Tracking planned output method of wind-storage combined system

Through the dual-layer control and deep learning network prediction methods, wind power output is optimized, wind farm volatility and energy storage compensation economic problems are solved, and the stable operation and economic benefits of the wind-storage joint system are achieved.

CN118971182APending Publication Date: 2024-11-15GUONENG JILIN NEW ENERGY DEVELOPMENT CO LTD
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411015705.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The randomness and volatility of the wind farm make it difficult for it to respond to the grid frequency regulation instructions, affecting the grid stability, and the single energy storage compensation lacks economicality, making it difficult to achieve the optimal coordinated operation of the wind-storage joint system.

Method used

Using a two-layer control method, through point-by-point rolling optimization and model prediction control, combining the historical data of the wind farm and the charging and discharging capabilities of the energy storage system, the wind power output is optimized, the fluctuation is suppressed, and the wind power output is predicted through the deep learning network to realize wind- The tracking plan of the combined storage system is contributed.

Benefits of technology

The fluctuations in wind power output have been stabilized, the operating stability and economic benefits of the wind-storage joint system have been improved, and the punishments caused by wind loss and insufficient power have been reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118971182A_ABST
    Figure CN118971182A_ABST
Patent Text Reader

Abstract

The invention discloses a planned output tracking method of a wind-storage combined system, solves the problem of how to determine an optimal wind-storage coordinated operation strategy, and belongs to the field of new energy (wind power) power generation and energy storage. According to the method, rolling optimization is carried out after fluctuation stabilization is carried out on actual output of the wind power plant through double-layer control, in upper-layer control, on the basis of a wind power fluctuation limit value, a point-by-point rolling optimization mode is adopted, stored energy participates in adjustment to reduce fluctuation, and the fluctuation stabilization effect is achieved; in lower-layer control, a model prediction control method is adopted, data of planned output and actual output of a wind power plant are utilized, charging and discharging capacity of energy storage is combined, and a wind-storage combined system double-layer coordinated operation strategy for tracking planned output is researched by taking a minimum prediction error as a target; and finally, verifying the tracking accuracy through experiments, and obtaining each evaluation index.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method for tracking planned output of a wind-storage combined system, and belongs to the field of new energy (wind power) power generation and energy storage. Background Art

[0002] Wind energy is a clean energy with randomness and volatility, which is often affected by various factors and changes, including geographical location, weather conditions, seasonal changes, surface openness and other influencing conditions. Wind power is usually provided to power grids and users as a source of electricity, but the uncertainty brought about by the randomness and volatility of wind power often leads to insufficient ability of wind power stations to respond to power grid frequency modulation instructions, making it difficult to ensure that the stability quality of the power grid meets the requirements. In order to ensure that the output power of wind farms can supply the power grid and power load, an emerging technology is urgently needed to adjust the wind power output. In view of the challenges faced by high-proportion renewable energy grid connection, configuring energy storage power stations is an effective means to solve the problem of wind abandonment and maintain stable operation of the system. In this case, the introduction of energy storage technology and the coordinated operation of wind farms form a wind-storage joint system. Through the power dispatching capacity of energy storage, the shortage / redundancy of wind power is compensated / absorbed, thereby reducing the penalty caused by insufficient power and the loss of wind abandonment. Given that the development of wind-storage joint technology is not yet mature, and the single energy storage compensation does not reflect the economic efficiency well, it has become an important topic to find the optimal wind-storage coordinated operation strategy in the scenario of comprehensive consideration of multiple factors. Summary of the invention

[0003] In order to determine the optimal wind-storage coordinated operation strategy, the present invention provides a method for tracking planned output of a wind-storage combined system.

[0004] A method for tracking planned output of a wind-storage combined system of the present invention comprises:

[0005] S1. Get the optimal planned output P plan (n+k), n represents the current sampling time, k is a positive integer;

[0006] S2. Determine the real-time output P of the wind farm actual (n) Whether it is in the energy storage fluctuation range [P lower_limit (n),P upper_limit (n)], if yes, go to S5, if no, go to S3; P upper_limit (n), P lower_limit (n) are the upper and lower limits constrained by energy storage fluctuations;

[0007] S3, when the wind farm outputs real-time power P actual (n) greater than P upper_limit (n), the energy storage system is charged and the process goes to S4, otherwise the energy storage system is discharged and the process goes to S4;

[0008] S4. Construct the objective function for smoothing fluctuations and energy storage constraints. After solving the objective function under the energy storage constraints, the real-time wind power output P after smoothing fluctuations is obtained. actual' (n), the objective function of smoothing fluctuations is:

[0009] maxαAcc-βCDc

[0010] Among them, the accuracy of energy storage in smoothing fluctuations Energy storage charge and discharge times ratio P actual' (n) is the real-time wind power output after smoothing fluctuations, α and β are weight coefficients, Indicates the rate of change of the total amount of energy storage;

[0011] S5. Establish a prediction model to obtain the real-time predicted wind power output P at time n+k adjust (n+k|n);

[0012] S6. According to the obtained P adjust (n+k|n), with the goal of minimizing the tracking error and the energy storage output at the time n+k, the wind power compensation objective function and energy storage constraint conditions are established, and the wind power compensation objective function is solved under the energy storage constraint conditions to obtain the output P after wind power compensation adjust (n);

[0013] The objective function of wind power compensation is:

[0014]

[0015] Where N represents the total number of sampling moments, ΔE BESS (n+k|n) is the predicted output of energy storage at time n+k, and γ is the proportion of energy storage output in the objective function;

[0016] S7, output power P after wind power compensation adjust (n).

[0017] As a preferred embodiment, in S5, the state of charge SOC(n) of the energy storage system is used as the state variable x(n) of the control system, and the charging and discharging power ΔE of the energy storage system is BESS (n) is used as the control variable u(n), and the output P after wind power compensation is adjust (n) is the output of the system y(n), and the wind power prediction error ε(n) = P actual' (n)-P plan (n) is used as the disturbance r(n), a prediction model is established, and the MPC control method is adopted to obtain the state equation of the output y(n) at time n+k:

[0018]

[0019]

[0020] Among them, SOC(n+k|n) and SOC(n+k-1|n) are the predicted values ​​of the energy storage charge state at time n+k and time n+k-1 respectively.

[0021] Preferably, in S6, the objective function and the constraints are solved by quadratic programming, and the objective function is converted into:

[0022]

[0023] Where u is the optimization variable vector, including ΔE at all sampling times BESS (n), H is the coefficient matrix of the quadratic term, and f is the coefficient vector of the linear term.

[0024] As a preferred choice, the optimal planned output P is obtained plan (n+k) method:

[0025] A training set of a wind power prediction model is constructed based on historical data of wind power stations. The input data of each training sample in the training set includes the output value of the wind power station at each sampling point within any continuous M hours, and the output data includes the optimal planned output at the kth moment after M consecutive hours.

[0026] Using the training set to train the wind power prediction model, and obtaining a trained wind power prediction model;

[0027] The wind farm real-time output P at the current moment actual (n) is input into the wind power prediction model, and the wind power prediction model obtains the optimal planned output P plan (n+k), the wind power prediction model is a deep learning network.

[0028] Preferably, the energy storage constraint in S4 is:

[0029]

[0030] Where, ΔE BESS (n) is the change in charging and discharging power of the energy storage system, E BESS (n) is the energy storage power, ΔE BESS,max , ΔE BESS,min = ΔE BESS The upper and lower limits of (n), E BESS,max 、E BESS,min For E BESS The upper and lower limits of (n).

[0031] Preferably, the energy storage constraint condition in S6 is:

[0032]

[0033] where ΔE BESS (n) is the change in charging and discharging power of the energy storage system, E BESS (n) is the energy storage power, ΔE BESS,max , ΔE BESS,min = ΔE BESS The upper and lower limits of (n), E BESS,max 、E BESS,min For E BESS The upper and lower limits of (n).

[0034] As a preferred embodiment, during the charging and discharging process of the energy storage system in S3, the charging power P at time n is c (n) and discharge power P f (n) respectively satisfy:

[0035]

[0036] Where P c,max and P f,max are the maximum charging and discharging power of the energy storage system, σ is the self-discharge rate of the energy storage system, η c , η d are the charging and discharging efficiency of the energy storage system, E BESS,max is the maximum capacity of the energy storage system, E BESS,min is the minimum capacity of the energy storage system, ΔE BESS (n) is the change in charging and discharging power of the energy storage system at time n-1, and Δt is the duration between two adjacent times.

[0037] The beneficial effects of the present invention are as follows: the present invention adopts double-layer control to smooth the fluctuation of the actual output of the wind farm and then performs rolling optimization. In the upper-layer control, based on the wind power fluctuation limit, a point-by-point rolling optimization method is adopted, and the energy storage participates in the regulation to reduce the fluctuation, thereby achieving the effect of smoothing the fluctuation; in the lower-layer control, the model predictive control method is adopted, using the data of the planned output and actual output of the wind farm, combined with the charging and discharging capacity of the energy storage, with the minimum prediction error as the goal, and the double-layer coordinated operation strategy of the wind-storage combined system that tracks the planned output is studied; finally, the tracking accuracy is verified through experiments, and various evaluation indicators are obtained. The present invention is mainly used for the output control of wind power stations equipped with energy storage power stations, so as to achieve the optimal output of the wind-storage system and realize the optimal economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the method of the present invention;

[0039] Figure 2 It is a schematic diagram of the principle of MPC control of the present invention;

[0040] Figure 3 It is a framework diagram of upper and lower layer data transmission in the MPC control process of the present invention;

[0041] Figure 4 This is the wind power output prediction curve using the long short-term memory method. The horizontal axis is the number of sampling points, and the vertical axis is the predicted output.

[0042] Figure 5 It is a curve chart for tracking planned output comparison;

[0043] Figure 6 It is the wind power plan, real-time output and energy storage output curve;

[0044] Figure 7 It is a curve diagram of energy storage charge state change. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0047] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0048] The purpose of this embodiment is to propose a method for tracking the planned output of a wind-storage combined system, which is to optimize the coordinated operation of the wind-storage combined system that tracks the planned output, so as to better predict and optimize the actual output. The method for tracking the planned output of a wind-storage combined system in this embodiment includes:

[0049] Step 1: Get the optimal planned output P plan (n+k), n represents the current sampling time, k is a positive integer;

[0050] This implementation method can use a deep learning network as a wind power prediction model, and construct a training set of the wind power prediction model based on historical data of the wind power station. The input data of each training sample in the training set includes the output value of the wind power station at each sampling point within any continuous M hours, and the output data includes the optimal planned output at the kth moment after M consecutive hours.

[0051] Using the training set to train the wind power prediction model, and obtaining a trained wind power prediction model;

[0052] The wind farm real-time output P at the current moment actual (n) is input into the wind power prediction model, and the wind power prediction model obtains the optimal planned output P plan (n+k);

[0053] The wind power prediction model of this embodiment can use the long short-term memory (LSTM) neural network to perform ultra-short-term power prediction on the historical output data of the wind farm. By appropriately increasing the proportion of training sets, learning rate and training rounds, the accuracy of prediction can be maximized without significantly extending the simulation time.

[0054] The dispatch center or national policy will provide real-time wind farm output P actual (n), the real-time output of the wind farm provided by the dispatch center P actual (n) often has strong volatility. Therefore, in order to better track the planned output, it is necessary to smooth the wind power output. Here, the model prediction method is adopted. In the upper-level control, the limit of wind power fluctuation is introduced to judge whether it is necessary to smooth the fluctuation operation. The fluctuation limit ratio is selected as a to calculate the fluctuation range, where the lower limit and upper limit of the fluctuation range are:

[0055] P lower_limit (n) = P actual (n-1)×(1-a)

[0056] P upper_limit (n) = P actual (n-1)×(1+a)

[0057] Step 2: Determine the real-time output P of the wind farm actual (n) Whether it is in the energy storage fluctuation range [P lower_limit (n),P upper_limit (n)], if yes, go to step 5, if no, go to step 3; P upper_limit (n), P lower_limit (n) are the upper and lower limits constrained by energy storage fluctuations:

[0058] Step 3: When the wind farm outputs real-time power P actual (n) greater than P upper_limit (n), charge the energy storage system and go to step 4; otherwise, discharge the energy storage system and go to step 4;

[0059] Step 4: Energy storage participates in regulating wind power fluctuations. The maximum capacity and initial state of charge of energy storage have been given by the energy storage department. At the same time, considering the influencing factor of energy storage life, the highest accuracy of energy storage after smoothing fluctuations and the smallest possible number of charge and discharge times are combined as the objective function. Energy storage and wind power output constraints are used as constraints to construct an upper-level mathematical model to regulate energy storage output. The upper-level mathematical model includes:

[0060] The objective function for smoothing fluctuations and energy storage constraints are constructed. After solving the objective function under the energy storage constraints, the real-time wind power output P after smoothing fluctuations is obtained. actual' (n), the objective function of smoothing fluctuations is:

[0061] maxαAcc-βCDc

[0062] Among them, the accuracy of energy storage in smoothing fluctuations Energy storage charge and discharge times ratio P actual' (n) is the real-time wind power output after smoothing fluctuations, α and β are weight coefficients, Indicates the rate of change of the total amount of energy storage;

[0063] The fluctuation smoothing objective function consists of two parts, namely maximizing the accuracy of fluctuation smoothing and minimizing the number of energy storage charging times, where Acc is the accuracy of energy storage fluctuation smoothing, that is, the proportion of points within the limit range after energy storage smoothes fluctuations to the total reference points, and CDc is the proportion of energy storage charging and discharging times. The number of energy storage charging and discharging times is the main indicator for measuring the life of energy storage, and is the proportion of energy storage charging and discharging times (that is, the points of charge and discharge conversion of the energy storage system) to all points. The two objective functions are both percentages in terms of values, but as the power plant's expectations for different goals change, the proportion between the two will also change accordingly. Therefore, the relationship between the two cannot be simply added together. The weight coefficients α and β are introduced to take into account both subjective and objective factors for real-time adjustment.

[0064] In the case of considering multi-objective optimization, the volatility of wind power output is reduced as much as possible, and a relatively smooth wind power real-time output value P is obtained. actual' (n), the energy storage constraint in this process is:

[0065]

[0066] Where, ΔE BESS (n) is the change in charging and discharging power of the energy storage system, E BESS (n) is the energy storage power, ΔE BESS,max , ΔE BESS,min = ΔE BESS The upper and lower limits of (n), E BESS,max 、E BESS,min For E BESS The upper and lower limits of (n).

[0067] The rest are upper and lower limit constraints. The change of energy storage in this process is as follows, which needs to be carried out with the rolling of sampling points:

[0068] S(n+1)=S(n)-ΔE BESS (n)×Δt

[0069] In the formula, S(n) is the total energy storage at time n, S(n+1) is the total energy storage at time n+1, and Δt is the duration of the interval;

[0070] The upper control layer performs optimization in a shorter time period to ensure the real-time response capability of the system. The execution results and real-time data of the upper control layer are fed back to the lower control layer to help the lower control layer perform more accurate optimization in the next cycle. And output a smoother wind-storage combined system output curve.

[0071] Step 5: In the lower-level control, the model prediction method is used to track and predict the real-time wind power output after smoothing the fluctuation. The state of charge SOC(n) of the energy storage system is used as the state variable x(n) of the control system, and the charging and discharging power ΔE of the energy storage system is used as the charging and discharging power ΔE of the energy storage system. BESS (n) is used as the control variable u(n), and the output P after wind power compensation is adjust (n) is the output of the system y(n), and the wind power prediction error ε(n) = P actual' (n)-P plan (n) is used as the disturbance r(n) to establish a prediction model to obtain the real-time predicted wind power output P at time n+k adjust (n+k|n), the prediction model has state equation and output equation as follows:

[0072]

[0073] P adjust (n) = P actual' (n)+ΔE BESS (n)

[0074] Where SOC(n) and SOC(n+1) are the energy storage charge states at time n and time n+1 respectively.

[0075] The system's output y(n) state model formula at time n+k is:

[0076]

[0077]

[0078] In the formula, SOC(n+k|n) and SOC(n+k-1|n) are the predicted values ​​of the energy storage charge state at time n+k and time n+k-1 respectively;

[0079] Step 6: According to the obtained P adjust (n+k|n), with the goal of minimizing the tracking error and the energy storage output at the time n+k, comprehensively considering the limitation of energy storage capacity, optimizing the real-time output tracking, establishing the wind power compensation objective function and energy storage constraint conditions, solving the wind power compensation objective function under the energy storage constraint conditions, and obtaining the output P after wind power compensation adjust (n);

[0080] The objective function of wind power compensation is:

[0081]

[0082] Where N represents the total number of sampling moments, ΔE BESS (n+k|n) is the predicted output of energy storage at time n+k, and γ is the proportion of energy storage output in the objective function;

[0083] The energy storage constraints are:

[0084]

[0085] where ΔE BESS (n) is the change in charging and discharging power of the energy storage system, E BESS (n) is the energy storage power, ΔE BESS,max , ΔE BESS,min = ΔE BESS The upper and lower limits of (n), E BESS,max 、E BESS,min For E BESS The upper and lower limits of (n).

[0086] Solve the output P after wind power compensation adjust (n) Method: Minimizing the sum of squares of prediction errors is a typical quadratic programming problem. Quadratic programming can effectively express this objective function, making the optimization process have clear mathematical and practical significance. Solving it through quadratic programming is converted to:

[0087]

[0088] Where u is the optimization variable vector, including ΔE at all time nodes BESS (n), H is the coefficient matrix of the quadratic term, f is the coefficient vector of the linear term, and the constraints remain unchanged.

[0089] Step 7: Output wind power compensation output P adjust (n). The output tracks the coordinated output of the optimized wind farm and energy storage, and gives the energy storage output value for each data point, thereby completing the optimal allocation of the actual output of the wind power station and calculating the evaluation index parameters.

[0090] Data transmission between upper and lower layers:

[0091] Figure 3 This is a data transmission framework diagram for double-layer control. The upper and lower layers are connected through a central data bus. The real-time power data, energy storage system status data, and fluctuation smoothing results calculated by the upper layer are sent to the central data bus, and then go through a series of data serialization and deserialization to reduce the transmission bandwidth occupancy.

[0092] Energy storage system constraints:

[0093] Taking the energy storage state of the system into consideration, it can be concluded that the energy storage constraint of the system is within a certain range, in which the battery energy storage can minimize the life attenuation. Consider the following energy storage constraints:

[0094]

[0095] 0≤SOC(n)≤1

[0096] Among them, SOC(n) is the state of charge at the nth point. The specific constraints will be pointed out in the following content. The overall constraints of the system are:

[0097]

[0098] Energy storage system charging and discharging power P c (n) and P f (n) respectively satisfy:

[0099]

[0100] Where P c,max and P f,max are the maximum charging and discharging power of the energy storage system, P c (n) and P f (n) are the charging and discharging power of the energy storage system at time n, σ is the self-discharge rate of the energy storage system, η c , η d are the charging and discharging efficiency of the energy storage system, E BESS,max is the maximum capacity of the energy storage system, E BESS,min is the minimum capacity of the energy storage system, E BESS (n-1) is the remaining power of the energy storage system at the n-1 reference point, and Δt is the time difference between time n and time n-1.

[0101] Energy storage system battery health status:

[0102] Since the energy storage system is always in the process of charging and discharging, the battery health status of the energy storage system must be tested. There are many definitions of the battery health status. The most common definition method is to define it from the perspective of capacity:

[0103]

[0104] In the formula, Q max (n) is the current maximum discharge capacity of the battery, and C is the capacity of the battery when it leaves the factory.

[0105] Evaluation indicators:

[0106] This experiment consists of daily root mean square error RMSE, mean absolute error MAE, and maximum tracking error δ max The experimental results are evaluated by four error indicators: tracking accuracy Acc', and the formulas are:

[0107]

[0108]

[0109] Where n is the number of points tracked by rolling optimization. This experiment has 96 points per day. i is the actual value of the output at each moment, is the average output at n moments, P lower_limit (n), P upper_limit (n) and P adjust (n) are the lower and upper limits of real-time output fluctuations, and the coordinated output of the optimized wind-storage combined system. The specific evaluation results will be pointed out in the test verification and analyzed.

[0110] Verification test:

[0111] In order to verify the proposed ultra-short-term wind power output prediction model implemented by LSTM neural network, actual data is used to control the operation model. Figure 4 As shown in the figure, the input data of wind power output is generally given by the wind power dispatching center. The experiment obtains one data point every 15 minutes. The experiment is set to the wind power output of the previous 7 days, with a total of 7×96=672 wind power data. The output data is 96 wind power data for 1 day, which are generally the predicted optimal output.

[0112] Next, the wind-storage joint system optimization strategy is coordinated and the planned output and actual output are comprehensively considered. After the upper-level optimization, the tracking planned output curve is obtained as follows: Figure 5As shown in the figure, the red dotted line is the planned output curve of the wind farm, and the shaded part in the figure is the upper and lower limits of the planned output after the deviation coefficient is introduced. According to the relevant literature on national policies and wind farm economic benefits, it can be concluded that the economic benefits can be optimized by default within the planned deviation range. At the same time, the blue curve is the actual output given by the wind power station on the same day, with a rolling interval of 4 hours, and the real-time output is continuously input to the dispatching center. The optimized green curve is the output curve after energy storage participates in tracking optimization. The specific tracking optimization process has been given in the previous article. It can be seen that after energy storage optimization, most points have entered the optimal planned value range, but due to the capacity constraints of the energy storage itself, the regulation capacity is limited. Finally, the black curve is the wind-storage output after leveling, and the number of points within the range has risen sharply.

[0113] Figure 6 It is a curve that tracks the output of each individual in the planned output process. Figure 5 A detailed interpretation of the graph is given in Figure 1. The first sub-picture is a comparison of the two inputs. It can be seen that there is a large deviation in the data before the input. The second sub-picture shows the actual action of energy storage participating in the output. The third picture is the value after tracking optimization, which is more intuitive.

[0114] Figure 7 It is to evaluate the energy storage status. We can see the charging and discharging process curve of the energy storage. As the tangent of the SOC curve changes, it means the state switching between charging / discharging, that is, the number of charging and discharging times, which has a great impact on the life and health of the energy storage. It can be seen that the change of energy storage is very small, and in the case of short-term tracking, the life of this energy storage is very long, and its loss can often be ignored.

[0115] Finally, some parameters of energy storage are not shown in the figure. Among them, the indicators representing the error category are generally low, which shows that the invented algorithm has good stability; the accuracy rate representing accuracy is generally high, above 90%, and most points are within the planned range, which shows that the algorithm has good accuracy.

[0116] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in other described embodiments.

Claims

1. A method for tracking planned output of a wind-storage combined system, characterized in that: The method comprises: S1. Get the optimal planned output P plan (n+k), n represents the current sampling time, k is a positive integer; S2. Determine the real-time output P of the wind farm actual (n) Whether it is in the energy storage fluctuation range [P lower_limit (n),P upper_limit (n)], if yes, go to S5, if no, go to S3; P upper_limit (n), P lower_limit (n) are the upper and lower limits constrained by energy storage fluctuations; S3, when the wind farm outputs real-time power P actual (n) greater than P upper_limit (n), the energy storage system is charged and the process goes to S4, otherwise the energy storage system is discharged and the process goes to S4; S4. Construct the objective function for smoothing fluctuations and energy storage constraints. After solving the objective function under the energy storage constraints, the real-time wind power output P after smoothing fluctuations is obtained. actual' (n), the objective function of smoothing fluctuations is: maxαAcc-βCDc Among them, the accuracy of energy storage in smoothing fluctuations Energy storage charge and discharge times ratio P actual' (n) is the real-time wind power output after smoothing fluctuations, α and β are weight coefficients, Indicates the rate of change of the total amount of energy storage; S5. Establish a prediction model to obtain the real-time predicted wind power output P at time n+k adjust (n+k|n); S6. According to the obtained P adjust (n+k|n), with the goal of minimizing the tracking error and the energy storage output at the time n+k, the wind power compensation objective function and energy storage constraint conditions are established, and the wind power compensation objective function is solved under the energy storage constraint conditions to obtain the output P after wind power compensation adjust (n); The objective function of wind power compensation is: Where N represents the total number of sampling moments, ΔE BESS (n+k|n) is the predicted output of energy storage at time n+k, and γ is the proportion of energy storage output in the objective function; S7, output power P after wind power compensation adjust (n).

2. The method for tracking planned output of a wind-storage combined system according to claim 1, characterized in that: In S5, the state of charge SOC(n) of the energy storage system is used as the state variable x(n) of the control system, and the charging and discharging power ΔE of the energy storage system is BESS (n) is used as the control variable u(n), and the output P after wind power compensation is adjust (n) is the output of the system y(n), and the wind power prediction error ε(n) = P actual' (n)-P plan (n) is used as the disturbance r(n), a prediction model is established, and the MPC control method is adopted to obtain the state equation of the output y(n) at time n+k: Among them, SOC(n+k|n) and SOC(n+k-1|n) are the predicted values ​​of the energy storage charge state at time n+k and time n+k-1 respectively.

3. The method for tracking planned output of a wind-storage combined system according to claim 1, characterized in that: In S6, the objective function and constraints are solved by quadratic programming, and the objective function is converted into: Where u is the optimization variable vector, including ΔE at all sampling times BESS (n), H is the coefficient matrix of the quadratic term, and f is the coefficient vector of the linear term.

4. The method for tracking planned output of a wind-storage combined system according to claim 1, characterized in that: Get the optimal planned output P plan (n+k) method: A training set of a wind power prediction model is constructed based on historical data of wind power stations. The input data of each training sample in the training set includes the output value of the wind power station at each sampling point within any continuous M hours, and the output data includes the optimal planned output at the kth moment after M consecutive hours. Using the training set to train the wind power prediction model, and obtaining a trained wind power prediction model; The wind farm real-time output P at the current moment actual (n) is input into the wind power prediction model, and the wind power prediction model obtains the optimal planned output P plan (n+k), the wind power prediction model is a deep learning network.

5. The method for tracking planned output of a wind-storage combined system according to claim 1, characterized in that: The energy storage constraints in S4 are: Where, ΔE BESS (n) is the change in charging and discharging power of the energy storage system, E BESS (n) is the energy storage power, ΔE BESS,max , ΔE BESS,min ΔE BESS The upper and lower limits of (n), E BESS,max 、E BESS,min For E BESS The upper and lower limits of (n).

6. The method for tracking planned output of a wind-storage combined system according to claim 1, characterized in that: The energy storage constraint in S6 is: where ΔE BESS (n) is the change in charging and discharging power of the energy storage system, E BESS (n) is the energy storage power, ΔE BESS,max , ΔE BESS,min ΔE BESS The upper and lower limits of (n), E BESS,max 、E BESS,min For E BESS The upper and lower limits of (n).

7. The method for tracking planned output of a wind-storage combined system according to claim 1, characterized in that: During the charging and discharging process of the energy storage system in S3, the charging power P at time n c (n) and discharge power P f (n) respectively satisfy: Where P c,max and P f,max are the maximum charging and discharging power of the energy storage system, σ is the self-discharge rate of the energy storage system, η c , η d are the charging and discharging efficiency of the energy storage system, E BESS,max is the maximum capacity of the energy storage system, E BESS,min is the minimum capacity of the energy storage system, ΔE BESS (n) is the change in charging and discharging power of the energy storage system at time n-1, and Δt is the duration between two adjacent times.

8. A computer-readable storage device storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for tracking planned output of a wind-storage combined system according to any one of claims 1 to 7 are implemented.

9. A device for tracking planned output of a wind-storage combined system, comprising a storage device, a processor, and a computer program stored in the storage device and executable on the processor, characterized in that: The processor executes the computer program to implement the steps of the method for tracking planned output of a wind-storage integrated system according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for tracking planned output of a wind-storage combined system as claimed in any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Power prediction method, system and equipment based on wind power plant super-capacity energy storage and medium

    CN120150133A

  • Power Prediction Method, System, Device and Medium Based on Supercapacitor Energy Storage in Wind Farm

    CN120150133B

  • Energy storage coordination management method, system and equipment for wind power generation power control and medium

    CN120749844A