Control method, system, electronic device and storage medium for wind farm cluster

Abstract

The present application discloses a control method, system, electronic device and storage medium for a wind storage farm cluster, and the technical field to which it belongs is wind power generation technology. The control method for the wind storage farm cluster includes: dividing the wind storage farms of the wind storage farm cluster into one first wind storage farm and multiple second wind storage farms; constructing a strategy game model including the first wind storage farm and the second wind storage farm; determining the current power dispatching demand, and performing iterative game calculations in the strategy game model according to the current power dispatching demand; determining the probability distribution information when the iterative game calculation result reaches a stable state; determining the target probability of the first wind storage farm adopting the zero determinant strategy according to the probability distribution information, and controlling the power of the wind storage farm cluster under constraints. The present application can reasonably dispatch the power of the wind storage farm cluster.

Description

Technical Field

[0001] The present application relates to the field of wind power generation technology, and in particular to a control method, system, electronic equipment and storage medium for a wind farm cluster. Background Art

[0002] With the rapid growth of renewable energy, especially the large-scale deployment of wind and solar power, the power system is undergoing significant changes. However, the intermittent and uncertain nature of renewable energy has brought new challenges to the power system. Wind farms, as a multi-energy hybrid system that integrates wind power, energy storage, and other energy sources, provide an important solution to these challenges. The development of wind power technology faces some common challenges, one of which is the lack of active response to collective decision-making, which makes it difficult to achieve reasonable power dispatch in the actual implementation of the dispatch plan.

[0003] Therefore, how to reasonably dispatch the power of a wind farm cluster is a technical problem that those skilled in the art currently need to solve. Summary of the Invention

[0004] The purpose of this application is to provide a control method for a wind-storage power farm cluster, a control system for a wind-storage power farm cluster, an electronic device and a storage medium, which can reasonably dispatch the power of the wind-storage power farm cluster.

[0005] To solve the above technical problems, the present application provides a control method for a wind farm cluster, comprising:

[0006] The wind storage farms in the wind storage farm cluster are divided into a first wind storage farm and a plurality of second wind storage farms; wherein the first wind storage farm is a wind storage farm directly connected to the dispatching center, and the second wind storage farms are other wind storage farms in the wind storage farm cluster except the first wind storage farm;

[0007] Constructing a strategic game model including the first wind-storage power farm and the second wind-storage power farm; wherein the strategic game model is a game model for describing the strategy adopted by the first wind-storage power farm and the second wind-storage power farm to control the benefits of the wind-storage power farm cluster;

[0008] Determining a current power dispatching demand, and performing an iterative game calculation in the strategy game model according to the current power dispatching demand;

[0009] Determining probability distribution information when the iterative game calculation results reach a stable state; wherein the probability distribution information is used to describe the probability that the first wind-storage power farm and the second wind-storage power farm adopt a target strategy in the current round of iterative game calculation, and the target strategy adopted by the first wind-storage power farm is a zero-determinant strategy;

[0010] The target probability of the first wind-storage power plant adopting the zero-determinant strategy is determined based on the probability distribution information, and the power of the wind-storage power plant cluster is controlled under constraints; wherein the constraint condition is that the conditional probability of the first wind-storage power plant adopting the zero-determinant strategy to control the efficiency of the wind-storage power plant cluster is the target probability.

[0011] Optionally, also include:

[0012] Determining a first allocation coefficient according to the waveform similarity between the equivalent output curve and the actual output curve of each wind-storage power farm; wherein the first allocation coefficient is positively correlated with the waveform similarity;

[0013] Determining a second allocation coefficient according to the excess wind energy utilization of each wind storage farm; wherein the second allocation coefficient is positively correlated with the excess wind energy utilization;

[0014] Determining a third allocation coefficient according to the voltage deviation value of each wind farm; wherein the third allocation coefficient is negatively correlated with the voltage deviation value; the voltage deviation value is determined according to the absolute value of the difference between the grid-connected node voltage and the preset voltage;

[0015] Calculate the Shapley value benefit allocation coefficient of each wind-storage power plant according to the first allocation coefficient, the second allocation coefficient and the third allocation coefficient;

[0016] The benefit distribution result of each wind-storage power plant is calculated according to the Shapley value benefit distribution coefficient.

[0017] Optionally, after calculating the benefit distribution result of each wind-storage power plant according to the Shapley value benefit distribution coefficient, the method further includes:

[0018] The grid-connected electricity price strategy of each wind-storage power farm is calculated according to the benefit allocation result.

[0019] Optionally, calculating the Shapley value benefit allocation coefficient of each wind-storage power plant according to the first allocation coefficient, the second allocation coefficient, and the third allocation coefficient includes:

[0020] Calculate weight coefficients of the first distribution coefficient, the second distribution coefficient, and the third distribution coefficient by an entropy method;

[0021] The first allocation coefficient, the second allocation coefficient, and the third allocation coefficient are weightedly calculated based on the weight coefficient to obtain the Shapley value benefit allocation coefficient of each wind-storage power plant.

[0022] Optionally, also include:

[0023] Determining a power regulation boundary of each of the wind farms;

[0024] Constructing a wind speed model of the wind turbine generator set of the wind storage power farm according to the power regulation boundary;

[0025] Determining an induction coefficient according to the wind speed model; wherein the induction coefficient is used to describe the influence of the wake effect between the wind turbines on the wind speed;

[0026] Determining a wind speed error between an actual wind speed of the wind turbine generator set and a target wind speed;

[0027] Performing empirical mode decomposition on the wind speed error to obtain an induction coefficient error;

[0028] Performing an inverse calculation on the inductance error using a gated cyclic unit to obtain a unit control strategy;

[0029] The rotor speed and pitch angle of the wind turbine generator set are adjusted according to the turbine control strategy.

[0030] Optionally, determining the power regulation boundary of each wind farm includes:

[0031] Determining the upper limit of active power regulation of each wind power storage farm according to the mechanical power limit value of the wind turbine generator set;

[0032] Determining the lower limit of active power regulation of each wind power storage farm according to the current speed, current pitch angle, maximum speed and maximum pitch angle of the wind turbine generator set;

[0033] The upper limit of reactive power regulation and the lower limit of reactive power regulation of each wind power storage farm are determined according to the power factor of the wind turbine generator set.

[0034] Optionally, before performing inversion calculation on the inductance error using a gated recurrent unit, the method further includes:

[0035] Modal components of wind speed data are extracted, and the gated recurrent unit is trained using the modal components so that the gated recurrent unit learns the corresponding relationship between control parameters and induction coefficients; wherein the control parameters include rotor speed and pitch angle.

[0036] The present application also provides a control system for a wind farm cluster, the system comprising:

[0037] A role division module is used to divide the wind storage farms in the wind storage farm cluster into one first wind storage farm and multiple second wind storage farms; wherein the first wind storage farm is a wind storage farm directly connected to the dispatching center, and the second wind storage farm is other wind storage farms in the wind storage farm cluster except the first wind storage farm;

[0038] A model building module, configured to build a strategic game model including the first wind-storage power farm and the second wind-storage power farm; wherein the strategic game model is a game model for describing the strategy used by the first wind-storage power farm and the second wind-storage power farm to control the benefits of the wind-storage power farm cluster;

[0039] A game calculation module, configured to determine a current power dispatching demand and perform iterative game calculations in the strategy game model according to the current power dispatching demand;

[0040] a probability calculation module, configured to determine probability distribution information when the iterative game calculation results reach a stable state; wherein the probability distribution information is used to describe the probability that the first wind-storage farm and the second wind-storage farm adopt a target strategy in the current round of iterative game calculation, and the target strategy adopted by the first wind-storage farm is a zero determinant strategy;

[0041] A power control module is used to determine the target probability of the first wind-storage power farm adopting a zero-determinant strategy based on the probability distribution information, and to control the power of the wind-storage power farm cluster under constraints; wherein the constraint condition is that the conditional probability of the first wind-storage power farm adopting a zero-determinant strategy to control the benefits of the wind-storage power farm cluster is the target probability.

[0042] Furthermore, it also includes:

[0043] A benefit allocation module is used to determine a first allocation coefficient based on the waveform similarity between the equivalent output curve and the actual output curve of each wind storage farm; wherein the first allocation coefficient is positively correlated with the waveform similarity; and is also used to determine a second allocation coefficient based on the excess wind energy utilization of each wind storage farm; wherein the second allocation coefficient is positively correlated with the excess wind energy utilization; and is also used to determine a third allocation coefficient based on the voltage deviation value of each wind storage farm; wherein the third allocation coefficient is negatively correlated with the voltage deviation value; the voltage deviation value is determined based on the absolute value of the difference between the grid-connected node voltage and the preset voltage; and is also used to calculate the Shapley value benefit allocation coefficient of each wind storage farm based on the first allocation coefficient, the second allocation coefficient and the third allocation coefficient; and is also used to calculate the benefit allocation result of each wind storage farm based on the Shapley value benefit allocation coefficient.

[0044] Furthermore, it also includes:

[0045] The electricity price strategy calculation module is used to calculate the grid-connected electricity price strategy of each wind-storage power farm according to the benefit allocation result calculated for each wind-storage power farm according to the benefit allocation coefficient.

[0046] Furthermore, the process of the benefit distribution module calculating the Shapley value benefit distribution coefficient of each wind-storage power farm based on the first distribution coefficient, the second distribution coefficient and the third distribution coefficient includes: calculating the weight coefficient of the first distribution coefficient, the second distribution coefficient and the third distribution coefficient by the entropy method; performing weighted calculation on the first distribution coefficient, the second distribution coefficient and the third distribution coefficient based on the weight coefficient to obtain the Shapley value benefit distribution coefficient of each wind-storage power farm.

[0047] Furthermore, it also includes:

[0048] A unit regulation module is used to determine the power regulation boundary of each wind storage farm; is also used to construct a wind speed model of the wind turbines of the wind storage farm based on the power regulation boundary; is also used to determine the induction coefficient based on the wind speed model; wherein the induction coefficient is used to describe the influence of the wake effect between the wind turbines on the wind speed; is also used to determine the wind speed error between the actual wind speed of the wind turbine and the target wind speed; is also used to perform empirical mode decomposition on the wind speed error to obtain the induction coefficient error; is also used to perform inverse calculation on the induction coefficient error using a gated loop unit to obtain a unit control strategy; is also used to adjust the rotor speed and pitch angle of the wind turbine according to the unit control strategy.

[0049] Furthermore, the process of the unit regulation module determining the power regulation boundary of each wind storage farm includes: determining the upper boundary of active power regulation of each wind storage farm based on the mechanical power limit value of the wind turbine; determining the lower boundary of active power regulation of each wind storage farm based on the current speed, current pitch angle, maximum speed and maximum pitch angle of the wind turbine; and determining the upper boundary and lower boundary of reactive power regulation of each wind storage farm based on the power factor of the wind turbine.

[0050] Furthermore, it also includes:

[0051] A training module is used to extract modal components of wind speed data before using a gated cyclic unit to perform inversion calculation on the inductance coefficient error, and use the modal components to train the gated cyclic unit so that the gated cyclic unit learns the corresponding relationship between control parameters and inductance coefficient; wherein the control parameters include rotor speed and pitch angle.

[0052] The present application also provides a storage medium storing a computer program, which implements the steps of the control method for the wind farm cluster when the computer program is executed.

[0053] The present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the control method for the wind farm cluster when calling the computer program in the memory.

[0054] This application provides a control method for a wind farm cluster. This method divides the wind farm cluster into a first wind farm and multiple second wind farms, and constructs a strategy game model encompassing the first and second wind farms. This strategy game model is used to describe the strategic interactions between the first and second wind farms in controlling the efficiency of the wind farm cluster. Based on current power dispatch requirements, this application performs iterative game calculations within the strategy game model to simulate various strategies that the first and second wind farms might adopt in real-world situations. When the iterative game calculation results reach a stable state, the probability distribution information at that point is determined, and the target probability for the first wind farm to adopt a zero-determinant strategy is determined. The power of the wind farm cluster is then controlled based on this target probability. Through this approach, regardless of the strategy adopted by the second wind farm, the first wind farm, which adopts a zero-determinant strategy, can maintain the maximum efficiency of the wind farm cluster while meeting power dispatch requirements. Therefore, this application enables rational power dispatch of the wind farm cluster, reducing interference from uncertainties and variability in the external environment. The present application also provides a control system for a wind farm cluster, a storage medium, and an electronic device, which have the above-mentioned beneficial effects and will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0056] Figure 1 A flow chart of a control method for a wind-storage power plant cluster provided in an embodiment of the present application;

[0057] Figure 2 A schematic diagram of a rapid capacity reduction for active power regulation provided in an embodiment of the present application;

[0058] Figure 3 A technical roadmap for optimizing the active regulation capability of a wind farm based on multi-agent game-playing provided in an embodiment of the present application;

[0059] Figure 4 A receiving schematic diagram of an IEEE-69 node system provided in an embodiment of the present application;

[0060] Figure 5 A schematic diagram of a grid-connected electricity price provided in an embodiment of the present application;

[0061] Figure 6 A schematic diagram of a wind power output coefficient provided in an embodiment of the present application;

[0062] Figure 7 A schematic diagram of an energy storage output coefficient provided in an embodiment of the present application;

[0063] Figure 8 A schematic diagram of an energy storage charge state provided in an embodiment of the present application;

[0064] Figure 9 A schematic diagram of a wind power cluster wind curtailment rate provided in an embodiment of the present application;

[0065] Figure 10 A schematic diagram of a node system voltage amplitude provided in an embodiment of the present application;

[0066] Figure 11 A schematic diagram of the change in average inductance provided in an embodiment of the present application. DETAILED DESCRIPTION

[0067] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0068] See below Figure 1 , Figure 1 This is a flow chart of a control method for a wind farm cluster provided in an embodiment of the present application.

[0069] Specific steps may include:

[0070] S101: Dividing the wind storage power farms of the wind storage power farm cluster into one first wind storage power farm and multiple second wind storage power farms;

[0071] This embodiment can be applied to control equipment in a dispatch center of a wind-storage farm cluster, where the wind-storage farm cluster includes multiple wind-storage farms, each of which includes a wind farm and an energy storage power station, and each wind farm includes multiple wind turbines. This embodiment can determine the wind-storage farms included in the wind-storage farm cluster and classify the wind-storage farms to obtain a first wind-storage farm and multiple second wind-storage farms.

[0072] As a feasible implementation, this embodiment can designate one wind farm directly connected to the dispatch center as the first wind farm, and the remaining wind farms as the second wind farm. That is, the first wind farm is the one directly connected to the dispatch center, and the second wind farms are the remaining wind farms in the wind farm cluster, excluding the first. If a wind farm cluster contains multiple wind farms directly connected to the dispatch center, this embodiment can designate the one directly connected to the dispatch center as the first wind farm.

[0073] S102: Constructing a strategy game model including the first wind-storage power farm and the second wind-storage power farm;

[0074] The strategic game model describes how the first and second wind-storage farms employ strategies to control the efficiency of a wind-storage farm cluster. In this model, the first wind-storage farm is the leader, and the second wind-storage farm is the follower. The leader sets benchmarks or rules, and its strategy influences the efficiency of the entire cluster. Followers can adjust their own strategies based on the leader's strategy and the behavior of other followers.

[0075] S103: Determine a current power dispatching demand, and perform an iterative game calculation in the strategy game model according to the current power dispatching demand;

[0076] Among them, this embodiment can determine the current power dispatch demand based on the real-time load of the power system, the supply and demand situation of the power market, and the dispatch of the power grid operator. Before performing the iterative game calculation, the strategy game model can be initialized. In the iterative game calculation, the first wind storage power farm and the second wind storage power farm can update their own strategies based on the current power dispatch demand and the other party's strategy. Specifically, the first wind storage power farm first selects the optimal strategy based on the power dispatch demand and its own benefit function, and notifies the second wind storage power farm; after observing the strategy of the first wind storage power farm, the second wind storage power farm adjusts its own strategy based on its own benefit function and constraints. The above process is repeated until a stable state is reached.

[0077] S104: Determine the probability distribution information when the iterative game calculation result reaches a stable state;

[0078] The probability distribution information is used to describe the probability of the first wind-storage farm and the second wind-storage farm adopting the target strategy in the iterative game calculation of the current round. The target strategy adopted by the first wind-storage farm is the zero-determinant strategy (ZD strategy), and the target strategy adopted by the second wind-storage farm can be any strategy.

[0079] In order to determine whether the iterative game calculation results have reached a stable state, this embodiment can set corresponding judgment criteria, such as the frequency change of strategy selection in 5 consecutive rounds of iteration is less than a preset value, or the change rate of total benefit in 5 consecutive rounds is less than a preset value.

[0080] In a stable state, this embodiment can count the number of times the first wind-storage farm and the second wind-storage farm adopt their respective target strategies, and calculate the proportion of these times to the total number of iterations, thereby obtaining probability distribution information.

[0081] S105: Determine the target probability of the first wind-storage farm adopting the zero determinant strategy according to the probability distribution information, and control the power of the wind-storage farm cluster under constraints;

[0082] After obtaining the probability distribution information, the target probability of the first wind-storage farm adopting the zero-determinant strategy (i.e., the probability of the first wind-storage farm adopting the zero-determinant strategy in the iterative game) can be queried from the determined probability distribution information. In this embodiment, a constraint condition can be set based on the probability distribution information. The constraint condition specifies that the conditional probability of the first wind-storage farm adopting the zero-determinant strategy to control the wind-storage farm cluster efficiency is the target probability. This constraint condition ensures that the first wind-storage farm achieves the target probability in actual iterations.

[0083] This embodiment can control the power of each wind storage farm in the wind storage farm cluster, and in the control process ensure that the conditional probability of the first wind storage farm using the zero determinant strategy to control the efficiency of the wind storage farm cluster is the target probability.

[0084] This embodiment divides a wind farm cluster into a first wind farm and multiple second wind farms. A strategic game model encompassing the first and second wind farms is constructed to describe the strategic interactions between the first and second wind farms in controlling the efficiency of the wind farm cluster. Based on current power dispatch requirements, this embodiment performs iterative game calculations within the strategic game model to simulate various strategies that the first and second wind farms might adopt in real-world situations. When the iterative game calculation results reach a stable state, the probability distribution at that point is determined, along with the target probability for the first wind farm to adopt the zero-determinant strategy. The power of the wind farm cluster is then controlled based on this target probability. This approach allows the first wind farm, which employs the zero-determinant strategy, to maintain the maximum efficiency of the wind farm cluster while meeting power dispatch requirements, regardless of the strategy adopted by the second wind farm. Therefore, this embodiment enables rational power dispatch of the wind farm cluster, mitigating interference from uncertainties and variability in the external environment.

[0085] The above solution is described below through an embodiment of practical application. In this embodiment, the leader is the first wind-storage farm in the above embodiment, and the follower is the second wind-storage farm in the above embodiment. The specific solution is as follows:

[0086] Assuming that all wind farms in a wind farm cluster are equipped with energy storage power stations that meet the construction requirements, this is a wind-storage farm cluster. The power control of this wind-storage farm cluster is modeled as an iterative game with N players. Each wind-storage farm can determine its own continuous output power in the game. Assuming that the wind-storage farm directly connected to the dispatch center is the controlling player in the game, that is, the leader The other N-1 wind farms are followers. In order to achieve the best utilization of the output power of the wind farm, each wind farm is equipped with energy storage equipment to support or store wind energy when the output of the wind farm is insufficient or excessive. The strategy is expressed as ,in 、 express Similarly, for the other N-1 followers The strategy is expressed as ,in 、 express Active output and power factor angle. When a wind farm cluster jointly completes a power dispatch task, the cluster leader first receives power instructions from the dispatch center. Assuming sufficient wind speed conditions and excessive output power from each wind farm, wind curtailment will occur. Excess wind energy needs to be stored or utilized. To avoid unit tripping and possible frequent starts and stops within the wind farm, energy conversion losses during charging and discharging of energy storage stations should be considered, and dispatch strategies should be rationally optimized to address wind curtailment and achieve greater public benefits for the wind farm cluster.

[0087] This embodiment provides a wind farm upper-level strategy game model, which is described in detail as follows:

[0088] Benefits of wind turbines in wind farms and energy storage benefits Define them separately:

[0089] (1);

[0090] (2);

[0091] Where, is the grid-connected electricity price; and are the active dispatch response coefficients of the leader and other followers, respectively; The power demand of the dispatching center (i.e., power dispatching demand); and are the maximum active power output capabilities of the leader and follower wind farms respectively; It is the sum of the maximum active power output of the wind power cluster.

[0092] (3);

[0093] (4);

[0094] (5);

[0095] Where, and are the energy storage benefits of the leader and follower, respectively; Charging efficiency for energy storage; and is the utilization rate of excess wind energy in the leader and follower wind farms. The adjustment of this utilization rate is the core of the strategy of this layer model. The game strategy can be further expressed as:

[0096] (6);

[0097] Therefore, the economic benefits of wind farm clusters Can be defined as:

[0098] (7);

[0099] For the zero-rank strategy, no matter what game strategy the followers adopt, the leader who adopts the zero-rank strategy can unilaterally maintain the total economic benefits of the wind power cluster at the maximum value while meeting the system scheduling requirements. Therefore, it can be assumed that the followers all adopt the average strategy, that is, the probability of choosing any action in each round is the same. and followers' benefits They can be defined as:

[0100] (8);

[0101] Where, and Represent the auxiliary costs of the leader and follower respectively, which can be expressed as follows:

[0102] (9);

[0103] In the above formula, is the unit cost of energy storage auxiliary services.

[0104] In each round of iterative calculation, the expected benefit of the leader in the game model is and the expected benefits to followers It can be expressed as:

[0105] (10);

[0106] Where, represents the probability distribution of the leader and follower adopting the corresponding strategy in round h. According to the properties of Markov stationary distribution, this distribution is consistent with the probability distribution of the corresponding strategy in round h-1. There is the following relationship matrix M:

[0107] (11);

[0108] (12);

[0109] (13);

[0110] Where, and are the conditional probabilities of the leader and follower in round h-1, respectively. Indicates that the cluster strategy is When the leader selects The conditional probability of Indicates that the cluster strategy is When the follower chooses When the probability distributions of the hth round and the h-1th round are equal, the iterative process reaches a stable point, that is, the steady-state distribution is obtained. , according to the definition of zero determinant strategy, the expected benefit of the leader at this time is and the expected benefits to followers The following linear relationship can be maintained:

[0111] (14);

[0112] Where v1, v2 and v3 are constant parameters.

[0113] At this point, for a multi-agent multi-strategy model with a leader and multiple followers, the goal is to find the maximum economic benefit of the wind farm cluster. The scheduling policy problem can be transformed into the following maximization of the total expected benefit The optimization problem is:

[0114] (15);

[0115] (16);

[0116] The conditional probability when the leader adopts the zero-rank strategy to control the total expected benefit of the wind farm is Sometimes, there are , the maximum benefit of the cluster can converge to , then the optimization objective can be simplified to formula (16). This is the target probability mentioned above. is the utilization rate of excess wind energy in the leader’s wind farm in round h, is the utilization rate of excess wind energy of the leader’s wind farm in round h-1, is the benefit of the N-1th follower.

[0117] In this optimization process, when the leader adopts the zero-rank strategy, the follower's strategy will not change. The maximum expected benefit of the wind power cluster under certain conditions is .

[0118] As a feasible implementation method, the benefits can be distributed among wind farms in the following ways:

[0119] Step A1: determining a first allocation coefficient according to the waveform similarity between the equivalent output curve and the actual output curve of each wind power storage farm;

[0120] Wherein, the first allocation coefficient is positively correlated with the waveform similarity;

[0121] Step A2: determining a second allocation coefficient according to the excess wind energy utilization of each wind storage farm;

[0122] Wherein, the second allocation coefficient is positively correlated with the excess wind energy utilization;

[0123] Step A3: determining a third allocation coefficient according to the voltage deviation value of each wind power storage farm;

[0124] The third allocation coefficient is negatively correlated with the voltage deviation value; the voltage deviation value is determined according to the absolute value of the difference between the grid-connected node voltage and the preset voltage;

[0125] Step A4: calculating the Shapley value benefit allocation coefficient of each wind-storage power plant according to the first allocation coefficient, the second allocation coefficient, and the third allocation coefficient;

[0126] Specifically, this embodiment can calculate the weight coefficients of the first distribution coefficient, the second distribution coefficient and the third distribution coefficient by the entropy method; based on the weight coefficients, the first distribution coefficient, the second distribution coefficient and the third distribution coefficient are weightedly calculated to obtain the Shapley value benefit distribution coefficient of each wind storage power farm.

[0127] Step A5: Calculating the benefit distribution result of each wind-storage power plant according to the Shapley value benefit distribution coefficient.

[0128] Furthermore, after calculating the benefit allocation result of each wind-storage power plant according to the Shapley value benefit allocation coefficient, the grid-connected electricity price strategy of each wind-storage power plant can also be calculated according to the benefit allocation result.

[0129] The following illustrates the above solution through the actual application of the wind farm cluster lower-level revenue game model:

[0130] In the upper-level strategy game model of wind-storage power plants, the output strategy that maximizes the expected benefits while fully utilizing excess wind energy is calculated by determining the leaders and followers, but ignores the contribution of the wind farm to the stability of the node system and to reducing wind curtailment.

[0131] This embodiment can use the cooperative game model to allocate the marginal contributions of N participants, using the improved Shaply value method. For a wind farm cluster, the participants of the cooperative game model are , the strategy set of the wind farm is the grid-connected electricity price strategy , The grid-connected electricity price for the leader, For the grid-connected electricity price of followers, due to the particularity of the zero-sum determinant strategy, the lower-level cooperative game model will not change the maximum economic benefit, but the distribution of Shaply values ​​can be improved in combination with the contribution of wind farms, so as to re-formulate the grid-connected electricity price strategy. , and then fed back to the upper-level strategy model, so that while ensuring maximum benefits, the strategy results have the characteristics of considering the stability of the node system and reducing the contribution of wind curtailment.

[0132] The Shapley value of participant i is defined as:

[0133] (17);

[0134] In the above formula, S represents the alliance composed of all participants, and v (S) represents the total benefits of all participants in the alliance.

[0135] (18);

[0136] In the Shaply method calculation formula, member i has (|S|-1)! possible rankings when participating in alliance S, where |S| represents the number of members in alliance S, and there are (n-|S|)! possible rankings for the remaining (n-|S|) members. The different ranking combinations of all members participating in the calculation divided by the random ranking combinations of n members is the weight of the benefits that member i should share with the entire alliance, which is recorded as [(|S|-1)!(n-|S|)!] / (n!).

[0137] is the Shapley value of participant i; is the total benefit of the alliance formed except for participant i; is the weighting factor; is the total number of possible permutations of the game participants. Therefore, the Shapley value considers the average marginal contribution under all possible sequences, ensuring fair distribution of benefits. Considering the contribution of the wind farm to the stability of the node system, the Shapley allocation result is reconstructed mainly from the following aspects:

[0138] Aspect 1: Similarity between the total load waveforms of the wind power cluster and the alliance.

[0139] For the node system where the wind power cluster is located, when the wind power output change trend is similar to the load waveform, that is, the less the exchange with regional thermal power units and energy storage energy is, the lower the impact on the stability of the power system itself and the cost is correspondingly reduced. The equivalent output of each node wind farm is calculated using the equal power-following load method. , calculate the waveform similarity of equivalent output and actual output curves using the cosine similarity algorithm , the greater the waveform similarity, the more the wind farm meets the load demand and the greater its contribution to reducing electricity costs. The corresponding allocation coefficient The bigger. This is the first distribution coefficient mentioned above.

[0140] (19);

[0141] (20);

[0142] In the above formula, represents the actual output of the wind farm node, T represents the total time interval, and t is the time point corresponding to the current moment.

[0143] Aspect 2: The level of excess wind energy utilization in wind power clusters.

[0144] In the case of excess wind power, considering the minimum output requirement of thermal power plants, in order to reduce wind curtailment, if the excess wind energy can be connected to the grid or charged to energy storage during the entire calculation period, that is, the excess wind energy is fully utilized through reasonable scheduling, the excess wind energy utilization can be calculated. For the redistribution of alliance benefits, the more the excess wind energy is fully utilized, the more benefits can be distributed. The corresponding distribution coefficient is The bigger. This is the second distribution coefficient mentioned above.

[0145] (twenty one);

[0146] (twenty two);

[0147] Where, is the unit coefficient corresponding to the time scale; and They are excess wind energy utilization coefficient, maximum output, dispatch response coefficient and dispatch instruction respectively.

[0148] Aspect 3: Steady-state level of node system.

[0149] In the node system, the power flow process will cause losses in the transmission line, thereby reducing the voltage level at the remote node. The wind storage system can be used as a reactive power supplement device to support the node voltage, reduce the voltage support from the regional thermal power balance node, and reduce the inductive voltage drop in the transmission line. Therefore, by calculating the error between the actual voltage of the wind farm grid-connected node and 1p.u. For the redistribution of alliance benefits, the smaller the error, the more effective the use of excess wind energy to support local node voltage. The corresponding allocation coefficient is The smaller. is the voltage deviation value mentioned above, is the third distribution coefficient mentioned above.

[0150] (twenty three);

[0151] (twenty four);

[0152] In the above formula, It represents the actual voltage of the wind farm grid-connected node i at time t, i represents the serial number of the wind farm, t represents the voltage at that moment; T represents the calculation cycle time.

[0153] In summary, the improved Shapley value benefit allocation coefficient , improved allocation results And the corresponding grid-connected electricity price strategy for the previous period of optimization allocation for:

[0154] (25);

[0155] (26);

[0156] (27);

[0157] Where, represents the Shapley value before improvement.

[0158] 、 、 is the weight coefficient, which is determined by the entropy method. The greater the information entropy of the factor, the greater its weight; For the The economic benefits of a wind farm.

[0159] As a feasible implementation method, the wind turbine generator system can be adjusted in the following ways:

[0160] Step B1: determining the power regulation boundary of each wind power storage farm;

[0161] Among them, the power regulation boundary includes the active power regulation boundary and the reactive power regulation boundary, which can be determined specifically in the following ways: determining the upper active power regulation boundary of each wind storage farm (also known as the active power regulation upper boundary) according to the mechanical power limit value of the wind turbine; determining the lower active power regulation boundary of each wind storage farm (also known as the active power regulation lower boundary) according to the current speed, current pitch angle, maximum speed and maximum pitch angle of the wind turbine; determining the upper reactive power regulation boundary and the lower reactive power regulation boundary (also known as the reactive power regulation boundary) of each wind storage farm according to the power factor of the wind turbine.

[0162] Step B2: constructing a wind speed model of the wind turbine generator set of the wind storage farm according to the power regulation boundary;

[0163] Step B3: determining an induction coefficient according to the wind speed model; wherein the induction coefficient is used to describe the influence of the wake effect between the wind turbines on the wind speed;

[0164] Step B4: determining a wind speed error between an actual wind speed of the wind turbine generator set and a target wind speed;

[0165] Step B5: performing empirical mode decomposition on the wind speed error to obtain an induction coefficient error;

[0166] Step B6: using a gated cyclic unit to perform an inverse calculation on the inductance error to obtain a unit control strategy;

[0167] Step B7: adjusting the rotor speed and pitch angle of the wind turbine generator set according to the turbine generator set control strategy.

[0168] As a feasible implementation method, before using the gated cyclic unit to perform inversion calculation on the induction coefficient error, the modal components of the wind speed data can also be extracted, and the gated cyclic unit can be trained using the modal components so that the gated cyclic unit learns the correspondence between control parameters and induction coefficient; wherein the control parameters include rotor speed and pitch angle.

[0169] The above implementation is described below using a wind turbine active regulation model in actual application.

[0170] The upper limit of active power regulation is calculated as follows:

[0171] Affected by factors such as weather, the output of wind turbines is random and fluctuating, resulting in the frequency regulation capability of wind farms achieved through load reduction operation being time-varying. This will bring about a series of problems. Therefore, it is particularly important to accurately evaluate the frequency regulation capability of wind farms.

[0172] When the wind farm receives the dispatch instruction from the regional power grid dispatch center When the power of each unit in the wind farm needs to be increased, the active power regulation capacity of the wind farm can be obtained. , its specific calculation formula can be written as:

[0173] (28);

[0174] Where, is the theoretical maximum output of the i-th wind turbine, is the current actual output of the i-th unit, and It can be directly obtained through the data acquisition and monitoring system of the wind farm, where n is the number of wind turbines.

[0175] The upper limit of wind farm regulation is By mechanical power limit Determine as follows:

[0176] (29);

[0177] Where, is the air density of the wind farm; is the impeller radius of the i-th unit; is the wind speed of the i-th unit; 0° pitch angle , optimal tip speed ratio The corresponding wind energy utilization efficiency is shown below.

[0178] The lower limit of active power regulation is calculated as follows:

[0179] When the system frequency is higher than the rated frequency, the wind turbines need to reduce their own output to participate in frequency regulation to reduce the system frequency, that is, the dispatching instructions received by the wind farm The wind farm active power regulation capacity that can be continuously reduced by wind turbines during frequency regulation is changed to the following formula: ,in is the lower bound of the active regulation capacity of the i-th wind turbine. It can be seen that the key to the evaluation and calculation of the active regulation capacity of the wind farm is of confirmation.

[0180] (30);

[0181] There are two ways for wind turbines to reduce their own output:

[0182] (1) Through rotor overspeed control, the rotor speed is increased, causing the output power of the wind turbine to deviate from its maximum power tracking point , converting part of the electromagnetic energy into rotor kinetic energy, thereby reducing the active output of the unit.

[0183] (2) Increase the pitch angle by controlling the pitch angle To reduce the wind energy utilization coefficient , thereby reducing the wind energy captured by the unit and, in turn, reducing active power output. Pitch angle control can achieve load reduction control at all wind speeds, but due to the mechanical movement, the response speed is slow, and frequent pitch angle adjustments can increase wear on the wind turbine, shortening its lifespan and increasing operating costs.

[0184] Comparing the two methods, the wind energy lost by pitch angle control is directly discarded, reducing wind energy utilization. In contrast, rotor overspeed control, achieved through PWM control, has a faster response and can store some of the wind energy discarded by load shedding as rotational kinetic energy. However, overspeed control becomes ineffective when the rotor speed reaches 1.2 pu, making it unsuitable for control at all wind speeds. Each control method has its own advantages and disadvantages. To fully exploit the wind turbine's active power reduction capability, it is necessary to combine the two control methods to leverage their respective strengths: overspeed control in the maximum power tracking region and pitch angle control in the constant speed and constant power regions.

[0185] In the maximum power tracking area, the optimal rotor speed The maximum speed has not been reached , it is possible to increase the speed to reduce the active power output. Figure 2 , Figure 2 This is a schematic diagram of a fast capacity reduction of active power regulation provided in an embodiment of the present application. The intersection points in the figure include point A, point B and point C. The horizontal axis WS is the generator speed, the vertical coordinate P m is the active power of the unit, β is the pitch angle, β max is the maximum pitch angle, point A is the maximum power tracking point, point B is the maximum speed that can be achieved by overspeed control, and the difference in captured mechanical power between A and B is the rapid reduction capacity of the unit.

[0186] Taking into account the actual situation, the current unit operation point may not be the maximum power tracking point, and the current unit pitch angle may not be 0, but as long as the unit speed does not reach , then the unit has a certain rapid reduction capacity. Assuming the current speed of the unit is , the pitch angle is , radius is R, wind speed is v w , the current tip speed ratio is , the maximum tip speed ratio is , C p To improve the efficiency of wind energy utilization, the unit's capacity should be quickly reduced. The calculation formula is:

[0187] (31);

[0188] It can be seen that in the maximum power tracking area, the rapid reduction of the wind turbine capacity is only related to the current operating point of the wind turbine, specifically, it is only related to the speed, pitch angle and current wind speed. In this process, only the rotor speed of the wind turbine is changed, and its pitch angle is always remains unchanged, at this time the regulation capacity of the entire wind farm It can be expressed as:

[0189] (32);

[0190] Where, is the rapid reduction capacity of unit i. Due to the influence of the wake effect, the reduction capacity of different units is also different under the same operating conditions. The rapid reduction capacity of the unit originally running in the maximum power tracking mode at different wind speeds can be calculated. When the wind speed gradually increases to the rated wind speed, the corresponding tip speed ratio The closer to the optimal tip speed ratio , so the difference in wind energy utilization coefficient Getting smaller and smaller, It decreases gradually, and when the wind speed reaches the rated wind speed, its rapid downward adjustment capacity is 0.

[0191] At this time, in order to further reduce the output power of the wind turbine, the pitch angle control can be used to increase the pitch angle of the turbine. With pitch angle The relationship between The bigger, The smaller the pitch angle, the smaller the mechanical power captured by the unit. However, in order to avoid the wind turbine being disconnected from the grid due to too low mechanical power captured, the pitch angle There is a maximum value ,once , which may cause the unit to be disconnected from the grid or the load to be too high. The speed and pitch angle of point C reach the maximum value, which corresponds to the minimum output point of the unit. The power difference between B and C is the slow down capacity of the unit. , and its specific calculation formula is:

[0192] (33);

[0193] Slow capacity reduction is achieved by changing only the pitch angle, while the rotor speed remains constant. Less than the maximum pitch angle When the unit has slow capacity reduction. When considering the influence of two factors: one is the influence of mechanical load, and the other is the influence of power limitation when the unit is connected to the grid. The specific expression is:

[0194] (34);

[0195] Where, It is the maximum pitch angle of the group when the mechanical load is taken into account. is the maximum pitch angle of the group when active power limitation is considered, and the pitch angle Take the smaller of the two. The calculation formula is:

[0196] (35);

[0197] Indicates the slow capacity reduction of the i-th wind turbine.

[0198] Combined with the above analysis, we can know that the lower limit of the actual active power regulation capacity of wind turbines is for:

[0199] (36);

[0200] According to the relevant regulations on the allowable active power output limit of wind turbines in the case of grid connection, this article stipulates Not less than 15% of rated active output. , current pitch angle , maximum allowable speed and the maximum allowable pitch angle , then formula (30) can be written as:

[0201] (37);

[0202] At this time, the lower limit of the active regulation capacity of the entire wind farm is for

[0203] (38);

[0204] Considering that the wind farm needs to carry out maintenance, the wind farm units are not all running at all times, but some are running and some are shut down. When a unit is shut down, it will have no effect on the wind speed of the downstream units, and the wake of the entire wind farm will change. In order to deal with this situation, we define is the output coefficient of unit i after being affected by the wake; is the switching function of unit i: when unit i is running, ; When unit i stops operating, .

[0205] In summary, the active power regulation capacity of wind farms It can be described by the following formula:

[0206] (39);

[0207] The reactive power regulation boundary is calculated as follows:

[0208] The stator of the doubly fed asynchronous wind turbine is directly connected to the grid, and the power is transmitted to the grid through the stator and rotor ends. According to relevant regulations, the power factor of the wind turbine is The requirement of leading 0.95 to lagging 0.95 must be met, so the reactive power limit of the wind farm can be calculated according to the following formula:

[0209] (40);

[0210] Where: It is the upper limit of reactive power output of wind farm; It is the lower limit of reactive power output of wind farm; is the power factor angle of the i-th unit.

[0211] Relevant regulations require wind farms to operate under a restricted control strategy. However, in practice, factors such as inaccurate wake models, uncertainty about real-time wind speed and direction, and inaccurate interaction information can lead to errors in the wind farm's regulation capabilities.

[0212] The wind speed model of wind turbines is introduced below:

[0213] A wind farm typically includes multiple wind turbines that interact with each other. Different wind direction conditions and relative installation positions result in different output power. During wind farm operation, downstream wind turbines may be located in the wake of one or more upstream wind turbines, resulting in cumulative wake losses. Therefore, when calculating the power generated by a wind turbine, it is necessary to account for the fact that the actual wind speed of each wind turbine is insufficient due to the wake effect.

[0214] Based on the Park wake model, wind turbines Located in another wind turbine In the wake zone, Wind speed for:

[0215] (41);

[0216] Where, represents the air roughness coefficient (i.e., the slope of the wake expanding away from the wind turbine), represents the distance between wind turbines i and j in the wind direction, for wind speed; for Wake pair The resulting inductance; for Wind wheel diameter; For downstream Wind rotor swept area; for and Overlapping area of ​​wake region. Wind farm with 10 turbines For equivalent wind farm wind speed It can be expressed as:

[0217] (42);

[0218] Where v0 represents the wind speed of the first wind turbine facing the wind in the current wind speed direction; Therefore, when the wind turbine adjustment capability deviates from the expected value, the induction coefficient between the turbines can be adjusted by calculation. The control strategy controls the output capacity of the wind farm. At this time, according to the Bernoulli equation, under the wake effect, the wind farm Power generated As shown in the following formula:

[0219] (43);

[0220] The following introduces the induction coefficient inversion model based on EMD-GRU.

[0221] Inductance is a set of intermediate control variables that control the rotor speed and pitch angle Later, due to uncertainties such as inaccurate wake model, real-time wind speed and direction, the wind speed of other wind turbines is affected, namely:

[0222] (44);

[0223] Where, The corresponding function is formula (41); is the intermediate control variable between fans i and j and The mapping model between them.

[0224] represents the generator speed ω of wind turbine j j and the generator speed ω of wind turbine i i A sequential set of represents the pitch angle β of wind turbine j j Pitch angle β with i i A sequential set of , .

[0225] By controlling the intermediate variables to timely adjust the error with the actual output, the actual control strategy is deduced through the inversion model and fed back to the unit control system, so that the unit's control strategy is transformed from passive control to active participation in regulating the output capacity of the wind farm.

[0226] The Gated Recurrent Unit (GRU) neural network is a simplified and improved version of the LSTM network. Its main difference from the LSTM is that a single gating unit simultaneously controls the forgetting factor and the decision to update the state unit. It also solves the vanishing gradient problem that occurs during traditional RNN training using the BPTT algorithm. It better captures features in time series, enables modeling of long-term time series, and is more adaptable to complex and fluctuating data such as wind speed. Its mathematical model is shown below:

[0227] (45);

[0228] (46);

[0229] (47);

[0230] (48);

[0231] Where, 、 They are reset gate and update gate respectively; is the Logistic function; 、 They are the GRU output at the current moment and the output at the previous moment respectively; is a candidate status; Input for the current moment; They are the weight matrices for resetting the gate, updating the gate, and obtaining the candidate states.

[0232] In summary, the modal components of wind speed data are extracted by EMD (Empirical Mode Decomposition), and the GRU (Gated Recurrent Unit) neural network model is used to train the The mapping relationship Secondly, the optimization search algorithm combined with the two-dimensional Halton sequence is used to invert The active control strategy of the unit (i.e. the unit control strategy mentioned above); then through the existing The function characterizes the wind speed under each modal component, and finally superimposes the results to output the wind speed, thereby proactively adjusting the output in a timely manner to better fit the wind farm power curve of the scheduling requirements.

[0233] See Figure 3 , Figure 3 This is a technical roadmap for optimizing the active regulation capability of wind farms based on a multi-agent game, as provided in this application. The figure shows wind farms 1 through N, each containing energy storage batteries and connected to grid connection points 1 through n. This solution primarily consists of three parts: the first is the upper-level strategy game model for the wind farm, the second is the lower-level revenue game model for the wind farm cluster, and the third is the wind turbine active regulation model.

[0234] The implementation process of the wind turbine active regulation model includes: calculating the wind farm regulation boundary, wind farm scheduling strategy instructions, wind farm actual response output, calculating the strategy response error, so that the strategy response error can be input into the inversion model; in the inversion model, calculating the wind speed error corresponding to each unit, and inputting the unit wind speed error into the EMD calculation function , the corresponding induction coefficient error is calculated by GRU inversion to obtain the corresponding unit control strategy change.

[0235] The implementation process of the upper-level strategy game model of wind-storage power plants is as follows: determine the leader and follower, calculate the excess power of each wind-storage power plant, calculate the wind turbine benefit, energy storage benefit and cluster benefit, calculate the leader and follower benefits, calculate the corresponding relationship matrix, and calculate the strategy distribution V corresponding to rounds h-1 and h. h-1 and V h; Determine whether the probability distribution of rounds h-1 and h is equal. If not, adjust the strategy. If so, obtain the steady-state distribution v s ; The leader strategy is a zero-rank strategy, so that there is an unchanging maximum benefit - (v3) max The real-time strategy is transmitted to the numerical dispatch center, and the benefit strategy can also be transmitted to the wind power cluster's underlying revenue game model. The wind power cluster's underlying revenue game model can upload grid-connected electricity price adjustment information to update node grid-connected electricity prices.

[0236] The implementation process of the lower-level revenue game model of the wind-storage power plant is as follows: determine that the leader and followers in the upper-level model are all cooperative game participants, set the leader to adopt a zero-rank strategy, set the followers to adopt a full-cooperative strategy, and calculate the Shaply value y i ; Calculate the distribution coefficient α corresponding to the similarity between the cluster and the total load waveform w , calculate the allocation coefficient α corresponding to the cluster excess wind energy utilization level e , calculate the distribution coefficient α corresponding to the steady-state level of the node system u ; Calculate the improved Shaply value , adjust the grid-connected electricity price strategy c i g .

[0237] As can be seen, this embodiment provides a wind farm cluster control strategy and a two-tier model for secondary revenue distribution. This embodiment also provides a definition of wind turbine regulation capability and the associated induction coefficient inversion model. This embodiment also provides an overall regulation process for implementing wind farm control strategies into wind turbines.

[0238] Wind farms with storage have become a key approach to addressing renewable energy volatility and power system scheduling issues. To fully tap their potential, various entities involved in wind farms are dispatched and coordinated through a collaborative and competitive approach to achieve frequency stability, balance supply and demand, and improve energy efficiency.

[0239] Currently, various methods have been proposed for optimizing wind farm power. For example, some studies have utilized situational awareness technology and optimized instructions to improve resilience to environmental interference. Other studies have used principal component analysis and clustering techniques to provide robust scheduling solutions for the uncertainty of wind power output. Other studies have focused on day-ahead power scheduling of wind farm clusters, using specific techniques to reduce wind curtailment by considering power forecast deviations and uncertainty in regulation capacity. Other studies have focused on emergency control strategies for wind farms, such as emergency active power control, to reduce the number and power of units removed. Furthermore, research involves adjusting the operating strategies of doubly-fed wind turbines, optimizing the frequency regulation participation of energy storage-based wind farms, and analyzing the capacity of wind, solar, and storage power generation systems from the perspective of dispatching load demand to evaluate their power output.

[0240] While existing research covers a wide range of topics, most focuses on power optimization within wind farms, with less attention paid to coordination between wind turbines and wind farms. To address this gap, some studies have employed multi-agent consensus and game theory methods to propose strategies for controlling wind turbine and wind farm power. Furthermore, some studies, based on cooperative and non-cooperative game theory, have explored collaborative operation strategies between microgrids and the power grid to improve the absorption rate of photovoltaic power generation. Market participation strategies for energy storage power plants have also been investigated, aiming to explore optimal operation and bidding methods. Furthermore, optimizing the potential role of energy storage systems in grid frequency regulation has become a focus, with cooperative game models being constructed to reduce costs and improve efficiency. Furthermore, to increase the absorption rate of distributed generation and rationally distribute energy-sharing benefits, some studies have introduced optimization strategies based on modified Shapley values ​​for further research.

[0241] In the field of wind power generation, dealing with uncertainty and variability is one of the main challenges facing power dispatch. Currently, some engineering applications attempt to use control technology to address these issues, but this is usually accompanied by high costs. With the development of wind farm technology, each wind farm is more focused on pursuing its own interests. Therefore, it is increasingly important to introduce game theory into the power dispatch problem of wind farm clusters. This field focuses on the strategy of each wind farm, aiming to maximize its own benefits. Game theory, as one of the effective methods to resolve contradictions in group decision-making, provides a useful approach to solving this problem. In large-scale wind farm projects, economic evaluation has become particularly important to enhance the position of wind power generation in market competition, which is crucial for the sustainable development of wind farm clusters.

[0242] One of the key issues in wind power generation is coping with its uncertainty and variability, especially the challenges in power scheduling. Currently, some engineering applications attempt to use control technology to solve these problems, but this is usually accompanied by high costs. Considering the intelligence of wind farms, each wind farm is more inclined to pursue its own interests. Therefore, it becomes very valuable to introduce game theory into the power scheduling problem of wind farm clusters. In this context, focusing on the strategies adopted by each wind farm to maximize its own benefits has become the core of the research. Game theory is considered to be one of the effective methods to resolve contradictions in group decision-making in order to improve the competitiveness of wind power generation in the increasingly mature electricity market competition. Solving this problem is crucial to the sustainable development of wind farm clusters.

[0243] To implement wind farm dispatch instructions, wind turbines control speed and pitch angle to achieve a wind farm output strategy that meets dispatch requirements. However, in practice, factors such as inaccurate wake models, uncertainty about real-time wind speed and direction, and inaccurate interaction information can lead to errors in the wind farm's regulation capabilities. Consequently, the power generation boundaries of wind farm energy management systems often differ from actual capabilities. To mitigate this discrepancy, the actual output curve reported for dispatch is conservative. Therefore, leveraging the regulation potential of wind storage farms plays a crucial role in improving the economic efficiency of wind farms.

[0244] This example constructs a model to assess the active regulation capability of wind-storage farms. This model clarifies the regulation capability boundaries of each wind-storage farm from a cluster perspective. Using an induction coefficient inversion model based on EMD-GRU, it enables each unit within the wind farm to proactively regulate in response to the farm's regulation needs. This model helps fully utilize the regulation potential of wind-storage farms and improves the efficiency and accuracy of dispatch responses.

[0245] This embodiment also proposes a two-level optimization model based on multi-agent game theory. The upper-level model utilizes zero-rank game theory to develop a wind farm cluster output plan that maximizes economic benefits and reduces wind energy waste. The lower-level model considers the impact of strategies on system stability and reducing wind curtailment, innovatively applying an improved Shapley value method. Through a dynamic feedback mechanism, the economic benefit allocation results and adjusted grid-connected electricity prices of the lower-level model are fed back to the upper-level model, iteratively optimizing the dispatch strategy of the wind farm cluster and rationally allocating economic benefits between wind and storage power plants, thereby improving the fairness and incentives of benefit distribution.

[0246] The following example uses the improved IEEE-69 node system to verify the results. Figure 4 , Figure 4 This is a schematic diagram of an IEEE-69 node system provided in an embodiment of the present application. The figure shows a thermal power plant. The numbers 1 to 69 in the figure represent the connection points or branch points of the power network. WS1 to WS10 represent wind power plants. The capacity of the thermal power units is 500MVA. Node 1 is the thermal power balance node in the system. The wind power plant access location is as follows: Figure 4 As shown in the figure, different wind farms are equipped with 4-6MW wind turbines. According to current requirements, this article assumes that the minimum output of the thermal power unit is 30% of the rated capacity, that is, the minimum technical output reaches 150MVA; the power capacity of the energy storage configured in the wind storage farm is configured as 15% of the wind power capacity, and the energy capacity is configured as a 2h power supply capacity.

[0247] Assuming that WS1 is directly connected to the dispatch center, it is determined as the leader of the wind power cluster, and the other wind farms are followers. The improved convex inner approximation method is used to calculate in the improved IEEE69 node system. After iterative calculation, the optimal balance is obtained as follows: Figure 5The system electricity price level is shown. Figure 5 This is a schematic diagram of a grid-connected strategy electricity price provided in an embodiment of the present application, where the horizontal axis is time T, the vertical axis is the wind farm number, and the vertical axis is the grid-connected electricity price.

[0248] The dispatching output of wind power storage farm is as follows: Figure 6 and Figure 7 As shown, Figure 6 A schematic diagram of wind power output coefficient provided in an embodiment of the present application is shown in FIG. Figure 7 A schematic diagram of energy storage output coefficient provided in an embodiment of the present application is shown in FIG. Figure 6 and Figure 7 The horizontal axis is time T, the vertical axis is the output coefficient, and WS1~WS10 represent each wind storage power farm.

[0249] It can be seen that in the interval between the 45th and 62nd time points, the proposed strategy considers reducing the shutdown risk of thermal power units. When the wind turbine output level cannot meet the maximum demand load, the energy storage maintains a certain output level, but no wind power energy is stored in the energy storage power station. Figure 8 As shown, Figure 8 A schematic diagram of energy storage charge state provided in an embodiment of the present application is shown. Figure 8 The horizontal axis is the wind farm number, the vertical axis is time T, and the vertical axis is state of charge SOC / %.

[0250] The charge state of the energy storage power station dropped rapidly between the 45th and 53rd time points and maintained a low charge level for a period of time. The energy storage's charging and discharging strategy actively responded to the load and the overall output strategy of the wind turbine, leaving capacity space for absorbing excess wind energy in subsequent wind energy surplus situations.

[0251] Figure 9 A schematic diagram of the wind power cluster wind curtailment rate provided in the embodiment of the present application, with the horizontal axis being time T and the vertical axis being the total wind curtailment rate (%). The figure shows the difference between the traditional control strategy and the control strategy of this paper. Figure 9 Using data from a wind farm with storage, the proposed active regulation strategy for wind farms successfully reduced wind curtailment by a maximum of 8% and an average of 3%, significantly improving the passive curtailment experienced under traditional strategies. This demonstrates that active regulation strategies have achieved significant results in improving wind power resource utilization.

[0252] Figure 10 A schematic diagram of a node system voltage amplitude provided in an embodiment of the present application is shown. Figure 10 The horizontal axis is time T, the vertical axis is the node number, and the vertical axis is the voltage amplitude U / pu. Figure 11 This is a schematic diagram of the change of average inductance provided in an embodiment of the present application. Figure 11The horizontal axis is time T, the vertical axis is the wind farm number, and the vertical axis is the voltage amplitude U / pu.

[0253] according to Figure 10 and Figure 11 Data from the wind farm show that the wind farm adjusts its output level by actively adjusting the inductance. Compared with the traditional strategy, the active adjustment strategy increased the maximum voltage amplitude at some nodes by 2.09% relative to the original maximum amplitude. However, the average voltage deviation of the system decreased by 6.17%. Therefore, the active adjustment strategy proposed in this paper not only does not negatively impact the voltage level of the node system, but also has a certain positive gain in actual conditions.

[0254] See Table 1 for the improved Shaply value calculation results. According to the data in Table 1, after implementing the active regulation strategy, the average profit of each wind-storage farm within the node increased by 9.06%. In particular, wind-storage farms WS7, WS9, and WS10 saw their profits increase by over 11% due to their high contributions to waveform similarity and wind energy utilization. Although WS1 failed to fully utilize wind energy due to coordinated control, reducing its additional profit, it still achieved a 3.84% profit increase due to its contribution to waveform similarity.

[0255] Table 1 Improved Shaply value calculation results

[0256]

[0257] This shows that the active regulation strategy not only helps improve the average revenue of each wind-storage farm, but also achieves significant economic benefits in different scenarios. This further verifies the effectiveness of the active regulation strategy and demonstrates its potential for broad application in optimizing wind-storage farm operations.

[0258] The present invention provides a control system for a wind farm cluster, including:

[0259] A role division module is used to divide the wind storage farms in the wind storage farm cluster into one first wind storage farm and multiple second wind storage farms; wherein the first wind storage farm is a wind storage farm directly connected to the dispatching center, and the second wind storage farm is other wind storage farms in the wind storage farm cluster except the first wind storage farm;

[0260] A model building module, configured to build a strategic game model including the first wind-storage power farm and the second wind-storage power farm; wherein the strategic game model is a game model for describing the strategy used by the first wind-storage power farm and the second wind-storage power farm to control the benefits of the wind-storage power farm cluster;

[0261] A game calculation module, configured to determine a current power dispatching demand and perform iterative game calculations in the strategy game model according to the current power dispatching demand;

[0262] a probability calculation module, configured to determine probability distribution information when the iterative game calculation results reach a stable state; wherein the probability distribution information is used to describe the probability that the first wind-storage farm and the second wind-storage farm adopt a target strategy in the current round of iterative game calculation, and the target strategy adopted by the first wind-storage farm is a zero determinant strategy;

[0263] A power control module is used to determine the target probability of the first wind-storage power farm adopting a zero-determinant strategy based on the probability distribution information, and to control the power of the wind-storage power farm cluster under constraints; wherein the constraint condition is that the conditional probability of the first wind-storage power farm adopting a zero-determinant strategy to control the benefits of the wind-storage power farm cluster is the target probability.

[0264] This embodiment divides a wind farm cluster into a first wind farm and multiple second wind farms. A strategic game model encompassing the first and second wind farms is constructed to describe the strategic interactions between the first and second wind farms in controlling the efficiency of the wind farm cluster. Based on current power dispatch requirements, this embodiment performs iterative game calculations within the strategic game model to simulate various strategies that the first and second wind farms might adopt in real-world situations. When the iterative game calculation results reach a stable state, the probability distribution at that point is determined, along with the target probability for the first wind farm to adopt the zero-determinant strategy. The power of the wind farm cluster is then controlled based on this target probability. This approach allows the first wind farm, which employs the zero-determinant strategy, to maintain the maximum efficiency of the wind farm cluster while meeting power dispatch requirements, regardless of the strategy adopted by the second wind farm. Therefore, this embodiment enables rational power dispatch of the wind farm cluster, mitigating interference from uncertainties and variability in the external environment.

[0265] Furthermore, it also includes:

[0266] A benefit allocation module is used to determine a first allocation coefficient based on the waveform similarity between the equivalent output curve and the actual output curve of each wind storage farm; wherein the first allocation coefficient is positively correlated with the waveform similarity; and is also used to determine a second allocation coefficient based on the excess wind energy utilization of each wind storage farm; wherein the second allocation coefficient is positively correlated with the excess wind energy utilization; and is also used to determine a third allocation coefficient based on the voltage deviation value of each wind storage farm; wherein the third allocation coefficient is negatively correlated with the voltage deviation value; the voltage deviation value is determined based on the absolute value of the difference between the grid-connected node voltage and the preset voltage; and is also used to calculate the Shapley value benefit allocation coefficient of each wind storage farm based on the first allocation coefficient, the second allocation coefficient and the third allocation coefficient; and is also used to calculate the benefit allocation result of each wind storage farm based on the Shapley value benefit allocation coefficient.

[0267] Furthermore, it also includes:

[0268] The electricity price strategy calculation module is used to calculate the grid-connected electricity price strategy of each wind-storage power farm according to the benefit allocation result calculated for each wind-storage power farm according to the benefit allocation coefficient.

[0269] Furthermore, the process of the benefit distribution module calculating the Shapley value benefit distribution coefficient of each wind-storage power farm based on the first distribution coefficient, the second distribution coefficient and the third distribution coefficient includes: calculating the weight coefficient of the first distribution coefficient, the second distribution coefficient and the third distribution coefficient by the entropy method; performing weighted calculation on the first distribution coefficient, the second distribution coefficient and the third distribution coefficient based on the weight coefficient to obtain the Shapley value benefit distribution coefficient of each wind-storage power farm.

[0270] Furthermore, it also includes:

[0271] A unit regulation module is used to determine the power regulation boundary of each wind storage farm; is also used to construct a wind speed model of the wind turbines of the wind storage farm based on the power regulation boundary; is also used to determine the induction coefficient based on the wind speed model; wherein the induction coefficient is used to describe the influence of the wake effect between the wind turbines on the wind speed; is also used to determine the wind speed error between the actual wind speed of the wind turbine and the target wind speed; is also used to perform empirical mode decomposition on the wind speed error to obtain the induction coefficient error; is also used to perform inverse calculation on the induction coefficient error using a gated loop unit to obtain a unit control strategy; is also used to adjust the rotor speed and pitch angle of the wind turbine according to the unit control strategy.

[0272] Furthermore, the process of the unit regulation module determining the power regulation boundary of each wind storage farm includes: determining the upper boundary of the active power regulation of each wind storage farm based on the mechanical power limit value of the wind turbine; determining the lower boundary of the active power regulation of each wind storage farm based on the current speed, current pitch angle, maximum speed and maximum pitch angle of the wind turbine; and determining the upper boundary and lower boundary of the reactive power regulation of each wind storage farm based on the power factor of the wind turbine.

[0273] Furthermore, it also includes:

[0274] A training module is used to extract modal components of wind speed data before using a gated cyclic unit to perform inversion calculation on the inductance coefficient error, and use the modal components to train the gated cyclic unit so that the gated cyclic unit learns the corresponding relationship between control parameters and inductance coefficient; wherein the control parameters include rotor speed and pitch angle.

[0275] Since the embodiments of the system part correspond to the embodiments of the method part, please refer to the description of the embodiments of the method part for the embodiments of the system part, and will not be repeated here.

[0276] This application also provides a storage medium having a computer program stored thereon. When executed, the computer program can implement the steps provided in the above embodiments. The storage medium may include: a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, an optical disk, or other medium capable of storing program code.

[0277] The present application also provides an electronic device that may include a memory and a processor, wherein the memory stores a computer program, and when the processor calls the computer program in the memory, the steps provided in the above embodiment can be implemented. Of course, the electronic device may also include various network interfaces, a power supply, and other components.

[0278] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the various embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of this application, several improvements and modifications can be made to this application, and these improvements and modifications also fall within the scope of protection of this application.

[0279] It should also be noted that, in this specification, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

Claims

1. A control method for a wind farm cluster, characterized in that: include: The wind storage farms in the wind storage farm cluster are divided into a first wind storage farm and a plurality of second wind storage farms; wherein the first wind storage farm is a wind storage farm directly connected to the dispatching center, and the second wind storage farms are other wind storage farms in the wind storage farm cluster except the first wind storage farm; Constructing a strategic game model including the first wind-storage power farm and the second wind-storage power farm; wherein the strategic game model is a game model for describing the strategy adopted by the first wind-storage power farm and the second wind-storage power farm to control the benefits of the wind-storage power farm cluster; Determining a current power dispatching demand, and performing an iterative game calculation in the strategy game model according to the current power dispatching demand; Determining probability distribution information when the iterative game calculation results reach a stable state; wherein the probability distribution information is used to describe the probability that the first wind-storage power farm and the second wind-storage power farm adopt a target strategy in the current round of iterative game calculation, and the target strategy adopted by the first wind-storage power farm is a zero-determinant strategy; The target probability of the first wind-storage power plant adopting the zero-determinant strategy is determined based on the probability distribution information, and the power of the wind-storage power plant cluster is controlled under constraints; wherein the constraint condition is that the conditional probability of the first wind-storage power plant adopting the zero-determinant strategy to control the efficiency of the wind-storage power plant cluster is the target probability.

2. The control method of the wind farm cluster according to claim 1, characterized in that: Also includes: Determining a first allocation coefficient according to the waveform similarity between the equivalent output curve and the actual output curve of each wind-storage power farm; wherein the first allocation coefficient is positively correlated with the waveform similarity; Determining a second allocation coefficient according to the excess wind energy utilization of each wind storage farm; wherein the second allocation coefficient is positively correlated with the excess wind energy utilization; Determining a third allocation coefficient according to the voltage deviation value of each wind farm; wherein the third allocation coefficient is negatively correlated with the voltage deviation value; the voltage deviation value is determined according to the absolute value of the difference between the grid-connected node voltage and the preset voltage; Calculate the Shapley value benefit allocation coefficient of each wind-storage power plant according to the first allocation coefficient, the second allocation coefficient and the third allocation coefficient; The benefit distribution result of each wind-storage power plant is calculated according to the Shapley value benefit distribution coefficient.

3. The control method of the wind farm cluster according to claim 2, characterized in that: After calculating the benefit distribution result of each wind-storage power plant according to the Shapley value benefit distribution coefficient, the method further includes: The grid-connected electricity price strategy of each wind-storage power farm is calculated according to the benefit allocation result.

4. The control method of a wind farm cluster according to claim 2, characterized in that: Calculating the Shapley value benefit allocation coefficient of each wind-storage power plant according to the first allocation coefficient, the second allocation coefficient, and the third allocation coefficient includes: Calculate weight coefficients of the first distribution coefficient, the second distribution coefficient, and the third distribution coefficient by an entropy method; The first allocation coefficient, the second allocation coefficient, and the third allocation coefficient are weightedly calculated based on the weight coefficient to obtain the Shapley value benefit allocation coefficient of each wind-storage power plant.

5. The control method of a wind farm cluster according to claim 1, characterized in that: Also includes: Determining a power regulation boundary of each of the wind farms; Constructing a wind speed model of the wind turbine generator set of the wind storage power farm according to the power regulation boundary; Determining an induction coefficient according to the wind speed model; wherein the induction coefficient is used to describe the influence of the wake effect between the wind turbines on the wind speed; Determining a wind speed error between an actual wind speed of the wind turbine generator set and a target wind speed; Performing empirical mode decomposition on the wind speed error to obtain an induction coefficient error; Performing an inverse calculation on the inductance error using a gated cyclic unit to obtain a unit control strategy; The rotor speed and pitch angle of the wind turbine generator set are adjusted according to the turbine control strategy.

6. The control method of a wind farm cluster according to claim 5, characterized in that: Determining the power regulation boundary of each wind farm includes: Determining the upper limit of active power regulation of each wind power storage farm according to the mechanical power limit value of the wind turbine generator set; Determining the lower limit of active power regulation of each wind power storage farm according to the current speed, current pitch angle, maximum speed and maximum pitch angle of the wind turbine generator set; The upper limit of reactive power regulation and the lower limit of reactive power regulation of each wind power storage farm are determined according to the power factor of the wind turbine generator set.

7. The control method of a wind farm cluster according to claim 5, characterized in that: Before performing inversion calculation on the inductance error by using a gated recurrent unit, the method further includes: Modal components of wind speed data are extracted, and the gated recurrent unit is trained using the modal components so that the gated recurrent unit learns the corresponding relationship between control parameters and induction coefficients; wherein the control parameters include rotor speed and pitch angle.

8. A control system for a wind farm cluster, characterized in that: include: A role division module is used to divide the wind storage farms in the wind storage farm cluster into one first wind storage farm and multiple second wind storage farms; wherein the first wind storage farm is a wind storage farm directly connected to the dispatching center, and the second wind storage farm is other wind storage farms in the wind storage farm cluster except the first wind storage farm; A model building module, configured to build a strategic game model including the first wind-storage power farm and the second wind-storage power farm; wherein the strategic game model is a game model for describing the strategy used by the first wind-storage power farm and the second wind-storage power farm to control the benefits of the wind-storage power farm cluster; A game calculation module, configured to determine a current power dispatching demand and perform iterative game calculations in the strategy game model according to the current power dispatching demand; a probability calculation module, configured to determine probability distribution information when the iterative game calculation results reach a stable state; wherein the probability distribution information is used to describe the probability that the first wind-storage farm and the second wind-storage farm adopt a target strategy in the current round of iterative game calculation, and the target strategy adopted by the first wind-storage farm is a zero determinant strategy; A power control module is used to determine the target probability of the first wind-storage power farm adopting a zero-determinant strategy based on the probability distribution information, and to control the power of the wind-storage power farm cluster under constraints; wherein the constraint condition is that the conditional probability of the first wind-storage power farm adopting a zero-determinant strategy to control the benefits of the wind-storage power farm cluster is the target probability.

9. The control system of the wind farm cluster according to claim 8, characterized in that: Also includes: A benefit allocation module is used to determine a first allocation coefficient based on the waveform similarity between the equivalent output curve and the actual output curve of each wind storage farm; wherein the first allocation coefficient is positively correlated with the waveform similarity; and is also used to determine a second allocation coefficient based on the excess wind energy utilization of each wind storage farm; wherein the second allocation coefficient is positively correlated with the excess wind energy utilization; and is also used to determine a third allocation coefficient based on the voltage deviation value of each wind storage farm; wherein the third allocation coefficient is negatively correlated with the voltage deviation value; the voltage deviation value is determined based on the absolute value of the difference between the grid-connected node voltage and the preset voltage; and is also used to calculate the Shapley value benefit allocation coefficient of each wind storage farm based on the first allocation coefficient, the second allocation coefficient and the third allocation coefficient; and is also used to calculate the benefit allocation result of each wind storage farm based on the Shapley value benefit allocation coefficient.

10. The control system of the wind farm cluster according to claim 9, characterized in that: Also includes: The electricity price strategy calculation module is used to calculate the grid-connected electricity price strategy of each wind-storage power farm according to the benefit allocation result calculated for each wind-storage power farm according to the benefit allocation coefficient.

11. The control system of the wind farm cluster according to claim 9, characterized in that: The process of the benefit distribution module calculating the Shapley value benefit distribution coefficient of each wind-storage power farm based on the first distribution coefficient, the second distribution coefficient and the third distribution coefficient includes: calculating the weight coefficient of the first distribution coefficient, the second distribution coefficient and the third distribution coefficient by the entropy method; performing weighted calculation on the first distribution coefficient, the second distribution coefficient and the third distribution coefficient based on the weight coefficient to obtain the Shapley value benefit distribution coefficient of each wind-storage power farm.

12. The control system of the wind farm cluster according to claim 8, characterized in that: Also includes: A unit regulation module is used to determine the power regulation boundary of each wind storage farm; is also used to construct a wind speed model of the wind turbines of the wind storage farm based on the power regulation boundary; is also used to determine the induction coefficient based on the wind speed model; wherein the induction coefficient is used to describe the influence of the wake effect between the wind turbines on the wind speed; is also used to determine the wind speed error between the actual wind speed of the wind turbine and the target wind speed; is also used to perform empirical mode decomposition on the wind speed error to obtain the induction coefficient error; is also used to perform inverse calculation on the induction coefficient error using a gated loop unit to obtain a unit control strategy; is also used to adjust the rotor speed and pitch angle of the wind turbine according to the unit control strategy.

13. The control system of the wind farm cluster according to claim 12, characterized in that: The process of the unit regulation module determining the power regulation boundary of each wind storage farm includes: determining the upper boundary of the active power regulation of each wind storage farm based on the mechanical power limit value of the wind turbine; determining the lower boundary of the active power regulation of each wind storage farm based on the current speed, current pitch angle, maximum speed and maximum pitch angle of the wind turbine; and determining the upper boundary and lower boundary of the reactive power regulation of each wind storage farm based on the power factor of the wind turbine.

14. The control system of the wind farm cluster according to claim 12, characterized in that: Also includes: A training module is used to extract modal components of wind speed data before using a gated cyclic unit to perform inversion calculation on the inductance coefficient error, and use the modal components to train the gated cyclic unit so that the gated cyclic unit learns the corresponding relationship between control parameters and inductance coefficient; wherein the control parameters include rotor speed and pitch angle.

15. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the processor calls the computer program in the memory, the steps of the control method of the wind-storage power plant cluster according to any one of claims 1 to 7 are implemented.

16. A storage medium, characterized in that The storage medium stores computer-executable instructions, which, when loaded and executed by a processor, implement the steps of the method for controlling a wind-storage power plant cluster according to any one of claims 1 to 7.

Images