Power grid distribution management system and method based on terminal consumption trend prediction

Abstract

The present invention discloses a power grid distribution management system and method based on terminal consumption trend prediction, which relates to the technical field related to the distribution system. The method comprises: obtaining a first power consumption record data set, performing power consumption backtracking, constructing a first consumption prediction model, performing consumption prediction on a first terminal power consumption area in a time zone to be managed, and generating a first theoretical consumption prediction result. Obtaining the heat index of the associated media, identifying the sudden increase in power consumption variation, and obtaining the sudden increase coefficient of variation. Obtaining the terminal distribution network, identifying the line loss rate, and performing feedback optimization on the first theoretical consumption prediction result with the first line loss rate and the sudden increase coefficient of variation to generate a first consumption prediction result. Using the first consumption prediction result, a distribution plan for the first terminal power consumption area is generated, and distribution management is performed. The technical problem that the power grid distribution management method in the prior art has a low prediction accuracy for the non-periodic impact on power demand, resulting in difficulty in improving the stability of power grid operation is solved.

Description

Technical Field

[0001] The present application relates to the technical field related to power distribution systems, and specifically to a power grid distribution management system and method based on terminal consumption trend prediction. Background Art

[0002] With the development of economy and the acceleration of urbanization, the load of power grid is showing a trend of rapid growth. Traditional power grid management methods mainly rely on historical power consumption data for load forecasting and distribution management. However, the load forecasting and distribution management based on historical power consumption data often ignore the non-periodic impact of external factors such as emergencies on power demand, resulting in insufficient prediction of power demand changes caused by emergencies, thus affecting the stability of the power grid.

[0003] Therefore, in the prior art, the power distribution management method of the power grid has a low accuracy in predicting the non-periodic impact on electricity demand, which makes it difficult to improve the stability of power grid operation. Summary of the invention

[0004] This application solves the technical problem that the power distribution management method in the prior art has a low prediction accuracy rate for the non-periodic impact on electricity demand, which makes it difficult to improve the stability of power grid operation by providing a power distribution management system and method based on terminal consumption trend prediction. It achieves more accurate prediction of power grid load changes and timely response to the impact of non-periodic external factors on electricity demand, thereby realizing intelligent and efficient power distribution management and ensuring the safe and stable operation of the power grid.

[0005] The first aspect of the present application provides a power grid distribution management system based on terminal consumption trend prediction, including: a record data set acquisition module, which is used to collect historical power consumption records of a first terminal power consumption area of ​​a first distribution network in a preset time zone to generate a first power consumption record data set. A prediction model acquisition module, which is used to trace the power consumption of the first power consumption record data set, obtain a first monthly month-on-month growth rate sequence of power consumption and multiple first monthly year-on-year growth rate sequences of power consumption, and train a first consumption prediction model. A prediction result acquisition module, which is used to perform consumption prediction on the first terminal power consumption area in the time zone to be managed through the first consumption prediction model, and generate a first theoretical consumption prediction result. A variation sudden increase coefficient acquisition module, which is used to set a media information monitoring window based on the time zone to be managed, identify associated media information for the first terminal power consumption area, obtain an associated media heat index, identify power consumption variation sudden increase, and obtain a variation sudden increase coefficient. A feedback optimization module is used to obtain the terminal power distribution network of the first terminal power consumption area, identify the line loss rate of the terminal power distribution network, and perform feedback optimization on the first theoretical consumption prediction result with the first line loss rate and the variation sudden increase coefficient to generate a first consumption prediction result. A power distribution management module is used to generate a power distribution plan for the first terminal power consumption area with the first consumption prediction result and perform power distribution management.

[0006] The second aspect of the present application provides a power distribution management method based on terminal consumption trend prediction, the method comprising: collecting historical power consumption records of a first terminal power consumption area of ​​a first distribution network in a preset time zone to generate a first power consumption record data set. The first power consumption record data set is backtracked to obtain a first power consumption month-on-month growth rate sequence and multiple first power consumption month-on-month growth rate sequences, and a first consumption prediction model is trained. The first consumption prediction model is used to perform consumption prediction on the first terminal power consumption area in the time zone to be managed, and a first theoretical consumption prediction result is generated. A media information monitoring window is set based on the time zone to be managed, and the associated media information of the first terminal power consumption area is identified, and the associated media heat index is obtained, and the power consumption variation sudden increase is identified to obtain the variation sudden increase coefficient. The terminal distribution network of the first terminal power consumption area is obtained, and the line loss rate of the terminal distribution network is identified. The first theoretical consumption prediction result is feedback optimized with the first line loss rate and the variation sudden increase coefficient to generate a first consumption prediction result. The distribution plan of the first terminal power consumption area is generated with the first consumption prediction result, and distribution management is performed.

[0007] The power grid distribution management system and method based on terminal consumption trend prediction proposed in this application is proposed to collect historical power consumption records of the first terminal power consumption area of ​​the first distribution network in a preset time zone to generate a first power consumption record data set. The first power consumption record data set is backtracked to obtain the first power consumption monthly month-on-month growth rate sequence and multiple first power consumption monthly year-on-year growth rate sequences, and the first consumption prediction model is trained. The first consumption prediction model is used to perform consumption prediction on the first terminal power consumption area in the time zone to be managed, and a first theoretical consumption prediction result is generated. A media information monitoring window is set based on the time zone to be managed, and the associated media information of the first terminal power consumption area is identified, the associated media heat index is obtained, and the power consumption variation sudden increase is identified to obtain the variation sudden increase coefficient. The terminal distribution network of the first terminal power consumption area is obtained, the line loss rate of the terminal distribution network is identified, and the first theoretical consumption prediction result is feedback optimized with the first line loss rate and the variation sudden increase coefficient to generate the first consumption prediction result. The distribution plan of the first terminal power consumption area is generated with the first consumption prediction result, and distribution management is performed. The invention solves the technical problem that the power distribution management method in the prior art has a low prediction accuracy rate for the non-periodic impact on power demand, which makes it difficult to improve the stability of power grid operation. It achieves a more accurate prediction of power grid load changes and timely response to the impact of non-periodic external factors on power demand, thereby realizing intelligent and efficient power distribution management and ensuring the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0009] Figure 1 A schematic diagram of the structure of a power grid distribution management system based on terminal consumption trend prediction provided in an embodiment of the present application.

[0010] Figure 2 A schematic flow chart of a power grid distribution management method based on terminal consumption trend prediction provided in an embodiment of the present application.

[0011] Explanation of the reference numerals: record data set acquisition module 11, prediction model acquisition module 12, prediction result acquisition module 13, variation sudden increase coefficient acquisition module 14, feedback optimization module 15, power distribution management module 16. DETAILED DESCRIPTION

[0012] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0013] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0014] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0015] The present application embodiment provides a power grid distribution management system and method based on terminal consumption trend prediction, such as Figure 2 As shown, the method includes:

[0016] Historical electricity consumption records are collected for a first terminal electricity consumption area of ​​a first distribution network in a preset time zone to generate a first electricity consumption record data set.

[0017] The first electricity consumption record data set is backtracked to obtain a first electricity consumption month-on-month growth rate sequence and multiple first electricity consumption month-on-month growth rate sequences, and a first consumption prediction model is trained.

[0018] The first consumption prediction model is used to perform consumption prediction on the first terminal power consumption area in the time zone to be managed, and a first theoretical consumption prediction result is generated.

[0019] Collect historical electricity consumption record data of the first terminal electricity consumption area in a preset time zone to generate a first electricity consumption record data set. The first terminal electricity consumption area is the corresponding area for consumption trend prediction, and the preset time zone is a preset time interval corresponding to the time interval to be managed, and the time interval is a preset interval, such as electricity consumption record data for 5 consecutive years in historical data. The first electricity consumption record data set includes real-time electricity consumption data and total electricity consumption data in different time intervals, as well as statistical data such as the acquisition time of the corresponding data. Subsequently, the first electricity consumption record data set is backtracked to obtain the first electricity consumption monthly month-on-month growth rate sequence and multiple first electricity consumption monthly year-on-year growth rate sequences, and train the first consumption prediction model. Among them, the first electricity consumption monthly month-on-month growth rate sequence is a sequence composed of the growth rates of electricity consumption for two consecutive months, such as the growth rate of January compared to December, the growth rate of February compared to January, the growth rate of March compared to February, and so on. The growth rate sequence finally formed is obtained to obtain the first electricity consumption monthly month-on-month growth rate sequence. The first monthly year-on-year growth rate sequence of electricity consumption is the growth rate at the same time in history, that is, the growth rate of the same month in adjacent years, such as the growth rate in January of each year in 10 years. Each month in each year corresponds to a growth rate, and the growth rates corresponding to 12 months in each year in 10 years constitute 12 sequences. Further, the first consumption prediction model is used to perform consumption prediction on the first terminal power consumption area in the time zone to be managed to generate a first theoretical consumption prediction result, which is the comprehensive change rate of electricity consumption in the first terminal power consumption area in the time interval to be managed predicted by the first consumption prediction model.

[0020] The method provided in the embodiment of the present application also includes:

[0021] The first prediction sub-model is constructed by training the first electricity consumption monthly month-on-month growth rate sequence.

[0022] The multiple first electricity consumption month-on-month growth rate sequences are used to train multiple second prediction sub-models.

[0023] The first prediction sub-model and the multiple second prediction sub-models are tested, and model integration and fusion are performed according to the test results to generate the first consumption prediction model.

[0024] The first electricity consumption record data set is backtracked to obtain the first electricity consumption month-on-month growth rate sequence and multiple first electricity consumption month-on-month growth rate sequences, and the first consumption forecasting model is trained, including: using the first electricity consumption month-on-month growth rate sequence to train and construct a first forecasting sub-model, wherein the first forecasting sub-model is constructed by a time series related model, such as a recurrent neural network model, a long short-term memory network model, etc. The multiple first electricity consumption month-on-month growth rate sequences are used to train multiple second forecasting sub-models, wherein the second forecasting sub-model is constructed by a time series related model. This embodiment uses a recurrent neural network model as a basis to construct a first prediction sub-model and a second prediction sub-model. The construction process is as follows: obtain a first electricity consumption monthly month-on-month growth rate sequence, and divide the sequence. For example, the growth rate in January is used as input data, and the growth rate in February is used as output data to constitute a set of training data. The growth rate in January and the growth rate in February are used as input data, and the growth rate in March is used as output data to constitute a set of training data. And so on, multiple sets of training data consisting of the first electricity consumption monthly month-on-month growth rate sequence are obtained, and supervised training is performed on the recurrent neural network model until the model meets the preset accuracy rate, and the training of the model is completed to obtain the first prediction sub-model. The second prediction sub-model is constructed in the same manner as the first prediction sub-model, wherein the training data of the second prediction sub-model is constructed using multiple first monthly year-on-year growth rate sequences of electricity consumption. When dividing the training data, the year-on-year growth rate in January last year is used as input data, and the year-on-year growth rate in January this year is used as output data to form a set of training data. The training data is acquired and the model is trained in this way until the model meets the preset accuracy rate. A second prediction sub-model is obtained, and subsequent training is performed according to the monthly data to obtain multiple second prediction sub-models, each of which corresponds to one month. The first prediction sub-model and the multiple second prediction sub-models are tested, and the model is integrated and fused according to the test results to generate the first consumption prediction model.

[0025] The method provided in the embodiment of the present application also includes:

[0026] The first prediction sub-model and the multiple second prediction sub-models are tested to obtain multiple groups of prediction results, and multiple groups of prediction deviations are calculated.

[0027] Determine whether the plurality of groups of prediction deviations are within a preset deviation range.

[0028] If so, the first prediction sub-model and the multiple second prediction sub-models are tested for accuracy identification based on the multiple groups of prediction results, and are integrated according to the test accuracy to generate the first consumption prediction model.

[0029] If not, a reconstruction instruction is generated to retrain the first prediction sub-model and the plurality of second prediction sub-models until the plurality of prediction deviations obtained by the test are within a preset deviation range.

[0030] The first prediction sub-model and the multiple second prediction sub-models are tested, and model integration and fusion are performed according to the test results, including: testing the first prediction sub-model and the multiple second prediction sub-models based on training data, obtaining multiple groups of prediction results, and calculating multiple groups of prediction deviations, wherein the prediction deviation is 1 minus the ratio of the test output result to the output result in the training data. It is judged whether the multiple groups of prediction deviations are within a preset deviation range, wherein the preset deviation range is a pre-set data deviation range. When the deviation range is less than the deviation range, the output result has a high accuracy, otherwise the output result has a low accuracy. If it is within the preset deviation range, the test accuracy of the first prediction sub-model and the multiple second prediction sub-models is identified based on the multiple groups of prediction results, the average test accuracy of the multiple second prediction sub-models, and the test accuracy of the first prediction sub-model are obtained, and the weight of the first prediction sub-model and the multiple second prediction sub-models is assigned based on the accuracy parameter, wherein the weight of the first prediction sub-model is the ratio of the test accuracy of the first prediction sub-model to the sum of the average test accuracy and the test accuracy of the first prediction sub-model. The weights of the multiple second prediction sub-models are 1 minus the weight of the first prediction sub-model, and the first prediction sub-model and the second prediction sub-model are fused based on the model weights to generate the first consumption prediction model. If it is not within the preset deviation range, a reconstruction instruction is generated to retrain the first prediction sub-model and the multiple second prediction sub-models until the multiple groups of prediction deviations obtained by the test are within the preset deviation range.

[0031] A media information monitoring window is set based on the time zone to be managed, associated media information is identified for the power consumption area of ​​the first terminal, an associated media heat index is obtained, a sudden increase in power consumption variation is identified, and a sudden increase coefficient of variation is obtained.

[0032] A terminal power distribution network of the first terminal power consumption area is obtained, a line loss rate of the terminal power distribution network is identified, and the first theoretical absorption prediction result is feedback-optimized using the first line loss rate and the variation sudden increase coefficient to generate a first absorption prediction result.

[0033] A power distribution plan for the first terminal power consumption area is generated based on the first consumption prediction result to perform power distribution management.

[0034] A media information monitoring window is set based on the time zone to be managed, wherein the time zone to be managed is the time interval for power grid distribution management, and the media information monitoring window is used to obtain related media information, and the related media information is various types of media information related to regional tourism extracted based on the media information monitoring window. The related media information is identified for the power consumption area of ​​the first terminal, and the related media heat index is obtained. Under normal circumstances, the month-on-year growth rate and the month-on-month growth rate can be accurately predicted. However, due to the development of social media, for example, a city is too popular on short video platforms, there may be non-periodic surges in traffic, which in turn leads to an increase in electricity demand. By obtaining the related media heat index and identifying the sudden increase in electricity consumption variation, the variation sudden increase coefficient is obtained, and the variation sudden increase coefficient is the predicted change parameter of the growth rate. By obtaining the variation sudden increase coefficient, the electricity consumption of the related media heat index is quantified, and the accuracy of power grid distribution management is further improved.

[0035] The terminal distribution network of the first terminal power consumption area is obtained, and the line loss rate of the terminal distribution network is identified. The line loss rate refers to the ratio of the power loss caused by resistance and other factors to the original power during the power transmission process. The line loss rate can be identified and obtained through the historical average line loss rate of the terminal distribution network. Further, the first theoretical absorption prediction result is feedback optimized with the first line loss rate and the variation sudden increase coefficient. When performing feedback optimization, the first theoretical absorption prediction result is added and divided by the variation sudden increase coefficient by the first line loss rate to obtain the absorption change result, and the product of the first prediction sub-model weight and the data of the adjacent time zone to be managed is combined with the product of the weights of multiple second prediction sub-models and the data of the time zone managed at the same time. The sum is multiplied by the absorption change result to generate the first absorption prediction result. The first absorption prediction result is the rate of change of the power of the first terminal power consumption area in the time zone to be managed. Finally, the first consumption prediction result is used to generate the power distribution plan for the first terminal power consumption area, that is, the specific power distribution data. The power distribution plan calculates 1 plus the first consumption prediction result, and obtains specific power distribution parameters according to the product of the calculation result and the power distribution plan parameter of the previous node to perform power distribution management. This solves the technical problem that the power distribution management method of the prior art has a low prediction accuracy for the non-periodic impact on power demand, which makes it difficult to improve the stability of power grid operation. It achieves a more accurate prediction of power grid load changes and a timely response to the impact of non-periodic external factors on power demand, thereby realizing intelligent and efficient power grid distribution management and ensuring the safe and stable operation of the power grid.

[0036] The method provided in the embodiment of the present application also includes:

[0037] Construct keyword features of the power consumption area of ​​the first terminal.

[0038] In the media information monitoring window, the keyword feature is used as a filtering element to filter the associated media information and obtain an information filtering result.

[0039] The information filtering result is used to identify user interaction behaviors and obtain an interaction behavior statistical matrix.

[0040] Perform heat analysis using the interactive behavior statistical matrix to obtain the associated media heat index.

[0041] The sudden increase in electricity consumption variation is identified based on the associated media heat index to generate the sudden increase coefficient.

[0042] The media information monitoring window is set based on the time zone to be managed, the associated media information of the first terminal power consumption area is identified, the associated media heat index is obtained, the power consumption variation and sudden increase are identified, and the variation and sudden increase coefficient is obtained, including: constructing the keyword features of the first terminal power consumption area, the keyword features include keywords such as place name keywords and local building keywords that can represent the power consumption area. Based on the associated media information collected by the media information monitoring window, the keyword features are used as filtering elements to filter the associated media information to obtain the information filtering results, and only the media information containing the power consumption area keywords is included in the information filtering results. Subsequently, the user interaction behavior is identified based on the information filtering results, wherein the user interaction behavior includes the amount of browsing data, the amount of likes, the amount of comments, the amount of forwarding, etc., and the interaction behavior statistical matrix is ​​obtained, and the interaction behavior statistical matrix contains specific statistical data of the user interaction behavior. The interaction behavior statistical matrix is ​​used to perform heat analysis to obtain the associated media heat index. Finally, the power consumption variation and sudden increase are identified based on the associated media heat index to generate the variation and sudden increase coefficient.

[0043] The method provided in the embodiment of the present application also includes:

[0044] Build preset interaction-heat conversion logic.

[0045] A thermal analyzer is constructed using the preset interaction-thermal conversion logic.

[0046] The interaction behavior statistical matrix is ​​analyzed by the heat analyzer to output the associated media heat index.

[0047] Performing heat analysis with the interactive behavior statistical matrix to obtain the associated media heat index includes: constructing a preset interactive-heat conversion logic, in which each interactive behavior data and a corresponding heat conversion parameter are included, such as 10 heat conversion parameters for every 1,000 forwardings. Constructing a heat analyzer with the preset interactive-heat conversion logic, the heat analyzer is used to analyze the interactive behavior statistical matrix based on the preset interactive-heat conversion logic, obtain the heat index parameters of each interactive behavior in the interactive behavior statistical matrix, and sum up all the interactive behavior heat index parameters to output the associated media heat index.

[0048] The method provided in the embodiment of the present application also includes:

[0049] A plurality of groups of associated media heat index samples are constructed, and a plurality of groups of electricity consumption sudden increase samples are collected with the plurality of groups of associated media heat index samples as constraints.

[0050] Based on the multiple groups of electricity consumption sudden increase samples and the multiple groups of associated media heat index samples, heat-sudden increase impact relationship identification is performed.

[0051] The associated media heat index is input into the heat-sudden increase influence relationship to generate the variation sudden increase coefficient.

[0052] Construct multiple groups of associated media heat index samples, and collect multiple groups of electricity consumption sudden increase samples corresponding to multiple groups of associated media heat index samples in the historical electricity consumption record data with the multiple groups of associated media heat index samples as constraints. The electricity consumption sudden increase samples are the sudden increase electricity consumption growth rate data under the influence of the associated media heat index samples. For example, the original forecast growth rate for this month is 3%. When the sudden increase electricity consumption growth rate data obtained based on the associated media heat index, that is, the electricity consumption sudden increase is 1%, the forecast growth rate for this month needs to be adjusted, and the adjusted forecast growth rate for this month is 4%. Subsequently, the heat-sudden increase influence relationship is identified based on the multiple groups of electricity consumption sudden increase samples and the multiple groups of associated media heat index samples, that is, the functional relationship between the multiple groups of electricity consumption sudden increase samples and the multiple groups of associated media heat index samples is fitted to obtain the heat-sudden increase influence relationship. Finally, the associated media heat index is input into the heat-sudden increase influence relationship to generate the variation sudden increase coefficient, which is the predicted change parameter of the growth rate. By obtaining the variation sudden increase coefficient, the electricity consumption of the related media heat index is quantified, further improving the accuracy of power grid distribution management.

[0053] The method provided in the embodiment of the present application also includes:

[0054] A load adjustment span is calculated for the power distribution plan of the first terminal power consumption area and the power distribution plan of the adjacent previous node to obtain a first load span.

[0055] Determine whether the first load span is greater than a preset adjustment step, and if so, generate a segmented adjustment instruction.

[0056] According to the segmented adjustment instruction, segmented power distribution management is performed on the first load span according to the preset adjustment step.

[0057] The load adjustment span of the power distribution scheme of the first terminal power consumption area and the power distribution scheme of the adjacent previous node is calculated. The load adjustment span is the difference between the current power distribution scheme and the power distribution scheme of the adjacent previous node, and the first load span is obtained. It is judged whether the first load span is greater than the preset adjustment step. The preset adjustment step is the preset step adjustment range. For example, when the preset adjustment step is set to n megawatts, when the first load span is greater than the n megawatts, the corresponding load span is higher and has exceeded the maximum span range of the power grid load adjustment. When the load is adjusted according to this plan, the instantaneous electrical parameters change greatly, and the distribution network is prone to abnormal damage. Otherwise, the load span is within the maximum span range of the power grid load adjustment, and the corresponding plan can be directly adopted for power distribution management. If the first load span is greater than the preset adjustment step, a segmented adjustment instruction is generated. The segmented adjustment instruction is used to perform segmented adjustments to the power distribution plan of the first terminal power consumption area based on the power distribution plan of the adjacent previous node, such as dividing it into multiple adjustment stages, and the adjustment amplitude of the power distribution plan in each adjustment stage is the same, thereby reducing the instantaneous pressure of the distribution adjustment on the power grid load, and thus improving the operating efficiency and stability of the power grid.

[0058] In the above, refer to Figure 2 The present invention describes in detail a method for managing power distribution in a power grid based on terminal consumption trend prediction according to an embodiment of the present invention. Figure 1 A power grid distribution management system based on terminal consumption trend prediction according to an embodiment of the present invention is described.

[0059] According to the power grid distribution management system based on terminal consumption trend prediction according to the embodiment of the present invention, the technical problem that the power grid distribution management method in the prior art has a low prediction accuracy of the non-periodic impact on electricity demand, resulting in difficulty in improving the stability of power grid operation is solved. It achieves more accurate prediction of power grid load changes and timely response to the impact of non-periodic external factors on electricity demand, thereby realizing intelligent and efficient power grid distribution management and ensuring the safe and stable operation of the power grid. The power grid distribution management system based on terminal consumption trend prediction includes: a record data set acquisition module 11, a prediction model acquisition module 12, a prediction result acquisition module 13, a variation sudden increase coefficient acquisition module 14, a feedback optimization module 15, and a distribution management module 16.

[0060] The record data set acquisition module 11 is used to collect historical power consumption records of a first terminal power consumption area of ​​a first distribution network in a preset time zone to generate a first power consumption record data set.

[0061] The prediction model acquisition module 12 is used to trace the electricity consumption of the first electricity consumption record data set, obtain the first electricity consumption monthly chain growth rate sequence and multiple first electricity consumption monthly year-on-year growth rate sequences, and train the first consumption prediction model.

[0062] The prediction result acquisition module 13 is used to perform a consumption prediction on the first terminal power consumption area in the time zone to be managed by using the first consumption prediction model to generate a first theoretical consumption prediction result.

[0063] The variation sudden increase coefficient acquisition module 14 is used to set a media information monitoring window based on the time zone to be managed, identify the associated media information of the first terminal power consumption area, obtain the associated media heat index, identify the power consumption variation sudden increase, and obtain the variation sudden increase coefficient.

[0064] The feedback optimization module 15 is used to obtain the terminal distribution network of the first terminal power consumption area, identify the line loss rate of the terminal distribution network, and feedback optimize the first theoretical absorption prediction result with the first line loss rate and the variation sudden increase coefficient to generate a first absorption prediction result.

[0065] The power distribution management module 16 is used to generate a power distribution plan for the first terminal power consumption area based on the first consumption prediction result and perform power distribution management.

[0066] The specific configuration of the prediction model acquisition module 12 will be described in detail below. The prediction model acquisition module 12 may further include: performing a backtracking of the first electricity consumption record data set, obtaining a first electricity consumption month-on-month growth rate sequence and multiple first electricity consumption month-on-month growth rate sequences, and training a first consumption prediction model, including: using the first electricity consumption month-on-month growth rate sequence to train and construct a first prediction sub-model. Using the multiple first electricity consumption month-on-month growth rate sequences to train multiple second prediction sub-models. Testing the first prediction sub-model and the multiple second prediction sub-models, integrating and fusing the models according to the test results, and generating the first consumption prediction model.

[0067] Below, the specific configuration of the prediction model acquisition module 12 will continue to be described in detail. The prediction model acquisition module 12 further includes: testing the first prediction sub-model and the multiple second prediction sub-models, and performing model integration fusion according to the test results, including: testing the first prediction sub-model and the multiple second prediction sub-models, obtaining multiple groups of prediction results, and calculating and obtaining multiple groups of prediction deviations. Determine whether the multiple groups of prediction deviations are within the preset deviation range. If so, the first prediction sub-model and the multiple second prediction sub-models are tested for accuracy based on the multiple groups of prediction results, and are fused according to the test accuracy to generate the first consumption prediction model. If not, a reconstruction instruction is generated to retrain the first prediction sub-model and the multiple second prediction sub-models until the multiple groups of prediction deviations obtained by the test are within the preset deviation range.

[0068] Below, the specific configuration of the variation sudden increase coefficient acquisition module 14 will be described in detail. The variation sudden increase coefficient acquisition module 14 may further include: setting a media information monitoring window based on the time zone to be managed, identifying the associated media information of the first terminal power consumption area, obtaining the associated media heat index, identifying the variation sudden increase in power consumption, and obtaining the variation sudden increase coefficient, including: constructing keyword features of the first terminal power consumption area. In the media information monitoring window, the keyword features are used as filtering elements to filter the associated media information and obtain information filtering results. User interaction behavior is identified based on the information filtering results to obtain an interaction behavior statistical matrix. Heat analysis is performed based on the interaction behavior statistical matrix to obtain the associated media heat index. Based on the associated media heat index, the variation sudden increase in power consumption is identified to generate the variation sudden increase coefficient.

[0069] The specific configuration of the variation sudden increase coefficient acquisition module 14 will be described in detail below. The variation sudden increase coefficient acquisition module 14 further includes: performing heat analysis with the interactive behavior statistical matrix to obtain the associated media heat index, including: constructing a preset interaction-heat conversion logic. Constructing a heat analyzer with the preset interaction-heat conversion logic. Analyzing the interactive behavior statistical matrix with the heat analyzer and outputting the associated media heat index.

[0070] The specific configuration of the variation sudden increase coefficient acquisition module 14 will be described in detail below. The variation sudden increase coefficient acquisition module 14 further includes: identifying the variation sudden increase of electricity consumption based on the associated media heat index, and generating the variation sudden increase coefficient, including: constructing multiple groups of associated media heat index samples, and collecting multiple groups of electricity consumption sudden increase samples with the multiple groups of associated media heat index samples as constraints. Identifying the heat-sudden increase impact relationship based on the multiple groups of electricity consumption sudden increase samples and the multiple groups of associated media heat index samples. Inputting the associated media heat index into the heat-sudden increase impact relationship to generate the variation sudden increase coefficient.

[0071] The specific configuration of the power distribution management module 16 will be described in detail below. The power distribution management module 16 further includes: performing load adjustment span calculation on the power distribution plan of the first terminal power consumption area and the power distribution plan of the adjacent previous node to obtain a first load span. Determine whether the first load span is greater than a preset adjustment step, and if so, generate a segmented adjustment instruction. According to the segmented adjustment instruction, perform segmented power distribution management on the first load span according to the preset adjustment step.

[0072] The power grid distribution management system based on terminal consumption trend prediction provided by the embodiment of the present invention can execute the power grid distribution management method based on terminal consumption trend prediction provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0073] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or the server, and the various units and modules included are only divided according to functional logic, but are not limited to the above divisions, as long as the corresponding functions can be achieved. In addition, the specific names of the functional units are only for the convenience of distinguishing from each other, and are not used to limit the scope of protection of the present invention.

[0074] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A power distribution management system based on terminal consumption trend prediction, characterized in that: The system comprises: A record data set acquisition module, used to collect historical power consumption records of a first terminal power consumption area of ​​a first distribution network in a preset time zone to generate a first power consumption record data set; A prediction model acquisition module, configured to perform a backtracking of the first electricity consumption record data set to obtain a first electricity consumption month-on-month growth rate sequence and a plurality of first electricity consumption month-on-month growth rate sequences, and train a first consumption prediction model; A prediction result acquisition module, used to perform a consumption prediction on the first terminal power consumption area in the to-be-managed time zone by using the first consumption prediction model, and generate a first theoretical consumption prediction result; A variation sudden increase coefficient acquisition module is used to set a media information monitoring window based on the time zone to be managed, identify the associated media information of the first terminal power consumption area, obtain the associated media heat index, identify the power consumption variation sudden increase, and obtain the variation sudden increase coefficient; A feedback optimization module is used to obtain a terminal power distribution network in the first terminal power consumption area, identify a line loss rate of the terminal power distribution network, and perform feedback optimization on the first theoretical consumption prediction result based on the first line loss rate and the variation sudden increase coefficient to generate a first consumption prediction result; A power distribution management module, configured to generate a power distribution plan for the first terminal power consumption area based on the first consumption forecast result, and perform power distribution management; The variation sudden increase coefficient acquisition module is also used for: Constructing keyword features of the power consumption area of ​​the first terminal, wherein the keyword features are keywords such as place name keywords and local building keywords that can represent the power consumption area; In the media information monitoring window, the keyword feature is used as a filtering element to filter the associated media information to obtain an information filtering result, wherein the associated media information is various types of media information related to regional tourism extracted based on the media information monitoring window; Using the information filtering results to identify user interaction behaviors, and obtain an interaction behavior statistical matrix; Performing heat analysis based on the interactive behavior statistical matrix to obtain the associated media heat index; The sudden increase in electricity consumption variation is identified based on the associated media heat index to generate the sudden increase coefficient.

2. The system according to claim 1, characterized in that The prediction model acquisition module is also used for: The first prediction sub-model is constructed by training the first electricity consumption month-on-month growth rate sequence; Using the multiple first electricity consumption month-on-month growth rate sequences to train multiple second prediction sub-models; The first prediction sub-model and the multiple second prediction sub-models are tested, and model integration and fusion are performed according to the test results to generate the first consumption prediction model.

3. The system according to claim 2, characterized in that The prediction model acquisition module is also used for: Testing the first prediction sub-model and the plurality of second prediction sub-models to obtain a plurality of groups of prediction results, and calculating and obtaining a plurality of groups of prediction deviations; Determining whether the plurality of groups of prediction deviations are within a preset deviation range; If yes, based on the multiple groups of prediction results, test accuracy identification is performed on the first prediction sub-model and the multiple second prediction sub-models, and they are integrated according to the test accuracy to generate the first consumption prediction model; If not, a reconstruction instruction is generated to retrain the first prediction sub-model and the plurality of second prediction sub-models until the plurality of prediction deviations obtained by the test are within a preset deviation range.

4. The system according to claim 1, characterized in that The variation sudden increase coefficient acquisition module is also used for: Build preset interaction-heat conversion logic; Constructing a heat analyzer with the preset interaction-heat conversion logic; The interaction behavior statistical matrix is ​​analyzed by the heat analyzer to output the associated media heat index.

5. The system according to claim 1, wherein: The variation sudden increase coefficient acquisition module is also used for: Constructing multiple groups of related media heat index samples, and collecting multiple groups of electricity consumption sudden increase samples based on the multiple groups of related media heat index samples as constraints; Identifying a heat-sudden increase impact relationship based on the multiple groups of electricity consumption sudden increase samples and the multiple groups of associated media heat index samples; The associated media heat index is input into the heat-sudden increase influence relationship to generate the variation sudden increase coefficient.

6. The system according to claim 1, wherein: The power distribution management module is also used for: Perform load adjustment span calculation on the power distribution plan of the first terminal power consumption area and the power distribution plan of the adjacent previous node to obtain a first load span; Determine whether the first load span is greater than a preset adjustment step, and if so, generate a segmented adjustment instruction; According to the segmented adjustment instruction, segmented power distribution management is performed on the first load span according to the preset adjustment step.

7. A method for managing power distribution in a power grid using the power distribution management system based on terminal consumption trend prediction according to claim 1, characterized in that: The method comprises: Collect historical electricity consumption records of a first terminal electricity consumption area of ​​a first distribution network in a preset time zone to generate a first electricity consumption record data set; Performing a backtracking of the electricity consumption of the first electricity consumption record data set, obtaining a first electricity consumption month-on-month growth rate sequence and multiple first electricity consumption month-on-month growth rate sequences, and training a first consumption prediction model; Performing a consumption prediction on the first terminal power consumption area in the time zone to be managed by using the first consumption prediction model to generate a first theoretical consumption prediction result; A media information monitoring window is set based on the time zone to be managed, associated media information is identified for the power consumption area of ​​the first terminal, an associated media heat index is obtained, a sudden increase in power consumption variation is identified, and a sudden increase coefficient of variation is obtained; Acquire a terminal power distribution network in the first terminal power consumption area, identify a line loss rate of the terminal power distribution network, perform feedback optimization on the first theoretical consumption prediction result using the first line loss rate and the variation sudden increase coefficient, and generate a first consumption prediction result; Generating a power distribution plan for the first terminal power consumption area based on the first consumption forecast result, and performing power distribution management; The method further comprises: Constructing keyword features of the power consumption area of ​​the first terminal, wherein the keyword features are keywords such as place name keywords and local building keywords that can represent the power consumption area; In the media information monitoring window, the keyword feature is used as a filtering element to filter the associated media information to obtain an information filtering result, wherein the associated media information is various types of media information related to regional tourism extracted based on the media information monitoring window; Using the information filtering results to identify user interaction behaviors, and obtain an interaction behavior statistical matrix; Performing heat analysis based on the interactive behavior statistical matrix to obtain the associated media heat index; The sudden increase in electricity consumption variation is identified based on the associated media heat index to generate the sudden increase coefficient.

Images