Day-ahead and intra-day economic dispatching method for wind and light storage system

Abstract

The invention discloses a day-ahead and intra-day economic dispatching method for a wind and light storage system, which relates to the technical field of wind and light storage and comprises the following steps of: establishing a dynamic aging model based on historical operation data of a battery, fitting actual attenuation characteristics of the battery through machine learning, acquiring the health state of the battery in different operation states and the change rule of the battery along with time, and calculating the economic dispatching result. Therefore, the basis of the dynamic adjustment strategy is formed. Energy storage charging and discharging are optimized through health factors, the system monitors parameters such as SOH and SOC in real time, tasks are dynamically adjusted, battery aging is reduced, stability is improved, the SOC strategy is optimized in combination with deep reinforcement learning, the service life is prolonged, and the maintenance cost is reduced. Rolling optimization scheduling and experience playback are adopted, a system can dynamically adjust a strategy, the wind and light absorption rate is improved, power grid balance is optimized, thermal power dependence is reduced, scheduling cost is reduced, extreme weather adaptability is enhanced, and green and low-carbon development of electric power is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of wind-solar-storage, and specifically to a day-ahead and intra-day economic dispatch method for a wind-solar-storage system. Background Art

[0002] The day-ahead and intra-day economic dispatch of a wind-solar-storage system refers to a combined optimization dispatch method for a power system operation, based on wind energy, photovoltaic, and energy storage systems. Due to the significant uncertainty of wind and solar power generation, traditional day-ahead dispatch plans often struggle to accurately match actual power generation and load demands. Therefore, adjustments and optimizations are required within the day. Specifically, first, an optimization dispatch plan is formulated based on meteorological forecasts, grid load forecasts, and energy storage status on the day-ahead (i.e., the previous day). However, due to possible deviations between the predicted values and the actual values, during the operation of the next day, the actual situations of wind and solar power generation and load are monitored in real time, and the dispatch strategy is continuously adjusted and optimized. This method relies on the idea of rolling optimization, that is, based on the latest real-time data, the day-ahead dispatch results are corrected, so that the final dispatch plan is closer to the actual operation state, thereby reducing the control error caused by deviations and improving the economy and stability of the system operation.

[0003] The existing technologies have the following deficiencies: The cumulative error of the dynamic aging effect of the energy storage system in day-ahead and intra-day dispatch is a problem that is easily overlooked but may bring serious consequences. Existing wind-solar-storage dispatch models usually assume that the charge-discharge efficiency of the energy storage battery is constant and dispatch according to a fixed charge-discharge curve, while ignoring the non-linear aging effect of the battery caused by factors such as temperature changes and current surges during high-frequency charge-discharge processes. This aging not only reduces the actual available capacity of the battery but may also cause the gradual accumulation of energy balance errors in the dispatch optimization calculation. Especially after long-term operation, there may be a deviation between the available capacity of the battery and the dispatch plan, resulting in the energy storage being unable to release power as planned at critical moments and affecting the system stability. More seriously, in extreme cases, over-discharge may trigger internal short circuits or thermal runaway in the battery, increasing the risk of equipment damage and even safety accidents.

[0004] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The objective of the present invention is to provide a day-ahead and intra-day economic dispatch method for a wind-solar-storage system. Through a health factor-driven energy storage charge-discharge optimization method, the system can monitor parameters such as SOH, SOC, temperature, and current in real time, dynamically adjust the charge-discharge tasks, reduce the unbalanced aging of the battery, improve the system stability, and combine with deep reinforcement learning to optimize the SOC management strategy, extend the battery life, and reduce the maintenance cost. In addition, by adopting a rolling optimization dispatch and experience replay mechanism, the system can dynamically adjust the dispatch strategy based on factors such as real-time load demand and wind-solar prediction error, improve the wind-solar accommodation rate, optimize the grid energy balance, reduce the dependence on thermal power, lower the dispatch cost, enhance the adaptability to extreme weather, and promote the green and low-carbon development of the power system, so as to solve the problems in the above-mentioned background technology.

[0006] To achieve the above objective, the present invention provides the following technical solution: A day-ahead and intra-day economic dispatch method for a wind-solar-storage system, comprising the following steps:

[0007] Based on the historical operation data of the battery, establish a dynamic aging model, fit the actual attenuation characteristics of the battery through machine learning, and obtain the battery health state under different operating states and its variation law over time, so as to form the basis for the dynamic adjustment strategy;

[0008] In the day-ahead dispatch stage, combined with the wind-solar power generation prediction data, the grid load prediction data, and the health state of the energy storage system, use a multi-objective optimization method to calculate the optimal charge-discharge plan of the energy storage, ensure that while maximizing the wind-solar accommodation, reduce the impact of high-rate charge-discharge on the battery life, and provide a benchmark plan for the intra-day dispatch;

[0009] During the intra-day dispatch process, use high-precision sensors and state estimation methods to obtain the actual output of the wind-solar power generation, the real-time demand of the grid load, and the actual available capacity of the energy storage system, compare with the day-ahead dispatch plan, identify the existing deviations, and determine whether it is necessary to adjust the energy storage charge-discharge plan;

[0010] Based on the battery health state parameters calculated by the energy storage dynamic aging model, adjust the dispatch strategy in actual operation, correct the charge-discharge plan by introducing a health factor, prevent the battery from overcharging and discharging, and at the same time optimize the load distribution of different battery packs, so that the overall attenuation of the energy storage system is balanced, extend the service life and reduce the dispatch deviation caused by uncertainty;

[0011] Based on the real-time monitoring data and the aging compensation control mechanism, adopt a rolling horizon optimization method to dynamically correct the day-ahead dispatch plan, recalculate the optimal energy storage charge-discharge strategy for a future period at regular time intervals, make it adapt to the dynamic changes of the wind-solar output and the load demand, and reduce the energy balance problem caused by the prediction error;

[0012] Combining historical operation data and real-time scheduling feedback, the reinforcement learning algorithm is used to dynamically adjust the scheduling parameters, enabling the energy storage system to adapt to the long-term changes in the battery aging state, improving the energy storage utilization efficiency, and optimizing the scheduling strategies under different operation scenarios, thereby reducing the cumulative error during the long-term operation process and making the energy storage system intelligent and economical.

[0013] Preferably, based on the historical operation data of the battery, a dynamic aging model is established. By using machine learning to fit the actual attenuation characteristics of the battery, the state of health of the battery under different operation states and its variation law over time are obtained to form the basis for the dynamic adjustment strategy. The specific steps are as follows:

[0014] Construct a battery operation database, collect and clean the key data of charge and discharge current, voltage, and temperature, remove outliers and perform standardization processing to ensure data integrity and consistency;

[0015] Screen the key feature variables affecting battery aging, and use correlation analysis to optimize variable selection to improve the accuracy of modeling;

[0016] Use machine learning algorithms to train the SOH prediction model, optimize the hyperparameters and evaluate the prediction accuracy to ensure that the model accurately reflects the changes in the battery state of health;

[0017] Integrate the SOH prediction model into the energy storage management system, continuously optimize it using a rolling update mechanism, and adjust the charge and discharge strategies in combination with optimization algorithms to extend the battery life and improve the system stability.

[0018] Preferably, in the day-ahead scheduling stage, a multi-objective optimization method is used to calculate the optimal charge and discharge plan of the energy storage, ensuring that while maximizing the absorption of wind and light, the impact of high-rate charge and discharge on the battery life is reduced, and providing a benchmark plan for the intra-day scheduling. The specific steps are as follows:

[0019] Integrate the wind and light power generation prediction, power grid load prediction, and energy storage state of health data, and use uncertainty modeling methods to reduce the impact of prediction errors on scheduling decisions;

[0020] Construct a multi-objective optimization model that takes into account wind and light absorption, battery aging, and economic costs, and use the Pareto optimization algorithm to dynamically balance different optimization objectives;

[0021] Based on mixed-integer linear programming, optimize the charge and discharge time, power, and path of the energy storage, and use the MPC rolling optimization method to improve the scheduling adaptability;

[0022] Formulate a benchmark scheduling plan including dynamic SOC window adjustment and intelligent balanced charge and discharge strategies, and continuously correct and optimize it in the intra-day scheduling stage in combination with the rolling optimization algorithm.

[0023] Preferably, during the intraday scheduling process, high-precision sensors and state estimation methods are used to obtain the actual output of wind and photovoltaic power generation, the real-time demand of the power grid load, and the actual available capacity of the energy storage system, and compare them with the day-ahead scheduling plan to identify existing deviations, and the specific steps to determine whether to adjust the energy storage charge and discharge plan are as follows:

[0024] Use high-precision sensors and state estimation methods to obtain wind and photovoltaic output, load demand, and the state of the energy storage system, and improve the measurement accuracy through data filtering and machine learning methods;

[0025] Compare the actual operating status with the day-ahead scheduling plan, calculate the key deviation indicators of wind and photovoltaic output, load demand, and the energy storage system, and evaluate the impact of the deviation on system operation;

[0026] Based on the MPC rolling optimization method, recalculate the future charge and discharge strategy, optimize the SOC window, and improve the economy and the life of the energy storage system by combining grid scheduling and market electricity prices;

[0027] Execute the optimized scheduling plan, feedback control the operating status of the energy storage in real time, and use reinforcement learning technology to optimize the long-term scheduling strategy to improve the intelligent level of the system.

[0028] Preferably, based on the battery health state parameters calculated by the energy storage dynamic aging model, adjust the scheduling strategy in actual operation, correct the charge and discharge plan by introducing a health factor to prevent overcharging and over-discharging of the battery, and at the same time optimize the load distribution of different battery packs to make the overall attenuation of the energy storage system balanced, extend the service life and reduce the scheduling deviation caused by uncertainty. The specific steps are as follows:

[0029] Based on the key factors of battery SOH, SOC, and cycle times, calculate the HF value through the weighted comprehensive evaluation method to measure the battery health state and provide a basis for scheduling optimization;

[0030] Optimize the charge and discharge task allocation of the battery pack according to the HF value, adjust the SOC working window to ensure that the batteries in good health state undertake more loads, and achieve overall aging balance of the energy storage system;

[0031] Adopt a day-ahead to intraday rolling optimization mechanism, dynamically calculate the future scheduling strategy at the same interval, and combine with the adaptive control algorithm to make the energy storage scheduling more flexible and economical;

[0032] Establish a closed-loop monitoring mechanism, use reinforcement learning methods to optimize the long-term scheduling strategy, so that the energy storage system can be continuously optimized under different operating environments, extend the battery life and improve the overall operating efficiency.

[0033] Preferably, based on real-time monitoring data and an aging compensation control mechanism, a rolling horizon optimization method is used to dynamically correct the day-ahead scheduling plan. The optimal energy storage charge and discharge strategy for a future period is recalculated at regular time intervals to adapt to the dynamic changes in wind and solar power output and load demand, reducing the energy balance problems caused by prediction errors. The specific steps are as follows:

[0034] At the current moment, based on wind and solar power generation, grid load, and the state of the energy storage system, a rolling optimization objective function is established to minimize the aging loss of the energy storage system while ensuring system energy balance. Over the prediction horizon T p , the wind and solar power generation and load demand at future moments are predicted, and combined with the health factor of the battery, the available power of the energy storage system and the optimal scheduling strategy are calculated. The optimization objective is defined as follows:

[0035]

[0036] , where J is the optimization objective function, measuring the comprehensive cost of the energy storage scheduling strategy, t k is the current moment, T p is the prediction horizon, t is the optimization moment, representing each discrete time point in the rolling optimization process, P b (t) is the charge and discharge power of the energy storage system at moment t, SOC(t) is the state of charge of the energy storage system at moment t, SOC ref is the target SOC value, HF(t) is the battery health factor, used to measure the health state of the battery, λ 1 is the energy storage power regulation weight, controlling the power fluctuation of the energy storage system, λ 2 is the SOC deviation penalty weight, λ 3 is the battery aging impact weight;

[0037] Based on the optimization objective function J, constraint conditions are constructed to ensure that the optimized scheduling strategy is feasible and meets the system operation requirements. Over each prediction horizon T p , MPC adopts a rolling optimization strategy, that is, at the current moment t k the optimal energy storage charge and discharge strategy is calculated, but only the charge and discharge instructions within the current optimization period are executed. Subsequently, the next optimization period is entered, the prediction data is updated, and the optimization problem is solved again. The energy balance equation is as follows:

[0038]

[0039] , where SOC(t + 1) is the state of charge at the next moment, η c is the charging efficiency, η d is the discharging efficiency, and are the optimal charging power and discharging power at time t, respectively, obtained by optimizing and solving the objective function J, and C b is the rated capacity of the battery, which determines the SOC change rate.

[0040] Preferably, by combining historical operation data and real-time scheduling feedback, the reinforcement learning algorithm is used to dynamically adjust the scheduling parameters, enabling the energy storage system to adapt to the long-term changes in the battery aging state, improving the energy storage utilization efficiency, and optimizing the scheduling strategies under different operation scenarios, thereby reducing the cumulative error during the long-term operation process and making the energy storage system intelligent and economical. The specific steps are as follows:

[0041] First, define the state, action, and reward functions of reinforcement learning, and construct the state transition equation to enable the model to learn the optimal scheduling strategy of the energy storage system under different operating environments;

[0042] State S t includes the current state of health of the battery SOH, the remaining charge SOC, the battery temperature T b , the predicted value of wind and solar power generation P wp , the grid load demand P load , and the market electricity price C m key information, defined as:

[0043] S t ={SOH t , SOC t , T b , P wp , P load , C m}

[0044] Action A t represents the scheduling decision of the energy storage system at the current moment;

[0045] A t ={P ch , P dis , ΔSOC}

[0046] , where P ch is the charging power, P dis is the discharging power, and ΔSOC is the SOC adjustment strategy;

[0047] The state transition equation is based on the battery dynamic aging model to calculate the SOH and temperature changes at the next moment:

[0048] SOH t+1 =SOH t -α·f(T b , P ch , P dis )

[0049] Among them, SOH t is the state of health of the battery at time t, SOH t+1 is the state of health of the battery at the next moment, α is the aging factor, f(T b , P ch , P dis ) represents the attenuation rate of the battery caused by temperature, charge and discharge power, T b is the battery temperature, P ch is the charging power, P dis is the discharging power;

[0050] The reward function is designed to optimize the scheduling strategy, and a composite reward function is designed, which is defined as:

[0051]

[0052] , where, R t is the reward function, P curtail is the curtailed wind and solar power, β 1 is the battery life optimization weight, β 2 is the wind and solar accommodation optimization weight, β 3 is the economic cost optimization weight, which measures the role of market electricity price and charge and discharge cost in the optimization goal.

[0053] Preferably, after determining the state S t , state A t and the reward function R t , deep reinforcement learning based on the Actor-Critic architecture is used for policy training, so that the energy storage system adapts to the battery aging state and optimizes the scheduling strategy under different scenarios;

[0054] The Actor network is responsible for generating the state A t from the state S t , and the formula is as follows:

[0055] A t =π θ (S t )+N t

[0056] , where, π θ (S t ) is the parameterized policy, N t is the exploration noise, which is used to balance exploration and exploitation;

[0057] The Critic network is used to evaluate whether the policy of the Actor network is optimal and calculate the Q value, and the calculation formula is as follows:

[0058] Q(S t , A t )=R t+γQ(S t+1 ,A t+1 )

[0059] where Q(S t ,A t ) is the state - action value function, i.e., the Q - value, and γ is the discount factor used to control the importance of future rewards. Q(S t+1 ,A t+1 ) is the Q - value estimate at the next time step t + 1;

[0060] Experience replay: To improve the training stability, the past state - action - reward - state transition data is stored in the experience replay pool and randomly sampled for training to avoid the gradient oscillation problem caused by data correlation;

[0061] Policy update: Gradient descent is used to optimize the Actor and Critic networks, and the policy parameters θ are updated to maximize the cumulative reward. The update formula is as follows:

[0062]

[0063] where θ is the training parameter of the Actor network, η is the learning rate, is the parameter update direction based on the policy gradient.

[0064] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0065] Through the energy storage charge - discharge optimization method driven by the health factor (HF), the present invention can effectively reduce the unbalanced aging phenomenon of energy storage batteries and improve the long - term operation stability of the entire energy storage system. In traditional scheduling methods, all battery packs often operate according to the same charge - discharge strategy and cannot be dynamically adjusted according to the health status of individual batteries, resulting in some batteries accelerating aging due to high - rate discharge, long - term high - SOC operation, or abnormal temperature, ultimately affecting the reliability of the entire system. In this solution, by real - time monitoring parameters such as SOH, SOC, temperature, and current, the battery health factor (HF) is calculated, and the charge - discharge task allocation is intelligently adjusted according to the HF value to ensure that batteries in better health states undertake greater power, while batteries with more serious aging reduce deep discharge and high - rate operation. This dynamic optimization strategy can extend the overall life of the energy storage system, reduce scheduling errors caused by the decline of individual batteries, and thus improve the system stability. In addition, combined with the deep reinforcement learning (DRL) technology, the system can continuously learn and optimize the SOC management strategy during long - term operation, such as dynamically adjusting the SOC window and optimizing the charge - discharge rate, etc., to minimize the battery aging rate to the greatest extent. This can not only reduce the maintenance and replacement costs of energy storage batteries but also improve the long - term return on investment of the energy storage system, making the wind - solar - energy - storage system have better economic feasibility and operation reliability.

[0066] The present invention combines rolling optimal scheduling (MPC) with an experience replay mechanism (Replay Buffer), enabling the system to dynamically adjust the charge and discharge strategies according to factors such as real-time load demand, market electricity price, and prediction errors of wind and light, thereby improving the absorption rate of wind and light and optimizing the power grid energy balance. Specifically, during the intraday scheduling stage, the system can perform rolling optimal calculations based on the latest load and wind and light output data every 15 - 30 minutes to ensure that the scheduling plan is always consistent with the actual operating state and reduce the scheduling deviation caused by prediction errors. In addition, through reinforcement learning technology, the system can continuously optimize the scheduling strategies under different meteorological conditions and load characteristics during long-term operation, improving the system's adaptability to extreme weather or sudden load fluctuations. For example, when encountering continuous low wind and low light weather, the system can utilize historical experience data to automatically switch to a more optimal energy storage management strategy, such as charging in advance, reducing the discharge power, and optimizing the charging timing, to ensure the stable operation of the power grid. Ultimately, this optimization scheme can improve the utilization rate of wind and light power generation while reducing the dependence on standby thermal power units, lowering the power grid scheduling cost, and increasing the penetration rate of new energy in the power system, promoting the power system to develop towards a greener and lower-carbon direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0068] Figure 1 It is a method flow chart of the day-ahead and intraday economic scheduling method for the wind-solar-storage system of the present invention. DETAILED IMPLEMENTATION MANNER

[0069] Now, the exemplary embodiments will be described more comprehensively with reference to the drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0070] The present invention provides a Figure 1 day-ahead and intraday economic scheduling method for a wind-solar-storage system, characterized by including the following steps:

[0071] Based on the historical operation data of the battery, a dynamic aging model including factors such as temperature influence, number of cycles, charge-discharge rate, etc. is established. By using machine learning to fit the actual attenuation characteristics of the battery, the State of Health (SOH) of the battery under different operating states and its variation law over time are obtained, so as to form the basis for a dynamic adjustment strategy;

[0072] The specific steps for establishing a dynamic aging model based on the historical operation data of the battery, including factors such as temperature influence, number of cycles, charge-discharge rate, etc., and using machine learning to fit the actual attenuation characteristics of the battery, obtaining the State of Health (SOH) of the battery under different operating states and its variation law over time, so as to form the basis for a dynamic adjustment strategy are as follows:

[0073] Construct a battery operation database, collect and clean key data such as charge-discharge current, voltage, and temperature, remove outliers and perform standardization processing to ensure data integrity and consistency;

[0074] First, construct a battery operation database and collect battery data during long-term operation, including key parameters such as charge-discharge current, voltage, temperature, number of cycles, charge-discharge rate, change in State of Charge (SOC), and State of Health (SOH). The data sources can be historical operation records, laboratory accelerated aging test data, or real-time monitoring data of large-scale energy storage systems. During the data collection process, outlier detection is required to eliminate invalid data caused by sensor errors, communication failures, etc., and interpolation, smoothing filtering, etc. are used to fill in missing data to ensure data integrity and reliability. At the same time, standardize the data from different sources to ensure that all data is within the same scale range, making subsequent modeling more accurate.

[0075] Screen key characteristic variables that affect battery aging, such as the growth rate of internal resistance and the capacity attenuation rate, and use correlation analysis to optimize variable selection to improve the accuracy of modeling;

[0076] In data preprocessing, characteristic variables related to battery aging are extracted, such as the growth rate of battery internal resistance, the attenuation rate of the maximum available capacity, the charge-discharge efficiency change curve, the impact of temperature change on the capacity retention rate, etc. Statistical analysis methods (such as correlation analysis, principal component analysis PCA) can be used for feature extraction to screen out the variables that have the greatest impact on SOH. In addition, through interpretive AI techniques such as grey relational analysis or SHAP (Shapley Additive Explanations), the contribution degree of each variable to the change of SOH can be analyzed, redundant variables can be removed, and only the features that have a significant impact on the battery health state are retained. For example, some batteries may be particularly sensitive to temperature changes, while others are more sensitive to the discharge rate. Therefore, the selection of characteristic variables may vary under different application scenarios.

[0077] Use machine learning algorithms to train the SOH prediction model, optimize the hyperparameters and evaluate the prediction accuracy to ensure that the model accurately reflects the changes in the battery health state;

[0078] Select machine learning algorithms suitable for processing time series data, such as long short-term memory network (LSTM), random forest regression (Random Forest Regression), or gradient boosting tree (GBDT), etc., to train the historical data and establish an SOH prediction model. When modeling, the data set needs to be divided into a training set, a validation set, and a test set, and the cross-validation method is used to optimize the model hyperparameters to improve the generalization ability. After the model training is completed, the prediction error of the model can be evaluated based on the validation set and the test set. For example, metrics such as mean squared error (MSE), root mean squared error (RMSE), and R 2 Coefficient of determination are used to evaluate the prediction accuracy. If the prediction error of the model is large, the accuracy can be improved by adjusting the model hyperparameters, increasing the characteristic variables, or using a more complex deep learning network (such as Transformer or CNN-LSTM hybrid architecture).

[0079] Integrate the SOH prediction model into the energy storage management system, continuously optimize it using a rolling update mechanism, and adjust the charge-discharge strategy in combination with optimization algorithms to extend the battery life and improve the system stability.

[0080] In the model deployment stage, the trained SOH prediction model needs to be integrated into the Energy Management System (EMS) to achieve real-time monitoring and prediction of the battery health state. Meanwhile, a rolling update mechanism is adopted to regularly perform online training on the model using the latest operation data, enabling it to adapt to the long-term changes in the battery aging state. In addition, parameter optimization methods such as Bayesian optimization or genetic algorithms can be combined to dynamically adjust the charge and discharge strategies of the energy storage system according to the current SOH of the battery, for example, reducing the frequency of high-rate charge and discharge or optimizing the charge and discharge window, so as to extend the battery service life. Finally, this model can not only be used for energy storage scheduling optimization, but also for early warning of possible battery failures, improving the safety and economy of the wind-solar-storage system.

[0081] In the day-ahead scheduling stage, combining the wind-solar power generation prediction data, grid load prediction data, and the health state of the energy storage system, a multi-objective optimization method is used to calculate the optimal charge and discharge plan of the energy storage, ensuring that while maximizing the wind-solar power consumption, the impact of high-rate charge and discharge on the battery life is reduced, and providing a benchmark plan for the intra-day scheduling;

[0082] In the day-ahead scheduling stage, the specific steps of using a multi-objective optimization method to calculate the optimal charge and discharge plan of the energy storage, ensuring that while maximizing the wind-solar power consumption, the impact of high-rate charge and discharge on the battery life is reduced, and providing a benchmark plan for the intra-day scheduling are as follows:

[0083] Integrate the wind-solar power generation prediction, grid load prediction, and energy storage health state data, and use uncertainty modeling methods to reduce the impact of prediction errors on the scheduling decision-making;

[0084] First, integrate the wind-solar power generation prediction data, grid load prediction data, and the health state data of the energy storage system to provide comprehensive information support for scheduling optimization. The wind-solar power generation prediction data is usually generated by meteorological models, historical power generation data, and short-term prediction algorithms (such as time series ARIMA models, LSTM, etc.). The grid load prediction is analyzed by combining historical load curves, weather factors, electricity consumption habits, etc. The health state data of the energy storage system is based on the aforementioned SOH prediction model, and through machine learning analysis of factors such as temperature, cycle times, charge and discharge rates, etc., the remaining available capacity, maximum allowable charge and discharge power, and the optimal charge and discharge window of the battery are dynamically evaluated. In addition, in order to cope with the errors in wind-solar power generation and load prediction, methods such as Monte Carlo simulation and Kalman filtering can be used to model the uncertainty, making the scheduling plan more robust and reducing the impact of prediction errors on the charge and discharge decision-making.

[0085] Construct a multi-objective optimization model that takes into account wind-solar power consumption, battery aging, and economic costs, and use the Pareto optimization algorithm to dynamically balance different optimization objectives;

[0086] After data fusion is completed, it is necessary to construct a multi-objective optimization model to take into account multiple factors such as maximizing wind and solar power consumption, minimizing battery aging, and optimizing the system operation cost. The optimization objectives can be set as follows: ①Maximize the self-consumption ratio of wind and solar power generation to avoid energy waste caused by curtailment of wind and solar power; ②Minimize battery aging loss and extend the life through measures such as restricting high-rate charge and discharge, controlling the SOC operating range, and balancing the load of the battery pack; ③Reduce the system operation cost, including peak shaving and valley filling to reduce the power purchase cost, and reducing the maintenance cost brought by frequent start and stop. To achieve these goals, a weight-based Pareto multi-objective optimization method can be used to dynamically balance the optimization objectives under different operation scenarios. In addition, constraint conditions need to consider the maximum charge and discharge power of the battery, the minimum SOC limit, the grid power balance constraint, etc., to ensure that the optimization results are feasible and meet the physical limitations.

[0087] Based on mixed integer linear programming, optimize the charge and discharge time, power and path of energy storage, and combine the MPC rolling optimization method to improve the scheduling adaptability;

[0088] Based on the multi-objective optimization model, use mixed integer linear programming (MILP) to solve the optimal scheduling scheme. During this process, the system needs to consider the following key factors: ①Determine the best charge and discharge time window to ensure that the battery charges when the electricity price is low and the photovoltaic output is high, and discharges during the peak load period or when the wind and solar output is low; ②Restrict high-rate charge and discharge, and dynamically adjust the maximum charge and discharge power according to the SOH prediction model to prevent premature aging; ③Optimize the charge and discharge path of the energy storage system. For example, in an energy storage system with multiple battery packs, give priority to using the battery packs with better health status to balance aging and optimize the overall life. During the calculation process, the rolling horizon optimization method (MPC) can be used to recalculate the optimal charge and discharge strategy for a period of time in the future at regular time intervals to adapt to the dynamic changes of wind and solar forecasts and improve the adaptability of the plan.

[0089] Formulate a benchmark scheduling scheme including dynamic SOC window adjustment and intelligent balanced charge and discharge strategy, and continuously correct and optimize it in the intraday scheduling stage in combination with the rolling optimization algorithm.

[0090] After obtaining the optimal charge-discharge plan, it is used as the benchmark scheduling scheme to provide an optimization reference for intraday scheduling. This benchmark scheme should not only include the charge-discharge time and power curve of the energy storage, but also contain battery health management strategies, such as dynamic SOC window adjustment (flexibly adjusting the SOC operating range according to the SOH state) and intelligent balanced charge-discharge strategy (allocating different charge-discharge tasks for battery packs with different aging degrees). In addition, in order to adapt to the uncertainties of intraday wind-solar power generation and load, the benchmark scheme can be combined with the day-ahead to intraday rolling optimization algorithm, and continuously updated and optimized during the intraday scheduling stage to reduce the control error caused by deviations. Finally, this optimization scheme can not only improve the wind-solar accommodation ratio, reduce the system operation cost, but also extend the service life of the energy storage system, providing long-term guarantee for the stable operation of the wind-solar-storage system.

[0091] During the intraday scheduling process, high-precision sensors and state estimation methods are used to obtain the actual output of wind-solar power generation, the real-time demand of the grid load, and the actual available capacity of the energy storage system, and compare them with the day-ahead scheduling plan to identify existing deviations, so as to determine whether it is necessary to adjust the energy storage charge-discharge plan;

[0092] During the intraday scheduling process, high-precision sensors and state estimation methods are used to obtain the actual output of wind-solar power generation, the real-time demand of the grid load, and the actual available capacity of the energy storage system, and compare them with the day-ahead scheduling plan to identify existing deviations, and the specific steps to determine whether it is necessary to adjust the energy storage charge-discharge plan are as follows:

[0093] Use high-precision sensors and state estimation methods to obtain wind-solar output, load demand and energy storage system status, and improve the measurement accuracy through data filtering and machine learning methods;

[0094] First of all, high-precision sensors and data acquisition systems (such as SCADA, PMU, etc.) are used to obtain the actual output of wind-solar power generation, the real-time demand of the grid load, and the current operating status of the energy storage system, including key parameters such as battery SOC, SOH, temperature, voltage, and current. These data are preprocessed through edge computing nodes or cloud computing platforms to remove outliers, filter and denoise, and use data assimilation techniques (such as Kalman filtering, particle filtering, etc.) to improve the measurement accuracy. In addition, combining historical data and real-time monitoring data, a state estimation model is constructed, and the true values of unmeasured variables are inferred through machine learning or Bayesian inference methods. For example, for a large-scale energy storage system, if some battery pack sensors are abnormal, its state can be estimated through the data of other normal batteries and the aging model to ensure the accuracy of scheduling decisions.

[0095] Compare the actual operating status with the day-ahead scheduling plan, calculate the key deviation indicators of wind-solar output, load demand and energy storage system, and evaluate the impact of the deviation on system operation;

[0096] After data acquisition and state estimation are completed, it is necessary to compare the actual operating state with the day-ahead scheduling plan and calculate key deviation indicators, such as: ①wind and solar power output deviation (actual power generation vs predicted power generation), ②load demand deviation (real-time load vs predicted load), ③energy storage system deviation (actual available battery capacity vs predicted available capacity). For different types of deviations, further analysis is required: if the wind and solar power output deviation is large, it may be necessary to increase the energy storage discharge power to balance the system power; if the load demand deviation is large, it may be necessary to adjust the energy storage charge and discharge timing; if the available capacity of the energy storage system is lower than expected, it may be necessary to adjust the charge and discharge rate or switch to a battery pack with a better health state. In addition, to improve the robustness of the scheduling decision-making, probability analysis methods, such as Monte Carlo simulation, can be used to calculate the energy gap or excess that the deviation may cause and evaluate the risk levels of different scheduling adjustment schemes, ensuring that the adjustment strategy can meet the supply-demand balance without exacerbating the aging of the energy storage system or increasing the operating cost.

[0097] Recalculate the future charge and discharge strategy based on the MPC rolling optimization method, optimize the SOC window, and improve the economy and the lifespan of the energy storage system by combining grid scheduling and market electricity prices;

[0098] Based on the deviation analysis results, use the real-time rolling optimization method (Model Predictive Control, MPC) to recalculate the optimal energy storage charge and discharge strategy for a future period. Specifically, this optimization process needs to comprehensively consider: ①the current SOC, SOH, and maximum allowable charge and discharge power of the energy storage system to ensure that the adjustment strategy will not damage the battery health; ②grid scheduling requirements to avoid frequent adjustment of the charge and discharge power resulting in system instability; ③market electricity prices (if there is a power market) to ensure that the energy storage scheduling is economically optimal. This optimization calculation usually uses methods such as mixed integer linear programming (MILP) or deep reinforcement learning (DRL) to optimize the charge and discharge decision-making on the premise of considering battery health. For example, if it is found that the wind and solar power output is lower than expected, the system can charge additionally during low electricity price periods to compensate for the future power gap, or reduce charging during the load valley to avoid excessive battery cycling. In addition, the SOC window can be dynamically adjusted during the optimization process, that is, the maximum and minimum SOC ranges are adjusted according to the battery health state, so as to extend the service life of the energy storage system.

[0099] Execute the optimized scheduling plan, provide real-time feedback to control the operating state of the energy storage, and use reinforcement learning technology to optimize the long-term scheduling strategy to improve the intelligent level of the system.

[0100] After the optimization calculation is completed, a new energy storage charge and discharge plan is generated and sent to the energy storage management system (EMS) to control the energy storage device to operate according to the adjusted strategy. To ensure the effective execution of the adjustment plan, the system will perform feedback control in a short cycle (such as 15 minutes or 30 minutes) to monitor the response of the energy storage system in real time. If a new deviation is found (such as the energy storage failing to discharge normally according to the adjustment plan), the system will trigger an exception handling mechanism, such as re-optimizing the scheduling, adjusting the backup power strategy, or notifying manual intervention. In addition, reinforcement learning technology can be used to enable the system to continuously self-learn during long-term operation and optimize the scheduling strategy in different environments. For example, during long-term operation, the system can gradually learn the wind and light output patterns under specific weather conditions and the impact of different SOC management strategies on battery life, thereby continuously improving the intelligence level of scheduling and ultimately achieving efficient, stable, and low-cost economic scheduling of the wind-solar-storage system.

[0101] Based on the battery health state parameters calculated by the energy storage dynamic aging model, adjust the scheduling strategy during actual operation. By introducing a health factor (Health Factor, HF), correct the charge and discharge plan to prevent overcharging and over-discharging of the battery. At the same time, optimize the load distribution of different battery packs to make the overall attenuation of the energy storage system balanced, extend the service life, and reduce the scheduling deviation caused by uncertainty.

[0102] Based on the battery health state parameters calculated by the energy storage dynamic aging model, adjust the scheduling strategy during actual operation. By introducing a health factor (Health Factor, HF), correct the charge and discharge plan to prevent overcharging and over-discharging of the battery. At the same time, optimize the load distribution of different battery packs to make the overall attenuation of the energy storage system balanced, extend the service life, and reduce the scheduling deviation caused by uncertainty. The specific steps are as follows:

[0103] Based on the key factors of battery SOH, SOC, and number of cycles, calculate the HF value through the weighted comprehensive evaluation method to measure the battery health state and provide a basis for scheduling optimization.

[0104] The Health Factor (HF) is a dynamically adjusted parameter for measuring the health state of a battery, which is calculated based on multiple key parameters such as the battery's SOH (State of Health), SOC (State of Charge), number of cycles, temperature, charge and discharge rate, etc. First, through the energy storage dynamic aging model, the health state of the battery is monitored and predicted in real time, and the core variables affecting the battery life are extracted. For example, some batteries may have a relatively fast increase in internal resistance due to long-term high-rate discharge, while others may have lithium metal precipitation due to long-term stay in a high SOC state, affecting safety. Subsequently, the weighted comprehensive evaluation method (such as the entropy weight method, Analytic Hierarchy Process AHP) is used to calculate the HF value of each battery. The HF range is usually set between 0 and 1. The lower the HF value, the higher the degree of battery aging, and the charge and discharge load needs to be reduced. Batteries with higher HF values can undertake more charge and discharge tasks to extend the service life of the entire energy storage system.

[0105] Optimize the charge and discharge task allocation of the battery pack according to the HF value, adjust the SOC working window, ensure that batteries in good health state undertake more loads, and achieve overall aging balance of the energy storage system;

[0106] After calculating the HF value, it is necessary to optimize the charge and discharge task allocation based on the health state of different battery packs to prevent over-aging of individual battery packs and achieve overall health balance of the energy storage system. Specifically, the dispatching system allocates the charge and discharge power according to the HF value. Batteries with high HF are given priority to undertake larger power, while batteries with low HF appropriately reduce the power output or reduce the number of deep charge and discharge cycles. In addition, the SOC working window can also be dynamically adjusted. For example, for batteries in poor health state, a narrower SOC range (such as 30%-70%) is set to reduce the impact of high SOC residence time and deep charge and discharge on them, while for batteries in good health state, a larger SOC range (such as 20%-90%) is allowed to improve the overall dispatching flexibility of the system. This process can use non-linear optimization methods (such as Particle Swarm Optimization PSO, Mixed Integer Linear Programming MILP) or reinforcement learning algorithms (such as Deep Reinforcement Learning DRL) to optimize the load allocation strategy and continuously adjust it during operation to make the system adapt to the dispatching requirements under different health states.

[0107] Adopt a day-ahead - intra-day rolling optimization mechanism to dynamically calculate the future dispatching strategy at regular intervals, combined with an adaptive control algorithm, to make the energy storage dispatching more flexible and economical;

[0108] On the basis of optimizing the load distribution, it is also necessary to combine the day-ahead and intra-day rolling optimization mechanism to adjust the charge and discharge plan in real time to adapt to the uncertainty of wind and solar power generation and the fluctuation of load demand. Specifically, the dispatching system will recalculate the optimal energy storage charge and discharge strategy for the future dispatching period at regular intervals (such as 15 minutes, 30 minutes) and perform dynamic optimization based on the HF value. During the rolling optimization process, the dispatching system needs to: ① Combine the latest wind and solar power generation prediction data to adjust the charging plan to avoid the high SOC state lasting too long; ② Adjust the discharge plan according to the grid load demand to prevent the battery from bearing too much load in the low SOH state; ③ Set the battery rotation strategy to ensure that the usage frequency of each battery pack is balanced and avoid the accelerated capacity decay of individual batteries due to long-term non-participation in charge and discharge. In this process, adaptive control algorithms (such as fuzzy control, Bayesian optimization) can be used to enable the system to learn automatically and continuously optimize the dispatching strategy based on historical data and real-time operating conditions to maximize the economic benefits and lifespan of the energy storage system.

[0109] Establish a closed-loop monitoring mechanism and use reinforcement learning methods to optimize the long-term dispatching strategy, enabling the energy storage system to continuously optimize under different operating environments, extend the battery life, and improve the overall operating efficiency.

[0110] After the adjusted charge and discharge strategy is executed, the system needs to monitor the change of the battery health status in real time to form a closed-loop control mechanism to ensure that the optimization effect meets the expectations. If it is found that the SOH of a certain battery pack still drops too fast, or the system has over-dispatching in some load scenarios, it is necessary to further optimize the HF calculation method or adjust the load distribution strategy of the battery pack. In addition, reinforcement learning techniques (such as Deep Q-Network DQN) can be used to enable the system to continuously iterate and optimize during long-term operation, improving the adaptability of the energy storage system under different operating environments. For example, the system can analyze the best HF dispatching strategy under different seasons and different wind and solar output patterns, and continuously optimize parameters such as the SOC window and charge and discharge rate according to historical data to ensure that the system is always in the best health state during long-term operation. Ultimately, this continuous optimization mechanism can reduce energy storage aging, improve the wind and solar power accommodation rate, optimize the economy, and ensure the safe and stable operation of the power grid.

[0111] Based on real-time monitoring data and the aging compensation control mechanism, use the Model Predictive Control (MPC) method to dynamically correct the day-ahead dispatching plan, recalculate the optimal energy storage charge and discharge strategy for a period of time in the future at regular intervals, so that it can adapt to the dynamic changes of wind and solar power output and load demand, and reduce the energy balance problem caused by prediction errors;

[0112] Based on real-time monitoring data and an aging compensation control mechanism, the Model Predictive Control (MPC) method is used to dynamically correct the day-ahead scheduling plan. The optimal energy storage charge and discharge strategy for a future period is recalculated at regular time intervals to adapt to the dynamic changes in wind and solar power generation and load demand, reducing the energy balance problems caused by prediction errors. The specific steps are as follows:

[0113] At the current moment, based on wind and solar power generation, grid load, and the state of the energy storage system, a rolling optimization objective function is established to minimize the aging loss of the energy storage system while ensuring system energy balance. During the prediction time horizon T p , the wind and solar power generation and load demand at future moments are predicted, and combined with the health factor of the battery, the available power of the energy storage system and the optimal scheduling strategy are calculated. The optimization objective is defined as follows:

[0114]

[0115] , where J is the optimization objective function, measuring the comprehensive cost of the energy storage scheduling strategy, t k is the current moment, i.e., the starting time point of the MPC calculation, T p is the prediction time horizon, i.e., the future time range considered when optimizing the scheduling plan, usually T p is between 1 hour and 24 hours, depending on the accuracy of wind and solar power prediction and the response speed of the energy storage system. t is the optimization moment, representing each discrete time point during the rolling optimization process. P b (t) is the charge and discharge power of the energy storage system at moment t, SOC(t) is the state of charge of the energy storage system at moment t, SOC ref is the target SOC value, usually set between 50% - 70% to reduce the deep charge and discharge of the battery and improve its service life. HF(t) is the battery health factor, used to measure the health state of the battery, usually ranging from 0 - 1: HF(t) = 1 indicates the best health state of the battery, allowing the maximum charge and discharge power; HF(t) < 1 indicates battery aging and the charge and discharge power needs to be reduced; when HF(t) ≈ 0, the battery may have been severely aged or damaged, and charge and discharge need to be minimized. λ 1 is the energy storage power regulation weight, controlling the power fluctuation of the energy storage system. λ 2 is the SOC deviation penalty weight, controlling the SOC to be maintained near the target value SOC ref , and λ 3 is the battery aging impact weight, controlling the charge and discharge power when HF is low;

[0116] The optimized objective function aims to minimize the energy storage power fluctuation, reduce the deviation of SOC from the reference value, and dynamically adjust the charge and discharge power to adapt to batteries with different aging degrees.

[0117] Based on the optimized objective function J, constraint conditions are constructed, including SOC constraints, battery power constraints, energy balance equations, etc., to ensure the feasibility of the optimized scheduling strategy and meet the system operation requirements in each prediction horizon T. p At the current time t, the MPC adopts a rolling optimization strategy. k The optimal energy storage charge and discharge strategy is calculated, but only the charge and discharge instructions within the current optimization period are executed. Subsequently, it enters the next optimization period, updates the prediction data, and re-solves the optimization problem. The energy balance equation is as follows:

[0118]

[0119] where SOC(t + 1) is the state of charge at the next moment, which is the SOC of the energy storage system calculated at the moment t + 1, representing the impact of the optimized scheduling strategy on the battery energy state. η c is the charging efficiency, representing the energy conversion efficiency of the battery during the charging process. Usually, η c < 1, and η d is the discharging efficiency, representing the energy conversion efficiency of the battery during the discharging process. Usually, η d < 1. and are the optimal charging power and discharging power at the moment t, respectively, which are obtained by optimizing and solving the objective function J, and satisfy (Only charging or discharging is allowed at the same moment), and C b is the rated capacity of the battery, which determines the SOC change rate.

[0120] Through the MPC rolling optimization, the system can correct the prediction error in real time, improve the adaptability of the scheduling, optimize the battery health state while satisfying the energy balance, and extend the service life of the energy storage system.

[0121] Combined with historical operation data and real-time scheduling feedback, the reinforcement learning algorithm (such as deep reinforcement learning DRL) is used to dynamically adjust the scheduling parameters, enabling the energy storage system to adapt to the long-term changes in the battery aging state, improving the energy storage utilization efficiency, and optimizing the scheduling strategy under different operation scenarios, so as to reduce the cumulative error during the long-term operation process and make the energy storage system intelligent and economical.

[0122] Combined with historical operation data and real-time scheduling feedback, the specific steps to dynamically adjust the scheduling parameters using reinforcement learning algorithms (such as deep reinforcement learning DRL) to make the energy storage system adapt to the long-term changes in the battery aging state, improve the energy storage utilization efficiency, and optimize the scheduling strategies under different operation scenarios, so as to reduce the cumulative error during long-term operation and make the energy storage system intelligent and economical are as follows:

[0123] First, define the state, action, and reward functions of reinforcement learning, and construct the state transition equation to enable the model to learn the optimal scheduling strategy of the energy storage system under different operating environments;

[0124] State S t Includes the current state of health (SOH) of the battery, state of charge (SOC), battery temperature T b , predicted value of wind and solar power generation P wp , grid load demand P load and market electricity price C m Key information, defined as:

[0125] S t ={SOH t , SOC t , T b , P wp , P load , C m}

[0126] Action A t Represents the scheduling decision of the energy storage system at the current moment;

[0127] A t ={P ch , P dis , ΔSOC}

[0128] , where P ch is the charging power, P dis is the discharging power, and ΔSOC is the SOC adjustment strategy;

[0129] The state transition equation is based on the battery dynamic aging model to calculate the SOH and temperature change at the next moment:

[0130] SOH t+1 =SOH t -α·f(T b , P ch , P dis )

[0131] where SOH t is the state of health of the battery at time t, SOH t+1 is the state of health of the battery at the next moment, α is the aging factor, f(T b , Pch , P dis ) represents the attenuation rate of the battery caused by temperature, charge and discharge power. T b is the battery temperature, P ch is the charging power, P dis is the discharging power;

[0132] The reward function is designed to optimize the scheduling strategy. A composite reward function is designed, including multiple objectives such as battery life optimization, grid power balance, and wind-solar accommodation rate, and is defined as:

[0133]

[0134] , where R t is the reward function, P curtail is the curtailed wind and solar power, representing the wind and solar power generation that cannot be accommodated due to insufficient energy storage or load. β 1 is the battery life optimization weight, which measures the importance of the change in the battery health state (SOH) in the reward function. β 2 is the wind-solar accommodation optimization weight, which measures the impact of the wind-solar power generation utilization rate on the scheduling strategy. β 3 is the economic cost optimization weight, which measures the role of the market electricity price and the charge and discharge cost in the optimization objective;

[0135] Through this reward function, the reinforcement learning system can learn how to adjust the charge and discharge strategy in different environments to achieve life optimization, economic optimization, and maximum wind-solar accommodation.

[0136] After determining the state S t , state A t and the reward function R t , deep reinforcement learning (Deep Deterministid Policy Gradient, DDPG) based on the Actor-Critic architecture is used for policy training, so that the energy storage system adapts to the battery aging state and optimizes the scheduling strategy in different scenarios;

[0137] The Actor network (policy network) is responsible for generating state A t from state S t , and the formula is as follows:

[0138] A t = π θ (S t ) + N t

[0139] , where π θ (S t ) is the parameterized policy, N tIt is to explore the use of noise (such as Ornstein-Uhlenbeck noise) to balance exploration and exploitation;

[0140] The Critic network (value network) is used to evaluate whether the strategy of the Actor network is optimal and calculate the Q value. The calculation formula is as follows:

[0141] Q(S t , A t ) = R t +γQ(S t+1 , A t+1 )

[0142] , where Q(S t , A t ) is the state-action value function, that is, the Q value, which represents the expected value of the future cumulative reward after taking action A in state S t . The Critic network is used to estimate this value and guide the Actor network to update the strategy to maximize the future return. γ is the discount factor, which is used to control the importance of future rewards. Q(S t , A t+1 ) is the Q value estimate at the next moment t+1, which represents the long-term cumulative return of the system after taking a new action A in the new state S t+1 ; t+1 t+1

[0143] Experience replay (Replay Buffer) To improve the training stability, the past state-action-reward-state transition data is stored in the experience replay pool and randomly sampled for training to avoid the gradient oscillation problem caused by data correlation;

[0144] Policy update: Gradient descent is used to optimize the Actor and Critic networks, and the policy parameters θ are updated to maximize the cumulative reward. The update formula is as follows:

[0145]

[0146] , where θ are the training parameters of the Actor network, including the weights and bias terms of the neural network, η is the learning rate, is the parameter update direction based on the policy gradient.

[0147] Finally, through reinforcement learning training, the system can dynamically adjust the scheduling parameters according to different operating scenarios, enabling the energy storage system to adapt to the battery aging state, improve the energy storage utilization efficiency, and optimize the long-term economy of the wind-solar-storage system.

[0148] ​​Embodiment 1: In the economic dispatch process of a wind-solar-storage system, accurately predicting the wind and solar power output and reasonably arranging the charge-discharge strategy of the energy storage system are important means to optimize energy utilization efficiency and improve economy. However, due to the uncertainty of wind and solar power generation, the dynamic changes of grid load, and the aging characteristics of the energy storage system itself, traditional fixed-rule dispatch methods often cannot adapt to complex operating environments. Therefore, this embodiment proposes an intelligent dispatch system based on Deep Reinforcement Learning (DRL), which can combine historical operation data with real-time dispatch feedback, autonomously learn the optimal charge-discharge strategy under different operating scenarios, and dynamically adjust the dispatch parameters of the energy storage system to optimize the overall operation effect of the wind-solar-storage system.

[0149] In the day-ahead dispatch stage, the system first obtains wind and solar power generation prediction data, grid load prediction data, and the health status information of the energy storage system, and uses a multi-objective optimization method to calculate the optimal energy storage charge-discharge plan. The optimization objectives include maximizing wind and solar power consumption, minimizing the battery aging rate, reducing the grid power purchase cost, etc., to ensure a balance between economy and equipment life. The calculated dispatch plan will be used as a reference plan for the intra-day dispatch stage.

[0150] In the intra-day dispatch stage, the system uses high-precision sensors to continuously monitor parameters such as wind and solar power output, grid load changes, energy storage SOC (state of charge), SOH (health state), and battery temperature, and uses state estimation techniques (such as Kalman filtering) to process the collected data to eliminate noise and outliers and improve data accuracy. When a large deviation is found between the actual state of the energy storage system and the day-ahead dispatch plan, the system will trigger a dynamic adjustment mechanism to recalculate the energy storage charge-discharge strategy for a future period of time to reduce dispatch errors. For example, when the actual wind and solar power output is much lower than the predicted value, the system can actively charge during low electricity price periods to avoid future power shortages, and when the wind and solar power output is higher than the predicted value, the system can increase the discharge power to make full use of the excess clean energy.

[0151] To enable the system to continuously optimize the scheduling strategy, deep reinforcement learning (DRL) is introduced into the scheduling process, enabling the system to continuously learn the best charge-discharge decisions from historical data and real-time feedback. DRL adopts a state-action-reward mechanism. Whenever the system executes a certain scheduling strategy, it calculates the benefits of the current decision based on key indicators such as battery life changes, wind and solar power accommodation ratios, and electricity costs, and optimizes future scheduling decisions based on long-term benefits. For example, if a certain charge-discharge strategy causes the battery's state of health (SOH) to rapidly decline in a short period, the system will reduce the selection of similar scheduling decisions in subsequent training to optimize the battery life. As the training progresses, the system will form a set of adaptive intelligent scheduling strategies that can dynamically adjust the charge-discharge power and SOC management strategies even under different environmental conditions to maximize the operating efficiency of the system.

[0152] Finally, the system can not only improve the economy and security of the wind-solar-storage system, but also significantly reduce the scheduling deviation caused by prediction errors, making the energy storage system more adaptable and capable of long-term optimization.

[0153] Specific implementation method 2: During the long-term operation process, the aging effect of energy storage batteries is a key issue affecting the scheduling accuracy and system life. Traditional scheduling methods usually rely on fixed charge-discharge strategies and ignore the dynamic changes in the battery health state, which may cause some batteries to age prematurely and affect the operating efficiency of the entire energy storage system. To solve this problem, this implementation method proposes an energy storage charge-discharge optimization method based on the health factor (HF). By real-time monitoring the battery health state and dynamically adjusting the load distribution of different battery packs, the overall attenuation of the energy storage system is balanced, and the long-term stability of the equipment is improved.

[0154] During the operation of the system, the energy storage management system (EMS) will real-time monitor key parameters such as the SOC, SOH, battery temperature, and charge-discharge rate of each battery pack, and calculate the health factor (HF) of each battery based on the energy storage dynamic aging model. The HF value is usually set between 0 and 1. The higher the value, the better the battery health state, and the more suitable it is to undertake more charge-discharge loads. Batteries with lower HF values should reduce the frequency of deep charge-discharge to reduce the aging speed.

[0155] During the scheduling decision-making process, the system optimally allocates the battery load based on the HF value to ensure that batteries in better health states undertake higher power, while batteries in poorer health states reduce their loads. For example, the system can dynamically adjust the SOC operating window. For battery packs with a lower SOH, it shrinks their SOC operating range (such as 30% - 70%) and reduces the dwell time at high SOC to mitigate lithium deposition. For batteries with a higher SOH, a larger SOC range (such as 20% - 90%) is allowed to improve energy utilization efficiency. Additionally, in large-scale energy storage systems, different battery packs may adopt different charge and discharge strategies. The system can classify and manage the battery packs according to the HF value, preferentially using battery packs in better health states to balance the battery aging rate and extend the overall lifespan of the system.

[0156] Through this method, the unbalanced aging phenomenon of energy storage batteries can be effectively reduced, the risk of premature retirement of individual battery packs can be lowered, and meanwhile, the scheduling flexibility and long-term operation stability of the entire wind-solar-storage system can be improved. Specific implementation method 3:

[0158] During the actual operation of the wind-solar-storage system, due to the continuous changes in factors such as wind and solar power output, grid load, and market electricity price, fixed scheduling strategies often cannot adapt to complex dynamic environments. Therefore, to ensure the long-term optimal operation of the energy storage system, this implementation method proposes a method that combines rolling optimization scheduling with an experience replay mechanism. Through real-time data updates and historical experience learning, the scheduling strategy can be continuously optimized to improve the operation efficiency of the wind-solar-storage system.

[0159] During the scheduling process, the system adopts the model predictive control (MPC) method of rolling horizon optimization and recalculates the optimal charge and discharge strategy for a certain period in the future every certain time interval (such as 15 minutes) to ensure that the scheduling plan always matches the actual system state. For example, in a short-term electricity market environment, the system can adjust the charge and discharge strategy according to the real-time market electricity price, charging during low-price periods and discharging during high-price periods to improve economic efficiency. At the same time, to improve the long-term stability of the scheduling strategy, the system introduces an experience replay mechanism (ReplayBuffer) to store the state-action-reward data during past operations and refer to the historical best scheduling strategy when making future decisions. For example, in extreme weather conditions (such as continuous rainy days or high-temperature weather), the system can call the best scheduling plan in past similar environments and adjust the current charge and discharge plan to better adapt to the current environmental changes.

[0160] Ultimately, this solution can improve the utilization efficiency of the energy storage system, reduce scheduling errors, and enable the energy storage system to achieve the optimal economic scheduling strategy under different operating environments.

[0161] The energy storage charge and discharge optimization method driven by the health factor (HF) of the present invention can effectively reduce the unbalanced aging phenomenon of energy storage batteries and improve the long-term operation stability of the entire energy storage system. In traditional scheduling methods, all battery packs often operate according to the same charge and discharge strategy and cannot be dynamically adjusted according to the health status of individual batteries, resulting in accelerated aging of some batteries due to high-rate discharge, long-term high SOC operation, or abnormal temperature, ultimately affecting the reliability of the entire system. In this solution, by real-time monitoring parameters such as SOH, SOC, temperature, and current, the battery health factor (HF) is calculated, and the charge and discharge task allocation is intelligently adjusted according to the HF value to ensure that batteries in better health states undertake greater power, while batteries with more serious aging reduce deep discharge and high-rate operation. This dynamic optimization strategy can extend the overall life of the energy storage system, reduce scheduling errors caused by the decline of individual batteries, and thus improve the stability of the system. In addition, combined with the deep reinforcement learning (DRL) technology, the system can continuously learn and optimize the SOC management strategy during long-term operation, such as dynamically adjusting the SOC window, optimizing the charge and discharge rate, etc., to minimize the battery aging rate to the greatest extent. This can not only reduce the maintenance and replacement costs of energy storage batteries but also improve the long-term return on investment of the energy storage system, making the wind-solar-storage system have better economic feasibility and operation reliability.

[0162] The present invention combines model predictive control (MPC) with the replay buffer mechanism to enable the system to dynamically adjust the charge and discharge strategy according to factors such as real-time load demand, market electricity price, and wind-solar prediction error, thereby improving the wind-solar accommodation rate and optimizing the grid energy balance. Specifically, during the intraday scheduling stage, the system can perform rolling optimization calculations based on the latest load and wind-solar output data every 15 - 30 minutes to ensure that the scheduling plan is always consistent with the actual operating state and reduce the scheduling deviation caused by prediction errors. In addition, through reinforcement learning technology, the system can continuously optimize the scheduling strategy under different meteorological conditions and load characteristics during long-term operation, improving the system's adaptability to extreme weather or sudden load fluctuations. For example, when encountering continuous low-wind and low-light weather, the system can use historical experience data to automatically switch to a better energy storage management strategy, such as charging in advance, reducing the discharge power, and optimizing the charging timing, etc., to ensure the stable operation of the grid. Ultimately, this optimization scheme can improve the utilization rate of wind-solar power generation while reducing the dependence on standby thermal power units, lowering the grid scheduling cost, and increasing the penetration rate of new energy in the power system, promoting the power system to develop in a more green and low-carbon direction.

[0163] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0164] The above has only described certain exemplary embodiments of the present invention by way of illustration. Without doubt, for those of ordinary skill in the art, various modifications can be made to the described embodiments in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0165] It should be noted that in this text, if there are relative terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0166] It should be understood that in various embodiments of the present application, the magnitude of the serial numbers of the above processes does not mean the order of execution is prior or subsequent. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0167] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this text can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0168] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0169] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0170] In addition, in each embodiment of the present application, each functional unit may be integrated into one processing unit, may exist separately as individual physical units, or two or more units may be integrated into one unit.

[0171] As described above, this is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0172] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. A method for economic dispatching of a wind, solar and energy storage system on the day before, characterized in that: The following steps are involved: Based on the historical operation data of the battery, a dynamic aging model is established. Through machine learning, the actual attenuation characteristics of the battery are fitted to obtain the battery health status under different operating conditions and its change pattern over time, so as to form the basis for dynamic adjustment strategies. In the day-ahead dispatching stage, the optimal charging and discharging plan for energy storage is calculated using a multi-objective optimization method, combining wind and solar power generation forecast data, grid load forecast data, and the health status of the energy storage system. This ensures that the impact of high-rate charging and discharging on battery life is reduced while maximizing wind and solar power consumption, and provides a benchmark solution for intraday dispatching. During the intraday dispatch process, high-precision sensors and state estimation methods are used to obtain the actual output of wind and solar power generation, the real-time demand of the grid load, and the actual available capacity of the energy storage system. These are then compared with the day-ahead dispatch plan to identify any deviations and determine whether the energy storage charging and discharging plan needs to be adjusted. Based on the battery health status parameters calculated by the dynamic aging model of energy storage, the scheduling strategy in actual operation is adjusted. By introducing the health factor to correct the charging and discharging plan, the battery is prevented from being overcharged and discharged. At the same time, the load distribution of different battery groups is optimized to balance the overall attenuation of the energy storage system, extend the service life and reduce the scheduling deviation caused by uncertainty. Based on real-time monitoring data and aging compensation control mechanism, the rolling time domain optimization method is used to dynamically modify the day-ahead dispatch plan, and the optimal energy storage charging and discharging strategy for a period of time in the future is recalculated at set time intervals to adapt to the dynamic changes in wind and solar power output and load demand, reducing energy balance problems caused by prediction errors. Combining historical operating data and real-time dispatch feedback, the reinforcement learning algorithm is used to dynamically adjust the dispatch parameters, so that the energy storage system can adapt to the long-term changes in the battery aging state, improve the energy storage utilization efficiency, and optimize the dispatch strategy under different operating scenarios, thereby reducing the cumulative error in the long-term operation process and making the energy storage system intelligent and economical.

2. The method for economic dispatching of a wind-solar-storage system within a day ahead according to claim 1 is characterized in that ,Based on the historical operation data of the battery, a dynamic aging model is established, and the actual attenuation characteristics of the battery are fitted through machine learning, and the battery health status under different operating conditions and its change law over time are obtained to form the basis of the dynamic adjustment strategy. The specific steps are as follows: Build a battery operation database, collect and clean key data on charge and discharge current, voltage, and temperature, remove outliers, and standardize processing to ensure data integrity and consistency; Screen the key characteristic variables that affect battery aging, use correlation analysis to optimize variable selection, and improve modeling accuracy; Use machine learning algorithms to train the SOH prediction model, optimize hyperparameters, and evaluate prediction accuracy to ensure that the model accurately reflects changes in battery health status; The SOH prediction model is integrated into the energy storage management system, and a rolling update mechanism is used for continuous optimization. The charging and discharging strategies are adjusted in combination with the optimization algorithm to extend battery life and improve system stability.

3. The method for day-ahead economic dispatch of a wind-solar-storage system according to claim 1 is characterized in that In the day-ahead dispatching stage, a multi-objective optimization method is used to calculate the optimal charging and discharging plan for energy storage, ensuring that the impact of high-rate charging and discharging on battery life is reduced while maximizing the wind and solar power consumption, and providing a benchmark solution for intraday dispatching. The specific steps are as follows: Integrate wind and solar power generation forecasts, grid load forecasts, and energy storage health status data, and use uncertainty modeling methods to reduce the impact of forecast errors on scheduling decisions; Construct a multi-objective optimization model that takes into account wind and solar power consumption, battery aging and economic costs, and use the Pareto optimization algorithm to dynamically balance different optimization objectives; Based on mixed integer linear programming, the energy storage charging and discharging time, power and path are optimized, and the MPC rolling optimization method is combined to improve the scheduling adaptability; Develop a benchmark scheduling plan that includes dynamic SOC window adjustment and intelligent balanced charging and discharging strategies, and combine it with a rolling optimization algorithm to continuously revise and optimize it during the intraday scheduling stage.

4. The method for day-ahead economic dispatch of a wind-solar-storage system according to claim 1 is characterized in that During the intraday dispatch process, high-precision sensors and state estimation methods are used to obtain the actual output of wind and solar power generation, the real-time demand of the grid load, and the actual available capacity of the energy storage system, and compare them with the day-ahead dispatch plan to identify the existing deviations and determine whether the energy storage charging and discharging plan needs to be adjusted. The specific steps are as follows: Use high-precision sensors and state estimation methods to obtain wind and solar power output, load demand and energy storage system status, and improve measurement accuracy through data filtering and machine learning methods; Compare the actual operating status with the day-ahead dispatch plan, calculate the key deviation indicators of wind and solar power output, load demand and energy storage system, and evaluate the impact of the deviation on system operation; Recalculate future charging and discharging strategies based on the MPC rolling optimization method, optimize the SOC window, and combine grid dispatch and market electricity prices to improve economic efficiency and energy storage system life; Execute the optimized dispatch plan, control the energy storage operation status with real-time feedback, and use reinforcement learning technology to optimize the long-term dispatch strategy to improve the intelligence level of the system.

5. The method for day-ahead economic dispatch of a wind-solar-storage system according to claim 1 is characterized in that ,Based on the battery health status parameters calculated by the dynamic aging model of energy storage, the scheduling strategy in actual operation is adjusted. By introducing the health factor to correct the charging and discharging plan, the battery is prevented from overcharging and discharging. At the same time, the load distribution of different battery groups is optimized to make the overall attenuation of the energy storage system balanced, extend the service life and reduce the scheduling deviation caused by uncertainty. The specific steps are as follows: Based on key factors such as battery SOH, SOC, and number of cycles, the HF value is calculated through a weighted comprehensive evaluation method to measure the battery health status and provide a basis for scheduling optimization; Optimize the charge and discharge task distribution of the battery pack according to the HF value, adjust the SOC working window, ensure that batteries in good health bear more loads, and achieve overall aging balance of the energy storage system; Adopting the day-ahead and day-intraday rolling optimization mechanism, dynamically calculating the future dispatch strategy at the same time interval, combined with the adaptive control algorithm, makes energy storage dispatch more flexible and economical; Establish a closed-loop monitoring mechanism and use reinforcement learning methods to optimize long-term scheduling strategies, so that the energy storage system can be continuously optimized under different operating environments, extend battery life and improve overall operating efficiency.

6. The method for day-ahead economic dispatch of a wind-solar-storage system according to claim 1 is characterized in that ,Based on real-time monitoring data and aging compensation control mechanism, the rolling time domain optimization method is used to dynamically modify the day-ahead dispatch plan, and the optimal energy storage charging and discharging strategy for a period of time in the future is recalculated at set time intervals to adapt to the dynamic changes in wind and solar output and load demand. The specific steps to reduce the energy balance problem caused by prediction errors are as follows: At the current moment, based on wind and solar power generation, grid load and energy storage system status, a rolling optimization objective function is established to minimize the aging loss of the energy storage system while ensuring the energy balance of the system. p In the system, the wind and solar power generation and load demand in the future are predicted, and the available power of the energy storage system and the optimal scheduling strategy are calculated in combination with the health factor of the battery. The optimization objective is defined as follows: ， Where J is the optimization objective function, which measures the comprehensive cost of the energy storage dispatch strategy, and t k is the current moment, T p is the prediction time domain, t is the optimization time, which means each discrete time point in the rolling optimization process, P b (t) is the charge and discharge power of the energy storage system at time t, SOC(t) is the state of charge of the energy storage system at time t, SOC ref is the target SOC value, HF(t) is the battery health factor, which is used to measure the health status of the battery, λ1 is the energy storage power regulation weight, which controls the power fluctuation of the energy storage system, λ2 is the SOC deviation penalty weight, and λ3 is the battery aging impact weight; Based on the optimization objective function J, constraints are constructed to ensure that the optimization scheduling strategy is feasible and meets the system operation requirements. p , MPC adopts a rolling optimization strategy, that is, at the current time t k Calculate the optimal energy storage charging and discharging strategy, but only execute the charging and discharging instructions within the current optimization cycle, then enter the next optimization cycle, update the prediction data, and re-solve the optimization problem. The energy balance equation is as follows: ， Where SOC(t+1) is the state of charge at the next moment, η c is the charging efficiency, η d is the discharge efficiency, and They are the optimal charging power and discharging power at time t, which are obtained by optimizing the objective function J, C b It is the rated capacity of the battery, which determines the rate of change of SOC.

7. The method for day-ahead economic dispatch of a wind-solar-storage system according to claim 1, characterized in that: Combining historical operation data and real-time dispatch feedback, the reinforcement learning algorithm is used to dynamically adjust the dispatch parameters, so that the energy storage system can adapt to the long-term changes in the battery aging state, improve the energy storage utilization efficiency, and optimize the dispatch strategy under different operation scenarios, thereby reducing the cumulative error in the long-term operation process. The specific steps to make the energy storage system intelligent and economical are as follows: First, the state, action and reward functions of reinforcement learning are defined, and the state transfer equation is constructed so that the model can learn the optimal scheduling strategy of the energy storage system under different operating environments; Status S t Including the current battery health status SOH, remaining power SOC, battery temperature T b , wind and solar power generation prediction value P wp , power grid load demand P load and the market electricity price C m Key information, defined as: S t ={SOH t ,SOC t ,T b ,P wp ,P load ,C m } Action A t Represents the dispatch decision of the energy storage system at the current moment; A t ={P ch ,P dis ,ΔSOC}, Among them, P ch is the charging power, P dis is the discharge power and ΔSOC is the SOC adjustment strategy; The state transfer equation is based on the battery dynamic aging model to calculate the SOH and temperature change at the next moment: SOH t+1 =SOH t -α·f(T b ,P ch ,P dis ) Among them, SOH t is the battery health status at time t, SOH t+1 is the battery health status at the next moment, α is the aging factor, f(T b , P ch , P dis ) represents the battery attenuation rate caused by temperature and charge / discharge power, T b is the battery temperature, P ch is the charging power, P dis is the discharge power; The reward function is designed to optimize the scheduling strategy and design a composite reward function, which is defined as: ， Among them, R t is the reward function, P curtail is the abandoned wind and solar power, β1 is the battery life optimization weight, β2 is the wind and solar power consumption optimization weight, and β3 is the economic cost optimization weight, which measures the role of market electricity prices and charging and discharging costs in the optimization target.

8. The method for economic dispatching of a wind-solar-storage system in the day ahead according to claim 7, characterized in that: In determining the state S t , State A t and the reward function R t Finally, deep reinforcement learning based on the Actor-Critic architecture is used for strategy training to make the energy storage system adaptive to the battery aging state and optimize the scheduling strategy in different scenarios; The Actor network is responsible for the state S t Generate state A t , the formula is as follows: A t =π θ (S t )+N t , Among them, π θ (S t ) is a parameterized strategy, N t is the exploration noise, used to balance exploration and exploitation; The Critic network is used to evaluate whether the strategy of the Actor network is optimal and calculate the Q value. The calculation formula is as follows: Q(S t ,A t )=R t +γQ(S t+1 ,A t+1 ), Among them, Q(S t , A t ) is the state-action value function, i.e., Q value, γ is the discount factor used to control the importance of future rewards, Q(S t+1 , A t+1 ) is the Q value estimate at the next time t+1; Experience replay: To improve training stability, the past state-action-reward-state transition data is stored in the experience replay pool, and randomly sampled from it for training to avoid gradient oscillation caused by data correlation. Strategy update, using gradient descent to optimize the Actor and Critic networks, update the strategy parameter θ to maximize the cumulative reward, the update formula is as follows: ， Where θ is the training parameter of the Actor network, η is the learning rate, It is the parameter update direction based on policy gradient.

Images