Optimal configuration method for offshore wind power generation and energy storage system

Abstract

The invention discloses an optimal configuration method for an offshore wind power generation and energy storage system. The method comprises the following steps: respectively establishing a wind power plant power output model and an energy storage system model; based on the capacity of the wind power plant and the capacity of the energy storage system, establishing a capacity configuration optimization problem of the wind storage system; establishing an optimization objective function based on the capacity configuration optimization problem of the wind storage system; and based on the optimization objective function, solving a capacity configuration optimization problem of the wind storage system by adopting a deep Q network reinforcement learning algorithm, and obtaining an optimization configuration result. The wind power volatility and the adjusting capacity of the energy storage system are combined, the wind storage capacity optimization model based on the deep Q network is designed, the capacity configuration of the wind power and energy storage system is optimized, the wind power absorption capacity is improved, and the economical efficiency and stability of the system are ensured.

Description

Technical Field

[0001] This invention belongs to the field of power system optimization configuration, specifically relating to a method for optimizing the configuration of offshore wind power generation and energy storage systems. Background Technology

[0002] Coastal cities, with their abundant wind energy resources, have become important areas for wind energy development. With the continuous advancement and application of wind power technology, wind farms in coastal cities have gradually achieved large-scale integration of renewable energy. Wind power can not only reduce dependence on traditional fossil fuels but also significantly reduce carbon emissions generated during power generation. However, as the proportion of wind energy in the power system gradually increases, the volatility and uncertainty of wind power are becoming key challenges to the stable operation of the system. The intermittency and randomness of wind power can lead to power fluctuations, frequency instability, and reduced power quality in the power grid, placing higher demands on the balance and stability of the power system.

[0003] Currently, the dispatching methods for wind farms in coastal cities typically rely on a single energy source, lacking comprehensive dispatching and coordination of wind power and energy storage systems. In scenarios involving combined wind and energy storage operation, existing capacity allocation methods fail to fully utilize the complementarity between wind energy and energy storage systems, making it difficult to cope with system shocks caused by wind power fluctuations. This leads to problems such as low wind energy utilization, idle energy storage equipment, and even wind curtailment of renewable energy, failing to meet the stability and economic requirements of the power grid under a high proportion of renewable energy integration. Summary of the Invention

[0004] Purpose of the invention: To overcome the shortcomings of existing technologies, this invention provides an optimized configuration method for offshore wind power generation and energy storage systems. By combining the volatility of wind power with the regulation capability of energy storage systems, a wind-storage capacity optimization model based on deep Q-networks is designed to optimize the capacity configuration of wind power and energy storage systems, improve the absorption capacity of wind power, and ensure the economic efficiency and stability of the system.

[0005] Technical Solution: To achieve the above objectives, this invention provides a method for optimizing the configuration of an offshore wind power generation and energy storage system, comprising the following steps:

[0006] S1: Establish the power output model of the wind farm and the energy storage system model respectively;

[0007] S2: Based on the wind farm capacity and energy storage system capacity, establish the capacity configuration optimization problem of the wind-storage system;

[0008] S3: Based on the capacity configuration optimization problem of wind-storage systems, establish an optimization objective function;

[0009] S4: Based on the optimization objective function, a deep Q-network reinforcement learning algorithm is used to solve the capacity configuration optimization problem of the wind storage system and obtain the optimization configuration result.

[0010] Furthermore, the establishment of the wind farm power output model in step S1 includes:

[0011] The wind turbine power curve is a mapping relationship between wind speed v and wind power output P, ​​and it follows the following rules:

[0012] When the wind speed is lower than the starting wind speed v s At that time, the wind turbine's output power is zero;

[0013] When the wind speed reaches the starting wind speed v s And less than the rated wind speed v r At that time, the output power of the wind turbine gradually increases;

[0014] When the wind speed reaches the rated wind speed v r At that time, the power output of the wind turbine reaches its maximum value Pr;

[0015] When the wind speed exceeds the cutoff wind speed v c At that time, the wind turbine stopped generating electricity, and the power output was zero;

[0016] Therefore, the following formula (1) is established:

[0017]

[0018] Where P(t) is the output power of the wind turbine at time t, v(t) is the wind speed at time t, and v s It's the starting fan speed, v r It is the rated wind speed; v c It cuts off the wind speed, P r This is the rated power of the wind turbine generator set;

[0019] A single wind farm consists of multiple wind turbines, and its total power output is the sum of the power outputs of all the wind turbines. For N wind turbines, the total power output P of the wind farm is... total (t) is:

[0020]

[0021] Among them, P w,i (t) is the output power of the i-th wind turbine at time t; N is the number of wind turbines in the wind farm.

[0022] Furthermore, the establishment of the energy storage system model in step S1 includes:

[0023] Energy storage systems are primarily used to balance the gap between wind power output and load demand. They absorb excess wind power through charging and compensate for unmet load demands through discharging. The power output model, maximum charge / discharge power, and energy storage limits of energy storage devices are detailed below:

[0024] The charging and discharging power P of the energy storage system s (t) is represented as:

[0025]

[0026] Among them, P charge (t) is the charging power of the energy storage system at time t; P discharge (t) is the discharge power of the energy storage system at time t;

[0027] Energy State E of Energy Storage System s (t) represents the energy storage device's charge at time t. Affected by charging and discharging power, its change within each time step is calculated using the following formula:

[0028]

[0029] Where, η charge It refers to charging efficiency; η discharge It represents the discharge efficiency; Δt is the time step.

[0030] Energy State E of Energy Storage System s (t) is limited by energy storage capacity; let the maximum storage capacity of the energy storage system be E. s,max The minimum energy storage capacity is E s,min for:

[0031] E s,min ≤E s (t)≤E s,max (5)

[0032] Furthermore, in the energy storage system model of step S1, to prevent overcharging or over-discharging of the energy storage system, the charging power P... charge (t) and discharge power P discharge (t) is limited by the maximum charge and discharge power:

[0033]

[0034] Among them, P s,max It is the maximum charging and discharging power of the energy storage system.

[0035] Furthermore, the establishment of the capacity configuration optimization problem of the wind-storage system in step S2 includes:

[0036] The optimization objective of the capacity configuration optimization problem of wind-storage systems is to select a suitable wind farm capacity C. w and energy storage system capacity C s This minimizes the system's operating costs and ensures maximum wind power absorption.

[0037] The capacity configuration optimization problem is represented as:

[0038]

[0039] Where, f(C) w C s ) is the objective function, which aims to minimize the operating cost of the system while satisfying the system constraints.

[0040] Furthermore, the objective function optimized in step S3 includes two parts: system operating cost and renewable energy absorption rate, as detailed below:

[0041] System operating costs:

[0042] System operating cost C total The cost includes two parts: the charging and discharging cost of the energy storage system and the grid electricity purchase cost. The charging and discharging cost of the energy storage system comprises the energy loss and expense incurred during charging and discharging of the energy storage equipment. The grid electricity purchase cost is the cost of purchasing electricity from the grid when wind power and energy storage systems cannot meet load demand. The objective function for the operating cost is expressed as:

[0043]

[0044] Among them, C battery (t) is the cost of the energy storage device's charging and discharging power, C grid (t) is the cost of purchasing electricity from the grid;

[0045] Renewable energy integration rate:

[0046] The objective function for renewable energy absorption rate is defined as follows:

[0047]

[0048] Wherein, min(P) w (t),L(t)) represents the actual wind power absorbed. Maximizing the absorption rate helps reduce wind power waste.

[0049] The optimized objective function f(Cw,Cs) combination is as follows:

[0050] f(Cw,Cs)=αC total -βη renewable (10)

[0051] Here, α and β are the first and second weighting coefficients, respectively, used to balance cost and absorption rate.

[0052] Furthermore, the process of solving the capacity configuration optimization problem of the wind-storage system using a deep Q-network reinforcement learning algorithm in step S4 includes:

[0053] A1: Define the state space and action space, and design the reward function;

[0054] A2: Based on the state space, action space, and reward function, a deep Q-network algorithm is used to solve the capacity optimization problem. By approximating the Q-value function through a neural network, the optimal wind power and energy storage capacity configuration can be found.

[0055] Further, step A1 specifically includes:

[0056] In reinforcement learning models, the agent's goal is to make the optimal decision based on the system's current state. To achieve this, we first define the system's state space, action space, and reward function.

[0057] The state space is:

[0058] s(t)=[P w (t),E s (t),L(t),P grid (t)](11)

[0059] Among them, P w (t) is the output power of the wind farm; E s (t) represents the energy state of the energy storage system; L(t) represents the load demand of the system; P grid (t) represents the power purchased by the power grid;

[0060] The action space a(t) represents the charging and discharging power P of the energy storage system. s (t), which is the agent's decision variable, whose action space is affected by the maximum charging and discharging power P of the energy storage system. s,max The constraints and motion space are:

[0061]

[0062] The design of the reward function includes:

[0063] The core of reinforcement learning is the reward function, which the agent optimizes to achieve the optimal policy. The reward function includes the operating cost (Equation 8) and the renewable energy absorption rate (Equation 9). Taking into account the weighting coefficients α and β of the operating cost and the renewable energy absorption rate, the objective function is expressed as:

[0064] R(t) = -C total (t)+γηrenewable (t) (13)

[0065] Here, γ is a discount factor used to balance long-term and short-term rewards.

[0066] Furthermore, step A2 specifically includes:

[0067] B1: The Q-value function Q(s(t), a(t)) represents the expected reward for taking action a(t) in state s(t). The Q-value update formula is:

[0068] Q(s(t),a(t))←Q(s(t),a(t))+a(R(t)+γmax a(t+1) Q(s(t+1),a(t+1))-Q(s(t),a(t))) (14)

[0069] Where a is the learning rate; max a(t+1) Q(s(t+1),a(t+1)) is the maximum Q value at the next time step, representing the best action to be chosen in the next state;

[0070] B2: A neural network is used to approximate the Q-value function. The input of the neural network is the system state s(t), and the output is the Q-value of each action. The neural network gradually optimizes the Q-value function through training, so that the agent can select the optimal charging and discharging strategy under different wind speeds, loads and other conditions.

[0071] Furthermore, the training process of the neural network in step B2 includes:

[0072] C1: Initialize the Q-value function and randomly initialize the neural network parameters;

[0073] C2: At each time step t, select action a(t) based on the current state s(t);

[0074] C3: Execute the action to obtain a new state s(t+1) and a reward R(t);

[0075] C4: Update the Q-value function and optimize the neural network;

[0076] C5: Repeat the training process until the optimal policy is converged.

[0077] This invention proposes an innovative wind-storage capacity optimization algorithm, aiming to improve the overall operating efficiency of wind farms by optimizing the capacity configuration of wind power and energy storage devices. The algorithm focuses on improving wind power absorption capacity, combining the regulation characteristics of energy storage systems to scientifically adjust wind and storage capacities, optimize power dispatch strategies, and reduce wind curtailment while lowering system operating costs. Through reasonable wind and storage capacity configuration, the system burden caused by wind power output fluctuations can be effectively addressed, improving grid stability and operational security. This invention not only provides an efficient wind and storage capacity optimization method for coastal city wind farms but also serves as a reference for other power systems with high proportions of renewable energy integration, possessing high theoretical value and broad application prospects.

[0078] The present invention has the following characteristics:

[0079] By combining the volatility of wind power with the regulation capability of energy storage systems, and fully exploring the complementary potential of the two through the deep Q-network algorithm, the synergistic optimization of wind-storage systems is achieved.

[0080] The capacity configuration of wind power and energy storage systems has been optimized, maximizing the absorption rate of renewable energy, reducing wind power waste, and significantly improving the wind power absorption rate.

[0081] By using a deep Q-network reinforcement learning algorithm, the system's capacity can be self-learned and optimized, enabling it to cope with the uncertainty of wind power output and load fluctuations, thereby improving the system's stability and economy.

[0082] Based on the construction of a capacity optimization problem model for wind power and energy storage systems, this invention fully explores the complementary potential of combining the volatility of wind power with the regulation capabilities of energy storage systems to optimize the capacity configuration of wind power and energy storage systems, maximize wind power absorption rate and minimize system operating costs. It designs a wind-storage capacity optimization model based on deep Q-networks and uses a deep Q-network learning algorithm to optimize the scheduling of wind and storage capacity, thereby effectively addressing the uncertainty of wind power output, improving wind power absorption capacity, and ensuring the economic efficiency and stability of the system.

[0083] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0084] 1. Maximize wind power absorption rate: By rationally allocating the capacity of wind power and energy storage systems, wind curtailment is reduced, effectively increasing the absorption rate of renewable energy and promoting the development of low-carbon power systems.

[0085] 2. Reduce system operating costs: By adopting a deep Q-network reinforcement learning algorithm, the configuration strategy of wind farms and energy storage systems is optimized, which reduces energy loss during the charging and discharging process of energy storage equipment and the cost of purchasing electricity from the grid.

[0086] 3. Adaptive scheduling strategy: The training process based on deep Q-network enables the agent to adjust the wind and storage capacity configuration in real time, optimize the scheduling strategy, and improve the system's adaptability to wind power output and load fluctuations.

[0087] 4. Improve system robustness: This invention utilizes the self-learning mechanism of the deep Q-network algorithm to automatically optimize the scheduling strategy under uncertain environments, thereby enhancing the system's adaptability to uncertainties and disturbances. Attached Figure Description

[0088] Figure 1 This is a schematic diagram of the training process of a deep Q-network.

[0089] Figure 2 The image shows the results of the configuration optimization. Detailed Implementation

[0090] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0091] Example 1:

[0092] This embodiment provides a method for optimizing the configuration of an offshore wind power generation and energy storage system, including the following steps:

[0093] S1: Establish the power output model of the wind farm and the energy storage system model respectively;

[0094] The establishment of the wind farm power output model includes:

[0095] The wind turbine power curve is a mapping relationship between wind speed v and wind power output P, ​​and it follows the following rules:

[0096] When the wind speed is lower than the starting wind speed v s At that time, the wind turbine's output power is zero;

[0097] When the wind speed reaches the starting wind speed v s And less than the rated wind speed v r At that time, the output power of the wind turbine gradually increases;

[0098] When the wind speed reaches the rated wind speed v r At that time, the power output of the wind turbine reaches its maximum value Pr;

[0099] When the wind speed exceeds the cutoff wind speed v c At that time, the wind turbine stopped generating electricity, and the power output was zero;

[0100] Therefore, the following formula (1) is established:

[0101]

[0102] Where P(t) is the output power of the wind turbine at time t, v(t) is the wind speed at time t, and v s It's the starting fan speed, v r It is the rated wind speed; v c It cuts off the wind speed, P r This is the rated power of the wind turbine generator set;

[0103] A single wind farm consists of multiple wind turbines, and its total power output is the sum of the power outputs of all the wind turbines. For N wind turbines, the total power output P of the wind farm is... total (t) is:

[0104]

[0105] Among them, P w,i (t) is the output power of the i-th wind turbine at time t; N is the number of wind turbines in the wind farm.

[0106] The establishment of the energy storage system model includes:

[0107] Energy storage systems are primarily used to balance the gap between wind power output and load demand. They absorb excess wind power through charging and compensate for unmet load demands through discharging. The power output model, maximum charge / discharge power, and energy storage limits of energy storage devices are detailed below:

[0108] The charging and discharging power P of the energy storage system s (t) is represented as:

[0109]

[0110] Among them, P charge (t) is the charging power of the energy storage system at time t; P discharge (t) is the discharge power of the energy storage system at time t;

[0111] Energy State E of Energy Storage System s (t) represents the energy storage device's charge at time t. Affected by charging and discharging power, its change within each time step is calculated using the following formula:

[0112]

[0113] Where, η charge It refers to charging efficiency; η discharge It represents the discharge efficiency; Δt is the time step.

[0114] Energy State E of Energy Storage Systems (t) is limited by energy storage capacity; let the maximum storage capacity of the energy storage system be E. s,max The minimum energy storage capacity is E s,min for:

[0115] E s,min ≤E s (t)≤E s,max (5)

[0116] In the energy storage system model, to prevent overcharging or over-discharging of the energy storage system, the charging power P... charge (t) and discharge power P discharge (t) is limited by the maximum charge and discharge power:

[0117]

[0118] Among them, P s,max It is the maximum charging and discharging power of the energy storage system.

[0119] S2: Based on the wind farm capacity and energy storage system capacity, establish the capacity configuration optimization problem of the wind-storage system;

[0120] The establishment of the capacity configuration optimization problem for wind-storage systems includes:

[0121] The optimization objective of the capacity configuration optimization problem of wind-storage systems is to select a suitable wind farm capacity C. w and energy storage system capacity C s This minimizes the system's operating costs and ensures maximum wind power absorption.

[0122] The capacity configuration optimization problem is represented as:

[0123]

[0124] Where, f(C) w C s ) is the objective function, which aims to minimize the operating cost of the system while satisfying the system constraints.

[0125] S3: Based on the capacity configuration optimization problem of wind-storage systems, establish an optimization objective function;

[0126] The optimization objective function includes two parts: system operating cost and renewable energy integration rate, as detailed below:

[0127] System operating costs:

[0128] System operating cost C totalThe cost includes two parts: the charging and discharging cost of the energy storage system and the grid electricity purchase cost. The charging and discharging cost of the energy storage system comprises the energy loss and expense incurred during charging and discharging of the energy storage equipment. The grid electricity purchase cost is the cost of purchasing electricity from the grid when wind power and energy storage systems cannot meet load demand. The objective function for the operating cost is expressed as:

[0129]

[0130] Among them, C battery (t) is the cost of the energy storage device's charging and discharging power, C grid (t) is the cost of purchasing electricity from the grid;

[0131] Renewable energy integration rate:

[0132] The objective function for renewable energy absorption rate is defined as follows:

[0133]

[0134] Wherein, min(P) w (t),L(t)) represents the actual wind power absorbed. Maximizing the absorption rate helps reduce wind power waste.

[0135] The optimized objective function f(Cw,Cs) combination is as follows:

[0136] f(Cw,Cs)=αC total -βη renewable (10)

[0137] Here, α and β are the first and second weighting coefficients, respectively, used to balance cost and absorption rate.

[0138] S4: Based on the optimization objective function, a deep Q-network reinforcement learning algorithm is used to solve the capacity configuration optimization problem of the wind storage system and obtain the optimization configuration result;

[0139] In this embodiment, the process of using a deep Q-network reinforcement learning algorithm to solve the capacity configuration optimization problem of the wind storage system includes:

[0140] A1: Define the state space and action space, and design the reward function;

[0141] In reinforcement learning models, the agent's goal is to make the optimal decision based on the system's current state. To achieve this, we first define the system's state space, action space, and reward function.

[0142] The state space is:

[0143] s(t)=[P w (t),E s (t),L(t),P grid(t)](11)

[0144] Among them, P w (t) is the output power of the wind farm; E s (t) represents the energy state of the energy storage system; L(t) represents the load demand of the system; P grid (t) represents the power purchased by the power grid;

[0145] The action space a(t) represents the charging and discharging power P of the energy storage system. s (t), which is the agent's decision variable, whose action space is affected by the maximum charging and discharging power P of the energy storage system. s,max The constraints and motion space are:

[0146]

[0147] The design of the reward function includes:

[0148] The core of reinforcement learning is the reward function, which the agent optimizes to achieve the optimal policy. The reward function includes the operating cost (Equation 8) and the renewable energy absorption rate (Equation 9). Taking into account the weighting coefficients α and β of the operating cost and the renewable energy absorption rate, the objective function is expressed as:

[0149] R(t) = -C total (t)+γη renewable (t) (13)

[0150] Here, γ is a discount factor used to balance long-term and short-term rewards.

[0151] A2: Based on the state space, action space, and reward function, a deep Q-network algorithm is used to solve the capacity optimization problem. By approximating the Q-value function through a neural network, the optimal wind power and energy storage capacity configuration can be found.

[0152] Step A2 specifically includes:

[0153] B1: The Q-value function Q(s(t), a(t)) represents the expected reward for taking action a(t) in state s(t). The Q-value update formula is:

[0154] Q(s(t),a(t))←Q(s(t),a(t))+a(R(t)+γmax a(t+1) Q(s(t+1),a(t+1))-Q(s(t),a(t))) (14)

[0155] Where a is the learning rate; max a(t+1) Q(s(t+1),a(t+1)) is the maximum Q value at the next time step, representing the best action to be chosen in the next state;

[0156] B2: A neural network is used to approximate the Q-value function. The input of the neural network is the system state s(t), and the output is the Q-value of each action. The neural network gradually optimizes the Q-value function through training, so that the agent can select the optimal charging and discharging strategy under different wind speeds, loads and other conditions.

[0157] like Figure 1 As shown, the training process of the neural network in step B2 includes:

[0158] C1: Initialize the Q-value function and randomly initialize the neural network parameters;

[0159] C2: At each time step t, select action a(t) based on the current state s(t);

[0160] C3: Execute the action to obtain a new state s(t+1) and a reward R(t);

[0161] C4: Update the Q-value function and optimize the neural network;

[0162] C5: Repeat the training process until the optimal policy is converged.

[0163] The final optimized configuration result obtained in this implementation is as follows: Figure 2 As shown.

[0164] Example 2:

[0165] To verify the effectiveness and impact of the method of this invention, this embodiment uses typical daily load forecast data and wind power data from a coastal area in eastern China as a basis to analyze and compare three operating modes of a microgrid, as detailed below:

[0166] 1) Microgrid operation without optimization analysis;

[0167] 2) Operational analysis of wind farms with microgrid configuration;

[0168] 3) Operational analysis of microgrid configuration with wind farms and energy storage systems;

[0169] In this embodiment, the costs of the upper-level public grid power purchase cost and energy storage maintenance cost under the three operating modes of the microgrid are shown in Table 1.

[0170] Table 1 Comparison of optimization results under different operating modes

[0171]

[0172] Table 1 shows the optimized costs of various types of microgrids in coastal areas under different operating modes, indicating that the proposed solution offers the best economic and environmental benefits. In operating mode 1, the system relies entirely on grid power purchases, with a total cost of 98,108 yuan, where the purchase cost accounts for the entire cost. Since no wind power or energy storage equipment is involved, both wind power operation and maintenance and energy storage maintenance costs are zero, and there is no wind curtailment. This mode demonstrates a high-cost solution that relies entirely on the grid. In operating mode 2, wind power begins to participate in power generation, reducing the total cost to 42,103 yuan. The purchase cost is 14,613 yuan, effectively reducing the grid's power purchase demand. Although the wind power operation and maintenance cost is 27,580 yuan, there is no energy storage system, so the energy storage maintenance cost is zero. At this point, 37MW of wind power is curtailed during wind power generation, indicating that energy storage has failed to absorb excess wind power. In operating mode 3, wind power and energy storage systems work together, further reducing the total cost to 34,055 yuan, with the purchase cost at only 2,459 yuan, demonstrating the synergistic effect of wind power and energy storage. The operation and maintenance cost of wind power was 27,580 yuan, while the operation and maintenance cost of energy storage was 4,016 yuan. The amount of wind curtailment decreased to 19MW, indicating that the energy storage system more effectively absorbed wind power.

[0173] With the introduction of wind power and energy storage, the system's electricity purchase cost is significantly reduced, and wind curtailment is decreased, but this is accompanied by an increase in operation and maintenance costs. Operation mode 3 demonstrates relatively ideal economic efficiency and effectiveness, showcasing the positive impact of optimized wind power and energy storage configuration on system performance. Therefore, the optimization scheme proposed in this invention effectively improves the overall economic efficiency and sustainable development of the microgrid.

Claims

1. A method for optimizing the configuration of an offshore wind power generation and energy storage system, characterized in that, Includes the following steps: S1: Establish the power output model of the wind farm and the energy storage system model respectively; S2: Based on the wind farm capacity and energy storage system capacity, establish the capacity configuration optimization problem of the wind-storage system; S3: Based on the capacity configuration optimization problem of wind-storage systems, establish an optimization objective function; S4: Based on the optimization objective function, a deep Q-network reinforcement learning algorithm is used to solve the capacity configuration optimization problem of the wind storage system and obtain the optimization configuration result.

2. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 1, characterized in that, The establishment of the wind farm power output model in step S1 includes: The wind turbine power curve is a mapping relationship between wind speed v and wind power output P, ​​and it follows the following rules: When the wind speed is lower than the starting wind speed v s At that time, the wind turbine's output power is zero; When the wind speed reaches the starting wind speed v s And less than the rated wind speed v r At that time, the output power of the wind turbine gradually increases; When the wind speed reaches the rated wind speed v r At that time, the power output of the wind turbine reaches its maximum value Pr; When the wind speed exceeds the cutoff wind speed v c At that time, the wind turbine stopped generating electricity, and the power output was zero; Therefore, the following formula (1) is established: Where P(t) is the output power of the wind turbine at time t, v(t) is the wind speed at time t, and v s It's the starting fan speed, v r It is the rated wind speed; v c It cuts off the wind speed, P r This is the rated power of the wind turbine generator set; A single wind farm consists of multiple wind turbines, and its total power output is the sum of the power outputs of all the wind turbines. For N wind turbines, the total power output P of the wind farm is... total (t) is: Among them, P w,i (t) is the output power of the i-th wind turbine at time t; N is the number of wind turbines in the wind farm.

3. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 2, characterized in that, The establishment of the energy storage system model in step S1 includes: The charging and discharging power P of the energy storage system s (t) is represented as: Among them, P charge (t) is the charging power of the energy storage system at time t; P discharge (t) is the discharge power of the energy storage system at time t; Energy State E of Energy Storage System s (t) represents the energy storage device's charge at time t. Affected by charging and discharging power, its change within each time step is calculated using the following formula: Where, η charge It refers to charging efficiency; η discharge It represents the discharge efficiency; Δt is the time step. Energy State E of Energy Storage System s (t) is limited by energy storage capacity; let the maximum storage capacity of the energy storage system be E. s,max The minimum energy storage capacity is E s,min for: E s,min ≤E s (t)≤E s,max (5)。 4. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 3, characterized in that, In the energy storage system model of step S1, to prevent overcharging or over-discharging of the energy storage system, the charging power P... charge (t) and discharge power P discharge (t) is limited by the maximum charge and discharge power: Among them, P s,max It is the maximum charging and discharging power of the energy storage system.

5. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 4, characterized in that, The establishment of the capacity configuration optimization problem of the wind storage system in step S2 includes: The optimization objective of the wind-storage system capacity configuration optimization problem is to select the wind farm capacity C. w and energy storage system capacity C s This minimizes the system's operating costs and ensures maximum wind power absorption. The capacity configuration optimization problem is represented as: Where, f(C) w C s ) is the objective function, which aims to minimize the operating cost of the system while satisfying the system constraints.

6. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 5, characterized in that, The objective function optimized in step S3 includes two parts: system operating cost and renewable energy absorption rate, as detailed below: System operating costs: System operating cost C total The cost includes two parts: the charging and discharging cost of the energy storage system and the grid electricity purchase cost. The charging and discharging cost of the energy storage system comprises the energy loss and expense incurred during charging and discharging of the energy storage equipment. The grid electricity purchase cost is the cost of purchasing electricity from the grid when wind power and energy storage systems cannot meet load demand. The objective function for the operating cost is expressed as: Among them, C battery (t) is the cost of the energy storage device's charging and discharging power, C grid (t) is the cost of purchasing electricity from the grid; Renewable energy integration rate: The objective function for renewable energy absorption rate is defined as follows: Wherein, min(P) w (t), L(t)) represents the actual wind power absorbed; The optimized objective function f(Cw,Cs) combination is as follows: f(Cw,Cs)=αC total -be renewable (10) Here, α and β are the first and second weighting coefficients, respectively, used to balance cost and absorption rate.

7. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 6, characterized in that, The process of using a deep Q-network reinforcement learning algorithm to solve the capacity configuration optimization problem of the wind storage system in step S4 includes: A1: Define the state space and action space, and design the reward function; A2: Based on the state space, action space, and reward function, a deep Q-network algorithm is used to solve the capacity optimization problem. By approximating the Q-value function through a neural network, the optimal wind power and energy storage capacity configuration can be found.

8. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 7, characterized in that, Step A1 specifically includes: Define the system's state space, action space, and reward function; The state space is: s(t)=[P w (t),E s (t),L(t),P grid (t)](11) Among them, P w (t) is the output power of the wind farm; E s (t) represents the energy state of the energy storage system; L(t) represents the load demand of the system; P grid (t) represents the power purchased by the power grid; The action space a(t) represents the charging and discharging power P of the energy storage system. s (t), which is the agent's decision variable, whose action space is affected by the maximum charging and discharging power P of the energy storage system. s,max The constraints are: The design of the reward function includes: The core of reinforcement learning is the reward function, which the agent optimizes to achieve the optimal policy. The reward function includes the operating cost (Equation 8) and the renewable energy absorption rate (Equation 9). Taking into account the weighting coefficients α and β of the operating cost and the renewable energy absorption rate, the objective function is expressed as: R(t)=-C total (t)+γη renewable (t) (13) Here, γ is a discount factor used to balance long-term and short-term rewards.

9. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 8, characterized in that, Step A2 specifically includes: B1: The Q-value function Q(s(t), a(t)) represents the expected reward for taking action a(t) in state s(t). The Q-value update formula is: Q(s(t),a(t))←Q(s(t),a(t))+a(R(t)+γmax a(t+1) Q(s(t+1),a(t+1))-Q(s(t),a(t))) (14) Where a is the learning rate; max a(t+1) Q(s(t+1),a(t+1)) is the maximum Q value at the next time step, representing the best action to be chosen in the next state; B2: Use a neural network to approximate the Q-value function. The input of the neural network is the system state s(t), and the output is the Q-value of each action. The neural network gradually optimizes the Q-value function through training, so that the agent can select the optimal charging and discharging strategy under different states.

10. The method for optimizing the configuration of an offshore wind power generation and energy storage system according to claim 9, characterized in that, The training process of the neural network in step B2 includes: C1: Initialize the Q-value function and randomly initialize the neural network parameters; C2: At each time step t, select action a(t) based on the current state s(t); C3: Execute the action to obtain a new state s(t+1) and a reward R(t); C4: Update the Q-value function and optimize the neural network; C5: Repeat the training process until the optimal policy is converged.

Images