A method and system for determining a power selling scheme of a power seller

Abstract

This invention provides a method for determining electricity sales plans for electricity retailers, comprising: predicting load values ​​and marginal market electricity prices using an artificial neural network system to obtain predicted marginal electricity prices and predicted load values; obtaining the error between the predicted load values ​​and historical actual market electricity prices, and correcting the predicted marginal electricity prices and predicted load values ​​based on the error; and determining the optimal electricity purchase plan based on the corrected predicted marginal electricity prices and predicted load values. This plan results in reasonable electricity sales, maximizes energy utilization, and ensures reasonable costs for electricity retailers.

Description

Technical Field

[0001] This invention belongs to the technical field of electricity purchase and sale strategies for e-commerce retailers. The invention relates to a method and system for determining electricity sales schemes for e-commerce retailers. Background Technology

[0002] Currently, inaccurate estimates of electricity prices in the medium- and long-term markets and the spot market lead to inaccurate purchase and sale of electricity. Inaccurate sales of electricity result in insufficient energy supply to the load or a large amount of energy waste, which in turn leads to waste of production resources and unreasonable costs for electricity retailers. Summary of the Invention

[0003] In addressing the shortcomings of existing technologies, such as inaccurate estimations of medium- and long-term market and spot market electricity prices leading to inaccurate electricity purchase and sales, and the resulting insufficient energy supply or significant energy waste, both of which result in wasted production resources and unreasonable costs for electricity retailers, this invention provides a method for determining electricity sales plans for electricity retailers. The specific steps are as follows:

[0004] Based on the artificial neural network system, the load value and the marginal market electricity price are predicted to obtain the predicted marginal electricity price and the predicted load value;

[0005] The error is obtained by comparing the predicted load value and the historical actual value with the predicted marginal electricity price, and the predicted marginal electricity price and the predicted load value are corrected based on the error.

[0006] Based on the revised predicted marginal electricity price and predicted load value, the optimal electricity purchase plan is determined.

[0007] Preferably, determining the optimal electricity purchase plan based on the corrected predicted marginal electricity price and predicted load value includes:

[0008] The electricity purchase volume in the medium- and long-term market and the spot market is determined based on the predicted load value;

[0009] Calculate the electricity purchase cost in the medium- and long-term market and the electricity purchase cost in the spot market based on the electricity purchase volume in the medium- and long-term market and the revised predicted marginal electricity price;

[0010] The electricity purchase cost of electricity retailers is calculated based on the medium- and long-term market electricity purchase cost, the spot market electricity purchase cost, and electricity purchase cost constraints.

[0011] Based on the aforementioned electricity purchase cost, determine the optimal electricity purchase volume plan;

[0012] The electricity purchase cost constraints include: the total electricity volume actually traded by the electricity retailers in the medium- and long-term market or the spot market; the ratio of the maximum total electricity purchase volume of the electricity retailers in the medium- and long-term market and the spot market to the actual total electricity purchase volume; and the electricity volume purchase constraints for the electricity retailers participating in spot market transactions.

[0013] Preferably, the calculation of the medium- and long-term market electricity purchase cost is based on the electricity purchased in the medium- and long-term market and the spot market, and the corrected predicted marginal electricity price. and spot market electricity purchase costs As shown in the following formula:

[0014]

[0015]

[0016] In the formula, For the medium- to long-term electricity price during period t, This refers to the actual purchase volume of e-commerce platforms in the medium-to-long-term market during the t-period. This refers to the actual amount of electricity purchased by the e-commerce platform in the spot market during the t-hour period. This is the spot market quote obtained by the electricity retailer i based on the predicted marginal electricity price and the probability of successful bidding. This is a measure of whether an online retailer (i) can successfully bid.

[0017] Preferably, the calculation of the electricity purchase cost for the electricity retailer based on the medium- and long-term market electricity purchase cost, the spot market electricity purchase cost, and the electricity purchase cost constraint is as follows:

[0018]

[0019] In the formula, The total cost of electricity purchased by the e-commerce platform. For the medium- and long-term market electricity purchase costs of e-commerce platform i, Electricity purchase cost in the spot market for e-commerce platform i.

[0020] Preferably, determining the optimal electricity purchase volume based on the electricity purchase cost includes:

[0021] Calculate profit based on the aforementioned electricity purchase cost;

[0022] The optimal electricity purchase plan is determined based on maximizing profits.

[0023] Preferably, the profit calculation is based on the electricity purchase cost of the retailer. As shown in the following formula:

[0024]

[0025] In the formula, The total revenue during period T. The total cost of purchasing electricity for e-commerce retailers.

[0026] Preferably, the calculation of the total revenue of the electricity retailer includes: calculating the total electricity sold by the electricity retailer based on the actual electricity purchased in the spot market and the actual electricity purchased in the medium and long term market, and setting time-of-use fixed electricity price packages and peak-valley electricity price packages;

[0027] Based on the time-of-use fixed electricity price package and the peak-valley electricity price package, calculate the probability of choosing the fixed electricity price package and the probability of choosing the peak-valley electricity price package.

[0028] Based on the probability of choosing a fixed-price electricity package and the probability of choosing a peak-valley electricity package, calculate the total revenue of the electricity retailer during the scheduled period.

[0029] Preferably, the calculation of the probability of selecting a fixed-price electricity package and the probability of selecting a peak-valley electricity package includes:

[0030] The probability of choosing a fixed electricity price package The calculation is as follows:

[0031]

[0032] In the formula, k represents the user's sensitivity to price differences, and α and β represent the weighted average electricity price of fixed electricity price and peak-valley electricity price, respectively.

[0033] The probability of selecting the peak-valley electricity pricing package is:

[0034] Preferably, the total revenue The calculation is shown in the following formula:

[0035]

[0036] In the formula, This represents the actual total load. For fixed electricity price packages, This refers to the peak-valley electricity pricing package.

[0037] Preferably, the step of determining the optimal electricity purchase scheme based on profit maximization includes:

[0038] Calculate the electricity sales risk based on the aforementioned profit;

[0039] Based on profit maximization and the aforementioned electricity sales risks, the optimal solution is obtained.

[0040] Preferably, the electricity sales risk The calculation is shown in the following formula:

[0041]

[0042] In the formula, μ represents the probability that the possible loss is within the acceptable range. The maximum loss that can be tolerated.

[0043] Preferably, the process of obtaining the optimal transaction plan based on profit maximization and the electricity sales risk includes:

[0044]

[0045] In the formula, ρ is the risk aversion factor. For profit.

[0046] Preferably, the prediction of load value and marginal market electricity price based on the artificial neural network system to obtain the predicted marginal electricity price and predicted load value includes:

[0047] The load values ​​at the same time the day before the forecast period, the load values ​​at the same time the day last week, the load values ​​of the previous period, and whether the forecast period is a holiday are input into the artificial neural network system. The artificial neural network system outputs the forecast load value for the forecast period.

[0048] The market marginal electricity price of the same period on the day before the forecast period, the same period on the same day last week, the forecast load, the forecast period, and the previous period's electricity price are input into the artificial neural network system, and the artificial neural network system outputs the forecast marginal electricity price.

[0049] Preferably, the step of obtaining the error between the predicted value of the load value and the historical actual value of the marginal market electricity price, and correcting the predicted marginal electricity price and the predicted load value based on the error, includes:

[0050] The load error is obtained by comparing the predicted load value with the actual load value input to the artificial neural network system, and the predicted load value is corrected by comparison based on the load error.

[0051] The electricity price error is obtained by comparing the predicted marginal electricity price with the actual market marginal electricity price input into the artificial neural network system, and the predicted marginal electricity price is corrected by comparison based on the electricity price error.

[0052] Based on the same concept, the present invention provides a system for determining the electricity sales scheme of a retail e-commerce platform, including: a prediction module, a correction module and a scheme module;

[0053] The prediction module is used to predict the load value and the marginal market electricity price based on an artificial neural network system, so as to obtain the predicted marginal electricity price and the predicted load value.

[0054] The correction module is used to obtain the error between the predicted value of the load value and the historical actual value of the marginal market electricity price, and to correct the predicted marginal electricity price and the predicted load value based on the error.

[0055] The scheme module is used to determine the optimal electricity purchase scheme based on the corrected predicted marginal electricity price and predicted load value.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] 1. This invention provides a method for determining electricity sales plans for retail operators, comprising: predicting load values ​​and marginal market electricity prices based on an artificial neural network system to obtain predicted marginal electricity prices and predicted load values; obtaining the error between the predicted load values ​​and historical actual values ​​of the marginal market electricity prices, and correcting the predicted marginal electricity prices and predicted load values ​​based on the error; and determining the optimal electricity purchase plan based on the corrected predicted marginal electricity prices and predicted load values. This ensures reasonable costs for retail operators.

[0058] 2. This invention provides a method and system for determining electricity sales schemes for e-commerce retailers, resulting in reasonable electricity sales volume and maximizing energy utilization. Attached Figure Description

[0059] Figure 1 This is a flowchart of the method of the present invention;

[0060] Figure 2 A schematic diagram of an artificial neural network provided in an embodiment of the present invention;

[0061] Figure 3 A general flowchart of the method of the present invention provided for embodiments of the present invention;

[0062] Figure 4 A schematic diagram illustrating the probability of successful electricity purchase provided in an embodiment of the present invention;

[0063] Figure 5 The system structure diagram provided for this invention. Detailed Implementation

[0064] The present invention will be further described below with reference to the accompanying drawings:

[0065] Example 1:

[0066] This embodiment provides a method and system for determining electricity sales schemes for e-commerce retailers, as illustrated in the flowchart of the present invention. Figure 1 The specific solution is as follows:

[0067] Step 1: Based on the artificial neural network system, predict the load value and the marginal market electricity price to obtain the predicted marginal electricity price and the predicted load value;

[0068] Step 2: Obtain the error between the predicted load value and the historical actual value of the market marginal electricity price, and correct the predicted marginal electricity price and the predicted load value based on the error;

[0069] Step 3: Based on the corrected predicted marginal electricity price and predicted load value, determine the optimal electricity purchase plan;

[0070] Step 1: Based on an artificial neural network system, the load value and the marginal market electricity price are predicted to obtain the predicted marginal electricity price and the predicted load value.

[0071] The methods and ideas for e-commerce transaction strategies that incorporate artificial neural network prediction include the following steps:

[0072] (1) An artificial neural network system is used to predict total load and marginal market electricity price;

[0073] (2) Correct the prediction data based on the prediction error;

[0074] (3) Calculate the electricity purchase strategy of the electricity retailer, where the electricity purchase target includes two parts: the medium- and long-term market and the spot market;

[0075] (4) Analyze the electricity sales strategies of electricity retailers. The electricity sales packages offered by electricity retailers include fixed-price packages and peak-valley price packages.

[0076] (5) The CVar method is used to assess the risks of electricity sales and to derive the final electricity sales strategy.

[0077] Specifically, in step (1) above, an Artificial Neural Network (ANN) is used for predictive analysis. Artificial Neural Networks are inspired by biological nervous systems. Essentially, the connections between elements largely determine the network's function. By adjusting the connection (weight) values ​​between elements, the neural network can be trained to perform specific functions. In load and electricity price forecasting, many input / target pairs are typically needed to train the neural network. Figure 2 This illustrates the situation. The network is adjusted based on a comparison of the output and the target until the network output matches the target.

[0078] In fitting problems, neural networks are mapped between a set of numerical inputs and a set of numerical targets. Neural network fitting tools consist of feedforward networks, hidden neurons, and linear output neurons. They can adapt well to multidimensional mapping problems given consistent data and a sufficient number of hidden neurons. Choosing the right number of neurons is crucial for accurate prediction; too few neurons lead to insufficient accuracy, requiring an increase in the number of neurons; too many neurons result in overfitting, significantly degrading accuracy, necessitating a reduction in the number of neurons.

[0079] Specifically, in step (1) above, such as Figure 3As shown, the artificial neural network (ANN) prediction stage is divided into two parts. The first part is load forecasting. The input data for the ANN in load forecasting are the load at the same time the previous day, the load at the same time last week, the load of the previous time period, the period to be predicted, and whether it is a holiday. The output data is the predicted total load value P. total，f (t). The second part is the marginal electricity price forecast. The input data of the ANN are the electricity price of the same period of the previous day, the electricity price of the same period of the previous week, the forecast load, the period to be forecasted, and the electricity price of the previous period. The output data is the forecasted marginal electricity price.

[0080] In this invention, different numbers of neurons are selected, and historical data of the output and input values ​​are used to trial and error to determine the number of neurons N. The number of neurons is then confirmed based on the measurement accuracy. Finally, the number of neurons N is confirmed.

[0081] Step 2: Obtain the error between the predicted load value and the historical actual value of the marginal market electricity price, and correct the predicted marginal electricity price and the predicted load value based on the error:

[0082] Specifically, in step (2), the predicted load value and marginal electricity price are compared with historical data to obtain the following relationship:

[0083]

[0084]

[0085] In equation (1), This represents the actual total load. ε is the predicted total load. total The error between the predicted and actual load values ​​is given in equation (2). This is the actual marginal electricity price. For the predicted marginal electricity price, ε S This represents the error between the predicted and actual electricity prices. We can assume that the prediction error follows a normal distribution, i.e., we have... Where μ1 and μ2 are the mathematical expectations of the normal distributions N1 and N2, respectively. These are the variances of the normal distributions N1 and N2, respectively. The mean and standard deviation of the normal distribution can be obtained by statistically analyzing historical predicted and actual values.

[0086] Specifically, in step (2), the average of the actual load and the electricity price should be:

[0087]

[0088]

[0089] Specifically, in step (2), such as Figure 4 As shown, if the actual marginal electricity price of the system is λ S (t), the price quoted by the e-commerce platform i is λ0. Figure 4 In the middle, the vertical line represents the marginal electricity value λ. S (t), then the right side of the vertical line (λ0≥λ) S If (t) is a successful bid, then the probability of successful bidding for e-commerce platform i is γ=P(λ0≥λ). S (t)) (5)

[0090] The acceptable value of γ varies among different retailers. For a conservative market participant, γ might be relatively large, requiring an increase in λ0 to increase the likelihood of successful bidding. Conversely, retailers who prefer high-risk, high-return strategies have a higher risk tolerance, allowing for a relatively smaller value for γ and an even smaller value for λ0.

[0091] Specifically, in step (3), it is assumed that all electricity is traded in the medium- to long-term market or the spot market. Then, the actual total electricity volume in time period t satisfies the following relationship:

[0092] P total (t)=P B (t)+P s (t) (6)

[0093]

[0094] In equation (6), P total (t) represents the actual total electricity consumption during time period t, P B (t) represents the actual medium- and long-term market electricity volume during time period t, P s (t) represents the actual spot market electricity volume during time period t; in equation (7), This represents the total electricity actually purchased by e-commerce platform i during time period t. This refers to the amount of electricity that e-commerce platform i actually purchases from the medium- and long-term market during the t-period. This refers to the amount of electricity that e-commerce platform i actually purchases from the spot market during the t-hour period.

[0095] Step 3: Based on the corrected predicted marginal electricity price and predicted load value, determine the optimal electricity purchase plan:

[0096] Specifically, in step (3), the maximum total electricity that the retailer i can purchase from the medium- and long-term market and the spot market has a certain proportional relationship with the actual total electricity, which is as follows:

[0097]

[0098] Specifically, in step (3), the calculation expression for the electricity purchase cost from the medium- and long-term market by the electricity retailer during the period from 1 to T (t < T) is as follows:

[0099]

[0100] In equation (9), For the medium- and long-term market electricity purchase costs of e-commerce platform i, For the medium- to long-term electricity price during period t, This refers to the actual medium- and long-term electricity purchases by e-commerce platforms during period t.

[0101] Specifically, in step (3), the amount of electricity purchased by e-commerce platform i participating in spot market transactions has the following relationship:

[0102]

[0103] This refers to the actual amount of electricity purchased by the e-commerce platform in the spot market during the t-hour period.

[0104] Specifically, in step (3), the spot market electricity purchase cost of retailer i is expressed by the following expression:

[0105]

[0106] In equation (11), For e-commerce platforms, based on predicted electricity prices And the probability of a successful bid, γ, yields the spot market price. Let i be the metric indicating whether the online retailer i can successfully bid. Its expression is:

[0107]

[0108] Specifically, in step (3), the total electricity purchase cost for retailer i is:

[0109]

[0110] Specifically, in step (4), the total electricity consumption of e-commerce platform i on the electricity sales side is...

[0111]

[0112] Specifically, in step (4), it is assumed that the electricity retailer i has a fixed electricity price on the electricity retail side. Peak-valley (time-of-use) electricity pricing The expression is as follows:

[0113]

[0114] In equation (15), T peak T normal T vally These represent peak hours, normal hours, and valley hours, respectively. The electricity prices are for peak hours, normal hours, and off-peak hours, respectively.

[0115] Specifically, in step (4), it is assumed that the user selects an electricity price package based solely on the weighted average electricity price, and considering a certain user sensitivity k to price differences, the probability of the user choosing a fixed contract is... for:

[0116]

[0117] In equation (16), α and β represent the weighted average electricity price of fixed price and peak-valley price, respectively.

[0118] Specifically, in step (4), the total revenue of e-commerce platform i during time period T is:

[0119]

[0120] Specifically, in step (5), according to equations (13) and (17), the total profit of e-commerce platform i during time period T can be obtained. The expression:

[0121]

[0122] Specifically, in step (5), the potential loss of e-commerce platform i's purchasing and selling strategy during time period T is L, and the maximum tolerable loss is L. The probability of a potential loss being within acceptable limits is μ, and the relationship is as follows:

[0123]

[0124] Specifically, in step (5), the conditional value-at-risk model of the transaction strategy of the retailer i is as follows:

[0125]

[0126] Specifically, in step (5), the final global optimization model of the present invention is obtained:

[0127]

[0128] In equation (21), ρ is the risk aversion factor. Risk-averse e-commerce sellers take a larger ρ, while risk-loving e-commerce sellers take a smaller ρ.

[0129] Example 2:

[0130] like Figure 3 The flowchart shown is a process for e-commerce transaction strategies that incorporate artificial neural network predictions. The steps are explained in detail below.

[0131] Step 1: Use an artificial neural network system to predict the total load and the marginal market electricity price to obtain the predicted total load and predicted electricity price.

[0132] like Figure 3 As shown, the artificial neural network (ANN) prediction stage is divided into two parts. The first part is load forecasting. The input data for the ANN in load forecasting are the load at the same time the previous day, the load at the same time last week, the load of the previous time period, the period to be predicted, and whether it is a holiday. The output data is the predicted total load value P. total，f (t). The second part is the marginal electricity price forecast. The input data of the ANN are the electricity price of the same period of the previous day, the electricity price of the same period of the previous week, the forecast load, the period to be forecasted, and the electricity price of the previous period. The output data is the forecasted marginal electricity price.

[0133] In this invention, different numbers of neurons are selected, and historical data of the output and input values ​​are used to trial and error to determine the number of neurons N. The number of neurons is then confirmed based on the measurement accuracy. Finally, the number of neurons N is confirmed.

[0134] Step 2: Correct the predicted data based on the prediction error.

[0135] By comparing the predicted load value and marginal electricity price with historical data, the following relationship can be obtained:

[0136]

[0137]

[0138] In equation (1), This represents the actual total load. ε is the predicted total load. total The error between the predicted and actual load values ​​is given in equation (2). This is the actual marginal electricity price. For the predicted marginal electricity price, ε S This represents the error between the predicted and actual electricity prices. We can assume that the prediction error follows a normal distribution, i.e., we have... Where μ1 and μ2 are the mathematical expectations of the normal distributions N1 and N2, respectively. These are the variances of the normal distributions N1 and N2, respectively. The mean and standard deviation of the normal distribution can be obtained by statistically analyzing historical predicted and actual values.

[0139] Therefore, the average of actual load and electricity price should be:

[0140]

[0141]

[0142] like Figure 4 As shown, if the actual marginal electricity price of the system is λ S (t), the price quoted by the e-commerce platform i is λ0. Figure 4 In the middle, the vertical line represents the marginal electricity value λ. S (t), then the right side of the vertical line (λ0≥λ) S If (t) is a successful bid, then the probability of successful bidding for e-commerce platform i is γ=P(λ0≥λ). S (t)) (5)

[0143] The acceptable value of γ varies among different retailers. For a conservative market participant, γ might be relatively large, requiring an increase in λ0 to increase the likelihood of successful bidding. Conversely, retailers who prefer high-risk, high-return strategies have a higher risk tolerance, allowing for a relatively smaller value for γ and an even smaller value for λ0.

[0144] Step 3: Calculate the electricity purchase cost in the medium- and long-term market and the electricity purchase cost in the spot market separately, and add them together to obtain the electricity purchase cost for the electricity retailer.

[0145] Assuming all electricity is traded in the medium- to long-term market or the spot market, then the actual total electricity volume in time period t satisfies the following relationship:

[0146] P total (t)=P B (t)+P s (t) (6)

[0147]

[0148] In equation (6), P total (t) represents the actual total electricity consumption during time period t, P B (t) represents the actual medium- and long-term market electricity volume during time period t, P s (t) represents the actual spot market electricity volume during time period t; in equation (7), This represents the total electricity actually purchased by e-commerce platform i during time period t. This refers to the amount of electricity that e-commerce platform i actually purchases from the medium- and long-term market during the t-period. This refers to the amount of electricity that e-commerce platform i actually purchases from the spot market during the t-hour period.

[0149] The maximum total electricity that e-commerce platform i can purchase from the medium- and long-term market and the spot market has a certain proportional relationship δ with the actual total electricity. The relationship is as follows:

[0150]

[0151] The formula for calculating the cost of electricity purchased by a retailer from the medium- to long-term market during the period 1 to T (t < T) is as follows:

[0152]

[0153] In equation (9), For the medium- and long-term market electricity purchase costs of e-commerce platform i, For the medium- to long-term electricity price during period t, This refers to the actual medium- and long-term electricity purchases by e-commerce platforms during period t.

[0154] The following relationships exist between the purchase volume of e-commerce platforms participating in spot market transactions:

[0155]

[0156] Let be the actual electricity volume purchased by electricity retailer i in the spot market during time period t. The spot market electricity purchase cost of electricity retailer i is expressed by the following expression:

[0157]

[0158] In equation (11), For e-commerce platforms, based on predicted electricity prices And the probability of a successful bid, γ, yields the spot market price. Let i be the metric indicating whether the online retailer i can successfully bid. Its expression is:

[0159]

[0160] The total cost of electricity purchase for e-commerce platform i is:

[0161]

[0162] Step 4: Calculate the revenue from the fixed electricity price package and the revenue from the peak-valley electricity price package separately, and add them together to get the electricity sales revenue of the electricity retailer.

[0163] The total electricity volume of e-commerce platform i on the electricity sales side is

[0164]

[0165] Assume that electricity retailer i has a fixed electricity price on the retail side. Peak-valley (time-of-use) electricity pricing The expression is as follows:

[0166]

[0167] In equation (15), T peak T normal T vally These represent peak hours, normal hours, and valley hours, respectively. The electricity prices are for peak hours, normal hours, and off-peak hours, respectively.

[0168] In this invention, it is assumed that users select electricity packages solely based on the weighted average electricity price, and a certain user sensitivity k to price differences is considered, along with the probability that users will choose a fixed contract. for:

[0169]

[0170] In equation (16), α and β represent the weighted average electricity price of fixed price and peak-valley price, respectively.

[0171] The total revenue of e-commerce platform i during time period T is:

[0172]

[0173] Step 5: Use the CVar method to assess the risks of electricity sales, derive the final global optimization model, and thus obtain the final electricity sales strategy.

[0174] Based on equations (13) and (17), the total profit of e-commerce platform i during time period T can be obtained. The expression:

[0175]

[0176] The potential loss of e-commerce platform i's purchasing and selling strategy during time period T is L, and the maximum tolerable loss is... The probability of a potential loss being within acceptable limits is μ, and the relationship is as follows:

[0177]

[0178] The conditional value-at-risk model for the transaction strategy of e-commerce platform i is then:

[0179]

[0180] Finally, the final global optimization model of this invention is obtained:

[0181]

[0182] In equation (13), ρ is the risk aversion factor. Risk-averse e-commerce sellers take a larger ρ, while risk-loving e-commerce sellers take a smaller ρ.

[0183] Example 3:

[0184] Based on the same inventive concept, this invention also provides a system for determining the electricity sales scheme of an e-commerce retailer. Since the principle of these devices in solving technical problems is similar to that of a method for determining the electricity sales scheme of an e-commerce retailer, the repetitions will not be repeated.

[0185] The following is combined Figure 5The system structure diagram of the system is introduced. A system for determining the electricity sales scheme of a retail e-commerce platform is characterized by including: a prediction module, a correction module and a scheme module;

[0186] The prediction module is used to predict the load value and the marginal market electricity price based on an artificial neural network system, so as to obtain the predicted marginal electricity price and the predicted load value.

[0187] The correction module is used to obtain the error between the predicted value of the load value and the historical actual value of the marginal market electricity price, and to correct the predicted marginal electricity price and the predicted load value based on the error.

[0188] The scheme module is used to determine the optimal electricity purchase scheme based on the corrected predicted marginal electricity price and predicted load value.

[0189] The solution module includes: a sub-module for electricity purchase, a sub-module for electricity purchase costs in the two markets, a sub-module for electricity purchase costs of retail e-commerce platforms, and a sub-module for optimal solution;

[0190] The electricity purchase submodule is used to determine the electricity purchase volume for the medium- and long-term market and the spot market based on the predicted load value;

[0191] The two-market electricity purchase cost submodule is used to calculate the electricity purchase cost in the medium- and long-term market and the electricity purchase cost in the spot market based on the electricity purchase volume in the medium- and long-term market and the corrected predicted marginal electricity price;

[0192] The electricity purchase cost submodule for electricity retailers is used to calculate the electricity purchase cost of electricity retailers based on medium- and long-term market electricity purchase costs, spot market electricity purchase costs, and electricity purchase cost constraints.

[0193] The optimal solution submodule is used to determine the optimal electricity purchase volume scheme based on the electricity purchase cost.

[0194] The electricity purchase cost constraints include: the total electricity volume actually traded by the electricity retailers in the medium- and long-term market or the spot market; the ratio of the maximum total electricity purchase volume of the electricity retailers in the medium- and long-term market and the spot market to the actual total electricity purchase volume; and the electricity volume purchase constraints for the electricity retailers participating in spot market transactions.

[0195] The optimal submodule of the scheme includes: a profit unit and an optimal unit;

[0196] The profit unit is used to calculate profit based on the electricity purchase cost;

[0197] The optimal unit is used to determine the optimal electricity purchase plan based on maximizing profit.

[0198] The optimal unit includes: an electricity sales risk subunit and an optimal solution subunit;

[0199] The electricity sales risk subunit is used to calculate the electricity sales risk based on the profit;

[0200] The optimal solution subunit is used to obtain the optimal solution based on profit maximization and the electricity sales risk.

[0201] The forecasting module includes: a load forecasting submodule and an electricity price forecasting submodule;

[0202] The load forecasting submodule is used to input the load value at the same time the day before the forecast period, the load value at the same time the day last week, the load value of the previous period, and whether the forecast period is a holiday into the artificial neural network system. The artificial neural network system outputs the forecast load value for the forecast period.

[0203] The electricity price prediction submodule is used to input the market marginal electricity price of the same period on the previous day, the same period on the same day last week, the predicted load, the period to be predicted, and the electricity price of the previous period into the artificial neural network system, and the artificial neural network system outputs the predicted marginal electricity price.

[0204] The correction module includes: a load forecast correction submodule and an electricity price forecast correction submodule;

[0205] The load prediction correction submodule is used to obtain the load error based on the predicted load value and the actual load value input to the artificial neural network system, and to compare and correct the predicted load value based on the load error.

[0206] The electricity price prediction correction submodule is used to obtain the electricity price error based on the predicted marginal electricity price and the actual market marginal electricity price input into the artificial neural network system, and to compare and correct the predicted marginal electricity price based on the electricity price error.

[0207] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0208] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0209] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0210] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0211] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A method for determining an electricity sales scheme by an online retailer, characterized in that, include: Based on the artificial neural network system, the load value and the marginal market electricity price are predicted to obtain the predicted marginal electricity price and the predicted load value; The error is obtained by comparing the predicted load value and the historical actual value with the predicted marginal electricity price, and the predicted marginal electricity price and the predicted load value are corrected based on the error. Based on the revised predicted marginal electricity price and predicted load value, determine the optimal electricity purchase plan; The method of predicting load value and marginal market electricity price based on artificial neural network system to obtain predicted marginal electricity price and predicted load value includes: The load values ​​at the same time the day before the forecast period, the load values ​​at the same time the day last week, the load values ​​of the previous period, and whether the forecast period is a holiday are input into the artificial neural network system. The artificial neural network system outputs the forecast load value for the forecast period. The market marginal electricity price of the same period on the day before the forecast period, the same period on the same day last week, the forecast load, the forecast period, and the electricity price of the previous period are input into the artificial neural network system, and the artificial neural network system outputs the forecast marginal electricity price. The error is obtained by comparing the predicted load value and the market marginal electricity price with the historical actual value, and the predicted marginal electricity price and predicted load value are corrected based on the error, as expressed by the following formula: In the formula, Indicates the actual total load; This represents the predicted total load. This indicates the error between the predicted value and the actual value; Indicates the actual marginal electricity price; This represents the predicted marginal electricity price; This represents the error between the predicted and actual electricity prices; where, , , They represent the normal distribution of errors, respectively. , The mathematical expectation, , They represent the normal distribution of errors, respectively. , The variance; The process of determining the optimal electricity purchase plan based on the corrected predicted marginal electricity price and predicted load value includes: The electricity purchase volume in the medium- and long-term market and the spot market is determined based on the predicted load value; Calculate the electricity purchase cost in the medium- and long-term market and the electricity purchase cost in the spot market based on the electricity purchase volume in the medium- and long-term market and the revised predicted marginal electricity price; The electricity purchase cost of electricity retailers is calculated based on the medium- and long-term market electricity purchase cost, the spot market electricity purchase cost, and electricity purchase cost constraints. Based on the aforementioned electricity purchase cost, determine the optimal electricity purchase volume plan; The electricity purchase cost constraints include: the total electricity volume actually traded by the electricity retailers in the medium- and long-term market or the spot market; the ratio of the maximum total electricity purchase volume of the electricity retailers in the medium- and long-term market and the spot market to the actual total electricity purchase volume; and the electricity volume purchase constraints for the electricity retailers participating in spot market transactions.

2. The method as described in claim 1, characterized in that, The calculation of medium- and long-term market electricity purchase costs is based on the electricity purchased in the medium- and long-term market and the spot market, and the revised predicted marginal electricity price. and spot market electricity purchase costs As shown in the following formula: In the formula, for Medium- and long-term electricity prices for different time periods For e-commerce At The actual electricity purchased in the medium- and long-term market during the period. For e-commerce At The actual amount of electricity purchased in the spot market during that period. For e-commerce The spot market price is obtained based on the predicted marginal electricity price and the probability of a successful bid. For e-commerce The measure of whether a bid can be successfully placed.

3. The method as described in claim 2, characterized in that, The calculation of electricity purchase costs for electricity retailers based on medium- and long-term market electricity purchase costs, spot market electricity purchase costs, and electricity purchase cost constraints is shown in the following formula: In the formula, The total cost of electricity purchased by the e-commerce platform. For e-commerce The medium- and long-term market electricity purchase costs, Online retail The cost of purchasing electricity in the spot market.

4. The method as described in claim 1, characterized in that, The process of determining the optimal electricity purchase volume based on the electricity purchase cost includes: Calculate profit based on the aforementioned electricity purchase cost; The optimal electricity purchase plan is determined based on maximizing profits.

5. The method as described in claim 3, characterized in that, The profit calculation based on the electricity purchase cost of e-commerce retailers As shown in the following formula: In the formula, The total revenue during period T. The total cost of purchasing electricity for e-commerce retailers.

6. The method as described in claim 5, characterized in that, The calculation of the total revenue of the electricity retailers includes: calculating the total electricity sales volume of the electricity retailers based on the actual electricity purchase volume in the spot market and the actual electricity purchase volume in the medium and long term market, and setting time-of-use fixed electricity price packages and peak-valley electricity price packages; Based on the time-of-use fixed electricity price package and the peak-valley electricity price package, calculate the probability of choosing the fixed electricity price package and the probability of choosing the peak-valley electricity price package. Based on the probability of choosing a fixed-price electricity package and the probability of choosing a peak-valley electricity package, calculate the total revenue of the electricity retailer during the scheduled period.

7. The method as described in claim 6, characterized in that, The calculation of the probability of selecting a fixed-price electricity package and the probability of selecting a peak-valley electricity package includes: The probability of choosing a fixed electricity price package The calculation is as follows: In the formula, For users' sensitivity to price differences, α and β represent the weighted average electricity price of fixed price and peak-valley price, respectively; The probability of selecting the peak-valley electricity pricing package is: .

8. The method as described in claim 7, characterized in that, Total revenue The calculation is shown in the following formula: In the formula, This represents the actual total load. For fixed electricity price packages, This refers to the peak-valley electricity pricing package.

9. The method as described in claim 4, characterized in that, The optimal electricity purchase plan based on profit maximization includes: Calculate the electricity sales risk based on the aforementioned profit; Based on profit maximization and the aforementioned electricity sales risks, the optimal solution is obtained.

10. The method as described in claim 9, characterized in that, The aforementioned electricity sales risks The calculation is shown in the following formula: In the formula, μ represents the probability that the possible loss is within the acceptable range. The maximum loss that can be tolerated.

11. The method as described in claim 10, characterized in that, The optimal trading scheme, derived based on profit maximization and electricity sales risk, includes: In the formula, As a risk aversion factor, For profit.

12. The method as described in claim 1, characterized in that, The process of obtaining the error between the predicted load value and the historical actual value of the marginal market electricity price, and correcting the predicted marginal electricity price and predicted load value based on the error, includes: The load error is obtained by comparing the predicted load value with the actual load value input to the artificial neural network system, and the predicted load value is corrected by comparison based on the load error. The electricity price error is obtained by comparing the predicted marginal electricity price with the actual market marginal electricity price input into the artificial neural network system, and the predicted marginal electricity price is corrected by comparison based on the electricity price error.

13. A system for determining electricity sales schemes for e-commerce retailers, characterized in that, include: Prediction module, correction module, and solution module; The prediction module is used to predict the load value and the marginal market electricity price based on an artificial neural network system, so as to obtain the predicted marginal electricity price and the predicted load value. The correction module is used to obtain the error between the predicted value of the load value and the historical actual value of the market marginal electricity price, and to correct the predicted marginal electricity price and the predicted load value based on the error. The scheme module is used to determine the optimal electricity purchase scheme based on the corrected predicted marginal electricity price and predicted load value; The forecasting module includes: a load forecasting submodule and an electricity price forecasting submodule; The load forecasting submodule is used to input the load value at the same time the day before the forecast period, the load value at the same time the day last week, the load value of the previous period, and whether the forecast period is a holiday into the artificial neural network system. The artificial neural network system outputs the forecast load value for the forecast period. The electricity price prediction submodule is used to input the market marginal electricity price of the same period on the day before the forecast period, the same period on the same day last week, the forecast load, the forecast period, and the electricity price of the previous period into the artificial neural network system, and the artificial neural network system outputs the predicted marginal electricity price. The error is obtained by comparing the predicted load value and the market marginal electricity price with the historical actual value, and the predicted marginal electricity price and predicted load value are corrected based on the error, as expressed by the following formula: In the formula, Indicates the actual total load; This represents the predicted total load. This indicates the error between the predicted value and the actual value; Indicates the actual marginal electricity price; This represents the predicted marginal electricity price; This represents the error between the predicted and actual electricity prices; where, , , They represent the normal distribution of errors, respectively. , The mathematical expectation, , They represent the normal distribution of errors, respectively. , The variance; The solution module includes: a sub-module for electricity purchase, a sub-module for electricity purchase costs in the two markets, a sub-module for electricity purchase costs of retail e-commerce platforms, and a sub-module for optimal solution; The electricity purchase submodule is used to determine the electricity purchase volume for the medium- and long-term market and the spot market based on the predicted load value; The two-market electricity purchase cost submodule is used to calculate the electricity purchase cost in the medium- and long-term market and the electricity purchase cost in the spot market based on the electricity purchase volume in the medium- and long-term market and the corrected predicted marginal electricity price; The electricity purchase cost submodule for electricity retailers is used to calculate the electricity purchase cost of electricity retailers based on medium- and long-term market electricity purchase costs, spot market electricity purchase costs, and electricity purchase cost constraints. The optimal solution submodule is used to determine the optimal electricity purchase volume scheme based on the electricity purchase cost; The electricity purchase cost constraints include: the total electricity volume actually traded by the electricity retailers in the medium- and long-term market or the spot market; the ratio of the maximum total electricity purchase volume of the electricity retailers in the medium- and long-term market and the spot market to the actual total electricity purchase volume; and the electricity volume purchase constraints for the electricity retailers participating in spot market transactions.

Images