Simulation method and device for newly added grid connection of photovoltaic energy storage system and medium

Abstract

The invention discloses a simulation method and device for newly added grid connection of a photovoltaic energy storage system and a medium, and relates to the technical field of electric energy storage systems, and the method comprises the steps: collecting the state data of each target data source dynamically changing with time in a target photovoltaic power station, and carrying out the classification based on each electricity price grade and an electricity limiting scene; extracting time sequence characteristics of state data of the target photovoltaic power station after newly added energy storage grid connection through a base learner; based on the correlation between the target time sequence characteristics and each electricity price level and the correlation between the target time sequence characteristics and each electricity limit scene, carrying out weight distribution dynamic adjustment on the target time sequence characteristics in each scene through an attention mechanism; taking the time sequence characteristics after the weight distribution is dynamically adjusted as input, performing operation simulation of the photovoltaic energy storage system in each electricity price level and an electricity limiting scene through a meta learning device, and outputting target scheduling parameters; and evaluating the simulation precision of the meta-learner based on a preset loss function, and outputting the meta-learner. The stability and reliability of power grid operation are ensured through accurate simulation.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical energy storage systems, and particularly to a simulation method, device and medium for newly connecting to the grid in a photovoltaic energy storage system. Background Art

[0002] With the increasing global demand for clean energy, photovoltaic power generation and energy storage technologies are playing an increasingly important role in the energy transition. However, photovoltaic power generation has significant volatility and intermittency characteristics, which pose challenges to the stable operation of the power grid. The electricity generated by the photovoltaic system not only depends on weather conditions but is also affected by day-night changes, resulting in unstable output power. This instability imposes an additional burden on the power grid, especially in the case of high-proportion renewable energy access to the grid, which may cause problems such as voltage fluctuations and frequency deviations, affecting the safety and reliability of the power system.

[0003] At the same time, as a key link in balancing power supply and demand, the scheduling strategy of the energy storage system directly affects the economy of the entire project. The energy storage system can store energy when power is in excess and release energy when power is in short supply, thereby smoothing the volatility of photovoltaic power generation. However, how to optimize the scheduling strategy of the energy storage system under different electricity price levels and power curtailment scenarios to maximize economic benefits is a complex and urgent problem to be solved. Traditional simulation methods and scheduling strategies often cannot accurately predict the actual power generation situation of the photovoltaic power station and the optimal charge-discharge strategy of the energy storage system, resulting in the inability to effectively guarantee the robustness and stability of the project investment return rate.

[0004] In addition, due to the complex and changeable operating environment of the photovoltaic energy storage system, involving the interaction of various factors, such as the spatio-temporal distribution of light resources, the technical parameters of energy storage devices, the changes in market electricity prices, etc., these all increase the difficulty of model construction and optimization. Summary of the Invention

[0005] In order to optimize the scheduling strategy of the photovoltaic energy storage system and ensure the robustness and stability of the project investment return rate, the present invention proposes a simulation method for newly connecting to the grid in a photovoltaic energy storage system, including the steps of: S1: Collect the state data that changes dynamically over time from each target data source in the target photovoltaic power station and classify it based on each electricity price level and power curtailment scenario; S2: Extract the time series features of the state data of the target photovoltaic power station after newly connecting the energy storage to the grid through a base learner; S3: Based on the correlation between the target time series features and each electricity price level and power curtailment scenario, dynamically adjust the weight distribution of the target time series features in each scenario through an attention mechanism; S4: Using the target time series features after dynamically adjusting the weight allocation as the input, perform operation simulations for each electricity price level and power curtailment scenario of the photovoltaic energy storage system through the meta-learner, and output the target scheduling parameters; S5: Evaluate the simulation accuracy of the meta-learner based on a preset loss function. Before the simulation accuracy is higher than the preset threshold, adjust the parameters of the meta-learner and return to step S4. After the simulation accuracy is higher than the preset threshold, output the meta-learner with adjusted parameters.

[0006] Furthermore, in step S1, the state data to be collected also includes energy storage capacity, installed capacity of the photovoltaic power station, reserve capacity, and light resources that change dynamically over time.

[0007] Furthermore, in step S2, the base learner is based on a recurrent neural network to extract the corresponding time series features of the state data of the target photovoltaic power station after the new energy storage is connected to the grid.

[0008] Furthermore, in step S3, dynamic weight allocation is performed through the following formula to obtain the adjusted target time series features: In the formula, is the weight of the current target time series feature at time t in the current scenario, is the attention score calculated through linear transformation from , is the adjusted target time series feature, is the total duration, is a constant with values from 1 to , is the current target time series feature at time t in the current scenario, is the attention score at time

[0009] Furthermore, in step S4, the target scheduling parameters include the actual utilization hours after grid connection and the grid-connected power generation curve .

[0010] Furthermore, the actual utilization hours after grid connection and the grid-connected power generation curve are obtained through the following formula: In the formula, is the adjusted target time series feature, is the meta-learner that takes as the optimization target and adjusts the actual utilization hours according to real-time gradient information to finally affect the formation of a closed-loop optimization of the grid-connected power generation curve.

[0011] Further, in the step S5, the preset loss function adopts the mean square error, which is expressed as the following formula: In the formula, is the error loss, is the total duration, is the true target time series feature at time t under actual operation, is the simulated target time series feature at time t obtained by the meta-learner simulation.

[0012] Further, in the step S5, the gradient descent method is used to adjust the parameters of the meta-learner.

[0013] The present invention further includes a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the simulation method for newly adding grid connection to a photovoltaic energy storage system are implemented.

[0014] It further includes a device for processing data, including: a memory, on which a computer program is stored; a processor, configured to execute the computer program in the memory to implement the steps of the simulation method for newly adding grid connection to a photovoltaic energy storage system.

[0015] Compared with the prior art, the present invention has at least the following beneficial effects: (1) The simulation method for newly adding grid connection to a photovoltaic energy storage system proposed by the present invention, through this multi-level model structure, especially the application of the attention mechanism, the model can focus on the most relevant state features and suppress irrelevant information, thereby improving the prediction accuracy. This enables the system to simulate the photovoltaic power generation and energy storage scheduling process in real time and accurately under different electricity price levels and power curtailment scenarios, ensuring the stability and reliability of the power grid operation; (2) By deeply analyzing and processing these multi-source parameters, the model can better understand and simulate the actual operation situation, optimize the scheduling strategy of the photovoltaic energy storage system, which not only improves the economic benefits of the project, but also enhances the anti-risk ability in the face of various uncertain factors, ensuring the robustness and reliability of the project investment return rate; (3) By smoothing the volatility of photovoltaic power generation, reducing the dependence on traditional energy, and increasing the proportion of renewable energy in the energy structure, it helps to promote the global transformation to a low-carbon economy; (4) The model based on deep learning has good generalization ability and can maintain a high prediction accuracy when facing unknown situations. By continuously verifying and optimizing the model parameters, the model can be continuously improved to adapt to the needs under different environments and conditions, ensuring its long-term effectiveness and reliability. Brief Description of the Drawings

[0016] Figure 1 It is a step diagram of a simulation method for newly adding grid connection to a photovoltaic energy storage system. Detailed Embodiment

[0017] The following are specific embodiments of the present invention and, in conjunction with the accompanying drawings, further describe the technical solutions of the present invention, but the present invention is not limited to these embodiments.

[0018] With the continuous growth of the global demand for clean energy, photovoltaic power generation and energy storage technologies play a crucial role in the energy transition. However, the volatility and intermittency characteristics of photovoltaic power generation, as well as the complexity of the energy storage system scheduling strategy, pose huge challenges to the stable operation of the power grid. To address these challenges and optimize the operation efficiency of the photovoltaic energy storage system, the present invention proposes a simulation method for newly adding grid connection to a photovoltaic energy storage system.

[0019] This embodiment will elaborate in detail on the specific implementation steps and technical key points of this method, aiming to provide a complete solution for researchers, engineers, and operators in related fields to achieve high-precision simulation and optimized scheduling of the photovoltaic energy storage system. As Figure 1 shown, the following are the main technical steps of this method: S1: Collect the state data that changes dynamically with time from each target data source in the target photovoltaic power station and classify it based on each electricity price level and power curtailment scenario; S2: Extract the time series features of the state data of the target photovoltaic power station after newly adding energy storage grid connection through a base learner; S3: Based on the correlation between the target time series features and each electricity price level and power curtailment scenario, dynamically adjust the weight allocation for the target time series features in each scenario through an attention mechanism; S4: Use the time series features with dynamically adjusted weight allocation as input, and perform operation simulation of the photovoltaic energy storage system under each electricity price level and power curtailment scenario through a meta-learner, and output the target scheduling parameters; S5: Evaluate the simulation accuracy of the meta-learner based on a preset loss function, and adjust the parameters of the meta-learner and return to step S4 before the simulation accuracy is higher than the preset threshold. After the simulation accuracy is higher than the preset threshold, output the meta-learner with adjusted parameters.

[0020] Specifically, in the photovoltaic energy storage virtual precise simulation method based on an attention mechanism deep learning model, data collection and classification are the basic links of the entire process. First, obtain the state data that changes dynamically with time from each data source in the target photovoltaic power station through an efficient data collection system. These data include light resources, energy storage capacity, installed capacity of the photovoltaic power station, and reserve capacity. In addition, it is also necessary to collect time series data that changes dynamically with time , where represents the state data of the multi-objective data source at time t, and T is the total duration.

[0021] To ensure the integrity and accuracy of the data, the data acquisition system needs to be highly accurate and real-time. Specifically, light resource data can be obtained through weather stations or satellite remote sensing technology, the energy storage capacity depends on the real-time monitoring system of energy storage devices, and the installed and standby capacity information of photovoltaic power plants can be extracted from the power plant management system. After organizing all these data, a comprehensive and detailed historical database is formed, laying a solid foundation for subsequent modeling.

[0022] Next, the collected data is classified based on each electricity price level and power rationing scenario. The dispatching strategies vary significantly under different electricity price levels and power rationing scenarios, so the data must be carefully classified. For example, during high electricity price periods, the energy storage system should give priority to releasing electricity to maximize economic benefits; while during low electricity price periods, it should try to store the excess photovoltaic power generation. At the same time, the power rationing scenario also needs special consideration. When power rationing occurs in the power grid, the energy storage system can act as a buffer to avoid power waste and maintain the stable operation of the system.

[0023] This refined data classification not only improves the quality of the model input data, but also enhances the accuracy and reliability of the model prediction. By classifying the data according to the electricity price level and power rationing scenario, the model can better understand the operation modes in various complex situations and generate more accurate dispatching plans accordingly. This helps to improve the investment return rate of the project and enhance the anti-risk ability in the face of market fluctuations and policy changes.

[0024] In addition, data classification also helps to optimize the dispatching strategy of the energy storage system. By deeply analyzing the data under different electricity price levels and power rationing scenarios, operators can formulate the best charge and discharge strategies, thereby reducing operating costs and improving economic benefits. For example, releasing the energy storage power during peak electricity price periods and storing the excess power generation during low electricity price periods can not only balance the power grid load but also achieve higher returns.

[0025] After completing data acquisition and classification, the next step is to capture the time series characteristics of the state data of the photovoltaic power plant after adding energy storage through feature extraction. This process uses the stacked ensemble (Stacking) two-layer model structure in deep ensemble learning, where the first-layer model is responsible for extracting valuable information from a large amount of multi-source parameters through base learners.

[0026] To effectively process time series data, the present invention selects Long Short-Term Memory Network (LSTM), Gated Recurrent Unit (GRU) algorithm, Simple Recurrent Neural Network (Simple RNN), and Bidirectional Recurrent Neural Network (Bidirectional RNN, Bi-RNN), and uses them in combination according to specific usage scenarios as the base learners of the first-layer model. These algorithms all belong to variants of Recurrent Neural Network (RNN) and are particularly suitable for processing sequence data with time dependence. Specifically: By introducing mechanisms such as forget gate, input gate, and output gate, LSTM can effectively capture dependencies over long time spans. The forget gate determines which information needs to be discarded, the input gate determines which new information needs to be stored, and the output gate controls the final output state. The state update equations are as follows: Forget gate:

[0027] Input gate:

[0028] Candidate state value:

[0029] State update:

[0030] Output gate: Hidden state update:

[0031] Among them, is the output of the forget gate, which determines which information needs to be discarded, is the weight matrix of the forget gate, is the hidden state at time t-1, is the input data at time t, is the bias term of the forget gate, is the Sigmoid activation function, which compresses the value to the interval (0, 1); is the output of the input gate, which determines which new information needs to be stored, is the weight matrix of the input gate, is the bias term of the input gate; is the new candidate state value, is the weight matrix of the candidate state value, is the bias term of the candidate state value, is the hyperbolic tangent activation function, which compresses the value to the interval (-1, 1); is the cell state at the current time t, is the cell state at time t-1, is the element-wise multiplication (Hadamard product); is the output of the output gate, controlling the state of the final output. is the weight matrix of the output gate. is the bias term of the output gate; is the hidden state at time t.

[0032] GRU simplifies the structure of LSTM by combining the forget gate and the input gate into an update gate and introducing a reset gate to control the degree of retention of historical information. Compared with LSTM, GRU usually has fewer parameters and higher computational efficiency. The state update equations are as follows: Update gate: Reset gate: Candidate activation: Hidden state update: Among them, is the output of the update gate, determining how much historical information to retain. is the weight matrix of the update gate. is the bias term of the update gate; is the output of the reset gate, determining how much historical information to discard. is the weight matrix of the reset gate. is the bias term of the reset gate; is the new candidate hidden state. is the weight matrix of the candidate hidden state. is the bias term of the candidate hidden state; is the hidden state at time t.

[0033] During the feature extraction process, these base learners can mine valuable information from time series data. For example, LSTM or GRU can capture the state change trend after the addition of new energy storage in a photovoltaic power station, identify the change patterns of power demand in different time periods, and the optimal charging and discharging times of energy storage devices. This ability enables the model to generate effective base learners and provide reliable inputs for the upper-level model.

[0034] In this way, feature extraction not only improves the model's ability to understand complex time-dependent relationships but also enhances the overall prediction performance of the model. Specifically, LSTM and GRU can better handle the volatility and intermittency problems in the operation of a photovoltaic power station, thereby generating more accurate state features. This is crucial for subsequent model calibration and optimization.

[0035] In addition, high-quality feature extraction brings significant beneficial effects. First, it improves the robustness and generalization ability of the model, ensuring that the model can still maintain a high prediction accuracy when facing unknown situations. Second, by capturing key time-series features, the model can generate more accurate scheduling plans, optimize the charge and discharge strategies of the energy storage system, reduce operating costs and improve economic benefits. Finally, the feature extraction process provides solid technical support for the entire simulation method, ensuring its reliability and effectiveness in practical applications.

[0036] After completing the feature extraction, the next step is to dynamically adjust the weight allocation of the target time-series features in each scenario through the attention mechanism based on the correlation between the target time-series features and each electricity price level and power curtailment scenario. This process (step S3) aims to improve the prediction accuracy of the model and better understand complex time-dependent relationships.

[0037] Here, the attention mechanism allows the model to focus on the most relevant state features while suppressing irrelevant information, thereby improving the accuracy of prediction. Specifically, the attention mechanism is introduced in the second-layer model, which enables the model to better handle complex situations under different electricity price levels and power curtailment scenarios by dynamically adjusting the weight allocation.

[0038] In the attention mechanism, it is first necessary to calculate the attention scores for each moment. These scores represent the importance of each moment and are used for subsequent weighted summation operations. The attention score is obtained through the following formula: In the formula, is the attention score at time t, is the weight matrix for calculating the attention score, is the current target time-series feature at time t in the current scenario (generated by the first-layer model), is the bias term for calculating the attention score, is the activation function, and usually the Softmax function is used to ensure that the sum of the weights at all moments is 1.

[0039] To ensure that the sum of the weights at each moment is 1, the Softmax function is usually used to normalize the attention scores:

[0040] Among them, is the weight of the current target time-series feature at time t in the current scenario, is the attention score at time i, and T is the total duration.

[0041] Finally, a new feature vector is obtained by performing a weighted sum on the original feature vector:

[0042] In the formula, is the weighted feature vector, that is, the time series feature after dynamic adjustment, and serves as the input of the second-layer model.

[0043] By introducing the attention mechanism and dynamically adjusting the weight allocation, the model can better capture the important information at different time steps. Especially in the scenarios of electricity price fluctuations and power rationing, this ability is particularly important. In this way, the model can dynamically adjust the charge and discharge strategies of the energy storage system according to the changes in real-time electricity price and power demand, so as to achieve higher economic benefits. For example, releasing the stored electricity of the energy storage system during the peak electricity price period and storing the excess generated electricity during the low electricity price period can not only balance the grid load but also achieve higher returns.

[0044] After completing the feature extraction and the weight allocation of the attention mechanism, the next step is to perform the operation simulation of the photovoltaic energy storage system under each electricity price level and power rationing scenario through the meta-learner and output the target scheduling parameters. This process aims to generate specific scheduling strategies by using the optimized feature vectors to maximize the economic benefits of the system and ensure the stable operation of the grid.

[0045] During the operation simulation process, the meta-learner receives the feature vectors extracted by the first-layer model and weighted by the attention mechanism as the input. These feature vectors contain the important information at each time step and can help the meta-learner better understand the complex time-dependent relationships. The meta-learner can adopt various machine learning algorithms, and common choices include linear regression, support vector machine (SVM), random forest (Random Forest), etc. Among them, the general expression of the meta-learner is as follows: Wherein, is the time series feature after dynamic adjustment, is as the optimization objective and adjusts the control parameters according to the real-time gradient information to finally affect the formation of the grid-connected power generation curve to form a closed-loop optimization meta-learner.

[0046] In order to better handle complex non-linear relationships, more advanced models such as Gradient Boosting Decision Tree (GBDT) or Neural Network are selected in the present invention. Based on these input feature vectors, the meta-learner simulates the operation of the photovoltaic energy storage system under different electricity price levels and power rationing scenarios, and generates detailed scheduling parameters such as the actual utilization hours after grid connection and the grid-connected power generation curve .

[0047] Specifically, the actual utilization hours after grid connection represent the actual effective power generation time length of the PV power station in a day. This parameter can be calculated by the following formula: where, is the grid-connected power generation power at time, is the maximum installed capacity of the PV power station. The grid-connected power generation curve describes the power generation power change of the PV power station every hour in a day. This curve can be directly predicted by the meta-learner or obtained by fitting historical data. The specific formula is as follows: Here, is the weight matrix of the meta-learner, is the bias term of the meta-learner, and in this formula, is the prediction function of the meta-learner.

[0048] To ensure the prediction accuracy of the meta-learner, it needs to be strictly evaluated and continuously optimized. A preset loss function (such as the mean square error MSE) is used to evaluate the performance of the meta-learner. The loss function is defined as follows: where, is the error loss, is the true target time series feature at time t under actual operation, is the simulated target time series feature at time t obtained by the meta-learner simulation.

[0049] To minimize the loss function and improve the performance of the model, the parameters of the meta-learner need to be adjusted. Commonly used optimization algorithms include the Gradient Descent method, the Stochastic Gradient Descent (SGD) method, the Adam optimizer, etc. Among them, the Gradient Descent method is a commonly used optimization algorithm. By calculating the gradient of the loss function with respect to the model parameters, the parameters are gradually adjusted to minimize the loss function. The specific formula is as follows: where, is the model parameter, is the learning rate, is the gradient of the loss function with respect to the parameter.

[0050] To ensure that the model can still maintain a high prediction accuracy when facing unknown situations, it is necessary to regularly update the dataset and retrain the model. During the model retraining process, hyperparameters (such as learning rate, batch size, etc.) can also be tuned to further improve the performance of the model. Common tuning methods include Grid Search and Random Search.

[0051] The present invention also includes a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the simulation method for newly adding grid connection to a photovoltaic energy storage system are implemented.

[0052] It also includes a device for processing data, including: A memory, on which a computer program is stored; A processor, configured to execute the computer program in the memory to implement the steps of the simulation method for newly adding grid connection to a photovoltaic energy storage system.

[0053] In summary, the simulation method for newly adding grid connection to a photovoltaic energy storage system proposed by the present invention shows significant benefits in many aspects in the photovoltaic energy storage virtual precise simulation method. First, in the data collection and classification stage, key parameters such as light resources, energy storage capacity, installed capacity of photovoltaic power plants, and reserve capacity are obtained through an efficient data collection system, and refined classification is carried out according to electricity price levels and power rationing scenarios, ensuring the integrity and accuracy of the data. This not only improves the basic quality of subsequent analysis but also provides a solid guarantee for generating accurate scheduling plans.

[0054] During the feature extraction process, choosing long short-term memory network (LSTM) or gated recurrent unit (GRU) as the base learner can effectively capture complex dependencies in time series data, thereby improving the prediction accuracy. In addition, introducing variants such as bidirectional recurrent neural network (Bi-RNN) and deep recurrent neural network (Deep RNN) further enhances the robustness and generalization ability of the model, making it still perform stably when facing noise and outliers.

[0055] The application of the attention mechanism is a major highlight of this method. By dynamically adjusting the weight distribution, the model can focus on the most relevant state features, suppress irrelevant information, and significantly improve the accuracy and reliability of the prediction. Especially in different electricity price levels and power rationing scenarios, the attention mechanism helps to optimize the charge and discharge strategies of energy storage devices, maximizing the economic benefits of the project and ensuring the balance of the grid load.

[0056] The application of the meta-learner in running simulations further improves the performance of the system. The meta-learner receives the feature vectors weighted by the attention mechanism and generates high-precision scheduling parameters, such as the actual utilization hours after grid connection and the grid-connected power generation curve. These parameters not only improve the scheduling accuracy and response speed of the photovoltaic energy storage system but also reduce the operating cost and increase the return on investment through a reasonable scheduling strategy.

[0057] In the model validation and optimization stage, loss functions such as mean squared error (MSE) are used to evaluate the model performance, and the model parameters are continuously adjusted through the gradient descent method or the Adam optimizer to ensure the minimum difference between the predicted value and the actual value. The dataset is updated regularly and the model is retrained, enabling the model to maintain a high prediction accuracy when facing unknown situations, enhancing the generalization ability and stability of the model.

[0058] Through comprehensive data processing, high-quality feature extraction, the application of the attention mechanism, efficient meta-learner running simulations, and model optimization, the present invention significantly improves the prediction accuracy and robustness of the photovoltaic energy storage system. Its optimized scheduling strategy not only reduces the operating cost and improves the economic benefits but also enhances the stability of the power grid, promoting the effective utilization and sustainable development of clean energy. The combined effect of this series of technologies provides strong support for the optimized scheduling of the photovoltaic energy storage system and has broad practical application prospects.

[0059] It should be noted that all directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention are only used to explain the relative position relationship and movement conditions between components in a specific posture (as shown in the drawings). If the specific posture changes, the directional indication will also change accordingly.

[0060] In addition, in the present invention, descriptions such as "first", "second", "one", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, "a plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0061] In the present invention, unless otherwise clearly defined and limited, terms such as "connection" and "fixation" should be understood in a broad sense. For example, "fixation" can be a fixed connection, a detachable connection, or integrated; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the internal connection of two components or the interaction relationship between two components, unless otherwise clearly limited. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0062] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

Claims

1. A simulation method for newly adding grid connection to a photovoltaic energy storage system, characterized in that, Including the steps: S1: Collect the state data that changes dynamically over time from each target data source in the target photovoltaic power station and classify it based on each electricity price level and power curtailment scenario; S2: Extract each time series feature of the state data of the target photovoltaic power station after the new energy storage is connected to the grid through the base learner; S3: Based on the correlation between the target time series features and each electricity price level and power curtailment scenario, dynamically adjust the weight allocation for the target time series features in each scenario through the attention mechanism; S4: Using the target time series features with dynamically adjusted weight allocation as input, perform operation simulations of the photovoltaic energy storage system under each electricity price level and power curtailment scenario through the meta-learner, and output the target scheduling parameters; S5: Evaluate the simulation accuracy of the meta-learner based on the preset loss function, and adjust the parameters of the meta-learner and return to step S4 before the simulation accuracy is higher than the preset threshold. After the simulation accuracy is higher than the preset threshold, output the meta-learner with adjusted parameters.

2. The simulation method for newly adding grid connection to a photovoltaic energy storage system according to claim 1, wherein, In the S1 step, the state data to be collected also includes energy storage capacity, installed capacity of the photovoltaic power station, reserve capacity, and light resources that change dynamically over time.

3. The simulation method for newly adding grid connection to a photovoltaic energy storage system according to claim 1, wherein, In the S2 step, the base learner is based on a recurrent neural network to extract each time series feature corresponding to the state data of the target photovoltaic power station after the new energy storage is connected to the grid.

4. The simulation method for newly adding grid connection to a photovoltaic energy storage system according to claim 1, wherein, In the S3 step, the following formula is used to dynamically adjust the weight allocation and obtain the adjusted target time series features: Wherein, is the weight of the current target temporal feature at time t in the current scenario, is the attention score calculated by linear transformation from , is the adjusted target temporal feature, is the total duration, is a constant taking values from 1 to , is the current target temporal feature at time t in the current scenario, is the attention score at time 5. The simulation method for newly adding grid connection in a photovoltaic energy storage system according to claim 1, wherein, In the step S4, the target scheduling parameter includes the actual utilization hours after grid connection and the grid-connected power generation curve .

6. The simulation method for newly adding grid connection to a photovoltaic energy storage system according to claim 5, characterized in that, The actual utilization hours after grid connection and the grid-connected power generation curve are obtained through the following formula: In the formula, is the adjusted target time series feature, is the optimization target, and adjusts the actual utilization hours according to the real-time gradient information to finally affect the grid-connected power generation curve to form a closed-loop optimized meta-learner.

7. The simulation method for newly adding grid connection to a photovoltaic energy storage system according to claim 1, characterized in that In the S5 step, the preset loss function uses the mean square error, which is expressed as the following formula: In the formula, is the error loss, is the total duration, is the true target time series feature at time t under actual operation, is the simulated target time series feature at time t obtained by simulation of the meta-learner.

8. The simulation method for newly adding grid connection to a photovoltaic energy storage system according to claim 1, wherein In the S5 step, the gradient descent method is used to adjust the parameters of the meta-learner.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the simulation method for the new grid connection of the photovoltaic energy storage system according to any one of claims 1 to 8.

10. A device for processing data, characterized in that, Including: A memory on which a computer program is stored; A processor for executing the computer program in the memory to implement the steps of the simulation method for the new grid connection of the photovoltaic energy storage system according to any one of claims 1 to 8.

Images