Multi-dimensional power demand prediction method and device, storage medium and electronic equipment

Abstract

The invention relates to a multi-dimensional power demand prediction method and device, a storage medium and electronic equipment, and relates to the technical field of power demand prediction. Multi-dimensional power demand prediction can be realized, and a reliable power demand prediction result is obtained. The method comprises the steps that electric power data of a target area are acquired, the electric power data comprise historical generating capacity and total distant view energy consumption, and the total distant view energy consumption represents a predicted value of an energy consumption demand of the target area within a preset duration; a data set of historical power generation is determined, power demand short-term prediction is carried out according to the data set, a short-term power generation prediction value of the target area is obtained, and the data set is obtained by arranging the historical power generation according to time; performing power demand medium-term prediction according to the data set and a preset sliding window to obtain a medium-term generating capacity prediction value of the target area; and performing long-term power demand prediction according to the total long-range energy consumption to obtain a long-term power generation prediction value of the target area.

Description

Technical Field

[0001] The present disclosure relates to the technical field of power demand prediction, and in particular, to a multi-dimensional power demand prediction method, device, storage medium and electronic device. Background Art

[0002] When planning and operating the power system, it is necessary to analyze the historical power generation data and conduct a comprehensive assessment of the factors that may affect the supply and demand of power in the future, that is, the power demand forecast. However, due to the influence of periodic and irregular changes, the historical power generation data will fluctuate greatly, and it is difficult to show the development trend, and it is impossible to obtain reliable power demand forecast results. Summary of the invention

[0003] In order to solve the deficiencies of the prior art, the present disclosure provides a multi-dimensional power demand prediction method, device, storage medium and electronic device.

[0004] In order to achieve the above objectives, in a first aspect, the present disclosure provides a method for predicting power demand, comprising:

[0005] Acquire power data of the target area, wherein the power data includes historical power generation and total future energy consumption, wherein the total future energy consumption represents a predicted value of energy consumption demand of the target area within a preset time period;

[0006] Determine a data set of the historical power generation, perform short-term power demand forecasting based on the data set, and obtain a short-term power generation forecast value of the target area, wherein the data set is obtained by arranging the historical power generation according to time;

[0007] Performing a medium-term forecast of power demand based on the data set and a preset sliding window or a preset attenuation coefficient to obtain a medium-term power generation forecast value for the target area;

[0008] A long-term forecast of electricity demand is performed based on the total prospective energy consumption to obtain a long-term power generation forecast value for the target area.

[0009] In a second aspect, the present disclosure provides a power demand prediction device, comprising:

[0010] An acquisition module is configured to acquire power data of a target area, wherein the power data includes historical power generation and total future energy consumption, wherein the total future energy consumption represents a predicted value of energy consumption demand of the target area within a preset time period;

[0011] A first execution module is configured to determine a data set of the historical power generation, perform short-term power demand forecasting based on the data set, and obtain a short-term power generation forecast value of the target area;

[0012] A second execution module is configured to perform a medium-term forecast of power demand based on the data set and a preset sliding window or a preset attenuation coefficient to obtain a medium-term power generation forecast value of the target area;

[0013] The third execution module is configured to perform long-term forecasting of electricity demand according to the total prospective energy consumption to obtain a long-term power generation forecast value of the target area.

[0014] In a third aspect, the present disclosure provides a non-temporary computer-readable storage medium having a computer program stored thereon, which implements the method described in the first aspect when executed by a processor.

[0015] In a fourth aspect, the present disclosure provides an electronic device, including:

[0016] a memory having a computer program stored thereon;

[0017] A processor is used to execute the computer program in the memory according to the method described in the first aspect.

[0018] In a fifth aspect, the present disclosure provides a computer program product, including a computer program, which implements the method described in the first aspect when executed by a processor.

[0019] Through the above technical solution, short-term power demand forecasts are made based on the historical power generation data set; medium-term power demand forecasts are made based on the historical power generation data set and a preset sliding window or preset attenuation coefficient; and long-term power demand forecasts are made based on the total energy consumption in the future, thus realizing multi-dimensional power demand forecasts. The medium-term power demand forecast based on the sliding window can avoid the impact of periodic and irregular changes in the data set on the medium-term power demand forecast results, thereby improving the accuracy and reliability of the medium-term power generation forecast value.

[0020] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present disclosure but do not constitute a limitation of the present disclosure. In the accompanying drawings:

[0022] Figure 1 is a flowchart of a method for predicting power demand according to an exemplary embodiment of the present disclosure.

[0023] Figure 2 FIG. 4 is a schematic diagram of sliding a sliding window according to an exemplary embodiment of the present disclosure.

[0024] Figure 3is a block diagram of a power demand prediction device according to an exemplary embodiment of the present disclosure.

[0025] Figure 4 is a block diagram of an electronic device according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0026] The specific implementation of the present disclosure is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the present disclosure, and is not used to limit the present disclosure.

[0027] It should be noted that all actions of acquiring signals, information or data in the present disclosure are carried out in compliance with the relevant data protection laws and policies of the country where the device is located and with the authorization given by the owner of the corresponding device.

[0028] It is worth noting that although the current power demand forecast can realize short-term power demand forecast, medium-term power demand forecast and long-term power demand forecast, the forecast process does not take into account the power constraint, which will lead to unreliable forecast results. And it is impossible to simultaneously realize the short-term, medium-term and long-term power demand forecast.

[0029] The inventors have discovered that in the process of medium-term forecasting of electricity demand, historical power generation data will fluctuate greatly due to the influence of periodic and irregular changes, making it difficult to show development trends and impossible to obtain reliable medium-term forecast results for electricity demand.

[0030] In view of this, the present disclosure provides a multidimensional electricity demand forecasting method, device, storage medium and electronic device, which can avoid the impact of periodic changes and irregular changes in historical power generation data sets on the medium-term forecast results of electricity demand, and improve the accuracy and reliability of the medium-term forecast results of electricity demand.

[0031] Figure 1 1 is a flow chart of a method for predicting power demand according to an exemplary embodiment of the present disclosure. The method can be applied to electronic devices, including but not limited to personal computers and mobile terminals. The method may include the following steps:

[0032] In step S101, power data of a target area is obtained, wherein the power data includes historical power generation and total future energy consumption, wherein the total future energy consumption represents a predicted value of energy consumption demand of the target area within a preset time period.

[0033] It is worth noting that the target area includes but is not limited to regions, cities, companies, etc. The historical power generation may include the power generation of the target area in the historical period, and the historical period is not limited to the target area in the past few months, years, decades, etc. The total prospective energy consumption represents the energy consumption of the target area in the forecasted or planned long-term. The total prospective energy consumption can be used for energy planning and policy making to evaluate the energy demand and supply of the target area in the long term in the future.

[0034] In step S102, a data set of the historical power generation is determined, and a short-term power demand forecast is performed based on the data set to obtain a short-term power generation forecast value of the target area. The data set is obtained by arranging the historical power generation according to time.

[0035] The data set may be obtained by sorting the historical power generation of the target area in chronological order from the beginning to the end.

[0036] In step S103, a medium-term forecast of power demand is performed based on the data set and a preset sliding window or a preset attenuation coefficient to obtain a medium-term power generation forecast value of the target area.

[0037] In step S104, a long-term forecast of power demand is performed based on the total future energy consumption to obtain a long-term power generation forecast value of the target area.

[0038] It is worth noting that the short-term forecast of electricity demand represents the forecast of electricity demand and supply in the next few hours to a few years, the medium-term forecast of electricity demand represents the forecast of electricity demand and supply in the next few months to ten years, and the long-term forecast of electricity demand represents the forecast of electricity demand and supply in the next few decades. In the embodiment of the present disclosure, the short-term power generation forecast value may include the power demand and supply of the target area in the next five years, the medium-term power generation forecast value may include the power demand and supply of the target area in the next ten years, and the long-term power generation forecast value may include the power demand and supply of the target area in the next thirty years.

[0039] In the embodiment of the present disclosure, short-term power demand forecasting is performed based on the data set of historical power generation; medium-term power demand forecasting is performed based on the data set of historical power generation and a preset sliding window or a preset attenuation coefficient; and long-term power demand forecasting is performed based on the total energy consumption in the future, which can simultaneously realize short-term, medium-term and long-term power demand forecasting, and realize multi-dimensional power demand forecasting. The medium-term power demand forecasting based on the sliding window can avoid the influence of the periodic changes and irregular changes of the data set on the medium-term power demand forecasting results, thereby improving the accuracy and reliability of the medium-term power generation forecast value.

[0040] In order to facilitate those skilled in the art to better understand the power demand forecasting method provided by the present disclosure, the steps of the method are described in detail below.

[0041] In a feasible implementation manner, in step S102, performing short-term forecasting of power demand according to the data set to obtain a short-term power generation forecast value of the target area may include:

[0042] After performing a level ratio inspection process on the data set, performing an accumulation process on the data set to obtain an intermediate data set, wherein the level ratio inspection process is used to adjust the level ratios corresponding to all adjacent data in the data set to be within a preset interval, and the level ratio represents the ratio between adjacent data;

[0043] Calculate the average value of all adjacent data in the intermediate data set to obtain a mean data set;

[0044] constructing a fourth prediction model based on the data set and the mean data set, wherein the fourth prediction model is used to predict short-term power generation of the target area based on the data set corresponding to the historical power generation of the target area;

[0045] A short-term power generation forecast value of the target area is obtained according to the fourth forecast model.

[0046] For example, for the dataset x (0) =(x (0) (1), x (0) (2), …, x (0) (T)) performs an accumulation process to obtain the intermediate data set x (1) (k)=(x (0) (1), x (0) (1)+x (0) (2), …, x (0) (1)+x (0) (T)), that is,

[0047] Calculate the intermediate data set x by the following calculation formula (1) The average value of all adjacent data in (1) (k), get the mean data set z (1) (k)=(z (1) (2), z (1) (3), …, z (1) (T)):

[0048] z (1) (k) = (0.5x (1) (k)+0.5x (1) (k-1), k=1, 2,…T.

[0049] According to the data set x (1) and mean data set z (1) Construct the differential equation:

[0050] x (1) (k)+az (1) (k)=b, k=2, 3,...,T.

[0051] The corresponding whitened differential equation is:

[0052]

[0053] Let u=(a,b) T , Y=(x (0) (2), x (0) (3),…,x (0) (T)) T , By using the least squares method, we can find To achieve the minimum value:

[0054]

[0055] Solve the differential equation to obtain the candidate short-term power generation forecast value of the target area:

[0056]

[0057] get:

[0058] According to the data set x (0) and candidate short-term power generation forecasts The reliability of the prediction results of the differential equation is verified using residual test and / or level ratio deviation test.

[0059] For the residual test, the residual ε(k) is calculated by the following formula:

[0060]

[0061] exist When ε(k) < the first threshold, it is determined that the prediction result of the differential equation meets the general requirements; if ε(k) < the second threshold, it is determined that the prediction result of the differential equation meets the higher requirements. Among them, the first threshold is 0.2, and the second threshold is 0.1. The first threshold and the second threshold can be preset according to the prediction accuracy to ensure that the second threshold is less than the first threshold. This disclosure does not limit this.

[0062] For the level ratio deviation test, according to the data set x (0) The corresponding step ratio deviation ρ(k) is obtained by the following calculation formula based on the step ratio λ(k) and the preset development coefficient d:

[0063]

[0064] If ρ(k) < the third threshold, it is determined that the prediction result of the differential equation meets the general requirements; if ρ(k) < the fourth threshold, it is determined that the prediction result of the differential equation meets the higher requirements. Among them, the third threshold is 0.2, the fourth threshold is 0.1, and the third threshold and the fourth threshold can also be other values, but it is necessary to ensure that the fourth threshold is less than the third threshold, which is not limited in this disclosure.

[0065] Optionally, performing a level ratio test on the data set may include:

[0066] Performing a level ratio test on the data set to obtain level ratios corresponding to all adjacent data in the data set, wherein the level ratio represents a ratio between adjacent data;

[0067] When any of the level ratios is not within the preset interval, a translation transformation process is performed on the data set until the level ratios corresponding to all adjacent data in the data set are within the preset interval.

[0068] The preset interval represents that the fluctuation range of the power generation forecast value covers the fluctuation range of the historical power generation. The preset interval can be determined by the following calculation formula:

[0069]

[0070] For example, the dataset x (0) =(x (0) (1), x (0) (2), …, x (0) (T)), for the data set x (0) Perform a level ratio test and convert the data set x (0) Substitute all adjacent data in into the following level ratio calculation formula to obtain the data set x (0) The level ratio λ(k) of all adjacent data in is:

[0071]

[0072] When any level ratio is not within the preset interval θ, the constant c is determined by the following calculation formula for the data set x (0) Perform translation transformation to obtain a new data set y (0) , until the transformed dataset y (0) The level ratios λ(k) corresponding to all adjacent data in θ fall within θ, and the data set y is determined. (0) Pass the level ratio test:

[0073] y (0) (k) = x (0)(k)+c, k=1, 2,...,T.

[0074] Dataset y (0) =(y (0) (1), y (0) (2), …, y (0) The level ratio of (T)) is:

[0075]

[0076] In the disclosed embodiment, by performing a level ratio test on the data set, the development trend of each data in the data set can have a certain regularity and stability, thereby improving the reliability of the results of subsequent short-term power demand forecasting based on the data set.

[0077] For example, the target area's power generation data from 2010 to 2020 is used as the data set y to solve the intermediate data set x (1) (k), calculate the level ratio of power generation in the target area and obtain the level ratio data shown in Table 1.

[0078] Table 1

[0079] years 2011 2012 2013 2014 2015 2016 2017 2018 2019 time 0.8937 0.9487 0.9282 0.9458 0.9896 0.9530 0.9333 0.9225 0.9547

[0080] It can be seen from Table 1 that the level ratio of the power generation from 2010 to 2020 is within the interval [0.8338, 1.1994], the level ratio test is passed, and there is no need to transform the power generation from 2010 to 2020.

[0081] A differential equation is established and solved by the least square method to obtain the predicted data and the coefficient a=-0.0533 and the coefficient b=43870 of the differential equation.

[0082] The relative error, residual or level ratio deviation between the predicted data and the data set as shown in Table 2 is determined to detect the reliability of the predicted data.

[0083] Table 2

[0084]

[0085] It can be seen from Table 2 that the relative error between the predicted value and the true value is small, and the residual and level ratio deviation test model results meet the high requirements, so the establishment of the model is relatively reasonable. The power generation data of the target area in 2021-2025 as shown in Table 3 are obtained by using differential equations.

[0086] Table 3

[0087] years 2021 2022 2023 2024 2025 Power generation 80693 85071 89662 94780 100378

[0088] In the embodiment of the present disclosure, the equal-dimensional grey complement method is used to perform short-term prediction of electricity demand based on the data set to obtain the short-term power generation forecast value of the target area without the need for additional auxiliary information. The calculation process of the constructed fourth prediction model is simple, can adapt to the nonlinear changes of the data set, and has a certain stability.

[0089] It is worth noting that the present disclosure can perform medium-term forecasting of electricity demand in two ways: method 1: performing medium-term forecasting of electricity demand by moving average method; method 2: performing medium-term forecasting of electricity demand by exponential smoothing method.

[0090] Regarding method one: medium-term forecast of electricity demand is carried out through moving average method, which may include simple moving average method, weighted moving average method and trend moving average method.

[0091] In a feasible implementation manner, in step S103, performing a medium-term forecast of power demand based on the data set and the preset sliding window to obtain a medium-term power generation forecast value of the target area may include:

[0092] Controlling the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, determining an average value of the data of the data set in the preset sliding window before the first sliding, and an average value of the data of the data set in the preset sliding window after each sliding, to obtain a first moving average value corresponding to the data set;

[0093] constructing a first prediction model with a hysteresis deviation correction function according to the first moving average value, wherein the first prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area;

[0094] A medium-term power generation forecast value of the target area is obtained according to the first forecast model.

[0095] Among them, the number of data that the sliding window can cover is less than the total number of data in the data set, and the length of the sliding window can also be called the number of items of the moving average. The preset sliding step size can be preset according to the prediction accuracy, but the sliding step size is less than the number of data that the sliding window can cover. Figure 2 As shown, when the preset sliding step is 1, the sliding window is controlled to slide to the right by one data number each time.

[0096] For example, the dataset x (0) =(x (0) (1), x (0) (2), …, x (0) (T)), taking the number of moving average items as m and the preset step size as 1, the first moving average corresponding to the data set can be determined by the following calculation formula

[0097]

[0098] When the data set x (0) When the trend of all data in the above example fluctuates above and below a certain horizontal line, a first prediction model with the function of correcting lag deviation is constructed based on the first moving average value.

[0099]

[0100] Wherein, t=m, m+1,…,T.

[0101] The standard error S of the prediction result can be determined by the following calculation formula:

[0102]

[0103] If S < the fifth threshold, it is determined that the prediction result of the first prediction model meets the general requirements; if ε(k) < the sixth threshold, it is determined that the prediction result of the first prediction model meets the higher requirements. The fifth threshold and the sixth threshold can be preset according to the prediction accuracy, and this disclosure does not limit this.

[0104] In the embodiment of the present disclosure, when the change trend of all data in a data set fluctuates above and below a certain horizontal line, a first prediction model can be established through a moving average method to predict the medium-term power generation of the target area based on the data set corresponding to the historical power generation of the target area. This can avoid the impact of periodic and irregular changes in the data set on the medium-term forecast results of electricity demand, and the calculation process is simple, reducing computing power pressure.

[0105] It is worth noting that when the time series of the data set shows a linear increase or decrease trend, establishing the first prediction model through a single moving average method will lead to a lag deviation in the prediction results. A second moving average is needed to correct the lag deviation and thus establish the first prediction model.

[0106] In a feasible implementation manner, constructing a first prediction model with a function of correcting hysteresis deviation according to the first moving average value may include:

[0107] Determining a second moving average value corresponding to the data set according to the first moving average value;

[0108] Determine a first smoothing coefficient and a second smoothing coefficient according to the second moving average, wherein the first smoothing coefficient represents an intercept corresponding to a change trend of the data set, and the second smoothing coefficient represents a slope corresponding to a change trend of the data set;

[0109] A first prediction model having a function of correcting hysteresis deviation is constructed according to the first smoothing coefficient and the second smoothing coefficient.

[0110] For example, determine a moving average:

[0111]

[0112] Based on the first moving average method, we can calculate the second moving average by the following formula:

[0113]

[0114] According to the first moving average and the second moving average Determine the first smoothing coefficient a t and the second smoothing coefficient b t :

[0115]

[0116] Combined with the lag deviation of the moving average, the following linear trend prediction model (i.e., the first prediction model) is constructed according to the first smoothing coefficient and the second smoothing coefficient:

[0117]

[0118] Among them, it can be known from the linear trend prediction model that when n=0:

[0119] a t =x (0) (T),

[0120] x (0) (Tm-1) = x (0) (T)-(m-1)b t ,

[0121] Combining the above calculation formula, we can get the first moving average value

[0122]

[0123] The above calculation formula is transformed into the following two calculation formulas:

[0124]

[0125] Finally we get:

[0126]

[0127] Therefore, according to the first moving average and the second moving average Determine the first smoothing coefficient a t and the second smoothing coefficient b t .

[0128] When the time series of a data set shows a linear increasing or decreasing trend, the disclosed embodiment can establish a first prediction model through the quadratic moving average method to predict the medium-term power generation of the target area based on the data set corresponding to the historical power generation of the target area, and can correct the moving average lag deviation, thereby improving the reliability and accuracy of the output results of the first prediction model.

[0129] It is worth mentioning that the single moving average method averages the data of each period in the sliding window during the sliding process, that is, each period of data is treated equally. However, the amount of information contained in each period of data is actually different. For example, recent data contains more information about future situations. Therefore, in the prediction process, the importance of each period of data needs to be considered.

[0130] In a feasible implementation manner, in step S103, performing a medium-term forecast of power demand based on the data set and the preset sliding window to obtain a medium-term power generation forecast value of the target area may include:

[0131] Controlling the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, and determining, according to a preset weight corresponding to each data in the data set, a weighted average value of the data of the data set in the preset sliding window before the first sliding, and a weighted average value of the data of the data set in the preset sliding window after each sliding, to obtain a weighted moving average value corresponding to the data set;

[0132] According to the weighted moving average, a second prediction model with a function of correcting hysteresis deviation is constructed, wherein the second prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area;

[0133] A medium-term power generation forecast value of the target area is obtained according to the second forecast model.

[0134] It is worth noting that the preset weight can be preset according to the time difference between the data generation time and the current time. The closer the data generation time is to the current time, the greater its weight, and the farther the data generation time is from the current time, the smaller its weight.

[0135] For example, the dataset x (0) =(x (0) (1), x (0) (2), …, x (0)The preset weights corresponding to each data in (T) are w1, w2, ..., w T . Calculate the data set x by the following formula (0) The corresponding weighted moving average M tw :

[0136]

[0137] Among them, M tw represents the weighted moving average of period t, w i Characterization x (0) The weight of (T-m+1).

[0138] In a feasible implementation manner, constructing a second prediction model with a function of correcting hysteresis deviation according to the weighted moving average value may include:

[0139] Substituting the weighted moving average into the following calculation formula, the second prediction model with the function of correcting the lag deviation is obtained:

[0140] The weighted moving average M tw Substitute the following calculation formula to obtain the function of correcting the lag deviation and construct the second prediction model

[0141]

[0142] in, Represents the predicted value of the (T+1)th period, M tw Represents the weighted moving average of period T.

[0143] In the embodiment of the present disclosure, a preset weight is assigned to each data in the data set, and a first prediction model is established through the weighted moving average method to predict the medium-term power generation of the target area based on the data set corresponding to the historical power generation of the target area. Since newer data has a higher weight, the lag of the preset result is reduced, and the change of data in the data set can be responded to more quickly, thereby shortening the prediction time.

[0144] Regarding the second method: medium-term forecast of electricity demand is carried out through exponential smoothing method, which may include single exponential smoothing method, double exponential smoothing method and triple exponential smoothing method.

[0145] In a feasible implementation manner, performing a medium-term forecast of power demand according to the data set and a preset attenuation coefficient to obtain a medium-term power generation forecast value of the target area may include:

[0146] Controlling the sliding window to slide according to a preset step length until the sliding window covers all the data in the data set, and determining an average value of the data in the data set in the sliding window after the last sliding;

[0147] According to the data set, the preset attenuation coefficient and the average value, a third prediction model with a hysteresis deviation correction function is constructed, wherein the third prediction model is used to predict the mid-term power generation of the target area according to the data set sequence corresponding to the historical power generation of the target area;

[0148] A medium-term power generation forecast value of the target area is obtained according to the third forecast model.

[0149] Among them, the number of data that can be covered by the sliding window is less than the total number of data in the data set, and the length of the sliding window can also be called the number of items of the moving average. The preset sliding step size can be preset according to the prediction accuracy, but the sliding step size is less than the number of data that can be covered by the sliding window. In the embodiment of the present disclosure, the preset sliding step size is taken as 1, that is, the sliding window is controlled to slide one data to the right each time.

[0150] For example, the data set y = (y1, y2, ..., y t ), the preset attenuation coefficient is a, 0<a<1, the number of data that can be covered by the sliding window each time is N, and an exponential smoothing formula is constructed:

[0151]

[0152] The average value of the data in the sliding window after the last slide is determined by the following calculation formula:

[0153]

[0154] The average As y t-N The best estimate of , we get:

[0155]

[0156] Let a = 1 / N, Instead of y t-N , we can get the following calculation formula:

[0157]

[0158] It can be seen that It is the weighted average of the data set, and the weight coefficients of each data are a, a(1-a), a(1-a) respectively. 2 , …, then we can get:

[0159]

[0160] Then the third prediction model is constructed

[0161]

[0162] Among them, the smoothing index value of the tth period after the last sliding of the sliding window is used as the prediction value of the t+1 period.

[0163] In the disclosed embodiment, the weight assigned to each data is determined by a preset decay coefficient, and the weight can decrease with the increase of time and will never disappear completely, so as to ensure that the constructed third prediction model can capture the long-term trend, so as to adapt to the different change rates of the data set, and in this way, the importance of each data in the data set is taken into account to improve the accuracy and reliability of the prediction results. And when a new data point appears, the first prediction model can be quickly updated without re-sorting the data.

[0164] It is worth noting that the above method will lead to obvious lag deviation in the prediction results when a linear trend appears in the time series of the data set. It is necessary to perform secondary exponential smoothing on the basis of the above method to correct the lag deviation.

[0165] In a feasible implementation manner, constructing a third prediction model with a function of correcting hysteresis deviation according to the data set, the preset attenuation coefficient and the average value may include:

[0166] Determine a first predicted value and a second predicted value corresponding to the data set according to the data set, the preset attenuation coefficient and the average value;

[0167] Determine a first horizontal coefficient and a first trend coefficient corresponding to the data set according to the first predicted value, the second predicted value, and the preset attenuation coefficient, wherein the first horizontal coefficient is used to smooth first data in the data set that does not change with time, and the first trend coefficient is used to smooth second data in the data set that changes with time;

[0168] The third prediction model with the function of correcting lag deviation is constructed according to the horizontal coefficient and the trend coefficient.

[0169] It is worth noting that the first prediction value represents the first exponential smoothing value corresponding to the data set subjected to the first exponential smoothing, and the second prediction value represents the second exponential smoothing value corresponding to the data set subjected to the second exponential smoothing.

[0170] For example, the first prediction value is calculated by the above-mentioned exponential smoothing formula: Based on the first predicted value, a second exponential smoothing is performed to obtain the second predicted value.

[0171]

[0172] Wherein, a represents the preset attenuation coefficient.

[0173] When the data set y has a straight-line trend starting from the tth period, according to the first predicted value of the tth period and the second predicted value Determine the first level coefficient a t and the first trend coefficient b t :

[0174]

[0175] According to the first level coefficient a t and the first trend coefficient b t , build the third prediction model

[0176]

[0177] In the embodiments of the present disclosure, when a linear trend appears in the time series of a data set, the lag deviation is corrected by the quadratic exponential smoothing method to better handle local fluctuations and long-term trends in the data set, and stable prediction results can be provided when the data set changes are more complex, thereby improving the reliability of the medium-term prediction results of electricity demand.

[0178] It is worth noting that when the above method shows a quadratic curve trend in the time series of the data set, it will lead to obvious lag deviation in the prediction results, and triple exponential smoothing is required to correct the lag deviation.

[0179] In a feasible implementation manner, constructing a third prediction model with a function of correcting hysteresis deviation according to the data set, the preset attenuation coefficient and the average value may include:

[0180] Determine, according to the data set, the preset attenuation coefficient and the average value, a third predicted value, a fourth predicted value and a fifth predicted value corresponding to the data set;

[0181] Determine, according to the third predicted value, the fourth predicted value, the fifth predicted value and the preset attenuation coefficient, a second horizontal coefficient, a second trend coefficient and a seasonal coefficient corresponding to the data set, wherein the second horizontal coefficient is used to smooth first data in the data set that does not change with time, the second trend coefficient is used to smooth second data in the data set that changes with time, and the seasonal coefficient is used to smooth periodic fluctuations of the data set;

[0182] The third prediction model with the function of correcting the lag deviation is constructed according to the second horizontal coefficient, the second trend coefficient and the seasonal coefficient.

[0183] For example, the third prediction value is calculated by the above-mentioned exponential smoothing formula Based on the third predicted value, a second exponential smoothing is performed to obtain the fourth predicted value. Based on the fourth prediction value, three exponential smoothings are performed to obtain the fifth prediction value.

[0184]

[0185] Wherein, a represents the preset attenuation coefficient.

[0186] When the data set y has a straight-line trend starting from the tth period, according to the first predicted value of the tth period Second prediction value and the third predicted value Determine the second level coefficient a t , the second trend coefficient b t and the seasonal coefficient c t :

[0187]

[0188] According to the second level coefficient a t , the second trend coefficient b t and the seasonal coefficient c t , build the third prediction model

[0189]

[0190] In the embodiment of the present disclosure, when a quadratic curve trend appears in the time series of a data set, the lag deviation is corrected by the cubic exponential smoothing method to better handle the local fluctuations and long-term trends in the data set. When the data set changes more complexly, stable prediction results can be provided, thereby improving the reliability of the medium-term prediction results of electricity demand.

[0191] The following uses the target area's power generation data from 1978 to 2020 as the data set y, and uses the quadratic exponential smoothing method to make a medium-term forecast of power demand. Take a = 0.3, and determine the first exponential smoothing value and the second exponential smoothing value based on the power generation data in 1978. According to the data set prediction, the predicted values ​​of the target area from 2021 to 2030 are obtained, and the predicted values ​​are shown in Table 4.

[0192] Table 4

[0193] years 2021 2022 2023 2024 2025 Power generation (kWh) 79575 83002 86429 89855 93282 years 2026 2027 2028 2029 2030 Power generation (kWh) 96708 100135 103561 106988 110415

[0194] The Logistic model (Logistic Growth Model) is used to model the growth process of power supply and demand, and the following Logistic model is obtained:

[0195]

[0196] Among them, t represents the time independent variable, P(t) represents the dependent variable, K, P0, and r all represent parameters to be identified, and satisfy the following relationship: P(∞) = k, P(0) = P0. And because both power supply and demand are integers, K, P0, and r are all real numbers greater than 0.

[0197] Denote the obtained data set as: Ω = {P(t i )|1 < i < N}, where N represents the number of data sets, and P(t i ) represents the actual power supply and demand in the t i th year in the data set.

[0198] Then, through the following least squares model, K, P0, and r are determined:

[0199]

[0200] Based on the least squares model, an additional constraint condition P(2060) = C is introduced, so as to limit the prediction result P(2060) in the long-term vision year 2060 to be equal to the judgment value C. The judgment value represents the estimated value of power demand predicted by scientific methods and models based on current data, trends, policy impacts, and possible future development scenarios. So that the identified first prediction model can not only fit historical data more accurately, but also conform to the judgment result of the long-term development scenario, and the above third prediction model can be solved by the interior point method.

[0201] In a feasible implementation manner, in step S104, the long-term power demand prediction based on the total远景能源消费总量 can include:

[0202] Multiply the total远景能源消费总量 by a preset conversion value and divide it by a preset consumption coefficient to obtain the long-term power generation prediction value of the target area. The preset consumption coefficient is the product of the coal consumption coefficient and the electro-thermal equivalent coefficient, and the electro-thermal equivalent coefficient represents the conversion of the consumption of various energy sources in the power generation process into the consumption of standard coal.

[0203] Among them, the preset conversion value can be determined according to the non-fossil energy consumption ratio conversion value. The sum of the preset conversion value and the non-fossil energy consumption ratio conversion value is 1. For example, when the non-fossil energy consumption ratio conversion value is 70%, the preset conversion value is 30%; when the non-fossil energy consumption ratio conversion value is 80%, the preset conversion value is 20%.

[0204] It is worth noting that energy research institutions and / or experts can analyze the data that affect the total energy consumption, such as gross domestic product, urbanization rate, fixed asset investment, etc., based on the historical energy consumption data of the target area, such as the consumption of various energy qualities, economic growth data, and population changes, and combine the energy policies, energy conservation and emission reduction targets, etc. of the target area to determine the total energy consumption of the target area in the next few decades, that is, the long-term total energy consumption.

[0205] For example, the forecast results of energy research institutions and well-known experts on the total energy consumption of the target area from 2030 to 2060 are determined, and the forecast results shown in Table 5 are obtained:

[0206] Table 5

[0207] time unit 2030 2035 2040 2050 2060 2015 XX 55 58 2017 XX 51.9 53.5 54.2 2017 XX ≤60 2018 XX 55-55.6 56.4-57.4 56.5-57.9 2020 XX 56-58 48-53 2021 XX 59 58 57 56 55 2021 XX 58 53 48

[0208] As shown in Table 5, relevant studies before 2020 did not consider the impact of the strategy, and basically believed that the target area would slowly peak at 5.8 billion tons of standard coal equivalent (tce) in 2050. The research results released after 2020 fully considered the promotion of energy conservation and consumption reduction by the strategy, predicting that the total energy consumption will reach a peak in 2030, and then slowly decline, and basically stabilize around 2060, roughly at 4.8-5.5 billion tce.

[0209] According to the data shown in Table 5, the target area's preset energy consumption in 2060 is determined to be an average of 4.8-5.5 billion tce, that is, 5.3 billion tce, and the coal consumption coefficient is determined to be 0.3046, and the electric heat equivalent coefficient is 1.229 kg / kwh. Substituting the preset energy consumption of 5.3 billion tce, the preset conversion value of 20%, the coal consumption coefficient of 0.3046, and the electric heat equivalent coefficient of 1.229 kg / kwh into the following calculation formula, the long-term power generation forecast value A of the target area is obtained:

[0210] A = tce[53x20% / (0.3046x1.299)]≈2.83 billion tce.

[0211] The present disclosure can provide corresponding prediction methods for the short-term, medium-term and long-term needs of different time sizes of the target area to meet the diverse decision-making and planning needs and realize the prediction of multi-dimensional power demand. Among them, the equal-dimensional gray recursive method is introduced to perform short-term power demand prediction based on the data set of historical power generation, without additional auxiliary data, and the short-term power generation prediction value obtained is relatively accurate when the data fluctuation is not large; according to the characteristics of the trend change of the data set of historical power generation, three moving average methods are introduced to perform medium-term power demand prediction based on a preset sliding window, and the influence of the periodic change and irregular change of the data set on the medium-term power demand prediction result is avoided based on the sliding window, thereby improving the accuracy and reliability of the medium-term power generation prediction value, or according to the change trend of the time series of the data set of historical power generation, one of the first exponential average method, the second exponential average method or the third exponential average method is used to perform medium-term power demand prediction, and the weight of each data is determined by the preset attenuation coefficient, so as to ensure that the first prediction model constructed can capture the long-term trend, so as to improve the accuracy and reliability of the prediction result. The introduction of preset energy consumption for long-term power demand prediction can more comprehensively consider the impact of future power demand, thereby improving the accuracy and reliability of the long-term power demand prediction result.

[0212] Based on the same inventive concept, the present disclosure also provides a power demand prediction device, such as Figure 3 As shown, the power demand prediction device includes an acquisition module 301 , a first execution module 302 , a second execution module 303 and a third execution module 304 .

[0213] The acquisition module 301 is configured to acquire power data of the target area, wherein the power data includes historical power generation and total future energy consumption, wherein the total future energy consumption represents a predicted value of energy consumption demand of the target area within a preset time period;

[0214] The first execution module 302 is configured to determine a data set of the historical power generation, perform short-term power demand forecasting based on the data set, and obtain a short-term power generation forecast value of the target area;

[0215] The second execution module 303 is configured to perform a medium-term forecast of power demand according to the data set and a preset sliding window or a preset attenuation coefficient to obtain a medium-term power generation forecast value of the target area;

[0216] The third execution module 304 is configured to perform long-term forecasting of power demand according to the total prospective energy consumption to obtain a long-term power generation forecast value of the target area.

[0217] In the embodiment of the present disclosure, short-term power demand forecasting is performed based on the data set of historical power generation; medium-term power demand forecasting is performed based on the data set of historical power generation and a preset sliding window or a preset attenuation coefficient; and long-term power demand forecasting is performed based on the total energy consumption in the future, which can simultaneously realize short-term, medium-term and long-term power demand forecasting, and realize multi-dimensional power demand forecasting. The medium-term power demand forecasting based on the sliding window can avoid the influence of the periodic changes and irregular changes of the data set on the medium-term power demand forecasting results, thereby improving the accuracy and reliability of the medium-term power generation forecast value.

[0218] Further, the second execution module 303 is configured to control the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, determine the average value of the data of the data set in the preset sliding window before the first sliding, and the average value of the data of the data set in the preset sliding window after each sliding, and obtain a first moving average value corresponding to the data set;

[0219] constructing a first prediction model based on the first moving average value, wherein the first prediction model is used to predict the mid-term power generation of the target area based on a data set corresponding to the historical power generation of the target area;

[0220] A medium-term power generation forecast value of the target area is obtained according to the first forecast model.

[0221] Further, the second execution module 303 is configured to control the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, and determine the weighted average value of the data of the data set in the preset sliding window before the first sliding, and the weighted average value of the data of the data set in the preset sliding window after each sliding, according to the preset weight corresponding to each data in the data set, to obtain the weighted moving average value corresponding to the data set;

[0222] According to the weighted moving average, a second prediction model with a function of correcting hysteresis deviation is constructed, wherein the second prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area;

[0223] A medium-term power generation forecast value of the target area is obtained according to the second forecast model.

[0224] Further, the second execution module 303 is configured to determine a second moving average value corresponding to the data set according to the first moving average value;

[0225] Determine a first smoothing coefficient and a second smoothing coefficient according to the second moving average, wherein the first smoothing coefficient represents an intercept corresponding to a change trend of the data set, and the second smoothing coefficient represents a slope corresponding to a change trend of the data set;

[0226] A first prediction model is constructed according to the first smoothing coefficient and the second smoothing coefficient.

[0227] Further, the second execution module 303 is configured to control the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, and determine the weighted average value of the data of the data set in the preset sliding window before the first sliding, and the weighted average value of the data of the data set in the preset sliding window after each sliding, according to the preset weight corresponding to each data in the data set, to obtain the weighted moving average value corresponding to the data set;

[0228] According to the weighted moving average, a second prediction model with a function of correcting hysteresis deviation is constructed, wherein the second prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area;

[0229] A medium-term power generation forecast value of the target area is obtained according to the second forecast model.

[0230] Furthermore, the second execution module 303 is configured to substitute the weighted moving average into the following calculation formula to obtain the second prediction model with the function of correcting the lag deviation:

[0231]

[0232] in, Represents the predicted value of the (T+1)th period, M tw Characterizes the weighted moving average.

[0233] Further, the second execution module 303 is configured to determine a first prediction value and a second prediction value corresponding to the data set according to the data set, the preset attenuation coefficient and the average value;

[0234] Determine a first horizontal coefficient and a first trend coefficient corresponding to the data set according to the first predicted value, the second predicted value, and the preset attenuation coefficient, wherein the first horizontal coefficient is used to smooth first data in the data set that does not change with time, and the first trend coefficient is used to smooth second data in the data set that changes with time;

[0235] The third prediction model with the function of correcting lag deviation is constructed according to the horizontal coefficient and the trend coefficient.

[0236] Further, the second execution module 303 is configured to determine a third prediction value, a fourth prediction value and a fifth prediction value corresponding to the data set according to the data set, the preset attenuation coefficient and the average value;

[0237] Determine, according to the third predicted value, the fourth predicted value, the fifth predicted value and the preset attenuation coefficient, a second horizontal coefficient, a second trend coefficient and a seasonal coefficient corresponding to the data set, wherein the second horizontal coefficient is used to smooth first data in the data set that does not change with time, the second trend coefficient is used to smooth second data in the data set that changes with time, and the seasonal coefficient is used to smooth periodic fluctuations of the data set;

[0238] The third prediction model with the function of correcting the lag deviation is constructed according to the second horizontal coefficient, the second trend coefficient and the seasonal coefficient.

[0239] Further, the first execution module 302 is configured to perform a level ratio check process on the data set and then perform an accumulation process on the data set to obtain an intermediate data set, wherein the level ratio check process is used to adjust the level ratios corresponding to all adjacent data in the data set to be within a preset interval, and the level ratio represents the ratio between adjacent data;

[0240] Calculate the average value of all adjacent data in the intermediate data set to obtain a mean data set;

[0241] Constructing a third prediction model with a hysteresis deviation correction function based on the data set and the mean data set, wherein the third prediction model is used to predict the short-term power generation of the target area based on the data set corresponding to the historical power generation of the target area;

[0242] A short-term power generation forecast value of the target area is obtained according to the third forecast model.

[0243] Furthermore, the third execution module 304 is configured to obtain the long-term power generation forecast value of the target area by dividing the product of the total future energy consumption and the preset conversion value by a preset consumption coefficient, wherein the preset consumption coefficient is the product of the coal consumption coefficient and the electric heat equivalent coefficient, and the electric heat equivalent coefficient represents the uniform conversion of the consumption of various energy sources in the power generation process into the consumption of standard coal.

[0244] Regarding the power demand prediction device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0245] Based on the same inventive concept, the present disclosure also provides an electronic device, including:

[0246] a memory having a computer program stored thereon;

[0247] A processor is used to execute the computer program in the memory to implement the above-mentioned power demand forecasting method.

[0248] In the embodiment of the present disclosure, short-term power demand forecasting is performed based on the data set of historical power generation; medium-term power demand forecasting is performed based on the data set of historical power generation and a preset sliding window or a preset attenuation coefficient; and long-term power demand forecasting is performed based on the total energy consumption in the future, which can simultaneously realize short-term, medium-term and long-term power demand forecasting, and realize multi-dimensional power demand forecasting. The medium-term power demand forecasting based on the sliding window can avoid the influence of the periodic changes and irregular changes of the data set on the medium-term power demand forecasting results, thereby improving the accuracy and reliability of the medium-term power generation forecast value.

[0249] Figure 4 FIG. 4 is a block diagram of an electronic device 400 according to an exemplary embodiment. Figure 4 As shown, the electronic device 400 may include: a processor 401 and a memory 402. The electronic device 400 may also include one or more of a multimedia component 403, an input / output (I / O) interface 404, and a communication component 405.

[0250] The processor 401 is used to control the overall operation of the electronic device 400 to complete all or part of the steps in the above-mentioned power demand prediction method. The memory 402 is used to store various types of data to support the operation of the electronic device 400, and these data may include, for example, instructions for any application or method used to operate on the electronic device 400, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 402 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Read-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 403 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 402 or sent through the communication component 405. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 404 provides an interface between the processor 401 and other interface modules, and the other interface modules may be keyboards, mice, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 405 is used for wired or wireless communication between the electronic device 400 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or a combination of one or more of them, is not limited here. Therefore, the corresponding communication component 405 may include: Wi-Fi module, Bluetooth module, NFC module, etc.

[0251] In an exemplary embodiment, the electronic device 400 can be implemented by one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned power demand forecasting method.

[0252] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned power demand forecasting method are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 402 including program instructions, and the above-mentioned program instructions can be executed by the processor 401 of the electronic device 400 to complete the above-mentioned power demand forecasting method.

[0253] In another exemplary embodiment, a computer program product is further provided. The computer program product includes a computer program executable by a programmable device. The computer program has a code portion for executing the above-mentioned power demand prediction method when executed by the programmable device.

[0254] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings; however, the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, a variety of simple modifications can be made to the technical solution of the present disclosure, and these simple modifications all fall within the protection scope of the present disclosure.

[0255] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0256] In addition, various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

Claims

1. A multidimensional power demand forecasting method, characterized in that: include: Acquire power data of the target area, wherein the power data includes historical power generation and total future energy consumption, wherein the total future energy consumption represents a predicted value of energy consumption demand of the target area within a preset time period; Determine a data set of the historical power generation, perform short-term power demand forecasting based on the data set, and obtain a short-term power generation forecast value of the target area, wherein the data set is obtained by arranging the historical power generation according to time; Performing a medium-term forecast of power demand based on the data set and a preset sliding window or a preset attenuation coefficient to obtain a medium-term power generation forecast value for the target area; A long-term forecast of electricity demand is performed based on the total prospective energy consumption to obtain a long-term power generation forecast value for the target area.

2. The multidimensional power demand forecasting method according to claim 1, characterized in that: The step of performing a medium-term forecast of power demand according to the data set and the preset sliding window to obtain a medium-term power generation forecast value of the target area includes: Controlling the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, determining an average value of the data of the data set in the preset sliding window before the first sliding, and an average value of the data of the data set in the preset sliding window after each sliding, to obtain a first moving average value corresponding to the data set; constructing a first prediction model with a hysteresis deviation correction function according to the first moving average value, wherein the first prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area; A medium-term power generation forecast value of the target area is obtained according to the first forecast model.

3. The multidimensional power demand forecasting method according to claim 1, characterized in that: The step of performing a medium-term forecast of power demand according to the data set and the preset sliding window to obtain a medium-term power generation forecast value of the target area includes: Controlling the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, and determining, according to a preset weight corresponding to each data in the data set, a weighted average value of the data of the data set in the preset sliding window before the first sliding, and a weighted average value of the data of the data set in the preset sliding window after each sliding, to obtain a weighted moving average value corresponding to the data set; According to the weighted moving average, a second prediction model with a function of correcting hysteresis deviation is constructed, wherein the second prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area; A medium-term power generation forecast value of the target area is obtained according to the second forecast model.

4. The multidimensional power demand forecasting method according to claim 2, characterized in that: The step of constructing a first prediction model with a function of correcting hysteresis deviation according to the first moving average value comprises: Determining a second moving average value corresponding to the data set according to the first moving average value; Determine a first smoothing coefficient and a second smoothing coefficient according to the second moving average, wherein the first smoothing coefficient represents an intercept corresponding to a change trend of the data set, and the second smoothing coefficient represents a slope corresponding to a change trend of the data set; A first prediction model having a function of correcting hysteresis deviation is constructed according to the first smoothing coefficient and the second smoothing coefficient.

5. The multi-dimensional power demand forecasting method according to claim 3, characterized in that: The second prediction model having the function of correcting the lag deviation is constructed according to the weighted moving average, comprising: Substituting the weighted moving average into the following calculation formula, the second prediction model with the function of correcting the lag deviation is obtained: in, Represents the predicted value of the (T+1)th period, M tw Represents the weighted moving average of period T.

6. The multi-dimensional power demand forecasting method according to claim 1, characterized in that: The step of performing a medium-term forecast of power demand according to the data set and the preset attenuation coefficient to obtain a medium-term power generation forecast value of the target area includes: Controlling the sliding window to slide according to a preset step length until the sliding window covers all the data in the data set, and determining an average value of the data in the data set in the sliding window after the last sliding; According to the data set, the preset attenuation coefficient and the average value, a third prediction model with a hysteresis deviation correction function is constructed, wherein the third prediction model is used to predict the mid-term power generation of the target area according to the data set corresponding to the historical power generation of the target area; A medium-term power generation forecast value of the target area is obtained according to the third forecast model.

7. The multi-dimensional power demand forecasting method according to claim 6, characterized in that: The constructing of a third prediction model with a function of correcting hysteresis deviation according to the data set, the preset attenuation coefficient and the average value comprises: Determine a first predicted value and a second predicted value corresponding to the data set according to the data set, the preset attenuation coefficient and the average value; Determine a first horizontal coefficient and a first trend coefficient corresponding to the data set according to the first predicted value, the second predicted value, and the preset attenuation coefficient, wherein the first horizontal coefficient is used to smooth first data in the data set that does not change with time, and the first trend coefficient is used to smooth second data in the data set that changes with time; The third prediction model with the function of correcting the lag deviation is constructed according to the horizontal coefficient and the trend coefficient.

8. The multi-dimensional power demand forecasting method according to claim 6, characterized in that: The constructing of a third prediction model with a function of correcting hysteresis deviation according to the data set, the preset attenuation coefficient and the average value comprises: Determine, according to the data set, the preset attenuation coefficient and the average value, a third predicted value, a fourth predicted value and a fifth predicted value corresponding to the data set; Determine, according to the third predicted value, the fourth predicted value, the fifth predicted value and the preset attenuation coefficient, a second horizontal coefficient, a second trend coefficient and a seasonal coefficient corresponding to the data set, wherein the second horizontal coefficient is used to smooth first data in the data set that does not change with time, the second trend coefficient is used to smooth second data in the data set that changes with time, and the seasonal coefficient is used to smooth periodic fluctuations of the data set; The third prediction model with the function of correcting the lag deviation is constructed according to the second horizontal coefficient, the second trend coefficient and the seasonal coefficient.

9. The multidimensional power demand forecasting method according to any one of claims 1 to 8, characterized in that: The performing short-term forecasting of power demand according to the data set to obtain a short-term power generation forecast value of the target area includes: After performing a level ratio inspection process on the data set, performing an accumulation process on the data set to obtain an intermediate data set, wherein the level ratio inspection process is used to adjust the level ratios corresponding to all adjacent data in the data set to be within a preset interval, and the level ratio represents the ratio between adjacent data; Calculate the average value of all adjacent data in the intermediate data set to obtain a mean data set; constructing a fourth prediction model based on the data set and the mean data set, wherein the fourth prediction model is used to predict short-term power generation of the target area based on the data set corresponding to the historical power generation of the target area; A short-term power generation forecast value of the target area is obtained according to the fourth forecast model.

10. The multi-dimensional power demand forecasting method according to any one of claims 1 to 8, characterized in that: The long-term forecast of power demand is performed according to the total prospective energy consumption to obtain the long-term power generation forecast value of the target area, including: The product of the total future energy consumption and the preset conversion value is divided by the preset consumption coefficient to obtain the long-term power generation forecast value of the target area. The preset consumption coefficient is the product of the coal consumption coefficient and the electric heat equivalent coefficient. The electric heat equivalent coefficient represents the uniform conversion of the consumption of various energy sources in the power generation process into the consumption of standard coal.

11. A multi-dimensional power demand forecasting device, characterized in that: include: An acquisition module is configured to acquire power data of a target area, wherein the power data includes historical power generation and total future energy consumption, wherein the total future energy consumption represents a predicted value of energy consumption demand of the target area within a preset time period; A first execution module is configured to determine a data set of the historical power generation, perform short-term power demand forecasting based on the data set, and obtain a short-term power generation forecast value of the target area; A second execution module is configured to perform a medium-term forecast of power demand based on the data set and a preset sliding window or a preset attenuation coefficient to obtain a medium-term power generation forecast value of the target area; The third execution module is configured to perform long-term forecasting of electricity demand according to the total prospective energy consumption to obtain a long-term power generation forecast value of the target area.

12. The multi-dimensional power demand forecasting device according to claim 10, characterized in that: The second execution module is configured to control the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, determine the average value of the data of the data set in the preset sliding window before the first sliding, and the average value of the data of the data set in the preset sliding window after each sliding, and obtain a first moving average value corresponding to the data set; constructing a first prediction model with a hysteresis deviation correction function according to the first moving average value, wherein the first prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area; A medium-term power generation forecast value of the target area is obtained according to the first forecast model.

13. The multi-dimensional power demand forecasting device according to claim 11, characterized in that: The second execution module is configured to control the preset sliding window to slide according to a preset step length until the preset sliding window covers all the data in the data set, and determine the weighted average value of the data of the data set in the preset sliding window before the first sliding, and the weighted average value of the data of the data set in the preset sliding window after each sliding according to the preset weight corresponding to each data in the data set, to obtain a weighted moving average value corresponding to the data set; According to the weighted moving average, a second prediction model with a function of correcting hysteresis deviation is constructed, wherein the second prediction model is used to predict the mid-term power generation of the target area according to a data set corresponding to the historical power generation of the target area; A medium-term power generation forecast value of the target area is obtained according to the second forecast model.

14. The multi-dimensional power demand forecasting device according to claim 12, characterized in that: The second execution module is configured to determine a second moving average value corresponding to the data set according to the first moving average value; Determine a first smoothing coefficient and a second smoothing coefficient according to the second moving average, wherein the first smoothing coefficient represents an intercept corresponding to a change trend of the data set, and the second smoothing coefficient represents a slope corresponding to a change trend of the data set; A first prediction model having a function of correcting hysteresis deviation is constructed according to the first smoothing coefficient and the second smoothing coefficient.

15. The multi-dimensional power demand forecasting device according to claim 13, characterized in that: The second execution module is configured to substitute the weighted moving average into the following calculation formula to obtain the second prediction model with the function of correcting the lag deviation: in, Represents the predicted value of the (T+1)th period, M tw Characterizes the weighted moving average.

16. The multi-dimensional power demand forecasting device according to claim 11, characterized in that: The second execution module is configured to control the sliding window to slide according to a preset step length until the sliding window covers all the data in the data set, and determine the average value of the data of the data set in the sliding window after the last sliding; According to the data set, the preset attenuation coefficient and the average value, a third prediction model with a hysteresis deviation correction function is constructed, wherein the third prediction model is used to predict the mid-term power generation of the target area according to the data set corresponding to the historical power generation of the target area; A medium-term power generation forecast value of the target area is obtained according to the third forecast model.

17. The multi-dimensional power demand forecasting device according to claim 16, characterized in that: The second execution module is configured to determine a first prediction value and a second prediction value corresponding to the data set according to the data set, the preset attenuation coefficient and the average value; Determine a first horizontal coefficient and a first trend coefficient corresponding to the data set according to the first predicted value, the second predicted value, and the preset attenuation coefficient, wherein the first horizontal coefficient is used to smooth first data in the data set that does not change with time, and the first trend coefficient is used to smooth second data in the data set that changes with time; The third prediction model with the function of correcting the lag deviation is constructed according to the horizontal coefficient and the trend coefficient.

18. The multi-dimensional power demand forecasting device according to claim 16, characterized in that: The second execution module is configured to determine a third prediction value, a fourth prediction value, and a fifth prediction value corresponding to the data set according to the data set, the preset attenuation coefficient, and the average value; Determine, according to the third predicted value, the fourth predicted value, the fifth predicted value and the preset attenuation coefficient, a second horizontal coefficient, a second trend coefficient and a seasonal coefficient corresponding to the data set, wherein the second horizontal coefficient is used to smooth first data in the data set that does not change with time, the second trend coefficient is used to smooth second data in the data set that changes with time, and the seasonal coefficient is used to smooth periodic fluctuations of the data set; The third prediction model with the function of correcting the lag deviation is constructed according to the second horizontal coefficient, the second trend coefficient and the seasonal coefficient.

19. The multi-dimensional power demand forecasting device according to any one of claims 11 to 18, characterized in that: The first execution module is configured to perform a level ratio check process on the data set and then perform an accumulation process on the data set to obtain an intermediate data set, wherein the level ratio check process is used to adjust the level ratios corresponding to all adjacent data in the data set to be within a preset interval, and the level ratio represents a ratio between adjacent data; Calculate the average value of all adjacent data in the intermediate data set to obtain a mean data set; constructing a fourth prediction model based on the data set and the mean data set, wherein the fourth prediction model is used to predict short-term power generation of the target area based on the data set corresponding to the historical power generation of the target area; A short-term power generation forecast value of the target area is obtained according to the fourth forecast model.

20. The multi-dimensional power demand forecasting device according to any one of claims 11 to 18, characterized in that: The third execution module is configured to obtain the long-term power generation forecast value of the target area by dividing the product of the total future energy consumption and the preset conversion value by the preset consumption coefficient. The preset consumption coefficient is the product of the coal consumption coefficient and the electric heat equivalent coefficient. The electric heat equivalent coefficient represents the uniform conversion of the consumption of various energy sources in the power generation process into the consumption of standard coal.

21. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

22. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the method according to any one of claims 1 to 10.

23. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

Images