A Prediction Method for the Aggregation Bidding Curve in the Peak Regulation Market of a Virtual Power Plant

Abstract

The present invention provides a method for predicting the aggregated bidding curve in the virtual power plant peak shaving market, including obtaining the operation data of the virtual power plant to be predicted; inputting it into a preset deterministic model for prediction to obtain the predicted value of the virtual power plant revenue; determining the virtual power plant benefit threshold for each segment in the preset segmentation criteria; when the virtual power plant benefit threshold for this segment is less than the predicted value of the virtual power plant revenue, performing constraints through a preset robust model to obtain a first predicted value; when the virtual power plant benefit threshold for this segment is less than the predicted value of the virtual power plant revenue, performing constraints through a preset opportunity model to obtain a second predicted value; combining multiple first predicted values and multiple second predicted values to obtain the final predicted value of the aggregated bidding curve in the virtual power plant peak shaving market. The present invention takes into account the prediction error of the virtual power plant for the clearing price in the peak shaving market, filling the gap in the current technology for constructing the bidding in the virtual power plant peak shaving market.

Description

Technical Field

[0001] The present invention relates to the technical field of power system automation, and particularly to a method for predicting an aggregated bidding curve for peak shaving in a virtual power plant market. Background Art

[0002] At present, the penetration rate of renewable energy is continuously increasing. To improve the power grid regulation ability, peak shaving and frequency modulation resources are urgently needed. A virtual power plant (VPP) can effectively regulate the flexibility of distributed energy inside, and can be used as an emerging peak shaving and frequency modulation resource. Therefore, market mechanisms encouraging virtual power plants to participate in the peak shaving market have been promulgated everywhere in China, and it is clear that virtual power plants can participate in the peak shaving market as third-party independent entities.

[0003] Therefore, to participate in the peak shaving ancillary service market, it is necessary to study the technology for constructing the peak shaving capacity bidding curve of virtual power plants. Since there are a large number of distributed energy sources with small capacities inside virtual power plants, it is necessary to consider the aggregated distributed energy characterization model of virtual power plants. On the other hand, the peak shaving market bidding curve of virtual power plants is a segmented bidding curve, and how to reflect the change trend of peak shaving capacity with the peak shaving bidding price is a problem to be solved. Summary of the Invention

[0004] The object of the present invention is to propose a method for predicting an aggregated bidding curve for peak shaving in a virtual power plant market, so as to solve the technical problem that the existing method cannot effectively reflect the change of peak shaving capacity with the peak shaving bidding price.

[0005] On the one hand, a method for predicting an aggregated bidding curve for peak shaving in a virtual power plant market is provided, including:

[0006] Obtaining the operation data of the virtual power plant to be predicted and the data of the peak shaving market bidding curve of the virtual power plant to be predicted;

[0007] Taking the obtained operation data of the virtual power plant as an input item and inputting it into a preset deterministic model for prediction to obtain a predicted value of the virtual power plant revenue;

[0008] Determining the benefit threshold of the virtual power plant for each segment in a preset segmentation criterion according to the data of the peak shaving market bidding curve of the virtual power plant to be predicted;

[0009] Comparing the benefit threshold of the virtual power plant for each segment with the predicted value of the virtual power plant revenue respectively. When the benefit threshold of the virtual power plant for this segment is less than the predicted value of the virtual power plant revenue, constraining the predicted value of the virtual power plant revenue through a preset robust model to obtain a first predicted value; when the benefit threshold of the virtual power plant for this segment is less than the predicted value of the virtual power plant revenue, constraining the predicted value of the virtual power plant revenue through a preset opportunity model to obtain a second predicted value;

[0010] Combine multiple first prediction values and multiple second prediction values according to the preset segmentation criteria to obtain the final predicted value of the virtual power plant peak shaving market aggregation bidding curve.

[0011] Preferably, the virtual power plant operation data at least includes the revenue from the purchase and sale of electricity by the virtual power plant, the revenue of the virtual power plant in the peak shaving market, the scheduling cost of the virtual power plant for regulating electric vehicles, and the scheduling cost of the virtual power plant for regulating temperature control loads.

[0012] Preferably, the deterministic model specifically includes:

[0013] B0 = Max B E +B F -C EV -C TCL

[0014] Among them, B0 represents the predicted value of the virtual power plant revenue, that is, the maximum revenue of the virtual power plant operator in the power market, B E represents the revenue from the purchase and sale of electricity by the virtual power plant, B F represents the revenue of the virtual power plant in the peak shaving market, C EV represents the scheduling cost of the virtual power plant for regulating electric vehicles, C TCL The fourth item represents the scheduling cost of the virtual power plant for regulating temperature control loads.

[0015] Preferably, calculate the scheduling cost of the virtual power plant for regulating electric vehicles according to the following formula:

[0016] P t VB,min ≤P t EV ≤P t VB,max

[0017]

[0018]

[0019] Among them, represents the upper limit of the aggregated energy of electric vehicles, represents the lower limit of the aggregated energy of electric vehicles, P t VB,max represents the upper limit of the aggregated power of electric vehicles, P t VB,min represents the lower limit of the aggregated power of electric vehicles, represents the scheduling cost of the virtual power plant for regulating electric vehicles, that is, the energy demand of electric vehicles at time t + 1, represents the energy of electric vehicles at time t, P t EV represents the power of electric vehicles at time t, Represents the step energy change caused by the entry and exit of the electric vehicle at time t+1.

[0020] Preferably, the scheduling cost of the virtual power plant for regulating the temperature control load is calculated according to the following formula:

[0021]

[0022] P t TCL,min ≤P t TCL ≤P t TCL,max

[0023] T TCL,min ≤T t ≤T TCL,max

[0024] Where, N w Represents the number of temperature control loads, T t Represents the temperature of the temperature control load at time t, T t+1 Represents the temperature of the temperature control load at time t+1, η, R, C a Represents the temperature control load parameter, P t TCL,min Represents the power upper limit, P t TCL,max Represents the power lower limit, T TCL,min Represents the temperature upper limit, T TCL,max Represents the temperature lower limit.

[0025] Preferably, the deterministic model also constrains the predicted value of the virtual power plant revenue according to the following formula:

[0026] P t B -P t S =P t EV +P t TCL -P t PV

[0027] Where, P t B Represents the power purchase of the virtual power plant, P t S Represents the power sale of the virtual power plant, P t EV Represents the electric vehicles in the virtual power plant, P t TCL Represents the temperature control load in the virtual power plant, P t PVRepresents the photovoltaic power of electric vehicles in a virtual power plant.

[0028] Preferably, the deterministic model further constrains the predicted value of the virtual power plant's revenue according to the following formula:

[0029] P t B -P t S -P t Base =P F

[0030] Where P t Base is the operating curve without considering peak shaving services, and P F is the peak shaving market bidding volume.

[0031] Preferably, the robust model specifically includes:

[0032]

[0033]

[0034]

[0035] Where α represents the fluctuation range of the uncertain quantity, P represents the decision variable, that is, the day-ahead operation plan of the virtual power plant, B represents the constraint on the decision variable, v represents the actual value of the uncertain variable, B R represents the threshold of the objective function, B0 represents the predicted value of the virtual power plant's revenue, ρ F represents the deviation coefficient of the uncertain variable, represents the average value of the deviation coefficient, and σ represents the robust factor.

[0036] Preferably, the chance model specifically includes:

[0037]

[0038]

[0039]

[0040] Where α represents the fluctuation range of the uncertain quantity, P represents the decision variable, that is, the day-ahead operation plan of the virtual power plant, B represents the constraint on the decision variable, v represents the actual value of the uncertain variable, B R represents the threshold of the objective function, B0 represents the predicted value of the virtual power plant's revenue, ρ F represents the deviation coefficient of the uncertain variable, represents the average value of the deviation coefficient, and σ represents the robust factor.

[0041] Preferably, the combination of a plurality of first prediction values and a plurality of second prediction values according to a preset segmentation standard includes:

[0042] Determine the time period in the preset segmentation standard corresponding to each first prediction value or each second prediction value;

[0043] Sort each first prediction value and each second prediction value according to the preset segmentation standard, and output the sorting result as the final prediction value of the virtual power plant peak shaving market aggregation bidding curve.

[0044] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0045] The prediction method of the virtual power plant peak shaving market aggregation bidding curve provided by the present invention aims at the problems of a large number and small capacity of distributed energy sources in the virtual power plant, considers the regulation costs of electric vehicles and thermostatic loads, establishes the aggregated regulation cost of electric vehicles and thermostatic loads, and proposes a bidding model for the virtual power plant in the peak shaving market; considering the prediction error of the virtual power plant for the clearing price in the peak shaving market, the information gap theory is used to generate a segmented bidding curve, which can ensure the established revenue of the virtual power plant in the peak shaving market and fill the gap in the current technology for constructing the virtual power plant peak shaving market bidding. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, obtaining other drawings without creative efforts still belongs to the scope of the present invention.

[0047] Figure 1 It is a schematic diagram of the main process of a prediction method for a virtual power plant peak shaving market aggregation bidding curve in an embodiment of the present invention.

[0048] Figure 2 It is a schematic diagram of the logic of a prediction method for a virtual power plant peak shaving market aggregation bidding curve in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0050] As Figure 1 and Figure 2 shown, it is a schematic diagram of an embodiment of a prediction method for a virtual power plant peak shaving market aggregation bidding curve provided by the present invention. In this embodiment, the method includes the following steps:

[0051] Obtain the virtual power plant operation data to be predicted and the virtual power plant peak shaving market bidding curve data to be predicted; specifically, the virtual power plant operation data at least includes the revenue of the virtual power plant's power purchase and sale, the revenue of the virtual power plant in the peak shaving market, the dispatching cost of the virtual power plant's regulation of electric vehicles, and the dispatching cost of the virtual power plant's regulation of temperature control loads. The virtual power plant peak shaving market bidding curve data is a segmented bidding curve.

[0052] Furthermore, input the obtained virtual power plant operation data as an input item into a preset deterministic model for prediction to obtain the predicted value of the virtual power plant revenue; that is, establish a deterministic model without considering uncertainty; the deterministic model includes: the electric vehicle aggregation model constraint, the temperature control load aggregation model constraint, and the peak shaving market bidding capacity constraint.

[0053] In a specific embodiment, the deterministic model specifically includes:

[0054] B0 = Max B E +B F -C EV -C TCL

[0055] Among them, B0 represents the predicted value of the virtual power plant revenue, that is, the maximum revenue of the virtual power plant operator in the power market, B E represents the revenue of the virtual power plant's power purchase and sale, B F represents the revenue of the virtual power plant in the peak shaving market, C EV represents the dispatching cost of the virtual power plant's regulation of electric vehicles, C TCL represents the fourth item, which represents the dispatching cost of the virtual power plant's regulation of temperature control loads. That is, when using conventional methods (such as probability methods, fuzzy programming methods, etc.) to handle the randomness of uncertain variables, a large amount of historical data of the uncertain variable needs to be obtained, and the error probability distribution between the predicted value and the actual value of the uncertain variable needs to be modeled, so as to determine the error random distribution of this uncertain variable according to the error probability model. The information gap decision theory is a non-probability method for dealing with uncertainty. A key difference between it and other probability decision theories is that this method obtains efficient, high-priority, and risk-averse solutions by modeling the interval error between the actual data and the predicted data, rather than through probability. The information gap decision theory is often used to solve uncertain scenarios that are difficult to accurately describe. Using the information gap decision model can quantify the uncertainty of data when the probability distribution and fluctuation range of the data are both unknown. Its basic idea is: while ensuring that the set target value is within the acceptable range of fluctuations above and below the benchmark target value, maximize the fluctuation range of the uncertain variable as much as possible to obtain a greater possibility of the target value that meets the conditions. The information gap decision model usually includes the following three parts: a system model, an uncertainty model, and a performance requirement:

[0056] (1) The system model is a mathematical model established for specific uncertainty problems and the decision variables to be solved, usually denoted by B(P,ν). Among them, P is the decision variable, which is the day-ahead operation plan of the virtual power plant, and ν and represent the actual value and the predicted value of the uncertain variable respectively. The system model represents the benefits obtained by the virtual power plant in the peak shaving market bidding.

[0057] (2) The uncertainty model is a mathematical expression of uncertain quantities, used to describe the gap between the actual data ν and the predicted data . Among them, ν={ρ f} indicates that the uncertain factors consider the day-ahead peak shaving market price, and is usually expressed as follows:

[0058]

[0059] In the formula: α—the fluctuation range of the uncertain quantity.

[0060] (3) According to different risk management strategies, the corresponding decision-making strategies can be selected in the information gap decision model according to the performance requirements needed in the actual situation. The common ones are the robust model decision-making strategy and the opportunity model decision-making strategy.

[0061] Specifically, the scheduling cost of the virtual power plant for regulating electric vehicles is calculated according to the following formula:

[0062] P t VB,min ≤P t EV ≤P t VB,max

[0063]

[0064]

[0065] Among them, represents the upper limit of the aggregated energy of electric vehicles, represents the lower limit of the aggregated energy of electric vehicles, P t VB,max represents the upper limit of the aggregated power of electric vehicles, P t VB,min represents the lower limit of the aggregated power of electric vehicles, represents the scheduling cost of the virtual power plant for regulating electric vehicles, that is, the energy demand of electric vehicles at time t + 1, represents the energy of electric vehicles at time t, P t EV represents the power of electric vehicles at time t, Indicates the step energy change caused by the entry and exit of electric vehicles at time t+1. It can be understood that when determining the aggregated electric vehicle model, the operating constraints of individual electric vehicles are considered first, including:

[0066]

[0067]

[0068] P i,t EV,max =min{(E i,t+1 EV,max -E i,t EV,min ) / Δt,P i max}t∈(To i ,Td i -1)

[0069] P i,t EV,min =min{(E i,t+1 EV,min -E i,t EV,max ) / Δt,0}t∈(To i ,Td i -1)

[0070] Among them, E i,t EV,max , E i,t EV,min are the upper and lower limits of the electric vehicle energy respectively, P i,t EV,max , P i,t EV,min are the upper and lower limits of the electric vehicle power respectively, To i , Td i are the entry and exit times of the electric vehicle respectively, P i max are the rated powers of the electric vehicle respectively.

[0071] Thus, the aggregated electric vehicle model is obtained:

[0072]

[0073]

[0074]

[0075]

[0076] Among them, P t VB,max , Pt VB,min They are the upper and lower limits of the aggregated energy and aggregated power of electric vehicles, respectively.

[0077] It is also necessary to consider the step energy change caused by the entry and exit of electric vehicles:

[0078]

[0079] Therefore, the aggregated electric vehicle model is obtained:

[0080] P t VB,min ≤P t EV ≤P t VB,max

[0081]

[0082]

[0083] In addition, since electric vehicle users prefer the operation mode of charging as soon as they stop, the virtual power plant needs to pay a dispatching cost for regulating electric vehicles.

[0084]

[0085] Among them, is the energy baseline formed when the electric vehicle operates in the mode of charging as soon as it stops.

[0086] Specifically, the dispatching cost of the virtual power plant for regulating the temperature control load is calculated according to the following formula:

[0087]

[0088] P t TCL,min ≤P t TCL ≤P t TCL,max

[0089] T TCL,min ≤T t ≤T TCL,max

[0090] Among them, N w represents the number of temperature control loads, T t represents the temperature of the temperature control load at time t, T t+1 represents the temperature of the temperature control load at time t + 1, and η, R, C a represent the parameters of the temperature control load, P t TCL,min represents the power upper limit, P t TCL,maxRepresents the lower power limit, T TCL,min Represents the upper temperature limit, T TCL,max Represents the lower temperature limit. It can be understood that when determining the aggregated temperature control load model,

[0091]

[0092] P t TCL,min ≤P t TCL ≤P t TCL,max

[0093] T TCL,min ≤T t ≤T TCL,max

[0094] Among them, N w is the number of temperature control loads, T t is the temperature of the temperature control load, η, R, C a are the parameters of the temperature control load, P t TCL,min , P t TCL,max are the upper and lower power limits, T TCL,min , T TCL,max are the upper and lower temperature limits.

[0095] In addition, since users are more inclined to the operation mode where the indoor temperature is at the set temperature, the virtual power plant needs to pay a dispatching cost for regulating the temperature control load.

[0096]

[0097] Among them, T Set is the set temperature.

[0098] Specifically, the deterministic model also constrains the predicted value of the virtual power plant's revenue according to the following formula:

[0099] P t B -P t S =P t EV +P t TCL -P t PV

[0100] Among them, P t B represents the power purchase of the virtual power plant, P t S represents the power sale of the virtual power plant, P t EVRepresents the electric vehicles in the virtual power plant, P t TCL Represents the temperature control load in the virtual power plant, P t PV Represents the photovoltaic power of the electric vehicles in the virtual power plant. That is, it is necessary to ensure power balance within the virtual power plant.

[0101] More specifically, the deterministic model also constrains the predicted value of the virtual power plant revenue according to the following formula:

[0102] P t B -P t S -P t Base =P F

[0103] Wherein, P t Base Is the operation curve without considering the peak shaving service, and P F Is the peak shaving market bid volume. That is, the peak shaving market bid capacity constraint.

[0104] Furthermore, determine the virtual power plant interest threshold for each segment in the preset segmentation standard according to the virtual power plant peak shaving market bidding curve data to be predicted; that is, according to the predicted clearing price of the peak shaving market and the segmented data of the bidding curve, the interest threshold of the virtual power plant during optimization for each segment can be directly determined according to the curve.

[0105] Furthermore, compare the virtual power plant interest threshold for each segment with the predicted value of the virtual power plant revenue respectively. When the virtual power plant interest threshold for this segment is less than the predicted value of the virtual power plant revenue, constrain the predicted value of the virtual power plant revenue through a preset robust model to obtain the first predicted value; when the virtual power plant interest threshold for this segment is less than the predicted value of the virtual power plant revenue, constrain the predicted value of the virtual power plant revenue through a preset opportunity model to obtain the second predicted value; that is, the information gap decision theory is used to cope with the economic risks in the power market, and the goal of the decision maker is to maximize the range of uncertain factors while meeting the key profit. The uncertain parameter considered is the clearing price of the peak shaving market. According to the predicted clearing price of the peak shaving market, the virtual power plant operator determines the day-ahead bid volume and the operation plan of the distributed energy, and considers its economic risks through the information gap decision theory.

[0106] In a specific embodiment, the robust model specifically includes:

[0107]

[0108]

[0109]

[0110] Among them, α represents the fluctuation range of uncertain quantities, P represents the decision variable, that is, the day-ahead operation plan of the virtual power plant, B represents the constraint on the decision variable, v represents the actual value of the uncertain variable, B R represents the threshold of the objective function, B0 represents the predicted value of the virtual power plant's revenue, ρ F represents the deviation coefficient of the uncertain variable, represents the average value of the deviation coefficient, and σ represents the robust factor. In the decision-making strategy of the robust model, the goal is to maximize the range of uncertain parameters while ensuring the minimum specified profit; among them, the constraint conditions are: power balance constraint, electric vehicle aggregation model constraint, temperature control load aggregation model constraint, peak shaving market bidding capacity constraint;

[0111] Specifically, the opportunity model specifically includes:

[0112]

[0113]

[0114]

[0115] Among them, α represents the fluctuation range of uncertain quantities, P represents the decision variable, that is, the day-ahead operation plan of the virtual power plant, B represents the constraint on the decision variable, v represents the actual value of the uncertain variable, B R represents the threshold of the objective function, B0 represents the predicted value of the virtual power plant's revenue, ρ F represents the deviation coefficient of the uncertain variable, represents the average value of the deviation coefficient, and σ represents the robust factor. In the decision-making strategy of the opportunity model, the goal is to minimize the range of uncertain parameters while the operator pursues higher profits; among them, the constraint conditions are: power balance constraint, electric vehicle aggregation model constraint, temperature control load aggregation model constraint, peak shaving market bidding capacity constraint.

[0116] Furthermore, according to the preset segmentation standard, multiple first prediction values and multiple second prediction values are combined to obtain the final predicted value of the virtual power plant's peak shaving market aggregation bidding curve.

[0117] Specifically, determine the time period in the preset segmentation standard corresponding to each first prediction value or each second prediction value; sort each first prediction value and each second prediction value according to the preset segmentation standard, and output the sorting result as the final predicted value of the virtual power plant's peak shaving market aggregation bidding curve.

[0118] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0119] The prediction method for the aggregated bidding curve in the virtual power plant peak shaving market provided by the present invention addresses the problem of a large number of distributed energy sources with small capacities within the virtual power plant. Considering the regulation costs of electric vehicles and thermostatic loads, it establishes the aggregated regulation costs of electric vehicles and thermostatic loads, and proposes a bidding model for the virtual power plant in the peak shaving market. Considering the prediction error of the virtual power plant for the clearing price in the peak shaving market, it uses the information gap theory to generate a segmented bidding curve, which can ensure the established revenue of the virtual power plant in the peak shaving market and fills the gap in the current technology for constructing the bidding in the virtual power plant peak shaving market.

[0120] The above-disclosed are only the preferred embodiments of the present invention, and of course, the scope of rights of the present invention cannot be limited thereby. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.

Claims

1. A prediction method for the peak shaving market aggregation bidding curve of a virtual power plant, characterized in that Including: Obtain the virtual power plant operation data to be predicted and the virtual power plant peak shaving market bidding curve data to be predicted; Take the obtained virtual power plant operation data as an input item and input it into a preset deterministic model for prediction to obtain a predicted value of the virtual power plant revenue; Determine the virtual power plant benefit threshold for each segment in the preset segmentation criteria according to the virtual power plant peak shaving market bidding curve data to be predicted; Compare the virtual power plant benefit threshold for each segment with the predicted value of the virtual power plant revenue respectively. When the virtual power plant benefit threshold for this segment is less than the predicted value of the virtual power plant revenue, constrain the predicted value of the virtual power plant revenue through a preset robust model to obtain a first predicted value; when the virtual power plant benefit threshold for this segment is less than the predicted value of the virtual power plant revenue, constrain the predicted value of the virtual power plant revenue through a preset opportunity model to obtain a second predicted value; Combine multiple first predicted values and multiple second predicted values according to the preset segmentation criteria to obtain the final predicted value of the virtual power plant peak shaving market aggregated bidding curve.

2. The method according to claim 1, wherein The virtual power plant operation data at least includes the revenue from the purchase and sale of electricity by the virtual power plant, the revenue of the virtual power plant in the peak shaving market, the dispatching cost of the virtual power plant for regulating electric vehicles, and the dispatching cost of the virtual power plant for regulating temperature control loads.

3. The method according to claim 2, wherein The deterministic model specifically includes: B0 = Max B E + B F - C EV - C TCL Among them, B0 represents the predicted value of the virtual power plant's revenue, that is, the maximum revenue of the virtual power plant operator in the electricity market, B E represents the revenue from the virtual power plant's electricity purchase and sale, B F represents the revenue of the virtual power plant in the peak shaving market, C EV represents the scheduling cost of the virtual power plant for regulating electric vehicles, C TCL The fourth item represents the scheduling cost of the virtual power plant for regulating the temperature control load.

4. The method according to claim 3, characterized in that, Calculate the dispatching cost of the virtual power plant for regulating electric vehicles according to the following formula: P t VB,min ≤P t EV ≤P t VB,max Among them, represents the upper limit of the aggregated energy of electric vehicles, represents the lower limit of the aggregated energy of electric vehicles, P t VB,max represents the upper limit of the aggregated power of electric vehicles, P t VB,min represents the lower limit of the aggregated power of electric vehicles, represents the dispatching cost for the virtual power plant to regulate electric vehicles, that is, the energy demand of electric vehicles at time t+1, represents the energy of electric vehicles at time t, P t EV represents the power of electric vehicles at time t, represents the step energy change caused by the in and out of electric vehicles at time t+1.

5. The method according to claim 3, wherein Calculate the dispatching cost of the virtual power plant for regulating temperature control loads according to the following formula: P t TCL,min ≤P t TCL ≤P t TCL,max T TCL,min ≤T t ≤T TCL,max Among them, N w represents the number of thermostatic loads, T t represents the temperature of the thermostatic load at time t, T t+1 represents the temperature of the thermostatic load at time t + 1, η, R, C a represent the parameters of the thermostatic load, P t TCL,min represents the upper power limit, P t TCL,max represents the lower power limit, T TCL,min represents the upper temperature limit, T TCL,max represents the lower temperature limit.

6. The method according to claim 3, wherein The deterministic model also constrains the predicted value of the virtual power plant revenue according to the following formula: P t B -P t S =P t EV +P t TCL -P t PV Among them, P t B represents the power purchase of the virtual power plant, P t S represents the power sale of the virtual power plant, P t EV represents the electric vehicles in the virtual power plant, P t TCL represents the temperature control load in the virtual power plant, P t PV represents the photovoltaic power of the electric vehicles in the virtual power plant.

7. The method according to claim 6, wherein The deterministic model also constrains the predicted value of the virtual power plant revenue according to the following formula: P t B -P t S -P t Base = P F Among them, P t Base is the operation curve without considering peak shaving services, and P F is the bid volume in the peak shaving market.

8. The method according to claim 3, characterized in that The robust model specifically includes: Among them, α represents the fluctuation range of uncertain quantities, P represents the decision variable, that is, the day-ahead operation plan of the virtual power plant, B represents the constraint on the decision variable, v represents the actual value of the uncertain variable, B R represents the threshold of the objective function, B0 represents the predicted value of the virtual power plant's revenue, ρ F represents the deviation coefficient of the uncertain variable, represents the average value of the deviation coefficient, and σ represents the robustness factor.

9. The method according to claim 3, wherein The opportunity model specifically includes: Among them, α represents the fluctuation range of uncertain quantities, P represents the decision variable, that is, the day-ahead operation plan of the virtual power plant, B represents the constraints on the decision variable, v represents the actual value of the uncertain variable, B R represents the threshold of the objective function, B0 represents the predicted value of the virtual power plant's revenue, ρ F represents the deviation coefficient of the uncertain variable, represents the average value of the deviation coefficient, and σ represents the robustness factor.

10. The method according to claim 1, characterized in that, The combining of multiple first predicted values and multiple second predicted values according to the preset segmentation criteria includes: Determine the time period in the preset segmentation criteria corresponding to each first predicted value or each second predicted value; Sort each first predicted value and each second predicted value according to the preset segmentation criteria, and output the sorting result as the final predicted value of the virtual power plant peak shaving market aggregated bidding curve.

Images