A time-scale-based adaptive wavelet packet method for damping wind power

Abstract

The application relates to a time scale-based adaptive wavelet packet method for suppressing wind power, which comprises the following steps: dividing a time domain based on a set time scale, performing adaptive wavelet packet decomposition, and suppressing data exceeding a limit of fluctuation by using an arithmetic average method in adjacent time domains; analyzing the correlation between power components of a hybrid energy storage device based on the characteristics of the energy storage device and the energy distribution of a node, obtaining power components of energy-type and power-type energy storages, and optimizing the capacity of the hybrid energy storage device in combination with maximum charging and discharging power and energy change; dynamically partitioning the SOC of the energy storage device, improving a membership function of fuzzy control, and adopting two-stage fuzzy control to optimize a hybrid energy storage power instruction, so as to reduce the over-limit of the SOC of the energy storage system. The application can effectively suppress power fluctuation, more accurately allocate hybrid energy storage components, effectively reduce the over-limit of the SOC, and improve power quality.

Description

Technical Field

[0001] This invention belongs to the technical field of wind power fluctuation mitigation strategies, and particularly relates to an adaptive wavelet packet method for mitigating wind power based on time scale. Background Technology

[0002] In the process of achieving carbon peaking and carbon neutrality goals in the power system, the traditional power system, centered on fossil fuels, is transforming and upgrading into a new power system centered on clean and low-carbon energy. However, wind power, as one of the representatives of new energy sources, exhibits fluctuations and uncertainties in its output, which can significantly impact power system stability, grid frequency, and power quality during grid connection. Therefore, effectively mitigating grid connection fluctuations has become an urgent problem to be solved. Currently, both energy storage and power storage systems have their own drawbacks, while hybrid energy storage systems combine the advantages of both, achieving economic and technological complementarity, and have become the preferred choice for mitigating wind power output fluctuations.

[0003] When configuring an energy storage system, it is crucial to rationally determine the grid-connected power of wind power and the coordinated allocation of fluctuating power among different energy storage devices. A wind power fluctuation mitigation strategy is needed that can rationally allocate power between energy-type and power-type energy storage and extend the service life of the energy storage system. Summary of the Invention

[0004] The purpose of this invention is to provide a time-scale-based adaptive wavelet packet smoothing method and system for wind power, which can rationally allocate power between energy-type energy storage and power-type energy storage, and extend the service life of the energy storage system.

[0005] This invention provides a time-scale-based adaptive wavelet packet smoothing method for wind power suppression. The method is designed for hybrid energy storage systems consisting of power-type and energy-type energy storage, and includes the following steps:

[0006] Step 1: Divide the time domain based on the set time scale and perform adaptive wavelet packet decomposition to reduce the local over-smoothing phenomenon when processing long-term wind power data. For the fluctuation limit that may occur at the boundary of adjacent time domains, perform volatility analysis on the unprocessed data to find the time domain boundary where the fluctuation limit is exceeded. Use the arithmetic mean method in adjacent time domains to smooth the data with the fluctuation limit exceeded.

[0007] Step 2: Based on the characteristics of energy storage devices and the energy distribution of nodes, the correlation between the power components of hybrid energy storage is analyzed using the energy entropy difference. By analyzing the energy entropy difference between each high-frequency component, the node with the largest difference in energy entropy during the change process is selected as the high-low frequency boundary point of hybrid energy storage. The power components of energy-type and power-type energy storage are obtained, and the hybrid energy storage capacity is optimized by combining the maximum charging and discharging power and energy change.

[0008] Step 3: By dynamically partitioning the energy storage SOC, the membership function of fuzzy control is improved, and two-stage fuzzy control is used to optimize the hybrid energy storage power command in order to reduce the SOC overrun of the energy storage system.

[0009] Furthermore, the adaptive wavelet packet decomposition in step 1 includes: selecting the db6 wavelet as the wavelet basis for the original wind power output power P. w The decomposition process is performed, with 4 hours chosen as the time scale. A typical day is divided into 6 time domains T(i). Wavelet packet decomposition is performed on each time domain T(i), and the minimum number of decomposition layers that meet the grid connection fluctuation standard is selected.

[0010] Furthermore, the energy storage system is a hybrid energy storage system consisting of flywheel energy storage and lithium iron phosphate batteries.

[0011] Furthermore, in step 1, the fluctuation limit is set according to national standards and based on the power grid frequency regulation characteristics, in accordance with the requirements of the power grid dispatching department. The maximum power fluctuation limit within the time window is used as the evaluation index, and the standard is defined as follows:

[0012] △P L =P max (t)-P min (t), t∈T L

[0013] In the formula, ΔP L For power fluctuation limits; T L P is a time window of length L; max (t), P min (t) represents T L The maximum and minimum wind power within the area.

[0014] Furthermore, the correlation analysis between the hybrid energy storage power components using the energy entropy difference described in step 2 is as follows:

[0015] H(i) = -P(i)log2P(i)

[0016] △H(i)=|H(i+1)-H(i)|

[0017] In the formula, (i) is the energy entropy value of the i-th node; P(i) is the proportion of the energy of the i-th node in the total energy; ΔH(i) is the energy entropy difference between the i-th and i+1-th nodes;

[0018]

[0019]

[0020] In the formula, cj,k The wavelet coefficients of signal x(t); E is a wavelet basis function; m Let E be the energy of each node; E is the sum of the energies of all nodes.

[0021] Repeating the analysis process, the node k with the largest difference in energy entropy during the change is taken as the power boundary between low-frequency and high-frequency power components, and the power P of energy-type and power-type energy storage is obtained by decomposition. b and P f As shown in the following formula:

[0022]

[0023] Furthermore, the optimization of hybrid energy storage capacity by combining maximum charge / discharge power and energy change described in step 2 is specifically expressed as follows:

[0024] Maximum charge / discharge power:

[0025]

[0026] In the formula, P b_max and P f_max These are the maximum charge and discharge reference powers for battery and flywheel energy storage, respectively; P b_ref and P f_ref These are the reference power values ​​for battery and flywheel energy storage, respectively; η b and η f The charging and discharging efficiencies are for battery and flywheel energy storage, respectively.

[0027] Energy changes in energy storage:

[0028]

[0029] The improved energy storage capacity is:

[0030]

[0031] In the formula, S ocu S ocl分别 The upper and lower limits of SOC are defined as follows: for energy storage, the upper and lower limits of SOC are 0.8 and 0.2 respectively; for power storage, the upper and lower limits of SOC are 0.9 and 0.1 respectively. C This refers to the rated capacity of the energy storage.

[0032] Furthermore, step 3 includes: considering the characteristics of hybrid energy storage, dynamically partitioning the SOC of the energy storage; for different partitions, using different fuzzy control membership functions and rules to adjust the power output of the flywheel energy storage and the battery; and using the centroid method to defuzzify the membership values ​​to obtain the corrected power regulation coefficient K of the energy storage. f(t), thereby optimizing the overall system performance; the revised power storage power command is:

[0033] P f (t)=K f (t)×P f_ref (t)

[0034] The difference before and after the power command correction is:

[0035] △P f (t)=(1-K f (t))×P f_ref (t)

[0036] The difference ΔP between power storage power command and the value before and after the correction. f (t) Compensated by energy storage, the corrected energy storage power command is:

[0037] P b (t)=K b (t)×(P b_ref (t)+△P f (t)).

[0038] The present invention also provides a time-scale-based adaptive wavelet packet smoothing wind power system, including a wind power smoothing module, wherein the wind power smoothing module executes the time-scale-based adaptive wavelet packet smoothing wind power method.

[0039] The present invention also provides a non-transitory computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the time-scale-based adaptive wavelet packet smoothing method for wind power generation.

[0040] The present invention also provides an electronic device, comprising:

[0041] The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes these computer instructions to perform the time-scale-based adaptive wavelet packet smoothing method for wind power generation.

[0042] By employing the above scheme, the time-scale-based adaptive wavelet packet smoothing method and system for wind power can effectively smooth power fluctuations, more accurately allocate hybrid energy storage components, and effectively reduce SOC exceedances, thereby improving power quality.

[0043] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Description of the Drawings

[0044] Figure 1 It is a system structure diagram of a hybrid energy storage of lithium iron phosphate and flywheel;

[0045] Figure 2 It is a flowchart of the method for suppressing wind power by adaptive wavelet packet based on time scale according to the present invention;

[0046] Figure 3 It is a comparison chart of the 10-minute volatility of wind power processed by different wavelet packet methods;

[0047] Figure 4 It is a diagram of the power distribution result;

[0048] Figure 5 It is the membership function of the flywheel;

[0049] Figure 6 It is the membership function of the lithium iron phosphate battery;

[0050] Figure 7 It is a schematic structural diagram of an electronic device according to the present invention. Detailed Embodiments

[0051] The following further describes the detailed embodiments of the present invention in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0052] The output of wind power has volatility and uncertainty, which will greatly affect the stability of the power system, the grid frequency and the power quality during the grid connection process. Installing an energy storage device in a wind farm can timely suppress the fluctuation of wind power, reduce the impact on the grid, and improve the power quality of wind power. To achieve the above purposes, this embodiment proposes a method for suppressing wind power by adaptive wavelet packet based on time scale.

[0053] Refer Figure 2 As shown, the specific steps of the suppression strategy in this embodiment are as follows:

[0054] First, select an appropriate time scale to divide the time domain, perform adaptive wavelet packet decomposition, reduce the local over-suppression phenomenon that occurs when processing long-term wind power data, and suppress the possible fluctuation over-limit at the boundary of adjacent time domains;

[0055] Secondly, considering the characteristics of the energy storage device and the node energy distribution, use the energy entropy difference to analyze the correlation between the power components of the hybrid energy storage, select the point with the largest entropy difference as the power demarcation point of the hybrid energy storage, obtain the power components of the flywheel energy storage and the lithium iron phosphate battery, and optimize the hybrid energy storage capacity in combination with the maximum charge and discharge power and energy change;

[0056] Finally, by dynamically partitioning the energy storage SOC, improving the fuzzy control membership function, and optimizing the hybrid energy storage power command, the SOC overrun of the energy storage system is reduced. The specific technical solution is as follows:

[0057] (1) Select an appropriate time scale to divide the time domain, perform adaptive wavelet packet decomposition, and smooth out any potential fluctuations that may occur at the boundaries of adjacent time domains. This step includes the following:

[0058] (1.1) Divide the time scale appropriately to reduce the impact of the overall smoothing effect on the maximum fluctuation in a certain period, and avoid the wind power smoothing effect in other periods being poor, thereby increasing the burden on the hybrid energy storage system.

[0059] In this embodiment, a typical day is divided into 6 time domains T(i). Wavelet packet decomposition is performed on each time domain T(i), and the minimum number of decomposition layers that meets the grid connection fluctuation standard is selected as follows:

[0060]

[0061] Figure 1 This is a system structure diagram of a hybrid energy storage system combining lithium iron phosphate and flywheel. The system mainly consists of a wind turbine, a flywheel energy storage system, an electrochemical energy storage system, a transformer, and a busbar. In the diagram, P... b P is the reference power for charging and discharging of lithium iron phosphate batteries. f P is the reference power for charging and discharging flywheel energy storage, with discharge being positive and charging being negative. G P is the reference value for grid-connected wind power. HESS P is the reference power for charging and discharging of the hybrid energy storage system. W This provides the real-time output power of the wind farm.

[0062] The relationship between the above powers is shown below:

[0063]

[0064] (1.2) Perform adaptive wavelet packet decomposition to reduce the local over-smoothing phenomenon that occurs when processing long-term wind power data.

[0065] The 10-minute volatility graph of wind power after processing with the wavelet packet method is shown below. Figure 3 As shown.

[0066] The decomposition and reconstruction are as follows:

[0067] Decomposition Algorithm:

[0068]

[0069] Reconstruction Algorithm:

[0070]

[0071] (1.3) Smooth out fluctuations that may occur at the boundaries of adjacent time domains.

[0072] To mitigate potential fluctuations at adjacent time domain boundaries, volatility analysis is performed on unprocessed data to identify time domain boundaries where fluctuations exceed limits, and these boundaries are then mitigated. This avoids the impact of boundary effects that may occur at adjacent time domain boundaries when wavelet basis functions are limited within the time domain, thus preventing them from affecting the overall performance.

[0073] Since wavelet packet decomposition across time domains may result in significant discontinuities or fluctuations at time domain boundaries, it is necessary to calculate whether the power at the boundaries of adjacent time domains meets the grid connection fluctuation standard. If not, the scale of the smoothing window is set to L, and L=1. The arithmetic mean of the current power within the smoothing window and the previous (L-1) historical values ​​is then calculated. Assuming the current grid-connected power of wind power, calculate whether it meets the grid connection fluctuation standard. If it does not meet the standard, let L = L + 1, continue to take the arithmetic average of the fluctuation limit exceeding the limit within the smoothing window and the historical value, and repeat the steps until the grid connection fluctuation standard is met. Figure 3 As shown.

[0074] Fluctuation limits need to be set according to national standards and the power grid frequency regulation characteristics, as well as the requirements of the power grid dispatching department. The maximum power fluctuation limit within a time window (MPFR) is used as the evaluation index, and this standard is defined as follows:

[0075] △P L =P max (t)-P min (t), t∈T L

[0076] In the formula, ΔP L —Power fluctuation limit; T L —A time window of length L; P max (t), P min (t)——T L The maximum and minimum wind power within the area.

[0077] (2) Considering the characteristics of energy storage devices and the energy distribution of nodes, the correlation between the power components of hybrid energy storage is analyzed using the energy entropy difference. The power boundary point of hybrid energy storage is selected to obtain the power components, and the hybrid energy storage capacity is optimized by combining the maximum charge and discharge power and energy change. This step includes the following:

[0078] (2.1) Calculate the energy entropy. The maximum entropy difference between two adjacent nodes is selected as the power boundary point for hybrid energy storage. The calculation results of the energy entropy difference are shown in Table 1:

[0079] Table 1 Calculation results of energy entropy difference

[0080]

[0081] From Table 1, we can see that the maximum value of the energy entropy difference is ΔH. 2,3 Then S 6,1:2 Composed of low-frequency components, S 6,3:63 The high-frequency component is composed of a low-frequency component, which is then distributed as a power signal to the lithium iron phosphate battery. The high-frequency component is distributed as a power signal to the flywheel energy storage. The power distribution result is as follows: Figure 4 As shown.

[0082] (2.2) Optimize hybrid energy storage capacity by combining maximum charge and discharge power and energy change.

[0083] Considering that the power components of the two parts have been obtained, this section configures the capacity of the hybrid energy storage system by combining the large charge and discharge power and energy change. The maximum power and energy change considered in the configuration are calculated by the following parts.

[0084] Maximum charge / discharge power:

[0085]

[0086] In the formula, P b_max and P f_max The maximum charge / discharge reference power for battery and flywheel energy storage; P b_ref and P f_ref Reference power for energy storage in batteries and flywheels; η b and η f The charging and discharging efficiency of energy storage for batteries and flywheels.

[0087] Energy changes in energy storage:

[0088]

[0089] The configuration results are shown in Table 2:

[0090] Table 2 Configuration Results

[0091]

[0092] (3) By dynamically partitioning the energy storage SOC, improving the fuzzy control membership function, and optimizing the hybrid energy storage power command, the SOC overrun of the energy storage system is reduced. This step includes the following:

[0093] (3.1) Dynamically partition the energy storage SOC, improve the fuzzy control membership function, and optimize the power command. This step includes the following:

[0094] The energy storage state of charge (SOC) is dynamically partitioned, and the membership function of the SOC is improved (e.g., the membership function of the flywheel and the lithium iron phosphate battery are shown in the figure). Figure 5 , Figure 6 As shown), it is set into three regions, namely: S OC1 ∈[0.4,0.6]、S OC2 ∈[S ocl ,0.4)∪(0.6,S ocu ] and S OC3 ∈(0~S ocl )∪(S ocu ,1), the specific correction rules are as follows: when S OC =S OC1 The energy storage system maintains a power allocation scheme based on energy entropy difference, without needing to adjust the current power command of the energy storage; when S OC OC1 Furthermore, energy storage still requires high-power discharge or S OC >S OC1 However, when energy storage still requires high-power charging, the fuzzy controller will control S. OC Optimization was performed. The optimized energy storage SOC range is shown in Table 3:

[0095] Table 3 Optimized Energy Storage SOC Range

[0096]

[0097]

[0098] Finally, the centroid method is used to defuzzify the membership values, and the corrected power is obtained.

[0099] This time-scale-based adaptive wavelet packet method for smoothing wind power fluctuations first selects an appropriate time scale to divide the time domain, performs adaptive wavelet packet decomposition, and uses an arithmetic averaging method in adjacent time domains to smooth out data exceeding the limit. Second, considering the characteristics of hybrid energy storage, it analyzes wind power fluctuations using energy entropy differences and rationally allocates the fluctuating power. Finally, taking into account the state of charge (SOC) and cycle life of the energy storage devices, it applies two-stage fuzzy control to optimize the power output commands of flywheel energy storage and lithium iron phosphate batteries. This method effectively smooths power fluctuations, more accurately allocates hybrid energy storage components, and effectively reduces SOC exceeding limits, thus improving power quality.

[0100] ​This embodiment also provides an adaptive wavelet packet power smoothing system for wind power based on time scales, including a wind power smoothing module that executes the method for smoothing wind power by adaptive wavelet packets based on time scales.

[0101] This embodiment also provides a non-transitory computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method for smoothing wind power by adaptive wavelet packets based on time scales.

[0102] As Figure 7 shown, this embodiment also provides an electronic device, including:

[0103] a memory 201 and a processor 202, which are communicatively connected to each other. The memory 201 stores computer instructions, and the processor 202 executes the computer instructions to execute the method for smoothing wind power by adaptive wavelet packets based on time scales.

[0104] The above is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that for those of ordinary skill in the art, without departing from the technical principles of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A time-scale-based adaptive wavelet packet smoothing method for wind power suppression, characterized in that, The method is for a hybrid energy storage system consisting of power-type energy storage and energy-type energy storage, and includes the following steps: Step 1: Divide the time domain based on the set time scale and perform adaptive wavelet packet decomposition to reduce the local over-smoothing phenomenon when processing long-term wind power data. For the fluctuation limit that may occur at the boundary of adjacent time domains, perform volatility analysis on the unprocessed data to find the time domain boundary where the fluctuation limit is exceeded. Use the arithmetic mean method in adjacent time domains to smooth the data with the fluctuation limit exceeded. Step 2: Based on the characteristics of energy storage devices and the energy distribution of nodes, the correlation between the power components of hybrid energy storage is analyzed using energy entropy difference. By analyzing the energy entropy difference between each high-frequency component, the node with the largest difference in energy entropy during the change process is selected as the high-low frequency boundary point of hybrid energy storage, obtaining the power components of energy-type and power-type energy storage. The hybrid energy storage capacity is then optimized by combining the maximum charge / discharge power and energy change. The analysis process for analyzing the correlation between the power components of hybrid energy storage using energy entropy difference is as follows: H(i) = -P(i)log2P(i) ΔH(i)=|H(i+1)-H(i)| In the formula, (i) is the energy entropy value of the i-th node; P(i) is the proportion of the energy of the i-th node in the total energy; ΔH(i) is the energy entropy difference between the i-th and i+1-th nodes; In the formula, c j,k The wavelet coefficients of the signal x(t); E is a wavelet basis function; m Let E be the energy of each node; E is the sum of the energies of all nodes. Repeating the analysis process, the node k with the largest difference in energy entropy during the change is taken as the power boundary between low-frequency and high-frequency power components, and the power P of energy-type and power-type energy storage is obtained by decomposition. b and P f As shown in the following formula: The optimization of hybrid energy storage capacity by combining maximum charge / discharge power and energy change is specifically expressed as follows: Maximum charge / discharge power: In the formula, P b_max and P f_max These are the maximum charge and discharge reference powers for battery and flywheel energy storage, respectively; P b_ref and P f_ref These are the reference power for battery and flywheel energy storage, respectively; η b and η f The charging and discharging efficiencies are for battery and flywheel energy storage, respectively. Energy changes in energy storage: The improved energy storage capacity is: In the formula, S ocu S ocl分别 The upper and lower limits of SOC are defined as follows: for energy storage, the upper and lower limits of SOC are 0.8 and 0.2 respectively; for power storage, the upper and lower limits of SOC are 0.9 and 0.1 respectively. C This refers to the rated capacity of the energy storage. Step 3: By dynamically partitioning the energy storage SOC, the membership function of fuzzy control is improved, and two-stage fuzzy control is used to optimize the hybrid energy storage power command in order to reduce the SOC overrun of the energy storage system.

2. The time-scale-based adaptive wavelet packet smoothing method for wind power reduction according to claim 1, characterized in that, The adaptive wavelet packet decomposition in step 1 includes: selecting the db6 wavelet as the wavelet basis for the original wind power output power P. w The decomposition process is performed, with 4 hours chosen as the time scale. A typical day is divided into 6 time domains T(i). Wavelet packet decomposition is performed on each time domain T(i), and the minimum number of decomposition layers that meet the grid connection fluctuation standard is selected.

3. The time-scale-based adaptive wavelet packet smoothing method for wind power reduction according to claim 2, characterized in that, The energy storage system is a hybrid energy storage system consisting of flywheel energy storage and lithium iron phosphate batteries.

4. The time-scale-based adaptive wavelet packet smoothing method for wind power reduction according to claim 3, characterized in that, In step 1, the fluctuation limit is set according to national standards and based on the power grid frequency regulation characteristics, in accordance with the requirements of the power grid dispatching department. The maximum power fluctuation limit within the time window is used as the evaluation index. The standard is defined as follows: ΔP L =P max (t)-P min (t),t∈T L In the formula, ΔP L For power fluctuation limits; T L P is a time window of length L; max (t), P min (t) represents T L The maximum and minimum wind power within the area.

5. The time-scale-based adaptive wavelet packet smoothing method for wind power reduction according to claim 4, characterized in that, Step 3 includes: considering the characteristics of hybrid energy storage, dynamically partitioning the SOC of the energy storage; for different partitions, adjusting the power output of the flywheel energy storage and the battery using different fuzzy control membership functions and rules; and using the centroid method to defuzzify the membership values ​​to obtain the corrected power regulation coefficient K of the energy storage. f (t), thereby optimizing the overall system performance; the revised power storage power command is: P f (t)=K f (t)×P f_ref (t) The difference before and after the power command correction is: ΔP f (t)=(1-K f (t))×P f_ref (t) The difference ΔP between power storage power command and the value before and after the correction. f (t) Compensated by energy storage, the corrected energy storage power command is: P b (t)=K b (t)×(P b_ref (t)+ΔP f (t))。 6. A time-scale-based adaptive wavelet packet smoothing system for wind power generation, characterized in that, It includes a wind power smoothing module, which executes the time-scale-based adaptive wavelet packet smoothing wind power method according to any one of claims 1-5.

7. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions, which, when executed by a processor, implement the time-scale-based adaptive wavelet packet smoothing method for wind power as described in any one of claims 1-5.

8. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform a time-scale-based adaptive wavelet packet smoothing method for wind power as described in any one of claims 1-5.

Images