Regional power load prediction method and device, and storage medium

Abstract

The invention discloses a regional power load prediction method and device and a storage medium, and the method comprises the steps: querying historical meteorological parameters and historical power load data of a plurality of historical time periods for a specified region, and querying future meteorological parameters of a future time period; preprocessing the power load data; if the preprocessing is completed, searching the most similar historical time period as the optimal time period for the future time period according to the historical meteorological parameters and the future meteorological parameters, and searching the adjacent and similar historical time period as the adjacent time period; and predicting future power load data of a future time period according to the historical power load data of the optimal time period and the adjacent time period. According to the embodiment, the appropriate historical time period is searched according to the similarity of the meteorological conditions and serves as a high-value reference object, and the influence of information forgetting is reduced, so that the accuracy of power load prediction is improved, and the operation and planning quality of a power grid is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grids, and in particular to a method, device and storage medium for predicting regional power load. Background Art

[0002] Grid planning and design is one of the tasks in the early decision-making stage of grid development, and load forecasting is one of the foundations of grid planning and design. Through accurate load forecasting, the plans of power generation equipment such as thermal power generation, photovoltaic power generation, and hydropower generation can be determined, and the start and stop of units can be arranged economically and reasonably, thereby reducing standby capacity and reducing power generation costs.

[0003] At present, one of the methods for power load forecasting is mainly based on the time series model. This method uses LSTM (Long Short Term Memory) to mine the laws of historical power load to predict future power load.

[0004] However, the LSTM model has the problem of information forgetting, and its accuracy in predicting future power loads based on the patterns of historical power loads is low, thus affecting the operation and planning of the power grid. Summary of the invention

[0005] In view of this, the present invention provides a regional power load forecasting method, device and storage medium to improve the accuracy of forecasting future power load.

[0006] A first aspect of the present invention provides a method for regional power load forecasting, comprising:

[0007] Query the historical meteorological parameters and historical power load data of multiple historical time periods for the specified area, as well as the future meteorological parameters of the future time period;

[0008] Preprocessing the historical power load data;

[0009] If the preprocessing is completed, then according to the historical meteorological parameters and the future meteorological parameters, the most similar historical time period is searched as the optimal time period for the future time period, and the adjacent and similar historical time periods are searched as the adjacent time periods;

[0010] The future power load data of the future time period is predicted based on the historical power load data of the optimal time period and the adjacent time periods.

[0011] A second aspect of the present invention provides a regional power load forecasting device, comprising:

[0012] The regional data acquisition and query module is used to query the historical meteorological parameters and historical power load data of multiple historical time periods for the specified area, as well as the future meteorological parameters of the future time period;

[0013] A preprocessing module, used for preprocessing the historical power load data;

[0014] A time period search module, for searching the most similar historical time period as the best time period and searching the adjacent and similar historical time periods as the neighboring time periods for the future time period according to the historical meteorological parameters and the future meteorological parameters if the preprocessing is completed;

[0015] The power load data prediction module is used to predict the future power load data of the future time period based on the historical power load data of the optimal time period and the adjacent time period.

[0016] A third aspect of the present invention provides an electronic device, the electronic device comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the regional power load forecasting method as described in the first aspect above.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the regional power load forecasting method as described in the first aspect above.

[0021] A fifth aspect of the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the regional power load forecasting method as described in the first aspect above.

[0022] In this embodiment, historical meteorological parameters and historical power load data of multiple historical time periods, as well as future meteorological parameters of future time periods, are queried for a designated area; the power load data is preprocessed; if the preprocessing is completed, the most similar historical time period is found as the best time period based on the historical meteorological parameters and the future meteorological parameters, and adjacent and similar historical time periods are found as adjacent time periods; the future power load data of the future time period is predicted based on the historical power load data of the best time period and the adjacent time period. This embodiment searches for a suitable historical time period based on the similarity of meteorological conditions as a high-value reference object, reduces the impact of information forgetting, thereby improving the accuracy of power load prediction and improving the quality of power grid operation and planning.

[0023] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 It is a flow chart of a regional power load forecasting method provided in Example 1 of the present invention.

[0026] Figure 2 It is a structural schematic diagram of a regional power load forecasting device provided in Example 2 of the present invention.

[0027] Figure 3 It is a structural schematic diagram of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can cover sequential implementations other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] Embodiment 1

[0031] See also Figure 1 , shows a flow chart of a regional power load forecasting method provided by Embodiment 1 of the present invention, the method can be executed by a regional power load forecasting device, the regional power load forecasting device can be implemented in the form of hardware and / or software, and the regional power load forecasting device can be configured in an electronic device. Figure 1 As shown, the method includes:

[0032] Step 101: query the historical meteorological parameters and historical power load data of multiple historical time periods, as well as future meteorological parameters of future time periods, for a designated area.

[0033] In this embodiment, multiple areas may be divided, and multiple time periods may be set for each area. The time period before and before the current time is the historical time period, and the time period after the current time is the future time period.

[0034] In the same area, since meteorological factors are correlated with power load, historical meteorological parameters and historical power load data for multiple historical time periods, as well as future meteorological parameters for future time periods, can be queried for the specified area to assist in predicting future power load data for future time periods.

[0035] Among them, the historical meteorological parameters of the historical time period are the real recorded meteorological parameters, the future meteorological parameters of the future time period are the meteorological parameters predicted by the weather forecast platform, and the types of the historical meteorological parameters of the historical time period are the same as the types of the future meteorological parameters of the future time period, for example, the maximum, minimum and mean values ​​of temperature and humidity, the maximum, minimum and average values ​​of irradiance, and so on.

[0036] Step 102: pre-process the historical power load data.

[0037] In this embodiment, the historical power load data of each historical time period can be preprocessed to ensure the smoothness and consistency of the order of magnitude of the historical power load data in the time series, improve the quality of the historical power load data, and thus improve the accuracy of subsequent predictions of future power load data in future time periods.

[0038] In the specific implementation, in the historical power load data, abnormal load parameters are manifested in various forms, which can be roughly divided into: 1) missing values ​​or zero values ​​appear in the historical power load data; 2) sudden increase or sudden decrease in the historical power load data.

[0039] For the former case, if there are load parameters with zero values ​​and / or missing values ​​in the historical power load data, the load parameters with zero values ​​are deleted, and the load parameters with missing values ​​and / or deleted zero values ​​are completed.

[0040] Furthermore, for missing values ​​and zero values ​​appearing in the historical power load data, the load parameters that are zero values ​​in the historical power load data can be deleted to form missing data; then, several matching parameters that are closest to the missing data are selected, and the weighted average is used as the missing load parameter.

[0041] The weighted average is expressed as: p(d)=w1p(d1)+w2p(d2)+w3p(d3)+…+w m p(d m ).

[0042] Where p(d) is the dth missing load parameter, p(d1), p(d2), p(d3), …, p(d m ) is the selected d1th load parameter, d2th load parameter, d3th load parameter, ..., d3th load parameter, etc., which are closest to the dth load parameter. m load parameters; w1, w2, w3, ..., w m are the d1th load parameter, the d2th load parameter, the d3th load parameter, ..., the d m The weight of each load parameter can be calculated using methods such as the Euclidean distance between the load parameters.

[0043] For the latter case, if there are abnormally increased or decreased load parameters in the historical power load data, callback processing is performed on the abnormally increased or decreased load parameters.

[0044] Since the historical power load data shows obvious periodic (especially seasonal) characteristics, the load parameters that have increased or decreased abnormally can be called back based on the periodic characteristics to correct the load parameters that have increased or decreased abnormally.

[0045] In a specific implementation, the power load is p(h,n), where h represents the sequence number in a time period, h=1,2,…,H, and n represents the sequence number of the time period, n=1,2…,N.

[0046] Calculate the load parameters p(h,n) with the same serial number in the historical power load data to calculate the average value E(h) and the mean square error V(h), where:

[0047] Calculate the ratio of the load deviation to the square root of the mean square error as the mutation rate of the load parameter δ(h,n); the load deviation is the absolute value of the difference between the load parameter and the average value, then,

[0048] If the mutation rate δ(h,n) is greater than or equal to the preset mutation threshold, then on the one hand, the adjacent load parameters are assigned to the first or last, that is, if n = 1, then If n = N, then On the other hand, the average value of adjacent load parameters is calculated to assign the value to the non-first or non-last load parameter, that is, if 1<n<N, then To implement callback processing for load parameters.

[0049] If the mutation rate δ(h,n) is less than the preset mutation threshold, the load parameters are kept unchanged and no callback is performed on the load parameters, that is,

[0050] Furthermore, the degree of identification and callback can be changed by adjusting the size of the mutation threshold: the larger the mutation threshold, the lower the degree of callback.

[0051] In addition, since the historical power load data is periodic, after the historical power load data is processed, it can be post-processed under the horizontal coordinate to ensure that the power load data within a time period has a certain smoothness.

[0052] In a specific implementation, a sliding first window is added to the historical power load data before the callback processing, and the load parameters in the window are expressed as p(n)=[p(1)p(2),…,p(15)]. The median of the load parameters is calculated in the first window to obtain the first power load sequence p 1 (n).

[0053] For example, the first power load sequence p 1 (n) can be expressed as:

[0054]

[0055] A sliding second window is added to the historical power load data after callback processing, with the first power load sequence p 1 (n) is used as a reference, and the median of the load parameters is calculated in the second window to obtain the second power load sequence p 2 (n).

[0056] For example, the second power load sequence p 2 (n) can be expressed as:

[0057]

[0058] Smoothing the plurality of second power load sequences to obtain a third power load sequence p p (n).

[0059] Exemplarily, the smoothing process can be expressed as:

[0060] p p (n) = 0.2p 2 (n-1)+0.6p 2 (n)+0.2p 2 (n+1)

[0061] The change rate ζ(n) on the same load parameter is calculated for the historical power load data after callback processing and the third power load sequence.

[0062] Exemplarily, the rate of change ζ(n) can be expressed as:

[0063]

[0064] If the mutation rate ζ(n) is greater than or equal to the preset change threshold, it is determined that the callback processing of the load parameter is effective, and the load parameter is maintained as the newly generated value.

[0065] If the mutation rate ζ(n) is less than the preset change threshold, it is determined that the callback processing of the load parameter is invalid, the callback processing of the load parameter is ignored, and the load parameter is kept at the original value.

[0066] Step 103: If the preprocessing is completed, the most similar historical time period is found as the best time period for the future time period based on the historical meteorological parameters and the future meteorological parameters, and adjacent and similar historical time periods are found as the neighboring time periods.

[0067] When the preprocessing of the power load data is completed, the historical meteorological parameters of each historical time period can be compared with the future meteorological parameters of the future time period, so as to find the most similar historical time period as the optimal time period for the future time period and find adjacent and similar historical time periods as neighboring time periods.

[0068] In one embodiment of the present invention, step 103 may include the following steps:

[0069] Step 1031: extract meteorological features from historical meteorological parameters and future meteorological parameters respectively.

[0070] In this embodiment, the same meteorological features may be extracted from the historical meteorological parameters and the future meteorological parameters, for example, the maximum daily power load, the daily power load variance, the daily power load standard deviation, the daily power load variation coefficient, and the like.

[0071] Step 1032: cluster the historical time periods according to the meteorological characteristics of the historical meteorological parameters to obtain multiple clusters representing weather types.

[0072] In this embodiment, a clustering algorithm such as K-means may be used to cluster historical time periods according to the meteorological characteristics of historical meteorological parameters to obtain multiple clusters representing weather types.

[0073] Among them, weather types include sunny, sunny to cloudy, cloudy, rainy, foggy and hazy, etc.

[0074] Step 1033: Calculate the Euclidean distance between the meteorological characteristics of the future meteorological parameters and the center point of each cluster.

[0075] Each cluster has a center point, and the Euclidean distance between the meteorological characteristics of the future meteorological parameters and the center point of each cluster can be calculated.

[0076] The calculation formula of Euclidean distance is as follows:

[0077]

[0078] Among them, i represents one of the weather types; j represents the jth parameter value; x0 represents the future meteorological parameters, x i represents the center point of the i-th cluster.

[0079] Step 1034: Divide the future time periods into clusters with the smallest Euclidean distance.

[0080] In this embodiment, each Euclidean distance is compared, and the future time period is divided into the cluster with the smallest Euclidean distance.

[0081] Step 1035: In the same cluster, find the most similar historical time period for the future time period as the optimal time period, and find adjacent and similar historical time periods as the neighboring time periods.

[0082] In the cluster where the future time period is located, the future meteorological parameters of the future time period are compared with the historical meteorological parameters of each historical time period, and the most similar historical time period is found as the best time period for the future time period, and the adjacent and similar historical time periods are found as the adjacent time periods.

[0083] In a specific implementation, in the same cluster, the future meteorological parameters and the various historical meteorological parameters are normalized to be dimensionless.

[0084] If the normalization process is completed, the correlation coefficient between the future meteorological parameters and each historical meteorological parameter is calculated using algorithms such as Pearson and MSE (Mean Square Error).

[0085] Exemplarily, the correlation coefficients between the parameter values ​​of the future meteorological parameters and the parameter values ​​of the historical meteorological parameters are calculated, and the average value of the correlation coefficients is calculated as the correlation coefficient.

[0086] Within the range of all historical time periods, the historical time period with the highest correlation coefficient is determined as the optimal time period.

[0087] Within the range of N (N is a positive integer, such as 7, 10, 15, etc.) historical time periods sorted before the future time period, determine the historical time period with the highest correlation coefficient as the adjacent time period.

[0088] Step 104: predict future power load data for a future time period based on the historical power load data for the best time period and adjacent time periods.

[0089] In practical applications, the historical power load data of the best time period and the historical power load data of the adjacent time periods provide references from two different directions, so that the future power load data of the future time period can be predicted by combining the two.

[0090] In one embodiment of the present invention, step 104 may include the following steps:

[0091] Step 1041: Perform empirical mode decomposition on the historical power load data of the optimal time period and the adjacent time periods to obtain the total component of the intrinsic mode function.

[0092] In this embodiment, empirical mode decomposition may be performed on the historical power load data of the optimal time period and the adjacent time periods to obtain the total component of the intrinsic mode function (IMF).

[0093] Step 1042: Reconstruct the total component of the intrinsic mode function using sample entropy to obtain high-frequency components and oscillation components.

[0094] Taking into account that complex and fluctuating data such as historical power load data have a large number of IMF total components after a single decomposition, and the degree of chaos of some IMF total components is similar, in order to reduce the amount of prediction calculations and improve the prediction efficiency, in this embodiment, time series entropy is introduced to characterize the degree of chaos of the primary component (i.e., the IMF total component), and a sample entropy algorithm with stronger representation ability is selected to reconstruct the IMF total component to obtain high-frequency, oscillating and other components for modeling.

[0095] Furthermore, the sample entropy algorithm has a better performance in characterizing the degree of data chaos, which can make the boundary values ​​of the reconstructed sequence components clearer and the increase in computational cost is not obvious. Therefore, choosing to reconstruct the total components of the IMF after a one-time decomposition of the empirical mode decomposition algorithm based on the sample entropy algorithm can ensure the accurate reconstruction of the mode, thereby improving the subsequent prediction efficiency.

[0096] Step 1043: Perform variational mode decomposition on the high-frequency components to obtain intrinsic mode function sub-components.

[0097] Considering that the sample entropy value of the high-frequency component reconstructed by the sample entropy algorithm is still large, it still has strong volatility and randomness, which is not conducive to load prediction. At the same time, since the empirical mode decomposition algorithm has limited decomposition ability for high-frequency noisy signals, and the variational mode decomposition is not affected by the selection of local extreme points, it can better handle the noise and interference in the signal. Therefore, this embodiment also proposes to use the variational mode decomposition method to perform secondary variational mode decomposition on the high-frequency component to obtain IMF sub-components, deeply mine data features, and improve the prediction accuracy of high-frequency components.

[0098] Step 1044: predict the first sub-load data of the future time period according to the intrinsic mode function sub-component.

[0099] In this embodiment, a deep learning model may be called to predict the first sub-load data of a future time period based on the IMF sub-component.

[0100] In a specific implementation, a long short-term memory network LSTM may be loaded, wherein the long short-term memory network LSTM may use the grey wolf algorithm to optimize the hyperparameters to obtain the optimal number of hidden layer neurons and T, a, A, and C parameters.

[0101] The intrinsic mode function IMF sub-component is input into the long short-term memory network LSTM to predict the first sub-load data of the future time period.

[0102] Since there are obvious differences in power load under different meteorological conditions, this embodiment can effectively predict the regional power load under different meteorological conditions. The empirical mode decomposition algorithm can reduce the volatility of regional historical power load under extreme weather conditions. The gray wolf algorithm optimizes the parameters of LSTM, which can improve the accuracy of LSTM prediction under complex meteorological conditions, and thus effectively predict the regional power load under different meteorological conditions.

[0103] Step 1045: predict the second sub-load data of the future time period according to the oscillation component.

[0104] In this embodiment, a machine learning model may be called to predict the second sub-load data of a future time period based on the oscillation component.

[0105] In a specific implementation, a multiple linear regression model may be loaded, and the oscillation component may be input into the multiple linear regression model to predict the second sub-load data of a future time period.

[0106] Step 1046: Combine the first sub-load data and the second sub-load data into future power load data for a future time period.

[0107] In this embodiment, the first sub-load data and the second sub-load data may be combined into future power load data of a future time period in a linear or nonlinear manner.

[0108] In this embodiment, historical meteorological parameters and historical power load data of multiple historical time periods, as well as future meteorological parameters of future time periods, are queried for a designated area; the power load data is preprocessed; if the preprocessing is completed, the most similar historical time period is found as the best time period based on the historical meteorological parameters and the future meteorological parameters, and adjacent and similar historical time periods are found as adjacent time periods; the future power load data of the future time period is predicted based on the historical power load data of the best time period and the adjacent time period. This embodiment searches for a suitable historical time period based on the similarity of meteorological conditions as a high-value reference object, reduces the impact of information forgetting, thereby improving the accuracy of power load prediction and improving the quality of power grid operation and planning.

[0109] Embodiment 2

[0110] See also Figure 2 , shows a schematic diagram of the structure of a regional power load prediction device provided by Embodiment 2 of the present invention. Figure 2 As shown, the device comprises:

[0111] The regional data acquisition and query module 201 is used to query the historical meteorological parameters and historical power load data of multiple historical time periods for a designated area, as well as future meteorological parameters of future time periods;

[0112] A preprocessing module 202, used for preprocessing the historical power load data;

[0113] A time period searching module 203 is used for searching the most similar historical time period as the best time period and searching the adjacent and similar historical time periods as the neighboring time periods for the future time period according to the historical meteorological parameters and the future meteorological parameters if the preprocessing is completed;

[0114] The power load data prediction module 204 is used to predict the future power load data of the future time period based on the historical power load data of the optimal time period and the adjacent time periods.

[0115] In one embodiment of the present invention, the preprocessing module 202 includes:

[0116] A data cleaning module, for deleting the load parameters with zero values ​​and / or missing values ​​if there are load parameters with zero values ​​and / or missing values ​​in the historical power load data, and for completing the load parameters with missing values ​​and / or deleted zero values;

[0117] The callback processing module is used to perform callback processing on the load parameters that have abnormally increased or decreased if there are load parameters that have abnormally increased or decreased in the historical power load data.

[0118] In one embodiment of the present invention, the callback processing module is further used for:

[0119] Calculate the average value and mean square error of the load parameters with the same sequence number in the historical power load data;

[0120] Calculate the ratio between the load deviation and the square root of the mean square error as the mutation rate of the load parameter; the load deviation is the absolute value of the difference between the load parameter and the average value;

[0121] If the mutation rate is greater than or equal to a preset mutation threshold, the adjacent load parameters are assigned to the first or last, and the average value of the adjacent load parameters is calculated to assign the values ​​to the non-first or non-last load parameters, so as to implement callback processing for the load parameters;

[0122] Adding a sliding first window to the historical power load data before callback processing, calculating the median of the load parameter in the first window, and obtaining a first power load sequence;

[0123] Adding a second slidable window to the historical power load data after callback processing, and calculating the median of the load parameter in the second window with reference to the first power load sequence to obtain a second power load sequence;

[0124] performing smoothing processing on a plurality of said second electric load sequences to obtain a third electric load sequence;

[0125] Calculate the change rate of the historical power load data after callback processing and the third power load sequence on the same load parameter;

[0126] If the mutation rate is greater than or equal to a preset change threshold, it is determined that the callback processing of the load parameter is effective;

[0127] If the mutation rate is less than a preset change threshold, it is determined that the callback processing of the load parameter is invalid, and the callback processing of the load parameter is ignored.

[0128] In one embodiment of the present invention, the time period finding module 203 includes:

[0129] A meteorological feature extraction module, used to extract meteorological features from the historical meteorological parameters and the future meteorological parameters respectively;

[0130] A clustering module, used for clustering the historical time period according to the meteorological characteristics of the historical meteorological parameters to obtain a plurality of clusters representing weather types;

[0131] A Euclidean distance calculation module, used for calculating the Euclidean distance between the meteorological characteristics of the future meteorological parameters and the center point of each cluster;

[0132] A time period division module, used for dividing the future time period into the clusters with the smallest Euclidean distance;

[0133] The cluster search module is used to search the most similar historical time period as the best time period for the future time period in the same cluster, and to search the adjacent and similar historical time periods as the neighboring time periods.

[0134] In one embodiment of the present invention, the cluster search module is further used for:

[0135] In the same cluster, normalizing the future meteorological parameter and each of the historical meteorological parameters;

[0136] If the normalization process is completed, the correlation coefficient between the future meteorological parameter and each of the historical meteorological parameters is calculated;

[0137] Within the range of all the historical time periods, determining the historical time period with the highest correlation coefficient as the optimal time period;

[0138] Within the range of N historical time periods arranged before the future time period, the historical time period with the highest correlation coefficient is determined as the adjacent time period.

[0139] In one embodiment of the present invention, the power load data prediction module 204 includes:

[0140] A primary decomposition module, used for performing empirical mode decomposition on the historical power load data of the optimal time period and the adjacent time period to obtain the total component of the intrinsic mode function;

[0141] A reconstruction module, used to reconstruct the total component of the intrinsic mode function by using sample entropy to obtain a high-frequency component and an oscillation component;

[0142] A secondary decomposition module, used for performing variational mode decomposition on the high-frequency component to obtain an intrinsic mode function sub-component;

[0143] A first sub-load data prediction module, configured to predict the first sub-load data of the future time period according to the intrinsic mode function sub-component;

[0144] A second sub-load data prediction module, used for predicting the second sub-load data of the future time period according to the oscillation component;

[0145] The power load data fusion module is used to merge the first sub-load data and the second sub-load data into future power load data of the future time period.

[0146] In one embodiment of the present invention, the first sub-load data prediction module is further used for:

[0147] Load the long short-term memory network;

[0148] The intrinsic mode function sub-component is input into the long short-term memory network to predict the first sub-load data of the future time period.

[0149] In one embodiment of the present invention, the second sub-load data prediction module is further used for:

[0150] Load the multiple linear regression model;

[0151] The oscillation component is input into the multivariate linear regression model to predict the second sub-load data of the future time period.

[0152] The regional power load forecasting device provided in the embodiment of the present invention can execute the regional power load forecasting method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the regional power load forecasting method.

[0153] Embodiment 3

[0154] See also Figure 3 , shows a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0155] like Figure 3 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0156] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0157] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the regional power load forecasting method.

[0158] In some embodiments, the regional power load forecasting method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the regional power load forecasting method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the regional power load forecasting method in any other appropriate manner (e.g., by means of firmware).

[0159] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0160] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0161] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0162] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0163] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0164] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0165] Embodiment 4

[0166] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method for predicting regional power load provided in any embodiment of the present invention is implemented.

[0167] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0168] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0169] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for regional power load forecasting, characterized in that: include: Query the historical meteorological parameters and historical power load data of multiple historical time periods for the specified area, as well as the future meteorological parameters of the future time period; Preprocessing the historical power load data; If the preprocessing is completed, then according to the historical meteorological parameters and the future meteorological parameters, the most similar historical time period is searched as the optimal time period for the future time period, and the adjacent and similar historical time periods are searched as the adjacent time periods; The future power load data of the future time period is predicted based on the historical power load data of the optimal time period and the adjacent time periods.

2. The method according to claim 1, characterized in that The preprocessing of the historical power load data includes: If there are load parameters with zero values ​​and / or missing values ​​in the historical power load data, the load parameters with zero values ​​are deleted, and the load parameters with missing values ​​and / or deleted zero values ​​are supplemented; If there are load parameters that are abnormally increased or decreased in the historical power load data, callback processing is performed on the load parameters that are abnormally increased or decreased.

3. The method according to claim 2, characterized in that The callback processing of the load parameter that has abnormal increase or abnormal decrease includes: Calculate the average value and mean square error of the load parameters with the same sequence number in the historical power load data; Calculate the ratio between the load deviation and the square root of the mean square error as the mutation rate of the load parameter; the load deviation is the absolute value of the difference between the load parameter and the average value; If the mutation rate is greater than or equal to a preset mutation threshold, the adjacent load parameters are assigned to the first or last, and the average value of the adjacent load parameters is calculated to assign the values ​​to the non-first or non-last load parameters, so as to implement callback processing for the load parameters; Adding a sliding first window to the historical power load data before callback processing, calculating the median of the load parameter in the first window, and obtaining a first power load sequence; Adding a second slidable window to the historical power load data after callback processing, and calculating the median of the load parameter in the second window with reference to the first power load sequence to obtain a second power load sequence; performing smoothing processing on a plurality of said second electric load sequences to obtain a third electric load sequence; Calculate the change rate of the historical power load data after callback processing and the third power load sequence on the same load parameter; If the mutation rate is greater than or equal to a preset change threshold, it is determined that the callback processing of the load parameter is effective; If the mutation rate is less than a preset change threshold, it is determined that the callback processing of the load parameter is invalid, and the callback processing of the load parameter is ignored.

4. The method according to claim 1, characterized in that: The step of searching the most similar historical time period as the optimal time period and searching adjacent and similar historical time periods as the neighboring time periods for the future time period based on the historical meteorological parameters and the future meteorological parameters includes: Extracting meteorological characteristics from the historical meteorological parameters and the future meteorological parameters respectively; Clustering the historical time period according to the meteorological characteristics of the historical meteorological parameters to obtain a plurality of clusters representing weather types; Calculating the Euclidean distance between the meteorological characteristics of the future meteorological parameters and the center point of each cluster; dividing the future time period into the clusters with the smallest Euclidean distance; In the same cluster, the most similar historical time period is found as the optimal time period for the future time period, and the adjacent and similar historical time periods are found as the neighboring time periods.

5. The method according to claim 4, characterized in that The step of searching the most similar historical time period as the optimal time period for the future time period and searching the adjacent and similar historical time periods as the neighboring time periods in the same cluster includes: In the same cluster, normalizing the future meteorological parameter and each of the historical meteorological parameters; If the normalization process is completed, the correlation coefficient between the future meteorological parameter and each of the historical meteorological parameters is calculated; Within the range of all the historical time periods, determining the historical time period with the highest correlation coefficient as the optimal time period; Within the range of N historical time periods arranged before the future time period, the historical time period with the highest correlation coefficient is determined as the adjacent time period.

6. The method according to any one of claims 1 to 5, characterized in that The predicting of the future power load data of the future time period based on the historical power load data of the optimal time period and the adjacent time period includes: Performing empirical mode decomposition on the historical power load data of the optimal time period and the adjacent time period to obtain a total component of an intrinsic mode function; Reconstructing the total component of the intrinsic mode function by using sample entropy to obtain high-frequency components and oscillation components; Performing variational mode decomposition on the high frequency component to obtain an intrinsic mode function sub-component; Predicting the first sub-load data of the future time period according to the intrinsic mode function sub-component; predicting second sub-load data of the future time period according to the oscillation component; The first sub-load data and the second sub-load data are combined into future power load data for the future time period.

7. The method according to claim 6, characterized in that The predicting the first sub-load data of the future time period according to the intrinsic mode function sub-component includes: Load the long short-term memory network; The intrinsic mode function sub-component is input into the long short-term memory network to predict the first sub-load data of the future time period.

8. The method according to claim 6, characterized in that The predicting the second sub-load data of the future time period according to the oscillation component includes: Load the multiple linear regression model; The oscillation component is input into the multivariate linear regression model to predict the second sub-load data of the future time period.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the regional power load forecasting method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for regional power load forecasting according to any one of claims 1 to 8 is implemented.

Images