Offshore wind power collection-transmission integrated optimization method for energy storage

Abstract

The invention discloses an offshore wind power collection-transmission integrated optimization method for energy storage, and relates to the technical field of wind power plant optimization, comprising the following steps: selecting the capacity of an energy storage unit according to annual wind resource data of a wind power plant, and estimating the operation cost of energy storage; establishing a wind power plant model based on a power transmission problem of energy storage, and expressing the wind power plant model in a cost function form; an ADMM algorithm and a solution algorithm based on reinforcement learning are adopted to solve the cost function, and an optimized energy storage scheme is determined; and performing cost comparison with a scheme without adopting the energy storage unit to obtain a comparison result, and finally obtaining an optimization scheme result. According to the offshore wind power collection-power transmission integrated optimization method for energy storage, the power transmission efficiency can be improved, and the cost related to a power transmission cable can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of wind farm optimization, and in particular to an offshore wind power collection and transmission integrated optimization method for energy storage. Background Art

[0002] As an important component of renewable energy, wind farms play an increasingly critical role in the transformation of energy structure. Traditional wind farms usually use the method of arranging wind turbine cables and directly connecting to the grid point or connecting to the grid through an offshore booster station. Such a design requires that the transmission cables must be able to carry the rated capacity of the wind farm to ensure full load operation when the wind is strong. However, the randomness and uncertainty of wind resources means that wind farms are not always able to generate electricity that matches the rated capacity, resulting in the transmission cables not always being able to operate at full load. In the case of long-distance offshore power transmission, this design is not only inefficient, but also leads to huge transmission costs. Summary of the invention

[0003] The present invention solves the problem that the transmission cables used in traditional wind farms cannot always operate at full load, and proposes an offshore wind power collection-transmission integrated optimization method for energy storage, which can not only improve the transmission efficiency, but also help reduce the costs associated with transmission cables.

[0004] In order to achieve the above-mentioned object, the present invention adopts the following technical solution: an offshore wind power collection and transmission integrated optimization method for energy storage, comprising the following steps: S1, based on the annual wind resource data of the wind farm, select the capacity of the energy storage unit and estimate the operating cost of the energy storage; S2, establish a wind farm model based on the transmission problem of energy storage and express it in the form of a cost function; S3, using the ADMM algorithm and the reinforcement learning-based solution algorithm to solve the cost function and determine the optimal energy storage solution; S4, compare the cost with the solution without the energy storage unit to obtain the comparison result, and finally obtain the optimization solution result.

[0005] The model construction of this technical solution is based on the overall optimization of the wind farm. This global consideration can ensure that when pursuing the optimization of a certain aspect, it does not sacrifice the interests of other aspects, thereby achieving an overall reduction in the operating costs of the wind farm. According to the mathematical characteristics of the model, an advanced optimization algorithm is adopted, that is, a solution strategy based on ADMM (alternating direction multiplier method) and reinforcement algorithm, which efficiently handles non-convex and non-continuous problems in the model and ensures the accuracy and reliability of the solution process.

[0006] The present invention is further configured as follows: the wind farm model includes a first model and a second model, the first model includes the transmission cost between the energy storage location and the grid connection point location; the second model is equivalent to the cost function of the cable layout optimization problem.

[0007] In this technical solution, the first model is composed of the transmission cost between the energy storage location and the grid connection point, which can be regarded as a connected convex function; the second model mainly involves the optimization of the cable layout of the wind farm, which can be regarded as a non-convex and non-continuous function.

[0008] The present invention is further configured as follows: the objective function of the wind farm model is the sum of the transmission cost between the energy storage location and the grid connection point location, the cost function of the cable layout optimization problem, and the operating cost of the energy storage.

[0009] In this technical solution, the optimization of wind farm cable layout is a path combination optimization problem. The shortest collection path to the energy storage will change when the energy storage location is different.

[0010] The present invention is further configured as follows: Step S3 comprises: S31, convert the cost function into an augmented Lagrangian equivalent form, and introduce the ADMM algorithm to iteratively update the variables x and z and the Lagrangian multiplier λ; S32, using the gradient method to solve the minimization problem of the continuous variable x; S33, uses reinforcement learning algorithm to solve the cable optimization problem; S34, multiple iterations are performed to finally determine the optimal energy storage location.

[0011] In this technical solution, in order to enable the problem to be processed by the ADMM algorithm, an auxiliary variable z is introduced, and the objective function and constraints of the problem are converted and transformed into an augmented Lagrangian equivalent form; then the gradient method and reinforcement learning algorithm are used to solve the problem respectively. After multiple iterations, the optimal energy storage position is finally determined, and the optimal connection method is selected for all wind turbines, so that the power transmission efficiency of the system is maximized and the cable loss is minimized.

[0012] The present invention is further configured as follows: Step S4 comprises: S41, comparing the economic costs of the optimized energy storage solution and the solution without the energy storage unit, and determining to adopt the optimized energy storage solution after the transmission cable exceeds a kilometers, where a is greater than 0; S42, determining the final optimization solution result, which includes the optimized energy storage solution adopted when the transmission cable exceeds a kilometers, the location of the energy storage unit construction, and the wind farm cable layout result.

[0013] In this technical solution, the optimization solution result is finally determined by comparing the energy storage solutions.

[0014] The present invention is further configured as follows: Step S33 includes: S331, define the state, action, and reward function of reinforcement learning; initialize the reinforcement learning model; S332, action selection and state transition; S333, calculating the fitness value of each state; S334, updating strategies using reinforcement learning methods; S335, iterative update, output the optimal cable connection solution.

[0015] In this technical solution, the above-mentioned algorithm is specifically used to solve the cable optimization problem.

[0016] The present invention is further configured as follows: the state includes wind turbine coordinates, energy storage position and current cable connection scheme; the action is to adjust the cable connection structure, and the reward function measures the contribution of each action to the overall layout optimization.

[0017] In this technical solution, the state, action and reward function of reinforcement learning are first defined, then the state space and action space are initialized, the initial cable connection scheme is randomly generated, and the strategy is initialized.

[0018] The present invention is further configured as follows: Step S332 includes: Based on the current status, the strategy selects an action, including the connected wind turbines and the selected cable type; then the status of the wind turbines in the wind farm is updated.

[0019] The present invention is further configured as follows: Step S333 includes: S3331, generating a corresponding cable connection structure according to the wind turbine and energy storage position in the current state; S3332, using the connection scheme generated by the cable, calculate the target value in this state according to the reward function; S3333, the target value is used as the fitness value of this state, which is equivalent to the reward value.

[0020] In this technical solution, the cable connection structure is first generated, and then the target value in this state is calculated.

[0021] The present invention is further configured as follows: Step S335 includes: S3351, updating the local optimal value and the global optimal value according to the current fitness value; S3352, by repeating steps S332-S334, the strategy is continuously adjusted until the fitness value reaches a maximum or a preset iteration stop condition is met; S3353, after multiple iterations, outputs the optimal cable connection solution.

[0022] In this technical solution, after iteration until the corresponding conditions are met, the optimal solution is output.

[0023] The present invention can bring the following beneficial effects: 1. The model construction of the present invention is based on the overall optimization of the wind farm, taking into account the cost of power collection, energy storage and transmission to form a global optimization model; this global consideration can ensure that when pursuing the optimization of a certain aspect, the interests of other aspects are not sacrificed, thereby achieving an overall reduction in the operating cost of the wind farm; 2. Aiming at the mathematical characteristics of the model, the present invention adopts an advanced optimization algorithm - a solution strategy based on ADMM and reinforcement algorithm; by combining these two algorithms, the present invention can efficiently handle non-convex and non-continuous problems in the model, ensuring the accuracy and reliability of the solution process; 3. The present invention can not only determine the optimal location of the energy storage unit through the optimization algorithm, but also optimize the connection mode of the internal cables of the wind farm; the reasonable layout of the energy storage unit is crucial to improving the energy utilization efficiency of the wind farm, and the optimized connection of the internal cables can reduce the loss in the energy transmission process, further improving the economy and reliability of the system; 4. The present invention can also provide decision-making guidance on transmission distance and energy storage construction; through efficient calculation, the present invention can determine the most suitable transmission distance for building energy storage units, which has important guiding significance for the design and planning of wind farms; in summary, the present invention demonstrates its unique advantages in reducing wind farm operating costs, improving energy utilization efficiency and optimizing decision support through the construction of a global optimization model, the design of an efficient optimization algorithm and decision-making guidance on transmission distance and energy storage construction; these advantages can not only improve the economy and reliability of wind farms, but also promote the wind power industry to develop in a more efficient, economical and environmentally friendly direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a flow chart of an offshore wind power collection and transmission integrated optimization method for energy storage in the present application.

[0025] Figure 2 It is a schematic diagram of an embodiment of an offshore wind power collection and transmission integrated optimization method for energy storage in the present application. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] With the transformation of the global energy structure and the rapid development of renewable energy, offshore wind farms, as an important part of clean energy, are increasingly developing in scale and offshore. In the process of exploring deep-sea wind resources, my country faces a series of technical challenges and economic considerations. The production capacity of offshore wind farms is significantly affected by weather conditions, and wind power generation has obvious intermittent and random uncertainties. These characteristics make it difficult to achieve efficient use of the transmission cables connecting wind farms to the power grid.

[0028] In existing wind farm designs, the selection of transmission cables is usually based on the rated capacity of the wind farm. However, due to the intermittent and uncertain nature of wind power generation, transmission cables often cannot continue to work at rated power, resulting in waste of resources and increased costs. For deep-sea wind farms, the cost of long-distance transmission is particularly significant.

[0029] In this case, the design and optimization of energy storage units become particularly important. Energy storage units can not only reduce transmission costs, but also reduce the use of collection cables by directly connecting to the wind turbines inside the wind farm, thereby further reducing the collection costs. Therefore, designing an optimization algorithm to determine the optimal location of the energy storage unit is of great significance for reducing the collection and transmission costs of wind farms. In order to solve this problem, the present invention proposes a solution to add energy storage units to wind farms. The introduction of energy storage units, through their charging and discharging functions, can effectively balance the output power of wind farms, so that transmission cables can operate at lower power, thereby reducing transmission costs. At the same time, the construction cost and operating cost of the energy storage unit must also be taken into consideration to ensure the maximization of the overall economic benefits.

[0030] By constructing a refined optimization model, this scheme not only considers the optimal layout of energy storage units, but also covers the overall collection and transmission costs of the wind farm. To achieve this goal, a composite optimization algorithm combining the alternating direction multiplier method (ADMM) and the reinforcement learning (RL) algorithm is proposed. The algorithm can accurately determine the optimal installation location of the energy storage unit to ensure that the transmission cost is minimized while meeting the internal collection needs of the wind farm, thereby optimizing the collection and transmission costs of the wind farm as a whole.

[0031] In addition, this solution also provides a key decision reference: when the transmission distance of the wind farm exceeds a certain threshold, this solution will be particularly suitable. Through this innovative energy storage unit design and optimization strategy, wind farm operators can not only maximize cost-effectiveness, but also ensure the long-term stable operation of the wind farm and the continuous growth of economic benefits.

[0032] Research on reinforcement learning shows that without human intervention and guidance, reinforcement learning can still achieve great success, even surpassing the guidance of human experience. It has the advantages of fast solution, strong generalization ability, and high solution accuracy, providing a new idea and method for solving combinatorial optimization problems. Related work in the industry has gradually been carried out. In 2018, Lin applied RL to the vehicle management and dispatching problems in shared travel; in 2019, Mao applied RL methods to distributed cluster task scheduling; in 2020, Mirhoseini applied RL to chip layout design. Zhang applied RL to the NP-hard workshop scheduling problem. These studies all use RL methods to solve combinatorial optimization problems in real life. The core idea is: the function of RL sequential decision-making has a natural similarity with combinatorial optimization problems with sequential decision-making properties. The RL model can obtain a better strategy for the problem through the continuous interaction between the intelligent agent and the environment and gradually accumulate experience. In the case of a small number of samples or even no samples, it can quickly solve the optimization problems in real life through self-learning, thereby obtaining the solution to the optimization problem. RL enables the agent to interact with the environment continuously, and uses reward values ​​to motivate learning to obtain data, overcoming the problem of the cost of a large number of labels in supervised learning.

[0033] Although solutions for photovoltaic power generation and transmission modes have emerged in the field of photovoltaic power generation. These systems usually include photovoltaic modules, grid-connected inverters, AC-coupled inverters, batteries, and energy management systems (EMS). These components work together to effectively respond to the power demand of the power grid, achieve load peak shaving, and ensure the safety and reliability of electricity use. The successful application of photovoltaic energy storage systems provides valuable experience and reference for the design of wind farm energy storage units.

[0034] However, compared with photovoltaic power generation, the environmental conditions of offshore wind farms are more complex, including salt spray corrosion, wave impact, and wind force changes, which put higher requirements on the design and optimization of energy storage units. In addition, the scale of offshore wind farms is usually larger, and the layout and capacity configuration of energy storage units are also more complex.

[0035] Therefore, based on the prior art, the present invention further explores the economic benefits brought by the location selection of the energy storage unit in the planning of the wind farm collection and transmission system, because it is directly related to the design of the collection line, the transmission efficiency and the overall cable cost.

[0036] The optimization model and algorithm design of the present invention, through precise mathematical models and advanced reinforcement learning algorithms, can provide a powerful tool for wind farm operators to minimize construction investment costs and achieve efficient and stable operation of wind farms.

[0037] As an important component of renewable energy, wind farms play an increasingly critical role in the transformation of energy structure. Traditional wind farms usually use the method of arranging wind turbine cables and directly connecting to the grid point or connecting to the grid through an offshore booster station. Such a design requires that the transmission cables must be able to carry the rated capacity of the wind farm to ensure full load operation when the wind is strong. However, the randomness and uncertainty of wind resources means that wind farms are not always able to generate electricity that matches the rated capacity, resulting in the transmission cables not always being able to operate at full load. In the case of long-distance offshore power transmission, this design is not only inefficient, but also leads to huge transmission costs.

[0038] Example 1 In order to solve the above technical problems, the introduction of energy storage technology provides new possibilities for the efficient operation of wind farms. The energy storage system can balance the output power of the wind farm by storing excess electrical energy and releasing it when the wind is insufficient, so that the transmission cables can operate at full load more frequently. This strategy not only improves the efficiency of power transmission, but also helps to reduce the costs associated with transmission cables. When planning the location of the energy storage system, we must not only consider its impact on the cost of transmission cables, but also consider the collection cost and transmission cost of the entire wind farm system to comprehensively consider reducing the construction cost of the wind farm.

[0039] The optimal layout of energy storage units is also crucial to reducing the cost of electricity collection within a wind farm. A reasonable location selection can reduce the length and complexity of internal cables, thereby reducing the construction cost of the electricity collection system. Therefore, comprehensively considering the cost changes of both electricity collection and transmission and building a more reasonable design model is the key to maximizing the cost-effectiveness of wind farms.

[0040] Due to the unique structure of the problem, in the objective function of the problem, the transmission cost function is a continuous convex function, while the collection cost function is a combinatorial optimization problem, which is non-convex and non-continuous, which makes it impossible to use a simple gradient descent algorithm for optimization. At the same time, the approximate method for cable layout optimization also has the need to improve the optimality of the solution and increase the speed of solution. Therefore, it is necessary to design an efficient optimization algorithm based on the problem model.

[0041] This embodiment mainly includes the following steps.

[0042] Step S1, first, according to the obtained annual wind resource data about the wind farm, the capacity of the energy storage unit is selected, and then the operating cost of the energy storage can be estimated.

[0043] Step S2, then, a wind farm model is created according to the transmission problem of energy storage, and the model is represented in the form of a cost function.

[0044] The above-mentioned wind farm model mainly includes a first model and a second model. The first model mainly includes the transmission cost between the energy storage location and the grid connection point location, and the second model is equivalent to the cost function of the cable layout optimization problem.

[0045] For the modeling of wind farm transmission problems based on energy storage, the first model mentioned above is composed of the transmission cost between the energy storage location and the grid connection point location, which can be regarded as a continuous convex function; the second model is composed of the wind farm cable layout optimization affected by the energy storage location. The wind farm cable layout optimization is a path combination optimization problem. The shortest collection path to the energy storage will change when the energy storage location is different. This part is a non-convex and non-continuous function.

[0046] The objective function of the wind farm model is the sum of the transmission cost between the energy storage location and the grid connection point, the cost function of the cable layout optimization problem, and the operating cost of the energy storage. Using the variable x to represent the location information of the energy storage, the cost function f(x) of the system is specifically expressed as: f(x)=f 1 (x)+f 2 (x)+c Among them, f 1 (x) = α||xa|| 2 represents the transmission cost between the energy storage location and the grid connection point, α is the unit cost of the cable, and ||xa|| 2 represents the distance between the energy storage location and the grid connection point, a represents the location of the grid connection point; f 2 (x) is the cost function of a cable layout optimization problem, which is a combinatorial optimization problem; c represents the cost of energy storage.

[0047] Step S3, solve the cost function according to the ADMM algorithm and the reinforcement learning solution algorithm, and determine the optimal energy storage solution, and select the optimal connection method for all wind turbines to maximize the power transmission efficiency of the system and minimize the cable loss.

[0048] Step S3 mainly includes the following sub-steps.

[0049] Step S31, first convert the cost function into an augmented Lagrangian equivalent form, and then introduce the ADMM algorithm to iteratively update the variable x, the auxiliary variable z and the Lagrangian multiplier λ.

[0050] In more detail, in order to enable the problem to be processed by the ADMM algorithm, an auxiliary variable z is introduced to transform the objective function and constraints of the problem into: The augmented Lagrangian equivalent form is: Where λ is the Lagrange multiplier and ρ>0.

[0051] In the ADMM algorithm, x, z, and the Lagrange multiplier λ need to be updated iteratively.

[0052] (1) Given the current z k and the Lagrange multiplier λ k , update x:

[0053] (2) Given the current x k+1 and the Lagrange multiplier λ k , update z:

[0054] (3) Update the Lagrange multiplier λ: λ k+1 =λ k +ρ(x k+1 -z k+1 ).

[0055] Step S32, directly using the gradient method to solve the minimization problem of the continuous variable x. More specifically, the formula in (1) above is a minimization problem of the continuous variable x, which can be solved using the gradient method and its displayed analytical solution can be written out.

[0056] Step S33: In order to solve the cable optimization problem, a reinforcement learning algorithm is used to solve it. More specifically, the formula in (2) above is a cable optimization problem. In order to solve this cable optimization problem, a reinforcement learning method is used.

[0057] This step specifically includes the following process.

[0058] Step S331, define the state, action and reward function of reinforcement learning; then initialize the reinforcement learning model.

[0059] More detailed, State (State,S t ), which mainly includes wind turbine coordinates, energy storage location and current cable connection scheme. t) is defined as adjusting the cable connection structure, including the following possible operations: changing the connection path between a wind turbine and the energy storage location; reconnecting the lines between different wind turbines and selecting the cable type.

[0060] Reward function (Reward, R t ), the reward function measures the contribution of each action to the overall layout optimization, the total length of the cable: the goal is to minimize the cable length, the impact of cable selection; the final reward function can be expressed as: R t =-(d(i,j)+π i +π j ) Where d(i,j) represents the distance between two fans, π i , π j is the penalty term added to the two fans.

[0061] Initializing the reinforcement learning model mainly includes: initializing the state space and action space; randomly generating the initial cable connection scheme S 0 , initialization strategy π 0 (a|s), can select an action.

[0062] Step S332, action selection and state transition; according to the current state S t , by strategy π 0 ( A t | S t )Select an action A t , including the connected wind turbines and the selected cable type; then update the status S of the wind turbines in the wind farm t .

[0063] Step S333, calculate the fitness value of each state; it includes the following process.

[0064] Step S3331, generating a corresponding cable connection structure according to the current status of the wind turbine and energy storage position.

[0065] Step S3332, using the connection scheme generated by the cable, calculate the target value in this state according to the reward function in the above steps, that is, the various components in the reward function.

[0066] Step S3333, the target value is used as the fitness value of this state, which is equivalent to the reward value.

[0067] Step S334, using a reinforcement learning method to update the strategy; specifically: Where α is the learning rate, γ is the discount factor, Q(s t ,a t ) means in state s t Next, perform action a t value.

[0068] Step S335, iterative update, outputting the optimal cable connection solution; it mainly includes the following process.

[0069] Step S3351, updating the local optimal value and the global optimal value according to the current fitness value; Step S3352, by repeating steps S332-S334, the strategy is continuously adjusted until the fitness value reaches a maximum or a preset iteration stop condition is met; Step S3353: After multiple iterations, the optimal cable connection solution is output.

[0070] Step S34, after satisfying the convergence condition or the iteration time, after multiple iterations, the optimal energy storage position is finally determined. More specifically, the equations in (1), (2) and (3) are iterated until the convergence condition of the approximate solution is satisfied or the set iteration time is satisfied; after multiple iterations, the optimal energy storage position is determined.

[0071] Step S4, compare the cost of the above-mentioned optimized energy storage solution and the solution without energy storage unit to obtain the comparison result, finally obtain the optimized solution result, and perform the optimized layout according to the optimized solution result. In more detail, step S4 mainly includes the following sub-steps.

[0072] Step S41, comparing the economic costs of the optimized energy storage solution and the solution without using the energy storage unit, and determining to use the optimized energy storage solution after the transmission cable exceeds a kilometers, where a is greater than 0.

[0073] Step S42, determining the final optimization solution result, which includes the optimized energy storage solution adopted when the transmission cable exceeds a kilometers, the construction location of the energy storage unit, and the wind farm cable layout result.

[0074] refer to Figure 2 , a simple example is given to illustrate the above process of this embodiment. It is assumed that there is a wind farm, which includes several wind turbines, including C1, C2, C3, C4 and energy storage devices connected to these wind turbines. The cable connection path between the wind turbines needs to be optimized to improve the transmission efficiency of electricity and reduce the overall cost. The decision of each wind turbine not only affects its connection with other wind turbines, but also affects the charging and discharging state of the energy storage device. This problem is a combinatorial optimization problem affected by the location of the energy storage, in which each wind turbine makes a decision based on the current state.

[0075] Wind turbine C1 makes its own decision based on the current status (wind turbine connection status, connection status with energy storage, and cable selection), for example, connecting to C2 or C4. The reward for connecting to C2 is 50, and the reward for connecting to C4 is -50. Figure 2 Iterate to get the current optimal choice.

[0076] Each time a wind turbine makes a choice, it is rewarded or penalized based on the outcome. For example, if the connected wind turbine brings about an efficiency improvement in power transmission (such as reduced losses or enhanced load balancing), a higher reward (such as +50) will be given; if the chosen connection results in excessive cable losses, a negative reward (such as -50) will be given.

[0077] The present embodiment involves an integrated optimization method for offshore wind power collection and transmission for energy storage, which can achieve the optimal layout of cables inside the wind farm based on the energy storage unit and minimize the cost of collection, storage and transmission of the entire wind farm.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention is described in detail with reference to the embodiments, it should be understood by those skilled in the art that any modification or equivalent replacement of the technical solutions of the present invention does not depart from the spirit and scope of the technical solutions of the present invention and should be included in the scope of the claims of the present invention.

Claims

1. An offshore wind power collection and transmission integrated optimization method for energy storage, characterized in that: The following steps are involved: S1, based on the annual wind resource data of the wind farm, select the capacity of the energy storage unit and estimate the operating cost of the energy storage; S2, establish a wind farm model based on the transmission problem of energy storage and express it in the form of a cost function; S3, using the ADMM algorithm and the reinforcement learning-based solution algorithm to solve the cost function and determine the optimal energy storage solution; S4, compare the cost with the solution without the energy storage unit to obtain the comparison result, and finally obtain the optimization solution result.

2. The offshore wind power collection and transmission integrated optimization method for energy storage according to claim 1, characterized in that: The wind farm model includes a first model and a second model. The first model includes the transmission cost between the energy storage location and the grid connection point location; the second model is equivalent to the cost function of the cable layout optimization problem.

3. The offshore wind power collection and transmission integrated optimization method for energy storage according to claim 1 or 2, characterized in that: The objective function of the wind farm model is the sum of the transmission cost between the energy storage location and the grid connection point location, the cost function of the cable layout optimization problem, and the operating cost of the energy storage.

4. The offshore wind power collection and transmission integrated optimization method for energy storage according to claim 3, characterized in that: The step S3 comprises: S31, convert the cost function into an augmented Lagrangian equivalent form, and introduce the ADMM algorithm to iteratively update the variables x and z and the Lagrangian multiplier λ; S32, using the gradient method to solve the minimization problem of the continuous variable x; S33, uses reinforcement learning algorithm to solve the cable optimization problem; S34, multiple iterations are performed to finally determine the optimal energy storage location.

5. An offshore wind power collection and transmission integrated optimization method for energy storage according to claim 1, 2 or 4, characterized in that: The step S4 comprises: S41, comparing the economic costs of the optimized energy storage solution and the solution without the energy storage unit, and determining to adopt the optimized energy storage solution after the transmission cable exceeds a kilometers, where a is greater than 0; S42, determining the final optimization solution result, which includes the optimized energy storage solution adopted when the transmission cable exceeds a kilometers, the location of the energy storage unit construction, and the wind farm cable layout result.

6. The offshore wind power collection and transmission integrated optimization method for energy storage according to claim 4, characterized in that: The step S33 comprises: S331, define the state, action and reward function of reinforcement learning; initialize the reinforcement learning model; S332, action selection and state transition; S333, calculating the fitness value of each state; S334, updating strategies using reinforcement learning methods; S335, iterative update, output the optimal cable connection solution.

7. The offshore wind power collection and transmission integrated optimization method for energy storage according to claim 6, characterized in that: The state includes wind turbine coordinates, energy storage location and current cable connection scheme; the action is to adjust the cable connection structure, and the reward function measures the contribution of each action to the overall layout optimization.

8. An offshore wind power collection and transmission integrated optimization method for energy storage according to claim 6 or 7, characterized in that: The step S332 includes: Based on the current status, the strategy selects an action, including the connected wind turbines and the selected cable type; then the status of the wind turbines in the wind farm is updated.

9. The offshore wind power collection and transmission integrated optimization method for energy storage according to claim 7, characterized in that: The step S333 includes: S3331, generating a corresponding cable connection structure according to the wind turbine and energy storage position in the current state; S3332, using the connection scheme generated by the cable, calculate the target value in this state according to the reward function; S3333, the target value is used as the fitness value of this state, which is equivalent to the reward value.

10. An offshore wind power collection and transmission integrated optimization method for energy storage according to claim 6 or 7, characterized in that: The step S335 includes: S3351, updating the local optimal value and the global optimal value according to the current fitness value; S3352, by repeating steps S332-S334, the strategy is continuously adjusted until the fitness value reaches a maximum or a preset iteration stop condition is met; S3353, after multiple iterations, outputs the optimal cable connection solution.

Images