Energy storage system scheduling method and device and storage medium

Abstract

The invention discloses an energy storage system scheduling method and device and a storage medium. The method comprises the following steps: acquiring various bidding behaviors of each target object in a frequency modulation market; the various bidding behaviors are obtained according to historical bidding records of the corresponding target objects in the frequency modulation market; according to the current market data of the frequency modulation market, calculating the income of each type of bidding behavior in the frequency modulation market, so as to predict the bidding information of each target object in the frequency modulation market; inputting the bidding information of each target object in the frequency modulation market into a frequency modulation market clearing model to predict bid winning information of the target energy storage system in the frequency modulation market; the frequency modulation market clearing model takes the lowest frequency modulation service cost as a target; inputting the bid winning information into a pre-constructed decision model, and calculating to obtain an optimal scheduling strategy of the target energy storage system; the decision model aims at maximizing the sum of benefits in the electric energy market and the frequency modulation market. According to the invention, the scheduling capability of a single energy storage system can be optimized.

Description

Technical Field

[0001] The present application relates to the technical field of energy storage systems, and in particular to an energy storage system scheduling method, device and storage medium. Background Art

[0002] With the rapid development of renewable energy, energy storage systems play an increasingly important role in the electricity market. Existing energy storage scheduling models mainly focus on how to combine new energy sources such as electricity, thermal energy and hydrogen energy with the optimized operation of energy storage itself to improve energy efficiency and system stability. Or from the perspective of independent system operators in the electricity market, optimizing the scheduling of new energy generation and adjustable loads in the entire region, and studying regional optimization scheduling methods under the uncertainty of new energy generation have certain limitations. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a method, device and storage medium for scheduling an energy storage system, so as to solve the problem that the prior art cannot optimize the scheduling of a single energy storage power station.

[0004] In order to achieve the above-mentioned object, the first aspect of the present application provides a method for scheduling an energy storage system, the method comprising:

[0005] Obtain various bidding behaviors of each target object in the frequency modulation market; various bidding behaviors are obtained according to the historical bidding records of the corresponding target object in the frequency modulation market;

[0006] According to the current market data of the frequency modulation market, the benefits of each type of bidding behavior in the frequency modulation market are calculated respectively to predict the bidding information of each target object in the frequency modulation market;

[0007] Input the bidding information of each target object in the frequency regulation market into the frequency regulation market clearing model to predict the winning bid information of the target energy storage system in the frequency regulation market; the frequency regulation market clearing model aims to minimize the frequency regulation service cost;

[0008] The winning bid information is input into a pre-built decision model to calculate the optimal dispatching strategy for the target energy storage system; the decision model aims to maximize the sum of the benefits in at least the electric energy market and the frequency regulation market.

[0009] In the embodiment of the present application, the step of constructing the frequency modulation market clearing model includes: constructing the frequency modulation market clearing model according to formula (1):

[0010]

[0011] Among them, C Reg The cost of frequency modulation services, For each target object, the winning bid capacity, For each target object, is the bidding capacity unit price for each target object, The bidding mileage unit price for each target object, is the winning bid capacity of the target energy storage system, The winning mileage of the target energy storage system, is the winning capacity unit price of the target energy storage system, is the winning mileage unit price of the target energy storage system, Ω is the set of various bidding behaviors of multiple target objects in the frequency regulation market, i is the number of the target object, t is the clearing time, and T is the total scheduling time.

[0012] In the embodiment of the present application, the method further includes: constraining the winning mileage of each target object and the winning capacity of each target object according to formula (2), formula (3) and formula (4):

[0013]

[0014]

[0015]

[0016] in, For each target object's bidding capacity, is the dual variable corresponding to the lower bound constraint of the bid capacity for each target object, is the dual variable corresponding to the upper limit constraint of the bid capacity of each target object, For each target object, the bidding mileage is the dual variable corresponding to the lower limit constraint of the mileage of each target object, is the dual variable corresponding to the upper limit constraint of the winning mileage for each target object, is the FM mileage multiplier for each target object, is the dual variable corresponding to the constraint;

[0017] And / or, according to formula (5), formula (6) and formula (7), the winning mileage of the target energy storage system and the winning capacity of the target energy storage system are constrained:

[0018]

[0019]

[0020]

[0021] in, is the bid capacity of the target energy storage system, is the bidding mileage of the target energy storage system, is the frequency regulation mileage multiplier of the target energy storage system, is the dual variable corresponding to the minimum value constraint of the target energy storage system’s winning capacity, is the dual variable corresponding to the maximum value constraint of the target energy storage system’s bid capacity, is the dual variable corresponding to the minimum mileage constraint of the target energy storage system, is the dual variable corresponding to the maximum mileage constraint of the target energy storage system, is the dual variable corresponding to the maximum value constraint of the winning capacity of the target energy storage system corresponding to the winning mileage;

[0022] And / or, constrain the capacity and mileage required by the power grid system according to formula (8) and formula (9):

[0023]

[0024]

[0025] in, is the capacity required by the power grid system at time t, is the dual variable corresponding to the upper limit constraint of the bidding capacity for each target object, is the mileage required by the power grid system at time t, The dual variable corresponding to the upper limit constraint of the bidding mileage for each target object.

[0026] In the embodiment of the present application, the step of constructing the decision model includes: constructing the decision model according to formula (10):

[0027]

[0028] Among them, C E is the benefit value of the target energy storage system, which includes the electric energy benefit and frequency regulation benefit of the target energy storage system. is the power purchased by the target energy storage system at time t, is the electricity sales power of the target energy storage system at time t, is the unit price of electricity purchased by the target energy storage system at time t, is the electricity selling price of the target energy storage system at time t, is the capacity compensation unit price, is the unit price of photovoltaic consumption compensation, Δt is the scheduling time interval, To clear the capacity unit price, is the winning bid capacity of the target energy storage system, To clear the mileage unit price, is the winning mileage of the target energy storage system, c j The degradation cost.

[0029] In an embodiment of the present application, the method further includes: constraining the charging and discharging power of the target energy storage system according to formula (11), formula (12), formula (13) and formula (14):

[0030]

[0031]

[0032]

[0033]

[0034] Among them, θ t is the binary variable for charging and discharging of the target energy storage system, is the upper limit of the charging power of the target energy storage system, is the upper limit of the discharge power of the target energy storage system, is the upper limit of the bidding capacity of the target energy storage system at time t, The bid capacity reported for the target energy storage system, The bid mileage declared for the target energy storage system, is the call rate of the frequency regulation mileage of the target energy storage system;

[0035] And / or, constraining the charge of the target energy storage system according to formula (15), formula (16), formula (17) and formula (18):

[0036]

[0037]

[0038] SOC min ≤∑ j∈J soc j,t ≤SOC max (18)

[0039] Among them, soc j,t is the charge of the target energy storage system at time t in segment j, is the upper limit of the charge in segment j, SOC min is the minimum charge value of the target energy storage system, SOC max is the maximum value of the charge of the target energy storage system, E cap is the capacity of the target energy storage system, η is the charging and discharging efficiency of the target energy storage system;

[0040] And / or, constrain the declared price of the target energy storage system according to formula (19) and formula (20):

[0041]

[0042]

[0043] in, is the lower limit of the declared capacity unit price of the target energy storage system, The upper limit of the declared capacity unit price of the target energy storage system, is the lower limit of the declared mileage unit price of the target energy storage system, The upper limit of the declared mileage unit price for the target energy storage system.

[0044] In the embodiment of the present application, according to the current market data of the frequency modulation market, respectively calculating the revenue of each type of bidding behavior in the frequency modulation market includes: determining the revenue of each target object according to formula (21):

[0045]

[0046] in, is the benefit of each target object at time t, a t is the set of bidding behaviors of each target object at time t, s t is the current market data at time t, and f is the long short-term memory neural network.

[0047] In an embodiment of the present application, historical bidding records include historical bidding capacity unit price, historical bidding mileage unit price, historical bidding capacity and historical bidding mileage; bidding information includes bidding capacity unit price, bidding mileage unit price, bidding capacity and bidding mileage; winning bid information includes winning bid mileage and winning bid capacity; current market data includes market load, clearing capacity unit price, clearing mileage unit price, clearing capacity and clearing mileage.

[0048] In an embodiment of the present application, obtaining various bidding behaviors of each target object in the FM market includes: clustering historical bidding records based on the K-mediods clustering algorithm to determine the various bidding behaviors of each target object in the FM market; the step of predicting the bidding information of each target object in the FM market includes: determining the state transition probability of each bidding behavior according to the revenue of each bidding behavior in the FM market; and predicting the bidding information of each target object in the FM market based on the state transition probability and the roulette algorithm.

[0049] A second aspect of the present application provides an energy storage system scheduling device, comprising:

[0050] a memory configured to store instructions; and

[0051] The processor is configured to call the instructions from the memory and implement the above-mentioned energy storage system scheduling method when executing the instructions.

[0052] A third aspect of the present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned energy storage system scheduling method.

[0053] Through the above technical solution, various bidding behaviors of each target object in the frequency modulation market are obtained. Various bidding behaviors can be obtained based on the historical bidding records of the corresponding target object in the frequency modulation market; according to the current market data of the frequency modulation market, the benefits of each type of bidding behavior in the frequency modulation market are calculated respectively to predict the bidding information of each target object in the frequency modulation market; the bidding information of each target object in the frequency modulation market is input into the frequency modulation market clearing model. The frequency modulation market clearing model requires the minimization of the auxiliary cost of frequency modulation. Therefore, when the bidding information is known, the winning bid information of the target energy storage system in the frequency modulation market can be predicted; the winning bid information is input into the pre-built decision model to calculate the optimal scheduling strategy of the target energy storage system. In this way, the production scheduling of a single target energy storage system can be optimized to improve the benefits of the target energy storage system.

[0054] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings:

[0056] Figure 1 A flowchart of a method for scheduling an energy storage system according to an embodiment of the present application is schematically shown;

[0057] Figure 2 A framework diagram of a decision model according to an embodiment of the present application is schematically shown;

[0058] Figure 3 A framework diagram of a target energy storage system bidding model according to an embodiment of the present application is schematically shown;

[0059] Figure 4a and Figure 4b A diagram schematically shows the result of clearing the winning bid capacity of the energy storage system according to an embodiment of the present application;

[0060] Figure 5a and Figure 5b A diagram schematically shows a result of bidding for energy storage system capacity according to an embodiment of the present application;

[0061] Figure 6a and Figure 6b A diagram schematically shows a quotation result diagram of the bidding mileage of the energy storage system according to an embodiment of the present application;

[0062] Figure 7a and Figure 7b A diagram schematically shows a charge amount change result diagram of an energy storage system according to an embodiment of the present application;

[0063] Figure 8 The structure of a scheduling device for an energy storage system according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0065] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0066] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0067] Figure 1 A flowchart of a method for scheduling an energy storage system according to an embodiment of the present application is schematically shown; Figure 2 A framework diagram of a decision model according to an embodiment of the present application is schematically shown; Figure 3 A target energy storage system bidding model framework diagram according to an embodiment of the present application is schematically shown. Figure 1 , Figure 2 and Figure 3 As shown, an embodiment of the present application provides a method for scheduling an energy storage system, which may include the following steps.

[0068] Step 101: Obtain various bidding behaviors of each target object in the frequency modulation market; various bidding behaviors are obtained according to the historical bidding records of the corresponding target object in the frequency modulation market.

[0069] In an embodiment of the present application, the target energy storage system may be an energy storage power station. In the frequency regulation market, the target object may refer to other market entities that participate in the market bidding together with the energy storage power station. These entities may include thermal power plants, photovoltaic power stations, wind farms and other equipment or systems that can provide frequency regulation services. The frequency regulation market is part of the power auxiliary service market and is used to balance the frequency fluctuations of the power grid. Participants provide frequency response services to the power grid by adjusting power output or consumption to ensure stable operation of the power grid. Various types of bidding behaviors may be specific bidding methods for each target object in the frequency regulation market, which may include capacity unit price, mileage unit price, declared capacity and mileage, etc. according to market rules.

[0070] Step 102: According to the current market data of the frequency modulation market, the revenue of each type of bidding behavior in the frequency modulation market is calculated respectively, so as to predict the bidding information of each target object in the frequency modulation market.

[0071] In the embodiment of the present application, the revenue of each type of bidding behavior in the frequency modulation market may refer to the economic revenue of different bidding behaviors in the current market environment. By calculating the revenue of each type of bidding behavior in the current market and combining the behavior pattern of the target object (such as historical bidding records and revenue tendency), its bidding information in the current market is predicted.

[0072] Step 103: Input the bidding information of each target object in the frequency regulation market into the frequency regulation market clearing model to predict the winning bid information of the target energy storage system in the frequency regulation market; the frequency regulation market clearing model aims to minimize the frequency regulation service cost.

[0073] In the embodiment of the present application, the frequency regulation market clearing model can be used to simulate the resource allocation mechanism of the frequency regulation market, combine the bidding information of all participants, calculate the winning capacity, mileage and corresponding price of each subject, and predict the winning result of the target energy storage system in the frequency regulation market, including the winning capacity, winning mileage and winning price.

[0074] Step 104: Input the winning bid information into a pre-built decision model to calculate the optimal dispatching strategy for the target energy storage system; the decision model aims to maximize the sum of the revenue in at least the electric energy market and the frequency regulation market.

[0075] In the embodiment of the present application, the optimal scheduling strategy may refer to the bidding information of the target energy storage system. When the target energy storage system is declared according to the optimal scheduling strategy, it can win the bid with a greater probability and obtain the maximum benefit. Therefore, by linking the winning bid information with the benefit, the bidding information of the target energy storage system can be dynamically adjusted and the production scheduling can be optimized to ensure the competitiveness of the target energy storage system in the market and maximize the benefit.

[0076] Through the above technical solution, various bidding behaviors of each target object in the frequency modulation market are obtained. Various bidding behaviors can be obtained based on the historical bidding records of the corresponding target object in the frequency modulation market; according to the current market data of the frequency modulation market, the benefits of each type of bidding behavior in the frequency modulation market are calculated respectively to predict the bidding information of each target object in the frequency modulation market; the bidding information of each target object in the frequency modulation market is input into the frequency modulation market clearing model. The frequency modulation market clearing model requires the minimization of the auxiliary cost of frequency modulation. Therefore, when the bidding information of the target object is known, the winning bid information of the target energy storage system in the frequency modulation market can be predicted; the winning bid information is input into the pre-built decision model to calculate the optimal scheduling strategy of the target energy storage system. In this way, the production scheduling of a single target energy storage system can be optimized to improve the benefits of the target energy storage system.

[0077] In the embodiment of the present application, the step of constructing the frequency modulation market clearing model includes: constructing the frequency modulation market clearing model according to formula (1):

[0078]

[0079] Among them, C Reg The cost of frequency modulation services, For each target object, the winning bid capacity, For each target object, is the bidding capacity unit price for each target object, The bidding mileage unit price for each target object, is the winning bid capacity of the target energy storage system, The winning mileage of the target energy storage system, is the winning capacity unit price of the target energy storage system, is the winning mileage unit price of the target energy storage system, Ω is the set of various bidding behaviors of multiple target objects in the frequency regulation market, i is the number of the target object, t is the clearing time, and T is the total scheduling time.

[0080] In the embodiment of the present application, the method further includes: constraining the bid mileage of each target object and the bid capacity of each target object according to formula (2), formula (3) and formula (4):

[0081]

[0082]

[0083]

[0084] in, For each target object's bidding capacity, is the dual variable corresponding to the lower bound constraint of the bid capacity for each target object, is the dual variable corresponding to the upper limit constraint of the bid capacity of each target object, For each target object, the bidding mileage is the dual variable corresponding to the lower limit constraint of the mileage of each target object, is the dual variable corresponding to the upper limit constraint of the winning mileage for each target object, is the FM mileage multiplier for each target object, is the dual variable corresponding to the constraint;

[0085] And / or, according to formula (5), formula (6) and formula (7), the winning mileage of the target energy storage system and the winning capacity of the target energy storage system are constrained:

[0086]

[0087]

[0088]

[0089] in, is the bid capacity of the target energy storage system, is the bidding mileage of the target energy storage system, is the frequency regulation mileage multiplier of the target energy storage system, is the dual variable corresponding to the minimum value constraint of the target energy storage system’s winning capacity, is the dual variable corresponding to the maximum value constraint of the target energy storage system’s bid capacity, is the dual variable corresponding to the minimum mileage constraint of the target energy storage system, is the dual variable corresponding to the maximum mileage constraint of the target energy storage system, is the dual variable corresponding to the maximum value constraint of the winning capacity of the target energy storage system corresponding to the winning mileage;

[0090] And / or, constrain the capacity and mileage required by the power grid system according to formula (8) and formula (9):

[0091]

[0092]

[0093] in, is the capacity required by the power grid system at time t, is the dual variable corresponding to the upper limit constraint of the bidding capacity for each target object, is the mileage required by the power grid system at time t, The dual variable corresponding to the upper limit constraint of the bidding mileage for each target object.

[0094] In the embodiment of the present application, the step of constructing the decision model includes: constructing the decision model according to formula (10):

[0095]

[0096] Among them, C E is the benefit value of the target energy storage system, which includes the electric energy benefit and frequency regulation benefit of the target energy storage system. is the power purchased by the target energy storage system at time t, is the electricity sales power of the target energy storage system at time t, is the unit price of electricity purchased by the target energy storage system at time t, is the electricity selling price of the target energy storage system at time t, is the capacity compensation unit price, is the unit price of photovoltaic consumption compensation, Δt is the scheduling time interval, To clear the capacity unit price, is the winning bid capacity of the target energy storage system, To clear the mileage unit price, is the winning mileage of the target energy storage system, c j The degradation cost.

[0097] In an embodiment of the present application, the method further includes: constraining the charging and discharging power of the target energy storage system according to formula (11), formula (12), formula (13) and formula (14):

[0098]

[0099]

[0100]

[0101]

[0102] Among them, θ t It is a binary variable for charging and discharging the target energy storage system, 0 means discharging, 1 means charging, is the upper limit of the charging power of the target energy storage system, is the upper limit of the discharge power of the target energy storage system, is the upper limit of the bidding capacity of the target energy storage system at time t, The bid capacity reported for the target energy storage system, The bid mileage declared for the target energy storage system, is the call rate of the frequency regulation mileage of the target energy storage system;

[0103] And / or, constraining the charge of the target energy storage system according to formula (15), formula (16), formula (17) and formula (18):

[0104]

[0105]

[0106]

[0107] Among them, soc j,t is the charge of the target energy storage system at time t in segment j, is the upper limit of the charge in segment j, SOC min is the minimum charge value of the target energy storage system, SOC max is the maximum value of the charge of the target energy storage system, E cap is the capacity of the target energy storage system, and η is the charging and discharging efficiency of the system;

[0108] And / or, constrain the declared price of the target energy storage system according to formula (19) and formula (20):

[0109]

[0110]

[0111] in, is the lower limit of the declared capacity unit price of the target energy storage system, The upper limit of the declared capacity unit price of the target energy storage system, is the lower limit of the declared mileage unit price of the target energy storage system, The upper limit of the declared mileage unit price for the target energy storage system.

[0112] In the embodiment of the present application, according to the current market data of the frequency modulation market, respectively calculating the revenue of each type of bidding behavior in the frequency modulation market includes: determining the revenue of each target object according to formula (21):

[0113]

[0114] in, is the benefit of each target object at time t, a t is the set of bidding behaviors of each target object at time t, st is the current market data at time t, and f is the long short-term memory neural network.

[0115] In an embodiment of the present application, historical bidding records include historical bidding capacity unit price, historical bidding mileage unit price, historical bidding capacity and historical bidding mileage; bidding information includes bidding capacity unit price, bidding mileage unit price, bidding capacity and bidding mileage; winning bid information includes winning bid mileage and winning bid capacity; current market data includes market load, clearing capacity unit price, clearing mileage unit price, clearing capacity and clearing mileage.

[0116] In an embodiment of the present application, obtaining various bidding behaviors of each target object in the FM market includes: clustering historical bidding records based on the K-mediods clustering algorithm to determine the various bidding behaviors of each target object in the FM market; the step of predicting the bidding information of each target object in the FM market includes: determining the state transition probability of each bidding behavior according to the revenue of each bidding behavior in the FM market; and predicting the bidding information of each target object in the FM market based on the state transition probability and the roulette algorithm.

[0117] In the embodiment of the present application, K-mediods is a clustering algorithm based on unsupervised learning, which is used to divide historical bidding records into several categories to reflect different market behavior patterns. The state transition probability represents the probability that the bidding information of the target object changes between historical categories. The roulette algorithm is a random selection algorithm that randomly samples the bidding information category of the next step according to the state transition probability, and the probability of the selection method is proportional to the state transition probability. Through the clustering of historical bidding records and state transition probability modeling, the roulette algorithm is used to generate simulated bidding behavior of the target object. This provides data support for the input of the decision model and provides the bidding information of the target object for the frequency modulation market clearing model.

[0118] The following is a specific embodiment:

[0119] 1. Simulate and generate bidding information of the target object.

[0120] 1. Predict the bidding behavior of the target object based on the profit function (LSTM deep learning network). The data in the input layer includes the market information of the FM market. t (as shown in formula 22) and the historical bidding information a of the target object t (As shown in formula 23).

[0121]

[0122]

[0123] Among them, D t is the load in the current market, are the clearing capacity unit price and clearing mileage unit price of the frequency regulation market, They are the frequency regulation capacity and frequency regulation mileage required by the frequency regulation market respectively.

[0124] The output layer is shown in formula (24):

[0125]

[0126] in, Bid is the income that the target object can obtain under the above bidding information. θ It is a neural network that needs to be trained.

[0127] The revenue model can be a long short-term memory neural network (LSTM), which is a special recurrent neural network (RNN) used to process time series data. It can remember historical data for a long time and make predictions at the current time point, and is used in the revenue model to capture market trends and behavioral patterns. The historical bidding information is input into the revenue model to predict the target object's possible bidding information in the frequency regulation market, providing input data for the optimized bidding strategy of the target energy storage system, thereby improving the bidding success rate and revenue.

[0128] 2. Specifically, the construction process of the profit function is as follows.

[0129] (1) Forget gate. The forget gate can decide which information should be forgotten or retained from the cell state. It decides to retain or discard the information in the cell state by looking at the output of the previous time step and the input of the current time step. As shown in formula (25):

[0130] f t =σ(W f ·[h t-1 ,x t ]+b f ) (25)

[0131] Among them, f t is the output of the forget gate. σ is the sigmoid activation function, and the output is compressed between 0 and 1; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step, that is, the market information s mentioned above t and the historical bidding information of the target object a t ; b f is the bias term of the forget gate.

[0132] (2) Input gate. The input gate consists of two parts: the sigmoid layer determines which values ​​will be updated; the tanh layer creates a new vector of candidate values ​​that will be added to the state. This is shown in formulas (26) and (27).

[0133] i t =σ(W i ·[ t-1 ,x t ]+b i ) (26)

[0134]

[0135] Among them, i t is the output of the input gate; is a candidate value; W i and W C is the weight matrix of the input gate and candidate value; b i and b C is the offset.

[0136] (3) Cell state. The cell state is part of the above-mentioned revenue model (i.e., LSTM network), which carries information about the observation sequence and is updated with time steps. The update of the cell state is jointly determined by the forget gate, input gate, and output gate. As shown in formula (28).

[0137]

[0138] Among them, C t is the cell state at the current time step. C t-1 is the cell state at the previous time step.

[0139] (4) Output gate. Satisfies formula (29) and formula (30).

[0140] o t =σ(W o ·[- t-1 ,x t ]+b o ) (29)

[0141] t=o t *tan(C t ) (30)

[0142] Among them, t is the output of the output gate; t is the hidden state of the current time step, that is, the feature representation required for the benefit of the target object; W o is the weight matrix of the output gate.

[0143] (5) Training process: Forward propagation: The input sequence data is input into the LSTM network in time steps to obtain the output. Calculate the error: After the last time step of the sequence or each time step, the output of the LSTM is compared with the true label, and the cross entropy is used to calculate the error between the predicted value and the actual value. Backward propagation: Using the calculated error, the gradient of each parameter is calculated through the back propagation algorithm. After the gradient calculation is completed, the Adam algorithm is used to update the weight and bias parameters of the network. Weight update: According to the calculated gradient and learning rate, the weights and bias items in the network are updated to minimize the prediction error. Iterative optimization: Repeat the above process until the performance of the model on the training set reaches a satisfactory level or the predetermined training cycle is reached.

[0144] 3. The K-mediods clustering algorithm is used to cluster the bidding behavior of the target object, as follows.

[0145] (1) At the beginning of the algorithm, K data points need to be randomly selected as the initial cluster centers.

[0146] (2) For each data point in the data set, calculate its distance to each cluster center and assign it to the cluster center with the closest distance. Use Euclidean distance as the distance metric, as shown in formula (31):

[0147]

[0148] Where x is the data point, c i is the i-th cluster center, d is the dimension of the data, x j and c ij are x and c respectively i The value in the jth dimension.

[0149] (3) For each cluster, recalculate its cluster center. The new cluster center is the point with the smallest sum of distances from all other points in the cluster, as shown in formula (32):

[0150]

[0151] Among them, the set of data points of each cluster is represented as S i ;c i is the new cluster center of the i-th cluster, and x is the cluster center of S i data points.

[0152] (4) Repeat steps (2) and (3) above until one of the following two termination conditions is met: the distance between the new cluster center and the old cluster center is less than the preset threshold ε or the maximum number of iterations T is reached. max .

[0153] 4. Based on the above-mentioned benefit function and the result of the bidding behavior clustering of the target object, the bidding information of the target object is simulated and generated.

[0154] Using the generated cluster bidding curve, the estimated revenue calculation of the bidding behavior obtained by clustering is performed through the trained revenue function to obtain the state transition probability:

[0155] Take action a t From state s t To state s t+1 The transition probability is expressed as shown in formula (33):

[0156]

[0157] Among them, r(s t ,a t ) is the bidding object in s t Take action a in state t The benefits that can be obtained, Pr(s t |a t ,s t+1 ) is the required transition probability.

[0158] Through the generated data, a corresponding target object bidding data set is generated according to the roulette method.

[0159] 2. Construct a bidding and auction model for energy storage systems.

[0160] 1. The bidding model of the energy storage system satisfies formula (10).

[0161]

[0162] Among them, C E is the benefit value of the target energy storage system, which includes the electric energy benefit and frequency regulation benefit of the target energy storage system. is the power purchased by the target energy storage system at time t, is the electricity sales power of the target energy storage system at time t, is the unit price of electricity purchased by the target energy storage system at time t, is the electricity selling price of the target energy storage system at time t, is the capacity compensation unit price, is the unit price of photovoltaic consumption compensation, Δt is the scheduling time interval, To clear the capacity unit price, is the winning bid capacity of the target energy storage system, To clear the mileage unit price, is the winning mileage of the target energy storage system, c jThe economic benefits of the energy storage power station are the benefits from participating in the electricity market plus the benefits from the frequency regulation market, minus the degradation costs of the energy storage equipment.

[0163] The constraints on the power of the energy storage system satisfy formula (11), formula (12), formula (13) and formula (14).

[0164]

[0165]

[0166]

[0167]

[0168] Among them, θ t It is a binary variable for charging and discharging the target energy storage system, 0 means discharging, 1 means charging, is the upper limit of the charging power of the target energy storage system, is the upper limit of the discharge power of the target energy storage system, is the upper limit of the bidding capacity of the target energy storage system at time t, is the adjusted bid capacity, is the adjusted bid mileage, is the call rate of the frequency regulation mileage of the target energy storage system.

[0169] The constraints on the charge of the energy storage system satisfy formula (15), formula (16), formula (17) and formula (18).

[0170]

[0171] soc j,0 =soc j,T (16)

[0172]

[0173] SOC min ≤∑ j∈J soc j,t ≤SOC max (18)

[0174] Among them, soc j,t is the charge of the target energy storage system at time t in segment j, is the power purchase of the target energy storage system at time t, is the electricity sales power of the target energy storage system at time t, is the upper limit of the charge in segment j, SOC minis the minimum value of the charge of the target energy storage system, SOC max is the maximum value of the charge of the target energy storage system, Δt is the scheduling time interval, E cap is the capacity of the target energy storage system, and η is the charging and discharging efficiency of the system;

[0175] The bidding price constraints for energy storage systems in the frequency regulation market satisfy formulas (19) and (20).

[0176]

[0177]

[0178] in, is the lower limit of the declared capacity unit price of the target energy storage system, is the upper limit of the declared capacity unit price of the target energy storage system, is the lower limit of the declared mileage unit price of the target energy storage system, The upper limit of the unit price of mileage reported for the target energy storage system.

[0179] 2. The frequency regulation market clearing model satisfies formula (1). In the current model for energy storage systems participating in the electric energy market and the frequency regulation auxiliary service market, the electricity price in the electric energy market is fixed by the time-of-use electricity price. Here, the frequency regulation auxiliary service market clearing model is simulated. The frequency regulation auxiliary service market clearing is sorted by comprehensive quotations, taking into account the system frequency regulation capacity demand and mileage demand.

[0180]

[0181] Among them, C Reg The cost of frequency modulation services, is the bidding capacity of the target object, The bidding mileage for the target object, is the bid capacity unit price of the target object, is the bid mileage unit price of the target object, is the winning bid capacity of the target energy storage system, is the winning mileage of the target energy storage system, is the winning capacity unit price of the target energy storage system, is the winning mileage unit price of the target energy storage system, Ω is the set of various bidding behaviors of multiple target objects in the frequency modulation market, i is the target object number, t is the clearing time, and T is the total scheduling time.

[0182] Constrain the winning bid capacity and mileage of the target object to satisfy formula (2), formula (3) and formula (4).

[0183]

[0184]

[0185]

[0186] in, For each target object's bidding capacity, is the dual variable corresponding to the lower bound constraint of the bid capacity for each target object, is the dual variable corresponding to the upper limit constraint of the bid capacity of each target object, For each target object, the bidding mileage is the dual variable corresponding to the lower limit constraint of the mileage of each target object, is the dual variable corresponding to the upper limit constraint of the winning mileage for each target object, is the FM mileage multiplier for each target object, is the dual variable corresponding to this constraint.

[0187] Constraints are imposed on the bid-winning capacity and mileage of the energy storage system to satisfy formulas (5), (6) and (7).

[0188]

[0189]

[0190]

[0191] in, is the winning bid capacity of the target energy storage system, is the winning mileage of the target energy storage system, is the bid capacity of the target energy storage system, is the bid mileage of the target energy storage system, is the frequency regulation mileage multiplier of the target energy storage system, is the dual variable corresponding to the minimum value constraint of the target energy storage system’s winning capacity, is the dual variable corresponding to the maximum value constraint of the target energy storage system’s bid capacity, is the dual variable corresponding to the minimum mileage constraint of the target energy storage system, is the dual variable corresponding to the maximum mileage constraint of the target energy storage system, It is the dual variable corresponding to the maximum value constraint of the winning capacity of the target energy storage system corresponding to the winning mileage.

[0192] The capacity and mileage required by the power grid system are constrained to satisfy equations (8) and (9).

[0193]

[0194]

[0195] in, are the capacity and mileage requirements of the power grid system frequency regulation auxiliary service at time t respectively; Formula (8) is used to constrain the overall winning frequency regulation capacity of the energy storage system to be greater than or equal to the frequency regulation capacity required by the power grid system; Formula (9) is used to constrain the overall winning frequency regulation mileage of the energy storage system to be greater than or equal to the frequency regulation mileage required by the power grid system.

[0196] 3. Model evaluation.

[0197] 1. Evaluation of the revenue model (LSTM deep learning network).

[0198] (1) Evaluation indicators: In this model, the main considerations are the bidding behaviors based on time coupling and load forecasting. Therefore, the accuracy of the prediction of this model mainly considers the accuracy of the reward function prediction in LSTM. The root mean square error (RMSE) and the coefficient of determination R2 (R-Square) are used as the accuracy evaluation indicators of the reward function prediction model. As shown in formula (34) and formula (35).

[0199]

[0200]

[0201] Among them, the smaller the RMSE index, the better the prediction effect, and the closer the R2 index is to 1, the better the effect.

[0202] (2) Comparison model settings.

[0203] In order to verify the superiority of the load forecasting method based on the LSTM network deep learning algorithm proposed in the above scheme, three forecasting models are set for comparative analysis. Model 1: LSTM model considering load, market price factors and bidding behavior, Model 2: LSTM model without considering load factors, Model 3: Fully connected neural network model considering load, market price factors and bidding behavior, predict the profit function, and the test set is the Australian electricity market bidding and clearing data from January 2023 to December 2023.

[0204] (3) Comparison of prediction results.

[0205] The data comes from the Australian electricity market, as shown in Table 1. Table 1 shows the comparison of the evaluation indicators of the root mean square error and the determination coefficient R2 of the model proposed in the above technical solution and the two comparative models. It can be seen from Table 1 that the load forecasting model proposed in the present invention is superior to the comparative forecasting model in both evaluation indicators. From the perspective of the root mean square error, the load forecasting model proposed in the present invention is improved by 3.21% and 4.82% compared with Model 2 and Model 3 respectively; from the perspective of the determination coefficient, the load forecasting model proposed in the present invention is improved by 4.1% and 8.52% compared with Model 1 and Model 2 respectively, which effectively proves the advantage of the bidding revenue forecasting method proposed in the present invention in terms of prediction accuracy.

[0206] Table 1

[0207]

[0208]

[0209] 2. Analysis of bidding strategies for energy storage systems.

[0210] Figure 4a and Figure 4b A diagram schematically shows the result of clearing the winning bid capacity of the energy storage system according to an embodiment of the present application; Figure 5a and Figure 5b A diagram schematically shows a result of bidding for energy storage system capacity according to an embodiment of the present application; Figure 6a and Figure 6b A diagram schematically shows a quotation result diagram of the bidding mileage of the energy storage system according to an embodiment of the present application; Figure 7a and Figure 7b The following diagram schematically shows the charge amount change result of the energy storage system according to the embodiment of the present application. Figure 4a , Figure 4b , Figure 5a , Figure 5b , Figure 6a , Figure 6b , Figure 7a and Figure 7b In the clearing model simulating the energy storage system, two cases are discussed: Case 1: the price of the frequency regulation market is slightly higher than that of the electric energy market; Case 2: the prices of the two are similar. Figure 4a , Figure 5a , Figure 6a and Figure 7a For case 1, Figure 4b , Figure 5b , Figure 6b and Figure 6b For situation 2.

[0211] exist Figure 4a and Figure 4b In, R cap_esIt represents the winning capacity of the energy storage system, Rocap is the bidding capacity of the target object, PregcapD is the capacity required by the power grid system, and Rocapmax is the sum of the capacities declared by all participants. When the prices of the two are close and in the above situation 1, the winning capacity of the energy storage system is basically equal to its upper limit (120 units), and the price of the frequency regulation market is generally higher than that of the electric energy market. Therefore, the energy storage system tends to invest its capacity in the frequency regulation market. In situation 2, the winning capacity of the energy storage system did not reach the maximum value every hour, and there was a trade-off between the frequency regulation market and the electric energy market. The winning capacity was reduced at some times, such as the 7th, 8th, and 10th hours.

[0212] Figure 5a and Figure 5b The capacity bid line charts of the energy storage system and the considered bidding opponents in two cases are shown in Figure 2. Figure 6a and Figure 6b The mileage quotation line charts in two cases. Target objects 1 to 10 are the bid price lines of the 10 bidding opponents (i.e., target objects) under consideration. The bidding price of the energy storage system every hour is marked by circles, which reflects the price market game between the energy storage system and the bidding opponents within a 24-hour time scale. It can be seen that in case 1, the energy storage system can obtain more winning bids through lower capacity and mileage declared prices. At this time, the benefits obtained from the lower bids are still higher than the benefits that can be brought by participating in the electric energy market. In case 2, taking the 10th hour as an example, there are many competitors with lower bids in the power grid system. The energy storage system can obtain more profits by participating in the electric energy market at this time. Therefore, unlike case 1, the declared mileage price is not lower than the target object. The capacity part participates in the frequency regulation market and part participates in the electric energy market.

[0213] Figure 7a and Figure 7b The following is a graph showing the change in charge of the energy storage system under two conditions. Focusing on the SOC at the 10th hour, in condition 1, the discharge depth of the energy storage system is 0.2, which is higher than 0.1 in condition 2. The energy storage system has more energy participating in the electric energy market and making more profits through peak-valley arbitrage. In addition, because the power grid system as a whole abandons wind and solar power during the noon period, charging at this time can get more price compensation. Therefore, in both cases, the energy storage chooses to charge during this period, with a lower charging price, in preparation for subsequent discharge arbitrage.

[0214] Figure 8 The structure block diagram of a scheduling device for an energy storage system according to an embodiment of the present application is schematically shown. Figure 8 As shown, an embodiment of the present application provides an energy storage system scheduling device, which may include:

[0215] Memory 810, configured to store instructions; and

[0216] The processor 820 is configured to call instructions from the memory 810 and implement the above-mentioned energy storage system scheduling method when executing the instructions.

[0217] An embodiment of the present application also provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned energy storage system scheduling method.

[0218] Through the above technical solution, the processor 820 obtains various bidding behaviors of each target object in the frequency modulation market. Various bidding behaviors can be obtained based on the historical bidding records of the corresponding target object in the frequency modulation market; the processor 820 calculates the revenue of each type of bidding behavior in the frequency modulation market based on the current market data of the frequency modulation market, so as to predict the bidding information of each target object in the frequency modulation market; the processor 820 inputs the bidding information of each target object in the frequency modulation market into the frequency modulation market clearing model, and the frequency modulation market clearing model requires the minimization of the auxiliary cost of frequency modulation. Therefore, in the case of knowing the bidding information, the processor 820 can predict the winning bid information of the target energy storage system in the frequency modulation market; the processor 820 inputs the winning bid information into the pre-built decision model to calculate the optimal scheduling strategy of the target energy storage system. In this way, the production scheduling of a single target energy storage system can be optimized to improve the revenue of the target energy storage system.

[0219] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0220] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0221] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0222] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0223] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0224] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0225] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0226] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0227] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for scheduling an energy storage system, characterized in that: The method comprises: Obtaining various bidding behaviors of each target object in the frequency modulation market; the various bidding behaviors are obtained according to the historical bidding records of the corresponding target object in the frequency modulation market; According to the current market data of the frequency modulation market, the revenue of each type of bidding behavior in the frequency modulation market is calculated respectively, so as to predict the bidding information of each target object in the frequency modulation market; Inputting the bidding information of each target object in the frequency regulation market into the frequency regulation market clearing model to predict the winning bid information of the target energy storage system in the frequency regulation market; the frequency regulation market clearing model aims to minimize the frequency regulation service cost; The winning bid information is input into a pre-built decision model to calculate the optimal dispatching strategy for the target energy storage system; the decision model aims to maximize the sum of the revenue in at least the electric energy market and the frequency regulation market.

2. The method according to claim 1, characterized in that The steps of constructing the frequency modulation market clearing model include: The frequency modulation market clearing model is constructed according to formula (1): Among them, C Reg The cost of frequency modulation services, is the winning bid capacity of each target object, The winning mileage for each target object, is the bidding capacity unit price of each target object, is the bid mileage unit price for each target object, is the winning bid capacity of the target energy storage system, is the winning mileage of the target energy storage system, is the winning capacity unit price of the target energy storage system, is the winning mileage unit price of the target energy storage system, Ω is the set of various bidding behaviors of multiple target objects in the frequency modulation market, i is the number of the target object, t is the clearing time, and T is the total scheduling time.

3. The method according to claim 2, characterized in that The method further comprises: According to formula (2), formula (3) and formula (4), the winning mileage of each target object and the winning capacity of each target object are constrained: in, is the bidding capacity of each target object, is the dual variable corresponding to the lower limit constraint of the bid capacity of each target object, is the dual variable corresponding to the upper limit constraint of the winning capacity of each target object, The bid mileage for each target object, is the dual variable corresponding to the lower limit constraint of the mileage of each target object, is the dual variable corresponding to the upper limit constraint of the winning mileage of each target object, is the frequency modulation mileage multiplier of each target object, is the dual variable corresponding to the constraint; And / or, constraining the winning mileage of the target energy storage system and the winning capacity of the target energy storage system according to formula (5), formula (6) and formula (7): in, is the bid capacity of the target energy storage system, is the bid mileage of the target energy storage system, is the frequency regulation mileage multiplier of the target energy storage system, is the dual variable corresponding to the minimum value constraint of the winning capacity of the target energy storage system, is the dual variable corresponding to the maximum value constraint of the winning capacity of the target energy storage system, is the dual variable corresponding to the minimum mileage constraint of the target energy storage system, is the dual variable corresponding to the maximum mileage constraint of the target energy storage system, is a dual variable corresponding to the maximum value constraint of the winning capacity of the target energy storage system corresponding to the winning mileage; And / or, constrain the capacity and mileage required by the power grid system according to formula (8) and formula (9): in, is the capacity required by the power grid system at time t, The dual variable corresponding to the upper limit constraint of the bidding capacity of each target object, is the mileage required by the power grid system at time t, The dual variable corresponding to the upper limit constraint of the bidding mileage of each target object.

4. The method according to any one of claims 1 to 3, characterized in that The steps of constructing the decision model include: The decision model is constructed according to formula (10): Among them, C E is the profit value of the target energy storage system, which includes the electric energy profit and frequency regulation profit of the target energy storage system. is the power purchased by the target energy storage system at time t, is the electricity sales power of the target energy storage system at time t, is the unit price of electricity purchased by the target energy storage system at time t, is the electricity selling price of the target energy storage system at time t, is the capacity compensation unit price, is the unit price of photovoltaic consumption compensation, Δt is the scheduling time interval, To clear the capacity unit price, is the winning bid capacity of the target energy storage system, To clear the mileage unit price, is the winning mileage of the target energy storage system, c j The degradation cost.

5. The method according to claim 4, characterized in that The method further comprises: According to formula (11), formula (12), formula (13) and formula (14), the charging and discharging power of the target energy storage system is constrained: Among them, θ t is the binary variable for charging and discharging of the target energy storage system, is the upper limit of the charging power of the target energy storage system, is the upper limit of the discharge power of the target energy storage system, is the upper limit of the bidding capacity of the target energy storage system at time t, the bid capacity declared for the target energy storage system, the bid mileage declared for the target energy storage system, is the call rate of the frequency regulation mileage of the target energy storage system; And / or, constraining the charge of the target energy storage system according to formula (15), formula (16), formula (17) and formula (18): Among them, soc j,t is the charge of the target energy storage system at time t in segment j, is the upper limit of the charge in segment j, SOC min is the minimum value of the charge of the target energy storage system, SOC max is the maximum value of the charge of the target energy storage system, E cap is the capacity of the target energy storage system, and η is the charge and discharge efficiency of the target energy storage system; And / or, constraining the declared price of the target energy storage system according to formula (19) and formula (20): in, is the lower limit of the declared capacity unit price of the target energy storage system, is the upper limit of the declared capacity unit price of the target energy storage system, is the lower limit of the declared mileage unit price of the target energy storage system, The upper limit of the unit price of mileage reported for the target energy storage system.

6. The method according to any one of claims 1 to 3, characterized in that The calculation of the revenue of each type of bidding behavior in the frequency modulation market according to the current market data of the frequency modulation market includes: The benefit of each target object is determined according to formula (21): in, is the income of each target object at time t, a t is the set of bidding behaviors of each target object at time t, s t is the current market data at time t, and f is the long short-term memory neural network.

7. The method according to claim 6, characterized in that The historical bidding records include historical bidding capacity unit price, historical bidding mileage unit price, historical bidding capacity and historical bidding mileage; the bidding information includes bidding capacity unit price, bidding mileage unit price, bidding capacity and bidding mileage; the winning bid information includes winning mileage and winning bid capacity; the current market data includes market load, clearing capacity unit price, clearing mileage unit price, clearing capacity and clearing mileage.

8. The method according to any one of claims 1 to 3, characterized in that The acquisition of various bidding behaviors of each target object in the frequency modulation market includes: Clustering the historical bidding records based on the K-mediods clustering algorithm to determine various bidding behaviors of each target object in the frequency modulation market; The step of predicting the bidding information of each target object in the frequency modulation market comprises: Determine the state transition probability of each type of bidding behavior according to the revenue of each type of bidding behavior in the frequency modulation market; Based on the state transition probability and the roulette algorithm, the bidding information of each target object in the frequency modulation market is predicted.

9. An energy storage system scheduling device, characterized in that: include: a memory configured to store instructions; as well as A processor is configured to call the instruction from the memory and implement the energy storage system scheduling method according to any one of claims 1 to 8 when executing the instruction.

10. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions for enabling a machine to execute the energy storage system scheduling method according to any one of claims 1 to 8.

Images