A data-driven power market transaction strategy optimization method and system

Abstract

The application relates to the technical field of power system automation, and discloses a data-driven power market transaction strategy optimization method and system. A provincial inter-market power market transaction simulation environment containing multiple market members is constructed, a transaction declaration boundary is calculated based on a deep reinforcement learning transaction declaration auxiliary decision model according to the constructed provincial inter-market power market simulation transaction environment, the model parameters are updated according to a transaction strategy benefit evaluation result, a corresponding provincial inter-provincial and provincial market transaction optimization model is constructed according to the calculated provincial inter-market power market transaction declaration boundary, and a final transaction declaration quantity of a target province in the provincial inter-market power market is obtained according to an optimization result of the model, so that the provincial inter-market power market transaction declaration quantity is automatically calculated, human resources are saved, the economic efficiency and cleanness of power grid operation of the target province are improved, and the economic benefits of the target province participating in the provincial inter-market power market are further improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system automation, in particular to a data-driven power market transaction strategy optimization method and system. BACKGROUND

[0002] The northwest region of China is rich in new energy resources (including wind power, photovoltaic, etc.), while the power load center is located in the central and eastern regions, and the power supply and demand presents a reverse distribution characteristic. In recent years, the new energy installed capacity of the northwest power grid has rapidly increased, and it has initially possessed the characteristics of a new power system with a high proportion of new energy power. Taking the Gansu power grid as an example, as of August 2024, its new energy installed capacity accounted for 62.98%, and the new energy power generation accounted for 37.06% (after deducting the outflow, the proportion is 30.94%), ranking second in the country. The power fluctuation and randomness caused by large-scale new energy access pose a serious challenge to regional power supply and demand balance. In order to alleviate the widening of the power supply and demand scissors difference and the shortage of flexible adjustment resources within the province, it is urgent to realize the optimal allocation of power resources in a larger range and carry out large-scale cross-provincial and cross-regional power trading. At present, the northwest power grid has initially established a regional electricity market in the northwest region and an auxiliary service trading mechanism covering peak shaving, standby and other transaction varieties. Considering that the provincial power market reform in the northwest region is also in the process of gradual deepening, in the future, a provincial-intra-provincial two-level power market mode with unified market and two-level operation will be formed.

[0003] The introduction of the two-level power market mode makes the coupling relationship between inter-provincial and intra-provincial power trading closer. Since most power users will not be able to directly participate in market bidding in the future for a long time, in the actual inter-provincial market operation process, most of the transactions are carried out by provincial power companies and their trading centers on behalf of intra-provincial power users, which means that provincial power companies need to take the inter-provincial power trading results as boundary conditions for carrying out intra-provincial power trading. The inter-provincial trading results will inevitably have some impact on intra-provincial trading. However, at present, most of the inter-provincial power market transactions are declared by dispatchers according to their own historical experience. In the period of sharp climbing / sliding of new energy power, it is easy to cause misjudgment of purchase and sale decisions, for example, too little external sales will cause insufficient generation space for intra-provincial new energy units, and the utilization rate of new energy will decrease; when the output level of new energy is lower than expected, insufficient repurchase will trigger the safeguard purchase mechanism (i.e. the regional power grid organizes emergency power support for the regional resources) and cause economic losses.

[0004] To address the aforementioned issues, this invention proposes a data-driven method for optimizing inter-provincial power market trading strategies. This method targets intraday trading in the inter-provincial power market and can automatically calculate the inter-provincial power purchase and sale transaction declaration volume for the next 4 hours (16 time periods) based on the target province's future load power supply demand and renewable energy consumption demand, taking into account factors such as the trading intentions of inter-provincial market participants and safe power flow transmission constraints. Compared to the existing manual declaration method, this method can significantly save human resources and improve economic efficiency. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a data-driven method and system for optimizing electricity market trading strategies. It can automatically calculate the inter-provincial electricity market trading volume for intraday transactions, saving manpower compared to the original manual reporting method. The provided inter-provincial electricity market trading strategy can also consider the intra-provincial electricity market trading needs of the target province, improving the economic efficiency and cleanliness of the target province's power grid operation. Based on deep reinforcement learning data-driven technology, it mines historical market transactions and renewable energy power forecasts, incorporating factors such as renewable energy / load forecast deviations and market participants' trading intentions into the trading strategy considerations, thereby further enhancing the economic benefits of the target province participating in the inter-provincial electricity market.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, this invention provides a data-driven method for optimizing electricity market trading strategies, comprising the following steps:

[0008] S1: Construct a simulated inter-provincial electricity market transaction environment that includes multiple market members. This step is used to simulate the response strategies of provincial power companies (or other provincial agents) other than the target province (i.e., the provincial power company served by this invention) to the transaction information declared by the target province, as well as the clearing process of the inter-provincial electricity market.

[0009] S2: Using the inter-provincial electricity market trading simulation environment built in S1, calculate the boundary of the trading declaration volume based on the trading declaration auxiliary decision model based on deep reinforcement learning, and update the model parameters according to the trading strategy benefit evaluation results;

[0010] S3: Based on the boundary of inter-provincial power market transaction declaration volume calculated in S2, construct the corresponding inter-provincial-intra-provincial two-level market transaction optimization model, and obtain the final transaction declaration volume submitted by the target province in the inter-provincial power market according to its optimization results.

[0011] Preferably, S1 includes the following S:

[0012] S1.1: Construct the approximate bidding function of the other provincial power companies (or other provincial agents) in the inter-provincial electricity market except for the target province:

[0013] (1)

[0014] (2)

[0015] (3)

[0016] (4)

[0017] (5)

[0018] The symbols in all the above formulas are for the bidding of the provincial power company (hereinafter referred to as the province ) in the time period , so the subscripts and are omitted. In formulas (1) and (2), and represent the load supply reserve and the new energy consumption reserve of the province in the time period , respectively. and represent the upper limit and the lower limit of the total power of all power generating units (thermal power, hydroelectric power, energy storage, etc.) except for new energy. is the upper limit of the total power of new energy units according to ultra-short-term prediction information. In formulas (3) and (4), and are the load supply reserve and the new energy consumption reserve after standardization, and represent the standardization basis values corresponding to the load supply reserve and the new energy consumption reserve, respectively. is a clipping function, which aims to control and within the range of [0, 1]. In formula (5), is the bidding amount of the province in the time period . If is positive, it means that the province buys electricity, and vice versa. and represent the minimum demand of the province for the power supply reserve and the new energy consumption reserve, respectively. The bidding amount in formula (5) is related to , in a piecewise linear function relationship, and the three segments in the function represent three cases: purchasing electricity when power supply margin is insufficient , purchasing electricity when new energy consumption margin is sufficient , and selling electricity when new energy consumption margin is insufficient . Parameters , and correspond to the starting values of the three function segments, respectively , and represent the slopes of the three function segments, respectively.

[0019] S1.2: Based on the transaction declaration amount calculated according to formula (5) and the historical actual declaration amount of the province , the undetermined parameters in formulas (3)-(5) (including , , , , , , , and ) are updated based on the gradient descent algorithm. Specifically, let the set of all undetermined parameters be , then can be represented as a function . The updating method of the parameter set is as follows:

[0020] (6)

[0021] (7)

[0022] wherein is the mean square error of the transaction declaration amount calculated according to formula (5) and the historical actual declaration amount is the updating step of the gradient descent method, is the partial derivative of to , is the partial derivative of to the parameter . Repeat the calculation of formulas (6) and (5) until converges to a stable level.

[0023] S1.3: Based on the parameters obtained in S1.2, the transaction declaration amount of market members in markets other than the target province is calculated according to formula (5). The inter-provincial power market simulation is performed combined with the transaction declaration amount of the target province (the specific calculation method is shown in S2), and the transaction amount result is obtained. In the simulation, the price of each province is the price declared by it in the day-ahead stage, and the inter-provincial power market simulation process refers to the current inter-provincial power market trading rules of Northwest China Power Grid. Based on the "high-low matching" mode, the transaction pair with the largest price difference is preferentially cleared until the transaction amount of the power buyer or seller is zero (when the price differences of multiple transaction pairs are the same, the transaction amount is proportionally distributed according to the transaction declaration amount), and the clearing price is the average price of the final cleared power buying and selling transaction pair.

[0024] Preferably, S2 includes the following S:

[0025] S2.1: Initialize the transaction declaration auxiliary decision-making model. The model is composed of 3 neural networks, including a value network, a strategy network and a historical state memory network. Based on normal probability distribution sampling, the parameters of the value network , the parameters of the strategy network and the parameters of the historical state memory network are randomly generated.

[0026] S2.2: Initialize the interaction data pool of the transaction declaration auxiliary decision-making model and the inter-provincial power market transaction simulation environment , and let be an empty set.

[0027] S2.3: Randomly select a historical arbitrary period and start calculation. For the period , the observation state vector of the transaction declaration decision-making model is constructed. The specific information contained in it is as follows:

[0028] (8)

[0029] Wherein, represents the total power upper limit prediction value of new energy units corresponding to the future 16 periods (each period lasts for 15 minutes, and there are 16 periods in total for the next 4 hours), represents the system load power prediction value of the future 16 periods, represents the total power plan of the interconnection line for the future 16 periods, represents the meteorological forecast information prediction value for the future 16 periods, represents the transaction quotation for the future 16 periods. and respectively represent the observed transaction amount and clearing price of the last period . and are the observed total power of new energy units and the observed system load power in the last time period . and are the observed total power of new energy units and the observed system load power in the last time period . and In addition to the data corresponding to the target province, the shared information of other market members in the inter-provincial power market also needs to be included, which can be obtained through the data acquisition and monitoring control system of the regional power grid. is the historical state information encoding output by the historical state memory network in the last time period .

[0030] S2.4: The observed state vector formed in S2.3 is input into the strategy network of the transaction declaration auxiliary decision-making model, and through this network, the inter-provincial power market transaction declaration amount boundary corresponding to the next 16 time periods is generated .

[0031] S2.5: The inter-provincial power market transaction declaration amount boundary generated in S2.4 is taken as the final transaction declaration amount and is passed to the inter-provincial power market transaction simulation environment formed in S1 to perform the inter-provincial power market simulation out of the clearing in S1.3, and the transaction electricity and the clearing price in time period are obtained.

[0032] S2.6: According to the inter-provincial power market clearing result in S2.5, the benefit evaluation of the transaction strategy in S3.3 is performed, which is specifically expressed using the following formula:

[0033] (9)

[0034] wherein, is the estimated benefit of the inter-provincial transaction strategy in time period , , , and The definitions of and are the same as the formulas (1)-(5) in S1. The first term in the formula represents the purchase and sale electricity cost, the second term represents the penalty for insufficient load supply margin in time period , and the third term represents the penalty for insufficient new energy consumption margin. and respectively represent the weights of the two penalty terms.

[0035] S2.7: The observed state vector ​​​Input into the historical state memory network while generating historical state information encoding , the sum power prediction value of new energy units and the system load power prediction value .

[0036] S2.8: Form a data packet according to the results formed in S2.3 to S2.7 and add it to the interaction data pool . The data packet specifically contains the following information:

[0037] (10)

[0038] wherein, is the observation state vector corresponding to the transaction period , which is constructed in the same way as .

[0039] S2.8.1: Repeat S2.3-2.8 until there is a considerable number of data packets in the experience pool.

[0040] S2.9: Use the data packets in the interaction data pool to update the policy network parameters, value network parameters and historical state memory network parameters of the transaction declaration decision model:

[0041] (11)

[0042] (12)

[0043] (13)

[0044] wherein, , , is the update step size of the parameter , , , , , is the update target of the parameter, , , , , , is the update gradient of the parameter , , , , , respectively represent as follows:

[0045] (14)

[0046] (15)

[0047] (16)

[0048] wherein, is a parameter the advantage value of the trading strategy before update in time period , and is a parameter the trading declaration volume before and after update corresponding probability ratio value, is an algorithm parameter for controlling the gradient descent amplitude. is the discount cumulative sum of the reward function in all data packets in the interaction data pool, is the output value of the value network in the trading declaration auxiliary decision-making model.

[0049] S2.10: Repeat S2.2-2.9. Until the parameters of each type of network in the trading declaration decision-making model converge to a stable level.

[0050] Preferably, the S3 comprises the following S:

[0051] S3.1: For the time period in which the target province currently needs to make inter-provincial power market declaration, construct the grid observation state vector (the calculation method is the same as S2.3). Input the observation state vector into the trading declaration decision-making model in S2.10 whose parameters have converged to a stable level, to generate the inter-provincial power market trading declaration volume boundary of the next 16 time periods (the calculation method is the same as S2.4). According to the differences in time period and trading object, it can be further expressed as the following matrix:

[0052] (17)

[0053] wherein represents the maximum trading declaration volume of the target province to the province in time period , and represents the maximum trading declaration volume of the target province to the province in time period , and so on. is the total number of market members that can trade with the target province.

[0054] S3.2: Construct the inter-provincial-intra-provincial two-level market trading optimization mathematical model of the target province as follows:

[0055] (18)

[0056] (19)

[0057] (20)

[0058] (twenty one)

[0059] (twenty two)

[0060] Where equation (18) is the objective function of the constructed mathematical model. The objective function of the mathematical model for clearing the electricity market within the target province is: Indicates the contact line At any moment The volume of inter-provincial electricity market transactions declared. The target province at any time The transaction quotes. Equations (19)-(23) are the constraints of the constructed mathematical model. Equation (19) is the power limit constraint of the connecting line after inter-provincial transactions. This indicates the initial communication power plan prior to inter-provincial transactions. This represents the revised tie-line power plan obtained after inter-provincial transactions. Equation (20) is the constraint on the tie-line power plan adjustment. It is all related to provinces A collection of connected connecting lines, The target province in S3.1 during the time period For provinces The maximum transaction declaration volume. Equation (21) is the power flow restriction constraint for power grid safety sections. It is a time period The corresponding power flow transfer factor matrix of power grid nodes-safety sections, It is a node At any moment Node injection power, It is the link-node correlation factor matrix. and These are the safety cross sections. At any moment The lower and upper limits of power flow transmission. Equation (22) is the power balance constraint of the target province, where each component includes the generating power of units within the province and the power of the tie line after inter-provincial power market transactions. It refers to all generating units (thermal power, hydropower, energy storage, etc.) except for new energy sources at any given time. Total power, It is a new energy unit at all times Total power, is the system load power of the target province at time .

[0061] S3.3: Solving the two-level market transaction optimization mathematical model constructed in S3.2 using a mathematical programming solver to obtain the transaction declaration amount of each tie line in the inter-provincial power market . If the inter-provincial power market requires transaction power declaration in units of tie lines, the transaction declaration amount can be directly submitted to the inter-provincial power market; if the inter-provincial power market requires declaration of purchase and sale of power in different market members, the final transaction declaration amount is calculated as follows:

[0062] (24)

[0063] wherein represents the final transaction declaration amount of the target province to the province at time period .

[0064] If the inter-provincial power market only allows declaration of total provincial purchase and sale of power, the final transaction declaration amount is calculated as follows:

[0065] (25)

[0066] wherein represents the final transaction declaration total amount of the target province at time period .

[0067] Preferably, in S1-3, an inter-provincial power market transaction simulation environment containing multiple market members is constructed, a transaction declaration auxiliary decision-making model is constructed and network parameter updating is performed based on the inter-provincial power market transaction simulation environment; a transaction declaration amount boundary is calculated using the transaction declaration auxiliary decision-making model after parameter convergence, an inter-provincial-intra-provincial two-level market transaction optimization model is constructed based on the transaction declaration amount boundary, and finally the inter-provincial power market transaction declaration amount is obtained by solving the model.

[0068] Preferably, the inter-provincial power market transaction simulation environment containing multiple market members, the approximate transaction declaration amount function construction method of the provincial power company (or provincial agent), and the updating method of the undetermined parameters in the function.

[0069] Preferably, the transaction declaration auxiliary decision-making model, the observation state vector construction method for inter-provincial power market transaction strategy optimization, the model structure containing a historical state memory network, the benefit evaluation method of the transaction strategy output by the model, and the parameter updating method of the neural network in the model.

[0070] Preferably, the inter-provincial-intra-provincial two-level market transaction optimization model adds a constraint condition of the output transaction declaration amount boundary of the transaction declaration auxiliary decision model, and a constraint condition of the power balance of the tie-line power after considering the power generation of the unit in the province and the inter-provincial power market transaction.

[0071] In a second aspect, the application discloses a data-driven power market transaction strategy optimization system, which comprises a construction module, an evaluation module and an optimization module.

[0072] The construction module is configured to construct an inter-provincial power market transaction simulation environment comprising multiple market members, and the construction module is connected to the evaluation module through a network.

[0073] The evaluation module is configured to calculate a transaction declaration amount boundary based on a deep reinforcement learning-based transaction declaration auxiliary decision model according to the constructed inter-provincial power market transaction simulation environment, and the evaluation module is connected to the optimization module through a network.

[0074] The optimization module is configured to construct an inter-provincial-intra-provincial two-level market transaction optimization model based on the obtained inter-provincial power market transaction declaration amount boundary, and obtain a final transaction declaration amount of a target province in the inter-provincial power market based on an optimization result of the inter-provincial-intra-provincial two-level market transaction optimization model.

[0075] In a third aspect, the application discloses a computer device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the data-driven power market transaction strategy optimization method when executing the computer program.

[0076] In a fourth aspect, the application discloses a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the data-driven power market transaction strategy optimization method when executed by a processor.

[0077] Compared with the prior art, the application provides a data-driven power market transaction strategy optimization method and system, which has the following beneficial effects:

[0078] The application is based on a data-driven power market transaction strategy optimization method, which builds an inter-provincial power market transaction simulation environment containing multiple market members, calculates the transaction declaration quantity boundary based on the deep reinforcement learning transaction declaration auxiliary decision model according to the built inter-provincial power market transaction simulation environment, updates the model parameters according to the transaction strategy benefit evaluation results, builds the corresponding inter-provincial-intra-provincial two-level market transaction optimization model according to the calculated inter-provincial power market transaction declaration quantity boundary, and obtains the final transaction declaration quantity of the target province in the inter-provincial power market according to the optimization results, so as to realize automatic calculation of the inter-provincial power market transaction declaration quantity, save human resources, improve the economic efficiency and cleanliness of the target provincial power grid, and further improve the economic benefit of the target province participating in the inter-provincial power market.

[0079] The application can automatically calculate the inter-provincial power market transaction declaration quantity for the next 4 hours every 15 minutes for the intra-day transaction of the inter-provincial power market, which saves human resources compared with the original manual declaration method; the inter-provincial power market transaction strategy provided by the application can take into account the intra-provincial power market transaction demand of the target province, improve the economic efficiency and cleanliness of the target provincial power grid. Specifically, in the period of large new energy power generation, the power transmission capacity to other provinces is increased to expand the new energy power generation space and improve the utilization rate of new energy; in the period of small new energy power generation, it is judged whether the load supply margin of the target province is sufficient, and if the margin is insufficient, the power purchase demand is initiated in the inter-provincial power market in advance to avoid the emergency power purchase cost caused by the untimely transaction declaration.

[0080] The application based on the data-driven technology of deep reinforcement learning can mine historical market transactions, new energy power prediction and other data, and can consider factors such as new energy / load prediction deviation and market member transaction willingness in the transaction strategy consideration scope, so as to further improve the economic benefit of the target province participating in the inter-provincial power market. BRIEF DESCRIPTION OF DRAWINGS

[0081] Figure 1 It is a flowchart of the data-driven power market transaction strategy optimization method in the embodiment of the application.

[0082] Figure 2 It is the total new energy power upper limit statistics of the target province in embodiment 1 of the application.

[0083] Figure 3 It is the system load power statistics of the target province in embodiment 1 of the application.

[0084] Figure 4 It is the transaction declaration and transaction of the target province participating in the inter-provincial power market in embodiment 1 of the application.

[0085] Figure 5 It is the tie-line power statistics of the target province in embodiment 1 of the application.

[0086] Figure 6 The total power upper limit and the planned power of the new energy in the target province in the embodiment 2 of the present application are counted.

[0087] Figure 7 The transaction declaration and transaction completion of the target province participating in the inter-provincial power market in the embodiment 2 of the present application are counted.

[0088] Figure 8 The transaction declaration and transaction completion of the target province participating in the inter-provincial power market in the embodiment 2 of the present application are counted.

[0089] Figure 9 The system flowchart in the embodiment of the present application is shown. DETAILED DESCRIPTION

[0090] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0091] Please refer to Figures 1-8 A data-driven based power market transaction strategy optimization method, comprising the following steps:

[0092] S1: constructing an inter-provincial power market transaction simulation environment containing multiple market members, which is used to simulate the response strategy of the provincial power company (or other provincial agent) other than the target province (i.e. the provincial power company served by the present application) to the transaction information declared by the target province, and the clearing process of the inter-provincial power market;

[0093] S1 comprises the following S:

[0094] S1.1: constructing the approximate transaction declaration amount function of the provincial power company (or other provincial agent) other than the target province in the inter-provincial power market:

[0095] (1)

[0096] (2)

[0097] (3)

[0098] (4)

[0099] (5)

[0100] The symbols in all the above formulas are for the provincial power company (Hereinafter, the province is referred to as ) in the time period , so the subscript and are omitted. In formula (1) and formula (2), and respectively represent the load supply reserve margin and the new energy consumption margin of the province in the time period . and respectively represent the upper limit of the total power and the lower limit of the total power of all power generation units (thermal power, hydropower, energy storage, etc.) except new energy. is the upper limit of the total power of the new energy unit according to the ultra-short-term prediction information. In formula (3) and formula (4), and are the load supply reserve margin and the new energy consumption margin after standardization, and respectively represent the standardization basis values corresponding to the load supply reserve margin and the new energy consumption margin. is a clipping function, whose purpose is to control and within the range of [0, 1]. In formula (5), is the trading declaration amount of the province in the time period . If is positive, it means that the province buys electricity, otherwise it means that it sells electricity. and respectively represent the minimum demand of the province for the power supply reserve margin and the new energy consumption margin. The declared trading amount in formula (5) has a piecewise linear function relationship with , The three segments in the function represent three cases: buying electricity when the power supply margin is insufficient , buying electricity when the new energy consumption margin is sufficient , and selling electricity when the new energy consumption margin is insufficient . Parameters , and respectively correspond to the starting values of the three function segments, , and respectively represent the slopes of the three function segments.

[0101] S1.2: The trading declaration amount calculated in combination with formula (5), and the historical actual declaration amount of the province , the undetermined parameters in equations (3)-(5) are updated based on the gradient descent algorithm , , , , , , and ). Specifically, let the set of all undetermined parameters be , then can be represented as a function . The update method of the parameter set is:

[0102] (6)

[0103] (7)

[0104] where is the mean square error of the transaction declaration volume calculated by equation (5) and the historical actual declaration volume. is the update step of the gradient descent method, is the partial derivative of , is the partial derivative of to the parameter . Repeat the calculation of equations (6) and (5) until converges to a stable level.

[0105] S1.3: Based on the parameters obtained in S1.2, calculate the transaction declaration volume of all market members outside the target province based on equation (5). Combine the transaction declaration volume of the target province (see S2 for specific calculation method), execute the inter-provincial power market simulation and get the transaction volume result. Among them, the price of each province is the price declared in the day-ahead stage, and the inter-provincial power market simulation process refers to the current inter-provincial power market trading rules of Northwest China Power Grid, based on the "high-low matching" method, the transaction with the largest price difference is cleared first, until the transaction volume of the buyer or seller is zero (when the price difference of multiple transaction pairs is the same, the transaction volume is distributed in proportion to the transaction declaration volume), and the clearing price is the average price of the final clearing buy-sell transaction pair

[0106] S2: Use the inter-provincial power market trading simulation environment constructed in S1 to calculate the transaction declaration volume boundary based on the deep reinforcement learning-based transaction declaration auxiliary decision-making model, and update the model parameters according to the transaction strategy benefit evaluation results;

[0107] S2 includes the following S:

[0108] S2.1: Initialize the transaction declaration auxiliary decision-making model. The model is composed of 3 neural networks, including a value network, a strategy network and a historical state memory network. Based on normal probability distribution sampling, the parameters of the value network , the parameters of the strategy network and the parameters of the historical state memory network are randomly generated.

[0109] S2.2: Initialize the interaction data pool of the transaction declaration auxiliary decision-making model and the inter-provincial power market transaction simulation environment , let be an empty set.

[0110] S2.3: Randomly select a historical arbitrary period and start calculation. For the period , the observation state vector of the transaction declaration decision-making model is constructed. The specific information contained in it is as follows:

[0111] (8)

[0112] Wherein, represents the total upper limit predicted value of the new energy unit power corresponding to the future 16 periods (each period lasts for 15 minutes, a total of 16 periods for 4 hours in the future), represents the system load power prediction value of the future 16 periods, represents the total power plan of the interconnection line for the future 16 periods, represents the meteorological forecast information prediction value for the future 16 periods, represents the transaction quotation for the future 16 periods. and respectively represent the observed transaction electricity and clearing price of the last period . and are the observed new energy unit total power upper limit prediction deviation and system load power prediction deviation of the last period . and are the observed new energy unit total power actual value and system load power actual value of the last period . Except for the data corresponding to the target province, and also need to include the shared information of other market members of the inter-provincial power market, which can be obtained through the data acquisition and monitoring control system of the regional power grid. is the historical state information code output by the historical state memory network in the last period .

[0113] S2.4: The observation state vector formed in S2.3 is input into the strategy network of the transaction declaration auxiliary decision-making model, and the inter-provincial power market transaction declaration amount boundary corresponding to the future 16 time periods is generated through the network . .

[0114] S2.5: The inter-provincial power market transaction declaration amount boundary generated in S2.4 is input into the transaction strategy network of the transaction declaration auxiliary decision-making model, and the inter-provincial power market transaction declaration amount boundary corresponding to the future 16 time periods is generated through the network . . . .

[0115] S2.6: According to the inter-provincial power market clearing result of S2.5, the benefit evaluation of the transaction strategy of S3.3 is carried out, which is specifically expressed by the following formula:

[0116] (9)

[0117] wherein, is the estimated benefit of the inter-provincial transaction strategy in time period , , , and are the same as the definitions of formulas (1)-(5) of S1. The first term in the formula represents the purchase and sale electricity cost, the second term represents the penalty for insufficient load supply margin in time period , and the third term represents the penalty for insufficient new energy consumption margin. and respectively represent the weights of the two penalty terms.

[0118] S2.7: The observation state vector is input into the historical state memory network, and the historical state information code , the total power prediction value of new energy units , and the system load power prediction value are generated at the same time.

[0119] S2.8: According to the results formed in S2.3 to S2.7, a data packet is formed and added to the interaction data pool . The data packet specifically contains the following information:

[0120] (10)

[0121] wherein, is the corresponding transaction time period The observation state vector is constructed in the same way as .

[0122] S2.8.1: Repeat S2.3-2.8 until there is a considerable number of data packets in the experience pool.

[0123] S2.9: Update the strategy network parameters, value network parameters, and historical state memory network parameters of the transaction declaration decision model using the data packets in the interaction data pool:

[0124] (11)

[0125] (12)

[0126] (13)

[0127] wherein, , , is the update step size of the parameter , , , , is the update target of the parameter, , , , , is the update gradient of the parameter , , , , respectively as follows:

[0128] (14)

[0129] (15)

[0130] (16)

[0131] wherein, is the advantage value of the transaction strategy before the update in the time period , is the probability ratio of the transaction declaration quantity before and after the update , is an algorithm parameter for controlling the gradient descent amplitude. is the reward function in all data packets in the interaction data pool​​​ The total cumulative discount This is the output value of the value network in the transaction reporting auxiliary decision-making model.

[0132] S2.10: Repeat S2.2-2.9 until the parameters of all networks in the transaction declaration decision model converge to a stable level;

[0133] S3: Based on the boundary of inter-provincial power market transaction declaration volume calculated in S2, construct the corresponding inter-provincial-intra-provincial two-level market transaction optimization model, and obtain the final transaction declaration volume submitted by the target province in the inter-provincial power market according to its optimization results;

[0134] S3 includes the following S:

[0135] S3.1: For the time period when the target province needs to submit inter-provincial electricity market applications. Construct power grid observation state vector (Calculation method is the same as S2.3). The observed state vector... The transaction declaration decision model, whose parameters have converged to a stable level in S2.10, generates the boundary of inter-provincial electricity market transaction declaration volume for the next 16 time periods. (Calculation method is the same as S2.4). Depending on the time period and the trading counterparty, This can be further represented as the following matrix:

[0136] (17)

[0137] in Indicates the target province during the time period For provinces The maximum transaction declaration volume, Indicates the target province during the time period For provinces The maximum transaction volume, and so on. It is the total number of market members who can trade with the target province.

[0138] S3.2: The following is the mathematical model for optimizing inter-provincial and intra-provincial market transactions in the target province:

[0139] (18)

[0140] (19)

[0141] (20)

[0142] (twenty one)

[0143] (twenty two)

[0144] Where equation (18) is the objective function of the constructed mathematical model. The objective function of the mathematical model for clearing the electricity market within the target province is: Indicates the contact line At any moment The volume of inter-provincial electricity market transactions declared. The target province at any time The transaction quotes. Equations (19)-(23) are the constraints of the constructed mathematical model. Equation (19) is the power limit constraint of the connecting line after inter-provincial transactions. This indicates the initial communication power plan prior to inter-provincial transactions. This represents the revised tie-line power plan obtained after inter-provincial transactions. Equation (20) is the constraint on the tie-line power plan adjustment. It is all related to provinces A collection of connected connecting lines, The target province in S3.1 during the time period For provinces The maximum transaction declaration volume. Equation (21) is the power flow restriction constraint for power grid safety sections. It is a time period The corresponding power flow transfer factor matrix of power grid nodes-safety sections, It is a node At any moment Node injection power, It is the link-node correlation factor matrix. and These are the safety cross sections. At any moment The lower and upper limits of power flow transmission. Equation (22) is the power balance constraint of the target province, where each component includes the generating power of units within the province and the power of the tie line after inter-provincial power market transactions. It refers to all generating units (thermal power, hydropower, energy storage, etc.) except for new energy sources at any given time. Total power, It is a new energy unit at all times Total power, The target province at any time The system load power.

[0145] S3.3: Use a mathematical programming solver to solve the two-level market transaction optimization mathematical model constructed in S3.2 to obtain the transaction declaration volume of each interconnection line in the inter-provincial power market. If the inter-provincial electricity market requires that the transaction volume be declared on a per-connection line basis, then the transaction declaration volume... The final transaction declaration quantity can be directly submitted to the inter-provincial power market, and if the inter-provincial power market requires declaration of the purchase and sale of electricity in different market members, the final transaction declaration quantity is calculated as follows:

[0146] (23)

[0147] Wherein represents the final transaction declaration quantity of the target province in the time period .

[0148] If the inter-provincial power market only allows declaration of the total purchase and sale of electricity in the province, the final transaction declaration quantity is calculated as follows:

[0149] (24)

[0150] Wherein represents the final transaction declaration quantity of the target province in the time period .

[0151] In S1~3, an inter-provincial power market transaction simulation environment containing multiple market members is constructed, a transaction declaration auxiliary decision-making model is constructed, and network parameter updating is performed based on the inter-provincial power market transaction simulation environment; the transaction declaration auxiliary decision-making model after parameter convergence is used to calculate the transaction declaration quantity boundary, an inter-provincial-intra-provincial two-level market transaction optimization model is constructed based on the transaction declaration quantity boundary, and finally the model is solved to obtain the inter-provincial power market transaction declaration quantity;

[0152] The inter-provincial power market transaction simulation environment containing multiple market members, the construction method of the approximate transaction declaration quantity function of the provincial power company (or the provincial agent), and the updating method of the undetermined parameters in the function;

[0153] The transaction declaration auxiliary decision-making model, the observation state vector construction method for inter-provincial power market transaction strategy optimization, the model structure containing a historical state memory network, the benefit evaluation method of the transaction strategy output by the model, and the parameter updating method of the neural network in the model;

[0154] The inter-provincial-intra-provincial two-level market transaction optimization model, the constraint condition construction method of the transaction declaration quantity boundary output by the transaction declaration auxiliary decision-making model, and the power balance constraint condition construction method considering the power of the intra-provincial unit and the power of the inter-provincial power market transaction after the tie line power;

[0155] As shown in Figure 9 , in another embodiment of the present application, a data-driven based power market transaction strategy optimization system implementation is provided, which includes a construction module, an evaluation module and an optimization module.

[0156] ​The construction module is configured to construct an inter-provincial electricity market transaction simulation environment containing multiple market members, and the construction module is connected to the evaluation module via a network;

[0157] The evaluation module is configured to calculate a transaction submission amount boundary based on a deep reinforcement learning-based transaction submission auxiliary decision-making model according to the constructed inter-provincial electricity market transaction simulation environment, and the evaluation module is connected to the optimization module via a network;

[0158] The optimization module is configured to construct a corresponding inter-provincial-intra-provincial two-level market transaction optimization model based on the obtained inter-provincial electricity market transaction submission amount boundary, and obtain a final transaction submission amount of a target province in an inter-provincial electricity market according to an optimization result of the model.

[0159] In another embodiment of the present application, a terminal device is provided, which includes a processor and a memory, the memory is configured to store a computer program, the computer program includes program instructions, and the processor is configured to execute the program instructions stored in the computer storage medium. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions to implement a corresponding method flow or a corresponding function; the processor in the embodiment of the present application can be used for the operation of the data-driven electricity market transaction strategy optimization method, including the following steps: constructing an inter-provincial electricity market transaction simulation environment containing multiple market members; using the constructed inter-provincial electricity market transaction simulation environment, calculating a transaction submission amount boundary based on a deep reinforcement learning-based transaction submission auxiliary decision-making model, and updating the model parameters according to the transaction strategy benefit evaluation result; based on the obtained inter-provincial electricity market transaction submission amount boundary, constructing a corresponding inter-provincial-intra-provincial two-level market transaction optimization model, and obtaining a final transaction submission amount of a target province in an inter-provincial electricity market according to an optimization result of the model.

[0160] In another embodiment of the present application, the present application further provides a storage medium, specifically a computer readable storage medium (Memory), which is a memory device in the terminal device, used for storing programs and data. It can be understood that the computer readable storage medium herein can include the built-in storage medium in the terminal device, and of course can also include the expansion storage medium supported by the terminal device. The computer readable storage medium provides a storage space, which stores the operating system of the terminal. Moreover, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and the instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory.

[0161] The one or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the data-driven power market transaction strategy optimization method in the above embodiments. The one or more instructions stored in the computer readable storage medium are loaded and executed by the processor to perform the following steps: constructing an inter-provincial power market transaction simulation environment including multiple market members; using the constructed inter-provincial power market transaction simulation environment, calculating a transaction submission amount boundary based on a deep reinforcement learning-based transaction submission auxiliary decision-making model, and updating the model parameters according to the transaction strategy benefit evaluation results; based on the calculated inter-provincial power market transaction submission amount boundary, constructing a corresponding inter-provincial-intra-provincial two-level market transaction optimization model, and obtaining the final transaction submission amount of the target province in the inter-provincial power market according to the optimization results of the model.

[0162] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take 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.) containing computer usable program code.

[0163] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0164] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0165] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0166] Finally, it should be noted that the above examples are merely used to illustrate the technical solutions of the present application rather than limit the technical solutions of the present application. Although the present application has been described in detail with reference to the above examples, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and any modification or replacement should be included in the protection scope of the claims of the present application.

[0167] For the purpose of explaining the present application, the technical solutions and advantages, the present application is further described in detail below with reference to the accompanying drawings.

[0168] The northwest Gansu power grid is selected as the target province for example analysis, which is specifically discussed in the following two examples.

[0169] (1) Example 1: January 24, 2024 (00:00-08:00)

[0170] AppendixFigure 2 This displays the predicted and measured curves of the total renewable energy power ceiling for the target province during the early morning hours of the day, with appendices. Figure 3 The actual system load curve and the predicted system load curve for the same time period are displayed, with appendix. Figure 4 This displays the inter-provincial electricity market declaration volume and actual transaction volume for each time period calculated using this invention, with appendix... Figure 5 The data displays the tie-line plan before the target province participates in the inter-provincial electricity market (day-ahead tie-line plan), the tie-line plan after participation (revised plan), and the historical actual tie-line power. It can be seen that the measured total power of renewable energy sources during this period is significantly lower than the predicted value, while the actual system load is significantly higher than the predicted value, indicating a severe shortage of power supply margin for the target province. Due to inaccurate forecasting information, the dispatcher failed to promptly request inter-provincial electricity market transactions, resulting in a negative tie-line power deviation, meaning the actual tie-line power is significantly lower than the day-ahead tie-line plan, with a total deviation of 7156 MWh. Based on the peak inter-provincial electricity market transaction price (450 RMB / MWh) and a 20% penalty for emergency power purchases, this would result in an additional economic loss of 644,000 RMB. After conducting inter-provincial electricity market transactions based on the calculation results of this invention, the deviation between the actual tie-line power and the revised plan is reduced to 2106 MWh, and the additional economic loss is reduced to 190,000 RMB, thus verifying the economic benefits of this invention in scenarios with insufficient load power supply margin.

[0171] (2) Example 2: Noon on July 11, 2024 (11:00-15:00)

[0172] Appendix Figure 6 This displays the predicted and measured curves of the total renewable energy power ceiling for the target province at noon on the same day, as well as the planned curves of renewable energy units before and after participating in the inter-provincial electricity market. (Appendix) Figure 7 This displays the inter-provincial electricity market declaration volume and actual transaction volume for each time period calculated using this invention, with appendix... Figure 8The inter-provincial tie-line plan before the target province participates in the inter-provincial power market (day-ahead tie-line plan), the inter-provincial tie-line plan after the target province participates in the inter-provincial power market (modified plan) and the actual tie-line power are shown. It can be seen that due to the insufficient adjustment capacity of flexible resources such as in-province thermal power in the period, the new energy unit plan curve is less than the upper limit curve, and therefore the target province has insufficient load consumption margin. At this time, the dispatcher does not raise an additional power selling request in the inter-provincial power market, and the new energy utilization rate is low. After the inter-provincial power market transaction according to the calculation result of the present application, the modified tie-line plan is higher than the day-ahead tie-line plan from 11:00 to 13:00, and therefore the additional planned new energy unit power generation reaches 1069 megawatt hours. According to the inter-provincial power market low valley transaction electricity price (150 yuan / megawatt hour), 160,000 yuan of power selling income is additionally generated, thereby verifying the economic benefits generated by the present application in the scenario of insufficient new energy consumption margin.

[0173] Although the embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and alterations can be made thereto without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A data-driven based power market trading strategy optimization method, characterized in that, The method comprises the following steps: S1: constructing an inter-provincial power market transaction simulation environment comprising multiple market members; S2: using the inter-provincial power market transaction simulation environment constructed in S1, calculating a transaction submission amount boundary based on a deep reinforcement learning-based transaction submission auxiliary decision-making model, and updating parameters of the model according to a transaction strategy benefit evaluation result; S3: based on the inter-provincial power market transaction submission amount boundary calculated in S2, constructing a corresponding inter-provincial-intra-provincial two-level market transaction optimization model, and obtaining a final transaction submission amount of a target province in the inter-provincial power market according to an optimization result of the model. The S1 comprises the following steps: S1.1: constructing an approximate transaction submission amount function of a market member other than the target province in the inter-provincial power market; (1) (2) (3) (4) (5) The symbols in all the above formulas are for the provincial power company In the time period , the following are for the provincial power company , which is referred to as the province , so the subscript is omitted and ; and and in formulas (1) and (2) represent the load supply reserve margin and the new energy consumption margin of the province in the time period ; and represent the upper limit and the lower limit of the total power of all power generators except new energy generators; is the upper limit of the total power of new energy generators according to ultra-short-term prediction information; and and in formulas (3) and (4) are the load supply reserve margin and the new energy consumption margin after standardization, and represent the standardization basis values corresponding to the load supply reserve margin and the new energy consumption margin; is a clipping function, which aims to control and within the range of [0, 1]; and in formula (5) is the trading declaration amount of the province in the time period ; if is positive, it represents the amount of electricity purchased by the province , and vice versa; and represent the minimum demand of the province for the power supply reserve margin and the new energy consumption margin; the declared trading amount in formula (5) has a piecewise linear function relationship with , The three segments in the function represent three cases: purchasing electricity when the power supply margin is insufficient, purchasing electricity when the new energy consumption margin is sufficient, and selling electricity when the new energy consumption margin is insufficient; and parameters , and represent the starting values of the three function segments, , and represent the slopes of the three function segments; S1.2: the transaction declaration volume calculated based on formula (5) , and the historical actual declaration volume of the province , the undetermined parameters in formula (3)-(5) are updated based on the gradient descent algorithm;​ S1.3: calculating the transaction submission amount of all market members other than the target province based on the parameters obtained in S1.2 and formula (5); The S2 comprises the following steps: S2.1: initializing the transaction submission auxiliary decision-making model; S2.2: Initialize the transaction declaration auxiliary decision model and the interaction data pool of the inter-provincial electricity market transaction simulation environment , let be the empty set; S2.3: Randomly select a historical arbitrary period and start the calculation; for the period , construct the observation state vector of the transaction declaration decision model ; S2.4: the observation state vector formed in S2.3 The strategy network input to the transaction declaration auxiliary decision model, through which the inter-provincial power market transaction declaration amount boundary corresponding to the future setting period is generated ; S2.5: Deriving interregional power market trade declaration volume bounds from S2.4 As the final trade declaration volume and passed to S1 to form an interregional power market trade simulation environment, perform S1.3 interregional power market simulation clearing to obtain the period of traded power volume and the clearing price ; S2.6: evaluating the benefit of the transaction strategy according to the inter-provincial power market clearing result in S2.5, which is specifically expressed by the following formula: (9) wherein, is the estimated benefit of inter-provincial trading strategy in time period , , , and are defined the same as S1 formula (1)-(5); the first term in the formula represents the cost of buying and selling electricity, the second term represents the penalty for insufficient load supply margin in time period , and the third term represents the penalty for insufficient new energy consumption margin; and represent the weights of the two penalty terms, respectively; S2.7: the observation state vector is input to the historical state memory network, while generating the historical state information encoding , the new energy unit total power prediction value and the system load power prediction value ; S2.8: Forming a data package as a result of S2.3 to S2.7 and adding to the interaction data pool ; data package Specifically containing the following information: (10) wherein, is the observation state vector for the corresponding transaction period is constructed in the same way as ​ S2.9: using the data packets in the interaction data pool to update the strategy network parameters, value network parameters and historical state memory network parameters of the transaction submission decision-making model: (11) (12) (13) wherein , , is a parameter , , is an update step size of , , is a parameter , , is an update target of , , is an update gradient of , , , , respectively as follows:​ (14) (15) (16) in, For parameters Trading strategies before the update during the time period The advantage value, For parameters Transaction declaration volume before and after update The corresponding probability ratio, Algorithm parameters used to control the magnitude of gradient descent; Estimated benefits for all data packets in the interactive data pool The total cumulative discount This refers to the output value of the value network in the transaction reporting auxiliary decision-making model. S2.10: repeating S2.2-2.9 until the parameters of various networks in the transaction submission decision-making model converge to a stable level.

2. The data-driven power market trading strategy optimization method of claim 1, wherein: The S3 comprises the following steps: S3.1: the time period in which the target province currently needs to make inter-provincial power market declaration , construct the grid observation state vector ; input the observation state vector into the trading declaration decision model in S2.10, where the parameters have converged to a stable level, to generate the inter-provincial power market trading declaration amount boundary corresponding to the future set period ; S3.2: constructing an inter-provincial-intra-provincial two-level market transaction optimization mathematical model of the target province as follows: (18) (19) (20) (21) (22) wherein formula (18) is the objective function of the constructed mathematical model, is the objective function of the mathematical model of the inter-provincial power market clearing in the target province, represents the tie-line at time , the inter-provincial power market transaction declaration quantity of the target province, is the transaction price of the target province at time ; formulae (19)-(22) are constraint conditions of the constructed mathematical model; formula (19) is the tie-line power limit constraint after inter-provincial transaction, represents the original tie-line power plan before inter-provincial transaction, represents the tie-line power plan obtained by modification after inter-provincial transaction; formula (20) is the tie-line power plan adjustment limit constraint, is the set of all tie-lines connected with the target province , is the maximum transaction declaration quantity of the target province in the target time period to the target province ; formula (21) is the power grid safety section flow limit constraint, is the power grid node-safety section flow transfer factor matrix corresponding to the time period , is the node injection power of node at time , is the tie-line-node correlation factor matrix, and are the lower and upper limits of the flow transmission of the safety section at time ; formula (22) is the power balance constraint of the target province, wherein each term contains the in-province unit power generation and the tie-line power after inter-provincial power market transaction, is the total power of all power generating units except new energy at time , is the total power of new energy units at time , is the system load power of the target province at time ; S3.3: Solving the two-stage market transaction optimization mathematical model built in S3.2 using a mathematical programming solver to obtain the transaction declaration quantity of each tie-line in the inter-provincial electricity market .

3. A data-driven based power market trading strategy optimization system based on the data-driven based power market trading strategy optimization method of any one of claims 1-2, characterized in that: The system comprises a construction module, an evaluation module and an optimization module; The construction module is configured to construct an inter-provincial power market transaction simulation environment comprising multiple market members, and the construction module is connected to the evaluation module through a network; The evaluation module is configured to calculate a transaction submission amount boundary based on a deep reinforcement learning-based transaction submission auxiliary decision-making model according to the constructed inter-provincial power market transaction simulation environment, and the evaluation module is connected to the optimization module through a network; The optimization module is configured to construct a corresponding inter-provincial-intra-provincial two-level market transaction optimization model based on the inter-provincial power market transaction submission amount boundary, and obtain a final transaction submission amount of a target province in the inter-provincial power market according to an optimization result of the model.

4. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the steps of the data-driven power market transaction strategy optimization method according to any one of claims 1-2.

5. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 4. The computer program is executed by the processor to realize the steps of the data-driven power market transaction strategy optimization method according to any one of claims 1-2.

Images