CN102946394B - Based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method in data acquisition analysis system - Google Patents

Abstract

The invention discloses in a kind of data acquisition analysis system (SCADA) based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method, according to wave character, nonuniform sampling is carried out to data in the refresh cycle at remote data acquisition terminal (RTU), extraction can reflect object of observation (node voltage in Power system dynamic process, branch road is meritorious or idle) data point of waveform characteristic is (as crest, trough), realize data compression, and through communication network, the characteristic point data of extraction and positional information thereof are sent to control centre, this dynamic process is reconstructed by cubic Hamiltonian symmetrical systems polynomial interopolation.The present invention can on existing SCADA system basis, and realize the supervision to electrical network multidate information by software upgrading, without the need to separately increasing hardware device, good economy performance, is conducive to dispatcher's in-service monitoring electrical network multidate information, ensures power grid security.<!--1-->

Description

Based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method in data acquisition analysis system

Technical field

The present invention relates to dispatching automation of electric power systems technology, particularly the compression transmission of Power system dynamic process data and reconstructing method in power system of data acquisition and supervisory control system (SCADA).

Background technology

Data acquisition analysis system (SupervisoryControlandDataAcquisition, SCADA) be a kind of supervisory control system come with computer and development communication technologies, be widely applied to the numerous areas such as electric power, water supply, oil, chemical industry, combustion gas and process control, the most general to apply in dispatching of power netwoks.It is by being arranged on the remote data acquisition terminal (RemoteTerminalUnit of power plant and transformer station, RTU) by monitored object---the state information (as electric current, voltage, Line Flow, on off state etc.) of electrical network feeds back to dispatching control center, abnormal state is reported to the police, and carries out decision-making for dispatcher and implement control offering help.Generally speaking, when electrical network normally runs, data acquisition analysis system energy complete and accurate ground reaction power grid state information, helps management and running personnel tracking in time, adjustment operation states of electric power system, has become the indispensable instrument of power scheduling.

Wide geographically tens even thousands of electric currents, voltage, frequency, trend and the switching-state information in a hundreds of transformer station and power plant are distributed in because data acquisition analysis system needs to realize monitoring, hold quantitative limitation by network service and computer disposal, display electric network state can only be refreshed in several seconds modes once.When electrical network is in normal operating condition and node voltage and Line Flow change little, data acquisition analysis system energy accurate response operation of power networks state.But when electrical network is short-circuited fault or other disturbance, the electric current of electrical network, voltage and Line Flow all may Rapid Variable Design, and now refreshing every several seconds the data acquisition analysis system once showing electric network state just cannot accurate response system dynamic course.This type of situation happens occasionally in the emergency process of actual electric network, and some typical Power system dynamic process can not get effective monitoring as low-frequency oscillation for a long time in data acquisition analysis system, has a strong impact on the timely correct decisions of dispatcher.

In digital transformer substation, remote data acquisition terminal receives each instantaneous electric current, voltage measurement data by communication network in transformer station from electronic mutual inductor, and calculates Line Flow.In fact, transformer substation side remote data acquisition terminal collect data integrity contain Power system dynamic process information.But because the communication bandwidth of data acquisition analysis system network limits, Power system dynamic process data integrity can not be sent to dispatching control center, display electric network state can only be refreshed, electrical network oscillatory process multidate information when being difficult to provide disturbance to management and running personnel every a few second.A kind of feasible selection of head it off is that dynamic process extracting data also transmits the data that can characterize dynamic process information on a small quantity within the Refresh Data cycle, to reduce the requirement to communication system, again at dispatching control center according to the data reconstruction Power system dynamic process received, reach the object of Electrical Power System Dynamic communicating information to dispatching control center.

Summary of the invention

Technical problem to be solved by this invention is: for above-mentioned the deficiencies in the prior art, there is provided in a kind of data acquisition analysis system based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method, it can improve the ability of its reflection system dynamic course on data acquisition analysis system existing hardware facility basis, to improve the observability of Power system dynamic process in data acquisition analysis system, improve the ability of dispatcher's solution of emergent event and quick correct decisions.

In order to solve the problems of the technologies described above, the technical solution adopted in the present invention is: based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method in a kind of data acquisition analysis system, the method is within a refresh cycle of data acquisition analysis system, at remote data acquisition terminal (RTU), adopt nonuniform sampling mode, to object of observation (i.e. node voltage, branch road is meritorious and branch road is idle) dynamic process data pick-up can reflect the data point (crest or trough) of Power system dynamic process oscillating characteristic, realize data compression, and the data point of extraction and positional information thereof are sent to dispatching control center through communication network, at dispatching control center end, the dynamic data of each time point of this object of observation is reconstructed to the data separate cubic Hamiltonian symmetrical systems polynomial interopolation received, to show dynamic process of electrical power system.

Wherein: the cubic Hamiltonian symmetrical systems multinomial adopted during reconstruct is:

$H_3\left(x\right)=\underset{i=k}{\overset{k+1}{\mathrm{\&Sigma;}}}[y_i{\mathrm{\&alpha;}}_i\left(x\right)+y_i^{\&prime;}{\mathrm{\&beta;}}_i\left(x\right)]$ ，

Each Interpolation-Radix-Function in above formula:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_k\left(x\right)=(1+2\frac{x-x_k}{x_{k+1}-x_k}){\left(\frac{x-x_{k+1}}{x_k-x_{k+1}}\right)}^2\\ {\mathrm{\&alpha;}}_{k+1}\left(x\right)=(1+2\frac{x-x_{k+1}}{x_k-x_{k+1}}){\left(\frac{x-x_k}{x_{k+1}-x_k}\right)}^2\end{array}\right.$

$\left\{\begin{array}{c}{\mathrm{\&beta;}}_k\left(x\right)=\left(x-x_k\right){\left(\frac{x-x_{k+1}}{x_k-x_{k+1}}\right)}^2\\ {\mathrm{\&beta;}}_{k+1}\left(x\right)=\left(x-x_{k+1}\right){\left(\frac{x-x_k}{x_{k+1}-x_k}\right)}^2\end{array}\right.$

Y _i, y _i ^'be respectively data point information, the first derivative values corresponding with time point in any node of extraction, k be more than or equal to zero integer.

On to be set forth in the method extracting object of observation oscillating characteristic data point (oscillation characteristics point) in a refresh cycle as follows:

1) this refresh period first moment and tail moment are as sampling number strong point;

2) 5% rated value person is less than without the need to carrying out sampling interpolation reconstruction in whole refresh period measured value fluctuation;

3) 5% rated value person is greater than to accumulated change amount in this period, all vibration crests, trough data point in the period can be tried to achieve, wherein, before and after crest data point, the data in moment are all less than this moment numerical value, and the data before and after trough data point are all greater than this time data;

4) for catching fault generation moment point, the point continuous three moment variable quantities being less than rear continuous three moment variable quantities 1/10th is defined as fault moment, as sampling number strong point.

Above-mentioned mention realize data compression, with less data (in this refresh cycle of the reflection namely extracted in Power system dynamic process the crest of object of observation oscillating characteristic or trough data) dynamic process of electrical power system (electric current, voltage and trend) information in the record data refresh cycle, and only transmit those data (those data are recovered by reconstructing at dispatching control center end), can transmitted data amount be reduced, reach and reduce communication system requirements and the object uploading defeated dynamic process of electrical power system information on existing system basis.Owing to only transmitting the data that the rear data volume of sampling significantly reduces, can effectively reduce electric power communication network traffic load.Institute of the present invention extracting method can be aided with necessary Software adapting in available data collection and supervisory control system basis to be implemented.Because the packet of institute's extracting method transmission is containing dynamic process of electrical power system information, so adopt the data acquisition analysis system of this method can react dynamic process of electrical power system truely and accurately.

Compare with existing lossless compression method (Huffman coding, arithmetic coding, run length encoding, LZW coding compression algorithm etc.), this method has outstanding advantage on compression ratio.Compare with the data lossy compression method method based on polynomial curve fitting, the renewal duration of this method to data acquisition analysis system is insensitive, the accuracy of reconstruct data all can be ensured under various renewal duration, and adopt higher order polynomial curve to exist to hold lattice phenomenons, when data acquisition analysis system elongates update time, data reconstruction precision significantly reduces.Compared with the data compression method based on uniform sampling/cubic spline interpolation, it is insensitive to there is the moment to disturbance in this method, and uniform sampling method is not when being drawn into characteristic point, and reconstruct data precision significantly reduces.

The present invention compared with the prior art, the most outstanding advantage to represent the dynamic process of electrical power system information that available data collection and supervisory control system cannot show, and helps the Power system dynamic process such as the recovery after the observation disturbance of management and running personnel simple, intuitive ground, vibration and unstability; Secondly, the method can realize in data acquisition analysis system existing hardware system-based, does not need to set up special communication network; Moreover, because the method only need carry out corresponding transformation to software systems, simple, specific monitoring target can be selected separately to implement flexibly.

Accompanying drawing explanation

Fig. 1 is emulation WEPRI36 node electrical network figure.

Fig. 2 is the comparison diagram of power fluctuation and the actual dynamic process of 31 nodes that available data collection and supervisory control system showed for the refresh cycle with 3 seconds.

Power system dynamic process design sketch in 18 seconds that Fig. 3 is oscillation characteristics point effect in 31 node actual power dynamic processes and data acquisition analysis system extract at remote collecting device end in the process of the present invention 18 seconds, dispatching control center end is reconstructed by oscillation characteristics point.

Embodiment

The present invention is based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method in data acquisition analysis system, within a refresh cycle of data acquisition analysis system, at remote data acquisition terminal, adopt nonuniform sampling mode, extraction to reflect in this refresh cycle object of observation (node voltage in Power system dynamic process, branch road is meritorious and branch road is idle) data point (crest and trough data) of oscillating characteristic, realize data compression, and those data points extracted and positional information thereof are sent to dispatching control center via communication network, at dispatching control center end, the dynamic data of each time point of object of observation in this refresh cycle is reconstructed to the data separate cubic Hamiltonian symmetrical systems polynomial interopolation received, with Restoration dynamics process data, and show dynamic process of electrical power system.

According to electrical network performance graph architectural feature under disturbance, the method extracting above-mentioned oscillating waveform characteristic point (oscillating characteristic data point) is as follows:

1, this refresh period first moment and tail moment are as sampling number strong point (i.e. oscillating waveform characteristic point);

2,5% rated value person is less than without the need to carrying out sampling interpolation reconstruction in whole refresh period measured value fluctuation;

3,5% rated value person is greater than to accumulated change amount in this period, all vibration crests, trough data point in the period can be tried to achieve, wherein, before and after crest data point, the data in moment are all less than this moment numerical value, and the data before and after trough data point are all greater than this time data;

4, for catching fault generation moment point, the point continuous three moment variable quantities being less than rear continuous three moment variable quantities 1/10th is defined as fault moment, as sampling number strong point.

N+1 in 1 refresh cycle is sent to the behavioral characteristics point of dispatching control center end, the time point x of its correspondence, x ₀< x ₁< x ₂< ... < x _n, the magnitude of voltage that each time point is corresponding and first derivative are y _m, y _m ^', m=0,1,2 ... n, can try to achieve the Hermite interpolation polynomial H of 2n+1 time _2n+1, make

$\left\{\begin{array}{c}{H}_{2n+1}\left(x_m\right)=y_m\\ H_{2n+1}^{\&prime;}\left(x_m\right)=y_m^{\&prime;}\end{array}\right.$

As n=1, H ₃x () is called cubic Hamiltonian symmetrical systems multinomial.Get interpolation time point x _k, x _k+1, structure Interpolation-Radix-Function a _i(x), β _i(x), i, j=k, k+1(0≤k≤n-1), each basic function is cubic polynomial, and meets the following conditions:

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_i\left(x_j\right)={\mathrm{\&delta;}}_{\mathrm{ij}},{\mathrm{\&alpha;}}_i^{\&prime;}\left(x_j\right)=0\\ {\mathrm{\&beta;}}_i\left(x_j\right)=0,{\mathrm{\&beta;}}_i^{\&prime;}\left(x_j\right)={\mathrm{\&delta;}}_{\mathrm{ij}}\end{array}\right.$

Utilize a _i(x) and β _ix () can construct cubic Hamiltonian symmetrical systems multinomial:

, y in formula _i, y _i ^'for time point x _icorresponding data point information, first derivative;

By the method for undetermined coefficients, can in the hope of Interpolation-Radix-Function

$\left\{\begin{array}{c}{\mathrm{\&alpha;}}_k\left(x\right)=(1+2\frac{x-x_k}{x_{k+1}-x_k}){\left(\frac{x-x_{k+1}}{x_k-x_{k+1}}\right)}^2\\ {\mathrm{\&alpha;}}_{k+1}\left(x\right)=(1+2\frac{x-x_{k+1}}{x_k-x_{k+1}}){\left(\frac{x-x_k}{x_{k+1}-x_k}\right)}^2\end{array}\right.$

$\left\{\begin{array}{c}{\mathrm{\&beta;}}_k\left(x\right)=\left(x-x_k\right){\left(\frac{x-x_{k+1}}{x_k-x_{k+1}}\right)}^2\\ {\mathrm{\&beta;}}_{k+1}\left(x\right)=\left(x-x_{k+1}\right){\left(\frac{x-x_k}{x_{k+1}-x_k}\right)}^2\end{array}\right.$

Then Successful construct cubic Hamiltonian symmetrical systems multinomial.Cubic Hamiltonian symmetrical systems multinomial is prior art, is just illustrated roughly at this.

Be described below in conjunction with specific embodiment.

See Fig. 1, be emulation WEPRI36 node electrical network figure, the inventive method is described as on the basis of test macro at this WEPRI36 node system.This is desired to make money or profit and test electrical network has been carried out to the transient stability emulation that continues 18 seconds with electric power system simulation analysis software, be engraved in circuit mid point between 30 and No. 31 nodes when being arranged at 2 seconds and carry out the three-phase symmetrical short trouble of lasting 0.1 second, and all Line Flows and node voltage are exported (per second comprise 100 time points) by 10 milliseconds of step-lengths, obtain Power system dynamic process data.

In conjunction with the comparison diagram see Fig. 2, Fig. 2 being power fluctuation and the actual dynamic process of 31 nodes that available data collection and supervisory control system showed for the refresh cycle with 3 seconds.Wherein, solid line is with 10 milliseconds of 31 node power dynamic process data exported for step-length (actual meritorious vibration), dotted line is by available data collection and supervisory control system data transmission method, the 31 node power change curves observed when refreshing display system state for every 3 seconds.At this, with the Refresh Data cycle initial transient data refresh display the whole Refresh Data cycle in electric network state, namely 0 to 3 second moment it is seen that 0 second moment voltage, 3 to 6 seconds moment, the rest may be inferred it is seen that 3 second moment voltage.As can be seen from Figure, the mode that shows of available data collection and supervisory control system refresh data obviously can not representation system dynamic process truely and accurately.

Again the voltage data the inventive method in Fig. 1 is processed below, shown the effect in Power system dynamic process information to show this method.

See Fig. 3, it is the oscillation characteristics point effect in 31 node actual power dynamic processes and data acquisition analysis system extract at remote collecting device end in the process of the present invention 18 seconds, Power system dynamic process design sketch in dispatching control center end reconstructed by oscillation characteristics point 18 seconds.Wherein solid line is be fluctuation data (actual meritorious vibration) in the 31 node active power 18 seconds that exports of step-length with 10 milliseconds, oscillation characteristics point (i.e. sampling number strong point) in little fork 31 node power that to be data acquisition analysis system extract at remote collecting device end with this method 2 refresh cycles (be 9 seconds for the refresh cycle, place's dotted line was 2 cycle separator bars in 9 seconds).The dotted line cubic Hamiltonian symmetrical systems method that to be data acquisition analysis system propose with this method is in the Power system dynamic process effect of 31 2 refresh cycles of node power that dispatching control center end is reconstructed by oscillation characteristics point.Wherein, be that step-length inputs cubic Hamiltonian symmetrical systems multinomial as independent variable with 10ms, the polynomial value obtained is the corresponding data reconstruction that just can realize data dispatch control centre end.

As can be seen here, this method can reach with a small amount of increase of amount of communication data the object significantly improving Power system dynamic process monitoring capability.

And available data collection and supervisory control system are with 1 magnitude of voltage characterization system state of each Refresh Data cycle, and when refresh time is extended for 9 seconds, the data volume of transmission is 9.Relative to the whole instantaneous voltage data of transmission, if by oscillating characteristic extract minutiae, the data volume in 9 seconds is generally 30 to 35, and compression ratio is between 30 to 25, and relative to existing system, the amount of communication data of increase is also few.Therefore, the relative available data collection of the data volume due to this method and supervisory control system increase not quite, and this method can be implemented in available data collection and supervisory control system basis.

Claims (2)

1. in a data acquisition analysis system based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method, it is characterized in that: the method is within a refresh cycle of data acquisition analysis system, at remote data acquisition terminal, adopt nonuniform sampling mode, the data point of Power system dynamic process oscillating characteristic can be reflected to object of observation dynamic process data pick-up, realize data compression, and the data point of extraction and positional information thereof are sent to dispatching control center through communication network, at dispatching control center end, the dynamic data of each time point of this object of observation is reconstructed to the data separate cubic Hamiltonian symmetrical systems polynomial interopolation received,

On be set forth in a refresh cycle to extract and can reflect that the method for the data point of object of observation oscillating characteristic in Power system dynamic process is as follows:

1) refresh period first moment and tail moment are as sampling number strong point;

2) at whole refresh period, measured value fluctuation is less than 5% rated value person, without the need to carrying out sampling interpolation reconstruction;

3) 5% rated value person is greater than to accumulated change amount in this period, tries to achieve all vibration crests, trough data point in this period; Wherein, before and after crest data point, the data in moment are all less than this moment numerical value, and the data before and after trough data point are all greater than this time data;

4) for catching fault generation moment point, the point that front continuous three moment variable quantities are less than rear continuous three moment variable quantities 1/10th is defined as fault moment, as sampling number strong point.

2. in data acquisition analysis system as claimed in claim 1 based on the Power system dynamic process data compression transmission of waveform feature and reconstructing method, it is characterized in that: described object of observation refers to that node voltage, branch road are meritorious and branch road is idle.