Day-ahead bidding decision-making method for flow battery energy storage power station

Abstract

The invention discloses a day-ahead bidding decision-making method for a flow battery energy storage power station, and the method comprises the steps: carrying out the compression of a historical frequency modulation signal scene, obtaining a typical scene, obtaining an objective function of the condition value-at-risk constraint of a risk aversion coefficient under the typical scene of the energy storage power station, and finally building the objective function of the condition value-at-risk constraint of the risk aversion coefficient. And obtaining the bidding discharge power of the energy storage power station in the day-ahead typical scene of the energy market, the bidding charge power of the energy storage power station in the day-ahead typical scene of the energy market, and the bidding capacity of the energy storage power station in the day-ahead typical scene of the frequency modulation market so as to assist in bidding decision making. According to the bidding strategy of the objective function based on the conditional value-at-risk constraint, the prediction accuracy of the current bidding power and the bidding capacity is high, the bidding risk caused by the uncertainty of the participation of the flow battery energy storage power station in the energy-frequency modulation market is reduced, and the efficiency of the participation of the flow battery energy storage power station in the electricity market is improved.

Description

Technical Field

[0001] This invention relates to the field of electricity market technology, and in particular to a day-ahead bidding decision-making method for flow battery energy storage power stations. Background Technology

[0002] With the increasing proportion of renewable energy generation in new power systems, stable grid operation has become a significant challenge. Flow batteries, as electrochemical energy storage systems characterized by fast response, high safety, long lifespan, and large-scale application, currently play a crucial role in ancillary service scenarios such as peak shaving and smoothing renewable energy consumption. With power market reforms, flow battery energy storage power stations are participating not only in the electricity market but also in the frequency regulation market; however, market participation requires prior bidding.

[0003] However, when energy storage power stations participate in electricity market transactions, they face problems such as market price fluctuations and uncertain frequency regulation signals, resulting in low accuracy in predicting day-ahead bidding power and capacity, and thus higher risks in day-ahead bidding. Summary of the Invention

[0004] This invention discloses a day-ahead bidding decision-making method for flow battery energy storage power stations to overcome the above-mentioned technical problems.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows: A day-ahead bidding decision-making method for flow battery energy storage power stations includes the following steps: S1: Obtain historical electricity market clearing prices and use an attention-based GRU neural network to obtain predicted day-ahead electricity market clearing prices; S2: Obtain historical FM signal scenarios, and use the probabilistic distance fast reduction method to obtain several typical scenarios and the probability of their occurrence; S3: Based on the predicted day-ahead electricity market clearing price, obtain the expected day-ahead revenue of the energy storage power station under typical scenarios; S4: Obtain the day-to-day operating cost of an energy storage power station in a typical scenario; S5: Based on the probability of typical scenarios occurring, the expected daily revenue of the energy storage power station under typical scenarios, and the daily operating cost of the energy storage power station under typical scenarios, obtain the conditional value of risk of the expected daily profit of the energy storage power station. S6: Based on the day-ahead expected revenue, day-ahead operating cost, and day-ahead expected profit of energy storage power stations in typical scenarios, establish an objective function constrained by the conditional value-at-risk (VAT) factor, considering the risk aversion coefficient, to obtain the bidding discharge power, bidding charging power, and bidding capacity of energy storage in typical day-ahead scenarios in the energy market; and use this VAT factor to assist in formulating bidding decision-making schemes.

[0006] Furthermore, the formula used to obtain the expected day-ahead return of an energy storage power station under typical scenarios is as follows:

[0007] In the formula: This indicates that energy storage power stations were recently... The first time period Expected returns for a typical scenario; This indicates that energy storage power stations are currently in the energy market. The first time period Expected returns for a typical scenario; This indicates that energy storage power stations are currently in the secondary frequency regulation market. The first time period Expected returns for a typical scenario; Index numbers for typical scenarios; in,

[0008]

[0009] In the formula: For the predicted date The first time period Clearing prices for a typical scenario; For energy storage power stations in the energy market day The first time period Discharge power for bidding in a typical scenario; For energy storage power stations in the energy market day The first time period The bidding charging power for a typical scenario; Indicates the length of the time period; For the day before The first time period The clearing price of frequency modulation capacity in a typical scenario; For the day before The first time period The clearing price of FM mileage in a typical scenario; For energy storage in the frequency regulation market The first time period Bidding capacity for a typical scenario; For energy storage The first time period Frequency modulation mileage in a typical scenario; This refers to the overall performance indicators of frequency modulation.

[0010] Furthermore, the formula used to obtain the current operating cost of an energy storage power station under typical scenarios is as follows:

[0011] In the formula: It indicates that the energy storage power station was recently at the A typical scenario Operating costs during a given time period; This indicates the investment cost of an energy storage power station; It indicates that the energy storage power station was recently at the A typical scenario Charging and discharging loss costs over a given period of time; It indicates that the energy storage power station was recently at the A typical scenario The lifespan degradation cost caused by the depth of charge and discharge over a given period; in,

[0012] In the formula: Cost per unit capacity of energy storage power station; This refers to the rated capacity of the energy storage power station. r The discount rate; For the float charging life of energy storage power stations; This represents the total number of bidding periods;

[0013]

[0014]

[0015] In the formula: This indicates the day-ahead charge and discharge loss cost of an energy storage power station participating in the energy market; This represents the day-ahead charge and discharge loss cost of an energy storage power station participating in the secondary frequency regulation market; Indicates the discharge identifier variable of the energy storage power station; This is represented as a charging identifier variable for energy storage power stations; Indicates the discharge efficiency of the energy storage power station; Indicates the charging efficiency of the energy storage power station; Indicates the electricity price for losses; express Accumulated value of positive frequency modulation signal over a given period; express Accumulated value of negative frequency modulation signal over a given period;

[0016] in:

[0017]

[0018] In the formula: Indicates the first Energy storage power stations in typical scenarios State of charge over a period of time; Indicates the first Energy storage power stations in typical scenarios State of charge over a period of time; The parameters representing the fitted energy storage characteristics; This indicates the unit power cost of an energy storage power station; Indicates the rated power of the energy storage power station; This indicates the number of energy storage cycles when the energy is charged and discharged at 100% depth. express t The change in electricity consumption resulting from time-of-use energy storage participating in frequency regulation; This indicates the rated energy of the energy storage power station.

[0019] Furthermore, the formula used to obtain the conditional value at risk of the expected day-ahead profit of an energy storage power station is as follows:

[0020]

[0021] In the formula: Value at risk (VAT) represents the conditional risk of the expected profit of an energy storage power station. To be at the risk confidence level Value at risk below; Risk confidence level; This represents the total number of typical scenarios; Index numbers for typical scenarios; For the first The probability of a typical scenario occurring; For the current bidding period, This represents the total number of bidding periods; Indicates intermediate calculation parameters; Represents unknown variables;

[0022]

[0023] In the formula: Indicates the first The loss exceeds the threshold in a typical scenario; Represents the loss boundary variable; This represents the penalty factor.

[0024] Furthermore, the objective function for the conditional value-at-risk constraint is established as follows:

[0025] In the formula: Risk aversion coefficient; Value at risk is the conditional risk of the expected profit of an energy storage power station.

[0026] Beneficial Effects: This invention provides a day-ahead bidding decision-making method for flow battery energy storage power stations. By compressing historical frequency regulation signal scenarios, typical scenarios are obtained, and then the bidding capacity of the energy storage power station under these typical scenarios is acquired. Finally, an objective function constrained by conditional risk value considering risk aversion coefficient is established to obtain the bidding discharge power, bidding charging power, and bidding capacity of the energy storage power station under typical day-ahead scenarios in the energy market, thus assisting in bidding decisions. This invention's bidding strategy based on the objective function constrained by conditional risk value considers market price forecast information, frequency regulation signal uncertainty simulation, and state-of-charge constraints at the end of the operating day, obtaining the Pareto optimal solution set considering the decision-maker's risk aversion, thereby assisting the decision-maker in formulating bidding decision schemes. This invention has high accuracy in predicting day-ahead bidding power and bidding capacity, reducing the bidding risk caused by the uncertainty of flow battery energy storage power stations participating in the energy-frequency regulation market, and improving the efficiency of flow battery energy storage power stations participating in the electricity market. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of the day-ahead bidding decision-making method for flow battery energy storage power stations according to the present invention; Figure 2 This is a schematic diagram of the internal structure of the GRU in an embodiment of the present invention; Figure 3 This is a schematic diagram of the GRU prediction model based on the attention mechanism in an embodiment of the present invention; Figure 4 This is a schematic diagram of the bidding process for a flow battery energy storage power station in an embodiment of the present invention; Figure 5 This is a day-ahead market forecast price curve in an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the probability of a typical scenario in an embodiment of the present invention; Figure 7 This is a schematic diagram of the energy accumulation of the frequency modulation signal in each time period in an embodiment of the present invention; Figure 8 This is a schematic diagram illustrating the predicted day-ahead bidding capacity of energy storage power stations in an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] This embodiment introduces a day-ahead bidding decision-making method for flow battery energy storage power stations, including the following steps: Figure 1 and Figure 4 As shown: S1: Obtain historical electricity market clearing prices and use an attention-based GRU neural network to obtain predicted day-ahead electricity market clearing prices.

[0031] Specifically, in this embodiment, the flow battery energy storage power station formulates its output plan for each time period based on historical data and frequency regulation reserved capacity one day before the operation date, with a time scale of 1 hour. Historical price data shows that frequency regulation clearing prices and clearing electricity prices exhibit potential patterns of change. However, historical data is vast, and manually extracting feature values ​​from the sequence would result in low efficiency, high workload, and insufficient accuracy.

[0032] This embodiment utilizes a GRU neural network combined with an attention mechanism to train and learn from historical data, uncovering potential patterns and trends in historical prices to predict future price movements. The GRU neural network is a variant of the LSTM neural network, offering a simpler model structure and reduced computational cost while maintaining the original performance. The GRU contains two gates: an update gate determines which information to discard and add, helping to capture long-term dependencies in time series data; and a reset gate determines which past information to forget, helping to capture short-term dependencies.

[0033] In this embodiment, the internal structure of the GRU is as follows: Figure 2 The diagram and mathematical derivation are shown below: (1) In the formula: To reset the gate control vector, from the previous time step... With the current input get; It is the sigmoid function; and These are the weights and biases for resetting the gate, which are automatically optimized by the backpropagation algorithm; This refers to the state at the previous moment; This is the current input; Indicates the current moment; (2) In the formula: For new input, when When it is 0, All from input This is equivalent to resetting. ;when When the value is 1, the state at the previous time step With the current input Together, new ; For the bias term of the new input; (3) In the formula: To update the gate control vector, used to control the state of the previous step. and new input For the new state vector The degree of impact; and These represent the weights and biases of the updated gate, respectively.

[0034] (4) in, Used to control the state of the previous moment. Signal, Used to control new input Signal; This refers to the current state.

[0035] Specifically, an attention-based GRU prediction model is employed, leveraging the GRU's advantage in processing time series data for training on a price dataset to uncover the changing patterns of frequency regulation clearing prices and clearing electricity prices. Weights are assigned to the new state vector output by the traditional GRU to improve the accuracy of market price predictions. The structure of the attention-based GRU prediction model is as follows: Figure 3 As shown.

[0036] This embodiment cleans the original historical price data (including time-series sequences of daily clearing electricity prices and daily frequency regulation capacity), inputs the processed data into a GRU neural network for training, then introduces an attention mechanism in the hidden layer, using the new state vector output by the GRU time-series prediction as the input to the attention mechanism, and finally calculates the next day's market price prediction value through a sigmoid function after weighted summation by the attention mechanism. The weight calculation process in the attention mechanism is as follows: 1) Construct a validation model to measure contribution: (5) In the formula: These are the attention probability distribution values; This is a rating vector used to measure... The degree of attention contribution to the overall output; The current state The weight coefficient matrix; This is the offset at the current moment; 2) Calculate the normalized weights corresponding to the output values. : (6) 3) Weighted summation yields the output value of the attention mechanism. V : (7) In the formula: This represents the output value of the attention mechanism; This represents the normalized weights corresponding to the output values; 4) Calculate the output layer Predicted value at time : (8) In the formula, sigmoid is the activation function; This is the training weight coefficient matrix from the attention mechanism layer to the output layer; The amount of bias to be trained; For output layer The predicted value at any given time; S2: Obtain historical FM signal scenarios, and use the probabilistic distance fast reduction method to obtain several typical scenarios and the probability of their occurrence; Specifically, the frequency modulation signals received by the energy storage power station are different every day, thus forming a scenario. In this embodiment, the frequency modulation signals received by the flow battery energy storage power station over a cumulative period of 365 days prior to the actual operating day are acquired, forming 365 frequency modulation signal scenarios on a daily basis.

[0037] Specifically, during day-ahead market bidding, due to the numerous and highly uncertain FM signal scenarios, 365 historical FM signal scenarios are used to generate typical scenarios and their corresponding probabilities using the probabilistic distance fast reduction method. This predicts the FM signal for the actual operating day, providing a basis for day-ahead bidding. The calculation steps are as follows: 1) Calculate the pairwise geometric distances between the 365 FM signal scenes to obtain a matrix consisting of the pairwise geometric distances between the 365 scenes. : (9) (10) In the formula: This represents a matrix representing the pairwise geometric distances between 365 scenes; For the first The first FM signal scenario and the first Geometric distance between FM signal scenes ; All of these represent the index number of the FM signal scene; and The first In the first frequency modulation signal scenario m The frequency modulation signal and the first In the first frequency modulation signal scenario m A frequency modulation signal; m This refers to the index number of the frequency modulation signal in the frequency modulation signal scenario; It is an absolute value.

[0038] Specifically, each FM signal scenario contains 43,200 (24*60*30) data points (i.e., FM signals). In this embodiment, the dispatch center sends out FM signals once every 2 seconds, so there are 24*60*30=43,200 FM signals in 24 hours.

[0039] 2) Generate the probability distance matrix Select the scene with the smallest sum of probabilities from the remaining scenes: (11) In the formula: Represents the probability distance matrix; This represents the geometric distance between the first FM signal scene and the y-th FM signal scene; This represents summing the geometric distances between the 365th FM signal scene and the 365th FM signal scene; Indicates transpose; 3) Using the scene set and the first A scenario The scene with the smallest geometric distance replace and will Add the probability to the scene Up, and will Cut away from the scene set to form a new scene set. The probability changes are as follows: (12) In the formula, and The first A scenario and the A scenario No. The probability at the next iteration To replace the second A scenario The probability of; All are index numbers for the scene; This is the index of the number of iterations in the probabilistic distance fast reduction method; 4) Continuously reduce the number of scenes until you get 10 scenes and the probability of each scene.

[0040] S3: Based on the predicted day-ahead electricity market clearing price, obtain the energy storage power station's [price] in the [day-ahead] phase. A typical scenario Daily expected return during the period ; Preferably, the energy storage power station is obtained in the first... A typical scenario The formula used for the expected daily return under the specified time period is as follows: (13) In the formula: This indicates that energy storage power stations were recently... The first time period Expected returns for a typical scenario; This indicates that energy storage power stations are currently in the energy market. The first time period Expected returns for a typical scenario; This indicates that energy storage power stations are currently in the secondary frequency regulation market. The first time period Expected returns for a typical scenario; Index numbers for typical scenarios; Specifically, the expected revenue before the energy market opens includes the expected revenue of energy storage power stations before the energy market opens. And expected revenue in the secondary FM market day .

[0041] in, (14) (15) In the formula: For the predicted date The first time period Clearing prices for a typical scenario; For energy storage power stations in the energy market day The first time period Discharge power for bidding in a typical scenario; For energy storage power stations in the energy market day The first time period The bidding charging power for a typical scenario; Indicates the length of the time period; For the day before The first time period The clearing price of frequency modulation capacity in a typical scenario; For the day before The first time period The clearing price of FM mileage in a typical scenario; For energy storage in the frequency regulation market The first time period Bidding capacity for a typical scenario; For energy storage The first time period Frequency modulation mileage in a typical scenario; The comprehensive performance index of frequency modulation is obtained by weighted summation of the adjustment rate, adjustment time, and adjustment accuracy.

[0042] S4: Acquiring energy storage power stations recently at the A typical scenario Operating costs during the time period ; Preferably, the energy storage power station's day-to-day data is obtained at the [date missing]. A typical scenario The formula used for operating costs during a given time period is as follows; (16) In the formula: It indicates that the energy storage power station was recently at the A typical scenario Operating costs during a given time period; This indicates the investment cost of an energy storage power station; It indicates that the energy storage power station was recently at the A typical scenario Charging and discharging loss costs over a given period of time; It indicates that the energy storage power station was recently at the A typical scenario The lifespan degradation cost caused by the depth of charge and discharge over a given period; Specifically, the day-ahead operating cost of an energy storage power station includes the energy capacity investment cost of the energy storage power station broken down into different bidding periods. Daily charge and discharge loss cost and the cost of lifespan degradation caused by depth of charge and discharge .

[0043] in, (17) In the formula: Cost per unit capacity of energy storage power station; This refers to the rated capacity of the energy storage power station. r The discount rate is 8% in this embodiment; For the float charging life of energy storage power stations; This represents the total number of bidding periods; (18) (19) (20) In the formula: This indicates the day-ahead charge and discharge loss cost of an energy storage power station participating in the energy market; This represents the day-ahead charge and discharge loss cost of an energy storage power station participating in the secondary frequency regulation market; This represents the discharge flag variable of the energy storage power station, with a value of [0,1]. This represents the charging identifier variable for the energy storage power station, with a value of [0,1]. Indicates the discharge efficiency of the energy storage power station; Indicates the charging efficiency of the energy storage power station; Indicates the electricity price for losses; express Accumulated value of positive frequency modulation signal over a given period; express Accumulated value of negative frequency modulation signal over a given period; Specifically, day-ahead charge and discharge loss costs include the day-ahead charge and discharge loss costs of energy storage power stations participating in the energy market. The current charging and discharging loss cost in the secondary frequency modulation market ; and These refer to the charging and discharging efficiencies of the energy storage power station, respectively. and All are [0,1] variables. When the energy storage station is charging, or when a frequency modulation signal is used... When the energy storage power station down-regulates its frequency, at this time , When the energy storage power station discharges, or frequency modulation signals... When the energy storage power station adjusts its frequency upwards, at this time , .

[0044] (twenty one) in: (twenty two) (twenty three) In the formula: Indicates the first Energy storage power stations in typical scenarios State of charge over a period of time; Indicates the first Energy storage power stations in typical scenarios State of charge over a period of time; The fitted energy storage characteristic parameters are generally obtained by fitting the relationship between the number of cycles and the depth of discharge using the operating data provided by the manufacturer. They are generally between 0.8 and 2.1, and in this embodiment, they are taken as 1.1. This indicates the unit power cost of an energy storage power station; Indicates the rated power of the energy storage power station; This indicates the number of energy storage cycles when the energy is charged and discharged at 100% depth. express t The change in electricity consumption resulting from time-of-use energy storage participating in frequency regulation; Indicates the rated energy of the energy storage power station; S5: Based on the probability of typical scenarios occurring, the energy storage power station in the... A typical scenario Daily expected return during the period The energy storage power station was recently at the A typical scenario Operating costs during the time period Value at risk for obtaining the expected day-ahead profit of an energy storage power station ; Preferably, the formula used to obtain the conditional value at risk of the expected day-ahead profit of an energy storage power station is as follows: (twenty four)

[0045] In the formula: Value at risk (VAT) represents the conditional risk of the expected profit of an energy storage power station. To be at the risk confidence level Value at risk below; Risk confidence level; This represents the total number of typical scenarios; Index numbers for typical scenarios; For the first The probability of a typical scenario occurring; For the current bidding period, the cycle is... , This represents the total number of bidding periods; in this example, the value is 24. Indicates intermediate calculation parameters; Represents unknown variables; Since the above equation is a nonlinear function, we can introduce auxiliary variables. By linearizing it using the Big M method, the CVaR calculation expression is obtained as follows: (25) (26) In the formula: Indicates the first The loss exceeds the threshold in a typical scenario; This represents the loss boundary variable, with a value of [0,1]. Indicates the penalty factor; Specifically, when hour, ;when hour, .

[0046] S6: According to the energy storage power station in the... A typical scenario Daily expected return during the period Energy storage power stations were recently at the A typical scenario Operating costs during the time period Conditional Value at Risk (VaR) of Expected Profits for Energy Storage Power Stations To establish an objective function that considers the conditional value-at-risk constraint based on the risk aversion coefficient, in order to obtain the energy storage power station's performance before the energy market date. The first time period Discharge power in a typical scenario Energy storage power stations are currently in the energy market. The first time period Bidding charging power in a typical scenario Energy storage in the frequency regulation market The first time period Bidding capacity in a typical scenario Based on the bidding discharge power, bidding charging power, and bidding capacity of energy storage power stations in typical day-ahead scenarios of the energy market, the bidding decision-making scheme can be formulated.

[0047] Specifically, Conditional Value at Risk (CVaR) can be used to measure investment risk factors. VaR represents the maximum potential loss of a financial investment over a complete investment period at a given confidence level, while CVaR represents the average loss exceeding VaR. For flow battery energy storage participating in the energy-frequency regulation market, factors such as the prediction error of the clearing price and the uncertainty of the frequency regulation signal will affect the bidding scheme. This embodiment uses CVaR as a risk indicator and establishes a bi-objective function incorporating CVaR as follows:

[0048] Since the current market objective function is a biobjective function including CVaR, a linear weighting method can be used to solve the biobjective optimization problem. By introducing weight coefficients, the biobjective problem can be transformed into a single-objective composite problem for optimization. The objective function formula with conditional value at risk constraints is as follows: (27) In the formula: This is the risk aversion coefficient, with a value range of [0,1]. The larger the value, the more risk-averse the bidding entity is. This indicates complete risk aversion; The smaller the value, the more the bidding entity values ​​profit. This indicates a strong preference for risk. The higher the risk aversion coefficient, the lower the expected utilization rate of the energy storage power station in the day-to-day market, which means that the bidding strategy of the energy storage power station is more conservative. Value at risk is the conditional risk of the expected profit of an energy storage power station.

[0049] In this embodiment, the objective function of the conditional risk value constraint is solved using a conventional Cplex solver. The specific solution process will not be described in detail here.

[0050] To verify the feasibility of the bidding strategy proposed in this embodiment, a 100MW / 200MWh flow battery energy storage power station is used for analysis. This energy storage power station meets the capacity requirements for participating in the energy market and frequency regulation ancillary services market. The objective function of the conditional value-at-risk constraint established in this embodiment is simulated on a computer with an Intel Core i5-12600kf processor. The model is solved using the commercial solver Cplex in Matlab R2023b. The relevant parameter settings for the selected energy storage power station are as follows: Table 1 Parameters of Flow Battery Energy Storage Power Station

[0051] The historical data is sourced from the PJM market, using electricity market price data from 2014. The day-ahead market price, frequency regulation capacity price, and mileage price for a given trading day are predicted using a GRU neural network model based on an attention mechanism, as shown below. Figure 5 As shown.

[0052] All FM signals from the US PJM market in 2014 (365 days in total) were selected, with 43,200 FM signals per day and a signal interval of 2 seconds. Using a probabilistic distance fast reduction algorithm, the total number of scenarios for the year was reduced to 10 typical scenarios. The probabilities of these 10 reduced scenarios are as follows: Figure 6 As shown, the cumulative energy of positive and negative frequency modulation signals in 10 typical scenarios is as follows: Figure 7 As shown.

[0053] Historical data and parameters are input into the objective function of the conditional value at risk constraint, and a linear weighted method is used, with a confidence level of [missing information]. Risk aversion coefficient is taken The model was solved using the existing solver Cplex in Matlab to obtain the bidding discharge power, bidding charging power, and bidding capacity in the frequency regulation market for the energy storage power station at various time periods. The results are plotted as follows: Figure 8 The energy storage power station shown is the current day's bidding capacity.

[0054] This implementation predicts actual daily electricity prices and frequency modulation signals using historical data, considers bidding risks and energy storage power station utilization rates, and uses a Cplex solver to solve the objective function of the conditional risk value constraint to generate day-ahead bidding plans. Compared to existing day-ahead bidding plan formation methods, this approach considers the risk aversion of energy storage power stations, ensuring that the energy storage power stations form day-ahead bidding plans with the highest utilization rates at an appropriate level of risk.

[0055] This embodiment presents a day-ahead bidding decision-making method for flow battery energy storage power stations. By employing a probabilistic distance reduction method, historical frequency regulation signal scenarios are compressed to obtain typical scenarios. Then, the method acquires the bidding discharge power, bidding charging power, and bidding capacity of the energy storage power station under these typical day-ahead scenarios in the energy market, considering the risk aversion coefficient. This assists in bidding decision-making. The bidding strategy based on the objective function of the conditional risk value constraint considers market price forecast information, frequency regulation signal uncertainty simulation, and state-of-charge constraints at the end of the operating day. It obtains the Pareto optimal solution set considering the decision-maker's risk aversion, thereby assisting the decision-maker in formulating bidding decision schemes. This invention has high accuracy in predicting day-ahead bidding power and capacity, reducing the bidding risk caused by the uncertainty of flow battery energy storage power stations participating in the energy-frequency regulation market, and improving the efficiency of flow battery energy storage power stations in participating in the electricity market.

[0056] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A day-ahead bidding decision-making method for flow battery energy storage power stations, characterized in that, Includes the following steps: S1: Obtain historical electricity market clearing prices and use an attention-based GRU neural network to obtain predicted day-ahead electricity market clearing prices; S2: Obtain historical FM signal scenarios, and use the probabilistic distance fast reduction method to obtain several typical scenarios and the probability of their occurrence; S3: Based on the predicted day-ahead electricity market clearing price, obtain the expected day-ahead revenue of the energy storage power station under typical scenarios; S4: Obtain the day-to-day operating cost of an energy storage power station in a typical scenario; S5: Based on the probability of typical scenarios occurring, the expected daily revenue of the energy storage power station under typical scenarios, and the daily operating cost of the energy storage power station under typical scenarios, obtain the conditional value of risk of the expected daily profit of the energy storage power station. S6: Based on the day-ahead expected revenue of the energy storage power station in typical scenarios, the day-ahead operating cost of the energy storage power station in typical scenarios, and the day-ahead expected profit of the energy storage power station, establish an objective function constrained by the conditional risk value considering the risk aversion coefficient, in order to obtain the bidding discharge power of the energy storage power station in typical day-ahead scenarios of the energy market, the bidding charging power of the energy storage power station in typical day-ahead scenarios of the energy market, and the bidding capacity of energy storage in typical day-ahead scenarios of the frequency regulation market. Based on the bidding discharge power, bidding charging power, and bidding capacity of energy storage in typical day-ahead scenarios of the energy market, the bidding decision-making scheme can be formulated.

2. The day-ahead bidding decision-making method for a flow battery energy storage power station according to claim 1, characterized in that, The formula used to obtain the expected day-ahead return of an energy storage power station in a typical scenario is as follows: In the formula: It indicates that energy storage power stations were recently The first time period Expected returns for a typical scenario; This indicates that energy storage power stations are currently in the energy market. The first time period Expected returns for a typical scenario; This indicates that energy storage power stations are currently in the secondary frequency regulation market. The first time period Expected returns for a typical scenario; Index numbers for typical scenarios; in, In the formula: For the predicted date The first time period Clearing prices in a typical scenario; For energy storage power stations in the energy market day The first time period Discharge power for a typical bidding scenario; For energy storage power stations in the energy market day The first time period The bidding charging power for a typical scenario; Indicates the length of the time period; For the day before The first time period The clearing price of frequency modulation capacity in a typical scenario; For the day before The first time period The clearing price of FM mileage in a typical scenario; For energy storage in the frequency regulation market The first time period Bidding capacity for a typical scenario; For energy storage The first time period Frequency modulation mileage in a typical scenario; This refers to the overall performance indicators of frequency modulation.

3. The day-ahead bidding decision-making method for a flow battery energy storage power station according to claim 2, characterized in that, The formula used to obtain the current operating cost of an energy storage power station under typical scenarios is as follows: In the formula: It indicates that the energy storage power station was recently at the A typical scenario Operating costs during a given time period; This indicates the investment cost of an energy storage power station; It indicates that the energy storage power station was recently at the A typical scenario Charging and discharging loss costs over a given period of time; It indicates that the energy storage power station was recently at the A typical scenario The lifespan degradation cost caused by the depth of charge and discharge over a given period; in, In the formula: Cost per unit capacity of energy storage power station; This refers to the rated capacity of the energy storage power station. r The discount rate; For the float charging life of energy storage power stations; This represents the total number of bidding periods; In the formula: This indicates the day-ahead charge and discharge loss cost of an energy storage power station participating in the energy market; This represents the day-ahead charge and discharge loss cost of an energy storage power station participating in the secondary frequency regulation market; Indicates the discharge identifier variable of the energy storage power station; This is represented as a charging identifier variable for energy storage power stations; Indicates the discharge efficiency of the energy storage power station; Indicates the charging efficiency of the energy storage power station; Indicates the electricity price for losses; express Accumulated value of positive frequency modulation signal over a given period; express Accumulated value of negative frequency modulation signal over a given period; in: In the formula: Indicates the first Energy storage power stations in typical scenarios State of charge over a period of time; Indicates the first Energy storage power stations in typical scenarios State of charge over a period of time; The parameters representing the fitted energy storage characteristics; This indicates the unit power cost of an energy storage power station; Indicates the rated power of the energy storage power station; This indicates the number of energy storage cycles when the energy is charged and discharged at 100% depth. express t The change in electricity consumption resulting from time-of-use energy storage participating in frequency regulation; This indicates the rated energy of the energy storage power station.

4. The day-ahead bidding decision-making method for a flow battery energy storage power station according to claim 3, characterized in that, The formula used to obtain the conditional value at risk for the expected day-ahead profit of an energy storage power station is as follows: In the formula: Value at risk (VAT) represents the conditional risk of the expected profit of an energy storage power station. To be at the risk confidence level Value at risk below; Risk confidence level; This represents the total number of typical scenarios; Index numbers for typical scenarios; For the first The probability of a typical scenario occurring; For the current bidding period, This represents the total number of bidding periods; Indicates intermediate calculation parameters; Represents unknown variables; In the formula: Indicates the first The loss exceeds the threshold in a typical scenario; Represents the loss boundary variable; This represents the penalty factor.

5. The day-ahead bidding decision-making method for a flow battery energy storage power station according to claim 3, characterized in that, The objective function for the conditional value at risk constraint is established as follows: In the formula: Risk aversion coefficient; Value at risk is the conditional risk of the expected profit of an energy storage power station.

Images