Maximum demand prediction method, device, electronic device and storage medium
Through the combination of daily average demand and maximum demand residual prediction models, the prediction accuracy of maximum demand is improved, and the problem of low prediction accuracy of maximum demand in the existing technology is solved, helping enterprises better choose electricity price types , reduce electricity bills.
Patent Information
- Application Number
- CN201910882806.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2019-09-18
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2039-09-18
AI Technical Summary
The accuracy of the maximum demand forecast in the prior art is low, which affects the accuracy of enterprises' choice of basic electricity prices and leads to uncertain electricity bill expenditures.
The pre-trained daily average demand prediction model and maximum demand residual prediction model are used to make biased predictions through the data reported on the smart meter to improve the accuracy of the maximum demand and avoid the risk of deviation.
It improves the accuracy of the forecast of maximum demand, helps enterprises more accurately choose capacity electricity prices or demand electricity prices, and reduces electricity bills.
Smart Images

Figure CN110633860B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer technology, and in particular to a maximum demand prediction method, device, electronic device and storage medium in the field of Internet of Things technology. Background Art
[0002] The two-part electricity price includes the basic electricity price and the kilowatt-hour electricity price. The basic electricity price is calculated based on the enterprise's transformer capacity (kVA) or the user's maximum demand (kW), commonly known as the capacity fee or demand fee. The basic electricity price is payable regardless of whether the enterprise uses electricity or not, and has nothing to do with the actual electricity consumption of the enterprise. At present, enterprises can choose whether to use the capacity electricity price or the demand electricity price for the basic electricity price every three months. Among them, when the actual maximum demand of the enterprise is lower than the maximum demand corresponding to the balance of capacity and demand (the capacity fee is equal to the demand fee), the enterprise's choice of the demand electricity price has certain cost savings compared to the capacity electricity price; when the actual maximum demand of the enterprise is higher than the maximum demand corresponding to the balance of capacity and demand (the capacity fee is equal to the demand fee), the enterprise's choice of the demand electricity price needs to pay more fees than the capacity electricity price. Therefore, predicting the maximum demand in the next three months is crucial for enterprises to choose capacity electricity price or demand electricity price.
[0003] In the prior art, the maximum demand in the next three months is usually predicted by multiplying the maximum demand in the past three months by a fixed coefficient to select the basic electricity price. Since the fixed coefficient is set by human experience, the accuracy of the maximum demand predicted by this method is low, which in turn affects the accuracy of the basic electricity price selected by the enterprise. Summary of the invention
[0004] The present application provides a maximum demand prediction method, device, electronic device and storage medium, which can improve the prediction accuracy of the maximum demand.
[0005] The first aspect of the present application provides a maximum demand prediction method, comprising:
[0006] Receive the maximum demand of the current time period and multiple demands of each day in the current time period reported by the smart meter; obtain the predicted daily average demand of each day in the next time period according to the multiple demands of each day in the current time period and the daily average demand prediction model, the daily average demand prediction model indicates the corresponding relationship between the daily average demand of each day and the predicted daily average demand, and the daily average demand of each day is the average of the multiple demands of each day in the current time period; obtain the predicted maximum demand residual of each day in the next time period according to the maximum demand, the daily average demand of each day in the current time period and the maximum demand residual prediction model, the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand of each day in the current time period and the maximum demand and the predicted maximum demand residual; obtain and output the predicted maximum demand of the next time period according to the predicted daily average demand of each day in the next time period and the predicted maximum demand residual of each day in the next time period.
[0007] In this embodiment, a biased prediction of the maximum demand is performed through the pre-trained daily average demand prediction model and the maximum demand residual prediction model, thereby improving the accuracy of the obtained maximum demand; and compared with the method of directly using the maximum demand prediction model, the risk of bias can be avoided and the accuracy of predicting the maximum demand can be further improved.
[0008] In one possible design, the maximum demand of the current time period includes the maximum demand of each month of the current time period, and predicting the predicted maximum demand residual of each day of the next time period based on the maximum demand, the daily average demand of each day of the current time period, and the maximum demand residual prediction model includes: obtaining the difference between the daily average demand of each day of the current time period and the maximum demand of the month corresponding to the daily average demand; inputting the difference into the maximum demand residual prediction model, and predicting the predicted maximum demand residual of each day of the next time period.
[0009] In this design, the maximum demand of each month in the current time period can be obtained to improve the accuracy of the difference between the daily average demand and the maximum demand of each day, thereby making the maximum demand residual prediction model obtained by training and testing have higher accuracy.
[0010] In one possible design, the predicted maximum demand for the next time period is obtained and output based on the predicted daily average demand for each day in the next time period and the residual of the predicted maximum demand for each day in the next time period, including: taking the sum of the predicted daily average demand for each day in the next time period and the residual of the predicted maximum demand for each day in the next time period as the predicted maximum demand for each day in the next time period; and obtaining and outputting the predicted maximum demand for the next time period based on the predicted maximum demand for each day in the next time period.
[0011] In this design, the sum of the predicted daily average demand for each day in the next time period and the residual of the predicted maximum demand for each day in the next time period is a biased prediction of the predicted maximum demand for each day in the next time period, thereby improving the accuracy of the obtained maximum demand.
[0012] In one possible design, the predicted maximum demand for the next time period is obtained and output based on the predicted maximum demand for each day of the next time period, including: obtaining the standard deviation and the mean of the predicted maximum demand for the next time period based on the predicted maximum demand for each day of the next time period; taking the sum of the mean of the predicted maximum demand and the standard deviation as the predicted maximum demand for the next time period, and outputting the predicted maximum demand for the next time period.
[0013] In this design, if the average of the predicted maximum demand for each day in the historical time period is used as the predicted maximum demand for the historical time period in this embodiment, combined with the actual maximum demand for the historical time period, it is determined that the average of the predicted maximum demand for each day in the historical time period is smaller than the actual maximum demand. Based on this, in this embodiment, the standard deviation can be added to the mean to improve the prediction accuracy of the maximum demand.
[0014] In one possible design, the method also includes: obtaining information of the smart meter, the information of the smart meter including the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise; using the daily average demand of each day in the current time period and the information of the smart meter as training test parameters, and training to obtain the daily average demand prediction model.
[0015] In one possible design, the maximum demand of the current time period includes the maximum demand of each month of the current time period, and the method further includes: obtaining the difference between the daily average demand of each day of the current time period and the maximum demand of the month to which the daily average demand belongs; using the daily average demand of each day of the current time period, the difference, and the information of the smart meter as training test parameters, and training to obtain the maximum demand residual prediction model.
[0016] In the above two designs, the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise are used as training parameters. The daily average demand prediction model and the maximum demand residual prediction model obtained through training can be applied to enterprises in the corresponding industry and / or region and are universal.
[0017] In a possible design, the daily average demand prediction model and the maximum demand residual prediction model are both recurrent neural network (RNN) models.
[0018] A second aspect of the present application provides a maximum demand prediction device, comprising:
[0019] The transceiver module is used to receive the maximum demand in the current time period and multiple demands for each day in the current time period reported by the smart meter.
[0020] A processing module, used to obtain the predicted daily average demand for each day in the next time period according to multiple demands for each day in the current time period and a daily average demand prediction model, wherein the daily average demand prediction model indicates the corresponding relationship between the daily average demand for each day and the predicted daily average demand, and the daily average demand for each day is the average of multiple demands for each day in the current time period; and to obtain the predicted maximum demand residual for each day in the next time period according to the maximum demand, the daily average demand for each day in the current time period and a maximum demand residual prediction model, wherein the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand for each day in the current time period and the maximum demand and the predicted maximum demand residual; and to obtain and output the predicted maximum demand for the next time period according to the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period.
[0021] In a possible design, the maximum demand in the current time period includes the maximum demand in each month of the current time period.
[0022] Correspondingly, the processing module is specifically used to obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand; input the difference into the maximum demand residual prediction model to predict the predicted maximum demand residual for each day in the next time period.
[0023] In one possible design, the processing module is specifically used to take the sum of the predicted daily average demand for each day in the next time period and the residual of the predicted maximum demand for each day in the next time period as the predicted maximum demand for each day in the next time period; and obtain and output the predicted maximum demand for the next time period based on the predicted maximum demand for each day in the next time period.
[0024] In one possible design, the processing module is specifically used to obtain the standard deviation and the mean of the predicted maximum demand for the next time period based on the predicted maximum demand for each day of the next time period; take the sum of the mean of the predicted maximum demand and the standard deviation as the predicted maximum demand for the next time period, and output the predicted maximum demand for the next time period.
[0025] In a possible design, the processing module is also used to obtain information of the smart meter, and the information of the smart meter includes the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise; the daily average demand of each day in the current time period and the information of the smart meter are used as training test parameters to train and obtain the daily average demand prediction model.
[0026] In a possible design, the processing module is also used to obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month to which the daily average demand belongs; the daily average demand of each day in the current time period, the difference, and the information of the smart meter are used as training test parameters to train and obtain the maximum demand residual prediction model.
[0027] Optionally, the daily average demand prediction model and the maximum demand residual prediction model are both recurrent neural network (RNN) models.
[0028] The beneficial effects of the maximum demand prediction device provided by the second aspect and various possible designs can be found in the beneficial effects brought about by the first aspect and various possible designs, and will not be elaborated here.
[0029] The third aspect of the present application provides an electronic device, comprising: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the electronic device executes the maximum demand prediction method of the first aspect above.
[0030] A fourth aspect of the present application provides a computer-readable storage medium having computer-executable instructions stored thereon. When the computer-executable instructions are executed by a processor, the maximum demand prediction method of the first aspect described above is implemented.
[0031] Other effects of the above optional manner will be described below in conjunction with specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present application.
[0033] Figure 1It is a schematic diagram of a scenario to which the maximum demand prediction method provided in this application is applicable;
[0034] Figure 2 It is a flow chart of Example 1 of the maximum demand prediction method provided by the present application;
[0035] Figure 3 It is a flow chart of Example 2 of the maximum demand prediction method provided by the present application;
[0036] Figure 4 It is a structural schematic diagram of the maximum demand prediction device provided by the present application;
[0037] Figure 5 The present application is a block diagram of an electronic device for implementing a maximum demand prediction method. DETAILED DESCRIPTION
[0038] The following is a description of exemplary embodiments of the present application in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.
[0039] Definitions of terms in this application:
[0040] Maximum demand: refers to the maximum average power retained among the average power used by users within a preset time range. The maximum demand within the preset time range can be determined by using the maximum average power.
[0041] For example, taking the example of a smart meter reporting average power every 15 minutes, if the smart meter determines that the average power corresponding to the current 15 minutes is greater than the average power corresponding to the previous 15 minutes, the average power corresponding to the current 15 minutes is temporarily determined as the average power used to obtain the maximum demand of the month, and the average power corresponding to the maximum demand is updated to the average power corresponding to the current 15 minutes; otherwise, the average power corresponding to the previous 15 minutes is used as the average power corresponding to the maximum demand of the month. That is, the smart meter will continuously update the maximum average power within the preset range, and at the end of the month, the maximum demand of the month is determined by the maximum average power. This application does not elaborate on how to determine the corresponding demand based on the average power, and it can refer to the relevant description in the prior art.
[0042] Daily average demand: refers to the daily average demand for a day determined by the average power used by users in a day. For example, taking the example of an electric meter reporting average power every 15 minutes, the daily average demand is determined based on the average of multiple average powers reported by the electric meter.
[0043] At present, the calculation of the electricity bill for corporate electricity consumption adopts a two-part electricity price, including the basic electricity price and the kilowatt-hour electricity price. Among them, the kilowatt-hour electricity price is calculated based on the actual electricity consumption of the enterprise and the corresponding electricity price. The kilowatt-hour electricity price implements a time-of-use electricity price, which is commonly known as the peak and valley electricity price, that is, the electricity price during peak hours is higher than that during valley hours. For enterprises, the kilowatt-hour electricity price is related to production and can be adjusted according to the peak and valley time periods to save electricity expenses. However, in actual production, it is affected by the type of production (such as some industries require uninterrupted production) and human factors, and adjustments will affect production, and there is no relatively universal method.
[0044] The basic electricity price is divided into capacity electricity price and demand electricity price. The basic electricity price is required to be paid regardless of whether the enterprise uses electricity or not, and has nothing to do with the actual electricity consumption of the enterprise. Among them, the capacity electricity price is related to the capacity of the enterprise's transformer, and the demand electricity price is related to the maximum power required by the enterprise for a period of time. There is a price difference between the capacity electricity price and the demand electricity price. At present, most enterprises choose the capacity electricity price because the capacity of the enterprise's transformer is fixed, and the monthly capacity electricity price of the enterprise remains unchanged, which can be used as a fixed expense for the enterprise's electricity bill. When the capacity and demand are balanced, the capacity electricity price is equal to the demand electricity price; when the maximum demand of the enterprise each month is less than the maximum demand corresponding to the capacity and demand balance, the enterprise will pay more electricity bills by adopting the capacity electricity price than by adopting the demand electricity price. That is, when the maximum demand is less than the maximum demand corresponding to the capacity and demand balance, it will be more cost-effective for the enterprise to adopt the demand electricity price to save enterprise expenses.
[0045] However, the maximum demand of an enterprise may be different every month. If the demand price is adopted, once the maximum demand of the enterprise exceeds the maximum demand corresponding to the capacity-demand balance, even if it only exceeds it by a quarter of an hour, more electricity charges will be incurred. Therefore, the key to whether an enterprise adopts demand price is to accurately predict the maximum demand of the enterprise in the future. When the predicted maximum demand is less than the maximum demand corresponding to the capacity-demand balance, the enterprise adopts the demand price; when the predicted maximum demand is greater than the maximum demand corresponding to the capacity-demand balance, the enterprise adopts the capacity price to reduce the enterprise's expenditure on electricity bills.
[0046] At present, enterprises can adjust the basic electricity price every three months. In the prior art, the maximum demand in the past three months is usually multiplied by a fixed coefficient as the predicted maximum demand, and the basic electricity price is determined to be a capacity price or a demand price based on the predicted maximum demand. However, the fixed coefficient in the prior art is set manually, so the accuracy of the predicted maximum demand is low.
[0047] In order to solve the problem of low accuracy of maximum demand prediction in the prior art, the present application provides a maximum demand prediction method, which performs biased prediction of maximum demand through a pre-trained daily average demand prediction model and a maximum demand residual prediction model, so as to achieve the purpose of improving the accuracy of predicting maximum demand.
[0048] Figure 1 This is a schematic diagram of the scenario to which the maximum demand prediction method provided in this application is applicable. Figure 1 As shown, the scenarios applicable to the maximum demand prediction method provided by the present application include: smart meters and maximum demand prediction devices. Among them, the smart meter sends the demand corresponding to the average power in the preset time period and the maximum demand in the current month or the preset time period (one month) to the maximum demand prediction device at preset time intervals (for example, 15 minutes), and the maximum demand prediction device is used to receive the demand data reported by the smart meter. The maximum demand prediction device can be a server or a terminal device, etc.
[0049] Optionally, the terminal device may include but is not limited to a mobile terminal device or a fixed terminal device. Mobile terminal devices include but are not limited to mobile phones, personal digital assistants (PDAs), tablet computers, portable devices (e.g., portable computers, pocket computers, or handheld computers), etc. Fixed terminal devices include but are not limited to desktop computers, etc.
[0050] The maximum demand prediction method provided by the present application is described below in conjunction with specific embodiments. Figure 2 It is a flow chart of Example 1 of the maximum demand prediction method provided in the present application. Figure 2 The execution subject of the method flow shown may be a maximum demand prediction device, which may be implemented by any software and / or hardware. Figure 2 As shown, the maximum demand prediction method provided in this embodiment may include:
[0051] S201, receiving the maximum demand of the current time period and multiple demands of each day of the current time period reported by the smart meter.
[0052] In this embodiment, the current time period may be a time period for the enterprise to adjust the basic electricity price, such as 3 months. In the current time period, the enterprise may adopt a capacity electricity price or a demand electricity price, which is not limited in this embodiment. In the current time period, the maximum demand prediction device may receive the maximum demand of the current time period reported by the smart meter and the daily average demand of each day in the current time period.
[0053] The maximum demand in the current time period may be the maximum demand in three months. The smart meter may report the maximum demand corresponding to the maximum power in the current time period at the end of the current time period. In a possible implementation, the maximum demand in the current time period may be the maximum demand in each month of the current time period. Correspondingly, the smart meter may report the maximum demand in each month of the current time period at the end of each month of the current time period.
[0054] Among them, the multiple demands of each day in the current time period recorded by the smart meter can be the multiple demands reported by the smart meter at preset time intervals (such as 15 minutes) every day, such as 96 demands reported by the smart meter in one day.
[0055] For example, the maximum demand for the current time period and multiple demands for each day of the current time period received by the maximum demand prediction device are shown in a table as shown in Table 1 below:
[0056] Table 1
[0057]
[0058] S202, based on multiple demands of each day in the current time period and the daily average demand prediction model, obtain the predicted daily average demand of each day in the next time period, the daily average demand prediction model indicates the corresponding relationship between the daily average demand of each day and the predicted daily average demand, and the daily average demand of each day is the average of the multiple demands of each day in the current time period.
[0059] In this embodiment, after receiving multiple demands for each day in the current time period reported by the smart meter, the daily average demand for each day in the current time period can be obtained, where the daily average demand for each day is the average of the multiple demands for each day in the current time period.
[0060] The maximum demand prediction device in this embodiment may store a daily average demand prediction model, which may use the daily average demand of each day in the current time period as a training parameter, or use the daily average demand of each day in the historical time period as a training parameter (it can be imagined that the closer the date to the next time period is, the more accurate the daily average demand prediction model obtained is), and is obtained based on neural networks, vector machines, Bayesian and other methods. Among them, the method of obtaining the daily average demand prediction model in this embodiment is specifically described below and will not be introduced here. The daily average demand prediction model indicates the correspondence between the daily average demand of each day and the predicted daily average demand.
[0061] In this embodiment, the daily average demand of each day in the current time period is input into the daily average demand prediction model, and the predicted daily average demand of each day in the next time period output by the daily average demand prediction model can be obtained. For example, the daily average demand prediction model outputs the predicted daily average demand of the enterprise for each day in the next three months.
[0062] S203, obtaining the predicted maximum demand residual for each day in the next time period based on the maximum demand, the daily average demand for each day in the current time period, and the maximum demand residual prediction model, wherein the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand and the maximum demand for each day in the current time period and the predicted maximum demand residual.
[0063] In this embodiment, after receiving multiple demands for each day in the current time period reported by the smart meter, the daily average demand for each day in the current time period can be obtained. Then, the difference between the daily average demand for each day in the current time period and the maximum demand for the current time period can be obtained, that is, the difference corresponding to each day in the current time period can be obtained.
[0064] The maximum demand prediction device in this embodiment may store a maximum demand residual prediction model, which may be obtained based on neural network, vector machine, Bayesian and other methods, using the difference between the daily average demand of each day in the current time period and the maximum demand at the current time as a training parameter. The method of obtaining the maximum demand residual prediction model in this embodiment is specifically described below and will not be introduced here. The maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand and the maximum demand of each day in the current time period and the predicted maximum demand residual.
[0065] In this embodiment, the difference between the daily average demand of each day in the current time period and the maximum demand of the current time period can be input into the maximum demand residual prediction model, and the predicted maximum demand residual of each day in the next time period output by the maximum demand residual prediction model can be obtained. The predicted maximum demand residual of each day in the next time period refers to the difference between the predicted daily average demand and the predicted maximum demand for each day in the next time period.
[0066] As shown in Table 2 below, Table 2 exemplarily displays in a tabular form the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period.
[0067] Table 2
[0068] Date of next period Forecasted daily average demand Predicted maximum demand residual 1st of X month of X year a A 2nd day of month of year b B X year X month 3 c C 4th of X month of X year d D
[0069] S204, obtaining and outputting the predicted maximum demand for the next time period according to the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period.
[0070] In this embodiment, the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period may be added according to the date to obtain the predicted maximum demand for each day in the next time period.
[0071] In a possible implementation, in this embodiment, the average of the predicted maximum demand of each day in the next time period can be used as the predicted maximum demand of the next time period.
[0072] For example, as shown in Table 2 above, the predicted maximum demand for the next time period in this embodiment is:
[0073] In order to more clearly illustrate the beneficial effects of the maximum demand prediction method in this embodiment, a possible method of predicting the maximum demand in the prior art is described:
[0074] In the prior art, the maximum demand in the current time period can be obtained, and the maximum demand is used as a training parameter to obtain a maximum demand prediction model. On the one hand, the amount of data used as a training parameter for the maximum demand in the current time period is relatively small. On the other hand, the predicted value output by the prediction model is an unbiased estimate of the sample distribution, and the deviation factor is not considered. Therefore, the accuracy of the maximum demand of the next time period obtained by only using the maximum demand prediction model in the prior art is low. In this embodiment, two prediction models are used, one is the daily average demand prediction model, and the other is the maximum demand residual prediction model. On the one hand, the problem of the small amount of data for training parameters in the prior art can be solved. On the other hand, in this embodiment, the maximum demand of the next time period is not directly obtained according to the prediction model, but the daily average demand of each day in the next time period and the predicted maximum demand residual of each day are obtained. Then, the daily average demand of each day in the next time period and the predicted maximum demand residual of each day are added, and the maximum demand of the next time period obtained by prediction is weighted by deviation, which effectively avoids the deviation risk.
[0075] It should be understood that there is no order distinction between S202 and S203 in this embodiment, and the two can be executed simultaneously.
[0076] The maximum demand prediction method provided in this embodiment includes: receiving the maximum demand of the current time period and the daily average demand of each day in the current time period reported by the smart meter; obtaining the predicted daily average demand of each day in the next time period according to the daily average demand of each day in the current time period and the daily average demand prediction model, the daily average demand prediction model indicates the corresponding relationship between the daily average demand and the predicted daily average demand; obtaining the predicted maximum demand residual of each day in the next time period according to the maximum demand, the daily average demand of each day in the current time period and the maximum demand residual prediction model, the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand and the maximum demand of each day in the current time period and the predicted maximum demand residual; obtaining and outputting the predicted maximum demand for the next time period according to the predicted daily average demand of each day in the next time period and the predicted maximum demand residual of each day in the next time period. The maximum demand prediction method provided in this embodiment makes a biased prediction of the maximum demand through a pre-trained daily average demand prediction model and a maximum demand residual prediction model, thereby improving the accuracy of the obtained maximum demand; and compared with the method of directly using the maximum demand prediction model, it can avoid the risk of bias and further improve the accuracy of predicting the maximum demand.
[0077] Based on the above embodiments, Figure 3 The maximum demand prediction method provided in this application is further described in detail. Figure 3 : is a flow chart of the second embodiment of the maximum demand prediction method provided by the present application. Figure 3 As shown, the maximum demand prediction method provided in this embodiment may include:
[0078] S301, receiving the maximum demand of the current time period and multiple demands of each day of the current time period reported by the smart meter.
[0079] S302, obtaining predicted daily average demand for each day in the next time period according to multiple demands for each day in the current time period and a daily average demand prediction model.
[0080] It should be understood that the implementation of S301-S302 in this embodiment can refer to the relevant description in the above-mentioned embodiments S201-S202, and will not be repeated here.
[0081] In this embodiment, the method of obtaining the daily average demand prediction model is also introduced simultaneously:
[0082] The prediction device for maximum demand in this embodiment may store information of the smart meter, wherein the information of the smart meter may include the identification of the smart meter, the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise, etc. After the prediction device for maximum demand receives the maximum demand for the current time period and multiple demands for each day in the current time period reported by the smart meter, the maximum demand for the current time period and multiple demands for each day in the current time period may be stored in correspondence with the information of the smart meter according to the reported identification of the smart meter.
[0083] The daily average demand of each day in the current time period and the information of the smart meter can be used as training test parameters to train and obtain the daily average demand prediction model. Optionally, the daily average demand prediction model in this embodiment is a recurrent neural network RNN model.
[0084] In this embodiment, onehot encoding is performed on dates (such as the date of each day in the current time period), information corresponding to each date (such as the daily average demand of each day, the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise), etc., and the encoded data is used as training test parameters to train and obtain a daily average demand prediction model.
[0085] S303, obtaining the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand.
[0086] It should be understood that there is no order distinction between S303 - S304 in this embodiment and the above S302 , and the two can be executed simultaneously.
[0087] The maximum demand of the current time period in this embodiment includes the maximum demand of each month of the current time period. Correspondingly, when obtaining the difference between the daily average demand and the maximum demand of each day of the current time period, the difference between the daily average demand of each day of the current time period and the maximum demand of the month corresponding to the daily average demand is obtained.
[0088] For example, if the current time period includes July, August and September, the maximum demand of the current time period includes the maximum demand of each month of July, August and September. The difference between the daily average demand and the maximum demand of each day in July is the difference between the daily average demand of each day in July and the maximum demand of July.
[0089] In this embodiment, the maximum demand of each month in the current time period is obtained, which can improve the accuracy of the difference between the daily average demand and the maximum demand of each day, thereby making the maximum demand residual prediction model obtained by training and testing have higher accuracy.
[0090] S304, inputting the difference into a maximum demand residual prediction model to predict the maximum demand residual for each day in the next time period.
[0091] In this embodiment, the daily average demand of each day in the current time period, the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand, and the information of the smart meter can be used as training test parameters to train and obtain the maximum demand residual prediction model. Optionally, the daily average demand prediction model in this embodiment is a recurrent neural network RNN model.
[0092] In this embodiment, onehot encoding is performed with dates (such as the date of each day in the current time period), information corresponding to each date (such as the difference between the daily average demand of each day and the maximum demand of the month corresponding to the daily average demand, the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise), etc., and the encoded data is used as training test parameters to train and obtain the maximum demand residual prediction model.
[0093] It is worth noting that in this embodiment, information such as the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise is used as training parameters. The daily average demand prediction model and the maximum demand residual prediction model obtained through training can be applied to enterprises in the corresponding industry and / or region to perform daily average demand prediction and maximum demand residual prediction.
[0094] Correspondingly, in this embodiment, the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand can be input into the maximum demand residual prediction model to obtain the predicted maximum demand residual of each day in the next time period output by the maximum demand residual prediction model.
[0095] S305: The sum of the predicted daily average demand for each day in the next time period and the residual of the predicted maximum demand for each day in the next time period is taken as the predicted maximum demand for each day in the next time period.
[0096] In this embodiment, after obtaining the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period according to the above-mentioned daily average demand prediction model and the maximum demand residual prediction model, the sum of the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period can be used as the predicted maximum demand for each day in the next time period.
[0097] In a possible implementation, the average of the predicted maximum demand of each day in the next time period may be used as the predicted maximum demand of the next time period.
[0098] S306, according to the predicted maximum demand of each day in the next time period, obtain the standard deviation and the mean of the predicted maximum demand for the next time period, and take the sum of the mean and the standard deviation as the predicted maximum demand for the next time period.
[0099] In order to improve the accuracy of the predicted maximum demand for the next time period, in this embodiment, the standard deviation and the mean of the predicted maximum demand for the next time period can be obtained based on the predicted maximum demand for each day in the next time period, and the sum of the mean and the standard deviation of the predicted maximum demand can be used as the predicted maximum demand for the next time period.
[0100] According to the average of the predicted maximum demand of each day in the historical time period predicted by the daily average demand prediction model and the maximum demand residual prediction model in this embodiment, combined with the actual maximum demand of the historical time period, it is determined that the average of the predicted maximum demand of each day in the historical time period directly obtained according to the daily average demand prediction model and the maximum demand residual prediction model is smaller than the actual maximum demand. Based on this, in this embodiment, the standard deviation can be added to the mean to further improve the prediction accuracy of the maximum demand.
[0101] S307, output the predicted maximum demand for the next time period.
[0102] In this embodiment, after the maximum demand prediction device obtains the predicted maximum demand for the next time period, the predicted maximum demand for the next time period can be output. Optionally, if the maximum demand prediction device has a display module or a playback module, it can display or play the predicted maximum demand for the next time period. Optionally, the maximum demand prediction device in this embodiment can also output the predicted maximum demand for the next time period to the terminal device of the enterprise, so that the enterprise can regularly determine the predicted maximum demand for the next time period, and then determine whether to use capacity electricity price or demand electricity price in the next time period.
[0103] In this embodiment, information such as the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise is used as training parameters, and the daily average demand prediction model and the maximum demand residual prediction model obtained through training can be applicable to enterprises in the corresponding industry and / or region; and in this embodiment, the maximum demand of each month in the current time period can be obtained to improve the accuracy of the difference between the daily average demand and the maximum demand of each day, thereby making the maximum demand residual prediction model obtained by training and testing have higher accuracy; further, in this embodiment, the standard deviation and the mean of the predicted maximum demand of the next time period are obtained based on the predicted maximum demand of each day in the next time period, and the sum of the mean and the standard deviation of the predicted maximum demand is used as the predicted maximum demand for the next time period, thereby further improving the accuracy of the predicted maximum demand and improving the accuracy of the enterprise's selection of the basic electricity price.
[0104] Figure 4 Schematic diagram of the structure of the maximum demand prediction device provided by this application. Figure 4 As shown, the maximum demand prediction device 400 includes: a transceiver module 401 and a processing module 402 .
[0105] The processing module 401 is used to receive the maximum demand in the current time period and multiple demands for each day in the current time period reported by the smart meter.
[0106] Processing module 402 is used to obtain the predicted daily average demand for each day in the next time period based on multiple demands for each day in the current time period and the daily average demand prediction model, the daily average demand prediction model indicates the corresponding relationship between the daily average demand for each day and the predicted daily average demand, and the daily average demand for each day is the average of multiple demands for each day in the current time period; and to obtain the predicted maximum demand residual for each day in the next time period based on the maximum demand, the daily average demand for each day in the current time period and the maximum demand residual prediction model, the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand and the maximum demand for each day in the current time period and the predicted maximum demand residual; and to obtain and output the predicted maximum demand for the next time period based on the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period.
[0107] Optionally, the maximum demand in the current time period includes the maximum demand in each month of the current time period.
[0108] Processing module 402 is specifically used to obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand; input the difference into the maximum demand residual prediction model to predict the predicted maximum demand residual for each day in the next time period.
[0109] Optionally, the processing module 402 is specifically used to take the sum of the predicted daily average demand for each day in the next time period and the residual of the predicted maximum demand for each day in the next time period as the predicted maximum demand for each day in the next time period; based on the predicted maximum demand for each day in the next time period, obtain and output the predicted maximum demand for the next time period.
[0110] Optionally, processing module 402 is specifically used to obtain the standard deviation and mean of the predicted maximum demand for the next time period based on the predicted maximum demand for each day in the next time period; take the sum of the mean and standard deviation of the predicted maximum demand as the predicted maximum demand for the next time period, and output the predicted maximum demand for the next time period.
[0111] Optionally, the processing module 402 is also used to obtain information about smart meters, including the industry corresponding to the enterprise to which the smart meter belongs and / or the region corresponding to the enterprise; the daily average demand of each day in the current time period and the information of the smart meter are used as training test parameters to train and obtain a daily average demand prediction model.
[0112] Optionally, the processing module 402 is also used to obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month to which the daily average demand belongs; the daily average demand of each day in the current time period, the difference, and the information of the smart meter are used as training test parameters to train and obtain the maximum demand residual prediction model.
[0113] Optionally, the daily average demand prediction model and the maximum demand residual prediction model are both recurrent neural network (RNN) models.
[0114] According to an embodiment of the present application, the present application also provides an electronic device and a readable storage medium. Figure 5 The present application is a block diagram of an electronic device for implementing a maximum demand prediction method.
[0115] like Figure 5 As shown, electronic devices are intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.
[0116] like Figure 5 As shown, the electronic device includes: one or more processors 501, a memory 502, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In other embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 501 is taken as an example.
[0117] The memory 502 is a non-transitory computer-readable storage medium provided in the present application. The memory stores instructions executable by at least one processor to enable the at least one processor to perform the maximum demand prediction method provided in the present application. The non-transitory computer-readable storage medium of the present application stores computer instructions, which are used to enable a computer to perform the maximum demand prediction method provided in the present application.
[0118] The memory 502 is a non-transient computer-readable storage medium that can be used to store non-transient software programs, non-transient computer executable programs and modules, such as program instructions / modules corresponding to the maximum demand prediction method in the embodiment of the present application. The processor 501 executes various functional applications and data processing of the server by running the non-transient software programs, instructions and modules stored in the memory 502, that is, the maximum demand prediction method in the above method embodiment is implemented.
[0119] The memory 502 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of an electronic device for executing the maximum demand prediction method, etc. In addition, the memory 502 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory 502 may optionally include a memory remotely arranged relative to the processor 501, and these remote memories may be connected to the electronic device for executing the maximum demand prediction method via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0120] The electronic device of the maximum demand prediction method may further include: an input device 503 and an output device 504. The processor 501, the memory 502, the input device 503 and the output device 504 may be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.
[0121] The input device 503 can receive input digital or character information, and generate key signal input related to user settings and function control of the electronic device for executing the maximum demand prediction method, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick and other input devices. The output device Y04 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The display device may include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display and a plasma display. In some embodiments, the display device may be a touch screen.
[0122] Various implementations of the systems and techniques described herein can be realized in digital electronic circuit systems, integrated circuit systems, dedicated ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0123] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for programmable processors and can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or means (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.
[0124] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).
[0125] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.
[0126] A computer system may include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship to each other.
[0127] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution disclosed in this application can be achieved, and this document is not limited here.
[0128] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.
Claims
1. A maximum demand prediction method, characterized in that: include: Receive the maximum demand of the current time period and multiple demands of each day of the current time period reported by the smart meter; According to the multiple demands of each day in the current time period and the daily average demand prediction model, the predicted daily average demand of each day in the next time period is obtained, the daily average demand prediction model indicates the corresponding relationship between the daily average demand of each day and the predicted daily average demand, and the daily average demand of each day is the average of the multiple demands of each day in the current time period; Obtaining the predicted maximum demand residual for each day of the next time period according to the maximum demand, the daily average demand of each day in the current time period, and the maximum demand residual prediction model, wherein the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand of each day in the current time period and the maximum demand and the predicted maximum demand residual; Obtain and output the predicted maximum demand for the next time period according to the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period; The maximum demand of the current time period includes the maximum demand of each month of the current time period, and predicting the predicted maximum demand residual of each day of the next time period according to the maximum demand, the daily average demand of each day of the current time period, and the maximum demand residual prediction model includes: Obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand; Inputting the difference into the maximum demand residual prediction model to predict the predicted maximum demand residual for each day in the next time period; The acquiring and outputting the predicted maximum demand for the next time period according to the predicted daily average demand for each day of the next time period and the predicted maximum demand residual for each day of the next time period includes: The sum of the predicted daily average demand for each day in the next time period and the predicted maximum demand for each day in the next time period is taken as the predicted maximum demand for each day in the next time period; According to the predicted maximum demand of each day in the next time period, obtaining and outputting the predicted maximum demand of the next time period; The acquiring and outputting the predicted maximum demand for the next time period according to the predicted maximum demand for each day of the next time period includes: According to the predicted maximum demand of each day in the next time period, obtain the standard deviation and the mean of the predicted maximum demand in the next time period; taking the sum of the predicted maximum demand mean and the standard deviation as the predicted maximum demand for the next time period, and outputting the predicted maximum demand for the next time period; The maximum demand in the current time period includes the maximum demand in each month of the current time period, and the method further includes: Obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month to which the daily average demand belongs; The daily average demand of each day in the current time period, the difference, and the information of the smart meter are used as training test parameters to train and obtain the maximum demand residual prediction model.
2. The method according to claim 1, characterized in that The method further comprises: Acquire information of the smart meter, where the information of the smart meter includes an industry corresponding to an enterprise to which the smart meter belongs and / or a region corresponding to the enterprise; The daily average demand of each day in the current time period and the information of the smart meter are used as training test parameters to train and obtain the daily average demand prediction model.
3. The method according to claim 1 or 2, characterized in that: The daily average demand prediction model and the maximum demand residual prediction model are both recurrent neural network (RNN) models.
4. A maximum demand prediction device, characterized in that: include: A transceiver module, used for receiving the maximum demand of the current time period and multiple demands of each day of the current time period reported by the smart meter; A processing module, used to obtain the predicted daily average demand for each day in the next time period according to multiple demands for each day in the current time period and the daily average demand prediction model, and to obtain the predicted maximum demand residual for each day in the next time period according to the maximum demand, the daily average demand for each day in the current time period and the maximum demand residual prediction model, and to obtain and output the predicted maximum demand for the next time period according to the predicted daily average demand for each day in the next time period and the predicted maximum demand residual for each day in the next time period; the daily average demand prediction model indicates the corresponding relationship between the daily average demand for each day and the predicted daily average demand, and the daily average demand for each day is the average of multiple demands for each day in the current time period; the maximum demand residual prediction model indicates the corresponding relationship between the difference between the daily average demand for each day in the current time period and the maximum demand and the predicted maximum demand residual; The maximum demand in the current time period includes the maximum demand in each month of the current time period; The processing module is specifically used to obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month corresponding to the daily average demand; Inputting the difference into the maximum demand residual prediction model to predict the predicted maximum demand residual for each day in the next time period; The processing module is specifically used to take the sum of the predicted daily average demand of each day in the next time period and the predicted maximum demand of each day in the next time period as the predicted maximum demand of each day in the next time period; According to the predicted maximum demand of each day in the next time period, obtaining and outputting the predicted maximum demand of the next time period; The processing module is specifically used to obtain the standard deviation and the mean of the predicted maximum demand of the next time period according to the predicted maximum demand of each day in the next time period; taking the sum of the predicted maximum demand mean and the standard deviation as the predicted maximum demand for the next time period, and outputting the predicted maximum demand for the next time period; The processing module is further used to obtain the difference between the daily average demand of each day in the current time period and the maximum demand of the month to which the daily average demand belongs; The daily average demand of each day in the current time period, the difference, and the information of the smart meter are used as training test parameters to train and obtain the maximum demand residual prediction model.
5. The device according to claim 4, characterized in that The processing module is further used to obtain information of the smart meter, where the information of the smart meter includes an industry corresponding to the enterprise to which the smart meter belongs and / or a region corresponding to the enterprise; The daily average demand of each day in the current time period and the information of the smart meter are used as training test parameters to train and obtain the daily average demand prediction model.
6. The device according to claim 4 or 5, characterized in that The daily average demand prediction model and the maximum demand residual prediction model are both recurrent neural network (RNN) models.
7. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 3.
8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 3.
Citation Information
Patent Citations
Electricity market monthly electricity utilization prediction method
CN105574607A
A supply chain demand prediction method based on big data
CN109685583A
A data prediction method and device
CN109697522A
Method and device for predicting electricity consumption, electronic device and computer readable medium
CN110210672A
Estimate power generation amount calculation device, program, and method
JP2015089196A
