CN115660214A - Method and device for predicting medium and long-term load of power system - Google Patents
Method and device for predicting medium and long-term load of power system Download PDFInfo
- Publication number
- CN115660214A CN115660214A CN202211415962.4A CN202211415962A CN115660214A CN 115660214 A CN115660214 A CN 115660214A CN 202211415962 A CN202211415962 A CN 202211415962A CN 115660214 A CN115660214 A CN 115660214A
- Authority
- CN
- China
- Prior art keywords
- prediction
- parameter
- target
- determining
- parameters
- 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
Images
Classifications
-
- 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
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a device for predicting medium and long-term load of a power system, wherein the method comprises the following steps: determining a first prediction parameter, a target parameter and a target year; determining the correlation degree between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter; according to the correlation degree, taking a sub-parameter of which the correlation degree exceeds a preset threshold value as a second prediction parameter; determining a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year. The selection of the sub-parameters is carried out through the calculation of the correlation degree between the prediction parameters and the target parameters, the screening of a plurality of prediction parameters most related to the current target parameters under different scenes is realized, the prediction algorithm is adaptively selected to complete the medium-term and long-term load prediction, and the flexibility of the load prediction is improved.
Description
Technical Field
The invention relates to the technical field of power system automation, in particular to a method and a device for predicting medium and long term load of a power system.
Background
The power load prediction is an important link of power system planning, and the planning of design, scheduling, construction, maintenance and other aspects of each infrastructure in the power system is completed by predicting the load value in a certain time period in the future. The time scale of the medium and long term load prediction is usually a long-term view year in the coming years or more, and provides basis for project plan implementation, power planning compilation, strategic planning and the like in a corresponding planning period. The existing power system medium and long term load forecasting means comprise a trend extrapolation method, a regression analysis method, a time series method, a grey forecasting method, an expert system, a wavelet analysis method, a neural network and other intelligent algorithms.
However, the prediction data adopted by these algorithms are not rich enough, and cannot comprehensively indicate the factors affecting the load change, and there are limitations in data selection and application scenarios, so these algorithms usually have a problem of flexibility.
Disclosure of Invention
The invention aims to solve the technical problems that prediction data adopted by a medium-and-long-term load prediction method is not rich enough, factors influencing load change cannot be comprehensively indicated, and the problems of data selection, application scene limitation and application flexibility exist.
In order to solve the above technical problem, a first aspect of the present invention discloses a method for predicting a medium-term load and a long-term load in an electrical power system, including:
determining a first prediction parameter, a target parameter and a target year;
wherein the first prediction parameters comprise a plurality of parameter sets, and the parameter sets respectively comprise a plurality of sub-parameters; the target parameter comprises at least one of the following parameters: the maximum load of the whole society, the electricity consumption of the whole society and the electricity consumption of the whole industry;
determining the correlation degree between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter;
according to the correlation degree, taking a sub-parameter of which the correlation degree exceeds a preset threshold value as a second prediction parameter;
determining a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year.
As an optional implementation manner, after the sub-parameter of which the correlation degree exceeds the preset threshold is taken as the second prediction parameter according to the correlation degree, the method further includes:
selecting at least two of the subparameters of the second prediction parameter as third prediction parameters;
determining a multivariate goodness-of-fit between the third prediction parameter and the target parameter, and determining a goodness-of-fit judgment threshold;
determining a multivariate goodness-of-fit residual error according to adjacent two iteration results;
if the multivariate goodness-of-fit residual tends to converge, judging whether the multivariate goodness-of-fit exceeds the goodness-of-fit judgment threshold; and if so, replacing the second prediction parameter with the third prediction parameter.
As an optional implementation manner, before replacing the second prediction parameter with the third prediction parameter, the method further includes:
determining a demand dimension for the second prediction parameter;
performing dimensionality reduction transformation on the third prediction parameter through a principal component analysis method according to the demand dimensionality to obtain a transformed third prediction parameter; the transformed third prediction parameters comprise principal component components of demand dimensions of the third prediction parameters;
the replacing the second prediction parameter with the third prediction parameter includes:
replacing the second prediction parameters with the transformed third prediction parameters.
As an optional implementation, the determining, according to the data of the historical year of the first prediction parameter and the data of the historical year of the target parameter, a correlation between each sub-parameter in the first prediction parameter and the target parameter includes:
determining parameter set correlations between the plurality of parameter sets and the target parameter;
determining at least one target parameter group according to the parameter group correlation;
and determining the correlation degree between each sub-parameter in the target parameter group and the target parameter according to the data of the historical years of the target parameter group and the data of the historical years of the target parameter.
As an alternative embodiment, the prediction algorithm includes at least two, and the obtaining the prediction result of the target parameter in the target year according to the prediction algorithm includes:
determining weights corresponding to the at least two prediction algorithms;
and obtaining a weighted prediction result of the target parameter in the target year according to the at least two prediction algorithms and the corresponding weights.
As an optional implementation, after determining the prediction algorithm according to the second prediction parameter, the method further includes:
determining an integration strategy according to the prediction algorithm;
integrating the prediction algorithm according to the integration strategy to obtain a strong learner based on the prediction algorithm;
and obtaining a prediction result of the target parameter under the target year according to the strong learner.
As an optional implementation, after obtaining the prediction result of the target parameter in the target year according to the prediction algorithm, the method further includes:
determining at least one site-to-be-selected point in the area to be planned according to the load prediction result;
and generating a load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm according to the quasi-addressing points.
As an optional implementation manner, the generating, according to the quasi addressing points, a load area corresponding to each quasi addressing point through a thieson polygon algorithm includes:
determining the load density of the area to be planned according to the load prediction result;
determining the expansion speed in the corresponding direction of the corresponding area in the Thiessen polygon generation process according to the load density, and generating the load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm;
wherein the generation process of the Thiessen polygon comprises the following steps: and expanding the area at a corresponding expansion speed in the outward direction along each radius by taking each address point to be selected as the center of a circle and each center of a circle as the reference until the area is contacted with the areas corresponding to other address points to be selected.
A second aspect of the present invention provides a medium-and-long-term load prediction apparatus for an electric power system, including:
the parameter determination module is used for determining a first prediction parameter, a target parameter and a target year;
wherein the first prediction parameters comprise a plurality of parameter sets, and the parameter sets respectively comprise a plurality of sub-parameters; the target parameter comprises at least one of the following parameters: the maximum load of the whole society, the electricity consumption of the whole society and the electricity consumption of the whole industry;
the relevancy determination module is used for determining the relevancy between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter;
the parameter determining module is further used for taking the sub-parameters of which the correlation degrees exceed a preset threshold value as second prediction parameters according to the correlation degrees;
the prediction module is used for determining a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year.
In an optional embodiment, the apparatus further includes a parameter replacement module, configured to, after the parameter determination module determines, according to the correlation, a sub-parameter whose correlation exceeds a preset threshold as the second prediction parameter,
selecting at least two of the sub-parameters of the second prediction parameter as third prediction parameters;
determining a multivariate goodness of fit between the third prediction parameter and the target parameter, and determining a goodness of fit judgment threshold;
determining a multivariate goodness-of-fit residual error according to the adjacent two iteration results;
if the multivariate goodness-of-fit residual tends to converge, judging whether the multivariate goodness-of-fit exceeds the goodness-of-fit judgment threshold; and if so, replacing the second prediction parameter with the third prediction parameter.
As an optional implementation manner, the parameter replacement module is further configured to, before replacing the second prediction parameter with the third prediction parameter,
determining a demand dimension for the second prediction parameter;
performing dimensionality reduction transformation on the third prediction parameter through a principal component analysis method according to the demand dimensionality to obtain a transformed third prediction parameter; the transformed third prediction parameters comprise main component components of demand dimensions of the third prediction parameters;
the replacing the second prediction parameter with the third prediction parameter includes:
replacing the second prediction parameters with the transformed third prediction parameters.
As an optional implementation manner, the specific manner of determining the correlation between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical year of the first prediction parameter and the data of the historical year of the target parameter by the correlation determination module includes:
determining parameter set correlations between the plurality of parameter sets and the target parameter;
determining at least one target parameter group according to the parameter group correlation;
and determining the correlation degree between each sub-parameter in the target parameter group and the target parameter according to the data of the historical years of the target parameter group and the data of the historical years of the target parameter.
As an optional implementation, the prediction algorithm includes at least two, and the specific manner of obtaining the predicted result of the target parameter in the target year by the prediction module according to the prediction algorithm includes:
determining weights corresponding to the at least two prediction algorithms;
and obtaining a weighted prediction result of the target parameter in the target year according to the at least two prediction algorithms and corresponding weights.
In an optional embodiment, the apparatus further comprises an ensemble learning module for determining a prediction algorithm according to the second prediction parameter after the prediction module determines the prediction algorithm,
determining an integration strategy according to the prediction algorithm;
integrating the prediction algorithm according to the integration strategy to obtain a strong learner based on the prediction algorithm;
and obtaining a prediction result of the target parameter under the target year according to the strong learner.
As an optional implementation manner, the device further comprises an addressing module for obtaining the prediction result of the target parameter in the target year after the prediction module obtains the prediction result of the target parameter in the target year according to the prediction algorithm,
determining at least one site-to-be-selected point in the area to be planned according to the load prediction result;
and generating a load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm according to the quasi-addressing points.
As an optional implementation manner, a specific manner in which the addressing module generates the load areas corresponding to the quasi addressing points through a thiessen polygon algorithm according to the quasi addressing points includes:
determining the load density of the area to be planned according to the load prediction result;
determining the expansion speed in the corresponding direction of the corresponding area in the Thiessen polygon generation process according to the load density, and generating the load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm;
wherein the generation process of the Thiessen polygon comprises the following steps: and expanding the area at a corresponding expansion speed in the outward direction along each radius by taking each address point to be selected as the center of a circle and each center of a circle as the reference until the area is contacted with the areas corresponding to other address points to be selected.
The third aspect of the present invention discloses another medium-and-long-term load prediction apparatus in an electric power system, the apparatus comprising:
a memory storing executable program code;
a processor coupled with the memory;
the processor calls the executable program codes stored in the memory to execute the method for predicting the medium and long term load of the power system disclosed by the first aspect of the invention.
In a fourth aspect of the present invention, a computer storage medium is disclosed, wherein the computer storage medium stores computer instructions, and when the computer instructions are called, the computer instructions are used for executing the method for predicting the medium and long term load of the power system disclosed in the first aspect of the present invention.
Compared with the prior art, the embodiment of the invention has the following beneficial effects: the selection of the sub-parameters is carried out through the calculation of the correlation degree between the prediction parameters and the target parameters, the screening of a plurality of prediction parameters most related to the current target parameters under different scenes is realized, the prediction algorithm is adaptively selected to complete the medium-term and long-term load prediction, the adopted prediction data is rich, the factors influencing the load change can be comprehensively indicated, the limitations of data selection and application scenes are avoided, and the flexibility of the load prediction is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flowchart illustrating a method for predicting a medium-term load and a long-term load in an electrical power system according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of a method for predicting a medium-long term load in an electrical power system according to a second embodiment of the present invention;
fig. 3 is a schematic structural diagram of a long-term load prediction apparatus in an electrical power system according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of another long-term load prediction apparatus in an electrical power system according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of a long-term load prediction apparatus in an electrical power system according to a fourth embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terms "first," "second," and the like in the description and claims of the present invention and in the above-described drawings are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, article, or article that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or article.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the invention. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein may be combined with other embodiments.
The power load prediction is an important link of power system planning, and the planning of design, scheduling, construction, maintenance and other aspects of each infrastructure in the power system is completed by predicting the load value in a certain time period in the future. The time scale of the medium and long term load prediction is usually a long-term view year in the coming years or more, and provides basis for project plan implementation, power planning compilation, strategic planning and the like in a corresponding planning period. The existing power system medium and long term load forecasting means comprise a trend extrapolation method, a regression analysis method, a time series method, a grey forecasting method, an expert system, a wavelet analysis method, a neural network and other intelligent algorithms.
Specifically, the trend extrapolation method requires a small amount of data and is convenient to use. However, the variation of the power load has diversity and randomness, so that it is difficult to find a suitable curve to express the variation rule of a specific system, and the fitting process is more complicated due to the influence of other factors. If a simple linear fit is used, it is difficult to take into account all the influencing factors; if polynomial fitting is used, the higher the number of fitting times, the more complex the formula is, and it is difficult to express the specific situation of load change. In addition, the trend extrapolation method has poor prediction effect on the inflection point of the load curve.
The principle of the regression analysis method is simple and easy to understand, but for a complex situation, various influence factors cannot be considered, an accurate model is difficult to establish, and the error is large for an area with impact load, so that the accuracy requirement cannot be met.
The model of the time series method must be identified by a stationary random process, i.e. the time series of the load is smoothed. After the model is established, the stable time sequence is analyzed, the model is identified, parameters are initialized, and the applicability of the model is detected. The model identification process mainly comprises the steps of analyzing the time sequence average value, the autocorrelation function and the partial correlation function, establishing a model structure and calculating parameters of the model structure. The load predicted by the method is continuous, the required data volume is small, and other influencing factors cannot be added. If the time series is not stationary, a smoothing process is performed, which makes the subsequent modeling identification difficult. For areas with large load fluctuation, the prediction difficulty is large.
The gray prediction method is generally suitable for predicting the condition that the load presents exponential change, and the load prediction result may not be ideal enough for other changing conditions, so that the prediction precision is reduced.
The intelligent algorithm includes an expert system method, a wavelet analysis method, an artificial neural network method and the like.
The expert system is difficult to convert the knowledge and experience of the expert into rules, the difficulty in establishing a knowledge base is high, the process of accurately analyzing data and obtaining results is time-consuming, the reestablishment of the knowledge during updating is complex, the maintainability is poor, and the determination of the influence of other factors is difficult.
For the load prediction of the power system, the wavelet transformation can decompose different frequency signals, select different wavelets for prediction according to loads with different properties, and then reconstruct the decomposed sequence to obtain a load prediction result. However, since reconstruction may cause error accumulation, the higher the accuracy requirement for the wavelet coefficient sequence, the more complex the model.
In conclusion, the prediction data adopted by the algorithms are not rich enough, the factors influencing the load change cannot be comprehensively indicated, and the data selection and application scenes are limited, so that the algorithms usually have the problem of insufficient prediction accuracy.
The invention discloses a method and a device for predicting medium and long term load of an electric power system, which select sub-parameters through the calculation of the correlation degree between prediction parameters and target parameters, realize the screening of a plurality of prediction parameters most related to the current target parameters under different scenes, and adaptively select a prediction algorithm to complete the medium and long term load prediction.
Example one
Referring to fig. 2, fig. 2 is a schematic flowchart illustrating a method for predicting a medium-term load and a long-term load in an electrical power system according to an embodiment of the present invention. As shown in fig. 2, the method for predicting the medium-and-long-term load in the power system may include the following operations:
s101, determining a first prediction parameter, a target parameter and a target year;
wherein the first prediction parameters comprise a plurality of parameter sets, and the plurality of parameter sets respectively comprise a plurality of sub-parameters; the target parameter comprises at least one of the following parameters: the maximum load of the whole society, the electricity consumption of the whole society and the electricity consumption of the whole industry. The target parameters can also comprise a plurality of sub-target parameters which can be determined according to the demand of specific load prediction.
S102, determining the correlation degree between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter;
according to the data of historical years, the degree of association between the first prediction parameter and the target parameter can be obtained, and the degree of association is measured through the degree of association.
As an optional implementation, S102 includes:
determining a parameter set correlation between the plurality of parameter sets and the target parameter;
the first prediction parameter is composed of a plurality of parameter sets reflecting different information, the parameter sets can comprise regional load data, regional power consumption data, industrial power consumption data, meteorological data and the like, and a plurality of sub-parameters can be further included under each parameter set, so that the parameter sets can be screened firstly, and the parameter set correlation degree between the plurality of parameter sets and the target parameter can be determined. If a sub-parameter with a higher correlation is present in a parameter set with a lower correlation, it is more likely to interpret this high correlation as a coupling effect of the sub-parameter with some sub-parameters in other parameter sets than an effect of the sub-parameter itself.
Determining at least one target parameter group according to the parameter group correlation;
after determining the correlation of the parameter sets, a threshold value may be set accordingly, and the parameter sets with the correlation greater than the threshold value are determined as the at least one target parameter set.
And determining the correlation degree between each sub-parameter in the target parameter group and the target parameter according to the data of the historical years of the target parameter group and the data of the historical years of the target parameter.
The overall correlation between the target parameter set and the target parameter is high, for example, when the meteorological data is determined as the target parameter set, it indicates that the meteorological data has stronger interpretability for the target parameter, so that partial sub-parameters can be screened out from the meteorological data.
By judging the relevance of the target parameter group and judging and screening the relevance of each sub-parameter in the target parameter group, the effectiveness of sub-parameter screening is improved, and the accuracy of a prediction result is further improved.
S103, according to the correlation degree, taking the sub-parameter of which the correlation degree exceeds a preset threshold value as a second prediction parameter;
as described above, a threshold of the correlation degree may be preset in each sub-parameter as a criterion of the correlation measure, and the sub-parameter exceeding the threshold may be considered as having a strong correlation with the target parameter, so that the sub-parameter having the correlation degree exceeding the preset threshold may be used as the second prediction parameter.
As an optional implementation manner, after S103, the method further includes:
selecting at least two of the sub-parameters of the second prediction parameter as third prediction parameters;
after the second prediction parameter is determined, the cumulative effect of each sub-parameter needs to be identified, and the number of sub-parameters to be applied by the model and the influence range of the cumulative effect need to be determined. For the initialization of the cumulative effect recognition, at least two of the sub-parameters of the second prediction parameter may be selected as the third prediction parameters.
Determining a multivariate goodness-of-fit between the third prediction parameter and the target parameter, and determining a goodness-of-fit judgment threshold;
the goodness of fit of the multivariate may be determined by reconciling the decision coefficients, with a coefficient closer to 1 indicating greater interpretability of the target parameter by the third prediction parameter. The goodness-of-fit judgment threshold can be determined according to an actual prediction task, one feasible value is 0.85, if the multi-element goodness-of-fit is greater than the threshold, it is indicated that the selected third prediction parameter has an accumulative effect on the target parameter, and the action range of the accumulative effect is the iteration number.
Determining a multivariate goodness-of-fit residual error according to the adjacent two iteration results;
and finishing iteration when the multivariate goodness-of-fit residual tends to be converged or is less than 0, and reselecting a group of third prediction parameters to identify the cumulative effect.
If the multivariate goodness-of-fit residual tends to converge, judging whether the multivariate goodness-of-fit exceeds the goodness-of-fit judgment threshold; and if so, replacing the second prediction parameter with the third prediction parameter.
A set of sub-parameters meeting the cumulative effect decision requirement can be used as parameters for determining a prediction algorithm and performing load prediction.
Through the selection of the sub-parameters in the second prediction parameter, the identification of the cumulative effect of the selected third prediction parameter is completed through the multiple fitting goodness, the selection effect of the sub-parameters can be improved, and the accuracy of the prediction result is further improved.
As an optional implementation manner, before replacing the second prediction parameter with the third prediction parameter, the method further includes:
determining a demand dimension for the second prediction parameter;
the demand dimension can be obtained through the iterative process of the cumulative effect judgment, and can also be preset through a specific application scene.
Performing dimensionality reduction transformation on the third prediction parameter through a principal component analysis method according to the demand dimensionality to obtain a transformed third prediction parameter; the transformed third prediction parameters comprise principal component components of demand dimensions of the third prediction parameters;
the principal component analysis method can determine the number of the most significant influencing dimensions in a set of parameters, and can also be used for determining the required dimensions. In addition, if the demand dimension is determined, the principal component analysis method may be further configured to perform a dimension reduction transformation on the third prediction parameter, where the transformed third prediction parameter includes the demand dimension principal component components of the third prediction parameter.
The replacing the second prediction parameter with the third prediction parameter includes:
replacing the second prediction parameter with the transformed third prediction parameter.
And the third prediction parameter after the dimensionality reduction of the principal component analysis method is finished can be replaced to the second prediction parameter, and the third prediction parameter after the dimensionality reduction is used as a parameter group for determining a prediction algorithm and carrying out load prediction.
The dimensionality reduction of the third prediction parameter is completed by a principal component analysis method through the determination of the demand dimensionality, the data applied by the prediction model is simplified, and the prediction efficiency and accuracy are improved.
S104, determining a prediction algorithm according to the second prediction parameter; according to the prediction algorithm, obtaining a prediction result of the target parameter in the target year;
the selection of the prediction algorithm may be performed based on data characteristics of sub-parameters in the second prediction parameter, and since load prediction is a regression task, the prediction algorithm may include at least one of: elastic regression, lasso regression, ridge regression, gaussian process regression, gated cyclic unit regression.
As an alternative embodiment, the prediction algorithm includes at least two kinds, and the S104 includes:
determining weights corresponding to the at least two prediction algorithms;
when multiple prediction algorithms are adopted to predict the same or a group of parameters, weights can be assigned to different algorithms according to principles and applicable scenes, so that the model has better applicability.
And obtaining a weighted prediction result of the target parameter in the target year according to the at least two prediction algorithms and the corresponding weights.
The predicted result may be a variation curve of the corresponding result, and accordingly, the variation curve is a weighted result of combining a plurality of prediction algorithms.
By the weighted prediction of a plurality of algorithms, the applicability and accuracy of the prediction are improved.
The embodiment provides a method for predicting medium and long-term load in an electric power system, which comprises the following steps: determining a first prediction parameter, a target parameter and a target year; determining the correlation degree between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter; according to the correlation degree, taking a sub-parameter of which the correlation degree exceeds a preset threshold value as a second prediction parameter; determining a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year. The selection of the sub-parameters is carried out through the calculation of the correlation degree between the prediction parameters and the target parameters, the screening of a plurality of prediction parameters most relevant to the current target parameters under different scenes is realized, the prediction algorithm is adaptively selected to complete the medium-term and long-term load prediction, and the flexibility of the load prediction is improved.
Example two
Referring to fig. 2, fig. 2 is a flowchart illustrating a method for predicting a medium-long term load in an electrical power system according to a second embodiment of the present invention. As shown in fig. 2, in accordance with any other embodiment, the method further includes:
s201, determining a first prediction parameter, a target parameter and a target year;
s202, according to the data of the historical years of the first prediction parameters and the data of the historical years of the target parameters, determining the correlation degree between each sub-parameter in the first prediction parameters and the target parameters;
s203, according to the correlation degree, taking the sub-parameter of which the correlation degree exceeds a preset threshold value as a second prediction parameter;
s204, determining a prediction algorithm according to the second prediction parameter;
in the second embodiment of the present invention, for descriptions of the same parts as those in step 104 in step 201, step 202, step 203, and step 204, please refer to the detailed descriptions of step 101 to step 104 in the first embodiment, which will not be described again in the second embodiment of the present invention.
S205, determining an integration strategy according to the prediction algorithm;
based on the model characteristics of the predictive algorithm, a corresponding integration strategy may be selected, which may include at least one of the following, based on whether the types of learners participating in the integration are the same: homogeneous integration and heterogeneous integration. In addition, the dividing manner of the integration policy may further include: the bagging method, the boosting method and the stacking method, and the specific integration strategy is not limited thereto.
S206, integrating the prediction algorithm according to the integration strategy to obtain a strong learner based on the prediction algorithm;
by integrating the respective advantages of a plurality of weak learners, the accuracy and robustness of the model can be improved, and a better prediction effect can be ensured in different application scenes.
And S207, obtaining a prediction result of the target parameter in the target year according to the strong learner.
As an optional implementation manner, after S104, the method further includes:
determining at least one site-to-be-selected point in the area to be planned according to the load prediction result;
and according to the determined medium-long term coincidence prediction result, at least one site to be selected of the corresponding year in the area to be planned can be obtained through a corresponding algorithm.
And generating a load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm according to the quasi-addressing points.
And according to the determined quasi addressing points, taking the determined quasi addressing points as reference points in the Thiessen polygons, generating the Thiessen polygons corresponding to the quasi addressing points, and forming load areas corresponding to the quasi addressing points.
The method for generating the Thiessen polygon is used for completing the determination of the planned site selection point of the transformer substation and the corresponding load area thereof through the load prediction result of the area to be planned, and the practicability of medium and long-term load prediction is improved.
As an optional implementation manner, the generating, according to the quasi site, a load area corresponding to each quasi site through a thiessen polygon algorithm includes:
determining the load density of the area to be planned according to the load prediction result;
the load prediction result can comprise the load density of each part in the area to be planned, and for the part with high load density, the power supply range of the substation should be reduced correspondingly.
Determining the expansion speed in the corresponding direction of the corresponding area in the Thiessen polygon generation process according to the load density, and generating the load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm;
wherein the generation process of the Thiessen polygon comprises the following steps: and expanding the area at a corresponding expansion speed in the outward direction along each radius by taking each to-be-addressed point as a circle center and taking each circle center as a reference until the area is contacted with the areas corresponding to other to-be-addressed points.
When the Thiessen polygon is generated by adopting a circular expansion method, under the condition of uneven load density, the expansion speed is in inverse proportion to the load density of the corresponding position, if the load density of the corresponding position is higher, the speed in the outward expansion direction along the radius needs to be correspondingly reduced, and the corresponding relation between the load density and the expansion speed can be determined by a fitting formula or an artificial intelligence method and the like.
The expanding speed of the Thiessen polygon is determined through the load density, the power supply range of each transformer substation is further determined, the accuracy and effectiveness of site selection and load area planning of the transformer substations are improved, and the practicability of medium-term and long-term load prediction is further improved.
The embodiment provides a method for forecasting loads in a power system for medium and long periods, which includes screening sub-parameters through correlation degrees, determining corresponding forecasting algorithms, and integrating the forecasting algorithms to obtain a strong learner for forecasting the loads, so that accuracy, robustness and practicability of load forecasting results are improved.
EXAMPLE III
A third embodiment of the present invention further provides a device for predicting a medium-and-long-term load in an electrical power system to implement the foregoing method, please refer to fig. 4, and fig. 4 is a schematic structural diagram of the device for predicting a medium-and-long-term load in an electrical power system disclosed in the third embodiment of the present invention. As shown in fig. 4, according to any of the other embodiments, the apparatus includes:
a parameter determining module 31, configured to determine a first prediction parameter, a target parameter, and a target year;
wherein the first prediction parameters comprise a plurality of parameter sets, and the plurality of parameter sets respectively comprise a plurality of sub-parameters; the target parameter includes at least one of the following parameters: the maximum load of the whole society, the electricity consumption of the whole society and the electricity consumption of the whole industry;
a correlation determination module 32, configured to determine, according to the data of the historical year of the first prediction parameter and the data of the historical year of the target parameter, a correlation between each sub-parameter in the first prediction parameter and the target parameter;
the parameter determining module 31 is further configured to, according to the correlation, use a sub-parameter of which the correlation exceeds a preset threshold as a second prediction parameter;
a prediction module 33, configured to determine a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year.
The selection of the sub-parameters is carried out through the calculation of the correlation degree between the prediction parameters and the target parameters, the screening of a plurality of prediction parameters most relevant to the current target parameters under different scenes is realized, the prediction algorithm is adaptively selected to complete the medium-term and long-term load prediction, and the flexibility of the load prediction is improved.
Referring to fig. 4, fig. 4 is a schematic structural diagram of a long-term load prediction apparatus in an electrical power system according to a third disclosure of the embodiment of the present invention, as an optional implementation manner, the apparatus further includes a parameter replacement module 34, configured to, after the parameter determination module 31 uses a sub-parameter of which the correlation degree exceeds a preset threshold as a second prediction parameter according to the correlation degree,
selecting at least two of the subparameters of the second prediction parameter as third prediction parameters;
determining a multivariate goodness-of-fit between the third prediction parameter and the target parameter, and determining a goodness-of-fit judgment threshold;
determining a multivariate goodness-of-fit residual error according to the adjacent two iteration results;
if the multivariate goodness-of-fit residual tends to converge, judging whether the multivariate goodness-of-fit exceeds the goodness-of-fit judgment threshold; and if so, replacing the second prediction parameter with the third prediction parameter.
Through the selection of the sub-parameters in the second prediction parameter, the identification of the cumulative effect of the selected third prediction parameter is completed through the multivariate goodness of fit, the selection effect of the sub-parameters can be improved, and the accuracy of the prediction result is further improved.
In an alternative embodiment, the parameter replacement module 34 is further configured to, before replacing the second prediction parameter with the third prediction parameter,
determining a demand dimension for the second prediction parameter;
performing dimensionality reduction transformation on the third prediction parameter through a principal component analysis method according to the demand dimensionality to obtain a transformed third prediction parameter; the transformed third prediction parameters comprise main component components of demand dimensions of the third prediction parameters;
the replacing the second prediction parameter with the third prediction parameter includes:
replacing the second prediction parameter with the transformed third prediction parameter.
The dimensionality reduction of the third prediction parameter is completed by a principal component analysis method through the determination of the demand dimensionality, the data applied by the prediction model is simplified, and the prediction efficiency and accuracy are improved.
As an optional implementation manner, the specific manner of determining the correlation between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical year of the first prediction parameter and the data of the historical year of the target parameter by the correlation determination module 32 includes:
determining parameter set correlations between the plurality of parameter sets and the target parameter;
determining at least one target parameter group according to the parameter group correlation;
and determining the correlation degree between each sub-parameter in the target parameter group and the target parameter according to the data of the historical years of the target parameter group and the data of the historical years of the target parameter.
By judging the relevance of the target parameter group and judging and screening the relevance of each sub-parameter in the target parameter group, the effectiveness of sub-parameter screening is improved, and the accuracy of a prediction result is further improved.
As an optional implementation, the prediction algorithm includes at least two types, and the specific manner for the prediction module 33 to obtain the prediction result of the target parameter in the target year according to the prediction algorithm includes:
determining weights corresponding to the at least two prediction algorithms;
and obtaining a weighted prediction result of the target parameter in the target year according to the at least two prediction algorithms and the corresponding weights.
By the weighted prediction of a plurality of algorithms, the applicability and accuracy of the prediction are improved.
As an alternative embodiment, as shown in fig. 4, the apparatus further includes an ensemble learning module 35, configured to, after the prediction module 33 determines the prediction algorithm according to the second prediction parameter,
determining an integration strategy according to the prediction algorithm;
integrating the prediction algorithm according to the integration strategy to obtain a strong learner based on the prediction algorithm;
and obtaining a prediction result of the target parameter under the target year according to the strong learner.
And screening the sub-parameters through the correlation degree, determining a corresponding prediction algorithm, and integrating the prediction algorithm to obtain a strong learner for load prediction, so that the accuracy, robustness and practicability of a load prediction result are improved.
As an alternative embodiment, as shown in fig. 4, the apparatus further includes an addressing module 36, configured to, after the prediction module 33 obtains the prediction result of the target parameter in the target year according to the prediction algorithm,
determining at least one site-to-be-selected point in the area to be planned according to the load prediction result;
and generating a load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm according to the quasi-addressing points.
The method for generating the Thiessen polygon is used for completing the determination of the planned site selection point of the transformer substation and the corresponding load area thereof according to the load prediction result of the area to be planned, and the practicability of medium and long term load prediction is improved.
As an optional implementation manner, a specific manner in which the addressing module generates the load areas corresponding to the planned addressing points through a thiessen polygon algorithm according to the planned addressing points includes:
determining the load density of the area to be planned according to the load prediction result;
determining the expanding speed in the corresponding direction of the corresponding area in the Thiessen polygon generation process according to the load density, and generating the load area corresponding to each planned addressing point through the Thiessen polygon algorithm;
wherein the generation process of the Thiessen polygon comprises the following steps: and expanding the area at a corresponding expansion speed in the outward direction along each radius by taking each address point to be selected as the center of a circle and each center of a circle as the reference until the area is contacted with the areas corresponding to other address points to be selected.
The expanding speed of the Thiessen polygon is determined through the load density, the power supply range of each transformer substation is further determined, the accuracy and effectiveness of site selection and load area planning of the transformer substations are improved, and the practicability of medium-term and long-term load prediction is further improved.
Example four
Referring to fig. 5, fig. 5 is a schematic structural diagram of a long-term load prediction apparatus in an electrical power system according to a fourth embodiment of the present invention. As shown in fig. 5, the device for predicting long-term load in an electrical power system may include:
a Processor (Processor) 291, the apparatus further including a Memory (Memory) 292 in which executable program code is stored; a Communication Interface 293 and bus 294 may also be included. The processor 291, the memory 292, and the communication interface 293 may communicate with each other via the bus 294. Communication interface 293 may be used for the transmission of information. The processor 291 is coupled to the memory 292, and the processor 291 may call logic instructions (executable program code) in the memory 292 to perform the method for predicting long-term load in the power system according to any of the embodiments.
Furthermore, the logic instructions in the memory 292 may be implemented in the form of software functional units and stored in a computer readable storage medium when sold or used as a stand-alone product.
The memory 292 is a computer-readable storage medium for storing software programs, computer-executable programs, such as program instructions/modules corresponding to the methods in the embodiments of the present application. The processor 291 executes the functional application and data processing by executing the software program, instructions and modules stored in the memory 292, so as to implement the method in the above method embodiments.
The memory 292 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal device, and the like. Further, the memory 292 may include a high speed random access memory and may also include a non-volatile memory.
An embodiment of the present invention further provides a computer-readable storage medium, where a computer execution instruction is stored in the computer-readable storage medium, and when the computer execution instruction is called, the computer execution instruction is used to implement the method described in any embodiment.
Embodiments of the present invention also disclose a computer program product comprising a non-transitory computer readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the method for predicting long-term load in an electric power system described in any of the embodiments.
The above-described embodiments of the apparatus are only illustrative, and the modules described as separate parts may or may not be physically separate, and the parts displayed as modules may or may not be physical modules, may be located in one place, or may be distributed on a plurality of network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above detailed description of the embodiments, those skilled in the art will clearly understand that each embodiment may be implemented by software plus a necessary general hardware platform, and may also be implemented by hardware. With this understanding in mind, the above technical solutions may essentially or in part contribute to the prior art, be embodied in the form of a software product, which may be stored in a computer-readable storage medium, including a Read-Only Memory (ROM), a Random Access Memory (RAM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), a One-time Programmable Read-Only Memory (OTPROM), an electronically Erasable Programmable Read-Only Memory (EEPROM), an optical Disc-Read (CD-ROM) or other storage medium capable of storing data, a magnetic tape, or any other computer-readable medium capable of storing data.
Finally, it should be noted that: the method and the apparatus for predicting medium and long term load in an electrical power system disclosed in the embodiments of the present invention are only preferred embodiments of the present invention, and are only used for illustrating the technical solutions of the present invention, rather than limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art; the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for predicting medium and long term load in an electric power system, the method comprising:
determining a first prediction parameter, a target parameter and a target year;
wherein the first prediction parameters comprise a plurality of parameter sets, and the parameter sets respectively comprise a plurality of sub-parameters; the target parameter comprises at least one of the following parameters: the maximum load of the whole society, the electricity consumption of the whole society and the electricity consumption of the whole industry;
determining the correlation degree between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter;
according to the correlation degree, taking a sub-parameter of which the correlation degree exceeds a preset threshold value as a second prediction parameter;
determining a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year.
2. The method according to claim 1, wherein after the sub-parameter with the correlation degree exceeding a preset threshold is used as the second prediction parameter according to the correlation degree, the method further comprises:
selecting at least two of the sub-parameters of the second prediction parameter as third prediction parameters;
determining a multivariate goodness of fit between the third prediction parameter and the target parameter, and determining a goodness of fit judgment threshold;
determining a multivariate goodness-of-fit residual error according to the adjacent two iteration results;
if the multivariate goodness-of-fit residual tends to converge, judging whether the multivariate goodness-of-fit exceeds the goodness-of-fit judgment threshold; and if so, replacing the second prediction parameter with the third prediction parameter.
3. The method of claim 2, wherein prior to replacing the second prediction parameter with the third prediction parameter, the method further comprises:
determining a demand dimension for the second prediction parameter;
performing dimensionality reduction transformation on the third prediction parameter through a principal component analysis method according to the demand dimensionality to obtain a transformed third prediction parameter; the transformed third prediction parameters comprise main component components of demand dimensions of the third prediction parameters;
the replacing the second prediction parameter with the third prediction parameter includes:
replacing the second prediction parameters with the transformed third prediction parameters.
4. The method of claim 1, wherein determining a correlation between each sub-parameter in the first prediction parameter and the target parameter based on the data for the historical year of the first prediction parameter and the data for the historical year of the target parameter comprises:
determining a parameter set correlation between the plurality of parameter sets and the target parameter;
determining at least one target parameter group according to the parameter group correlation;
and determining the correlation degree between each sub-parameter in the target parameter group and the target parameter according to the data of the historical years of the target parameter group and the data of the historical years of the target parameter.
5. The method of claim 1, wherein the predictive algorithm includes at least two, and wherein obtaining the predicted outcome of the target parameter at the target year according to the predictive algorithm includes:
determining weights corresponding to the at least two prediction algorithms;
and obtaining a weighted prediction result of the target parameter in the target year according to the at least two prediction algorithms and corresponding weights.
6. The method of claim 1, wherein after determining the prediction algorithm based on the second prediction parameter, the method further comprises:
determining an integration strategy according to the prediction algorithm;
integrating the prediction algorithm according to the integration strategy to obtain a strong learner based on the prediction algorithm;
and obtaining a prediction result of the target parameter under the target year according to the strong learner.
7. The method according to any one of claims 1-6, wherein after obtaining the predicted outcome of the target parameter at the target year according to the prediction algorithm, the method further comprises:
determining at least one site-to-be-selected point in the area to be planned according to the load prediction result;
and generating a load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm according to the quasi-addressing points.
8. The method of claim 7, wherein the generating the load regions corresponding to the proposed address points by the Thiessen polygon algorithm according to the proposed address points comprises:
determining the load density of the area to be planned according to the load prediction result;
determining the expansion speed in the corresponding direction of the corresponding area in the Thiessen polygon generation process according to the load density, and generating the load area corresponding to each quasi-addressing point through a Thiessen polygon algorithm;
wherein the generation process of the Thiessen polygon comprises the following steps: and expanding the area at a corresponding expansion speed in the outward direction along each radius by taking each to-be-addressed point as a circle center and taking each circle center as a reference until the area is contacted with the areas corresponding to other to-be-addressed points.
9. An apparatus for predicting medium and long term loads in an electrical power system, the apparatus comprising:
the parameter determination module is used for determining a first prediction parameter, a target parameter and a target year;
wherein the first prediction parameters comprise a plurality of parameter sets, and the parameter sets respectively comprise a plurality of sub-parameters; the target parameter comprises at least one of the following parameters: the maximum load of the whole society, the electricity consumption of the whole society and the electricity consumption of the whole industry;
the relevancy determination module is used for determining the relevancy between each sub-parameter in the first prediction parameter and the target parameter according to the data of the historical years of the first prediction parameter and the data of the historical years of the target parameter;
the parameter determining module is further used for taking the sub-parameters of which the correlation degrees exceed a preset threshold value as second prediction parameters according to the correlation degrees;
the prediction module is used for determining a prediction algorithm according to the second prediction parameter; and according to the prediction algorithm, obtaining the prediction result of the target parameter in the target year.
10. An apparatus for predicting a long-term load in an electric power system, the apparatus comprising:
a memory storing executable program code;
a processor coupled with the memory;
the processor invokes the executable program code stored in the memory to perform the method of long term load prediction in a power system as claimed in any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211415962.4A CN115660214A (en) | 2022-11-11 | 2022-11-11 | Method and device for predicting medium and long-term load of power system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211415962.4A CN115660214A (en) | 2022-11-11 | 2022-11-11 | Method and device for predicting medium and long-term load of power system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115660214A true CN115660214A (en) | 2023-01-31 |
Family
ID=85020393
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211415962.4A Pending CN115660214A (en) | 2022-11-11 | 2022-11-11 | Method and device for predicting medium and long-term load of power system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115660214A (en) |
-
2022
- 2022-11-11 CN CN202211415962.4A patent/CN115660214A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kuznetsova et al. | Reinforcement learning for microgrid energy management | |
| Pinto et al. | Strategic bidding in electricity markets: An agent-based simulator with game theory for scenario analysis | |
| Karabulut et al. | Long term energy consumption forecasting using genetic programming | |
| Kaboli et al. | An expression-driven approach for long-term electric power consumption forecasting | |
| Moallemi et al. | Robust decision making and Epoch–Era analysis: A comparison of two robustness frameworks for decision-making under uncertainty | |
| Lawson et al. | Bayesian framework for power network planning under uncertainty | |
| Unsihuay-Vila et al. | Electricity demand and spot price forecasting using evolutionary computation combined with chaotic nonlinear dynamic model | |
| CN114971094A (en) | Method, device, equipment and medium for predicting power load based on edge calculation | |
| CN117094535B (en) | Artificial intelligence-based energy supply management method and system | |
| Nguyen et al. | Acquisition of new technology information for maintenance and replacement policies | |
| Bierlaire et al. | A heuristic for nonlinear global optimization | |
| CN114692859A (en) | Data processing method and device, computing equipment and test simplification equipment | |
| CN115660214A (en) | Method and device for predicting medium and long-term load of power system | |
| Cortez et al. | Genetic and evolutionary algorithms for time series forecasting | |
| CN116051159A (en) | Prediction method, prediction device and storage medium for accessory demand | |
| CN116109339A (en) | Prediction method, prediction device and storage medium for accessory demand | |
| CN116051160A (en) | Prediction method, prediction device and storage medium for accessory demand | |
| CN112926777B (en) | Day-ahead load prediction method and system considering commercial building energy consumption behaviors | |
| CN111783487B (en) | Fault early warning method and device for card reader equipment | |
| Viana et al. | Load forecasting benchmark for smart meter data | |
| JP7424373B2 (en) | Analytical equipment, analytical methods and analytical programs | |
| Musa | A Hybrid Modern and Classical Algorithm for Indonesian Electricity Demand Forecasting | |
| KR102542488B1 (en) | Method and Apparatus for Predicting Appliance Power Usage State Using Non Intrusive Load Monitoring Based on Long-Short Term Memory | |
| US20230267007A1 (en) | System and method to simulate demand and optimize control parameters for a technology platform | |
| Liu | A study of short-term forecasting of raw material prices using a combined LSTM-ARIMA model |
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 |