CN114818861A

CN114818861A - Data processing method and device for phasor measurement device of power system and electronic equipment

Info

Publication number: CN114818861A
Application number: CN202210293539.5A
Authority: CN
Inventors: 于淼; 杨晨宇; 杜蔚杰; 张寿志; 胡敬轩; 李京霖
Original assignee: Beijing University of Civil Engineering and Architecture
Current assignee: Beijing University of Civil Engineering and Architecture
Priority date: 2022-03-23
Filing date: 2022-03-23
Publication date: 2022-07-29

Abstract

The invention discloses a method and a device for processing data of a phasor measurement unit of a power system and electronic equipment, wherein the method comprises the following steps: obtaining PMU data generated by a phasor measurement device; screening PMU data by adopting an iterative random forest algorithm to obtain first data; and filling the first data by adopting the trained dynamic neural network to obtain a data processing result. The dynamic neural network is adopted to regard PMU data as a time sequence for data filling, so that the accuracy and the efficiency of PMU data filling are improved, and the construction of complete PMU data is realized.

Description

Data processing method and device for phasor measurement device of power system and electronic equipment

Technical Field

The invention relates to the technical field of computers, in particular to a method and a device for processing data of a phasor measurement device of a power system and electronic equipment.

Background

In recent years, with the improvement of data mining capabilities such as national big data and artificial intelligence, the types and scales of data have become huge. Therefore, the analysis and prediction of the data provide an infinite space for strategic decision, future development and other aspects of the power grid enterprise. However, in actual operation of a power grid enterprise, a large amount of data is generated every second, and due to instrument precision limitation, noise influence and human errors, a large amount of data is lost or different. Pmu (phaser measurement unit) is a device used in the grid for measuring and outputting synchronous vectors and for dynamic recording. The device has real-time monitoring and recording functions, can perform quick-reading fault analysis, captures low-frequency oscillation of a power grid, and measures information such as a power angle of a generator in real time. However, when PMU data is generated, actual data is interfered by a plurality of adverse conditions, so that analysis of PMU data influences power grid analysis and stability analysis.

At present, for the problem of poor PMU data processing, data is predicted by a basic mathematical theory method or a modern machine learning method, but data screening is generally lacked in the PMU data prediction process, and the influence of the error data on the data prediction is often fatal. Meanwhile, the PMU is a value generated at one moment, so that the real-time performance and the accuracy of the PMU are very important.

In summary, there is a need for a data processing method for solving the above-mentioned problems in the prior art.

Disclosure of Invention

Due to the problems of the existing method, the invention provides a method and a device for processing data of a phasor measurement device of a power system and electronic equipment.

In a first aspect, the present invention provides a method for data processing of phasor measurement devices in a power system, including:

obtaining PMU data generated by a phasor measurement device;

screening the PMU data by adopting an iterative random forest algorithm to obtain first data;

and filling the first data by adopting the trained dynamic neural network to obtain a data processing result.

Further, screening the PMU data by using an iterative random forest algorithm to obtain first data includes:

establishing a plurality of decision trees according to the PMU data;

generating a random forest according to the decision trees;

and determining first data according to the random forest by adopting a CART algorithm and a random forest voting mechanism.

Further, the random forest voting mechanism is a minority-compliant majority mechanism.

Further, before the filling the first data with the trained dynamic neural network to obtain a data processing result, the method further includes:

establishing a time series model;

generating a dynamic neural network according to the time series model, the preset number of hidden nodes and the preset delay quantity;

and training the dynamic neural network to obtain the trained dynamic neural network.

Further, the training the dynamic neural network includes:

and training the dynamic neural network by adopting an LM algorithm.

In a second aspect, the present invention provides an apparatus for data processing of phasor measurement devices in a power system, including:

the acquisition module is used for acquiring PMU data generated by the phasor measurement unit;

the processing module is used for screening the PMU data by adopting an iterative random forest algorithm to obtain first data; and filling the first data by adopting the trained dynamic neural network to obtain a data processing result.

Further, the processing module is specifically configured to:

establishing a plurality of decision trees according to the PMU data;

generating a random forest according to the decision trees;

Further, the processing module is specifically configured to:

the random forest voting mechanism is a minority-compliant majority mechanism.

Further, the processing module is further configured to:

establishing a time sequence model before filling the first data by adopting the trained dynamic neural network to obtain a data processing result;

Further, the processing module is specifically configured to:

and training the dynamic neural network by adopting an LM algorithm.

In a third aspect, the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the processor implements the method for processing data of the phasor measurement device in the power system according to the first aspect.

In a fourth aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of power system phasor measurement device data processing according to the first aspect.

According to the technical scheme, the method, the device and the electronic equipment for processing the data of the phasor measurement unit of the power system provided by the invention have the advantages that aiming at small data fluctuation of the PMU, the iterative random forest algorithm is adopted to screen out bad data in the PMU data, the influence of noise on data prediction is reduced, and the accuracy of the data prediction is improved. The dynamic neural network is adopted to regard PMU data as a time sequence for data filling, so that the accuracy and the efficiency of PMU data filling are improved, and the construction of complete PMU data is realized.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

FIG. 1 is a system framework of a method for data processing of phasor measurement units of a power system according to the present invention;

FIG. 2 is a schematic flow chart of a data processing method for a phasor measurement unit of a power system according to the present invention;

FIG. 3 is a schematic flow chart of a data processing method for a phasor measurement unit of a power system according to the present invention;

FIG. 4 is a schematic flow chart of a data processing method for a phasor measurement unit of a power system according to the present invention;

FIG. 5 is a schematic diagram of the correlation between PMU data input and algorithm error provided by the present invention;

FIG. 6 is a graph of the autocorrelation of the algorithmic error provided by the present invention;

FIG. 7 is a schematic structural diagram of a data processing apparatus of a phasor measurement unit of a power system according to the present invention;

fig. 8 is a schematic structural diagram of an electronic device provided in the present invention.

Detailed Description

The following further describes embodiments of the present invention with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.

The method for processing data of the phasor measurement unit in the power system according to the embodiment of the present invention can be applied to a system architecture as shown in fig. 1, where the system architecture includes a phasor measurement unit 100 in the power system and a server 200.

Specifically, the server 200 is used to obtain PMU data generated by the phasor measurement device 100.

The server 200 is further configured to screen PMU data by using an iterative random forest algorithm to obtain first data.

The server 200 is further configured to fill the first data with the trained dynamic neural network to obtain a data processing result.

It should be noted that fig. 1 is only an example of a system architecture according to the embodiment of the present invention, and the present invention is not limited to this specifically.

Based on the above illustrated system architecture, fig. 2 is a schematic flow chart corresponding to a method for processing data of a phasor measurement device of an electrical power system according to an embodiment of the present invention, as shown in fig. 2, the method includes:

in step 201, PMU data generated by a phasor measurement device is obtained.

A Phasor Measurement Unit (PMU) is a phasor measurement unit formed by using GPS second pulses as a synchronous clock, can be used for measuring voltage vectors of each node of a power system in a transient process, and is an important device for ensuring safe operation of a power grid. PMU is installed at an important transformer substation and a power plant of the power system, and a real-time dynamic monitoring system of the power system is constructed.

PMU data is a time-varying value and can therefore be considered as a time series. For such a time series prediction, a future value is predicted from a past value in principle.

Step 202, screening PMU data by adopting an iterative random forest algorithm to obtain first data.

In the embodiment of the present invention, a flow of steps of screening PMU data by using an iterative random forest algorithm and obtaining first data is shown in fig. 3, which specifically includes:

in step 301, a plurality of decision trees are built based on PMU data.

Specifically, a random forest training set N is established for PMU data.

Further, m (m < N) data are taken from N randomly and with replacement, and this operation is repeated i times and recorded as a training subset. And respectively extracting k characteristic values from each training subset, and establishing a decision tree so as to obtain i decision trees.

In a possible implementation mode, a bagging algorithm is adopted to firstly extract m data from the whole sample data, a classifier is built according to k attributes of the m data, and the operation of building the classifier is repeated for i times.

In the above description, m, N, i, and k are positive integers.

Step 302, generating a random forest according to a plurality of decision trees.

Specifically, pruning is performed on each decision tree, and a random forest is generated after the fitting condition is eliminated.

And step 303, determining first data according to the random forest by using a CART algorithm and a random forest voting mechanism.

In the embodiment of the invention, a CART algorithm is adopted when the decision tree is selected.

It should be noted that a Classification And Regression Tree (CART) is one implementation of a decision Tree, And generally, there are three implementations of a decision Tree, namely, an ID3 algorithm, a CART algorithm, And a C4.5 algorithm.

Further, an ID3 algorithm and a C4.5 algorithm may also be used, which is not specifically limited in this embodiment of the present invention.

Specifically, the experience of calculating the PMU data D is as follows:

where D represents a sample set, C _k Representing a subset of samples in D that belong to class k.

Further, traversing all characteristics of PMU data, the following calculation is carried out for the characteristic A of PMU data:

1. the empirical conditional entropy H (D | A) of A on PMU data D is calculated.

2. And (3) calculating the information gain of the characteristic A, wherein the specific formula is as follows:

g(D,A)＝H(D)–H(D|A)

3. and selecting the characteristic with the largest Gini index as the current splitting characteristic.

It should be noted that a larger kini index indicates a stronger ability of this property to change from inconclusive to deterministic.

Further, the specific calculation formula of the kini index is as follows:

wherein p is _k Representing the probability of the sample belonging to the kth class.

Further, the quality of the decision tree is calculated by using a loss function. The specific calculation formula of the loss function is as follows:

C(T)＝∑ _t∈leaf N _t ·H(t)

in the embodiment of the invention, the complete tree is pruned firstly, the operation is repeated for a plurality of times until only the tree root is left, and the decision tree with the minimum loss function is selected.

It should be noted that the decision tree, although having good classification capability in training, still has uncertainty when testing unknown data, and may still have overfitting phenomenon, so that proper pruning is required.

In the embodiment of the invention, a random forest voting mechanism is adopted to select data, and bad data is deleted to obtain first data.

In one possible embodiment, the voting mechanism employed is a minority-compliant majority mechanism.

Specifically, the PMU data is correct and incorrect according to the voting results of the n classifiers.

It should be noted that the voting mechanism may also be a simple voting mechanism, a bayes voting mechanism, a laplacian smoothing mechanism, and the like, which is not specifically limited in this embodiment of the present invention.

According to the scheme, the PMU data are divided into missing data, bad data and accurate data by adopting the iterative random forest algorithm aiming at small PMU data fluctuation, the bad data are screened out, noise generated in the PMU data collecting process is screened out, the PMU data are preliminarily screened out before the PMU data prediction is carried out, the influence of the noise on the data prediction is reduced, and the accuracy of the data prediction is improved.

And step 203, filling the first data by adopting the trained dynamic neural network to obtain a data processing result.

In the embodiment of the invention, the dynamic neural network is adopted, so that the future data can be predicted according to the past value searching rule, and the PMU data is quantitatively analyzed.

Before the trained dynamic neural network is used for filling the first data, the step flow of the embodiment of the invention is shown in fig. 4, and the method specifically comprises the following steps:

step 401, a time series model is established.

Specifically, the first data that is newly generated data is denoted by X. Based on the data characteristic of X and the nature of PMU data, i.e., a non-linear time series, a model comprising X (t) and y (t), i.e., a model with inputs and outputs affecting the series, is selected for use.

Step 402, generating a dynamic neural network according to the time series model, the preset number of hidden nodes and the preset delay number.

Specifically, the embodiment of the invention determines an input function x (t) and y (t) to be output, and establishes a training set, a verification set and a test set according to a preset proportion.

For example, a training set, a validation set, and a test set are established at 70%, 15%, and 15% of the data, respectively.

It should be noted that the training set should be changed continuously in the model, the validation set is used to measure the generalization of the network, and the training is stopped when the generalization stops improving, and the test set is used to compare the training results.

Further, according to the previous stage test, a dynamic neural network is generated according to the time sequence model, the preset number of hidden nodes and the preset delay quantity.

For example, the number of hidden nodes and the delay number are set to 6 and 3, respectively.

And 403, training the dynamic neural network to obtain the trained dynamic neural network.

In one possible embodiment, the dynamic neural network is trained using the LM algorithm.

It should be noted that Levenberg-Marquardt (LM) is a numerical method for solving the nonlinear optimization problem, and the method is improved through continuous iteration until the optimal solution is approached.

Above-mentioned scheme uses the characteristic that LM algorithm can laminate PMU data to the at utmost. When the PMU data change is fast, the whole formula is close to the Gauss-Newton iteration method by using small parameters, and when the PMU data change rate is slow, the whole formula is close to the gradient method by using large parameters.

And further, judging whether retraining is needed or not according to the dynamic neural network generation data.

Further, the retraining metric is mainly based on error histograms, error autocorrelation, input error-related images, and the like.

According to the scheme, the LM algorithm is combined with the dynamic neural network to determine the PMU data as the time sequence for data filling, so that the accuracy and the efficiency of PMU data filling are improved, and the construction of complete PMU data is realized.

The embodiment of the invention compares the filling data with the original data, and fig. 5 is a schematic diagram of the correlation between the PMU data input and the algorithm error.

It can be seen from fig. 5 that their correlations are small, which indicates that the error of PMU data input and algorithm is small, and the filling condition of PMU data is reflected well, so that the algorithm can better fit the actual PMU data.

Fig. 6 is a graph of autocorrelation of an algorithm error, and it can be seen from fig. 6 that most of autocorrelation images are gathered at point 0, which shows that the random error is small and is less interfered by random error factors, so that the algorithm has high filling precision.

It can be seen from the figure that the embodiment of the invention greatly improves the accuracy, the computational efficiency, the adaptability and the like of PMU data prediction. The method has strong precision and adaptability, can effectively eliminate a large amount of bad data injection, and has good service performance.

Based on the same inventive concept, fig. 7 exemplarily illustrates a device for data processing of a power system phasor measurement unit according to an embodiment of the present invention, which can perform a flow of a method for data processing of a power system phasor measurement unit.

The apparatus, comprising:

an obtaining module 701, configured to obtain PMU data generated by a phasor measurement device;

the processing module 702 is configured to screen the PMU data by using an iterative random forest algorithm to obtain first data; and filling the first data by adopting the trained dynamic neural network to obtain a data processing result.

Further, the processing module 702 is specifically configured to:

establishing a plurality of decision trees according to the PMU data;

generating a random forest according to the decision trees;

Further, the processing module 702 is specifically configured to:

the random forest voting mechanism is a minority-compliant majority mechanism.

Further, the processing module 702 is further configured to:

Further, the processing module 702 is specifically configured to:

and training the dynamic neural network by adopting an LM algorithm.

Based on the same inventive concept, another embodiment of the present invention provides an electronic device, which specifically includes the following components, with reference to fig. 8: a processor 801, a memory 802, a communication interface 803, and a communication bus 804;

the processor 801, the memory 802 and the communication interface 803 complete mutual communication through the communication bus 804; the communication interface 803 is used for realizing information transmission between devices;

the processor 801 is configured to call a computer program in the memory 802, and the processor implements all the steps of the above method for processing the data of the power system phasor measurement unit when executing the computer program, for example, the processor implements the following steps when executing the computer program: obtaining PMU data generated by a phasor measurement device; screening the PMU data by adopting an iterative random forest algorithm to obtain first data; and filling the first data by adopting the trained dynamic neural network to obtain a data processing result.

Based on the same inventive concept, yet another embodiment of the present invention provides a non-transitory computer-readable storage medium having stored thereon a computer program, which when executed by a processor implements all the steps of the above-mentioned method for data processing of a phasor measurement device of a power system, for example, the processor implements the following steps when executing the computer program: obtaining PMU data generated by a phasor measurement device; screening the PMU data by adopting an iterative random forest algorithm to obtain first data; and filling the first data by adopting the trained dynamic neural network to obtain a data processing result.

In addition, the logic instructions in the memory may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a device for data processing of a phasor measurement unit of a power system, or a network device) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present invention. One of ordinary skill in the art can understand and implement it without inventive effort.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. Based on such understanding, the above technical solutions may be embodied in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for enabling a computer device (which may be a personal computer, a device for processing data of a phasor measurement apparatus of a power system, or a network device, etc.) to execute the method for processing data of the phasor measurement apparatus of the power system according to the embodiments or some parts of the embodiments.

In addition, in the present invention, terms such as "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.

Moreover, in the present invention, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

Furthermore, in the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A method for data processing of a phasor measurement unit of a power system, comprising:

obtaining PMU data generated by a phasor measurement device;

2. The method for data processing of phasor measurement devices of a power system according to claim 1, wherein the screening the PMU data using an iterative random forest algorithm to obtain first data comprises:

establishing a plurality of decision trees according to the PMU data;

generating a random forest according to the decision trees;

3. The method of power system phasor measurement device data processing according to claim 2, wherein the random forest voting mechanism is a minority-compliant majority mechanism.

4. The method for processing data of the phasor measurement unit of the electric power system according to claim 1, wherein before the padding of the first data with the trained dynamic neural network to obtain a data processing result, the method further comprises:

establishing a time series model;

5. The method of power system phasor measurement device data processing according to claim 4, wherein said training of a dynamic neural network comprises:

and training the dynamic neural network by adopting an LM algorithm.

6. An apparatus for data processing of a phasor measurement device of a power system, comprising:

the acquisition module is used for acquiring PMU data generated by the phasor measurement device;

7. The power system phasor measurement device data processing device according to claim 6, wherein the processing module is specifically configured to:

establishing a plurality of decision trees according to the PMU data;

generating a random forest according to the decision trees;

8. The power system phasor measurement device data processing apparatus according to claim 6, wherein said processing module is further configured to:

9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method according to any of claims 1 to 5 are implemented when the processor executes the program.

10. A non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 5.