Monthly electricity consumption prediction method considering newly installed capacity increasing and return and supplement factors

Abstract

The invention relates to a monthly electricity consumption prediction method considering new installation capacity increasing and return and supplement factors, which adopts a monthly electricity quantity report statistical value after purchase and sale as sample initial data to avoid the influence of inconsistent statistical periods on the electricity consumption prediction accuracy, and meanwhile, the data is data in recent two years, can better reflect the electricity consumption habits and rules of users, and can improve the electricity consumption prediction accuracy. The prediction stability is higher; according to the method, the initial value of the electric quantity in the corresponding month is corrected according to the non-government refund and supplement details of the user and the actual error month of the refund and supplement electric quantity, and the accuracy of electric quantity prediction is ensured from the data source; meanwhile, the corrected monthly electric quantity data is decomposed into stock user electric quantity and increment user electric quantity, the stock user electric quantity is predicted on the basis of analyzing the electric quantity change trend, and the influence of air temperature and holidays on the electric quantity is considered on the basis of the seasonal and continuous electric quantity change; and the electric quantity is predicted by combining the user research condition and the electricity consumption statistical condition of similar users, so that the prediction accuracy and stability are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of electricity consumption prediction, and in particular to a monthly electricity consumption prediction method taking into account new installation capacity increase and subsidy withdrawal factors. Background Art

[0002] Medium- and long-term electricity forecasting can determine the influencing factors and development trends of electricity demand in the power market, and provide technical support for power grid companies to conduct business analysis, formulate marketing service strategies, promote accurate electricity substitution, and predict the scale of electricity demand of commercial and industrial users who purchase electricity on behalf of them. Therefore, it has always been the focus of power marketing work. At present, most medium- and long-term electricity forecasts directly use monthly electricity data from several years of history as analysis samples, and use mathematical models and methods such as regression analysis, gray theory, and fuzzy clustering to analyze and find out the law of electricity changes, and then derive electricity consumption data for the next few months. In actual electricity forecasting work, it is found that such methods have the following defects:

[0003] 1) The sample data has not been preprocessed. Due to factors such as non-policy refunds, metering device failures, abnormal user files, and user electricity theft, the reported statistical values ​​of monthly electricity consumption for each electricity consumption category are not equivalent to the actual electricity consumption of users in historical months. Therefore, they cannot be directly used as the basic data for electricity consumption analysis and prediction. The initial data needs to be corrected and restored to the actual monthly electricity consumption data.

[0004] 2) The impact of the settlement cycle on electricity statistics is not considered. In the months when the purchase and sale synchronization is not implemented, the meter reading days of users in the power supply area are scattered at the beginning, middle and end of the month. The electricity issued by users is not the electricity consumption of the natural month, which is inconsistent with the statistical cycle of electricity consumption after the purchase and sale synchronization. It is difficult to restore it to the electricity consumption of the natural month and the accuracy is difficult to guarantee. If only the data after the purchase and sale synchronization is used, the data volume is small and it is difficult to support most prediction models and algorithms.

[0005] 3) The impact of newly installed and increased capacity users on electricity consumption is not considered. The electricity released by the newly installed equipment of the newly installed and increased capacity users is a pure increase in electricity consumption, which is bound to cause an increase in electricity consumption. The electricity consumption trends of new users and existing users are not completely consistent. Only by dividing the total electricity consumption into the electricity consumption of existing users and the electricity consumption of incremental users and analyzing them separately can we find out the specific reasons and changing rules for the changes in electricity consumption.

[0006] 4) The pattern of electricity consumption by users has changed fundamentally. With the continuous construction of new power systems and the in-depth advancement of the "dual carbon" goals, the rapid development and penetration of multiple loads such as distributed power sources, electric vehicles, user-side energy storage and microgrids have led to the transformation of loads from traditional rigid, pure consumption to flexible, production and consumption. This has greatly increased the complexity and difficulty of prediction, and traditional analysis methods and models have failed. Summary of the invention

[0007] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a monthly electricity consumption forecasting method that takes into account new installations, capacity increases and subsidy refund factors. The method is suitable for medium- and long-term electricity consumption forecasting for market-oriented users, proxy electricity purchasing users and non-market-oriented users, and is also suitable for electricity consumption forecasting for segmented electricity consumption categories and segmented industry categories, without being restricted by forecasting scenarios.

[0008] The present invention solves the technical problem by adopting the following technical solutions:

[0009] A method for predicting monthly power consumption taking into account new installation capacity increase and subsidy withdrawal factors comprises the following steps:

[0010] Step 1: Collect the monthly electricity consumption report statistics E from month n of year y-2 to month n of year y 0 And the details of non-policy refunds;

[0011] Step 2: According to the non-policy refund details, the monthly electricity consumption data is corrected according to the actual month of the electricity error to obtain the actual monthly electricity consumption E;

[0012] Step 3: Subdivide user X into an existing user set and a new user set according to the account establishment date;

[0013] Step 4: Split the actual monthly electricity consumption E by user according to the existing user set and the newly added user set;

[0014] Step 5: Predict the monthly electricity consumption of existing users;

[0015] Step 6: Predict the monthly electricity consumption of new users;

[0016] Step 7: Calculate the predicted monthly electricity consumption value based on the monthly electricity consumption of existing users predicted in step 5 and the monthly electricity consumption of new users predicted in step 6.

[0017] Moreover, the specific implementation method of step 3 is: according to the natural year, the total number of users X in the region is subdivided into an existing user set and a new user set. The existing user set is the users who were registered before y-2 years and have not been expanded. y-2 Indicates that the newly added user set is further subdivided into the user set X′ that was newly installed or increased in y-2 years y-2 , the user collection for new installation or capacity increase in year y-1 is X′ y-1 Indicates that the user set for new installation or capacity increase in year y is X′ y express.

[0018] Moreover, the specific implementation method of step 4 is: decompose the actual monthly electricity consumption E into the electricity consumption of existing users E(X) and the electricity consumption contributed by new users E(X′), where the electricity consumption from y-2 years to month i of year y is E (y-2)*i 、E (y-1)*i 、E y*i They are divided into:

[0019] E (y-2)*i =E (y-2)*i (X y-2 )+E (y-2)*i (X′ y-2 )

[0020] E (y-1)*i =E (y-1)*i (X y-2 )+E (y-1)*i (X′ y-2 )+E (y-1)*i (X′ y-1 )

[0021] E y*i =E y*i (X y-2 )+E y*i (X′ y-2 )+E y*i (X′ y-1 )+E y*i (X′ y )

[0022] Among them, i=1, 2,…, 12.

[0023] Moreover, the specific implementation method of step 5 is: predict the monthly electricity consumption E of the existing users in year y and month n+1 y*(n+1) (X y-2 ):

[0024]

[0025] Among them, ξ and γ are proportionality coefficients.

[0026] Moreover, the specific implementation method of step 6 is: predict the monthly electricity consumption E of the new user in year y-2 in month y+1 y*(n+1) (X′ y-2 ):

[0027]

[0028] Predict the monthly electricity consumption E of new users in year y-1 in month y+1 y*(n+1) (X′ y-1 ):

[0029] E y*(n+1) (X′ y-1 )=∑κ·S·T·D

[0030] Among them, S represents the user's operating capacity; κ represents the user's load rate; T represents the average daily electricity consumption hours; and D represents the number of days of electricity consumption.

[0031] Moreover, the specific implementation method of step 7 is: calculate the monthly electricity consumption forecast value E of month n+1 in year y y*(n+1) :

[0032] E y*(n+1) =E y*(n+1) (X y-2 )+E y*(n+1) (X′ y-2 )+E y*(n+1) (X′ y-1 )+E y*(n+1) (X′ y ).

[0033] The advantages and positive effects of the present invention are:

[0034] 1. The present invention adopts the statistical value of the monthly electricity consumption report after the purchase and sale are implemented as the initial sample data to avoid the influence of inconsistent statistical cycles on the accuracy of electricity prediction. At the same time, the data is data for the past two years, which can better reflect the user's electricity consumption habits and rules, and the prediction stability is higher; the method corrects the initial electricity consumption value of the corresponding month according to the actual error month of the refund and subsidy electricity according to the user's non-policy refund and subsidy details, so as to ensure the accuracy of electricity prediction from the data source; the present invention decomposes the corrected monthly electricity consumption data into the electricity consumption of existing users and the electricity consumption of incremental users, and predicts the electricity consumption of existing users on the basis of analyzing the electricity consumption change trends respectively, based on the seasonality and continuity of electricity consumption changes, considering the influence of temperature and holidays on electricity consumption, and proposes a month-on-month correction coefficient iterative algorithm, and predicts the electricity consumption of new users in combination with user surveys and electricity consumption statistics of similar users, so as to improve the prediction accuracy and stability.

[0035] 2. The present invention takes into account the impact of the settlement cycle on electricity statistics, and uses data from the same statistical cycle to make the prediction result more accurate. At the same time, the present invention uses refund and subsidy data to correct historical electricity report data, and the sample data is more accurate. The present invention divides the monthly electricity consumption into the electricity consumption of existing users and the electricity consumption of incremental users, and analyzes and predicts them separately to find out the specific reasons for the increase and decrease in electricity consumption, so that the prediction result is more in line with reality. The present invention proposes a month-on-month correction coefficient iterative algorithm to predict the electricity consumption of existing users, which requires a small number of samples, and the prediction result is highly accurate and stable. The present invention predicts the electricity consumption of new users based on user surveys and the electricity consumption of similar users, and the prediction effect is good. The present invention is suitable for medium- and long-term electricity consumption forecasts for market-oriented users, agency electricity purchasing users and non-market-oriented users, and is also suitable for electricity forecasts for subdivided electricity consumption categories and subdivided industry categories, without being restricted by prediction scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0037] The present invention is further described in detail below with reference to the accompanying drawings.

[0038] A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors, such as Figure 1 As shown, the following steps are included:

[0039] Step 1: Collect the monthly electricity consumption report statistics E from month n of year y-2 to month n of year y 0 And details of non-policy refunds.

[0040] Step 2: According to the non-policy refund details and the actual month of the electricity consumption error, correct the electricity consumption data for the current month to obtain the actual monthly electricity consumption E.

[0041] Step 3: User X is subdivided into an existing user set and a new user set based on the account establishment date.

[0042] According to the natural year, the total number of users X in the region is subdivided into the existing user set and the new user set. The existing user set is the users who were registered before y-2 years and have not been expanded. y-2 Indicates that the newly added user set is further subdivided into the user set X′ that was newly installed or increased in y-2 years y-2 , the user collection for new installation or capacity increase in year y-1 is X′ y-1 Indicates that the user set for new installation or capacity increase in year y is X′ y express.

[0043] Step 4: Split the actual monthly electricity consumption E by user based on the existing user set and the newly added user set.

[0044] The specific implementation method of step 4 is: decompose the actual monthly electricity consumption E into the electricity consumption of existing users E(X) and the electricity consumption contributed by new users E(X′), where the electricity consumption from y-2 years to month i of year y is E (y-2)*i 、E (y-1)*i 、E y*i They are divided into:

[0045] E (y-2)*i =E (y-2)*i (X y-2 )+E (y-2)*i (X′ y-2 ⑴

[0046] E (y-1)*i =E (y-1)*i (X y-2 )+E (y-1)*i (X′ y-2 )+E (y-1)*i (X′ y-1 ⑵

[0047] E y*i =E y*i (X y-2)+E y*i (X′ y-2 )+E y*i (X′ y-1 )+E y*i (X′ y )⑶

[0048] Among them, i=1, 2,…, 12.

[0049] Step 5: Predict the monthly electricity consumption of existing users.

[0050] Predict the monthly electricity consumption E of existing users in year y and month n+1 y*(n+1) (X y-2 ):

[0051]

[0052] Among them, ξ and γ are proportional coefficients, and the calculation method of ξ and γ is:

[0053] Collect the actual daily average temperature values ​​t of year y-2, month n and month n+1 of year y-1, and month n of year y, obtain the predicted daily average temperature value t′ of month n+1 of year y, and count the number of holidays H from year y-2 to month n and month n+1 of year y.

[0054]

[0055]

[0056]

[0057]

[0058]

[0059] Among them, R 1 、M 1 , Q 1 Represents the number of days from y-2 to y year n+1 month, R 2 、M 2 , Q 2 They represent the number of days from year y-2 to month n of year y. The value of month n varies in different regions.

[0060] Step 6: Predict the monthly electricity consumption of new users.

[0061] Predict the monthly electricity consumption E of new users in year y-2 in month y+1 y*(n+1) (X′ y-2 ):

[0062]

[0063] Predict the monthly electricity consumption E of new users in year y-1 in month y+1 y*(n+1) (X′ y-1 ):

[0064] E y*(n+1) (X′ y-1 )=∑κ·S·T·D ⑾

[0065] Among them, S represents the user's operating capacity; κ represents the user's load rate; T represents the average daily electricity consumption hours; and D represents the number of days of electricity consumption. When calculating specifically, for high-voltage users with an operating capacity greater than 1000kVA, it is recommended to determine the values ​​of κ, T, and D on a household basis based on the results of user surveys; for the remaining new users, it is recommended to divide the users into four categories for analysis: low-voltage residents, low-voltage non-residents-charging piles, low-voltage non-residents-others, and high voltage. The values ​​may vary in different regions with different load compositions and load characteristics.

[0066] Table 1 Table of values ​​of parameters for each user classification

[0067]

[0068]

[0069] Refer to the monthly electricity consumption E of month n+1 in year y-1 y*(n+1) (X′ y-1 ) is used to calculate the monthly electricity consumption E of new users in year y in month y+1 y*(n+1) (X′ y ).

[0070] E y*(n+1) (X′ y )=∑κ·S·T·D ⑿

[0071] Step 7: Calculate the predicted monthly electricity consumption value based on the monthly electricity consumption of existing users predicted in step 5 and the monthly electricity consumption of new users predicted in step 6.

[0072] Calculate the predicted monthly electricity consumption E for month n+1 in year y y*(n+1) :

[0073] E y*(n+1) =E y*(n+1) (X y-2 )+E y*(n+1) (X′ y-2 )+E y*(n+1) (X′ y-1 )+E y*(n+1) (X′ y ).

[0074] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation manner. Any other implementation manners derived by those skilled in the art based on the technical solution of the present invention also fall within the scope of protection of the present invention.

Claims

1. A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors, characterized by: The following steps are involved: Step 1: Collect the monthly electricity report statistics E0 and non-policy refund details from month n of year y-2 to month n of year y; Step 2: According to the non-policy refund details, the monthly electricity consumption data is corrected according to the actual month of the electricity consumption error to obtain the actual monthly electricity consumption E; Step 3: Subdivide user X into an existing user set and a new user set according to the account establishment date; Step 4: Split the actual monthly electricity consumption E by user according to the existing user set and the newly added user set; Step 5: Predict the monthly electricity consumption of existing users; Step 6: Predict the monthly electricity consumption of new users; Step 7: Calculate the predicted monthly electricity consumption value based on the monthly electricity consumption of existing users predicted in step 5 and the monthly electricity consumption of new users predicted in step 6.

2. A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors according to claim 1, characterized in that: The specific implementation method of step 3 is: according to the natural year, the total number of users X in the region is subdivided into an existing user set and a new user set. The existing user set is the users who were registered before y-2 years and have not been expanded. y-2 Indicates that the newly added user set is further subdivided into the user set X′ that was newly installed or increased in y-2 years y-2 , the user collection for new installation or capacity increase in year y-1 is X′ y-1 Indicates that the user set for new installation or capacity increase in year y is X′ y express.

3. A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors according to claim 2, characterized in that: The specific implementation method of step 4 is: decompose the actual monthly electricity consumption E into the electricity consumption of existing users E(X) and the electricity consumption contributed by new users E(X′), where the electricity consumption from y-2 years to month i of year y is E (y-2)*i 、E (y-1)*i 、E y*i They are divided into: AND (y-2)*i =And (y-2)*i (X y-2 )+E (y-2)*i (X′ y-2 ) AND (y-1)*i =And (y-1)*i (X y-2 )+E (y-1)*i (X′ y-2 )+E (y-1)*i (X′ y-1 ) AND y*i =And y*i (X y-2 )+E y*i (X′ y-2 )+E y*i (X′ y-1 )+E y*i (X′ y ) Among them, i=1, 2,…, 12.

4. A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors according to claim 3, characterized in that: The specific implementation method of step 5 is: predict the monthly electricity consumption E of the existing users in year y and month n+1 y*(n+1) (X y-2 ): Among them, ξ and γ are proportionality coefficients.

5. A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors according to claim 4, characterized in that: The specific implementation method of step 6 is: predict the monthly electricity consumption E of the new user in year y-2 in month y+1 y*(n+1) (X′ y-2 ): Predict the monthly electricity consumption E of new users in year y-1 in month y+1 y*(n+1) (X′ y-1 ): E y*(n+1) (X y ′ -1 )=∑κ·S·T·D Among them, S represents the user's operating capacity; κ represents the user's load rate; T represents the average daily electricity consumption hours; and D represents the number of days of electricity consumption.

6. A monthly electricity consumption prediction method considering new installation capacity increase and subsidy withdrawal factors according to claim 5, characterized in that: The specific implementation method of step 7 is: calculate the monthly electricity consumption forecast value E of month n+1 in year y y*(n+1) : AND y*(n+1) =And y*(n+1) (X y-2 )+E y*(n+1) (X′ y-2 )+E y*(n+1) (X′ y-1 )+E y*(n+1) (X′ y )。

Images