Optimal configuration method and device for hybrid new energy station generator set

Abstract

The invention provides an optimal configuration method and device for a hybrid new energy station generator set, and relates to the technical field of new energy stations, and the method comprises the steps: building an optimal configuration model of the hybrid new energy station generator set, and enabling the optimal configuration model to take the configuration of the generator set and an energy storage device as optimization variables, the purposes of improving the small interference stability margin and reducing the configuration number of the GFM generator set are achieved; the GAT and a DRL framework based on PPO are fused, and a DRL model based on GAT-PPO is obtained; and carrying out offline training on the DRL model based on the GAT-PPO by adopting the optimal configuration model, and outputting an optimal configuration strategy on line through the trained DRL model based on the GAT-PPO according to the external power grid strength and the hybrid new energy station topological information. The method can adapt to HRPS external power grid strength and internal topology changes, and the generated optimal configuration strategy not only can improve the small interference stability of the hybrid new energy station, but also can configure fewer GFM generator sets to obtain better economic benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy stations, and in particular to a method and device for optimizing the configuration of a hybrid new energy station generator set. Background Art

[0002] The proposal of the "dual carbon" strategic goal has promoted the development of renewable energy power generation in my country. However, the stability problems caused by the large-scale centralized access of renewable energy sources to the power grid have become increasingly prominent. The main manifestation is that when the grid-following (GFL) controlled generator set is connected to a weak power grid, it will cause a complex modal wide-band oscillation problem. In contrast, the grid-forming (GFM) controlled generator set has the ability to actively support the power grid and has stronger stability under a weak power grid. However, under a strong power grid, the GFM generator set may increase the risk of low-frequency oscillation and affect the stability of small disturbances. At the same time, too many GFM generator sets in the station will lead to higher operating costs. Therefore, how to optimize the capacity ratio and location distribution of GFM generator sets in hybrid renewable power stations (HRPS) is a research hotspot at this stage. Summary of the invention

[0003] The purpose of the present invention is to provide a method and device for optimizing the configuration of generator sets in a hybrid new energy station, which is used to solve the problems of the prior art on how to optimize the capacity ratio and location distribution of GFM generator sets in HRPS. It can adapt to the internal topology changes of HRPS and generate an optimized configuration strategy online, which can not only improve the small disturbance stability of HRPS, but also configure fewer GFM generator sets to obtain better economic benefits.

[0004] In order to achieve the above-mentioned object, in a first aspect, the present invention provides a method for optimizing configuration of a hybrid new energy station generator set, comprising:

[0005] Step 1: Establish an optimal configuration model for the generator set of a hybrid new energy station. The optimal configuration model uses the configuration of the generator set and the energy storage device as the optimization variable, with the goal of improving the small disturbance stability margin and reducing the number of GFM generator sets;

[0006] Step 2: Integrate GAT with the PPO-based DRL framework to obtain a GAT-PPO-based DRL model;

[0007] Step 3: Use the optimization configuration model to train the GAT-PPO-based DRL model offline. After the training, the GAT-PPO-based DRL model outputs the optimization configuration strategy online according to the external grid strength and the hybrid new energy station topology information.

[0008] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, the optimization configuration strategy includes the capacity ratio and location distribution of the GFM generator set and the GFL generator set in the hybrid new energy station, and the control type of the energy storage device.

[0009] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, step 2 specifically includes:

[0010] The state quantity is obtained according to the topological data and state data of the current hybrid new energy station;

[0011] Generate internal configuration actions according to the state quantity and act on the hybrid new energy station, and evaluate the internal configuration actions according to the change of the state of the hybrid new energy station;

[0012] The reward function is determined based on the relationship between the state quantity and the preset constraints.

[0013] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, the expression of the state quantity is:

[0014] w1(X)=min{ξ L1 ,ξ L2 ,...,ξ Lk}-0.05

[0015]

[0016]

[0017] S t =[S topo ,S h ]

[0018] In the formula, S t is the state quantity; s topo , S h They are the topological data and status data of the current hybrid new energy station respectively; w1, w2 and w3 are the low-frequency band small interference stability index, sub- / super-synchronous frequency band small interference stability index and operation cost index respectively; ξ Lk represents the damping ratio corresponding to the kth low-frequency closed-loop pole; ξ Sn and They represent the damping ratio and critical damping ratio of the closed-loop poles in the nth sub- / supersynchronous frequency bands respectively; N GFM is the number of GFM generator sets; X is the optimization variable, which represents the configuration vector of the generator sets and energy storage devices inside the hybrid new energy station.

[0019] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, the expression of the reward function is:

[0020]

[0021] In the formula, r is the reward value.

[0022] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, the DRL model based on GAT-PPO includes GAT, a strategy network and a value network; GAT is used to obtain a node feature set according to the state quantity of the current hybrid new energy station, the strategy network is used to generate configuration strategies for GFM generator sets, GFL generator sets and energy storage devices, and the value network is used to evaluate the advantages and disadvantages of the configuration strategies of GFM generator sets, GFL generator sets and energy storage devices.

[0023] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, step 3 specifically includes: characterizing the node characteristics of the hybrid new energy station through GAT, updating the strategy network and the value network, and applying the trained GAT-PPO-based DRL model online.

[0024] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, the strategy network is updated, including: inputting a node feature set represented by GAT into the strategy network to obtain the distribution of new strategies and old strategies; according to the distribution of the new strategies and the old strategies, respectively calculating the probability of selecting each action under the new strategy and the old strategy; dividing the probability under the new strategy by the probability under the old strategy to obtain a probability ratio; using the advantage function and the probability ratio to calculate the objective function value of the DRL model based on GAT-PPO; using the negative value corresponding to the objective function value as the loss function of the strategy network; using the loss function to update the parameters of the strategy network through back propagation until a new strategy that meets the shearing requirements is obtained.

[0025] According to a hybrid new energy station generator set optimization configuration method provided by the present invention, the value network is updated, including: inputting the node feature set represented by GAT into the value network, calculating the value function in the current state through the forward propagation mechanism, and calculating the loss function to perform gradient update on the value network.

[0026] In a second aspect, the present invention provides a hybrid new energy station generator set optimization configuration device, comprising:

[0027] Establishing a unit for establishing an optimal configuration model of a hybrid new energy station generator set, wherein the optimal configuration model uses the configuration of the generator set and the energy storage device as optimization variables, with the goal of improving the small disturbance stability margin and reducing the configuration number of the GFM generator set;

[0028] A fusion unit is used to fuse GAT with the PPO-based DRL framework to obtain a GAT-PPO-based DRL model;

[0029] The training output unit is used to perform offline training on the GAT-PPO-based DRL model using the optimized configuration model. After the training, the GAT-PPO-based DRL model outputs the optimized configuration strategy online according to the external power grid strength and the topology information of the hybrid new energy station.

[0030] The technical solution of the present invention has at least the following technical effects:

[0031] The present invention provides a hybrid new energy station generator set optimization configuration method and device, the method comprising: establishing an optimization configuration model for the hybrid new energy station generator set, the optimization configuration model uses the configuration of the generator set and the energy storage device as the optimization variable, with the goal of improving the small disturbance stability margin and reducing the configuration number of the GFM generator set; integrating GAT with the DRL framework based on PPO to obtain a DRL model based on GAT-PPO; using the optimization configuration model to perform offline training on the DRL model based on GAT-PPO, and outputting the optimization configuration strategy online based on the external power grid strength and the hybrid new energy station topology information through the DRL model based on GAT-PPO after training. The present invention can adapt to the changes in the external power grid strength and internal topology of HRPS, and the generated optimization configuration strategy can not only improve the small disturbance stability of the hybrid new energy station, but also configure fewer GFM generator sets to obtain better economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] In the attached picture:

[0034] Figure 1 Schematic diagram of the training process of the DRL model based on GAT-PPO of the present invention;

[0035] Figure 2 It is a schematic diagram of the structure of the HRPS for building a network / following a network including an energy storage device according to the present invention;

[0036] Figure 3 This is a comparison chart of the training reward curves of the GAT-PPO-based and PPO-based DRL models in scenario 1 of the present invention;

[0037] Figure 4 Optimized configuration diagram of GFM in HRPS based on GAT-PPO in three scenarios of the present invention;

[0038] Figure 5 The present invention is a flow chart of the method for optimizing the configuration of hybrid new energy station generator sets. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Some embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0041] The present invention has conducted in-depth research and found that the small-disturbance stability of HRPS is largely affected by factors such as the strength of the external power grid (short-circuit ratio), the capacity ratio of GFM / GFL generator sets, and the specific location of the GFM generator set in the HRPS. Therefore, with the configuration of the generator set and the energy storage device as the optimization variable, and with improving the small-disturbance stability margin and reducing the number of GFM generator sets as the optimization goal, an optimal configuration model for the capacity ratio and location distribution of the GFM generator set in the HRPS can be established. For the solution of the optimal configuration model, since the power grid strength of the external power system that affects the small-disturbance stability of the HRPS is time-varying, when the model-driven method is used for solution, it often faces the problems of low solution efficiency and heavy computational burden, and it is difficult to apply online. The method based on deep reinforcement learning (DRL) adopts the method of offline training and online application. The DRL model after training can quickly solve the optimal configuration model online.

[0042] At present, the small disturbance stability online optimization technology based on deep reinforcement learning focuses on the optimization of controller parameters, and is rarely applied to the optimal configuration of GFM / GFL generator sets in HRPS. Furthermore, since the topological structure of HRPS affects its small disturbance stability, it is necessary to take the topological changes of HRPS into consideration. Therefore, the present invention proposes a method for optimizing the configuration of HRPS generator sets, which is a method for online solving the optimal configuration of GFM / GFL generator sets in HRPS based on the DRL model of GAT-PPO.

[0043] The HRPS generator set optimization configuration method provided by the present invention can adapt to the internal topology changes of HRPS. The online generated optimization configuration strategy can not only improve the small disturbance stability of the hybrid new energy station, but also configure fewer GFM generator sets to obtain better economy. The method integrates the graph attention network (GAT) with stronger generalization ability for different topological structures with the deep reinforcement learning (DRL) framework based on the proximal policy optimization (PPO) algorithm to obtain a DRL model based on GAT-PPO, and then trains the optimization configuration model offline. After the training, the DRL model based on GAT-PPO outputs the optimization configuration strategy including the capacity ratio and location distribution of the GFM / GFL generator set in the HRPS and the control type of the energy storage device online according to the external power grid strength (short circuit ratio) and HRPS topology information.

[0044] See also Figure 5 The embodiment of the present invention provides a method for optimizing the configuration of a hybrid new energy station generator set, comprising:

[0045] Step 1: Establish an optimal configuration model for the generator set of a hybrid new energy station. The optimal configuration model uses the configuration of the generator set and the energy storage device as the optimization variable, with the goal of improving the small disturbance stability margin and reducing the configuration number of GFM generator sets;

[0046] Step 2: Integrate GAT with the PPO-based DRL framework to obtain a GAT-PPO-based DRL model;

[0047] Step 3: Use the optimization configuration model to train the DRL model based on GAT-PPO offline, and the DRL model based on GAT-PPO after training outputs the optimization configuration strategy online according to the external grid strength (short circuit ratio) and the topology information of the hybrid new energy station. The optimization configuration strategy includes the capacity ratio and location distribution of the GFM generator set and the GFL generator set in the hybrid new energy station, as well as the control type of the energy storage device.

[0048] Specifically, the design process of the DRL model based on GAT-PPO is as follows, that is, step 2 specifically includes:

[0049] 1) Design of state space

[0050] State quantity S t The topological data S of the current HRPS topo and status data S h Composition. Status data S hIncluding oscillation mode, damping threshold, small disturbance stability index (w1, w2) and operation cost index (w3) of low frequency band and sub / super synchronous frequency band. In order to study the interaction between HRPS and power grid, dq impedance modeling is used to calculate the characteristic value and damping ratio, and quantitatively evaluate the small disturbance stability and stability margin of the grid-connected system. Too many GFM generator sets in HRPS will lead to higher operating costs.

[0051] w1(X)=min{ξ L1 ,ξ L2 ,...,ξ Lk}-0.05 (1)

[0052]

[0053] S t =[S topo ,S h ] (4)

[0054] In the formula, ξ Lk represents the damping ratio corresponding to the kth low-frequency closed-loop pole; ξ Sn and They represent the damping ratio and critical damping ratio of the closed-loop poles in the nth sub- / supersynchronous frequency bands respectively; N GFM is the number of GFM generator sets; X is the optimization variable, and the present invention represents the configuration vector of the internal generator sets and energy storage devices of the hybrid new energy station.

[0055] 2) Design of action space

[0056] The action space is based on the current state S of HRPS t Generate internal configuration action A t , and acts on HRPS, evaluating the internal configuration action A according to the change of HRPS state t Specifically, the hybrid new energy station is used as the environment, and the configuration vector X of the generator set and energy storage device inside the HRPS at the current moment is t Treated as internal configuration action A t .

[0057] A t = [ X t ] (5)

[0058] 3) Design of reward function

[0059] The reward function is based on the state quantity S t The relationship with the preset constraints is determined, that is, the division is carried out according to the constraints and the objective function. The objective function is considered only when the constraints are met, otherwise the penalty value is returned; the current HRPS state quantity S tWhen the constraints are not met, a larger penalty value is directly returned, aiming to make the agent in DRL make decisions within the constraint domain. t Rewards are given when the constraints are met, aiming to enable the agent in DRL to find the optimal decision based on feasible decisions. Taking into account the small disturbance stability margin and the number of GFM generators, the expression of the reward function is as follows:

[0060]

[0061] In the formula, r is the reward value.

[0062] Specifically, the DRL model based on GAT-PPO proposed in the present invention mainly includes three parts: GAT, strategy network and value network. GAT is used to obtain the node feature set according to the state quantity of the current hybrid new energy station. The strategy network is used to generate the configuration strategy of GFM generator set, GFL generator set and energy storage device. The value network is used to evaluate the advantages and disadvantages of the configuration strategy of GFM generator set, GFL generator set and energy storage device.

[0063] The application process of the DRL model based on GAT-PPO proposed in the present invention mainly includes three steps, namely step 3 specifically includes: 1. Characterizing the node characteristics of the hybrid new energy station through GAT; 2. Updating the strategy network and the value network; 3. Applying the trained DRL model based on GAT-PPO online. Different from the traditional PPO algorithm, in order to perceive the changes in the HRPS topology structure, the state information of GAT-PPO needs to be updated through GAT to the node characteristics of HRPS (topology data S topo , the control type and status data of the generator set S h ) is used to characterize. In the training process, the current HRPS state quantity S t Input into GAT, output node feature set S gat Then, the node feature set S gat The inputs are respectively fed into the strategy network and the value network for training, and the internal configuration actions A of the internal generator set and the energy storage device of the HRPS are respectively output. t and to evaluate the internal configuration action A t The value function of advantages and disadvantages.

[0064] Among them, the update process of the policy network and the value network is as follows:

[0065] The node feature set S represented by GAT gatInput into the policy network to obtain the new and old policy distributions. According to the distribution of the new and old policies, calculate the probability of selecting each action under the two policies respectively. Then, divide the probability under the new policy by the probability under the old policy to obtain the probability ratio. Use the advantage function and the probability ratio to calculate the objective function value of the DRL model based on GAT-PPO. Then, use the negative value corresponding to the objective function value as the loss function of the policy network. Use the loss function to update the parameters of the policy network through back propagation until a new policy that meets the cutting requirements is obtained. The node feature set S represented by GAT gat Input into the value network, calculate the value function in the current state through the forward propagation mechanism. Then, calculate the loss function to update the gradient of the value network.

[0066] It is understandable that in the value network, the reward discount factor γ directly affects the expected long-term return evaluation in the value network. As the reward discount factor γ increases, GAT-PPO pays more attention to long-term rewards, which affects the policy update. When the loss function values ​​of the policy network and the value network are stable and close to a small value, and the moving average reward is positive and tends to be stable, it indicates that the model training has converged.

[0067] Finally, after training, the GAT-PPO-based DRL model outputs online the optimal configuration strategy including the capacity ratio and location distribution of GFM / GFL generator sets in HRPS and the control type of energy storage devices according to the external short-circuit ratio and internal topology information of HRPS.

[0068] Based on the same inventive concept, another embodiment of the present invention provides a hybrid new energy station generator set optimization configuration device, which corresponds to the method of the above embodiment, and includes:

[0069] Establishing a unit for establishing an optimal configuration model of a hybrid new energy station generator set, wherein the optimal configuration model uses the configuration of the generator set and the energy storage device as optimization variables, with the goal of improving the small disturbance stability margin and reducing the configuration number of the GFM generator set;

[0070] A fusion unit is used to fuse GAT with the PPO-based DRL framework to obtain a GAT-PPO-based DRL model;

[0071] The training output unit is used to perform offline training on the GAT-PPO-based DRL model using the optimized configuration model. After the training, the GAT-PPO-based DRL model outputs the optimized configuration strategy online according to the external power grid strength and the topology information of the hybrid new energy station.

[0072] The following is a specific application example of the present invention.

[0073] In order to verify the superiority of the online solution method for the optimal configuration strategy of the DRL model based on GAT-PPO proposed in this paper, Figure 2 The optimization configuration model is established for the grid-building / grid-following HRPS with energy storage device shown in the figure. The internal generator sets of the HRPS are all direct-drive wind turbines, which are divided into two types: grid-building control and grid-following control according to the control mode of the converter. The HRPS adopts a chain structure of multiple feeders, with a total of five feeders. Each feeder connects 18 wind turbines, including a total of 90 wind turbines, and the capacity of each wind turbine is 2MW. The wind turbines on the same feeder are connected by a bus line with a length of 0.43km. The inductance and resistance per unit length are 1.23mH / km and 0.17Ω / km respectively. A 20MW energy storage device is connected to the machine side at the PCC point (common connection point). In addition, the present invention further considers the topological form in which wind turbines at the ends of different feeders inside the hybrid new energy station can be interconnected. In order to explore the optimal configuration of GFM / GFL generator sets in HRPS under different grid strengths, the external grid short-circuit ratio K is designed. SCR There are three scenarios: 1, 2 and 3.

[0074] The optimal configuration model of the GFM / GFL generator set of the hybrid new energy station is as follows:

[0075] 1) Optimize variables

[0076] Vector X = [x1, x2, ..., x 91 ] characterizes the configuration of the GFM / GFL generator set and the energy storage device in the HRPS, so the number of optimization variables is 91. Among them, x k (k=1,2,...,90) is a 0 / 1 variable. When the value is 1, it means that the kth unit is a GFM generator set. When the value is 0, it means that the kth unit is a GFL generator set. 91 Indicates the control type of the configured energy storage device. When the value is 1, it indicates that the energy storage device is under GFM control. When the value is 0, it indicates that the energy storage device is under GFL control.

[0077] 2) Objective Function

[0078] The optimization goal is to maximize the small disturbance stability margin of the low frequency band and sub / super synchronous frequency band while minimizing the operating cost index. The small disturbance stability index of the low frequency band and sub / super synchronous frequency band is constructed based on their respective small disturbance stability margins as shown in formulas (1) and (2). Too many GFM generators in HRPS will lead to higher operating costs, as shown in formula (3). The optimization objective function is as follows:

[0079] {min{-w1(X)}; min{-w2(X)}; min{w3(X)}} (7)

[0080] Where w1, w2, w3 are the low-frequency band small interference stability index, sub- / super-synchronous band small interference stability index and operation cost index respectively; X is the optimization variable.

[0081] 3) Constraints

[0082] The following constraints need to be met: in the low frequency band, the damping ratio is greater than the stability threshold (5%); in the sub- / super-synchronous frequency band, the damping ratio is greater than the critical damping; the number of GFM generator sets is not less than 10 and not more than 50. The constraints can be expressed as:

[0083]

[0084] In the formula, f Lk , k, f St and t represent the frequency and number of low-frequency oscillation mode, and the frequency and number of sub / supersynchronous oscillation mode, respectively.

[0085] The DRL model based on GAT-PPO is used to solve the optimization configuration models in three different scenarios. Based on the parameter comparison experiment, the following parameters of the GAT-PPO method are determined: two layers of GAT are selected to characterize the topological structure of HRPS; in the structural setting of PPO, three layers of fully connected layers are selected, the total number of time steps is set to 100,000, the number of rounds is set to 2048, and the reward discount factor γ is set to 0.99. In different scenarios, the DRL model based on GAT-PPO and the DRL model based on PPO are used to solve the GFM generator capacity ratio and location distribution strategy obtained by the GFM / GFL generator set optimization configuration model in HRPS, and the corresponding objective function values ​​are shown in Table 1. The comparison diagrams of the training reward curves of the DRL model based on GAT-PPO and the DRL model based on PPO in the three scenarios are similar. Taking scenario 1 as an example, the comparison of the model training reward curves is Figure 3 The optimal configuration structure of GFM in HRPS of the DRL model based on GAT-PPO in three scenarios is shown as follows: Figure 4 shown.

[0086] Table 1 Optimal configuration strategy of GFM in HRPS under three scenarios

[0087]

[0088] As can be seen from Table 1, in scenario 1, although the number of GFM generators configured by the PPO-based DRL model and the GAT-PPO-based DRL model is the same, the stability margin of the GAT-PPO-based DRL model is 0.0601, which is 10.58% higher than the 0.0519 of the PPO-based DRL model. In scenarios 2 and 3, the number of GFM generators configured by the GAT-PPO-based DRL model is less than that of the PPO-based DRL model, and the stability margin of the GAT-PPO-based DRL model is 13.15% and 28.13% higher than that of the PPO-based DRL model, respectively. It can be seen that the optimal configuration strategy generated by the GAT-PPO-based DRL model not only reduces the number of GFM generators, but also improves the small disturbance stability margin.

[0089] Taking scenario 1 as an example, as the training progresses, Figure 3 The reward values ​​of the GAT-PPO-based and PPO-based DRL models in each round of training continue to increase, and gradually converge after about 400 rounds of training. Compared with the PPO-based DRL model, the GAT-PPO-based DRL model not only converges faster, but also improves the final convergence value of the reward function. This shows that GAT can use the attention mechanism to better extract key information and flexibly guide the agent in DRL to make the best decision.

[0090] from Figure 4 It can be seen that as the short-circuit ratio of the external grid increases from 1 to 3, the number of GFM generator sets in the HRPS decreases from 32 to 17, and the location of the GFM generator sets gradually transitions from the head end of the feeder to the end of the feeder. For example, in scenario 1, most of the GFM generator sets are configured at the front of each feeder, while in scenarios 2 and 3, most of the GFM generator sets are configured at the back of each feeder.

[0091] In summary, the hybrid new energy station generator set optimization configuration method and device proposed in the present invention adopts the DRL model based on GAT-PPO for online solution. Compared with other conventional DRL models, the PPO algorithm limits the strategy update step size through the probability ratio clipping technology, improves the stability and efficiency of the learning process, and has the advantages of fast convergence speed and wide adaptability; and GAT can effectively characterize the topological information of HRPS and the control type of the generator set, so that the optimization configuration strategy finally solved can better adapt to the topological changes of HRPS. Therefore, the present invention can adapt to the external power grid strength and internal topological changes of HRPS, and the generated optimization configuration strategy can not only improve the small disturbance stability of the hybrid new energy station, but also configure fewer GFM generator sets to obtain better economic benefits.

[0092] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the embodiments disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A hybrid new energy station generator set optimization configuration method, characterized in that: include: Step 1: Establish an optimal configuration model for the generator set of a hybrid new energy station, wherein the optimal configuration model uses the configuration of the generator set and the energy storage device as optimization variables, with the goal of improving the small disturbance stability margin and reducing the configuration number of the GFM generator set; Step 2: Integrate GAT with the PPO-based DRL framework to obtain a GAT-PPO-based DRL model; Step 3: Use the optimization configuration model to perform offline training on the GAT-PPO-based DRL model. After the training, the GAT-PPO-based DRL model outputs an optimization configuration strategy online according to the external power grid strength and the hybrid new energy station topology information.

2. The method for optimizing the configuration of hybrid new energy station generator sets according to claim 1, characterized in that: The optimization configuration strategy includes the capacity ratio and location distribution of the GFM generator set and the GFL generator set in the hybrid new energy station, and the control type of the energy storage device.

3. The method for optimizing the configuration of hybrid new energy station generator sets according to claim 2, characterized in that: The step 2 specifically includes: The state quantity is obtained according to the topological data and state data of the current hybrid new energy station; generating an internal configuration action according to the state quantity and acting on the hybrid new energy station, and evaluating the internal configuration action according to a change in the state of the hybrid new energy station; A reward function is determined according to the relationship between the state quantity and the preset constraint condition.

4. The method for optimizing configuration of hybrid new energy station generator sets according to claim 3 is characterized in that: The expression of the state quantity is: w1(X)=min{ξ L1 ,x L2 ,...,x Lk }-0.05 S t =[S topo ,S h ] In the formula, S t is the state quantity; S topo , S h They are the topological data and status data of the current hybrid new energy station respectively; w1, w2 and w3 are the low-frequency band small interference stability index, sub- / super-synchronous frequency band small interference stability index and operation cost index respectively; ξ Lk represents the damping ratio corresponding to the kth low-frequency closed-loop pole; ξ Sn and They represent the damping ratio and critical damping ratio of the closed-loop poles in the nth sub- / supersynchronous frequency bands respectively; N GFM is the number of GFM generator sets; X is the optimization variable, which represents the configuration vector of the generator sets and energy storage devices inside the hybrid new energy station.

5. The method for optimizing configuration of hybrid new energy station generator sets according to claim 4, characterized in that: The expression of the reward function is: In the formula, r is the reward value.

6. The method for optimizing the configuration of hybrid new energy station generator sets according to claim 3, characterized in that: The DRL model based on GAT-PPO includes GAT, a policy network and a value network; the GAT is used to obtain a node feature set according to the state quantity of the current hybrid new energy station, the policy network is used to generate configuration strategies for GFM generator sets, GFL generator sets and energy storage devices, and the value network is used to evaluate the advantages and disadvantages of the configuration strategies for the GFM generator sets, GFL generator sets and energy storage devices.

7. The method for optimizing configuration of hybrid new energy station generator sets according to claim 6, characterized in that: The step 3 specifically includes: characterizing the node characteristics of the hybrid new energy station through GAT, updating the strategy network and the value network, and applying the trained GAT-PPO-based DRL model online.

8. The method for optimizing configuration of hybrid new energy station generator sets according to claim 7, characterized in that: The updating strategy network comprises: inputting a node feature set represented by GAT into the strategy network to obtain the distribution of new strategies and old strategies; calculating the probability of selecting each action under the new strategy and the old strategy respectively according to the distribution of the new strategy and the old strategy; dividing the probability under the new strategy by the probability under the old strategy to obtain a probability ratio; calculating the objective function value of the DRL model based on GAT-PPO using the advantage function and the probability ratio; using the negative value corresponding to the objective function value as the loss function of the strategy network; and updating the parameters of the strategy network by back propagation using the loss function until a new strategy that meets the cutting requirements is obtained.

9. The method for optimizing configuration of hybrid new energy station generator sets according to claim 7, characterized in that: The updating of the value network includes: inputting the node feature set represented by GAT into the value network, calculating the value function in the current state through the forward propagation mechanism, and calculating the loss function to perform gradient update on the value network.

10. A hybrid new energy station generator set optimization configuration device, characterized in that: include: An establishing unit is used to establish an optimal configuration model of a hybrid new energy station generator set, wherein the optimal configuration model uses the configuration of the generator set and the energy storage device as an optimization variable, with the goal of improving the small disturbance stability margin and reducing the configuration number of the GFM generator set; A fusion unit is used to fuse GAT with the PPO-based DRL framework to obtain a GAT-PPO-based DRL model; A training output unit is used to use the optimization configuration model to perform offline training on the GAT-PPO-based DRL model, and the GAT-PPO-based DRL model after training outputs the optimization configuration strategy online according to the external power grid strength and the hybrid new energy station topology information.

Images