CN116663728A - Electric energy metering device error state prediction method and device - Google Patents
Electric energy metering device error state prediction method and device Download PDFInfo
- Publication number
- CN116663728A CN116663728A CN202310637392.1A CN202310637392A CN116663728A CN 116663728 A CN116663728 A CN 116663728A CN 202310637392 A CN202310637392 A CN 202310637392A CN 116663728 A CN116663728 A CN 116663728A
- Authority
- CN
- China
- Prior art keywords
- prediction
- sequence
- electric energy
- error
- metering device
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000004927 fusion Effects 0.000 claims abstract description 24
- 238000012847 principal component analysis method Methods 0.000 claims abstract description 13
- 238000007499 fusion processing Methods 0.000 claims abstract description 11
- 238000012549 training Methods 0.000 claims description 33
- 238000012545 processing Methods 0.000 claims description 19
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000004458 analytical method Methods 0.000 claims description 13
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 230000008859 change Effects 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000010998 test method Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 abstract description 4
- 239000000047 product Substances 0.000 description 9
- 229920001155 polypropylene Polymers 0.000 description 7
- 238000012360 testing method Methods 0.000 description 6
- 238000000513 principal component analysis Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000002939 conjugate gradient method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005672 electromagnetic field Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Computational Biology (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Operations Research (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Technology Law (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Educational Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention provides a method and a device for predicting error states of an electric energy metering device, which are used for differentiating a non-stationary time sequence to obtain a stationary time sequence; the method comprises the steps of analyzing a historical error sequence through a principal component analysis method to obtain a residual sequence of the historical error sequence, respectively carrying out rigid prediction and flexible prediction on the stable time sequence, carrying out nonlinear prediction on the residual sequence of the historical error sequence, carrying out weighted fusion processing on the prediction result to obtain a prediction value of an error state of the electric energy metering device, respectively predicting the residual sequences of the stable time sequence and the historical error sequence to obtain different prediction results, carrying out data fusion on the prediction results, considering the linear and nonlinear characteristics in the error sequence, enhancing the adaptability of the prediction results to sample data, improving the precision of the prediction results, enhancing the metering accuracy of the electric energy metering device, ensuring the safe and stable operation of an electric power system, and effectively realizing fair trade settlement of huge electric quantity.
Description
Technical Field
The invention relates to the field of electric energy metering, in particular to a method and a device for predicting an error state of an electric energy metering device.
Background
The electric energy metering device is used as an execution entity of electric energy trade settlement, and the metering accuracy is an absolute basis of electric energy trade fairness. With the further progress of the market reform of the power industry, how to ensure the accuracy of the electric energy metering device and maintain the fair and orderly operation of the power market is an important research subject of the power market. With the continuous development of the power grid scale, the main equipment of the power grid is increased, and a robust and intelligent large power grid is also improved.
The electric energy metering device is an important tool for accurately metering, accurately trade settlement, fair and fair transaction and checking of internal economic and technical indexes of the electric power system among power generation companies, power grid companies, power selling companies and electric power users, and the accuracy and stability of operation of the electric energy metering device are directly related to economic benefits of trade parties, and meanwhile, the trade parties have higher requirements on the accuracy and reliability of the electric energy metering device.
However, the metering accuracy of the electric energy metering device is affected by factors such as environmental temperature and humidity, secondary load, electromagnetic field, equipment aging and the like, a plurality of serious problems are exposed, the problems can directly lead to the reduction of the metering accuracy of the electric energy metering device, and the phenomenon of electric energy metering misalignment occurs in serious cases. The continuous operation of the misaligned electric energy metering device brings huge loss to metering trade settlement of three parties for delivery, so that the accuracy and fairness of electric energy metering are questioned, trade settlement problems and even legal disputes are extremely easy to generate, and meanwhile, system misoperation can be caused, and the stable operation of an electric power system is influenced.
The error level of the electric energy metering device is formed by integrating errors of 4 parts of an electric energy meter, a voltage and current transformer and a secondary circuit, and the time sequence of the electric energy metering error presents obvious non-stable change characteristics due to factor diversity and inherent relevance of error sources of all the parts, so that the operation trend of the electric energy metering device is difficult to predict.
Therefore, the operation trend of the electric energy metering device is difficult to predict, the accuracy of the generated error prediction result is low and the difficulty is high, and the metering accuracy of the electric energy metering device is influenced by the error prediction value with low accuracy, so that the trade settlement problem is very easy to generate.
Disclosure of Invention
The invention provides a method and a device for predicting the error state of an electric energy metering device, which can accurately predict the error state of the electric energy metering device, improve the accuracy of predicting the error state, enhance the metering accuracy of the electric energy metering device, ensure the safe and stable operation of an electric power system and effectively realize the fair trade settlement of huge electric quantity.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the invention provides a method for predicting error state of an electric energy metering device, which comprises the following steps,
acquiring a historical error sequence of the electric energy metering device;
judging whether the historical error sequence is a stable time sequence or not through a preset processing model, and performing differential processing on the non-stable time sequence when the historical error sequence is the non-stable time sequence to obtain the stable time sequence;
analyzing the historical error sequence through a preset analysis model by a principal component analysis method to obtain a residual sequence of the historical error sequence;
carrying out rigid prediction on the stable time sequence through a preset rigid prediction model to obtain a first prediction result, and carrying out flexible prediction on the stable time sequence through a preset flexible prediction model to obtain a second prediction result; carrying out nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result;
and carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device.
Preferably, the step of judging whether the historical error sequence is a stable time sequence through a preset processing model specifically comprises the step of carrying out stability verification on the historical error sequence through a PP (propene Polymer) test method.
Preferably, the rigidity prediction is performed on the stationary time sequence through a preset rigidity prediction model, and the obtaining the first prediction result specifically includes,
on a stationary time sequence, data with a time interval of Δt is truncated to construct m sets of data, each set of data is divided into rigid history windows Δt 1 And a rigid future window Δt 2 Wherein Δt=Δt 1 +t 2 ；
Respectively constructing a rigidity prediction feature and a rigidity prediction Target for each group of data, wherein the rigidity prediction feature comprises a rigidity history feature A and a rigidity future feature B; combining the rigidity prediction features of the m groups of data to serve as rigidity training set features; combining the rigidity prediction Target of m groups of data as a rigidity training set Target;
and training the stable time sequence by using the characteristics of the rigid training set and the rigid training set to obtain a rigid prediction model, and obtaining a first prediction result by using the rigid prediction model.
Preferably, the soft prediction is performed on the stable time sequence through a preset soft prediction model to obtain a second prediction result,
on a stationary time sequence, data with a time interval of deltat' is intercepted to construct n groups of data, and each group of data is divided into flexible history windows deltat 1 ' and Flexible future Window Δt 2 'wherein Δt' =t 1 ′+t 2 ′；
Respectively constructing a flexible prediction feature and a flexible prediction Target for each group of data, wherein the flexible prediction feature comprises a flexible history feature C and a flexible future feature D; combining the flexible prediction features of the n groups of data to serve as flexible training set features; merging flexible prediction targets of n groups of data to serve as flexible training sets Target;
and training the stable time sequence by using the flexible training set characteristics and the flexible training set to obtain a flexible prediction model, and obtaining a second prediction result by using the flexible prediction model.
Preferably, the analyzing the historical error sequence by the preset analysis model through the principal component analysis method to obtain the residual sequence of the historical error sequence specifically comprises,
the expression of the residual sequence of the history error sequence is as follows:
where E is a residual sequence of the historical error sequence characterizing the error change information,is a standardized matrix of the historical error sequence, I is an identity matrix, and P A For the load matrix +.>Is the transpose of the load matrix.
Preferably, the step of performing nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result specifically includes,
predicting a residual sequence of the historical error sequence by using a time sequence prediction model based on a multidimensional Taylor network;
the expression of the third prediction result is:
in which W is i i Lambda is the total term number of product term in expansion 1，q For weights before the q-th product term, e j Sigma, the variable in the product term q，j For variable e in the q-th product term j To the power of (3).
Preferably, the weighting fusion processing is performed on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device, which specifically comprises,
fusing the first prediction result, the second prediction result and the third prediction result by using a self-adaptive weighted data fusion algorithm to obtain a prediction value of the error state of the electric energy metering device;
the expression of the predicted value of the error state of the electric energy metering device is as follows:
Y 4 ＝ω 1 Y 1 +ω 2 Y 2 +ω 3 Y 3 ；
ω 1 +ω 2 +ω 3 ＝1；
wherein Y is 1 For the first prediction result, Y 2 For the second prediction result, Y 3 Y being the third predictor 4 Is the predicted value omega of the error state of the electric energy metering device 1 Weight, ω, of the first predictor 2 Weight, ω, of the second predictor 3 Is the weight of the third predictor.
The invention provides an error state prediction device of an electric energy metering device, which comprises,
the acquisition module is used for acquiring a historical error sequence of the electric energy metering device;
the preprocessing module is used for judging whether the historical error sequence is a stable time sequence or not through a preset processing model, and performing differential processing on the non-stable time sequence to obtain the stable time sequence;
the residual analysis module is used for analyzing the historical error sequence through a preset analysis model by using a principal component analysis method to obtain a residual sequence of the historical error sequence;
the prediction module is used for carrying out rigid prediction on the stable time sequence through a preset rigid prediction model to obtain a first prediction result, and carrying out flexible prediction on the stable time sequence through a preset flexible prediction model to obtain a second prediction result; carrying out nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result;
and the fusion module is used for carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device.
The invention proposes an electronic device comprising a processor and a memory, said memory storing at least one instruction, the instructions stored in said memory being executed to implement a method for predicting an error state of an electric energy metering device according to claim 1.
The present invention proposes a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method for predicting an error state of an electric energy metering device according to claim 1.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides a method for predicting error state of an electric energy metering device, which comprises the steps of differentiating a non-stationary time sequence to obtain a stationary time sequence; analyzing the historical error sequence through a principal component analysis method to obtain a residual sequence of the historical error sequence, respectively carrying out rigid prediction and flexible prediction on the stable time sequence to obtain a first prediction result and a second prediction result, carrying out nonlinear prediction on the residual sequence of the historical error sequence to obtain a third prediction result, carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result to obtain a prediction value of an error state of the electric energy metering device, respectively predicting the residual sequences of the stable time sequence and the historical error sequence to obtain different prediction results, carrying out data fusion on the prediction results, and considering the linear and nonlinear characteristics in the error sequence, enhancing the adaptability of the prediction results to sample data, improving the accuracy of the prediction results, enhancing the metering accuracy of the electric energy metering device, ensuring the safe and stable operation of an electric power system and effectively realizing fair trade settlement of huge electric quantity.
Furthermore, the invention analyzes the history error sequence by using a principal component analysis method, so that the characteristics of a nonlinear part in the history error sequence are prominently reflected, the correlation between data is reduced, the data space of the residual sequence of the generated history error sequence is simplified, and the calculation cost is reduced.
The invention provides an error state prediction device of an electric energy metering device, which integrates a stable time sequence with a long historical error sequence and a residual sequence of the historical error sequence, predicts the operation trend of the electric energy metering device, improves the accuracy of error prediction, ensures the metering accuracy of the electric energy metering device to be improved, further ensures the accuracy of trade settlement and avoids risks brought by trade settlement.
Drawings
Fig. 1 is a flowchart of a method for predicting an error state of an electric energy metering device according to the present invention.
Fig. 2 is a technical route of a method for predicting an error state of an electric energy metering device and a device thereof according to the present invention.
Fig. 3 is a diagram of a prediction process of a rigid prediction model of an error state prediction method and device for an electric energy metering device according to the present invention.
Fig. 4 is a block diagram of an apparatus for predicting an error status of an electric energy metering device according to the present invention.
Detailed Description
Reference will now be made in detail to the present embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein the accompanying drawings are used to supplement the description of the written description so that one can intuitively and intuitively understand each technical feature and overall technical scheme of the present invention, but not to limit the scope of the present invention.
In the description of the present invention, it should be understood that references to orientation descriptions such as upper, lower, front, rear, left, right, etc. are based on the orientation or positional relationship shown in the drawings, are merely for convenience of description of the present invention and to simplify the description, and do not indicate or imply that the apparatus or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the present invention.
In the description of the present invention, a number means one or more, a number means two or more, and greater than, less than, exceeding, etc. are understood to not include the present number, and above, below, within, etc. are understood to include the present number. The description of the first and second is for the purpose of distinguishing between technical features only and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of the technical features indicated.
In the description of the present invention, unless explicitly defined otherwise, terms such as arrangement, installation, connection, etc. should be construed broadly and the specific meaning of the terms in the present invention can be reasonably determined by a person skilled in the art in combination with the specific contents of the technical scheme.
The invention will now be described in further detail with reference to specific examples, which are intended to illustrate, but not to limit, the invention.
The invention provides a method for predicting error state of an electric energy metering device, referring to fig. 1 and 2, comprising,
acquiring a historical error sequence of the electric energy metering device;
judging whether the historical error sequence is a stable time sequence or not through a preset processing model, and performing differential processing on the non-stable time sequence when the historical error sequence is the non-stable time sequence to obtain the stable time sequence;
analyzing the historical error sequence through a preset analysis model by a principal component analysis method to obtain a residual sequence of the historical error sequence;
carrying out rigid prediction on the stable time sequence through a preset rigid prediction model to obtain a first prediction result, and carrying out flexible prediction on the stable time sequence through a preset flexible prediction model to obtain a second prediction result; carrying out nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result;
and carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device.
According to the method, the known historical error sequence is analyzed and processed, so that a stable time sequence and a residual sequence of the historical error sequence are obtained, the stable time sequence and the residual sequence of the historical error sequence are respectively predicted, different prediction results are obtained, the prediction results are subjected to data fusion, linear and nonlinear characteristics in the error sequence can be considered, adaptability of the prediction results to sample data is enhanced, accuracy of the prediction results is improved, further prediction of metering errors of the electric energy metering device is achieved, metering accuracy of the electric energy metering device is enhanced, effective prediction of an electric energy metering error state is achieved, accuracy is high, safe and stable operation of an electric power system is guaranteed, and fair trade settlement of huge electric quantity is effectively achieved. So that people can timely avoid adverse effects on fair and orderly operation of the electric power market due to errors.
In a specific embodiment of the present invention, referring to fig. 1, the determining, by a preset processing model, whether the historical error sequence is a stationary time sequence specifically includes performing stationary verification on the historical error sequence by a PP test method.
The stability check is carried out by using a unit root check, namely whether a unit root exists in a check sequence or not is checked by using the unit root check, and a PP check statistic is calculated by adopting a PP check method in the unit root check.
Wherein lambda is 2 (q) is a New-West estimator,for variance obtained when the DF test is performed in advance, gamma 0 For 0 th order auto-covariance, T is the time sequence length, MSE is the mean square error, and the test statistic is compared with a critical value table to make a judgment: given a significance level of 5%, a PP test statistic less than the critical table value indicates that the historical error sequence has no unity root and is a stationary sequence. The PP test statistic being greater than the critical table value indicates that the historical error sequence has a unity root and is a non-stationary sequence.
The differential processing method is to perform first-order differential and second-order differential processing on the history error sequence x (omega):
first order difference: delta 1 x(ω)＝x(ω+1)-x(ω)；
Second order difference: delta 2 x(ω)＝x(ω+2)-2x(ω+1)+x(ω)；
Where ω=1, 2, …, N represents the number of sampling points.
The randomness trend in the sequence is reduced through differential processing, fluctuation is eliminated, so that the data is stable, the existence of pseudo regression is avoided, and the method has practical and applicable significance.
In a specific embodiment of the present invention, referring to fig. 1 and fig. 3, the performing, by using a preset rigidity prediction model, rigidity prediction on the stationary time sequence to obtain a first prediction result specifically includes;
on a stationary time sequence, data with a time interval of Δt is truncated to construct m sets of data, each set of data is divided into rigid history windows Δt 1 And a rigid future window Δt 2 Wherein Δt=Δt 1 +Δt 2 ；
Respectively constructing a rigidity prediction feature and a rigidity prediction Target for each group of data, wherein the rigidity prediction feature comprises a rigidity history feature A and a rigidity future feature B; combining the rigidity prediction features of the m groups of data to serve as rigidity training set features; combining the rigidity prediction Target of m groups of data as a rigidity training set Target;
and training the stable time sequence by using the characteristics of the rigid training set and the rigid training set to obtain a rigid prediction model, and obtaining a first prediction result by using the rigid prediction model.
In a specific embodiment of the present invention, referring to fig. 1, the flexible prediction is performed on the stationary time sequence through a preset flexible prediction model to obtain a second prediction result;
on a stationary time sequence, data with a time interval of deltat 'are intercepted to construct n groups of data, and each group of data is divided into flexible history windows deltat' 1 And a flexible future window Δt' 2 Wherein Δt '=Δt' 1 +Δt′ 2 ；
Respectively constructing a flexible prediction feature and a flexible prediction Target for each group of data, wherein the flexible prediction feature comprises a flexible history feature C and a flexible future feature D; combining the flexible prediction features of the n groups of data to serve as flexible training set features; merging flexible prediction targets of n groups of data to serve as flexible training sets Target;
and training the stable time sequence by using the flexible training set characteristics and the flexible training set to obtain a flexible prediction model, and obtaining a second prediction result by using the flexible prediction model.
The electric energy metering device is gradually deteriorated in the operation process, so that error data at the time t-1 has small fluctuation relative to error data at the time t+1, the cost performance of the data which can be mined is low, and the data is intercepted in a variable time scale mode.
The time interval of the flexible prediction model is:
wherein a is a constant, deltaU T For the variance of the voltage deviation of the electric energy metering device in the time period T, kappa is a constant, T κ At the time of the kappa number,at t κ Time voltage deviation, ">For electric energy metering device t κ Actual voltage at time, ">Rated voltage of the power network at time ∈ ->Is the variance of the voltage deviation over time period T.
Wherein Δt is 1 ' historical window, Δt, which is a flexible predictive model 2 ' is the future window of the flexible predictive model.
In a specific embodiment of the present invention, referring to fig. 1, the preset analysis model analyzes the historical error sequence through a principal component analysis method to obtain a residual sequence of the historical error sequence;
the expression of the residual sequence of the history error sequence is as follows:
where E is a residual sequence of the historical error sequence characterizing the error change information,is a standardized matrix of the historical error sequence, I is an identity matrix, and P A For the load matrix +.>Is the transpose of the load matrix.
Specifically:
PCA processing is carried out on the history error sequence x (omega), and the steps are as follows:
acquiring a historical error sequence x (omega) epsilon R of an electric energy metering device n N is the number of samples;
performing standardization operation on x (omega) to obtain a standardized matrix
Solving forCovariance matrix C of (a);
taking the main component as 1, arranging the eigenvectors from large to small according to the eigenvalues, taking the first column as a load matrix P A ；
Residual sequence of historical error sequence
Wherein E is a residual sequence of a historical error sequence characterizing error variation information,is the main component of the error sequence.
Principal Component Analysis (PCA) is a commonly used method of data analysis that highlights a portion of the important features in the data space, reduces the dimensionality of the data space, and captures possible internal patterns and structures.
Principal component analysis is a means of dimension reduction, which reduces a plurality of nonlinear indexes in a history error sequence to a few important indexes, and stepwise regression retains nonlinear indexes with obvious influence in an original index system.
The basic principle of principal component analysis is to convert a historical error sequence into a new set of representations, called principal components. These principal components may represent the most important features of the historical error sequence, minimizing correlation therebetween. Since the correlation between the principal component important features is minimized, it can effectively simplify the data space of the residual sequence of the history error sequence.
In a specific embodiment of the present invention, referring to fig. 1, the pre-set analysis model analyzes the historical error sequence through a principal component analysis method to obtain a residual sequence of the historical error sequence specifically including,
predicting a residual sequence of the historical error sequence by using a time sequence prediction model based on a multidimensional Taylor network;
the expression of the third prediction result is:
in the method, in the process of the invention,lambda is the total term number of product term in expansion 1, For weights before the q-th product term, e j Sigma, the variable in the product term q,j For variable e in the q-th product term j To the power of (3).
The specific prediction process is as follows:
and carrying out data reconstruction on the residual sequence of the historical error sequence.
Let the residual sequence of the history error sequence be: e= { E (t), E (t-1), …, E (t-n) };
the method comprises the steps of generating an input vector and an output vector required by multidimensional Taylor network prediction according to the existing residual data, wherein the input vector and the output vector are specifically as follows:
input vector: e (t) =e 1 (t)，E 2 (t)，…，E i (t)] T ；
Output vector: e (t+1) wherein i is the dimension of the multidimensional Taylor network and t is the historical time.
Dividing the residual sequence of the reconstructed historical error sequence into two parts: training sets and test sets. The training set is used for optimizing model parameters, and the testing set is used for monitoring the performance of the built model;
and performing multidimensional Taylor network parameter learning of the training set. And carrying out parameter optimization by adopting a conjugate gradient method to obtain optimal MTN network parameters, thereby establishing an MTN prediction model.
And predicting the time sequence and carrying out prediction, and outputting a prediction result.
The multidimensional Taylor network is a three-layer forward network consisting of an input layer, a middle layer and an output layer, wherein the input layer is a controller input vector and is used for realizing the input of the network; the middle layer is a network processing layer and consists of each power product item unit and corresponding connection weight; and the output layer node is used for generating a control input of the controlled object.
The multidimensional Taylor network only utilizes the observation data of the nonlinear time sequence and the multidimensional Taylor network to learn and predict the time sequence under the condition that no priori knowledge of the system is needed, the generalization capability is strong, meanwhile, the multidimensional Taylor network can represent the dynamics characteristic in a general sense and can be arbitrarily approximate to a general nonlinear system, so that the state equation description of the nonlinear system is obtained. The prediction method based on the multidimensional Taylor network has the advantages of simple idea, convenient implementation and high approximation precision.
In a specific embodiment of the present invention, referring to fig. 1, the weighting fusion processing is performed on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of an error state of the electric energy metering device, which specifically includes,
the first prediction result, the second prediction result and the third prediction result are fused by utilizing a self-adaptive weighted data fusion algorithm, so that a prediction value of an error state of the electric energy metering device is obtained;
the expression of the predicted value of the error state of the electric energy metering device is as follows:
Y 4 ＝ω 1 Y 1 +ω 2 Y 2 +ω 3 Y 3 ；
ω 1 +ω 2 +ω 3 ＝1；
wherein Y is 1 For the first prediction result, Y 2 For the second prediction result, Y 3 Y being the third predictor 4 Is the predicted value omega of the error state of the electric energy metering device 1 Weight, ω, of the first predictor 2 Weight, ω, of the second predictor 3 Is the weight of the third predictor.
And fusing the first prediction result, the second prediction result and the third prediction result by using a self-adaptive weighted data fusion algorithm to obtain a prediction value of the error state of the electric energy metering device, so as to realize the prediction of the error state of the electric energy metering device.
The predicted values of the error states of the first predicted result, the second predicted result, the third predicted result and the electric energy metering device are respectively Y 1 、Y 2 、Y 3 、Y 4 ；
Y 4 ＝ω 1 Y 1 +ω 2 Y 2 +ω 3 Y 3 ；
ω 1 +ω 2 +ω 3 ＝1；
Wherein omega 1 、ω 2 、ω 3 Respectively representing the weights of the different prediction results.
Calculating average values of the first prediction result, the second prediction result and the third prediction resultAs a virtual standard predictor;
where l=1, 2, 3, softmax is the normalization function and b is the first threshold.
And calculating and distributing weights of the first prediction result, the second prediction result and the third prediction result through the virtual standard prediction value, and combining the calculated prediction value of the error state of the electric energy metering device by a weighted data fusion algorithm after distributing the weights, wherein the calculated prediction value of the error state of the electric energy metering device is used as the prediction result of the error state of the electric energy metering device.
The self-adaptive weighted fusion algorithm is an algorithm commonly used for multi-sensor data fusion, and can fuse data acquired by different sensors, so that a more accurate result is obtained. The main idea of the algorithm is to weight the data acquired by different sensors according to the reliability and accuracy of the data, and then fuse the weighted data to obtain a final result.
The basic principle of the self-adaptive weighted fusion algorithm is to weight the data acquired by different sensors, so that a more accurate result is achieved. Specifically, the algorithm can weight data acquired by different sensors according to the reliability and accuracy of the sensors. The reliability refers to stability and precision of the sensor in the working process, and the accuracy refers to the degree of coincidence between data acquired by the sensor and actual conditions.
The present invention provides an error state prediction device for an electric energy metering device, referring to fig. 4, comprising,
the acquisition module is used for acquiring a historical error sequence of the electric energy metering device;
the preprocessing module is used for judging whether the historical error sequence is a stable time sequence or not through a preset processing model, and performing differential processing on the non-stable time sequence to obtain the stable time sequence;
the residual analysis module is used for analyzing the historical error sequence through a preset analysis model by using a principal component analysis method to obtain a residual sequence of the historical error sequence;
the prediction module is used for carrying out rigid prediction on the stable time sequence through a preset rigid prediction model to obtain a first prediction result, and carrying out flexible prediction on the stable time sequence through a preset flexible prediction model to obtain a second prediction result; carrying out nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result;
and the fusion module is used for carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device.
And predicting the stable time sequence in the error sequence and the residual sequence of the historical error sequence respectively, and carrying out data fusion, so that the linear and nonlinear characteristics in the error sequence are considered, the adaptability of a prediction model to sample data is enhanced, and the prediction precision is improved.
The invention proposes an electronic device comprising a processor and a memory, said memory storing at least one instruction, the instructions stored in said memory being executed to implement a method for predicting the error state of an electric energy metering device according to claim 1.
The present invention proposes a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method for predicting an error state of an electric energy metering device according to claim 1.
While the invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiment, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.
Claims (10)
1. A method for predicting the error state of electric energy meter is characterized by comprising,
acquiring a historical error sequence of the electric energy metering device;
judging whether the historical error sequence is a stable time sequence or not through a preset processing model, and performing differential processing on the non-stable time sequence when the historical error sequence is the non-stable time sequence to obtain the stable time sequence;
analyzing the historical error sequence through a preset analysis model by a principal component analysis method to obtain a residual sequence of the historical error sequence;
carrying out rigid prediction on the stable time sequence through a preset rigid prediction model to obtain a first prediction result, and carrying out flexible prediction on the stable time sequence through a preset flexible prediction model to obtain a second prediction result; carrying out nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result;
and carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device.
2. The method for predicting an error state of an electric energy metering device according to claim 1, wherein the determining whether the historical error sequence is a stationary time sequence by a preset processing model specifically comprises performing stationary verification on the historical error sequence by a PP test method.
3. The method for predicting an error state of an electric energy meter according to claim 1, wherein the predicting rigidity of the stationary time series by a preset rigidity prediction model to obtain a first prediction result specifically comprises,
on a stationary time sequence, data with a time interval of Δt is truncated to construct m sets of data, each set of data is divided into rigid history windows Δt 1 And a rigid future window Δt 2 Wherein Δt=Δt 1 +t 2 ；
Respectively constructing a rigidity prediction feature and a rigidity prediction Target for each group of data, wherein the rigidity prediction feature comprises a rigidity history feature A and a rigidity future feature B; combining the rigidity prediction features of the m groups of data to serve as rigidity training set features; combining the rigidity prediction Target of m groups of data as a rigidity training set Target;
and training the stable time sequence by using the characteristics of the rigid training set and the rigid training set to obtain a rigid prediction model, and obtaining a first prediction result by using the rigid prediction model.
4. The method for predicting error status of an electric energy meter according to claim 1, wherein the soft prediction is performed on the stationary time series by a preset soft prediction model to obtain a second prediction result,
on a stationary time sequence, data with a time interval of deltat' is intercepted to construct n groups of data, and each group of data is divided into flexible history windows deltat 1 ' and Flexible future Window Δt 2 'wherein Δt' =t 1 ′+t 2 ′；
Respectively constructing a flexible prediction feature and a flexible prediction Target for each group of data, wherein the flexible prediction feature comprises a flexible history feature C and a flexible future feature D; combining the flexible prediction features of the n groups of data to serve as flexible training set features; merging flexible prediction targets of n groups of data to serve as flexible training sets Target;
and training the stable time sequence by using the flexible training set characteristics and the flexible training set to obtain a flexible prediction model, and obtaining a second prediction result by using the flexible prediction model.
5. The method for predicting an error state of an electric energy metering device according to claim 1, wherein the analyzing the historical error sequence by a principal component analysis method through a preset analysis model to obtain a residual sequence of the historical error sequence specifically comprises,
the expression of the residual sequence of the history error sequence is as follows:
where E is a residual sequence of the historical error sequence characterizing the error change information,is a standardized matrix of the historical error sequence, I is an identity matrix, and P A For the load matrix +.>Is the transpose of the load matrix.
6. The method for predicting an error state of an electric energy metering device according to claim 1, wherein the step of performing nonlinear prediction on the residual sequence of the historical error sequence by using a preset residual prediction model to obtain a third prediction result specifically comprises,
predicting a residual sequence of the historical error sequence by using a time sequence prediction model based on a multidimensional Taylor network;
the expression of the third prediction result is:
in which W is i i Lambda is the total term number of product term in expansion 1，q For weights before the q-th product term, e j Sigma, the variable in the product term q，j For variable e in the q-th product term j To the power of (3).
7. The method for predicting the error state of an electric energy metering device according to claim 1, wherein the weighting fusion processing is performed on the first, second and third predicted results through a preset fusion model to obtain the predicted value of the error state of the electric energy metering device, specifically comprising,
fusing the first prediction result, the second prediction result and the third prediction result by using a self-adaptive weighted data fusion algorithm to obtain a prediction value of the error state of the electric energy metering device;
the expression of the predicted value of the error state of the electric energy metering device is as follows:
Y 4 ＝ω 1 Y 1 +ω 2 Y 2 +ω 3 Y 3 ；
ω 1 +ω 2 +ω 3 ＝1；
wherein Y is 1 For the first prediction result, Y 2 For the second prediction result, Y 3 Y being the third predictor 4 Is the predicted value omega of the error state of the electric energy metering device 1 Weight, ω, of the first predictor 2 Weight, ω, of the second predictor 3 Is the weight of the third predictor.
8. An error state prediction device of an electric energy metering device is characterized by comprising,
the acquisition module is used for acquiring a historical error sequence of the electric energy metering device;
the preprocessing module is used for judging whether the historical error sequence is a stable time sequence or not through a preset processing model, and performing differential processing on the non-stable time sequence to obtain the stable time sequence;
the residual analysis module is used for analyzing the historical error sequence through a preset analysis model by using a principal component analysis method to obtain a residual sequence of the historical error sequence;
the prediction module is used for carrying out rigid prediction on the stable time sequence through a preset rigid prediction model to obtain a first prediction result, and carrying out flexible prediction on the stable time sequence through a preset flexible prediction model to obtain a second prediction result; carrying out nonlinear prediction on the residual sequence of the historical error sequence through a preset residual prediction model to obtain a third prediction result;
and the fusion module is used for carrying out weighted fusion processing on the first prediction result, the second prediction result and the third prediction result through a preset fusion model to obtain a prediction value of the error state of the electric energy metering device.
9. An electronic device comprising a processor and a memory, the memory storing at least one instruction, the instructions stored in the memory being executable to implement a method of predicting an error condition of an electric energy metering device as claimed in claim 1.
10. A computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, which when executed by a processor, implements a method for predicting an error state of an electric energy metering device according to claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310637392.1A CN116663728A (en) | 2023-05-31 | 2023-05-31 | Electric energy metering device error state prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310637392.1A CN116663728A (en) | 2023-05-31 | 2023-05-31 | Electric energy metering device error state prediction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116663728A true CN116663728A (en) | 2023-08-29 |
Family
ID=87727397
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310637392.1A Pending CN116663728A (en) | 2023-05-31 | 2023-05-31 | Electric energy metering device error state prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116663728A (en) |
-
2023
- 2023-05-31 CN CN202310637392.1A patent/CN116663728A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Petrone et al. | A review on model-based diagnosis methodologies for PEMFCs | |
| EP3477404A1 (en) | Power system model parameter conditioning tool | |
| CN109818349B (en) | Power grid robust state prediction method based on multidimensional state matrix sliding matching | |
| CN112731260B (en) | Online evaluation method for error state of voltage transformer based on concept drift recognition | |
| KR102646875B1 (en) | Mock cell construction method and mock cell construction device | |
| CN113255584B (en) | Fault diagnosis and monitoring system based on edge calculation | |
| CN114169631B (en) | Oil field power load management and control system based on data analysis | |
| Lin et al. | Reliability prediction of battery management system for electric vehicles based on accelerated degradation test: A semi-parametric approach | |
| CN116401532B (en) | Method and system for recognizing frequency instability of power system after disturbance | |
| CN115469260A (en) | Hausdorff-based current transformer abnormity identification method and system | |
| CN115408864A (en) | Electronic transformer error state self-adaptive prediction method, system and equipment | |
| CN116679211A (en) | Lithium battery health state prediction method | |
| Liu et al. | State of health estimation of lithium-ion batteries based on multi-feature extraction and temporal convolutional network | |
| CN117330963A (en) | Energy storage power station fault detection method, system and equipment | |
| CN116565861B (en) | Power distribution network reliability assessment method, system, equipment and medium | |
| CN116663728A (en) | Electric energy metering device error state prediction method and device | |
| CN114841222B (en) | Method for constructing fault detection model of high-voltage circuit breaker, detection method and power distribution system | |
| CN114814707A (en) | Intelligent ammeter stress error analysis method, equipment, terminal and readable medium | |
| CN115062533A (en) | Construction and prediction method and medium of multiple regression prediction model | |
| CN112098066A (en) | High-voltage shunt reactor fault diagnosis method and system based on gate control circulation unit | |
| CN113994276A (en) | Method for verifying system parameters of an energy system, method for operating an energy system, and energy management system for an energy system | |
| CN113657623B (en) | Power equipment state diagnosis effect determining method, device, terminal and storage medium | |
| CN117367751B (en) | Performance detection method and device for ultra-pulse thulium-doped laser | |
| CN116452070B (en) | Large-scale equipment health assessment method and device under multi-identification framework | |
| Zhang et al. | Performance Evaluation for Smart Electricity Meters Using Machine Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |