A virtual power plant day-ahead market bidding optimization method and related device

Abstract

The application belongs to the technical field of virtual power plants, and discloses a virtual power plant day-ahead market bidding optimization method and related devices; wherein the virtual power plant day-ahead market bidding optimization method comprises the following steps: taking the bidding electric quantity, the bidding price of each bidding period and the bidding electric quantity fluctuation of adjacent bidding periods as optimization variables, taking the maximization of benefits as an optimization target, considering the set constraint condition to create an optimization problem; solving the optimization problem to obtain a virtual power plant initial bidding strategy; based on the electric quantity delivery deviation, the bidding electric quantity in the virtual power plant initial bidding strategy is corrected to obtain an optimization bidding strategy of the virtual power plant for the next round of day-ahead market bidding. In the technical scheme disclosed by the application, a two-stage bidding optimization framework improvement means is adopted, accurate bidding and efficient clearing of the VPP in the day-ahead market are realized, the deviation assessment cost is reduced, and the market benefits are improved.

Description

Technical Field

[0001] This invention belongs to the field of virtual power plant technology, and specifically relates to a method and related apparatus for optimizing day-ahead market bidding in virtual power plants. Background Technology

[0002] Virtual power plants (VPPs), as an effective way to aggregate and manage electric vehicles, new energy sources, and energy storage systems, have become important participants in the electricity spot market. Explained, the operational characteristics of the aggregated resources within a VPP create its bidding space, which is a key factor influencing pricing strategies.

[0003] With the development of the electricity market, VPPs need to participate in day-ahead market bidding to achieve optimal resource allocation. However, due to factors such as the volatility of renewable energy output and the randomness of electric vehicle grid connection, accurately predicting the bidding space and formulating reasonable bidding strategies have become the core challenges to improving VPP market returns. Currently, there are still many shortcomings in the methods for predicting and optimizing VPP bidding space, mainly reflected in the following: existing technical solutions mostly adopt single models or static analysis, which are difficult to cope with the complex dynamic characteristics under the aggregation of multiple resources, resulting in a large deviation between the bid and the actual clearing, and increasing the cost of deviation assessment.

[0004] To illustrate, current research on VPP bidding space prediction and strategy optimization mainly focuses on the following areas: Traditional prediction models (such as BP neural networks), while capable of handling a certain scale of data, suffer from insufficient prediction accuracy in small sample situations, and have long training times, making them difficult to adapt to dynamically changing market environments. Specific exemplary existing technical solutions include... Figure 1 As shown. Furthermore, the single-stage bidding model determines the bidding strategy through only one optimization, failing to consider power deviations after market clearing. This can easily lead to significant discrepancies between actual delivery and the winning bid, resulting in high deviation assessment costs. Specific examples of existing technical solutions include... Figure 2 As shown. Furthermore, prediction methods that lack phase space reconstruction directly predict the original time series, failing to uncover hidden temporal correlations within the data, resulting in significant prediction errors and impacting the effectiveness of subsequent bidding strategies. Specific examples of existing technical solutions are shown below. Figure 3 As shown. In summary, the existing technical solutions described above cannot simultaneously meet the requirements of both bidding space prediction accuracy and market clearing deviation control, thus limiting the revenue improvement of VPPs in the day-ahead market. There is an urgent need to develop a new virtual power plant day-ahead market bidding optimization scheme. Summary of the Invention

[0005] The purpose of this invention is to provide a method and related apparatus for optimizing day-ahead market bidding for virtual power plants (VPPs), thereby addressing one or more of the technical problems existing in current VPP day-ahead market bidding, such as low accuracy in predicting bidding space, large discrepancies between bids and clearing prices, and insufficient coordination among multiple resources. The technical solution disclosed in this invention employs a two-stage bidding optimization framework improvement approach, achieving accurate bidding and efficient clearing of VPPs in the day-ahead market, reducing deviation assessment costs, and increasing market revenue.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for optimizing day-ahead market bidding for virtual power plants, comprising the following steps: Based on the selected virtual power plant and the optimization target date, obtain the total available bidding capacity; Using the bid volume, bid price, and bid volume fluctuations in adjacent bidding periods as optimization variables, and maximizing revenue as the optimization objective, an optimization problem is created considering the setting of constraints. Solving this optimization problem yields the initial bidding strategy for the virtual power plant. The revenue is the sum of the products of the bid volume and bid price in each bidding period. The constraints include capacity constraints, where the total available bidding capacity serves as the upper limit constraint on the total bid volume in each bidding period. Based on the initial bidding strategy of the virtual power plant, the virtual power plant's bids and competitors' bids are integrated to obtain a market bid set for the target period. Based on the market bid set, all bids are sorted from low to high price and the cumulative available power is calculated. Based on the cumulative available power and the market demand curve, the winning bid volume of the virtual power plant is determined. Based on the winning bid volume of the virtual power plant, the actual delivery volume of the virtual power plant is generated, taking into account output fluctuations. The difference between the winning bid volume of the virtual power plant and the actual delivery volume is calculated to obtain the power delivery deviation. Based on the power delivery deviation, the bid power in the initial bidding strategy of the virtual power plant is corrected to obtain an optimized bidding strategy for the virtual power plant for the next round of day-ahead market bidding.

[0007] A further improvement of the technical solution of the present invention is that, in the step of obtaining the total bidable capacity based on the selected virtual power plant and the optimization target day, the total bidable capacity is the sum of the photovoltaic predicted output on the optimization target day, the maximum output of electric vehicles the day before the optimization target day, and the maximum charge and discharge power of the battery on the optimization target day.

[0008] A further improvement to the technical solution of the present invention is that the step of obtaining the photovoltaic power forecast for the optimized target day includes: Based on the selected virtual power plant and the optimization target day, obtain the feature vector to be predicted; Based on the feature vector to be predicted, the photovoltaic power output is predicted using a trained Gaussian process regression model to obtain the optimized target day photovoltaic power output. The training steps of the Gaussian process regression model include: acquiring training sample data; constructing a Gaussian process regression model based on the training sample data and selecting a preset kernel function form; using the kernel function scale parameter and noise standard deviation as the model hyperparameters to be optimized; iteratively optimizing the model hyperparameters by maximizing the logarithmic marginal likelihood function; completing the training after satisfying the preset conditions; and obtaining the trained Gaussian process regression model.

[0009] A further improvement to the technical solution of the present invention lies in that, The training sample data includes a feature matrix and a label vector. The feature matrix consists of feature vectors from multiple historical time periods, and each feature vector includes at least the following feature data: the mean photovoltaic output of the day before the target prediction period, the standard deviation of the photovoltaic output of the day before the target prediction period, the maximum value of the photovoltaic output of the day before the target prediction period, the minimum value of the photovoltaic output of the day before the target prediction period, the hour number corresponding to the current prediction period, and the weekday information corresponding to the current prediction date. The label vector consists of actual photovoltaic output values ​​that correspond one-to-one with each feature vector in the feature matrix. The feature vector to be predicted The expression is: In the formula, This represents the average photovoltaic power output of the day before the target date. This represents the standard deviation of photovoltaic power output one day prior to the target date. This indicates the maximum photovoltaic output value on the day before the optimization target date; This represents the minimum photovoltaic power output on the day before the optimization target date; h This indicates the hour number corresponding to the forecast period for the target day; d This indicates the weekday information corresponding to the predicted target date; The preset conditions are one of the following three conditions: the change in the logarithmic marginal likelihood function is less than a preset threshold, the change in the kernel function scaling parameter and the noise standard deviation in adjacent iterations is less than a preset threshold, and the number of iterations reaches a preset maximum number.

[0010] A further improvement of the technical solution of the present invention is that the set constraints also include: price constraints and smoothness constraints; wherein, the price constraint specifically means that the bid price of the virtual power plant in each bidding period is limited to a preset fluctuation range of the predicted electricity price for the corresponding period; the smoothness constraint specifically means that the change range of the bid electricity volume of the virtual power plant in two adjacent bidding periods is within a set range.

[0011] A further improvement to the technical solution of the present invention is that the step of correcting the quoted electricity volume in the initial bidding strategy of the virtual power plant based on the electricity delivery deviation includes: Within multiple bidding periods, the power delivery deviation of virtual power plants is statistically analyzed, and the delivery deviation rate or average deviation level is calculated and used as a delivery deviation indicator. Based on the delivery deviation index, the quoted electricity volume in the initial bidding strategy of the virtual power plant is adjusted; wherein, when the delivery deviation index exceeds the allowable range, the quoted electricity volume for the next round of bidding for the virtual power plant is reduced; when the delivery deviation index is within the allowable range, the quoted electricity volume for the next round of bidding for the virtual power plant remains unchanged.

[0012] A further improvement to the technical solution of the present invention is that, in the step of correcting the bid volume in the initial bidding strategy of the virtual power plant based on the power delivery deviation to obtain an optimized bidding strategy for the virtual power plant for the next round of day-ahead market bidding, before or after correcting the bid volume in the initial bidding strategy of the virtual power plant based on the power delivery deviation, the method further includes: correcting the bid price in the initial bidding strategy of the virtual power plant based on the winning bid volume of the virtual power plant.

[0013] In a second aspect, the present invention provides a virtual power plant day-ahead market bidding optimization system, comprising: The total available bidding capacity acquisition unit is used to acquire the total available bidding capacity based on the selected virtual power plant and the optimization target date; The initial bidding strategy acquisition unit is used to create an optimization problem with the bid electricity volume, bid price, and bid electricity volume fluctuations in adjacent bidding periods as optimization variables, and with maximizing revenue as the optimization objective, considering set constraints; solve the optimization problem to obtain the initial bidding strategy for the virtual power plant; wherein, the revenue is the sum of the products of bid electricity volume and bid price in each bidding period; the set constraints include capacity constraints, wherein the capacity constraints use the total available bidding capacity as the upper limit constraint of the total bid electricity volume in each bidding period; The optimized pricing strategy acquisition unit is used to integrate virtual power plant quotations and competitor quotations based on the initial pricing strategy of the virtual power plant to obtain a market quotation set for the target period; based on the market quotation set, sort all quotations from low to high price and calculate the cumulative available power; based on the cumulative available power and the market demand curve, determine the winning bid volume of the virtual power plant; based on the winning bid volume of the virtual power plant, consider output fluctuations to generate the actual delivery volume of the virtual power plant; calculate the difference between the winning bid volume of the virtual power plant and the actual delivery volume to obtain the power delivery deviation; and based on the power delivery deviation, correct the quoted power volume in the initial pricing strategy of the virtual power plant to obtain an optimized pricing strategy for the virtual power plant for the next round of day-ahead market bidding.

[0014] In a third aspect, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the virtual power plant day-ahead market bidding optimization method as described in any one of the first aspects of the present invention.

[0015] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the virtual power plant day-ahead market bidding optimization method as described in any one of the first aspects of the present invention.

[0016] Compared with the prior art, the present invention has the following beneficial effects: Existing VPP day-ahead market bidding schemes have several shortcomings, primarily including: insufficient accuracy in predicting bidding space; traditional models (such as BP neural networks) are sensitive to small sample data and do not consider the chaotic characteristics of the data, leading to large prediction errors, especially with lag at curve inflection points; poor coordination between bidding and clearing; single-stage models do not dynamically link bidding strategies with the market clearing process, resulting in significant deviations between actual delivered power and winning bids, increasing deviation assessment costs; furthermore, the lack of multi-resource coupling modeling, failing to fully consider the coupling relationships between the power and electricity boundaries of electric vehicles, photovoltaics, and energy storage, making it difficult to accurately characterize the overall bidding space of VPPs; and weak market adaptability, unable to dynamically adjust strategies based on real-time market clearing results, resulting in insufficient responsiveness to electricity price fluctuations and load changes. In view of these problems, this invention discloses a virtual power plant day-ahead market bidding optimization method. Through a two-stage bidding optimization framework, it achieves accurate bidding and efficient clearing of VPPs in the day-ahead market, reducing deviation assessment costs, increasing market revenue, and ultimately improving the bidding efficiency and revenue stability of VPPs in the day-ahead market.

[0017] In a preferred embodiment of the present invention, a bidding space prediction model based on GPR is proposed; wherein, the GPR bidding space prediction method based on phase space reconstruction improves the prediction accuracy under small sample conditions by broadening the time series dimension to mine implicit information.

[0018] In a preferred embodiment of the present invention, a two-stage bidding-clearing collaborative optimization framework is further clarified, which dynamically links the bidding strategy with market clearing and reduces the cost of deviation assessment through power tracking.

[0019] In a preferred embodiment of the present invention, based on multi-resource coupling constraint modeling, the overall bidding space of VPP can be accurately characterized by integrating the power and energy boundaries of electric vehicles, photovoltaics, and energy storage. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of a traditional BP neural network prediction model in the existing technology; Figure 2 This is a schematic diagram of a single-stage bidding model in the existing technology; Figure 3 This is a schematic diagram of a prediction model in the prior art that does not include phase space reconstruction; Figure 4 This is a flowchart illustrating a method for optimizing day-ahead market bidding for virtual power plants, as described in an embodiment of the present invention. Figure 5 This is a schematic diagram illustrating the principle framework of a virtual power plant day-ahead market bidding optimization method in an embodiment of the present invention; Figure 6 This is a schematic diagram of the decomposition of each resource of the VPP in an embodiment of the present invention; Figure 7 This is a schematic diagram of the prediction of maximum photovoltaic power output (on a certain day) in an embodiment of the present invention; Figure 8 This is a scatter plot of actual and predicted values ​​in an embodiment of the present invention; Figure 9 This is a schematic diagram comparing the VPP bidding space with the market results in an embodiment of the present invention; Figure 10 This is a schematic diagram of market price and VPP pricing strategy in an embodiment of the present invention; Figure 11 This is a schematic diagram illustrating the deviation between actual delivery and the winning bid in an embodiment of the present invention; Figure 12 This is a schematic diagram of a virtual power plant day-ahead market bidding optimization system in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention; obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0023] Based on the technical solutions disclosed in the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0024] Please see Figure 4 The present invention provides a method for optimizing day-ahead market bidding for virtual power plants, comprising the following steps: Step 1: Based on the selected virtual power plant and the optimization target day, obtain the total available bidding capacity; in the exemplary preferred technical solution, the total available bidding capacity is the sum of the photovoltaic power output predicted on the optimization target day, the maximum power output of electric vehicles the day before the optimization target day, and the maximum charging and discharging power of the battery on the optimization target day; Step 2: Using the bid electricity volume, bid price, and bid electricity volume fluctuations in adjacent bidding periods as optimization variables, and maximizing revenue as the optimization objective, an optimization problem is created considering the setting of constraints (specifically, for example, using hourly bid electricity volume, hourly bid price, and adjacent hourly bid electricity volume fluctuations as optimization variables, and maximizing revenue as the optimization objective, an optimization problem is created considering the setting of constraints); the optimization problem is solved to obtain the initial bidding strategy for the virtual power plant; wherein, the revenue is the sum of the products of each hourly bid electricity volume and each hourly bid price; the setting of constraints includes capacity constraints, wherein the capacity constraints use the total available bidding capacity as the upper limit constraint of the total hourly bid electricity volume; in an exemplary optional technical solution, an optimization solver (such as the SQP algorithm) is used to solve the optimization problem; Step 3: Based on the initial bidding strategy of the virtual power plant, integrate the virtual power plant's bids with competitors' bids to obtain a market bid set for the target period; based on the market bid set, sort all bids from low to high and calculate the cumulative available power; based on the cumulative available power and the market demand curve, determine the winning bid volume of the virtual power plant; based on the winning bid volume of the virtual power plant, consider output fluctuations to generate the actual delivery volume of the virtual power plant; calculate the difference between the winning bid volume of the virtual power plant and the actual delivery volume to obtain the power delivery deviation; based on the power delivery deviation, correct the bid power in the initial bidding strategy of the virtual power plant to obtain an optimized bidding strategy for the virtual power plant for the next round of day-ahead market bidding.

[0025] In the specific example of the technical solution, the process for determining the clearing price and the winning bid volume for the virtual power plant is as follows: Construct a set of market quotes for the target time period; where, On the day before the optimization target date t Within a given time period, obtain quotes from virtual power plants. And the tiered pricing strategy of all competitors ;in, For virtual power plants in the first t The quoted price for a given time period is the transaction price declared by the virtual power plant during that time period. For virtual power plants in the first t The quoted electricity volume for a given period, i.e., the transaction volume declared by the virtual power plant during that period; i To identify competing entities (distinguishing different competitors, such as other power generation companies, electricity sales companies, etc.); k This is a tiered indicator (corresponding to different levels of "tiered pricing"); For the first i The first competitive entity in the... t Time period, number k Tiered pricing; For the first i The first competitive entity in the... t Time period, number k Tiered pricing for electricity consumption; The virtual power plant bids are combined with all tiered bids from each competitor to form the market bid set for that period.

[0026] Based on the market price set, a sorting process is performed; specifically, the market price set is sorted in ascending order of price to obtain a sorted price sequence.

[0027] Calculate the cumulative available electricity curve; wherein, according to the sorted price sequence, the electricity in the price sequence is added up item by item to calculate the cumulative available electricity at the corresponding price, and the cumulative available electricity curve is constructed.

[0028] Construct a market demand curve; whereby, based on the maximum and minimum demand values ​​set in the market before the current day, a linear function is used to construct the market demand curve for the target period, so that the quantity demanded decreases as the price increases.

[0029] Determine the market clearing price; among which, the cumulative available electricity curve is matched with the market demand curve to determine the quoted price corresponding to the first time when the cumulative available electricity is not less than the market demand, which is used as the market clearing price for that period.

[0030] Determine the winning bid volume of the virtual power plant; among which, compare the bid price of the virtual power plant with the clearing price: if the bid price of the virtual power plant is not higher than the clearing price, the corresponding bid volume is determined as the winning bid volume; otherwise, the winning bid volume of the virtual power plant for that period is zero.

[0031] In a specific exemplary technical solution, the process for generating actual delivery volume and calculating deviation assessment costs specifically includes: The actual delivered power of the virtual power plant is generated. Under the premise that the virtual power plant wins the bid, a random disturbance coefficient is introduced to characterize the fluctuation of the actual output, and the winning power is corrected to obtain the actual delivered power of the virtual power plant in that period.

[0032] Calculate the power delivery deviation; specifically, by comparing the power volume won in the virtual power plant bid with the actual power volume delivered, the difference between the two is calculated as the power delivery deviation for that period.

[0033] The deviation type is determined as follows: when the actual delivered electricity is less than the bid-winning electricity, it is determined as under-delivery deviation; when the actual delivered electricity is greater than the bid-winning electricity, it is determined as over-delivery deviation.

[0034] Calculate the deviation assessment cost; wherein, according to the determined deviation type, the calculation is carried out according to the preset deviation assessment rules: for underpayment deviation, the deviation assessment cost is calculated according to the penalty coefficient of the clearing price; for overpayment deviation, the corresponding settlement amount is calculated according to the settlement coefficient of the clearing price, thereby obtaining the deviation assessment cost.

[0035] Calculate the actual net revenue of the virtual power plant; the theoretical revenue is obtained by multiplying the winning bid volume of the virtual power plant by the market clearing price, and the deviation assessment cost is deducted to obtain the actual net revenue of the virtual power plant for that period.

[0036] In a specific exemplary technical solution, the pricing strategy adjustment process based on delivery results is as follows: Statistical analysis of historical delivery deviation indicators; in particular, statistical analysis of delivery deviations of virtual power plants across multiple time periods, calculation of delivery deviation rate or average deviation level, as a basis for adjusting pricing strategies.

[0037] Adjust the quoted electricity volume of virtual power plants; when the delivery deviation index exceeds the preset threshold, the quoted electricity volume of virtual power plants will be reduced; when the delivery deviation index is within the allowable range, the original quoted electricity volume will remain unchanged.

[0038] Adjust the bidding prices of virtual power plants; in particular, based on market clearing results and historical bidding data, the bidding prices of virtual power plants will be revised to limit them within the preset fluctuation range of the predicted electricity price.

[0039] A final bidding strategy for the virtual power plant is formed; wherein, the obtained revised bid volume and the obtained revised bid price are combined to form the final bidding strategy for the virtual power plant for the next round of day-ahead market bidding.

[0040] In the technical solution disclosed in this invention, the VPP bidding space prediction method broadens the time series dimension through phase space reconstruction and uses Gaussian process regression to predict the bidding space parameters for the next day. The two-stage VPP day-ahead bidding optimization method optimizes the bidding strategy based on the predicted bidding space in the first stage, and achieves power tracking through a market clearing model in the second stage, ultimately reducing deviation assessment costs.

[0041] In an optional technical solution of this invention embodiment, in step 1, the optimized target day photovoltaic power output is obtained based on a trained GPR model, and the specific steps are as follows: Step 1.1: Based on the selected virtual power plant and the optimization target day, obtain the feature vector to be predicted; Step 1.2: Based on the feature vector to be predicted obtained in Step 1.1, the photovoltaic output is predicted using the trained GPR model to obtain the optimized target day photovoltaic output. The feature vector to be predicted is represented as follows: ; In the formula, This represents the average photovoltaic power output of the day before the target date. This represents the standard deviation of photovoltaic power output one day prior to the target date. This indicates the maximum photovoltaic output value on the day before the optimization target date; This represents the minimum photovoltaic power output on the day before the optimization target date; h This indicates the hour number corresponding to the forecast period for the target day; d This indicates the weekday information corresponding to the predicted date of the target date.

[0042] In this embodiment of the invention, the training steps of the GPR model include: Obtain training sample data; wherein, the training sample data includes a feature matrix and a label vector; Based on the acquired training sample data, a Gaussian process regression model is constructed. A preset kernel function is selected, and the kernel function scale parameter and noise standard deviation are used as model hyperparameters to be optimized. The hyperparameters are iteratively optimized by maximizing the logarithmic marginal likelihood function to obtain a trained GPR model.

[0043] Explained in practice, the training sample data comes from the historical operating data of the virtual power plant, requiring no additional external data. During the training phase, the feature matrix consists of feature vectors from multiple historical time periods. Each feature vector includes at least the following feature data: the mean photovoltaic output of the day before the target prediction period; the standard deviation of the photovoltaic output of the day before the target prediction period; the maximum photovoltaic output of the day before the target prediction period; the minimum photovoltaic output of the day before the target prediction period; the hour number corresponding to the current prediction period; and the weekday information corresponding to the current prediction date. Explained, all of the above feature data can be calculated from the historical photovoltaic output data and timestamp information of the virtual power plant.

[0044] In addition, the label vector is composed of the actual photovoltaic power output value that corresponds one-to-one with each feature vector in the feature matrix, specifically: the actual photovoltaic power output value corresponding to the prediction period.

[0045] The feature vectors corresponding to multiple historical dates and time periods are combined in chronological order to form a training feature matrix; the actual photovoltaic power output values ​​of the corresponding time periods are combined to form a training label vector, thus forming a training sample dataset for training the GPR model.

[0046] In this embodiment of the invention, training is completed when preset conditions are met during iterative training. These preset conditions include at least one of the following: during GPR model training, the change in the logarithmic marginal likelihood function is less than a preset threshold; the changes in the kernel scaling parameter and noise standard deviation in adjacent iterations are less than preset thresholds; and the number of training iterations reaches a preset maximum number of iterations. Explained, when any of the above conditions is met, model training is stopped, and a trained GPR model is obtained.

[0047] In the optional technical solutions of the embodiments of the present invention, the setting of constraints further includes: price constraints, energy balance constraints and smoothness constraints, so as to ensure the rationality and feasibility of the pricing strategy.

[0048] In this embodiment of the invention, the price constraint specifically means that the price quoted by the virtual power plant in each bidding period should be limited to a preset fluctuation range of the predicted electricity price for the corresponding period, expressed as: ; In the formula, Indicates the virtual power plant in the first... t Price quotes for each time period; Indicates the first t Market-forecasted electricity prices for each time period; This is a price fluctuation factor used to limit the range of fluctuation in quoted prices.

[0049] In this embodiment of the invention, the energy balance constraint specifically means that the total bid-based electricity volume of a virtual power plant during the optimization period should not exceed the upper limit of its available energy for dispatchable resources during the corresponding period, expressed as: ; In the formula, Indicates the virtual power plant in the first... t Electricity volume quoted for each time period; Indicates the first t The photovoltaic forecast available output for each time period; Indicates the first t The maximum dispatchable output of electric vehicles after aggregation in a given time period; Indicates the first t The maximum discharge power of the energy storage system during a given time period; T represents the total number of time periods in the optimization cycle.

[0050] In this embodiment of the invention, the smoothness constraint specifically refers to: to avoid drastic fluctuations in the bid electricity volume of the virtual power plant in adjacent time periods, limiting the range of change in the bid electricity volume between two adjacent time periods, expressed as: ; In the formula, and These represent the quoted electricity volumes for two adjacent time periods; This indicates the maximum allowable power change threshold.

[0051] Please see Figure 5 The technical solution disclosed in the embodiments of the present invention has the following specific exemplary process: Step 1: Parameter setting and basic data generation; In this step, the time dimension parameters of the simulation analysis and the resource types and quantities contained in the virtual power plant are set. Based on the set parameters, the output data of various resources in the virtual power plant, the market price data of the corresponding time period, and the bidding data of competitors are generated through the preset mathematical model, which serve as the basic input data for subsequent predictive modeling and bidding optimization.

[0052] Step 2: Training the prediction model and predicting key parameters based on historical data; In this step, based on the historical data generated or collected in step 1, the training set and the test set are divided in chronological order; a feature matrix and label vector corresponding to the operating characteristics of the virtual power plant are constructed to train the prediction model; the trained prediction model is used to predict the key parameters of the optimization target day, and the accuracy of the prediction results is evaluated.

[0053] Step 3: Bidding optimization and market clearing simulation based on prediction results; In this step, based on the prediction results obtained in step 2, the capacity of various dispatchable resources within the virtual power plant is aggregated to determine the bidding space of the virtual power plant on the optimization target day. On this basis, a bidding optimization model with the goal of maximizing the revenue of the virtual power plant is established and solved to obtain the optimal bidding strategy of the virtual power plant. According to the optimal bidding strategy, the day-ahead market clearing process is simulated to calculate the winning bid volume, actual delivered volume, and corresponding deviation assessment cost of the virtual power plant, thereby determining the actual net revenue of the virtual power plant.

[0054] Step 4: Results Output and Performance Evaluation; In this step, the optimization results, such as the virtual power plant's pricing strategy, market clearing results, and net revenue, are output; and the performance of the method is comprehensively evaluated based on indicators such as prediction error, delivery deviation, and revenue level.

[0055] In this embodiment of the invention, to verify the feasibility and effectiveness of the virtual power plant day-ahead bidding method of the present invention, a simulation verification is performed using an RBTS38-node distribution system containing four virtual power plants as a case study. The specific process is as follows: Step 1, set the simulation time dimension parameters: num_days = 365 (simulating a one-year time span), num_hours = 24 (dividing each day into 24 hours). Set the resource type quantity parameters: num_resources = 3, explicitly considering three types of distributed resources: electric vehicles (EV), photovoltaics (PV), and batteries. The VPP resource decomposition is as follows: Figure 6 As shown.

[0056] In this step, the resource output data is generated as follows: In generating electric vehicle (EV) output data, the minimum output matrix for EVs is set as follows: ; The maximum available output of an electric vehicle is generated in the following way: ; In the formula, d represents the d-th day; h represents the h-th hour; This indicates the average charging and discharging power of an electric vehicle. This represents the standard deviation of the power output fluctuation of electric vehicles; This represents a standard normally distributed random variable used to simulate the randomness and uncertainty of the aggregate power output of electric vehicles over time.

[0057] Explained, the above formula is based on a normal distribution to simulate the random fluctuations of EV charging power, and the max function is used to ensure that the output value is non-negative, thereby reflecting the uncertainty of EV charging power.

[0058] (2) In the generation of photovoltaic (PV) power output data, the effective power output period of PV is set as: ; Construct seasonal factors to describe the effects of seasonal variation: ; Constructing the power output curve of a photovoltaic base: ; Introducing a random perturbation term: ; The maximum available output of photovoltaic power is expressed as: ; In the formula, a and b are both seasonal adjustment coefficients; The noise level of photovoltaic output is represented. Interpretatively, the above model is used to represent the intraday periodicity, seasonality, and random fluctuations of photovoltaic output.

[0059] (3) In the generation of battery output data, the upper and lower limits of battery charging and discharging power are set as follows: ; ; In the formula, negative values ​​represent charging power, and positive values ​​represent discharging power.

[0060] (4) During the generation of electricity price data, a daily cycle electricity price trend is constructed: ; After superimposing hourly fluctuations and random disturbances, the electricity price series is obtained: ; In the formula, This indicates hourly price fluctuations. This represents the random noise term; the lower limit for electricity price is set at $10 / MWh.

[0061] (5) Competitor pricing data generation: Three competitors are set up, and each competitor adopts a tiered pricing strategy. The pricing volume tiers are as follows: The corresponding quoted price is generated by adding random perturbations to the base price, and its minimum value is limited to no less than $15 / MWh.

[0062] Step 2, GPR model training and bidding space prediction.

[0063] First, the dataset is divided into a training set and a test set. The training set is the first... Heaven, that is The test set consists of the last 7 days, i.e. This is used for subsequent model training and performance verification.

[0064] Training data construction and calculation of the number of training samples: ; Extract the previous day's photovoltaic power output data: ; Construct a feature matrix and a label vector. The features include the mean, standard deviation, maximum, minimum, current hour, and day of the week of the previous day's PV output. Complete the construction of the matrix and label vector by iterating through each day and each hour.

[0065] GPR model training, calculating the kernel scale parameter kernel_scale=std(X) train (:)) and noise standard deviation .

[0066] Configure training options ,set up (Turn off display) (Maximum number of iterations) Then train the GPR model using the 'fitrgp' function.

[0067] Test set prediction, similar to the training set construction process, involves building the test set feature matrix X. train and label vector The trained GPR model is used for prediction to obtain the prediction results. X train .

[0068] For predictive performance evaluation, the mean squared error (MSE), root mean squared error (RMSE), and mean absolute error (MAE) are calculated using the following formulas: ; ; ; Output photovoltaic power generation forecast indicators such as RMSE and MAE, etc. Figure 7 and Figure 8 As shown.

[0069] Step 3, two-stage bidding optimization.

[0070] The first step is capacity aggregation to determine the target optimization date. and the previous day Extract the predicted output values ​​of photovoltaics, electric vehicles, and batteries. Contribute to the photovoltaic forecast for the target date. To maximize the output of the EVs from the previous day, Given the target day's maximum charge / discharge power of the battery, calculate the total available bidding capacity. This serves as an upper limit constraint on the subsequent bidding volume.

[0071] The first phase of optimization involves formulating a pricing strategy; among which... Create optimization problem The clear objective is to maximize profits. Optimization variables are defined as follows: quantity (Electricity consumption quoted per hour, minimum is 0) price (Hourly quoted price, minimum price is 0) delta (Electricity price fluctuations between adjacent hours, with a lower limit of 0). Set the objective function. That is, to maximize the sum of the products of the quoted electricity volume and the price.

[0072] Add constraints, the first of which is a capacity constraint: bid_opt.Constraints.capacity_limit=quantity≤total_capacity_pred; Ensure that the quoted electricity volume does not exceed the available capacity.

[0073] The second is price constraint: ; The electricity price for the day is predicted, and the quoted price is limited to fluctuating within ±10% of the predicted electricity price.

[0074] The third is the energy balance constraint: ; The total quoted electricity volume shall not exceed the sum of the base capacity and the energy storage capacity.

[0075] The fourth is the smoothness constraint: through delta The variable constrains the fluctuation of quoted electricity volume between adjacent hours; the formula is as follows: ; ; ; Configure the optimize solver options and set the algorithm to... sqp The algorithm has a maximum iteration limit of 200 times and a maximum function evaluation limit of 5000 times. By solving the optimization problem, it outputs the optimal electricity quotation and the optimal price quotation.

[0076] After obtaining the virtual power plant bidding strategy from the first phase of bidding optimization, market clearing simulations are performed for each time period on the optimization target day.

[0077] For each time period, virtual power plant bids and competitor bids are integrated to construct a total market bid pool. The total market bid pool is then sorted in ascending order of price, and the corresponding cumulative available power is calculated. A market demand curve is constructed, such that market demand linearly decreases from a preset maximum demand value to a minimum demand value as the price increases. By matching the cumulative available power with the market demand curve, the closest intersection price is determined as the market clearing price for that time period. Based on the relative position of the clearing price and the virtual power plant bids, the winning bid volume for each virtual power plant in each time period is determined: when a virtual power plant bid is before the clearing point, its corresponding bid volume is determined as the winning bid volume; when a virtual power plant bid is after the clearing point, its corresponding winning bid volume is zero.

[0078] Based on the awarded electricity volume obtained above, a disturbance factor is introduced to characterize the fluctuation of the virtual power plant's actual output, and the awarded electricity volume is corrected to obtain the actual delivered electricity volume of the virtual power plant in each time period. By comparing the actual delivered electricity volume with the awarded electricity volume, the delivery deviation of the virtual power plant in each time period is calculated and classified as under-generation deviation or over-generation deviation. On this basis, according to the preset deviation assessment rules, the corresponding deviation assessment cost is calculated, and the actual net profit of the virtual power plant in that time period is further calculated.

[0079] Based on the delivery deviation and deviation assessment costs obtained above, the original bidding strategy for virtual power plants is adjusted, specifically including: when the delivery deviation rate exceeds the preset threshold, the volume of subsequent bids for virtual power plants is reduced; when the deviation assessment costs are high, the bid prices of virtual power plants are revised to bring them closer to the market clearing price range; when virtual power plants fail to win bids for multiple consecutive periods, the bid prices are appropriately reduced to increase the probability of winning bids.

[0080] In this embodiment of the invention, the adjusted pricing strategy described above is used to re-participate in the market clearing process, and a comparative analysis is conducted between this adjusted strategy and the pricing strategy without adjustment. For example... Figures 9 to 11 As shown in the comparison results, the bidding strategy based on GPR prediction and two-stage bidding optimization has significant advantages in the following aspects: Under the adjusted bidding strategy, the winning bid volume of the virtual power plant is more matched with the actual delivered volume, and the delivery deviation rate is significantly reduced; the deviation assessment cost is significantly reduced, reducing the economic losses caused by the uncertainty of output of the virtual power plant; while ensuring the probability of winning the bid, the overall net income level of the virtual power plant is improved.

[0081] The technical solution disclosed in this invention, by introducing a pricing strategy adjustment mechanism based on prediction results and actual delivery feedback, achieves significant effects in at least the following aspects: (1) Improve the adaptability of the pricing strategy; Explain that the pricing strategy can be dynamically adjusted according to the market clearing results and delivery deviations, avoiding the risk of systemic deviation caused by a one-time pricing.

[0082] (2) Reduce deviation assessment costs; Explain, through the second-stage delivery deviation feedback mechanism, the actual delivered electricity volume is kept highly consistent with the winning bid electricity volume, effectively reducing deviation assessment costs.

[0083] (3) Improve the market revenue of virtual power plants; Explain, under the premise of ensuring deliverability, achieve synergistic optimization of the probability of winning the bid and the revenue of the bid, thereby improving the overall revenue level of virtual power plants in the day-ahead market.

[0084] (4) Enhance stability in multi-resource scenarios; interpretably, the bidding strategy can be dynamically adjusted according to changes in the internal resource structure of the virtual power plant, which is applicable to multiple resource combination scenarios.

[0085] (5) The simulation results verified the comprehensive advantages of the bidding method based on GPR prediction and two-stage bidding-clearing-feedback adjustment mechanism in terms of prediction accuracy, deviation control and economy.

[0086] To further explain, in addition to the two-stage optimization framework adopted in this invention, similar goals can be achieved through the following alternatives: Replacing GPR with an LSTM neural network, utilizing a long short-term memory network to process time-series data, suitable for large-sample scenarios, but with slightly lower prediction accuracy for small samples. Replacing two-stage optimization with Nash game theory, achieving bidding strategy optimization through multi-VPP non-cooperative game theory, suitable for multi-agent competition scenarios, but with higher computational complexity. Introducing reinforcement learning to dynamically adjust parameters, optimizing GPR hyperparameters in real time through deep reinforcement learning, improving the model's adaptability to market changes, but with higher training costs.

[0087] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not disclosed in the apparatus embodiments, please refer to the embodiments of the method of the present invention.

[0088] Please see Figure 12 In this embodiment of the invention, a virtual power plant day-ahead market bidding optimization system is provided, comprising: The total available bidding capacity acquisition unit is used to acquire the total available bidding capacity based on the selected virtual power plant and the optimization target date; The initial bidding strategy acquisition unit is used to create an optimization problem with the bid electricity volume, bid price, and bid electricity volume fluctuations in adjacent bidding periods as optimization variables, and with maximizing revenue as the optimization objective, considering set constraints; solve the optimization problem to obtain the initial bidding strategy for the virtual power plant; wherein, the revenue is the sum of the products of bid electricity volume and bid price in each bidding period; the set constraints include capacity constraints, wherein the capacity constraints use the total available bidding capacity as the upper limit constraint of the total bid electricity volume in each bidding period; The optimized pricing strategy acquisition unit is used to integrate virtual power plant quotations and competitor quotations based on the initial pricing strategy of the virtual power plant to obtain a market quotation set for the target period; based on the market quotation set, sort all quotations from low to high price and calculate the cumulative available power; based on the cumulative available power and the market demand curve, determine the winning bid volume of the virtual power plant; based on the winning bid volume of the virtual power plant, consider output fluctuations to generate the actual delivery volume of the virtual power plant; calculate the difference between the winning bid volume of the virtual power plant and the actual delivery volume to obtain the power delivery deviation; and based on the power delivery deviation, correct the quoted power volume in the initial pricing strategy of the virtual power plant to obtain an optimized pricing strategy for the virtual power plant for the next round of day-ahead market bidding.

[0089] In one embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used to execute the operation of a virtual power plant day-ahead market bidding optimization method.

[0090] In one embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the operating system of the terminal. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor, which can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM (Random Access Memory) or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the virtual power plant day-ahead market bidding optimization method in the above embodiments.

[0091] 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, optical storage, etc.) containing computer-usable program code.

[0092] 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.

[0093] 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 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0094] 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.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing day-ahead market bidding for virtual power plants, characterized in that, Includes the following steps: Based on the selected virtual power plant and the optimization target date, obtain the total available bidding capacity; Using the bid volume, bid price, and bid volume fluctuations in adjacent bidding periods as optimization variables, and maximizing revenue as the optimization objective, an optimization problem is created considering the setting of constraints. Solving this optimization problem yields the initial bidding strategy for the virtual power plant. The revenue is the sum of the products of the bid volume and bid price in each bidding period. The constraints include capacity constraints, where the total available bidding capacity serves as the upper limit constraint on the total bid volume in each bidding period. Based on the initial bidding strategy of the virtual power plant, the virtual power plant's bids and competitors' bids are integrated to obtain a market bid set for the target period. Based on the market bid set, all bids are sorted from low to high price and the cumulative available power is calculated. Based on the cumulative available power and the market demand curve, the winning bid volume of the virtual power plant is determined. Based on the winning bid volume of the virtual power plant, the actual delivery volume of the virtual power plant is generated, taking into account output fluctuations. The difference between the winning bid volume of the virtual power plant and the actual delivery volume is calculated to obtain the power delivery deviation. Based on the power delivery deviation, the bid power in the initial bidding strategy of the virtual power plant is corrected to obtain an optimized bidding strategy for the virtual power plant for the next round of day-ahead market bidding.

2. The virtual power plant day-ahead market bidding optimization method according to claim 1, characterized in that, In the step of obtaining the total available bidding capacity based on the selected virtual power plant and the optimization target date, the total available bidding capacity is the sum of the photovoltaic power output predicted on the optimization target date, the maximum power output of electric vehicles the day before the optimization target date, and the maximum charge and discharge power of the battery on the optimization target date.

3. The virtual power plant day-ahead market bidding optimization method according to claim 2, characterized in that, The steps for obtaining the predicted photovoltaic power output for the target day include: Based on the selected virtual power plant and the optimization target day, obtain the feature vector to be predicted; Based on the feature vector to be predicted, the photovoltaic power output is predicted using a trained Gaussian process regression model to obtain the optimized target day photovoltaic power output. The training steps of the Gaussian process regression model include: acquiring training sample data; constructing a Gaussian process regression model based on the training sample data and selecting a preset kernel function form; using the kernel function scale parameter and noise standard deviation as the model hyperparameters to be optimized; iteratively optimizing the model hyperparameters by maximizing the logarithmic marginal likelihood function; completing the training after satisfying the preset conditions; and obtaining the trained Gaussian process regression model.

4. The virtual power plant day-ahead market bidding optimization method according to claim 3, characterized in that, The training sample data includes a feature matrix and a label vector. The feature matrix consists of feature vectors from multiple historical time periods, and each feature vector includes at least the following feature data: the mean photovoltaic output of the day before the target prediction period, the standard deviation of the photovoltaic output of the day before the target prediction period, the maximum value of the photovoltaic output of the day before the target prediction period, the minimum value of the photovoltaic output of the day before the target prediction period, the hour number corresponding to the current prediction period, and the weekday information corresponding to the current prediction date. The label vector consists of actual photovoltaic output values ​​that correspond one-to-one with each feature vector in the feature matrix. The feature vector to be predicted The expression is: In the formula, This represents the average photovoltaic power output of the day before the target date. This represents the standard deviation of photovoltaic power output one day prior to the target date. This indicates the maximum photovoltaic output value on the day before the optimization target date; This represents the minimum photovoltaic power output on the day before the optimization target date; h This indicates the hour number corresponding to the forecast period for the target day; d This indicates the weekday information corresponding to the predicted target date; The preset conditions are one of the following three conditions: the change in the logarithmic marginal likelihood function is less than a preset threshold, the change in the kernel function scaling parameter and the noise standard deviation in adjacent iterations is less than a preset threshold, and the number of iterations reaches a preset maximum number.

5. The virtual power plant day-ahead market bidding optimization method according to claim 1, characterized in that, The set constraints also include: price constraints and smoothness constraints; wherein, the price constraint specifically means that the price quoted by the virtual power plant in each bidding period is limited to a preset fluctuation range of the predicted electricity price for the corresponding period; the smoothness constraint specifically means that the change range of the quoted electricity volume between two adjacent bidding periods of the virtual power plant is within a set range.

6. The virtual power plant day-ahead market bidding optimization method according to claim 1, characterized in that, The steps for correcting the quoted electricity volume in the initial pricing strategy of the virtual power plant based on the electricity delivery deviation include: Within multiple bidding periods, the power delivery deviation of virtual power plants is statistically analyzed, and the delivery deviation rate or average deviation level is calculated and used as a delivery deviation indicator. Based on the delivery deviation index, the quoted electricity volume in the initial bidding strategy of the virtual power plant is adjusted; wherein, when the delivery deviation index exceeds the allowable range, the quoted electricity volume for the next round of bidding for the virtual power plant is reduced; when the delivery deviation index is within the allowable range, the quoted electricity volume for the next round of bidding for the virtual power plant remains unchanged.

7. The virtual power plant day-ahead market bidding optimization method according to claim 1, characterized in that, In the step of correcting the bid volume in the initial bidding strategy of the virtual power plant based on the power delivery deviation to obtain an optimized bidding strategy for the virtual power plant for the next round of day-ahead market bidding, before or after correcting the bid volume in the initial bidding strategy of the virtual power plant based on the power delivery deviation, the method further includes: correcting the bid price in the initial bidding strategy of the virtual power plant based on the winning bid volume of the virtual power plant.

8. A virtual power plant day-ahead market bidding optimization system, characterized in that, include: The total available bidding capacity acquisition unit is used to acquire the total available bidding capacity based on the selected virtual power plant and the optimization target date; The initial bidding strategy acquisition unit is used to create an optimization problem with the bid electricity volume, bid price, and bid electricity volume fluctuations in adjacent bidding periods as optimization variables, and with maximizing revenue as the optimization objective, considering set constraints; solve the optimization problem to obtain the initial bidding strategy for the virtual power plant; wherein, the revenue is the sum of the products of bid electricity volume and bid price in each bidding period; the set constraints include capacity constraints, wherein the capacity constraints use the total available bidding capacity as the upper limit constraint of the total bid electricity volume in each bidding period; The optimized pricing strategy acquisition unit is used to integrate virtual power plant quotations and competitor quotations based on the initial pricing strategy of the virtual power plant to obtain a market quotation set for the target period; based on the market quotation set, sort all quotations from low to high price and calculate the cumulative available power; based on the cumulative available power and the market demand curve, determine the winning bid volume of the virtual power plant; based on the winning bid volume of the virtual power plant, consider output fluctuations to generate the actual delivery volume of the virtual power plant; calculate the difference between the winning bid volume of the virtual power plant and the actual delivery volume to obtain the power delivery deviation; and based on the power delivery deviation, correct the quoted power volume in the initial pricing strategy of the virtual power plant to obtain an optimized pricing strategy for the virtual power plant for the next round of day-ahead market bidding.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the virtual power plant day-ahead market bidding optimization method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the virtual power plant day-ahead market bidding optimization method as described in any one of claims 1 to 7.

Images