Power Optimization Allocation Control Method Based on Wind-Storage Combined Frequency Regulation System
By constructing a multi-objective optimization model and an online correction mechanism, the power of wind power and energy storage is dynamically allocated, which solves the problem of wind power fluctuation and energy storage response uncertainty in wind-storage joint frequency regulation, realizes precise collaborative optimization of wind power and energy storage, and improves grid frequency stability and resource utilization efficiency.
Patent Information
- Application Number
- CN202610368689.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-25
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2046-03-25
AI Technical Summary
Existing wind-storage joint frequency regulation control methods are unable to overcome the dual time-varying uncertainties of wind power fluctuation characteristics and energy storage response capabilities, resulting in the inability of power allocation strategies to achieve precise coordination and adaptive optimization, leading to problems such as overcharging and over-discharging of energy storage accelerating aging and wasting wind power regulation capabilities.
By synchronously acquiring grid frequency regulation commands, real-time wind farm output data, and energy storage system state of charge and adjustable power capacity data, volatility analysis and prediction error modeling are performed. The characteristics of wind power regulation capability and energy storage regulation margin are extracted, a multi-objective optimization model is constructed, the power of wind power and energy storage is dynamically allocated, the optimal allocation coefficient is determined, and the weight coefficients and constraint boundaries of the optimization model are adjusted through an online feedback correction mechanism.
It achieves optimized allocation of wind power and energy storage power while meeting the grid frequency regulation requirements, avoids energy storage aging, makes full use of wind power regulation capabilities, improves frequency regulation tracking accuracy and system stability, reduces response deviation, and enhances the grid frequency fluctuation support capability.
Smart Images

Figure CN121965539B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system frequency control technology, specifically to a power optimization allocation control method based on a wind-storage combined frequency regulation system. Background Technology
[0002] With the increasing penetration rate of wind power, the inertia of power systems is decreasing, and frequency stability issues are becoming increasingly prominent. Wind-storage combined frequency regulation, as an effective solution, compensates for the insufficient frequency regulation capabilities of wind power by leveraging the rapid response characteristics of energy storage systems, and has been widely adopted. Existing wind-storage combined frequency regulation control methods typically employ power allocation strategies based on fixed ratios or simple rules, such as allocating frequency regulation power in segments according to the state of charge of energy storage, or compensating for wind power prediction errors.
[0003] The existing technology has the following shortcomings: In the process of wind and energy storage joint frequency regulation, how to overcome the dual time-varying uncertainties of wind power fluctuation characteristics and energy storage response capabilities, and achieve precise coordination and adaptive optimization of power allocation strategies, is crucial to avoid the contradiction between the overcharging and over-discharging of energy storage leading to accelerated aging and the waste of wind power regulation capacity caused by traditional fixed-ratio allocation methods. Specifically, existing methods struggle to dynamically adjust the power carried by wind power based on the real-time fluctuation frequency and amplitude of wind power, and also cannot constrain the energy storage output depth by combining the current state of charge and response rate of energy storage. This results in the inability of the two to achieve an optimal state of coordination while meeting the grid frequency regulation requirements. Summary of the Invention
[0004] The purpose of this invention is to provide a power optimization allocation control method based on a wind-storage combined frequency regulation system to solve the problems mentioned above.
[0005] The objective of this invention can be achieved through the following technical solutions: The power optimization allocation control method based on the wind-storage combined frequency regulation system includes the following steps: S1, synchronously acquire the grid frequency regulation command sequence, the real-time power output data sequence of the wind farm, and the real-time state of charge and adjustable power capacity data sequence of the energy storage system; S2 performs volatility analysis and prediction error modeling on real-time power output data sequences to extract wind power regulation capability characteristics; and performs charge / discharge depth and response rate analysis on state of charge and adjustable power capacity data sequences to extract energy storage regulation margin characteristics. S3. Using the characteristics of wind power regulation capability and energy storage regulation margin as state constraints, and combining the grid frequency regulation command sequence, a multi-objective optimization model is constructed with the goal of maximizing frequency regulation tracking accuracy and minimizing energy storage lifetime loss. The multi-objective optimization model is then solved to obtain the optimal power allocation coefficient between wind power and energy storage. S4, based on the optimal power allocation coefficient, generates power adjustment commands for the wind farm and charging / discharging control commands for the energy storage system, and issues them for execution; S5 collects the actual frequency modulation response data after the execution command, calculates the deviation between the actual frequency modulation effect and the frequency modulation command, and corrects the objective function and constraints of the multi-objective optimization model online based on the deviation.
[0006] As a further aspect of the present invention: S2 specifically includes: Variational mode decomposition was performed on the real-time power output data sequence to obtain multiple modal components. The modal component with the highest energy proportion was selected as the main wave mode sequence. The Hilbert transform is performed on the main wave mode sequence to obtain the instantaneous frequency and instantaneous amplitude. The variance of the instantaneous frequency is calculated as the wind power wave frequency characteristic, and the mean of the instantaneous amplitude is calculated as the wind power wave amplitude characteristic, thus constituting the wind power regulation capability characteristic. Based on the characteristics of wind power fluctuation frequency, the length of the sliding window is set, and the extreme points of the state of charge sequence are detected within the sliding window. The change in state of charge between adjacent extreme points and the corresponding time interval are obtained, and the charging and discharging depth and response rate are obtained. Based on the characteristics of wind power fluctuation amplitude and combined with the current adjustable power capacity value, the proportion of power capacity required by energy storage is calculated to obtain the adjustable power capacity coefficient, which together with the depth of charge and discharge and the response rate constitutes the energy storage regulation margin characteristics.
[0007] As a further aspect of the present invention: the calculation process of the adjustable power capacity coefficient is as follows: The initial proportional coefficient is obtained based on the ratio of the wind power fluctuation amplitude characteristics to the preset benchmark amplitude; Based on the difference between the wind power fluctuation frequency characteristics and the rated response frequency of energy storage, the preset frequency response attenuation correspondence is queried to obtain the response attenuation coefficient. Multiply the initial proportional gain by the response decay coefficient to obtain the corrected proportional gain; Based on the ratio of the current adjustable power capacity to the rated power of the energy storage, the correction ratio coefficient is limited to obtain the adjustable power capacity coefficient.
[0008] As a further aspect of the present invention: the calculation process of the response attenuation coefficient is as follows: Calculate the absolute value of the difference between the wind power fluctuation frequency characteristic and the rated response frequency of energy storage, and use it as the frequency deviation. Based on the frequency deviation, the corresponding attenuation coefficient is found in the pre-stored frequency deviation and attenuation coefficient mapping table. The mapping table contains multiple frequency deviation ranges and the reference attenuation coefficient corresponding to each range. If the frequency deviation falls between two adjacent frequency deviation intervals, then the reference attenuation coefficients corresponding to these two intervals are linearly interpolated to obtain the interpolated attenuation coefficients. The found baseline attenuation coefficient or the attenuation coefficient calculated by interpolation is output as the response attenuation coefficient.
[0009] As a further aspect of the present invention: S3 specifically includes: Based on the fluctuation frequency and fluctuation amplitude characteristics in the wind power regulation capability characteristics, the frequency band of the power grid frequency regulation command sequence is segmented, and the command component that matches the wind power fluctuation characteristics is extracted as the wind power tracking target sequence. Based on the response rate and adjustable power capacity coefficient in the energy storage regulation margin characteristics, the grid frequency regulation command sequence is subjected to amplitude limiting processing to obtain the energy storage compensation target sequence. The first optimization objective is to minimize the deviation between the wind power tracking target sequence and the actual wind power output, and the second optimization objective is to minimize the deviation between the energy storage compensation target sequence and the actual energy storage output. A multi-objective optimization model is constructed using the upper limit of the fluctuation amplitude in the wind power regulation capability characteristics and the upper limit of the charge and discharge depth in the energy storage regulation margin characteristics as constraint boundaries. To solve the multi-objective optimization model: First, the Chebyshev aggregation method is used to aggregate the first and second optimization objectives into a single objective. Then, the golden section search method is used to iteratively optimize within the constraint boundary to obtain the power allocation coefficient that minimizes the aggregation objective, which is taken as the optimal power allocation coefficient.
[0010] As a further aspect of the present invention: solving the multi-objective optimization model specifically includes: Based on the positional relationship between the current state of charge value in the energy storage regulation margin characteristics and the preset ideal state of charge range, the first weighting coefficient of the first optimization objective and the second weighting coefficient of the second optimization objective are determined. Calculate the first deviation between the first optimization objective and its preset ideal value, and the second deviation between the second optimization objective and its preset ideal value. Take the maximum value between the product of the first deviation and the first weight coefficient and the product of the second deviation and the second weight coefficient as the Chebyshev aggregation objective. Within the search interval formed by the constraint boundary, the first and second trial points are selected according to the golden ratio. The Chebyshev aggregation target values corresponding to the two trial points are calculated respectively. The search interval is narrowed based on the comparison results of the two aggregation target values. The process continues until the length of the search interval is less than the preset convergence threshold. Then, the power allocation coefficient corresponding to the midpoint of the current search interval is taken as the optimal power allocation coefficient.
[0011] As a further aspect of the present invention: S4 specifically includes: The power values to be undertaken by wind power and energy storage are obtained by multiplying the optimal power allocation coefficient with the power demand of the current grid frequency regulation command. Based on the fluctuation frequency characteristics in the wind power regulation capability characteristics, the wind power load value is smoothed on a time scale to obtain the wind power adjustment command. Based on the response rate in the energy storage regulation margin characteristics, the change rate of the energy storage power value is limited to obtain the energy storage charging and discharging power command. The wind power adjustment command and the energy storage charging and discharging power command are sent synchronously to the wind farm controller and the energy storage system controller for execution.
[0012] As a further aspect of the present invention: obtaining the wind power adjustment command specifically includes: Based on the instantaneous frequency in the characteristics of wind power regulation capability, calculate the fluctuation period value at the current moment, and use the fluctuation period value as the initial length of the smoothing time window; Multiple historical sampling points of wind power load value are collected within the smoothing time window. After removing the maximum and minimum values, the arithmetic mean of the remaining sampling points is taken to obtain the smoothed power reference value at the current moment. Based on the deviation between the smoothed power benchmark value and the wind power carrying capacity value, and combined with the preset ramp rate limit, the correction is made, and the corrected value is used as the wind power adjustment command at the current moment.
[0013] As a further aspect of the present invention: S5 specifically includes: The actual power output of the wind farm and the actual power output of the energy storage system after the execution command are collected and compared with the grid frequency regulation command sequence to obtain the wind power tracking deviation value and the energy storage power tracking deviation value. The wind power tracking deviation value is compared with a preset first deviation threshold. If it exceeds the first deviation threshold, the weight coefficient of the first optimization objective in the multi-objective optimization model is increased. The energy storage power tracking deviation value is compared with the preset second deviation threshold. If it exceeds the second deviation threshold, the constraint boundary value of the upper limit of the energy storage charging and discharging depth in the multi-objective optimization model is reduced. The current state of charge value in the energy storage regulation margin feature is compared with the preset overcharge and over-discharge warning value. If the warning value is triggered, the weight coefficients of the first optimization objective and the second optimization objective are adjusted simultaneously.
[0014] The beneficial effects of this invention are: (1) This invention extracts the frequency and amplitude characteristics of wind power fluctuations and the state of charge and response rate characteristics of energy storage in real time, and constructs a multi-objective optimization model with frequency regulation tracking accuracy and energy storage lifespan loss as dual objectives. Under the premise of meeting the grid frequency regulation requirements, the power carried by wind power and energy storage is dynamically allocated to avoid accelerated aging of energy storage due to overcharging and discharging, while making full use of the regulation capacity of wind power to reduce wind curtailment. This reduces the energy storage cycle loss, improves the wind power absorption capacity, and achieves a balance between efficient resource utilization and stable system operation.
[0015] (2) This invention uses an online feedback correction mechanism to collect frequency regulation response data in real time and compare it with the command deviation, dynamically adjusting the weight coefficients and constraint boundary values in the optimization model. When the wind power tracking deviation exceeds the limit, the constraint weight is automatically increased; when the energy storage approaches the overcharge / overdischarge warning value, the target weight is simultaneously corrected, enabling the control strategy to adapt to changes in wind power fluctuation characteristics and time-varying energy storage status. This adaptive mechanism significantly improves the frequency regulation tracking accuracy under complex operating conditions, reduces response deviation, and enhances the wind-storage integrated system's support capability against grid frequency fluctuations. Attached Figure Description
[0016] The invention will now be further described with reference to the accompanying drawings.
[0017] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation
[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0019] Please see Figure 1 As shown, this invention is a power optimization allocation control method based on a wind-storage combined frequency regulation system, comprising the following steps: S1, synchronously acquire the grid frequency regulation command sequence, the real-time power output data sequence of the wind farm, and the real-time state of charge and adjustable power capacity data sequence of the energy storage system; S2 performs volatility analysis and prediction error modeling on real-time power output data sequences to extract wind power regulation capability characteristics; and performs charge / discharge depth and response rate analysis on state of charge and adjustable power capacity data sequences to extract energy storage regulation margin characteristics. S3. Using the characteristics of wind power regulation capability and energy storage regulation margin as state constraints, and combining the grid frequency regulation command sequence, a multi-objective optimization model is constructed with the goal of maximizing frequency regulation tracking accuracy and minimizing energy storage lifetime loss. The multi-objective optimization model is then solved to obtain the optimal power allocation coefficient between wind power and energy storage. S4, based on the optimal power allocation coefficient, generates power adjustment commands for the wind farm and charging / discharging control commands for the energy storage system, and issues them for execution; S5 collects the actual frequency modulation response data after the execution command, calculates the deviation between the actual frequency modulation effect and the frequency modulation command, and corrects the objective function and constraints of the multi-objective optimization model online based on the deviation.
[0020] In S1, the grid frequency regulation command sequence, the real-time output data sequence of the wind farm, and the real-time state of charge and adjustable power capacity data sequence of the energy storage system are acquired synchronously, specifically including: First, the remote communication device within the wind farm's substation receives real-time automatic power generation control commands from the power grid dispatch center. These commands are then arranged chronologically to form a power grid frequency regulation command sequence. Simultaneously, data acquisition and monitoring control system terminals deployed at each wind turbine tower base read the instantaneous active power value of each wind turbine using industrial Ethernet communication at a preset sampling period. The summation of the instantaneous active power values of all turbines is then arranged chronologically to obtain the wind farm's real-time output data sequence. Furthermore, the energy management controller of the energy storage system, using the controller area network bus protocol, reads the current state of charge percentage (SOP) value from the energy storage converter and the current maximum allowable charging power and maximum discharge power values from the battery management system. The SOP values are arranged chronologically to obtain a real-time SOP data sequence, and the maximum charging power and maximum discharge power values are arranged chronologically to obtain an adjustable power capacity data sequence. All four sequences were tagged with a unified time tag during data acquisition to ensure time synchronization between sequences and were stored in a real-time database for subsequent steps.
[0021] In S2, volatility analysis and prediction error modeling are performed on the real-time power output data sequence to extract wind power regulation capability characteristics; charge / discharge depth and response rate analysis are performed on the state of charge and adjustable power capacity data sequence to extract energy storage regulation margin characteristics, specifically including: First, a volatility analysis is performed on the real-time power output data sequence. The real-time power output data sequence is input into the iterative solution process of variational mode decomposition. The center frequency and bandwidth of each modal component are searched cyclically using the alternating direction multiplier method until a preset convergence accuracy condition is met, decomposing the original sequence into multiple modal components arranged from low to high frequency. The energy of each modal component is calculated by squaring the values at each sampling point of that component and summing them. The modal component with the highest percentage of energy value relative to the total energy of all modal components is selected as the main volatility mode sequence characterizing the main volatility characteristics of wind power.
[0022] Then, a Hilbert transform is performed on the main wave mode sequence. The main wave mode sequence is taken as the real part signal, and its Hilbert transform result is calculated through convolution as the imaginary part signal. An analytic signal is constructed from the real and imaginary parts. For each sampling time point, the change in the amplitude of the analytic signal per unit time is calculated as the instantaneous frequency at that time, and the magnitude of the analytic signal is calculated as the instantaneous amplitude at that time. After traversing the entire sequence, an instantaneous frequency sequence and an instantaneous amplitude sequence are obtained. The arithmetic mean of the squares of the deviations of each element in the instantaneous frequency sequence from the arithmetic mean of the sequence is calculated, i.e., the variance, and this variance value is used as the wind power fluctuation frequency characteristic. The arithmetic mean of each element in the instantaneous amplitude sequence is calculated, and this average value is used as the wind power fluctuation amplitude characteristic. The wind power fluctuation frequency characteristic and the wind power fluctuation amplitude characteristic are combined to form the wind power regulation capability characteristic.
[0023] Next, charge / discharge depth analysis is performed on the state of charge (POC) data sequence. Based on the numerical value of the wind power fluctuation frequency characteristic, its reciprocal is converted to time dimension. This time dimension value is multiplied by a preset sampling period coefficient to obtain the length of the sliding window. On the POC data sequence, the current window is defined starting from the beginning of the sequence according to the sliding window length. Maximum and minimum values are found within the window, and the position indices of all maximum and minimum values and their corresponding POC values are recorded. The absolute value of the difference in POC values between adjacent extreme points (whether maximum or minimum) is calculated sequentially as the charge / discharge depth value for one charge / discharge cycle. Simultaneously, the number of sampling points between adjacent extreme points is calculated and multiplied by the sampling period to obtain the response time interval corresponding to that charge / discharge cycle. After traversing the entire sequence, the charge / discharge depth sequence and the response time interval sequence are obtained. These two sequences are used as the basic data for charge / discharge depth and response rate in the energy storage regulation margin characteristics.
[0024] Finally, margin coefficients are calculated for the adjustable power capacity data sequence. The characteristic value of the wind power fluctuation amplitude is taken and divided by a pre-set reference amplitude value to obtain the initial proportional coefficient. The reference amplitude value is a fixed constant determined based on the statistical value of the maximum allowable fluctuation amplitude under frequency compliance conditions in the historical operating data of the wind farm. The absolute value of the difference between the wind power fluctuation frequency characteristic and the rated response frequency of the energy storage system is calculated, and this absolute value is used as the frequency deviation. A frequency deviation-attenuation coefficient mapping table is pre-constructed. This mapping table is obtained through experimental calibration: a unit step power command is applied to the energy storage system under different frequency deviations, and the time required for the actual output to reach 90% of the command value is measured. The ratio of the reciprocal of this time to the theoretical response frequency is used as the attenuation coefficient. The frequency deviation is divided into intervals of 0.05 Hz, and each interval corresponds to a reference value for the attenuation coefficient. The mapping table is consulted based on the frequency deviation: if the frequency deviation falls exactly within a certain interval, the attenuation coefficient benchmark value corresponding to that interval is directly taken; if the frequency deviation is between the boundaries of two adjacent intervals, the attenuation coefficient benchmark values corresponding to these two intervals are taken, and a linear weighted calculation is performed according to the distance ratio between the frequency deviation and the center values of the two intervals to obtain the interpolated attenuation coefficient. The initial proportional coefficient is multiplied by the attenuation coefficient to obtain the corrected proportional coefficient. The adjustable power capacity value at the current moment is read, that is, the absolute value of the maximum allowable charge and discharge power of the energy storage system, and divided by the rated power value of the energy storage system to obtain a capacity proportional factor between 0 and 1. The corrected proportional coefficient is compared with the capacity proportional factor: if the corrected proportional coefficient is less than the capacity proportional factor, the corrected proportional coefficient is taken as the adjustable power capacity coefficient; if the corrected proportional coefficient is greater than or equal to the capacity proportional factor, the capacity proportional factor is taken as the adjustable power capacity coefficient. The adjustable power capacity coefficient is combined with the aforementioned charge and discharge depth sequence and response time interval sequence to form a complete energy storage regulation margin feature.
[0025] In S3, the characteristics of wind power regulation capability and energy storage regulation margin are used as state constraints. Combined with the grid frequency regulation command sequence, a multi-objective optimization model is constructed with the objectives of maximizing frequency regulation tracking accuracy and minimizing energy storage lifetime loss. Solving this multi-objective optimization model yields the optimal power allocation coefficient between wind power and energy storage, specifically including: First, a wind power tracking target sequence is constructed. Based on the wind power fluctuation frequency characteristic value in the wind power regulation capability characteristics, a digital low-pass filter is designed. The passband cutoff frequency of this filter is set to half of the wind power fluctuation frequency characteristic value, and the stopband start frequency is set to twice the wind power fluctuation frequency characteristic value. The grid frequency regulation command sequence is input into this digital low-pass filter for convolution operation. The output sequence after filtering out high-frequency components is the wind power tracking target sequence that matches the wind power fluctuation characteristics. Simultaneously, based on the wind power fluctuation amplitude characteristic value, the output amplitude limit of this filter is set to ensure that the fluctuation amplitude of the output sequence does not exceed 1.2 times the wind power fluctuation amplitude characteristic value.
[0026] Secondly, an energy storage compensation target sequence is constructed. The response rate value in the energy storage regulation margin characteristics, i.e., the minimum value in the response time interval sequence, is read, and its reciprocal is taken as the maximum trackable frequency of the energy storage system. Based on the adjustable power capacity coefficient, the current allowable output limit of the energy storage system is calculated, i.e., the rated energy storage power multiplied by the adjustable power capacity coefficient. The grid frequency regulation command sequence is subtracted point-by-point from the wind power tracking target sequence to obtain a difference sequence. This difference sequence is then subjected to amplitude limiting processing: if the difference at a certain moment is greater than the current allowable output limit, the difference at that moment is replaced with the current allowable output limit; if the difference is less than a negative value of the current allowable output limit, it is replaced with a negative value of the current allowable output limit. The amplitude-limited sequence serves as the energy storage compensation target sequence.
[0027] Then, a multi-objective optimization model is constructed. The first optimization objective is to minimize the cumulative value of the sum of squares of the difference between the wind power tracking target sequence and the actual wind power output over the entire time window. The second optimization objective is to minimize the cumulative value of the sum of squares of the difference between the energy storage compensation target sequence and the actual energy storage output over the entire time window. An upper limit value of the fluctuation amplitude is extracted from the wind power regulation capability characteristics, which is 1.5 times the wind power fluctuation amplitude characteristic, as the maximum allowable fluctuation constraint boundary for wind power output. An upper limit value of the charge / discharge depth is extracted from the energy storage regulation margin characteristics, which is 1.2 times the maximum value in the charge / discharge depth sequence, as the maximum allowable depth constraint boundary for a single charge / discharge cycle of energy storage. These two constraint boundaries and the two optimization objectives together constitute the multi-objective optimization model.
[0028] Next, the Chebyshev aggregation weights are determined. The current state of charge (SOC) value in the energy storage regulation margin characteristic is read and compared with a preset ideal SOC range. The ideal SOC range is 20% to 80%. If the current SOC value is less than 20%, the first weight coefficient of the first optimization objective is set to 2.0, and the second weight coefficient of the second optimization objective is set to 1.0; if the current SOC value is greater than 80%, the first weight coefficient is set to 1.0, and the second weight coefficient is set to 2.0; if the current SOC value is between 20% and 80%, both the first and second weight coefficients are set to 1.5.
[0029] Then, the Chebyshev aggregation target is calculated. The preset ideal value for the first optimization target is set to zero, and the preset ideal value for the second optimization target is also set to 0. For any candidate power allocation coefficient value in the current iteration, the first deviation value between the actual wind power output and the wind power tracking target sequence under that candidate value is calculated; this is the calculation result of the first optimization target. The second deviation value between the actual energy storage output and the energy storage compensation target sequence under that candidate value is also calculated; this is the calculation result of the second optimization target. The product of the first deviation value and the first weighting coefficient is taken as the first weighted deviation, and the product of the second deviation value and the second weighting coefficient is taken as the second weighted deviation. The larger of the first weighted deviation and the second weighted deviation is taken as the Chebyshev aggregation target value corresponding to the current power allocation coefficient.
[0030] Finally, the golden ratio search method is used for optimization. The maximum allowable fluctuation boundary of wind power output and the maximum allowable depth boundary of energy storage are used as the lower and upper limits of the search interval for the power allocation coefficient. Within the search interval, the point at 0.618 of the golden ratio is selected as the first trial point, and the point at 0.382 of the golden ratio is selected as the second trial point. The Chebyshev aggregation target values corresponding to the first and second trial points are calculated respectively. The two aggregation target values are compared: if the aggregation target value of the first trial point is less than that of the second trial point, the portion from the second trial point to the upper limit in the search interval is discarded, and the upper limit is updated to the second trial point; otherwise, the portion from the lower limit to the first trial point is discarded, and the lower limit is updated to the first trial point. Trial points are selected again according to the golden ratio in the new search interval and compared, and this process is repeated iteratively. After each iteration, the length of the current search interval is calculated, which is the upper limit value minus the lower limit value. The iteration stops when the length of the search interval is less than a preset convergence threshold, which is set to one-hundredth of the original search interval length. Take the midpoint value of the search interval obtained in the last iteration, which is the arithmetic mean of the upper and lower limits, and output this midpoint value as the optimal power allocation coefficient.
[0031] In S4, based on the optimal power allocation coefficient, power adjustment commands for the wind farm and charging / discharging control commands for the energy storage system are generated and executed. Specifically, this includes: First, calculate the power contribution value of wind power and energy storage. Read the current grid frequency regulation command demand power, which is the active power value that the wind-storage combined system is required to increase or decrease in the automatic generation control command issued by the dispatch center. Multiply the optimal power allocation coefficient obtained in step S3 by the grid frequency regulation command demand power, and record the product as the wind power contribution value. Simultaneously, subtract the wind power contribution value from the grid frequency regulation command demand power, and record the difference as the energy storage contribution value. If the difference is positive, it indicates that the energy storage system needs to discharge to provide power support; if it is negative, it indicates that the energy storage system needs to charge to absorb excess power.
[0032] Next, a wind power adjustment command is generated. Based on the instantaneous frequency value in the wind power regulation capability characteristics (obtained in step S2 after performing a Hilbert transform on the main fluctuation mode sequence), the reciprocal of this instantaneous frequency value is used to calculate the fluctuation period value at the current moment, with the unit being seconds. This fluctuation period value is used as the initial length of the smoothing time window. Using the current moment as a reference, a period equal to the smoothing time window length is traced back, and the wind power load values at all sampling moments within this period are collected. These sampling values constitute a historical dataset. The sampling period is a pre-set fixed value, such as 100 milliseconds. The number of data points in this historical dataset is counted and recorded as the window sampling point count. All wind power load values in the historical dataset are sorted, and the maximum and minimum values are identified and removed from the dataset. The arithmetic mean of the remaining window sampling point count minus two data points after removing the maximum and minimum values is calculated. This summation is divided by the window sampling point count minus two, and the result is used as the smoothed power reference value for the current moment. Then, the deviation between the smoothed power reference value and the wind power load value at the current moment is calculated, i.e., the smoothed power reference value minus the wind power load value, and the absolute value of the deviation is taken. A preset wind farm ramp rate limit is read; this limit is the maximum allowable power change per unit time, preset based on the wind turbine's mechanical load capacity, in megawatts per second. The absolute value of the deviation is compared with the ramp rate limit: if the absolute value of the deviation is less than or equal to the ramp rate limit, the wind power adjustment command at the current moment is set to the smoothed power reference value; if the absolute value of the deviation is greater than the ramp rate limit, correction is made according to the direction of the deviation. When the smoothed power reference value is greater than the wind power load value, the wind power load value plus the ramp rate limit is used as the wind power adjustment command; when the smoothed power reference value is less than the wind power load value, the wind power load value minus the ramp rate limit is used as the wind power adjustment command. This correction process ensures that the wind power change rate never exceeds the turbine's safety limits.
[0033] Next, an energy storage charging / discharging power command is generated. The response rate in the energy storage regulation margin characteristic is read. This response rate is the rate of change corresponding to the minimum value in the response time interval sequence obtained in step S2, i.e., the maximum change capability of the energy storage power per unit time, measured in megawatts per second. The energy storage power value is compared with the actual energy storage output value at the previous moment to obtain the power change demand. This power change demand is divided by the response rate to obtain the theoretically required change time. If this change time is greater than a preset control period, the change rate of the energy storage power value needs to be limited: the energy storage charging / discharging power command at the current moment should be set to the actual energy storage output value at the previous moment plus or minus the product of the response rate and the control period, where the control period is the time interval between command issuance and execution, for example, 100 milliseconds. If the change time obtained by dividing the power change demand by the response rate is less than or equal to the control period, the energy storage power value is directly used as the energy storage charging / discharging power command at the current moment. The above processing ensures that the rate of change of power commands in the energy storage system does not exceed its actual response capability, thus avoiding tracking failure or equipment impact due to excessively rapid command changes.
[0034] Finally, the generated wind power adjustment commands and energy storage charging / discharging power commands are time-synchronized and sent to the wind farm controller via the remote communication device on the wind farm side using the industrial Ethernet protocol. Simultaneously, they are sent to each energy storage converter via the energy management controller of the energy storage system using the controller area network bus protocol. Upon receiving the commands, the wind farm controller and energy storage converters immediately execute power adjustments in the next control cycle, completing the power allocation and control of the wind-storage joint frequency regulation.
[0035] In S5, the actual frequency modulation response data after the execution command is collected, the deviation between the actual frequency modulation effect and the frequency modulation command is calculated, and the objective function and constraints of the multi-objective optimization model are corrected online based on the deviation. Specifically, this includes: First, the actual frequency regulation response data after the execution command is collected. Through the wind farm data acquisition and monitoring control system terminal, the actual active power value of each wind turbine at the end of the current control cycle is read and accumulated to obtain the actual output value of the wind farm. Through the energy management controller of the energy storage system, the actual output power value of each energy storage converter at the current moment is read and accumulated to obtain the actual output value of the energy storage system. These two actual output values are compared with the corresponding command values in the grid frequency regulation command sequence at the same moment in step S1: the difference between the actual wind power output value and the corresponding value in the wind power tracking target sequence is calculated, and its absolute value is taken as the wind power tracking deviation value; the difference between the actual energy storage output value and the corresponding value in the energy storage compensation target sequence is calculated, and its absolute value is taken as the energy storage power tracking deviation value. All deviation values are in megawatts.
[0036] Secondly, the weight coefficients of the first optimization objective are adjusted online based on the wind power tracking deviation value. A first deviation threshold is preset, which is determined statistically based on the maximum tracking deviation allowed when the frequency regulation compliance rate is not less than 95% in the historical operating data of the wind farm, for example, set as 2% of the wind farm's rated power. The current wind power tracking deviation value is compared with the first deviation threshold: if the wind power tracking deviation value is less than or equal to the first deviation threshold, the current weight coefficient of the first optimization objective remains unchanged; if the wind power tracking deviation value is greater than the first deviation threshold, the new weight coefficient of the first optimization objective is calculated according to the following formula: ; in, This represents the corrected weight coefficients of the first optimization objective. This represents the weight coefficient of the first optimization objective before correction. This indicates the wind power tracking deviation value at the current moment. This represents the first deviation threshold. According to this formula, the larger the deviation, the greater the increase in the weighting coefficient, thus forcing wind power output to track the target more closely in subsequent optimizations and improving frequency regulation accuracy.
[0037] Then, the constraint boundary value of the upper limit of energy storage charge-discharge depth is corrected online based on the energy storage power tracking deviation value. A second deviation threshold is preset, which is determined based on the maximum allowable tracking deviation that does not affect the expected lifespan in the energy storage system cycle life test data, for example, set to 3% of the rated energy storage power. The current energy storage power tracking deviation value is compared with the second deviation threshold: if the energy storage power tracking deviation value is less than or equal to the second deviation threshold, the current constraint boundary value of the upper limit of energy storage charge-discharge depth remains unchanged; if the energy storage power tracking deviation value is greater than the second deviation threshold, the new constraint boundary value of the upper limit of energy storage charge-discharge depth is calculated according to the following formula: ; in, This indicates the revised upper limit of the energy storage charge / discharge depth. This indicates the upper limit of the energy storage charge / discharge depth before the correction. This indicates the current energy storage power tracking deviation value. This represents the second deviation threshold. This formula limits the output of the energy storage system by reducing the upper limit of depth, thus avoiding accelerated lifespan degradation due to frequent deep charging and discharging.
[0038] Finally, the weighting coefficients are synchronously corrected based on the current state of charge (SOC) value in the energy storage regulation margin characteristics. Preset overcharge and over-discharge warning values are established, for example, 90% for overcharge and 10% for over-discharge. The current SOC value is compared with these two warning values: if the current SOC value is greater than or equal to the overcharge warning value, it indicates that the energy storage system is about to overcharge. In this case, the weighting coefficient of the first optimization objective is reduced to 0.8 times its original value, while the weighting coefficient of the second optimization objective is increased to 1.2 times its original value, prioritizing the reduction of energy storage charging demand; if the current SOC value is less than or equal to the over-discharge warning value, it indicates that the energy storage system is about to over-discharge. In this case, the weighting coefficient of the first optimization objective is increased to 1.2 times its original value, while the weighting coefficient of the second optimization objective is reduced to 0.8 times its original value, prioritizing the reduction of energy storage discharging demand; if the current SOC value is between the overcharge and over-discharge warning values, no correction is made. The revised weighting coefficients and constraint boundary values will be used to solve the multi-objective optimization model in the next control cycle, enabling adaptive adjustment of the control strategy.
[0039] The working principle of this invention is as follows: Simultaneously acquire grid frequency regulation commands, real-time wind power output, and energy storage state of charge and adjustable power capacity data sequences; perform variational mode decomposition and Hilbert transform on the wind power output sequence to extract wind power fluctuation frequency and amplitude characteristics as wind power regulation capability characteristics; perform sliding window extreme point detection on the energy storage state of charge sequence to obtain charge / discharge depth and response rate, and combine this with the adjustable power capacity coefficient to construct energy storage regulation margin characteristics; use the wind power regulation capability characteristics and energy storage regulation margin characteristics as state constraints, and combine them with the frequency regulation command sequence to construct a multi-objective optimization model with the objectives of maximizing frequency regulation tracking accuracy and minimizing energy storage lifetime loss; use Chebyshev aggregation method and golden section search method to solve for the optimal power allocation coefficient; generate wind power smoothing commands and energy storage limiting commands based on these coefficients and issue them for execution; finally, collect actual response data to calculate the tracking deviation, and dynamically adjust the weight coefficients and constraint boundaries in the optimization model based on the deviation to achieve adaptive adjustment of the control strategy.
[0040] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.
Claims
1. A power optimization allocation control method based on a wind-storage combined frequency regulation system, characterized in that, Includes the following steps: S1, synchronously acquire the grid frequency regulation command sequence, the wind farm real-time output data sequence, and the energy storage system real-time state of charge and adjustable power capacity data sequence; S2 performs volatility analysis and prediction error modeling on real-time power output data sequences to extract wind power regulation capability characteristics; Charge-discharge depth and response rate analysis were performed on the state of charge and adjustable power capacity data sequences to extract energy storage regulation margin characteristics, specifically including: Variational mode decomposition was performed on the real-time power output data sequence to obtain multiple modal components. The modal component with the highest energy proportion was selected as the main wave mode sequence. The Hilbert transform is performed on the main wave mode sequence to obtain the instantaneous frequency and instantaneous amplitude. The variance of the instantaneous frequency is calculated as the wind power wave frequency characteristic, and the mean of the instantaneous amplitude is calculated as the wind power wave amplitude characteristic, thus constituting the wind power regulation capability characteristic. Based on the characteristics of wind power fluctuation frequency, the length of the sliding window is set, and the extreme points of the state of charge sequence are detected within the sliding window. The change in state of charge between adjacent extreme points and the corresponding time interval are obtained, and the charging and discharging depth and response rate are obtained. Based on the characteristics of wind power fluctuation amplitude and combined with the current adjustable power capacity value, the proportion of power capacity required by energy storage is calculated to obtain the adjustable power capacity coefficient, which together with the depth of charge and discharge and the response rate constitutes the energy storage regulation margin characteristics. The calculation process for the adjustable power capacity coefficient is as follows: The initial proportional coefficient is obtained based on the ratio of the wind power fluctuation amplitude characteristics to the preset benchmark amplitude; Based on the difference between the wind power fluctuation frequency characteristics and the rated response frequency of energy storage, the preset frequency response attenuation correspondence is queried to obtain the response attenuation coefficient. Multiply the initial proportional gain by the response decay coefficient to obtain the corrected proportional gain; Based on the ratio of the current adjustable power capacity to the rated power of the energy storage, the correction ratio coefficient is limited to obtain the adjustable power capacity coefficient. S3. Using the characteristics of wind power regulation capability and energy storage regulation margin as state constraints, and combining the grid frequency regulation command sequence, a multi-objective optimization model is constructed with the goal of maximizing frequency regulation tracking accuracy and minimizing energy storage lifetime loss. The multi-objective optimization model is then solved to obtain the optimal power allocation coefficient between wind power and energy storage. S4, based on the optimal power allocation coefficient, generates power adjustment commands for the wind farm and charging / discharging control commands for the energy storage system, and issues them for execution; S5 collects the actual frequency modulation response data after the execution command, calculates the deviation between the actual frequency modulation effect and the frequency modulation command, and corrects the objective function and constraints of the multi-objective optimization model online based on the deviation.
2. The power optimization allocation control method based on a wind-storage combined frequency regulation system according to claim 1, characterized in that, The calculation process for the response attenuation coefficient is as follows: Calculate the absolute value of the difference between the wind power fluctuation frequency characteristic and the rated response frequency of energy storage, and use it as the frequency deviation. Based on the frequency deviation, the corresponding attenuation coefficient is found in the pre-stored frequency deviation and attenuation coefficient mapping table. The mapping table contains multiple frequency deviation intervals and the reference attenuation coefficient corresponding to each interval. If the frequency deviation falls between two adjacent frequency deviation intervals, then the reference attenuation coefficients corresponding to these two intervals are linearly interpolated to obtain the interpolated attenuation coefficients. The found baseline attenuation coefficient or the attenuation coefficient calculated by interpolation is output as the response attenuation coefficient.
3. The power optimization allocation control method based on a wind-storage combined frequency regulation system according to claim 1, characterized in that, S3 specifically includes: Based on the fluctuation frequency and fluctuation amplitude characteristics in the wind power regulation capability characteristics, the frequency band of the power grid frequency regulation command sequence is segmented, and the command component that matches the wind power fluctuation characteristics is extracted as the wind power tracking target sequence. Based on the response rate and adjustable power capacity coefficient in the energy storage regulation margin characteristics, the grid frequency regulation command sequence is subjected to amplitude limiting processing to obtain the energy storage compensation target sequence. The first optimization objective is to minimize the deviation between the wind power tracking target sequence and the actual wind power output, and the second optimization objective is to minimize the deviation between the energy storage compensation target sequence and the actual energy storage output. A multi-objective optimization model is constructed using the upper limit of the fluctuation amplitude in the wind power regulation capability characteristics and the upper limit of the charge and discharge depth in the energy storage regulation margin characteristics as constraint boundaries. To solve the multi-objective optimization model: First, the Chebyshev aggregation method is used to aggregate the first and second optimization objectives into a single objective. Then, the golden section search method is used to iteratively optimize within the constraint boundary to obtain the power allocation coefficient that minimizes the aggregation objective, which is taken as the optimal power allocation coefficient.
4. The power optimization allocation control method based on a wind-storage combined frequency regulation system according to claim 3, characterized in that, Solving the multi-objective optimization model specifically includes: Based on the positional relationship between the current state of charge value in the energy storage regulation margin characteristics and the preset ideal state of charge range, the first weighting coefficient of the first optimization objective and the second weighting coefficient of the second optimization objective are determined. Calculate the first deviation between the first optimization objective and its preset ideal value, and the second deviation between the second optimization objective and its preset ideal value. Take the maximum value between the product of the first deviation and the first weight coefficient and the product of the second deviation and the second weight coefficient as the Chebyshev aggregation objective. Within the search interval formed by the constraint boundary, the first and second trial points are selected according to the golden ratio. The Chebyshev aggregation target values corresponding to the two trial points are calculated respectively. The search interval is narrowed based on the comparison results of the two aggregation target values. The process continues until the length of the search interval is less than the preset convergence threshold. Then, the power allocation coefficient corresponding to the midpoint of the current search interval is taken as the optimal power allocation coefficient.
5. The power optimization allocation control method based on a wind-storage combined frequency regulation system according to claim 1, characterized in that, S4 specifically includes: The power values to be undertaken by wind power and energy storage are obtained by multiplying the optimal power allocation coefficient with the power demand of the current grid frequency regulation command. Based on the fluctuation frequency characteristics in the wind power regulation capability characteristics, the wind power load value is smoothed on a time scale to obtain the wind power adjustment command. Based on the response rate in the energy storage regulation margin characteristics, the change rate of the energy storage power value is limited to obtain the energy storage charging and discharging power command. The wind power adjustment command and the energy storage charging and discharging power command are sent synchronously to the wind farm controller and the energy storage system controller for execution.
6. The power optimization allocation control method based on a wind-storage combined frequency regulation system according to claim 5, characterized in that, The obtained wind power adjustment command specifically includes: Based on the instantaneous frequency in the characteristics of wind power regulation capability, calculate the fluctuation period value at the current moment, and use the fluctuation period value as the initial length of the smoothing time window; Multiple historical sampling points of wind power load value are collected within the smoothing time window. After removing the maximum and minimum values, the arithmetic mean of the remaining sampling points is taken to obtain the smoothed power baseline value at the current moment. Based on the deviation between the smoothed power benchmark value and the wind power carrying capacity value, and combined with the preset ramp rate limit, the correction is made, and the corrected value is used as the wind power adjustment command at the current moment.
7. The power optimization allocation control method based on a wind-storage combined frequency regulation system according to claim 1, characterized in that, S5 specifically includes: The actual power output of the wind farm and the actual power output of the energy storage system after the execution command are collected and compared with the grid frequency regulation command sequence to obtain the wind power tracking deviation value and the energy storage power tracking deviation value. The wind power tracking deviation value is compared with a preset first deviation threshold. If it exceeds the first deviation threshold, the weight coefficient of the first optimization objective in the multi-objective optimization model is increased. The energy storage power tracking deviation value is compared with the preset second deviation threshold. If it exceeds the second deviation threshold, the constraint boundary value of the upper limit of the energy storage charging and discharging depth in the multi-objective optimization model is reduced. The current state of charge value in the energy storage regulation margin feature is compared with the preset overcharge and over-discharge warning value. If the warning value is triggered, the weight coefficients of the first optimization objective and the second optimization objective are adjusted simultaneously.
Citation Information
Patent Citations
Wind-torage combined operation optimization method and system considering life and frequency modulation performance of energy storage
CN111049198A
