KR100824908B1 - A Sync Detection Method of Real Time and Non-real Time Signals in Dynamic Systems - Google Patents

Abstract

The present invention relates to a method for detecting a synchronization occurring in a dynamic system such as a power system in real time or non-real time. Since the vibration generated in a system in normal operation is attenuated by a constant braking coefficient and a time constant, the output signal of the system is It is expressed as the sum of several attenuation index functions. Therefore, in estimating the parameters included in the multiple signals and estimating the critical mode and residual included in the time series data, and comparing them, searching for tuning between the output signals is similar in the critical vibration mode included in the signal. By selecting a mode with frequency, a braking coefficient, and a time constant, and searching for tuning by comparing the magnitude and phase of these flows, tuning can be detected between signals having the same characteristics, thereby distinguishing system characteristics. By calculating global system oscillation, which can consider vibration as a vibration, it is possible to prevent the collapse of dynamic systems in advance, thereby reducing the social and economic cost of losses.

Description

A Sync Detection Method of Real Time and Non-real Time Signals in Dynamic Systems

1 is a schematic block diagram of hardware for carrying out the present invention;

2 is a flowchart illustrating an operation process of the present invention;

3 is a time series data waveform and vector diagram for tuning.

4 is a time series data waveform of an embodiment

5 is a vector diagram of flowing water of an embodiment

6 is a tuning detected in an embodiment of an actual dynamic system.

7 is a flow vector separated by frequency in an embodiment of a real dynamic system

<Explanation of symbols for the main parts of the drawings>

10: data input unit 20: parameter estimation unit

30: parameter selection section 40: tuning detection section

50: data output storage and transmission unit

The present invention relates to a method of detecting a synchronization phenomenon occurring in a dynamic system. A sync is a phenomenon in which two events occur at the same time repeatedly for a certain period of time. It must consist of a large number of oscillators, each of which must be of equal magnitude so that the phases are connected to each other to control the characteristics of the system. For example, because all devices in a power system operate at a constant frequency, the synchronous generator must always operate at a constant frequency, that is, each generator must stay in sync with each other. The controllers are operated so that all power devices included in the are operated at the same frequency. Synchronism usually considers only the frequency, but sync considers not only the frequency but also the phase of the signal. Therefore, the low frequency below 2.0Hz, which can consider the phase, is the main frequency of interest. Until now, the problem of low frequency has been mainly dealt with as a problem of microsignal stability. To solve this problem, we developed a model of dynamic system and solved the mode with the smallest braking coefficient through mathematical linearization process, eigenvalue, and eigenvector calculation. Major developments have been made to find and improve. However, the eigenvalue analysis using the linear model analyzes the characteristics of the system around a specific operating point without using the discrete signals actually acquired, so the concept of tuning is not applicable and discrete signal data is not used. It is impossible to interpret and the concept of global system oscillation is not applicable.

In detecting the synchronization of the dynamic system, which is impossible using the conventional method, since the vibration generated in the dynamic system is a discrete signal that is attenuated by a constant braking coefficient and a time constant, an important mode included in the discrete signal by applying the parameter estimation method ) And residuals are compared and compared to detect tuning between the output signals. In the important vibration mode included in a plurality of signals, a mode having a similar frequency and a braking coefficient is selected, and a tuning index is calculated by comparing the magnitude and phase of the flowing water to detect the tuning occurring in the dynamic system. The system detects the total vibration of the dynamic system directly from the synchronized time series data acquired in real time and non real time.

[Hardware for Carrying Out the Invention]

1 is a schematic block diagram of hardware for carrying out the present invention. As can be seen from the drawings, the hardware for carrying out the present invention includes a real-time data acquisition unit 210 and a stored discrete data input unit 220. ), An input processor 230, a calculator 240, a data outputter 250, a data transmitter 260, and a data storage 270.

The input processor 230 may analog-to-digital convert real time series data of an analog type to be input to the calculator 240 or input stored discrete time series data to the calculator 240. Enter it.

The calculator 240 estimates a parameter for time series data input from the input processor 230, calculates a tuning index, and detects tuning between signals of the dynamic system.

The data output unit 250 outputs the parameters calculated by the calculator 240, the tuning index, and the detected tuning state to the output device.

The data transmission unit 260 transmits the parameters calculated by the operation unit 240, the tuning index, and the detected tuning state to the remote device (server) through the established network.

The data storage unit 270 stores the parameters calculated by the calculator 240, the tuning index, and the detected tuning state in the storage device.

FIG. 2 is a block diagram schematically illustrating an operation process for performing the present invention. As can be seen with reference to the drawings, the operation process for performing the present invention includes a discrete time data input unit 10 and a parameter estimating unit ( 20), a selection unit 30 for selecting important parameters, a detection unit 40 for detecting tuning in the selected parameter, and a portion 50 for outputting, storing and transmitting the detected tuning result.

Hereinafter, specific details of the present invention performed through the hardware configured as described above will be described.

The present invention relates to a method for acquiring information for interpreting and controlling a dynamic system in an output signal of a dynamic system. The dynamic characteristics important for the dynamic system are mainly influenced by vibrations generated at a low frequency of 2.0 Hz or less. do.

For example, in the steady state of the power system, except for transients, vibrations due to the dynamic characteristics of generators occur mainly at low frequencies below 2.0 Hz. Such oscillation causes the power system to collapse. It leads to As a result, the dynamic customer is out of power, resulting in huge industrial and economic losses. Therefore, in order to analyze and control the vibration phenomenon in the power system, there is a need to prevent the collapse of the power system in advance by detecting tuning included in a plurality of output signals. .

In any system, the measured or calculated discrete data can be expressed as the sum of the functions of the product of the attenuation index function and the sinusoidal signal. If the amplitude and phase of the i-th cosine function in the signal y (t) are A _i and φ _i , respectively, and the attenuation constant and frequency are α _i and ω _i , respectively, they are expressed as follows.

If two signals y _1, if the k- th dominant mode in the y _2, that is the leading signal (residue) corresponding to the frequency ω _k A _k is the largest, is smaller the braking coefficient α _k, two signals are as follows: Can be represented.

Where φ ₁ and φ ₂ are the phases of the two signals.

Since these two signals have the same magnitude, frequency, and braking coefficient, they have the same vibration characteristics. If the phase difference φ ₁ - φ ₂ of the two signals is 180 °, the two signals are shown in Fig. 3 (a), and when the phase differences are 90 ° and 0 °, they are shown in Figs. 3 (b) and 3 (c), respectively. . 3 (a) is referred to as negative sync (-sync), and 3 (b) and 3 (c) are referred to as quasi sync (Qsync) and positive tuning ( positive sync, + sync).

Tuning treats frequency and phase of a signal as important variables, but considering the stability of a system in a complex dynamic system, not only frequency and phase but also the braking coefficient and flow of the signal are important factors. Therefore, when dealing with tuning in a dynamic system, it is necessary to consider frequency, braking factor, and corresponding flow and phase simultaneously. That is, in Equation 2 and Equation 3, A _k , α _k , ω _k , φ ₁ , and φ ₂ are important factors for determining a tuning phenomenon between two signals y ₁ and y ₂ .

When the complex mode z _i = e ^{λ i} is defined in order to simplify the flow and phase relationship in the discrete signal, Equations 2 and 3 are converted into a complex exponential form as follows.

In this equation, the coefficients B ₁ and B ₂ of the complex mode are values of the signals y ₁ and y ₂ represented by the complex exponential function with values corresponding to the same complex mode z _k . Therefore, the flow B ₁ and B ₂ according to the phase difference can be expressed as a vector, as shown in Figs. 3 (a) to 3 (c). In Fig. 3 (a), when the two time responses are -sync, the flow vector has a 180 ° phase difference, and when the signals y ₁ and y ₂ are Qsync and + sync, respectively, the flow vector is 90 ° and 0 °, respectively. Phase difference occurs.

Therefore, by calculating the specific mode and the number of flows corresponding to the mode in many signals, it is possible to quantitatively identify the tuning relationship between the signals, and to identify the degree of tuning among the multiple signals, a quantitative indicator is needed. . When the signal is divided into two groups according to the flow phase in the same frequency and braking coefficient mode, the flow vectors included in each group are B _i and B _j , and the sum of the flow vectors ΣB _l and ΣB _s is It can be represented as 6.

Where m and n are the number of streams separated into two groups. Assuming that ΣB _l is larger than ΣB _s , the sync index is defined as follows.

Here, S _m representing the size of the tuning index is referred to as a magnitude sync index, and φ representing the phase is referred to as a phase sync index.

The magnitude tuning index S _m represents the specific gravity between the tuned signals. If S _m is close to 1, the sum of the flow vector of the two signal groups is similar, so they participate in the total vibration similarly, and if it is close to 0, one signal group is dominantly participating in the total vibration. The phase tuning index φ represents the degree of tuning between the signals. Therefore, when the phase tuning indicator φ approaches 0 °, the two signal groups vibrate with + sync, and when near 180 °, the two signal groups vibrate with -sync.

When a disturbance occurs in a dynamic system, if the flows corresponding to a particular mode have a value larger than that of the other mode, the specific mode dominates the global system oscillation. The overall system vibration can be identified from the tuning index. If n tuning indices are calculated, the total vibration g (t) can be calculated by the following equation.

Α _i and ω _i are the i-th mode, respectively, and S _mi and φ _mi are the size tuning indicators and the phase tuning indicators corresponding to the i-th mode.

To detect tuning in a signal, we first need to estimate the parameters, including the same frequency, contained in each signal.The parametric method, which is based on statistical signal processing, is based on the ratio of a discrete Fourier transform. More accurate than the non-parametric method. The Prony method, which is a representative method of the parametric method, is a method of estimating a parameter by fitting an arbitrary signal to a linear combination of a complex mode, and converting a discrete signal into a complex mode as shown in Equations 4 and 5 below. In expressing, the mode is estimated by calculating the pole of the discrete signal from a linear prediction polynomial. In the linear predictive polynomial, the frequency and braking coefficients of the system generating discrete signals are determined, and the remaining parameters of the signal, the magnitude and phase of the flow, are represented by a Vandermonde matrix where each element consists of complex modes estimate from the mode equation).

When estimating the parameters from the time series data, many modes are calculated according to the degree of noise included or the size of the signal. Therefore, the estimated parameters should be appropriately separated according to the purpose of application. In the present invention, a tuning detection criterion for low frequency vibration, which is an issue in terms of a power system, is presented. In a dynamic system such as a power system, low frequency vibration is a vibration occurring in a frequency band below 2.0 Hz, a vibration in a frequency band above 1.0 Hz occurs locally, and a wide band mode that dominates the vibration of the entire system has a frequency band below 1.0 Hz. Occurs mainly in The local mode is included in other local signals, but the flow is small and has a similar phase difference. The wide mode has a similar magnitude and phase difference of about 180 ° because it distributes the vibration in the whole system.

In the present invention, since the tuning is searched in the low frequency vibration, the tuning selection criteria are set as follows. When the mode is λ = α + jβ, the braking coefficient is defined by Equation 9 and the time constant is expressed by Equation 10.

The relative flow is the sum of the flows of the vibration modes divided by the flows corresponding to each mode. When n vibration modes are selected, the relative residue is represented by Equation (11).

Relative flow represents the relative weight of each mode in the signal and is a criterion when determining the importance of the modes. In the present invention, the vibration mode selection criteria for detecting tuning using the variables shown in Equations 9, 10, and 11 are set as follows.

Range of frequency f: 0.1Hz <f <2.0Hz

Range of braking factor ζ: ζ <10

Range of time constant τ: τ> 2

Range of relative flow: B _i > 0.1% B

First, set the frequency band of interest for low frequency vibration, select the modes that satisfy the braking coefficient and the time constant, and finally select only the relatively high flow modes as the vibration mode.

Under this condition, we can search for the tuning that occurs in the whole system. As the range of the parameter set as above is the main concern in terms of low frequency vibration, the other vibration problem should change the range of the frequency of interest or the braking coefficient.

Tuning can distinguish the participating regions for the same mode according to time and disturbances, and can grasp the degree of coupling between local systems, thereby identifying the overall vibration of the dynamic system.

The following describes the tuning detection method according to the embodiment.

4 is an active power waveform of each generator in a power system in which four generators are connected to each other. Two generators (G1-G2, or G3-G4) are tuned to + sync, respectively, and two groups of generators are tuned to -sync and vibrating.

FIG. 5 shows a vector diagram of flowing water for a parameter near 0.57 Hz among the parameters estimated from the signals of FIG. 4. The flowing water B _l , B _s and the tuning index S _i of each generator group are as follows.

0.57 Hz: B _l = 0.012∠35.663 °

B _s = 0.012 ° -145.758 °

S _i = 1.0∠181.421 °

Since the phase tuning indicator is close to 180 °, the generators are precisely synchronized with each other as described above, and because the size tuning indicator is 1, the overall system vibration is almost zero.

6 shows an example of detecting and applying tuning in a dynamic system. In FIG. 6, the 0.7 Hz mode vibrates the same as the glory region and the Kori region, and the Uljin region and the Boryeong region vibrate opposite to the Kori glory region. In addition, the 1.0Hz mode shows that the Boryeong area vibrates opposite to the Gyeongin and Yeonggwang areas.

7 shows running water corresponding to the frequencies 0.7 Hz and 1.0 Hz, respectively. The sync index of each frequency shown in FIG. 7 is as follows.

0.7 Hz: S _i = 0.823∠149.3 °

1.0 Hz: S _i = 0.983∠179.1 °

From the tuning index, we can see that both frequencies in the signal are vibrating with -sync. In particular, at 1.0 Hz, the size tuning indicator is close to 1 and the phase tuning indicator is exactly at 180 °, indicating that the generators of the west coast are vibrating with -sync. In the frequency 0.7Hz mode, the internal generators in the southern and northern regions are vibrating + sync with each other, and the generators in the southern and northern regions are vibrating with -sync. In the frequency 1.0Hz mode, the Gyeongin and Yeonggwang area generators including Seoincheon are vibrating with + sync, and this generator group is vibrating with -sync with Boryeong-Taean area generators.

Since the tuning index S _i is close to 1 at both frequencies, it can be seen that the overall system vibration by each frequency is small and therefore the overall system is stable. If the tuning index is close to zero, the vibration of the whole system is influenced by the vibration in a particular region, which can have a big impact on the system even with small disturbances.

As shown in the above example, if the synchronization is identified, the vibration characteristics of each local system participating in the dynamic system can be identified, and the vibration characteristics of the entire system can be identified, so that the overall vibration of the dynamic system can be identified in real time. Take control action to prevent collapse.

The present invention is described above by illustrating specific embodiments, but the present invention is not limited to the above embodiments. Those skilled in the art can easily make various changes and modifications to the present invention, and it should be noted that such variations or modifications are included within the scope of the present invention as long as the features of the present invention are used.

As described above, the present invention is a method for detecting tuning occurring in a complex dynamic system, and selects a mode having a similar frequency, a braking coefficient, and a time constant in an important vibration mode included in a plurality of signals, By comparing the phases, the tuning of the dynamic system is detected. By detecting the synchronization of signals having the same characteristics in real time or non-real time, it is possible to grasp important system characteristics, so that appropriate control operations can be performed to prevent the collapse of dynamic systems.

Claims (3)

In detecting real-time and non-real-time tuning in order to prevent system collapse due to oscillation in a dynamic system, Estimating a parameter by inputting a time series data signal; Extracting a mode included in a frequency range and a braking coefficient range affecting dynamic system vibrations from the estimated parameters; Extracting a mode included in a time constant range affecting dynamic system vibration with respect to the extracted mode, selecting a mode included in a relative flow range affecting dynamic system vibration with respect to the extracted mode; Dividing the flow vectors into two groups by using a phase of the flow for the selected modes; Calculating the sum of each flow vector ΣB _l , ΣB _s for the separated flow vectors; 『

"from Calculating a sync index, Detecting a tuning characteristic for the dynamic system vibration in the calculated tuning index; And a step of outputting, storing and transmitting the detected result. However, in the above equation, Si is a tuning index, and ΣB _l and ΣB _s are the sums of the separated flow vectors, respectively. And Sm is the size tuning indicator and φ is the phase tuning indicator. The method according to claim 1 Relative flow 『

"from Tuning detection of real-time and non-real-time signals in a dynamic system comprising the step of calculating. In the above equation, B is a flow vector, B _i is an i-th flow vector, B _Σ is the sum of the flow vectors, and% B is a relative flow. In detecting real-time and non-real-time tuning to prevent system collapse due to vibrations in dynamic systems, Estimating a parameter by inputting a time series data signal; Extracting a mode included in a frequency range and a braking coefficient range affecting dynamic system vibrations from the estimated parameters; Extracting a mode included in a time constant range affecting dynamic system vibration with respect to the extracted mode, selecting a mode included in a relative flow range affecting dynamic system vibration with respect to the extracted mode; Separating the flow vectors into two groups by using the phase of flow for the selected modes, calculating a tuning index from the sum of the flow vectors for the separated flow vectors, The total vibration of the dynamic system is calculated from the calculated tuning index. 『

"from Calculation step, And outputting, storing, and transmitting the calculated result. In the above equation, α _i and ω _i are the real part (braking coefficient) and the imaginary part (frequency) of the i-th mode, respectively, and S _mi and φ _mi are the magnitude tuning index and the phase tuning corresponding to the i-th mode. It is an indicator.

Images