimation formula template creating means 31 creates a template of the estimation formula. As a template, formulas that represent characteristics that can be qualitatively estimated are prepared. For example, many of the control effects of the generator governor and the frequency components of the load have a strong first-order lag characteristic with respect to frequency fluctuations, so it is conceivable to use the first-order lag equation as a template. Therefore, as a model of this estimation formula, (1) that simulates the first-order delay that is the most main mode is used.

[0026]

[Equation 1]

Consider the following. This (Equation 1) is shown in FIG.
As shown in (a) and FIG. 5 (b), the frequency response f (t) is a mathematical expression in which the gain J, the offset K, and the time constant τ have a degree of freedom. Further, by providing an offset K as this mathematical expression, it becomes possible to cope with frequency response from a state where there is a deviation in frequency. Since these values depend to some extent on the grid, these values can be specified if the grid is actually operated. This is given as an initial value and used as a template for the estimation formula.

As another template, a method using a mathematical expression in which a plurality of the first-order lag terms described above are combined or an n-order expression is used can be considered.

Subsequently, the estimation formula error calculating means 32 calculates the error generated between the estimation formula and the observed value to calculate the validity of the estimation formula. Since the estimation formula is created by iterative calculation, each parameter at the number of iterations n is described with a subscript n.

[0030]

[Equation 2]

[0031]

[Equation 3]

The evaluation function G of the estimation formula is (Equation 2). However, the function g in (Equation 2) is (Equation 3). In these equations, tk is the time-series data k-th time data, Δ
fk means the frequency deviation data of the kth time series data.

Based on the above formulation, by adjusting each parameter so as to minimize the evaluation function G shown in (Equation 2), the most appropriate estimation formula for the observed value can be obtained. This minimization process is performed by iterative calculation in which the estimation formula correction means 34 adjusts the parameters while the estimation formula convergence determination 33 determines that the evaluation by (Equation 2) is insufficient.

In the estimation formula correction means 34, in order to reduce the evaluation value of (Equation 2), (Equation 2) is applied to each parameter.
ΔJn, ΔKn, and Δτn obtained by partial differentiation of are calculated, and the parameters are updated as in (Equation 4). δ and m are convergence coefficients.

[0035]

[Equation 4]

According to the estimation formula convergence determination 33, the iterative calculation of (Equation 2) to (Equation 4) is performed until the evaluation function converges or a predetermined fixed number of times. Then, the estimated convergence determination 3
The final parameter when 3 is satisfied is the most appropriate estimation formula for the observed value. With this, the frequency response prediction means 109 ends.

Next, the frequency correction control means 110 predicts the frequency at a future time point from the frequency response estimation formula created by the frequency response prediction means 109. This is calculated by substituting the target future time into the frequency response estimation formula.

When the predicted frequency value does not satisfy the operating condition of the frequency, the necessary amount of supply / demand imbalance is calculated and the corresponding control command is output.

Next, the operation contents of one embodiment of the present invention will be described with reference to FIG.
Will be explained.

FIG. 2 shows the frequency deviation of the lower bus bar 104 in one embodiment of the present invention. Time t
_{At 0} , the emergency shutoff command 106 is issued as a control command, and at this time, some frequency deviation has already occurred.

Then, at times t _{1 to} t ₅ , the lower bus 1
After the frequency state 04 is detected by the frequency multipoint observation means 107 and stored as the frequency time series data 108, the frequency response prediction means 109 described above is used until the evaluation function converges from the data at the times t _{1 to} t _5. Alternatively, iterative calculation is performed up to a predetermined number of times to obtain an estimation formula.
When it is expected that the frequency deviation will reach the lower limit value at a future time according to the obtained estimation formula, further calculation for eliminating the necessary supply / demand imbalance is performed, and at time t ₆ , a control command corresponding to this calculation is issued. It will be output and will move to a stable frequency deviation state.

Further, as in the above-described embodiment, not only the operating condition that the predicted frequency at a future time point is within a certain range is checked, but also the frequency at a plurality of time points is predicted and the rate of change thereof is used. As a condition, a method of judging the necessity of control may be considered.

Further, at times t _{1 to} t ₅ , when there are three points
It is possible to obtain an estimation formula if there is time data.
Therefore, data measured at the past time for each of these measurement times
Using, for example, at time t ₄ , data at times t _{1 to} t ₃
Use to calculate the estimation formula. Furthermore, this kind of measurement
By performing after these times, for example, at time t ₁₀
Using data at time t ₁ ~t ₉ to the time the time t ₂₀
estimation equation is obtained by using the data of t ₁₁ ~t _19. like this
By calculating the future frequency deviation at each time,
Each time the future frequency deviation situation of the power system elapses,
It is possible to calculate in more detail.
The condition such as the rate of change of frequency deviation at a specific time in the future
As a method to judge the necessity of system control,
It

In the embodiment of the present invention, when it is judged that the control is necessary, the method of calculating the control amount is to use the frequency characteristic 51 of the system as shown in FIG.
This is realized by obtaining the supply / demand imbalance amount 53 corresponding to 2.

Here, the frequency characteristic 51 of FIG. 6 shows a one-to-one correspondence relationship between the static frequency value after the response is converged and the supply and demand imbalance amount. Although the observed frequency is a dynamic frequency, the static frequency characteristic can be utilized by predicting the static frequency after the response converges from the frequency response estimation formula created by the frequency response prediction means 109. It is possible.

The frequency correction control means 110 based on the frequency multi-point observation means 108 and the frequency response prediction means 109 described above is completed by the frequency stability judgment 111, or the fixed time elapses. Repeat until.

Further, in the present invention, by installing the frequency stabilizing device described above in the power system, the frequency at the future point can be predicted in the middle of the frequency response, so that unnecessary control can be prevented. Moreover, since the effect of the implemented control can be foreseen, it becomes possible to add control in multiple stages in a short time. Therefore, it can be expected to prevent excessive control. This improves the certainty of control in maintaining the separated system. In addition, it is expected that economic loss can be suppressed by reducing the control amount.

Further, as another embodiment of the present invention, a frequency stabilizing device will be described in which a means as shown in FIG. 7 is adopted for the frequency response predicting means 109 in the above embodiment. As an outline of the frequency response predicting means of FIG. 7, in estimating the frequency response, a function of removing abnormal observation points from a plurality of observation points and performing prediction is added to the frequency response prediction means of FIG. It was done.

Based on the estimation formula created in steps 31 to 34 similar to FIG. 4, the observation point filter means 61 evaluates the estimation error between the estimation formula and the observation value for each observation point of the frequency time series data 108. Then, the processing for storing the data obtained by removing the observation points where the abnormality is judged is stored in the frequency time series processed data 62. Details of this processing are performed as shown in FIG.

First, the individual observation point evaluation means 71
The evaluation value according to (Equation 3) is calculated for all the observation points and stored as the individual observation point evaluation data 72. Next, the evaluation value distribution evaluation means 73 calculates the distribution of the evaluation values with all the observation points as a population, and only the observation points whose evaluation values are within a certain range with respect to the variance are used as the frequency time series processed data 62. Remember.

Based on the frequency-time-series processed data 62, the estimation formula error computing means 63, the estimation formula convergence determination 64, and the estimation formula correction means 65 perform the same calculation as the processing of FIG. Make a prediction.

Through the above processing, the frequency response prediction means 10
9. In addition to the above-described embodiment, in 9, the influence of noise and the like in the observation is removed, and the most main fluctuation is extracted when a state change occurs in the frequency response process observed due to load recovery or the like. Then, it becomes possible to estimate this.

[0053]

As described above, according to the present invention, it is possible to realize the control for keeping the frequency after the response convergence within the proper range at the midpoint of the process in which the response of the frequency continues. Since the control can be performed without waiting for the response to converge, a system that surely performs the required control in multiple stages becomes possible. As a result, excessive control can be prevented, so that the reliability of control is improved in terms of maintaining the frequency of the separated system. Further, it is possible to expect economical loss control by surely reducing the control amount.

Furthermore, since there is little dependence on the operational / operating states of the generators in the system and the parameters for pre-settling showing the characteristics of the load condition, the necessity of adjustment or repair work becomes low.

[Brief description of drawings]

FIG. 1 shows an embodiment of a power system frequency stabilizing device of the present invention.

FIG. 2 shows an example of frequency deviation shift by the power system frequency stabilizing device of the present invention.

FIG. 3 is a schematic diagram of frequency multipoint observation means of the frequency stabilizing device.

FIG. 4 is a schematic diagram of frequency response predicting means of the frequency stabilizing device.

FIG. 5 shows an example of a frequency response estimation formula template.

FIG. 6 shows an example of frequency characteristics.

FIG. 7 is a schematic diagram of observation point filtering frequency response prediction means of the frequency stabilizing device.

FIG. 8 is a schematic diagram of observation point filter means.

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Tomoji Nakamura Inventor Tomoji Nakamura 1-1-1, Kokubun-cho, Hitachi City, Ibaraki Hitachi Co., Ltd. Kokubun Factory (72) Inventor Mototaka Tsubouchi 2-chome, Suehiro-cho, Kariya City, Aichi Prefecture Machi 13-7 Chubu Electric Power Co., Inc. Kariya Electric Power Center (72) Inventor Masamichi Kojima, Daido Higashi 7, Tozamachi, Okazaki City, Aichi Prefecture Chubu Electric Power Co., Ltd. Okazaki Branch (56) Reference JP 60-194729 JP, A) JP 10-32927 (JP, A) JP-B 7-108063 (JP, B2) (58) Fields investigated (Int.Cl. ⁷ , DB name) H02J 3/00-5/00

Claims (8)

(57) [Claims] 1. A power system frequency stabilizing device for stabilizing a frequency of a power system, wherein when an accident occurs in the power system, a plurality of frequencies in the power system separated by the accident are observed in time series. Possible observation means, storage means for storing the observed frequencies observed by the observation means as time series data, and frequency response of the power system separated by the accident based on the frequency time series data by the storage means. Is derived as an equation including a first-order lag term with respect to time, and an arbitrary frequency prediction after the observation time is performed based on the derived approximation equation. 2. A power system frequency stabilizer for stabilizing the frequency of a power system, wherein when an accident occurs in the power system, a plurality of frequencies in the power system separated by the accident are observed in time series. Possible observation means, storage means for storing the observed frequencies observed by the observation means as time series data, and frequency response of the power system separated by the accident based on the frequency time series data by the storage means. Is a number including the n-th order expression (n = 1, 2, ...) for time.
Derived as the expression, the power system frequency stabilizing apparatus which is characterized in that any frequency estimate since observation time based on the derived approximate expression. 3. The power system frequency stabilizing device according to claim 1 , wherein each time a frequency in the power system is observed in a time series, the frequency time series data observed in the past is used. An electric power system frequency stabilizing device, characterized in that an approximate expression is derived for a frequency response of the electric power system. 4. Based on the estimated frequency response according to claim 1 or 2 , if the predicted frequency value deviates from a certain range by predicting the frequency value at a certain time point, A power system frequency stabilizing device, characterized in that control is performed to eliminate supply imbalance based on the power system. 5. The frequency response is measured by controlling only a necessary amount based on the predicted frequency value according to claim 4 , deleting the accumulated frequency time series data, and then observing the frequency in time series again. A power system frequency stabilizing device comprising means for performing control in the next step based on the result of the estimation, and performing the control stepwise by repeatedly performing the means. 6. The power system frequency stabilizing device according to claim 1 , wherein when deriving an approximate expression for the estimated frequency response based on the time-series data of the observed frequency, the frequency response estimation expression is used. A power system frequency stabilizing device, characterized in that it has a template. 7. The electric power system frequency stabilizing device according to any one of claims 1 to 6, for the estimated frequency response based on the time series data of said observation frequency, deriving an approximate expression In addition, the dominant reaction part is extracted from the time-series data of the observation frequency by evaluating the error between the estimation formula and the observation value at each observation point and eliminating the observation points that have an estimation error above a certain level. A power system frequency stabilizing device characterized in that the frequency response is estimated by. 8. A electric power system frequency stabilizing device according to any one of claims 1 to 7, for the estimated frequency response based on the time series data of said observation frequency, deriving an approximate expression In addition, when the error between the estimation formula and the observation value is evaluated for each observation point and the accumulated estimation error exceeds a certain level, the accumulated time series data of the observation point is deleted and a new observation is made after that point. An electric power system frequency stabilizer, which estimates frequency response from frequency time series data.