Multi-objective coordinated control method of energy storage power station for improving wind power consumption capacity

Abstract

The present application relates to the field of energy storage power station multi-objective collaborative control, and particularly relates to an energy storage power station multi-objective collaborative control method for improving wind power consumption capacity. The content includes: collecting multi-source data and preprocessing, based on the preprocessed multi-source data, constructing a wind power short-term prediction model to obtain a wind power prediction value; based on the wind power prediction value, generating a three-dimensional state matrix to determine the charging and discharging direction; based on the wind power prediction value and the preprocessed multi-source data, combining the charging and discharging direction to obtain the initial charging and discharging power; multi-objective optimization is performed on the initial charging and discharging power, and constraint verification and adjustment are performed to obtain the optimal charging and discharging power, which is converted into PCS and BMS instructions for execution. The problems of single wind power prediction data source, lack of dynamic correction, resulting in insufficient accuracy, and energy storage scheduling only relying on single supply and demand logic and one-sided constraint verification, resulting in low wind power consumption rate, large power grid fluctuation and fast energy storage battery life attenuation are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of multi-objective collaborative control of energy storage power stations, and in particular to a multi-objective collaborative control method for improving wind power consumption capacity of energy storage power stations. BACKGROUND

[0002] Wind power has become the core force of energy transformation, but the strong intermittency, volatility and local microclimate of wind energy can lead to problems such as difficulty in predicting wind power output, large grid impact, and low consumption rate. At present, the existing technology has formed a preliminary application system: the wind power prediction field has experienced the evolution from traditional statistical models (such as ARIMA, gray prediction) to machine learning models (such as SVM, random forest), and then to deep learning models (such as basic LSTM, GRU), among which ARIMA is suitable for short-period stationary data and has small prediction error in small wind conditions; SVM performs stably in small sample feature mapping, but has low efficiency in processing long time series data; basic LSTM can capture time sequence association, but it is not optimized for the characteristics of wind power "multi-factor coupling" and the key feature weight of wind speed is not strengthened; and data collection focuses on wind turbine SCADA data or regional meteorological data, and data preprocessing mainly uses simple deletion of outliers and mean filling of missing values, without forming a multi-source data collaborative processing mechanism; in the energy storage scheduling field, the mainstream adopts a fixed strategy of "peak clipping and valley filling" or a simple threshold control, such as setting the SOC 20%-80% as a safe interval to trigger charging and discharging; some advanced schemes integrate the grid load for preliminary judgment, but do not integrate wind power prediction, battery health (SOH) and other multi-dimensional information; the constraint check only covers the upper limit of charging and discharging power and the SOC range, and lacks linkage check of key indicators such as grid frequency / voltage deviation.

[0003] Therefore, the above-mentioned technology has the following limitations: first, the data source of the prediction model is single and lacks dynamic correction, resulting in insufficient accuracy; second, the scheduling logic is one-sided and the constraint check is not comprehensive, making it difficult to coordinate multiple target requirements. SUMMARY

[0004] The present application provides a multi-objective collaborative control method for improving wind power consumption capacity of energy storage power stations to solve the technical problems of insufficient accuracy due to single data source and lack of dynamic correction in the process of wind power prediction, and low wind power consumption rate, large grid fluctuation and fast degradation of energy storage battery life caused by single supply and demand logic in energy storage scheduling without considering grid stability and battery life and one-sided constraint check.

[0005] The multi-objective collaborative control method for improving wind power consumption capacity of energy storage power stations of the present application comprises the following steps:

[0006] S1. Collect meteorological data, wind turbine operation data, power grid operation data and energy storage battery state data, and perform data preprocessing to obtain preprocessed multi-source data; based on the preprocessed multi-source data, a wind power short-term prediction model is constructed to obtain a wind power prediction value;

[0007] S2. Based on the wind power prediction value, a three-dimensional state matrix is generated, and the charging and discharging direction is determined; based on the wind power prediction value and the preprocessed multi-source data, the initial charging and discharging power is calculated, combined with the charging and discharging direction, to obtain the initial charging power and the initial discharging power; the initial charging power and the initial discharging power are subjected to multi-objective optimization to obtain the charging power and the discharging power with the minimum total cost; the charging power and the discharging power with the minimum total cost are subjected to constraint verification and adjustment to obtain the optimal charging power and the optimal discharging power, and are converted into energy storage converter and battery management system instruction execution.

[0008] Preferably, the S1 specifically comprises:

[0009] In the construction process of the wind power short-term prediction model, based on the wind speed, wind speed difference and wind turbine equipment working condition parameters in the preprocessed multi-source data, a prediction model combining long short-term memory network and attention mechanism is used to obtain a wind power basic prediction value; a dynamic correction error term is introduced to dynamically correct the wind power basic prediction value to obtain the wind power prediction value.

[0010] Preferably, the S2 specifically comprises:

[0011] Based on the wind power prediction value, the real-time state of charge of the energy storage battery in the preprocessed multi-source data and the real-time load of the power grid in the preprocessed multi-source data, a three-dimensional state matrix is generated combined with a sliding window; based on the three-dimensional state matrix, combined with the power grid allowable fluctuation threshold, the charging and discharging direction is determined.

[0012] Preferably, the S2 specifically comprises:

[0013] The supply-demand difference between the wind power prediction value and the real-time load of the power grid in the preprocessed multi-source data is calculated, and a dynamic comprehensive weight coefficient is introduced to construct a basic adjustment term, combined with a power grid voltage constraint term and an energy storage state coordination term, to calculate the initial charging and discharging power.

[0014] Preferably, the S2 specifically comprises:

[0015] Based on the wind speed in the preprocessed multi-source data and the real-time frequency of the power grid in the preprocessed multi-source data, the influence of wind speed change and power grid frequency fluctuation on charging and discharging power is quantified to obtain a dynamic comprehensive weight coefficient.

[0016] Preferably, the S2 specifically comprises:

[0017] The energy storage state coordination term is constructed based on the battery health of the energy storage battery in the preprocessed multi-source data and the real-time state of charge of the energy storage battery in the preprocessed multi-source data through a linear weighted coupling method.

[0018] Preferably, S2 specifically includes:

[0019] Based on the initial charging power and the initial discharging power, the charging loss term and the discharging loss term are constructed in combination with the charging efficiency and the discharging efficiency in the preprocessed multi-source data, the multi-objective constraint optimization function is constructed in combination with the grid penalty term and the battery life loss term, and the charging power and the discharging power with the minimum total cost are obtained.

[0020] Preferably, S2 specifically includes:

[0021] The charging power and the discharging power with the minimum total cost are subjected to constraint verification, including charging and discharging power upper limit constraint, SOC safe range constraint, grid power quality constraint and charging and discharging mutual exclusion constraint; when any constraint fails, the failed constraint condition is located and the parameters are adjusted for recalculation until all constraints are satisfied, and the optimal charging power and the optimal discharging power are obtained; the optimal charging power and the optimal discharging power are converted into energy storage converter and battery management system instruction execution, and the multi-objective implementation is guaranteed.

[0022] The technical scheme of the present application has the following advantages:

[0023] 1. A three-dimensional state matrix capable of representing "wind power output-grid load-energy storage state" is constructed to accurately determine the charging and discharging direction and break through the traditional single factor driving logic; a multi-objective constraint optimization function is constructed to comprehensively consider the core demands of wind power consumption, grid stability and battery life, and the weight coefficients are determined by the analytic hierarchy process to realize the minimum total cost scheduling.

[0024] 2. The four core constraints cover the charging and discharging power upper limit, the SOC safe range, the grid power quality and the charging and discharging mutual exclusivity, and the backtracking adjustment mechanism is used to dynamically correct the parameters to avoid the risks of equipment overload, grid fluctuation and circulating current burnout, and the power-current closed-loop control method is used to realize the accurate execution of the instructions, which can adapt to sudden working conditions.

[0025] 3. While maximizing wind power consumption, the battery life is extended, the grid fluctuation penalty cost is reduced, the examination cost is reduced, the equipment operation and maintenance cost is optimized, and the safe and efficient operation of the new power system is facilitated. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 The flow chart of the energy storage power station multi-objective coordinated control method for improving wind power consumption capacity. DETAILED DESCRIPTION

[0027] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0029] The specific scheme of the multi-objective collaborative control method for the energy storage power station for improving the wind power consumption capacity provided by the present application will be specifically described below in conjunction with the drawings.

[0030] Referring to the drawings Figure 1 which shows a flow chart of the multi-objective collaborative control method for the energy storage power station for improving the wind power consumption capacity provided by an embodiment of the present application, the method comprises the following steps:

[0031] S1. Collecting meteorological data, wind turbine operation data, power grid operation data and energy storage battery state data, and performing data preprocessing to obtain preprocessed multi-source data; based on the preprocessed multi-source data, a wind power short-term prediction model is constructed to obtain a wind power prediction value;

[0032] First, data collection and preprocessing are performed, wherein the collected multi-source data covers four types of information of the wind farm: first, meteorological data collected through meteorological observation stations and high-precision meteorological sensor clusters deployed locally in the wind farm, including real-time wind speed, wind direction, wind speed difference, etc.; second, wind turbine operation data obtained through the sensor clusters of the wind turbine and the SCADA monitoring system of the wind farm, including historical wind speed time series data set of the wind farm, wind power, wind turbine equipment working condition parameters, factory rated value of the wind turbine equipment working condition parameters; third, power grid operation data obtained through the monitoring system of the power grid dispatching center, mainly including power grid real-time load, power grid real-time load change, power grid real-time voltage, power grid real-time frequency, rated frequency and power grid rated voltage, etc.; fourth, energy storage battery state data collected through the battery management system (BMS), including real-time state of charge, charging efficiency, discharging efficiency, real-time temperature, health degree and charging and discharging cycle number of the energy storage battery, etc.

[0033] Then the collected multi-source data is preprocessed according to the data source category (weather, wind turbine operation, power grid operation, and energy storage battery state), and the preprocessed multi-source data is obtained. Specifically, the 3σ criterion is used to eliminate abnormal data points, and the linear interpolation method is used to fill in the missing values in the multi-source data collection process. At the same time, all collected multi-source data is unified in data time granularity (such as 15 minutes per data point, i.e. one period), and finally the dimension difference between different multi-source data is eliminated through standardization processing such as Min-Max. In the preprocessing process of the wind turbine equipment operating parameter, multiple wind turbine equipment operating parameters (such as transmission system parameters and aerodynamic characteristic parameters) that are strongly related to wind energy capture efficiency and operation stability are selected. After Min-Max standardization processing of each parameter, the weights of each equipment operating parameter are determined through the analytic hierarchy process (AHP), and weighted coupling calculation is performed to obtain the preprocessed wind turbine equipment operating parameter. Finally, the structured data with unified structure, complete data, and precision up to standard, i.e. the preprocessed multi-source data, is formed.

[0034] After the multi-source data preprocessing is completed and the preprocessed multi-source data is output, a prediction model combining long short-term memory network (LSTM) and attention mechanism is used as the prediction core architecture, which integrates the time series feature extraction advantage of LSTM and the key data enhancement ability of attention mechanism, and introduces dynamic correction to build a wind power short-term prediction model to adapt to the characteristics of strong time series and multi-influence factor coupling of wind power. The specific implementation process is as follows: first, the preprocessed wind speed, wind speed difference, and wind turbine equipment operating parameter are input into a 3-layer LSTM to extract wind speed, operating condition, and other time series features. Then the time series features output by the 3-layer LSTM are transmitted to the attention layer to calculate the attention weights of each time series feature and perform weighted fusion to output the wind power basic prediction value. The wind power basic prediction value is dynamically corrected by combining the dynamic correction error term to obtain the wind power prediction value. In the wind power short-term prediction model, all LSTM gate weights (input gate, forget gate, and output gate) and attention layer weights are initialized using the Xavier initialization method to set the initial values. In the training process of the wind power short-term prediction model, the mean square error (MSE) is used as the loss function, the gradient is calculated through back propagation, and the Adam optimizer is used to update all model parameters adaptively until the loss converges to the minimum value.

[0035] The wind power short-term prediction model is as follows:

[0036] ,

[0037] wherein, is the time The predicted wind power output of the wind turbine units; For a moment Pre-treated wind speed at the location of the wind turbine; Indicates time Compared to the previous moment The difference in wind speed after pretreatment between them; These are the pre-processed operating parameters of the wind turbine equipment, used to represent the stationary characteristics of the wind turbine. This is the basic predicted value of wind power output based on the existing LSTM-attention mechanism prediction model; For a moment The dynamic correction error term for wind power forecasting, and the formula for calculating the dynamic correction error term: , This represents the average prediction bias within a historical time window, calculated at the current moment. Before The mean difference between the predicted and actual wind power values ​​(preprocessed wind power) at each time step is obtained. This is the length of the historical time window, and can be set to 5. For a moment The deviation between the real-time actual values ​​of the pre-processed wind turbine equipment operating parameters and the pre-processed factory rated values. Weighting coefficients After initialization using the Xavier method, optimization is performed by minimizing the prediction error (MSE loss): in each iteration, the gradient of the loss with respect to the weight coefficients is calculated, and the Adam optimizer updates the weight coefficients along the gradient direction until training converges, yielding the optimal coefficients. The range of values ​​is , The range of values ​​is ; This means that the dynamic correction error term is introduced into the wind power base prediction value by addition; the initialization and optimization methods of the above weight coefficients are well known to those skilled in the art and will not be described in detail here.

[0038] S2. Based on the predicted wind power, a three-dimensional state matrix is ​​generated, and the charging and discharging direction is determined. Based on the predicted wind power and preprocessed multi-source data, the initial charging and discharging power is calculated. Combined with the charging and discharging direction, the initial charging power and initial discharging power are obtained respectively. Multi-objective optimization is performed on the initial charging power and initial discharging power to obtain the charging power and discharging power with the minimum total cost. Constraint verification and adjustment are performed on the charging power and discharging power with the minimum total cost to obtain the optimal charging power and optimal discharging power, which are then converted into instructions for execution by the energy storage converter and battery management system.

[0039] Based on the wind power prediction value, combined with the pre-processed real-time state of charge of the energy storage battery, the pre-processed real-time load of the power grid, a three-dimensional state matrix is integrated. The integration process is to organize the above three types of data into a matrix in the time dimension. Then use the sliding window technology to organize the data at each time into a time series, thereby forming a state matrix. Based on the three-dimensional state matrix, the supply and demand balance logic is used as the core judgment criterion to determine the charging and discharging direction: when the wind power prediction value is greater than the pre-processed real-time load of the power grid, and the difference between the wind power prediction value and the pre-processed real-time load of the power grid exceeds the power grid allowable fluctuation threshold, it is determined that the wind power output is excessive, triggering the charging mode of the energy storage system matched with the wind farm construction. The power grid allowable fluctuation threshold is determined by the power grid operation standard, and is set to 5% of the pre-processed real-time load of the power grid; when the wind power prediction value is less than the pre-processed real-time load of the power grid, and the difference between the wind power prediction value and the pre-processed real-time load of the power grid exceeds the power grid allowable fluctuation threshold, it is determined that the wind power output is insufficient, triggering the discharging mode of the energy storage system matched with the wind farm construction; when the difference between the wind power prediction value and the pre-processed real-time load of the power grid is less than the power grid allowable fluctuation threshold, the energy storage system matched with the wind farm construction maintains standby state and real-time monitors the charging and discharging state changes; after determining the charging and discharging direction, based on the multi-objective optimization theory system, the charging and discharging power calculation formula is constructed to calculate the initial charging and discharging power, so as to break through the traditional single factor driven calculation logic, and the formula is as follows:

[0040] ,

[0041] wherein, is the initial charging and discharging power of the energy storage system at time . is the dynamic comprehensive weight coefficient at time , which is used to describe the influence of wind speed change and power grid frequency fluctuation on the charging and discharging power, and the specific calculation formula is wherein, are the wind speed state weight and the power grid state weight respectively, which are obtained by weighted least squares fitting based on historical wind turbine operation data and historical power grid operation data, and the value range is , , are the wind power operation state parameters and the power grid operation state parameters at time , , is the pre-processed wind speed at the location of the wind turbine at time . is the factory rated value of the pre-processed wind turbine equipment working condition parameter, , is the pre-processed wind turbine equipment working condition parameter at time a frequency deviation of the preprocessed real-time frequency of the power grid from the rated frequency, a preprocessed rated frequency; a time point a preprocessed real-time load of the power grid; a nonlinear regulation index for controlling the influence strength of the wind power supply-demand difference on the charging and discharging power, determined by least square fitting of the preprocessed wind power, the preprocessed real-time load difference of the power grid, and the preprocessed real-time state of charge of the energy storage battery , the value range of which is ; a basic regulation term, representing a supply-demand difference between the wind power prediction value and the preprocessed real-time load of the power grid, for avoiding too drastic wind power fluctuations that cause power grid shocks; a time point a difference between the preprocessed real-time voltage of the power grid and the preprocessed rated voltage of the power grid; a preprocessed rated voltage of the power grid; a power grid voltage constraint term, for avoiding that the charging and discharging operation of the energy storage system further aggravates the instability of the power grid operation when the power grid voltage is unstable; a time point a preprocessed real-time state of charge of the energy storage battery; a time point a preprocessed battery health degree; an energy storage state coordination term, constructed by a linear weighted coupling method based on the preprocessed battery health degree of the energy storage battery and the preprocessed real-time state of charge of the energy storage battery, the weight coefficient being obtained by least square fitting.

[0042] when the value of the initial charging and discharging power of the energy storage system is greater than zero, indicating that the energy storage system is in a discharging mode, in which the initial discharging power , and the initial charging power ; when the value of the initial charging and discharging power of the energy storage system is equal to zero, indicating that the energy storage system is in a basic supply-demand balance state, in which the energy storage system remains standby, ; when the value of the initial charging and discharging power of the energy storage system is less than zero, indicating that the energy storage system is in a charging mode, in which the initial charging power , and the initial discharging power . Subsequently, the initial charging power and the initial discharging power are input into a multi-objective constraint optimization function, and through multi-objective optimization calculation, the charging power and the discharging power with the minimum total cost are obtained, the multi-objective constraint optimization function taking as core targets the differentiation optimization of the charging and discharging process, the guarantee of power grid stability, and the prolongation of battery life, and being specific as follows:

[0043] ,

[0044] wherein, is a multi-objective constraint optimization function, which is a weighted sum of various costs (including charging loss, discharging loss, grid penalty, and battery life loss), and the optimization objective is to minimize the multi-objective constraint optimization function, i.e. ; is a weight coefficient of the multi-objective constraint optimization function, which is used to adjust the relative importance of each cost item, and is determined by a combination weighting method: first, based on the pre-processed historical energy storage battery state data and pre-processed historical grid operation data, a subjective weight is obtained by an analytic hierarchy process, and then, in combination with the historical grid operation data, an objective correction is made by an entropy weight method, and finally, a comprehensive weight of each cost item is obtained; is a charging loss item, which represents the cost of wind power consumption due to efficiency loss in the charging process of the energy storage system, and is calculated based on the initial charging power and the pre-processed charging efficiency , and the calculation formula is ; is a discharging loss item, which represents the cost of grid energy compensation due to efficiency loss in the discharging process of the energy storage system, and is calculated based on the initial discharging power and the pre-processed discharging efficiency , and the calculation formula is ; is a grid penalty item, which is used to penalize the grid fluctuation caused by energy storage scheduling, and is calculated based on the frequency deviation and voltage deviation by a mature quantitative model in the field of power system, is a comprehensive deviation value of the grid operation state caused by the charging and discharging operation of the energy storage system at time , which is calculated based on the frequency deviation and the voltage deviation by the existing power system optimization scheduling, grid stability analysis and automatic control system; is a battery life loss item, which reflects the influence of the charging and discharging process on the battery life, and is calculated by the existing battery degradation model based on the pre-processed real-time battery temperature and the pre-processed charging and discharging cycle number; is the pre-processed real-time battery temperature at time ; is the pre-processed battery charging and discharging cycle number at time .

[0045] Further, the total cost minimum charging power and discharging power and pre-processed multi-source data are checked one by one according to four core constraints, including: first, the upper limit constraint of charging and discharging power, to ensure that the charging and discharging power is not greater than the rated power of the energy storage converter PCS specified in the equipment manual, to avoid overloading and damaging the power device; second, the SOC safety range constraint, which is determined based on the cycle life characteristics of lithium ion batteries, to ensure that the real-time state of charge SOC of the energy storage battery is in the interval of 20%-80%; third, the power quality constraint of the power grid, which limits the frequency deviation ≤±0.2Hz and the voltage deviation ≤±5% according to the existing wind power grid connection technical standard, to ensure the stability of the power grid; and fourth, the charging and discharging exclusion constraint, which is limited based on the topology structure of the energy storage system, to ensure that only single mode operation is performed at the same time, to avoid equipment burning caused by charging and discharging loop circulation. If any constraint is not met (such as charging power of 260kW, exceeding the upper limit of the rated power of the energy storage converter PCS, or the real-time state of charge SOC of the battery = 82%, exceeding the upper limit of charging), the unpassed constraint condition needs to be located first, and the problem reason needs to be determined, that is, the charging and discharging power, SOC, power grid fluctuation or mutual exclusion problem, if the charging and discharging power is out of limit, the dynamic comprehensive weight coefficient in the charging and discharging power calculation formula is lowered , if the SOC is out of limit, the energy storage state coordination term in the charging and discharging power calculation formula is lowered , if the power grid fluctuation is out of limit, the weight coefficient of the power grid penalty term in the multi-objective constraint optimization function is increased ; if the charging and discharging exclusion occurs, the single operation mode is forced to be locked, the conflict power is cleared, and the compliance of the charging and discharging power is rechecked. Finally, the initial charging and discharging power is recalculated based on the adjusted parameters, and the multi-objective constraint optimization function is optimized, the constraint checking is repeated, and all constraints are met, and finally the total cost minimum charging power and discharging power that meet the safety compliance are output, that is, the optimal charging power and the optimal discharging power.

[0046] Finally, the mature power-current closed-loop control method in the field of power systems is adopted to convert the optimal charging power and the optimal discharging power into balanced instructions of the energy storage converter (PCS) and the battery management system (BMS), and the specific charging and discharging operation is performed by the driving energy storage converter (PCS) and the battery management system (BMS), which can adapt to various sudden operating conditions that may occur in actual operation while ensuring the stability of the power grid, prolonging the battery life and maximizing the wind power consumption.

[0047] In summary, the multi-objective collaborative control method for improving the wind power consumption capacity of the energy storage power station is completed.

[0048] The progressive nature of the specification and claims, with different embodiments, should not necessarily be viewed as having an ascending or descending order of precedence or suitability. The processes depicted in the figures do not necessarily require the particular order illustrated, as some steps can be performed in other orders or concurrently.

[0049] Each of the various embodiments described in this specification are described in progressive form, and the same or similar parts between each embodiment can be referred to each other, and each embodiment focuses on the difference from other embodiments.

[0050] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A multi-objective collaborative control method for energy storage power stations to enhance wind power absorption capacity, characterized in that, Includes the following steps: S1. Collect meteorological data, wind turbine operation data, power grid operation data, and energy storage battery status data, and perform data preprocessing to obtain preprocessed multi-source data; Based on the preprocessed multi-source data, a short-term wind power prediction model is constructed to obtain the predicted wind power value. S2. Based on the predicted wind power, a three-dimensional state matrix is ​​generated, and the charging and discharging direction is determined; based on the wind speed and real-time grid frequency in the preprocessed multi-source data, the impact of wind speed changes and grid frequency fluctuations on charging and discharging power is quantified to obtain dynamic comprehensive weighting coefficients; by calculating the supply and demand difference between the predicted wind power and the real-time grid load in the preprocessed multi-source data, and combining the dynamic comprehensive weighting coefficients, a basic adjustment term is constructed, and the initial charging and discharging power is calculated by combining the grid voltage constraint term and the energy storage state coordination term; The energy storage state coordination term is constructed using a linear weighted coupling method based on the battery health status and real-time state of charge of the energy storage battery in the preprocessed multi-source data. Based on the initial charge and discharge power, and combined with the charge and discharge direction, the initial charging power and initial discharge power are obtained respectively. Multi-objective optimization is performed on the initial charging power and initial discharge power to obtain the charging power and discharge power with the minimum total cost. Constraint verification and adjustment are performed on the charging power and discharge power with the minimum total cost to obtain the optimal charging power and optimal discharge power, which are then converted into instructions for execution by the energy storage converter and battery management system.

2. The multi-objective collaborative control method for energy storage power stations to enhance wind power absorption capacity according to claim 1, characterized in that, S1 specifically includes: In the process of constructing the short-term wind power prediction model, based on the wind speed, wind speed difference and wind turbine equipment operating parameters in the preprocessed multi-source data, the basic predicted value of wind power is obtained through a prediction model that combines long short-term memory network and attention mechanism; a dynamic correction error term is introduced to dynamically correct the basic predicted value of wind power to obtain the predicted value of wind power.

3. The multi-objective collaborative control method for energy storage power stations to enhance wind power absorption capacity according to claim 1, characterized in that, S2 specifically includes: Based on the predicted wind power, the real-time state of charge of energy storage batteries in the preprocessed multi-source data, and the real-time load of the power grid in the preprocessed multi-source data, a three-dimensional state matrix is ​​generated by combining a sliding window. Based on the three-dimensional state matrix and the allowable fluctuation threshold of the power grid, the charging and discharging direction is determined.

4. The multi-objective collaborative control method for energy storage power stations to enhance wind power absorption capacity according to claim 1, characterized in that, S2 specifically includes: Based on the initial charging power and initial discharging power, and combined with the charging efficiency and discharging efficiency from the preprocessed multi-source data, charging loss terms and discharging loss terms are constructed. Combined with the grid penalty term and battery life loss term, a multi-objective constrained optimization function is constructed, and the charging power and discharging power with the minimum total cost are obtained.

5. The multi-objective collaborative control method for energy storage power stations to enhance wind power absorption capacity according to claim 4, characterized in that, S2 specifically includes: The charging and discharging power with the minimum total cost are constrained and verified, including upper limit constraints on charging and discharging power, SOC safety range constraints, grid power quality constraints, and charging and discharging mutual exclusion constraints. When any constraint fails, the failed constraint condition is located, and the parameters are backtracked and recalculated until all constraints are satisfied, thus obtaining the optimal charging and discharging power. The optimal charging and discharging power are then converted into instructions for the energy storage converter and battery management system to ensure the achievement of multiple objectives.

Images