Calculation method of dynamic heat generation power of lithium batteries for smoothing new energy power fluctuations

Abstract

The present invention discloses a method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in the power of new energy sources, and belongs to the technical field of power system energy storage. The present invention performs cluster analysis on historical output data of new energy sources using an improved K-means algorithm based on the silhouette coefficient method to extract typical output scenarios. Taking the fluctuation rate of the grid-connected power of new energy sources satisfying the fluctuation rate constraint as the optimization goal, an adaptive wavelet packet decomposition algorithm is used to decompose and reconstruct the typical output scenarios, and an expected grid-connected power instruction is determined. Based on the expected grid-connected power instruction, an energy storage configuration model is constructed and solved taking into account the rated power configuration and the rated capacity configuration. The state of charge calculation formula at each moment in the operation cycle is solved, and functional relationships between the dynamic changes of the DC resistance and the working current with the state of charge are respectively constructed. The dynamic heat generation power of the lithium battery is then calculated using the Bernardi heat generation rate formula, and the characteristic mapping relationship between the dynamic heat source, the state of charge, and time is characterized by the dynamic heat generation formula of the lithium battery.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system energy storage, and specifically relates to a method for calculating the dynamic heat generation power of a lithium battery for smoothing new energy power fluctuations. Background Art

[0002] With the large-scale and high-proportion integration of renewable energy into the grid, the randomness, volatility, and intermittency of its power generation pose a serious challenge to the safe and stable operation of the power grid. Energy storage technology, as a key component of the new power system, has attracted considerable attention in this context. Among them, lithium iron phosphate energy storage (hereinafter referred to as "lithium battery energy storage") has become one of the fastest-growing energy storage technologies due to its high energy density and rapid response capabilities.

[0003] Lithium-ion battery energy storage is most commonly used to smooth out fluctuations in renewable energy power. Proper energy storage configuration can effectively mitigate fluctuations in renewable energy output and ensure the safe and stable operation of the power grid. However, lithium-ion batteries generate significant heat during frequent charging and discharging. Failure to effectively manage this heat can not only accelerate performance degradation and shorten battery lifespan, but can also lead to safety incidents such as thermal runaway. Therefore, accurately characterizing the dynamic thermal behavior of lithium-ion batteries is crucial to ensuring the safe operation of energy storage systems.

[0004] Existing research has largely focused on either energy storage configuration optimization or thermal management strategies, lacking a comprehensive approach that organically combines the two in practical application scenarios. Isolated research approaches struggle to fully and accurately characterize the thermal behavior of lithium batteries under complex operating conditions, limiting the overall safety and reliability of lithium battery energy storage systems. Summary of the Invention

[0005] The present invention addresses the problem of energy storage configuration optimization and thermal management strategies being insufficiently integrated in practical application scenarios, which in turn limits the overall safety and reliability of lithium-ion battery energy storage systems. This paper proposes a method for calculating the dynamic thermal power of lithium-ion batteries for smoothing renewable energy power fluctuations. The method uses a k-means algorithm to cluster historical renewable energy output data and extract typical output scenarios. Secondly, an adaptive wavelet packet decomposition algorithm is used to decompose and reconstruct these typical scenarios to identify the high-frequency fluctuation components that need to be smoothed. Furthermore, an energy storage configuration model and a dynamic thermal power calculation model for lithium-ion batteries that takes into account electrochemical-thermal coupling are constructed. The dynamic thermal power of lithium-ion batteries during operation is numerically calculated. Compared with existing methods, the present invention not only optimizes energy storage configuration to enhance the stability of renewable energy access to the grid, but also further quantifies the dynamic thermal behavior of lithium-ion batteries when used to smooth renewable energy power fluctuations, significantly improving the safety and reliability of energy storage systems. This invention provides technical support for the safe operation of lithium-ion battery energy storage systems in scenarios where renewable energy power fluctuations are smoothed.

[0006] To solve the above technical problems, the present invention provides a method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power, comprising the following steps:

[0007] S1. An improved K-means algorithm based on the silhouette coefficient method is used to perform cluster analysis on historical output data of new energy sources to extract typical output scenarios.

[0008] S2. Taking the new energy grid-connected power fluctuation rate meeting the fluctuation rate constraint as the optimization goal, an adaptive wavelet packet decomposition algorithm is used to decompose and reconstruct the typical output scenario and determine the expected grid-connected power command;

[0009] S3. Based on the expected grid-connected power command, considering the rated power configuration and the rated capacity configuration, constructing an energy storage configuration model and solving the energy storage configuration model to obtain an energy storage configuration result;

[0010] S4. Solving the initial state of charge based on the energy storage configuration result, and then combining the energy storage calculation formula of the energy storage system at any time to obtain the state of charge calculation formula at each time during the operation cycle;

[0011] S5. Construct functional relationships between the dynamic changes of DC resistance and working current with the state of charge respectively. Then, combine the state of charge calculation formula at each moment in the operating cycle, calculate the dynamic heat generation power of the lithium battery through the Bernardi heat generation rate formula, and use the dynamic heat generation formula of the lithium battery under energy storage to smooth fluctuations to characterize the characteristic mapping relationship between the dynamic heat source and the state of charge and time.

[0012] Preferably, the S1 includes:

[0013] S11. Obtain historical output data of new energy within a set time interval and perform statistics, and verify and supplement abnormal values ​​and missing values ​​through a sliding average;

[0014] S12, new energy historical output data after simplification through principal component analysis;

[0015] S13, setting the number range of cluster centers and performing iterative solution through the silhouette coefficient method to obtain the optimal number of clusters;

[0016] S14. Based on the optimal number of clusters, a K-means algorithm is used to obtain a typical scenario and generate a new energy output power of the typical scenario.

[0017] Preferably, the S2 includes:

[0018] S21. Define the volatility of new energy and consider the volatility constraint to determine whether energy storage needs to be configured and obtain the configuration result.

[0019] S22. Based on the configuration results, the grid-connected power of the new energy and the expected grid-connected power command are obtained by using the optimization objective of the grid-connected power fluctuation rate of the new energy to meet the fluctuation rate constraint and the adaptive wavelet packet decomposition algorithm. The grid-connected power fluctuation rate of the new energy mentioned here is the new energy power fluctuation rate mentioned above. The specific value can be referred to in formula (9) and the specification "Technical Provisions for Wind Farm Integration into the Power System Part 1: Onshore Wind Power".

[0020] Preferably, the configuration result includes: if the output power of the new energy in the typical scenario meets the grid-connected volatility constraint, then no energy storage is required and the system is directly connected to the grid; if the grid-connected volatility constraint is not met, then energy storage is configured. The grid-connected volatility constraint is that the power volatility of the new energy meets the specification "Technical Provisions for Wind Farms Connected to the Power System Part 1: Onshore Wind Power", as shown in formula (9).

[0021] Preferably, in S22, the output power that requires power smoothing is spectrally analyzed using a discrete Fourier transform to obtain the amplitude-frequency characteristic results for this output scenario. At a fixed scale, a wavelet packet function is defined to obtain a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, and the optimal number of decomposition layers is determined. The renewable energy output power is decomposed into high-frequency and low-frequency components, and the low-frequency component is used as the desired grid-connected power. The difference between the desired grid-connected power and the original renewable energy output power is used as the desired grid-connected power instruction. This output scenario can be understood as a typical renewable energy output scenario that requires energy storage. The amplitude-frequency characteristic results facilitate understanding the high-frequency and low-frequency components of the renewable energy output, which are used for subsequent wavelet packet decomposition. Equation (10) is the construction principle.

[0022] Preferably, in said S3, solving the energy storage configuration model includes solving the rated power and rated capacity of the energy storage system;

[0023] Solving the energy storage system rated power includes solving the energy storage system expected power based on the expected grid-connected power instruction, taking into account the energy storage comprehensive cycle efficiency and energy storage charging and discharging efficiency, obtaining the primary power instruction of the energy storage system, and then considering the satisfaction of the power balance constraint to obtain the energy storage system rated power;

[0024] Solving the rated capacity of the energy storage system includes solving the energy fluctuation of the energy storage system during its operation cycle compared to its initial state based on the primary power command, obtaining a calculation formula for the energy storage system's stored energy at any time, and then obtaining a state of charge expression while satisfying the upper and lower limit constraints of the energy storage system's charge and discharge. The energy storage system's rated capacity is obtained by calculating the ratio of the difference between the maximum energy fluctuation of the energy storage system during its entire operation cycle and the difference between its upper and lower limits of charge and discharge.

[0025] Preferably, the initial charge state is calculated in S4 by the following expression:

[0026] ;

[0027] Where, represents the initial charge state, is the rated capacity of the energy storage system, Indicates the upper limit of energy storage charging, Indicates the lower limit of energy storage discharge, It is the storage energy of the energy storage system at any time.

[0028] Preferably, in S4, the specific expression of DC resistance is:

[0029] ;

[0030] Where, are the polynomial coefficients; is the DC resistance, is an exponential function of the state of charge.

[0031] Preferably, in S4, the specific expression of DC resistance is:

[0032] ;

[0033] Where, is the rated capacity of the single cell; when t =1, ; The sampling interval for renewable energy output data; This is the data of parallel connection of battery cells; is the time corresponding to the sampling point; is the initial state of charge;

[0034] When calculating the DC resistance, the series and parallel connection of the energy storage cells is taken into account, and it is assumed that the charge state of each cell in the energy storage system remains consistent; it is assumed that the charging and discharging power of the lithium battery is constant during the sampling period.

[0035] Preferably, the state of charge calculation formula at each moment in the operation cycle is specifically expressed as follows:

[0036] ;

[0037] Where, is the rated capacity of the energy storage system; is the time point corresponding to the end of the sampling period; is the grid-connected power corresponding to the energy storage at time t.

[0038] As a preferred method, the dynamic heat generation formula of lithium batteries under energy storage to smooth fluctuations is specifically expressed as follows:

[0039] ;

[0040] where X(·), Y(·), and Z(·) are the abstract functions of dynamic heating power, operating current, and DC internal resistance, respectively; and F(·) is the abstract function of dynamic heating power obtained after solution.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. This solution considers the configuration and operation of energy storage for use in scenarios where renewable energy fluctuations are mitigated, and fully calculates the dynamic heat generation of lithium batteries. This approach derives the appropriate scale for energy storage deployment and the dynamic heat generation during operation. This provides an important foundation for "new energy storage allocation" and serves as a valuable reference for adapting to renewable energy grid integration and ensuring safe and stable battery operation. This method addresses deficiencies in calculating the dynamic heat generation of lithium batteries used in energy storage applications to mitigate renewable energy power fluctuations.

[0043] 2. This solution takes into full consideration energy storage configuration optimization and thermal management strategies and their combination, and innovatively expresses the dynamic heat source of lithium batteries as a function of the state of charge and time of the energy storage system in scenarios where new energy power fluctuations are smoothed. This makes it feasible to quantitatively characterize the dynamic heat source of lithium batteries in scenarios where energy storage is used to smooth new energy power fluctuations.

[0044] 3. This solution solves the dynamic DC internal resistance and operating current of the lithium battery, and then jointly solves the charge state at each moment in the operating cycle. It can more accurately predict the state parameters of the lithium battery and provide strong data support for the safe and stable operation of the energy storage system.

[0045] 4. This method is mainly used to smooth out fluctuations in renewable energy power and is suitable for the capacity configuration of energy storage for renewable energy. On the basis of ensuring the safe grid connection of renewable energy power generation and reducing the problem of wind and solar power abandonment caused by power generation fluctuations, it can also be further used as a safety guidance plan for the planning, scheduling and operation of power storage power stations on the power supply side to ensure a high proportion of renewable energy consumption.

[0046] 5. This solution quantitatively characterizes the dynamic heat source of lithium batteries in scenarios where energy storage is used to mitigate renewable energy power fluctuations. This is because the specific analytical expressions for the dynamic heat source, state of charge, and time are dynamic and difficult to accurately describe using specific functional expressions. Therefore, an abstract function is used to map the characteristic relationship between the dynamic heat source, state of charge, and time. This is used to quantitatively characterize the dynamic heat source of lithium batteries in scenarios where energy storage is used to mitigate renewable energy power fluctuations. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Schematic diagram of the system flow of the present invention;

[0048] Figure 2This is the software interface for calculating the dynamic thermal power of a lithium-ion battery in accordance with an embodiment of the present invention;

[0049] Figure 3 Schematic diagram of the decomposition of the adaptive wavelet packet decomposition algorithm of the embodiment of the present invention

[0050] Figure 4 The following is a schematic diagram of the solution results of typical scenarios of the embodiment of the present invention. Figure 1 ;

[0051] Figure 5 The following is a schematic diagram of the solution results of typical scenarios of the embodiment of the present invention. Figure 2 ;

[0052] Figure 6 This is a typical scenario result diagram of an implementation example of the present invention;

[0053] Figure 7 This is a schematic diagram of the calculation results of the output sequence of a typical scenario of an embodiment of the present invention;

[0054] Figure 8 The results of the amplitude-frequency characteristic analysis under various typical scenarios of the implementation examples of the present invention are as follows: Figure 1 ;

[0055] Figure 9 The results of the amplitude-frequency characteristic analysis under various typical scenarios of the implementation examples of the present invention are as follows: Figure 2 ;

[0056] Figure 10 The results of the amplitude-frequency characteristic analysis under various typical scenarios of the implementation examples of the present invention are as follows: Figure 3 ;

[0057] Figure 11 The results of the amplitude-frequency characteristic analysis under various typical scenarios of the implementation examples of the present invention are as follows: Figure 4 ;

[0058] Figure 12 The comparison results of the new energy output sequence before and after the stabilization of the embodiment of the present invention are shown as follows: Figure 1 ;

[0059] Figure 13 The comparison results of the new energy output sequence before and after the stabilization of the embodiment of the present invention are shown as follows: Figure 2 ;

[0060] Figure 14 The comparison results of the new energy output sequence before and after the stabilization of the embodiment of the present invention are shown as follows: Figure 3 ;

[0061] Figure 15 This is a schematic diagram of the output result window of an embodiment of the present invention;

[0062] Figure 16A power sequence curve diagram of an energy storage system according to an embodiment of the present invention;

[0063] Figure 17 This is a graph showing energy changes in an energy storage system according to an embodiment of the present invention;

[0064] Figure 18 A comparison chart of grid-connected fluctuation rates before and after energy storage configuration according to an embodiment of the present invention;

[0065] Figure 19 This is a schematic diagram of measured data of SOC-DC internal resistance of a lithium battery according to an embodiment of the present invention;

[0066] Figure 20 Module 4 outputs window result information of the embodiment of the present invention;

[0067] Figure 21 The best polynomial fitting relationship diagram containing SOC-R of an embodiment of the present invention;

[0068] Figure 22 This is a schematic diagram of the dynamic heat generation power results of the energy storage lithium battery in a typical scenario of an implementation example of the present invention. DETAILED DESCRIPTION

[0069] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0070] Example 1: Figure 1 - Figure 22 As shown in FIG, the calculation method of the dynamic heat generation power of lithium batteries for smoothing the power fluctuation of new energy sources includes:

[0071] Step 1: Clustering and Reduction of New Energy Output Scenario

[0072] Renewable energy output, such as wind and photovoltaic power, typically has high temporal resolution and a massive amount of data. Considering every specific output curve significantly increases computational complexity. Furthermore, renewable energy output is influenced by factors such as weather and season, exhibiting distinct spatiotemporal characteristics. Therefore, clustering and reduction can simplify the vast amount of wind and solar power output data into a few representative scenarios, reducing model complexity and computation time, thereby improving operational efficiency.

[0073] The K-means algorithm is a classic clustering algorithm that is widely used due to its high execution efficiency. However, the K-means algorithm requires the number of cluster centers to be obtained in advance, otherwise it may lead to an increase in the error of the clustering results. To this end, the present invention uses an improved K-means algorithm based on the silhouette coefficient method to cluster and reduce historical operation scenarios. Before clustering, the silhouette coefficient method is used to analyze the historical operation scenarios of new energy to determine the optimal number of cluster centers, and then the K-means algorithm is used to extract typical operation scenarios. The specific steps are as follows:

[0074] Step 1: Input a set of historical renewable energy output data and perform data cleaning and verification to avoid null values ​​or incorrect values ​​(such as negative numbers).

[0075] Step 2: Use principal component analysis (PCA) for dimensionality reduction. Renewable energy output has high dimensions and time series correlation. PCA projects high-dimensional data into a low-dimensional space while preserving the data's key information as much as possible, simplifying the data and revealing output patterns and characteristics, thereby improving computational efficiency.

[0076] Step 3: Use the silhouette coefficient method to determine the optimal number of clusters. Initially set a reasonable range for the number of cluster centers, iteratively solve the data after dimensionality reduction, and determine the optimal number of clusters using the silhouette coefficient method. Assume that each data point in the data set is i , the following is the silhouette coefficient S(i) The calculation process is:

[0077] (1) Calculate the average distance within the cluster a ( i ) —— Measuring Data Points i The consistency with the cluster it belongs to.

[0078]

[0079] Where: C i Yes i The cluster in which it is located; is the number of points in the cluster; d ( i,j ) is a dot i with dot j The distance between them.

[0080] (2) Calculate the average distance of the nearest cluster b ( i ) - measure data points i The distance to its nearest neighbor cluster.

[0081]

[0082] Where: C k is with i Cluster C i Different clusters; It is a cluster C k The number of points in .

[0083] (3) Calculate the silhouette coefficient S ( i )——Used to evaluate clustering quality. Its value is between [-1,1]. The larger the value, the better the clustering effect.

[0084]

[0085] (4) Overall silhouette coefficient S ——For the entire data set, the overall silhouette coefficient is the average of the silhouette coefficients of all data points. k Under the value S , we can find the value that maximizes the silhouette coefficient k value, which is generally considered to be the optimal number of clusters.

[0086]

[0087] Step 4: Obtain typical scenarios based on the K-means algorithm. The basic steps are as follows.

[0088] (1) Data preprocessing, including standardization and outlier filtering.

[0089] (2) Selection k (The best number of clusters determined in the previous step) cluster centers, recorded as

[0090] 、 … .

[0091] (3) Define the loss function J ( c,μ )——The sum of squared errors between each sample and the center point of the cluster to which it belongs.

[0092]

[0093] Where: Representative i samples; C i yes x i The cluster to which it belongs; Represents the center point corresponding to the cluster; M is the total number of samples.

[0094] (4) The loss function converges iteratively.

[0095] make t =0,1,2... is the number of iteration steps, repeat the following process until J Convergence. For each sample x i , assigning it to the nearest center.

[0096]

[0097] Where: argmin It refers to the parameter value that achieves the minimum value in the domain.

[0098] For each class center k , recalculate the center of the class.

[0099]

[0100] (5) Generate typical scene data. When the loss function no longer changes significantly, the algorithm converges and obtains k Typical scenes, each of which is represented by a cluster center point, generate k The output power of new energy in a typical scenario.

[0101] Step 2: Adaptive wavelet packet decomposition

[0102] Compared to traditional low-pass filtering algorithms, wavelet decomposition has good localization and multi-resolution characteristics in processing non-stationary mutation signals, making it more suitable for analyzing renewable energy power fluctuations. The wavelet packet transform is a further development based on the wavelet transform and can provide higher resolution than the wavelet transform. It overcomes the shortcomings of the wavelet transform, which has poor frequency resolution in high-frequency bands and poor time resolution in low-frequency bands. Through multi-level frequency band division, it adaptively improves the time-frequency resolution according to the characteristics of the analyzed signal, helping to obtain more detailed information about the signal. The specific steps are as follows:

[0103] Step 1: New energy target grid-connected power constraints

[0104] Renewable energy power fluctuation is a key indicator for describing the safety of renewable energy grid integration. To ensure the stability and safety of renewable energy grid integration, renewable energy power generation must meet grid connection fluctuation constraint standards when connected to the grid.

[0105] (1) Definition of new energy volatility:

[0106] The change (fluctuation rate) of the active power of renewable energy includes the 1-minute fluctuation rate and the 10-minute fluctuation rate, which can be solved according to the following formula:

[0107]

[0108] Where: and It is represented by the volatility within 1 minute and 10 minutes. Its specific mathematical meaning is the maximum fluctuation of new energy power ( 、 ) and minimum value ( 、 ) deviation.

[0109] (2) Volatility Constraint

[0110] Renewable energy volatility constraints are being promoted and applied according to the "Technical Provisions for Wind Farm Integration into the Power System Part 1: Onshore Wind Power" (GB / T 19963.1-2021). These constraints are closely related to the installed capacity of renewable energy stations. The following formula can be used to calculate the renewable energy volatility limit for different installed capacities.

[0111]

[0112] Where: P N is the installed capacity of new energy stations, MW.

[0113] (3) Determine whether energy storage is needed

[0114] Based on the above definition of fluctuation rate, analyze the raw power output in a typical scenario. If the raw power output meets the grid connection constraints, no energy storage is required and the system can directly connect to the grid. If the constraints are not met, energy storage is required to smooth out power fluctuations, and the system proceeds to the next step.

[0115] Step 2: Adaptive wavelet packet decomposition method

[0116] Wavelet packet decomposition can decompose renewable energy output into high-frequency and low-frequency components in different frequency bands, thereby capturing the characteristics of renewable energy power at different time scales. The relationship between high- and low-frequency components primarily reflects local variations and overall trends in renewable energy power. Renewable energy power is primarily concentrated in the low-frequency portion, while the high-frequency portion, which requires smoothing, accounts for a relatively small proportion. By combining the grid-connected optimization objective with wavelet packet decomposition, we can determine the renewable energy grid-connected power that meets the optimization objective, as well as the deviation that requires energy storage to smooth it out. This deviation can be used as a benchmark for subsequent energy storage configuration research.

[0117] (1) Spectrum analysis

[0118] First, you need to determine the sampling time of the data T s (The minimum sampling time in this invention should be less than 1 minute, in seconds), the sampling frequency is f s =1 / T s The total number of sampling points is N =86400 / T s Use the Discrete Fourier Transform (DFT) to perform spectrum analysis on the output data that needs to be smoothed to obtain the amplitude-frequency characteristics of the scenario. The DFT formula is as follows:

[0119]

[0120] Where: X(k) Represents the data after DFT transformation, k is the frequency index in the frequency domain, P(n) Indicates the power signal of the new energy. is the complex form of Euler's formula, which is equivalent to This formula realizes the time domain signal P (n) to frequency domain signal X(k) conversion.

[0121] (2) Wavelet packet decomposition and reconstruction

[0122] Wavelet packet decomposition is used to decompose the new energy power signal into high-frequency signal and low-frequency signal. Wavelet packet decomposition is obtained by further decomposing the high-frequency part obtained by wavelet decomposition. The decomposition result is to map the original signal to 2 n (n here represents the number of decomposition layers) wavelet packet subspaces, forming a complete binary tree in structure.

[0123] The mathematical definition of wavelet packet is as follows:

[0124]

[0125] Where: is a scaling function used to generate the low-frequency (nearsighted) component of the signal, { h n} n∈k is the low-pass filter coefficient; is a wavelet function used to generate the high-frequency (detail) component of the signal, { g n} n∈k is the high-pass filter coefficient, with g n =(-1) n h 1-n ; Z represents a set of integers. In practical applications, the coefficients are usually finite. By scaling factor 2 t and translation factors k, wavelet function can analyze signals at different time and frequency scales. In order to facilitate the representation of wavelet packet function, a new function symbol is introduced, = , = , and substituting it into formula (11) we get:

[0126]

[0127] At a fixed scale, the orthogonal scaling function = OK, passed 、 、 h n 、 g n The wavelet packet function is defined as follows:

[0128]

[0129] Based on the above wavelet packet function, the wavelet packet decomposition algorithm is shown as follows:

[0130]

[0131] In the formula 、 are the decomposition coefficients of low frequency and high frequency of the nth layer respectively.

[0132] The wavelet packet reconstruction algorithm is defined as follows:

[0133]

[0134] (3) Calculation of the optimal number of layers for wavelet packet decomposition

[0135] right The new energy output power of typical scenarios is decomposed into n-layer wavelet packets and reconstructed into n-layer 2 n The power components of the frequency bands are used to obtain the low-frequency components With high frequency components ( ), where the bandwidth of each signal band is f o Specific examples: Figure 3 As shown in the figure, this algorithm, a core method in the field of time-frequency analysis, can simultaneously perform multi-scale, layer-by-layer decomposition of high- and low-frequency signals. Its decomposition process strictly satisfies orthogonal completeness requirements, eliminating information redundancy while ensuring full coverage of signal features. Compared to traditional wavelet transforms, this algorithm has superior time-frequency localization capabilities and is particularly suitable for analyzing data scenarios with strong randomness and non-stationary characteristics, such as renewable energy output. It can provide high-precision time-frequency domain representation for subsequent feature extraction.

[0136]

[0137] Cyclic deepening of decomposition layers n , using the proposed grid-connected fluctuation constraint to control the low-frequency component Make a judgment. When the amplitude fluctuation of meets the grid-connected fluctuation constraint for the first time, the optimal decomposition level of wavelet packet decomposition can be determined. As the expected grid-connected power , and The difference between the two is used as the expected grid-connected power instruction of the energy storage system .

[0138] Step 3: Energy Storage Configuration Model

[0139] The energy storage configuration model must fully consider the operating characteristics of the energy storage system during its operation cycle, including upper and lower limits of charge and discharge, comprehensive cycle efficiency, power balance and other constraints. The steps to build the energy storage configuration model are as follows:

[0140] Step 1: Rated power configuration

[0141] (1) Expected power of energy storage system

[0142] Solve the expected grid-connected power command of the energy storage system based on the results of the previous step .

[0143]

[0144] Where: When it takes a positive value, it indicates that the energy storage is discharged; when it takes a negative value, it indicates that the energy storage is charged; Output power for new energy.

[0145] (2) Energy storage cycle efficiency constraints

[0146] During the charging and discharging process of the energy storage system, the comprehensive cycle efficiency of the energy storage should be considered. . 、 Represent the charging efficiency and discharging efficiency respectively. Assuming that the charging and discharging efficiencies are equal, then: .

[0147] (3) Primary power instruction of energy storage system

[0148] Considering the energy storage charging and discharging efficiency, the primary power instruction of the energy storage system is obtained .

[0149]

[0150] (4) Power balancing constraints

[0151] In order to ensure that the energy storage system can operate continuously and stably, the power balance constraint should be met, that is, the net charge and discharge amount is zero during the operation cycle. And because the comprehensive cycle efficiency of the energy storage system is always lower than 100%, the charge amount of the energy storage system is always lower than the discharge amount. Therefore, in order to meet the power balance constraint, it is necessary to intervene in the power command of the energy storage system and shift the primary power command downward as a whole. Constraints Power downshift The mathematical definition of is:

[0152]

[0153] (5) Rated power of energy storage system

[0154] After determining the power downshift amount, the grid-connected target power of the energy storage system is is the difference between the primary power command and the power downshift of the energy storage system, and the rated power of the energy storage system is the maximum grid-connected target power of the energy storage system.

[0155]

[0156] Step 2: Rated capacity configuration

[0157] To meet the needs of smoothing out fluctuations in renewable energy power, the energy storage system should be configured with sufficient capacity. The steps for configuring the rated capacity of the energy storage system are as follows:

[0158] (1) Energy fluctuation

[0159] By solving the energy fluctuations during the operation cycle of the energy storage system compared to the initial state, the storage energy of the energy storage system at any time can be obtained.

[0160]

[0161] (2) Upper and lower limit constraints on energy storage system charging and discharging

[0162] In order to avoid overcharge and overdischarge of energy storage system and ensure its safe operation, the upper and lower limits of charge and discharge of energy storage system should be defined. For the convenience of expression, the state of charge (SOC) is used to indicate the remaining power of energy storage system as a percentage of rated capacity. The energy state of the energy storage system is described by the percentage of

[0163]

[0164] Where: Indicates the upper limit of energy storage charging, Indicates the lower limit of energy storage discharge.

[0165] (3) Rated capacity of energy storage system

[0166] The ratio of the difference between the maximum energy fluctuation of the energy storage system during the entire operating cycle and the difference between its upper and lower limits of charge and discharge is calculated to be the rated capacity of the energy storage system configuration.

[0167]

[0168] Step 3: Initial state of charge

[0169] Initial state of charge It means that the initial state of charge should meet the upper limit constraint of the energy storage system in charge and discharge (avoiding over-limit), so it can be solved by the following formula.

[0170]

[0171] Step 4: Dynamic heat generation power calculation model of energy storage lithium battery taking into account electrochemical-thermal coupling

[0172] Lithium-ion energy storage batteries themselves are the primary factor affecting energy storage safety. As the core component of energy storage systems, batteries present potential thermal safety hazards under various complex operating conditions, presenting a safety challenge in practical applications. Therefore, to fundamentally address the thermal safety issues of lithium-ion batteries, research must be conducted on the inherent safety of batteries. Therefore, this paper uses single cells as the research object to construct a dynamic heat generation power calculation model for lithium-ion energy storage batteries.

[0173] Step 1: Lithium battery heat generation rate model

[0174] The heat generation rate of lithium batteries can be solved by the heat generation equation proposed by Bernardi.

[0175]

[0176] Where: q Indicates the heat generation power of the lithium battery, in W / m 3 ; V b Indicates the volume of the lithium battery in m 3 ; U ocv 、 U Respectively represent the open circuit voltage and operating voltage of the lithium battery; I Indicates the working current of the lithium battery, in A; R Indicates the DC internal resistance of the lithium battery, in Ω; T Indicates the ambient temperature of the lithium battery, in K; It represents the entropy thermal coefficient, and its value range is 1~2.8×10 -4 , unit is mV / K.

[0177] From the above formula, we can see that in order to accurately obtain the heat generation rate of the lithium battery, it is necessary to solve the DC internal resistance and operating current of the lithium battery.

[0178] Step 2: Lithium battery DC internal resistance

[0179] During the charging and discharging process of lithium batteries, the DC internal resistance ( R ) will fluctuate with the change of SOC. In order to refine the heat generation power of standard lithium batteries, it is necessary to fully consider the lithium battery R Different types of lithium batteries R The trend of SOC changes is roughly the same, but there are differences in the values. In this invention, the fluctuation measured data of the internal resistance of lithium batteries with SOC changes given by the manufacturer is fitted into a function expression through numerical fitting. R The fluctuation data changing with SOC has the best fitting relationship with the polynomial function, so its function expression can be numerically fitted using the polynomial function.

[0180]

[0181] Where: are the polynomial coefficients, where the value of n can be determined based on the minimum deviation of the numerical fitting.

[0182] Step 3: Lithium battery operating current

[0183] This paper selects the heat generation power of a single cell as the research object, considers the series and parallel connection of the energy storage cells, and solves the working current. Reasonable assumptions need to be made:

[0184] (1) Assume that the state of charge of each battery cell in the energy storage system remains consistent;

[0185] (2) Assume that the charging and discharging power of the lithium battery is constant during the sampling period.

[0186] Therefore, the working current of a single cell at each sampling moment during the operating cycle can be solved based on the following formula:

[0187]

[0188] Where: C N Indicates the rated capacity of a single cell, Ah; p Indicates the parallel connection of energy storage batteries.

[0189] Step 4: Calculate the dynamic heat generation power of the lithium battery

[0190] Equations (26) and (27) show that the DC resistance of a lithium-ion battery has an accurate functional relationship with its operating current and state of charge. Therefore, it is necessary to characterize the state of charge of the energy storage at any given moment. Equation (24) can be used to determine the initial state of charge of the energy storage system, and combined with Equation (21) to determine the state of charge at each moment during the operating cycle.

[0191]

[0192] Substituting the calculated state of charge at each moment of the operating cycle into Equations 26 and 27 yields the DC internal resistance and operating current at any point during the operating cycle. Simultaneously, Equation 25 yields a method that both smooths out fluctuations in renewable energy power and accurately quantifies the dynamic heat generation of lithium batteries in this scenario. Unlike conventional steady-state methods for calculating cell heat generation, this method provides a more representative and practical solution for dynamic cell heat generation, offering practical engineering significance.

[0193] This method innovatively expresses the dynamic heat source of lithium batteries as a function of the state of charge (SOC) and time (t) of the energy storage system in the scenario of smoothing the power fluctuation of new energy, as shown in Equation (29). This makes it feasible to quantitatively characterize the dynamic heat source of lithium batteries in the scenario of energy storage smoothing the power fluctuation of new energy.

[0194]

[0195] Where: X( · ) 、 Y( · ) 、 Z( · ) are the abstract functions of dynamic heating power, working current and DC internal resistance respectively; F( · ) is the abstract function of the dynamic heat generation power obtained after solution.

[0196] Here, the charge state, DC internal resistance, and operating current all change dynamically. Specifically, the charge state at different times considered in this application determines the dynamic DC internal resistance and dynamic operating current at different times. The dynamic DC internal resistance and dynamic operating current at different times are then used as variables to calculate the dynamic heat generation power. Simplifying the above process, it can be considered that the fundamental factor determining the dynamic heat generation power is the charge state at different times. Finally, the entire process is described by a mathematical formula to obtain Equation (29). This formula, as the core formula of this invention patent, further represents the heat generation power in Equation 25, the DC internal resistance in Equation 26, the operating current in Equation 27, and the charge state in Equation 28, completing the transition from an abstract concept to a concrete formula for the dynamic heat source.

[0197] Example 2: Figure 1-Figure 22 As shown, the overall idea is as follows Figure 1 As shown in the figure, the method for calculating the dynamic thermal power of lithium batteries used to smooth out renewable energy power fluctuations includes the following four steps: clustering and reduction of renewable energy output scenarios, adaptive wavelet packet decomposition, energy storage configuration modeling, and a dynamic thermal power calculation model for energy storage batteries that takes into account electrochemical-thermal coupling. Due to the large amount of calculations and the complexity of the process, a computer program was developed to automate the optimization of energy storage configuration and the calculation of dynamic thermal power for smoothing renewable energy power fluctuations, thereby improving computational efficiency and simplifying routine analysis.

[0198] Developed in Python, the software uses the output sequence of renewable energy stations as input and can quickly perform cluster analysis of typical renewable energy scenarios at different time scales, adaptive discrete wavelet decomposition, energy storage configuration, dynamic thermal power calculation of lithium batteries, and visualization.

[0199] Hardware environment: CPU: i5-8300H 2.30GHz; Memory: 16GB; Hard disk: 1TB

[0200] Software environment: Operating system: Windows 10; Supported software: Microsoft Office 2016.

[0201] Software composition: The software mainly consists of two parts, including the calculation module button on the left and the output result window on the right. Users can click on the modules in order to perform calculations according to the prompts in the output result window on the right, and enter the parameters according to the software prompts to complete the calculation. The output result window on the right will generate the corresponding conclusion information and the result saving path. Users can check or modify the calculation results according to the saving path, such as Figure 2 To facilitate reproduction, the author built a calculation software based on Python, which is divided into four steps and has a strong logical connection with the system flow chart of the present invention.

[0202] The first step in calculating the dynamic heat generation power of lithium batteries, used to smooth out fluctuations in new energy power, is clustering and reducing new energy output scenarios. The specific steps are as follows:

[0203] Collecting sample data: In the embodiment of the present invention, the second-level output data of an 80MW wind farm in a certain region in 2022 is used as sample data. The sampling time is 10 seconds, and a total of 8640 sampling points are collected every day.

[0204] The data input file consists of two parts: (1) produced by Excel software in the format of "csv", the stored data is the output sequence of the new energy station; (2) the user is based on the core parameters of the new energy station and the energy storage lithium battery, such as the rated installed capacity of the new energy station, the rated capacity of the lithium battery cell, the internal resistance parameters, etc.

[0205] Data preprocessing: Statistics are performed on the sample data, with the data dimensions being 8640 rows x 364 columns. Outliers (such as negative values) and missing values ​​(null values) are checked and supplemented using a sliding average to complete data cleaning.

[0206] Data dimensionality reduction based on principal component analysis (PCA): PCA is a commonly used dimensionality reduction technique. Its primary purpose is to eliminate redundancy and correlation, improve computational efficiency, and avoid the "curse of dimensionality." It also largely preserves the core features of the source data, making subsequent model training and analysis more efficient and accurate. In this embodiment of the present invention, PCA is used to reduce the dimensionality of the source data. By retaining the principal components that explain 95% of the variance, the data dimension is reduced while preserving as much important information as possible.

[0207] Optimal number of clusters: The dimensionality-reduced data is iteratively solved using the silhouette coefficient method, with the overall silhouette coefficient as the recommended basis to obtain the optimal number of clusters. The optimal number of clusters in the embodiment of the present invention is: k =4, that is, the typical scenario of the source data after clustering reduction should be 4 categories.

[0208] Clustering and reduction: clustering and reducing the original wind power output data based on the k-means algorithm;

[0209] After preparing the input file and entering the corresponding parameters as prompted, click the corresponding module calculation button in sequence to complete the calculation task. Finally, the required key parameters and corresponding output files will be generated. This software includes four calculation results, each of which includes a calculation table and an image. The table format is "csv" and the image format is "png".

[0210] Module 1: Clustering and Reduction of New Energy Output Scenario

[0211] There are two types of pictures, including the best clustering iteration results of the elbow method and the silhouette coefficient method and the typical scene results. Figure 4 and 5 The typical scenario number solution results are: Figure 6 The typical scenario result diagram is divided into 4 typical scenarios, with the horizontal axis representing time and the vertical axis representing the output of new energy;

[0212] The calculation table is the output sequence (output power) of the typical scenario, which is mainly the time series of the typical scenario after cluster reduction and the output power series of the new energy at the corresponding moment. Figure 7 Typical scenario output sequence calculation results.

[0213] The cluster reduction results are as follows Figure 4 and 5 As shown, a total of four typical scenarios are included. Based on the optimization analysis results of dual-criteria clustering, the elbow method analysis shows that when the number of clusters increases to 4, the curve of the sum of squares of the distance from the sample to the centroid shows a significant "elbow" inflection point, and its subsequent changes tend to be gentle; the silhouette coefficient method verification shows that when k=4, the system obtains a peak silhouette coefficient, and the coefficient attenuation between adjacent cluster numbers is large. The consistency verification of the two methods shows that when the number of new energy output scenarios is determined to be 4, the optimal balance between intra-class compactness and inter-class separation can be achieved. As can be seen from the figure, the power fluctuations of various typical scenarios are relatively obvious, among which typical scenario 4 has the most obvious fluctuations, typical scenarios 1-2 have relatively weak fluctuations, and typical scenario 3 has the weakest fluctuations. The typical original output scenarios of new energy are obtained through cluster reduction, and the following calculations are based on typical scenarios.

[0214] The second step of the aforementioned method for calculating the dynamic heat generation power of lithium batteries used to smooth out the power fluctuations of new energy sources is adaptive wavelet packet decomposition. The specific steps are as follows:

[0215] Determining the necessity of energy storage configuration: Based on the volatility constraint (Equation 9), determine whether energy storage is required for each typical scenario. Based on the aforementioned definition of new energy volatility and volatility limits, calculate the volatility of each typical scenario and determine whether energy storage is required to meet the grid-connected power requirements. The judgment results on the necessity of energy storage configuration for each scenario include the optimal wavelet basis function and the optimal number of decomposition layers.

[0216] The results show that Typical Scenario 3 does not require energy storage to meet the renewable energy grid-connected volatility limit. Typical Scenarios 1, 2, and 4 all require energy storage to meet the renewable energy grid-connected volatility limit.

[0217] Module 2: Adaptive Discrete Wavelet Transform

[0218] Spectral analysis: Using DFT to analyze the spectrum of typical scenarios requiring energy storage, the results show that wind power output is primarily concentrated in the low-frequency portion, with relatively low energy in the high-frequency portion. This aligns with wind speed characteristics: high-frequency wind speed variations have small amplitudes, while low-frequency wind speed variations have large amplitudes. Therefore, the low-frequency signal is used as the expected power value for grid-connected renewable energy, while the high-frequency power signal is smoothed by the energy storage system. This achieves smooth grid-connected power while also considering the impact on energy storage system performance.

[0219] There are two types of pictures, such as Figures 8-11 The chart includes analysis results of amplitude-frequency characteristics for various typical scenarios, as well as a spectrum of amplitude-frequency characteristics for typical scenarios (including full-frequency distribution and detailed localized representation of low-frequency segments). The horizontal axis represents frequency, and the vertical axis represents amplitude. Spectral analysis shows that renewable energy output exhibits significant energy concentration in low-frequency bands, while the amplitude distribution tends to be stable in high-frequency bands.

[0220] like Figure 12-14 The chart compares the output sequence of renewable energy before and after power stabilization. The vertical axis represents power (MW), and the horizontal axis represents time. The decomposition and reconstruction diagrams for Typical Scenario 1, Typical Scenario 2, and Typical Scenario 4 show the results of the renewable energy output sequence before and after power stabilization (Typical Scenario 3 does not require energy storage for power stabilization). The yellow line shows the result after power stabilization, and the blue line shows the result before power stabilization. It can be clearly seen that the output trend of renewable energy after power stabilization is much smoother than before power stabilization.

[0221] Adaptive wavelet packet decomposition: Due to space limitations, this embodiment uses typical scenario 1 as an example for illustration, and the specific steps for other scenarios are similar. Through adaptive wavelet packet decomposition, the optimal number of decomposition layers is determined. When the fluctuation characteristics of the low-frequency component meet the grid-connected constraint conditions, the appropriate number of decomposition layers can be determined. In this embodiment, for typical scenario 1, the optimal wavelet basis function is selected as 'db25', and the optimal number of decomposition layers is determined to be 4 layers. In this step, the expected grid-connected power of new energy can be obtained at the same time. and the expected grid-connected power command of the energy storage system .

[0222] The calculation table shows the amplitude-frequency characteristics calculation results for each typical scenario and the output sequence calculation results after discrete wavelet decomposition and reconstruction. The amplitude-frequency characteristics result graph shows the frequency and amplitude calculation results for typical scenarios 1, 2, and 4 (energy storage scenario required). The first column shows the frequency data, and the second column shows the amplitude data.

[0223] The discrete wavelet decomposition results include the original renewable energy power, the reconstructed renewable energy power, and the expected energy storage output power. These are typical scenarios 1, 2, and 4, respectively. The first column shows the original renewable energy output data, the second column shows the stabilized renewable energy output data, and the third column shows the power required to compensate for the energy storage (corresponding to the expected energy storage power).

[0224] Module 3: Energy Storage Configuration Model

[0225] The third step of the aforementioned method for calculating the dynamic heat generation power of lithium batteries for smoothing out fluctuations in new energy power is to construct an energy storage configuration model. The specific steps are as follows:

[0226] Energy storage configuration model optimization goal: Construct a dual-objective optimization model with the specific requirement that the power fluctuation rate of the combined grid-connected "wind power + energy storage" is lower than the limit value on the 1-minute and 10-minute time scales (using sliding values). Since the rated capacity of wind power in the embodiment of the present invention is 80MW, there is .

[0227] Energy storage configuration model constraints: To ensure safe and stable operation of the energy storage system, the model constraints in the embodiment of the present invention are as follows.

[0228] (1) Energy storage cycle efficiency constraint. Combined with the actual comprehensive cycle efficiency of the lithium battery energy storage system, the energy storage cycle efficiency is taken as ;

[0229] (2) Power balancing constraint, the power downshift is ;

[0230] (3) The upper and lower limits of charge and discharge are ;

[0231] (4) The initial state of charge constraint is .

[0232] Energy storage configuration model output: By solving the model, the energy storage configuration results under typical scenario 1 can be obtained.

[0233] (1) The rated power of the configured energy storage system is ;

[0234] (2) The rated capacity of the configured energy storage system is ;

[0235] (3) Comparison of grid-connected power fluctuations: Figure 18 It can be seen that the results of the solution based on this model meet the optimization objectives—— , , the grid-connected fluctuation of the model built by the present invention after configuring energy storage is significantly improved compared with the original fluctuation;

[0236] like Figure 18The figure shows a comparison of the 1-minute and 10-minute grid-connected fluctuations of renewable energy sources before and after energy storage deployment. The red line represents the maximum fluctuation specified by the standard. The gray line shows the fluctuation without energy storage. It can be clearly seen that the maximum fluctuations of 1 minute and 10 minutes are 42.5% and 58.4%, respectively. At some points, these fluctuations exceed the red line, indicating non-compliance with the standard and potentially leading to power curtailment. The blue line shows the fluctuation with energy storage deployment. It can be clearly seen that the maximum fluctuations of 1 minute and 10 minutes are reduced to 9.2% and 29.6%, respectively, meeting the standard. Without energy storage, the maximum fluctuations of 1 minute and 10 minutes are 42.5% and 58.4%, respectively. With energy storage deployment, the maximum fluctuations of 1 minute and 10 minutes are 9.2% and 29.6%, respectively. The improvements in the maximum fluctuations of 1 minute and 10 minutes are 78.4% and 49.3%, respectively, validating the effectiveness of the model. Therefore, this solution can be applied to energy storage deployment in scenarios where renewable energy power fluctuations are mitigated.

[0237] The information in the output result window includes charge and discharge efficiency, energy storage rated power, energy storage rated capacity, and initial state of charge. Figure 20 As shown in the figure, it includes the optimal polynomial order, the fitted polynomial equation, and the path to save the dynamic heat generation power calculation results.

[0238] There are two types of pictures, including energy storage system power sequence curves (energy storage system expected power, primary power command and grid-connected power command) such as Figure 16 As shown in the figure, the vertical axis represents power (MW) and the horizontal axis represents time (hours). The expected power command is connected to the results obtained by discrete wavelet decomposition. The primary power command of the energy storage system is the energy storage output sequence obtained by considering the energy storage system's charging and discharging efficiency. The grid-connected power command of the energy storage system is the energy storage output sequence obtained by considering power balance.

[0239] Energy storage system energy change curve (energy storage system cumulative energy curve and SOC curve) as shown in the figure Figure 17 As shown in the figure, the vertical axis represents energy (MWH) in the energy storage system's cumulative energy curve, and the horizontal axis represents time (hours). The vertical axis represents SOC (%) in the energy storage system's SOC curve, and the horizontal axis represents time (hours). The cumulative energy curve represents the actual energy change in MWh, while the SOC curve represents the per-unit value of energy change. Therefore, the two curves have the same trend.

[0240] Module 4: Calculation of dynamic thermal power of lithium batteries

[0241] The fourth step of the aforementioned method for calculating the dynamic heat generation power of lithium batteries for smoothing out new energy power fluctuations is to develop a dynamic heat generation power calculation model for energy storage batteries that takes into account electrochemical-thermal coupling. The specific steps are as follows:

[0242] Battery cell parameter input: In the embodiment of the present invention, lithium iron phosphate energy storage batteries are used. The parameters of the energy storage system are: the rated capacity of the single cell is 314Ah, the rated voltage is 3.2V, and the cell volume is 173mm×54mm×200mm (length×width×height) = 186840mm 3 The series-parallel connection mode of the energy storage system is 12P416S, and the operating environment temperature is 25℃.

[0243] According to the manufacturer's actual test results of DC internal resistance and SOC, the calculation requires the user to provide SOC-DC internal resistance measured data based on the actual situation of the energy storage lithium battery to be configured. Figure 19 As shown in the figure, it shows that the DC internal resistance changes with the SOC, rather than being constant.

[0244] The calculation information of the output window results includes the best fitting polynomial order-polynomial function relationship-calculation result storage information, etc. Figure 20 shown.

[0245] The functional relationship between DC internal resistance and SOC is fitted by numerical fitting. The best fit is a 7th-order polynomial function, and its functional relationship is as follows:

[0246] ;

[0247] The best polynomial fitting relationship diagram containing SOC-R is shown in the figure below. Figure 21 shown.

[0248] The reason for using polynomial fitting for cell internal resistance: Based on the changing trend of lithium battery cell internal resistance and SOC, academia and industry usually use polynomial fitting to numerically describe their relationship. Therefore, the same method is also used for data verification in the patent of this invention.

[0249] Determining the numerical values ​​of polynomial coefficients: The determination method is relatively conventional, and polynomial fitting can be performed using methods such as the least squares method, Lagrange interpolation method, Newton interpolation method, and programming language. Among them, the least squares method finds the best function matching the data by minimizing the sum of squares of errors; the Lagrange interpolation method constructs a polynomial function from known data points and makes predictions at unknown points. There are also many ways to implement it in programming languages. MATLAB provides the polyfit function, which can easily perform polynomial fitting.

[0250] The embodiments mentioned in the present invention are operated based on Python. The specific polynomial order is determined by iterative experiments, i.e., 2nd order, 3rd order... 7th order... 10th order. The specific order is determined by judging the error between the experimental point and the curve. The polynomial coefficients are obtained by fitting using np.polysit in the numpy library of Python data processing.

[0251] Electrochemical-thermal coupling model: Based on the aforementioned formula (25), an electrochemical-thermal coupling model is constructed. In order to solve the dynamic heat generation power of the lithium battery, the SOC, DC internal resistance, and operating current at any time must be solved first.

[0252] (1) SOC at any time: Based on the energy storage configuration results and the aforementioned formula (28), the SOC at any time is solved;

[0253] (2) DC internal resistance at any time: Substitute the polynomial function according to the SOC at any time to solve the DC internal resistance at any time;

[0254] (3) Working current at any time: Based on the above formula (27), solve the working current at any time;

[0255] (4) Dynamic heat generation power: Substituting the existing parameters into the electrochemical-thermal coupling model, the dynamic heat generation power during the operation cycle of the energy storage lithium battery can be solved. Figure 22 As shown in the figure. Following the solution steps, the dynamic heat generation power of the lithium battery in the scenario of smoothing new energy power fluctuations can be solved, as shown in the figure. The vertical axis represents heat generation power, and the horizontal axis represents time. It can be clearly seen that the time-varying characteristics of heat generation power are more obvious over time, and the results also verify the necessity of solving the dynamic heat generation power in this paper.

[0256] Dynamic heat generation power results: The dynamic heat generation power results of lithium batteries are as follows: Figure 22 As shown, combined Figure 18 Analysis of fluctuations in new energy output Figure 22 Analysis of the dynamic heat generation power trend of lithium batteries shows that during the 4h-14h period, due to the large fluctuation of wind power output at the corresponding time in this scenario, the energy storage demand increases, resulting in a large heat generation power; at other times, the wind power output is relatively stable, the energy storage demand is low, and the heat generation power is also small.

[0257] This method not only optimizes the energy storage configuration to enhance the stability of renewable energy access to the grid, but also further quantifies the dynamic thermal behavior of lithium batteries when used to smooth renewable energy power fluctuations, effectively improving the safety and reliability of the energy storage system and providing a data basis for lithium battery temperature change prediction and thermal management.

Claims

1. A method for calculating the dynamic heat generation power of lithium batteries for smoothing out fluctuations in new energy power, characterized in that: The steps include: S1. An improved K-means algorithm based on the silhouette coefficient method is used to perform cluster analysis on historical output data of new energy sources to extract typical output scenarios. S2. Taking the new energy grid-connected power fluctuation rate meeting the fluctuation rate constraint as the optimization goal, an adaptive wavelet packet decomposition algorithm is used to decompose and reconstruct the typical output scenario and determine the expected grid-connected power command; S3. Based on the expected grid-connected power command, considering the rated power configuration and the rated capacity configuration, constructing an energy storage configuration model and solving the energy storage configuration model to obtain an energy storage configuration result; S4. Solving the initial state of charge based on the energy storage configuration result, and then combining the energy storage calculation formula of the energy storage system at any time to obtain the state of charge calculation formula at each time during the operation cycle; S5. Construct functional relationships between the dynamic changes of DC resistance and working current with the state of charge respectively. Then, combine the state of charge calculation formula at each moment in the operating cycle, calculate the dynamic heat generation power of the lithium battery through the Bernardi heat generation rate formula, and use the dynamic heat generation formula of the lithium battery under energy storage to smooth fluctuations to characterize the characteristic mapping relationship between the dynamic heat source and the state of charge and time.

2. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out the fluctuation of new energy power according to claim 1, characterized in that: Said S1 comprises: S11. Obtain historical output data of new energy within a set time interval and perform statistics, and verify and supplement abnormal values ​​and missing values ​​through a sliding average; S12, new energy historical output data after simplification through principal component analysis; S13, setting the number range of cluster centers and performing iterative solution through the silhouette coefficient method to obtain the optimal number of clusters; S14. Based on the optimal number of clusters, a K-means algorithm is used to obtain a typical scenario and generate a new energy output power of the typical scenario.

3. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 1, characterized in that: The S2 includes: S21. Define the volatility of new energy and consider the volatility constraint to determine whether energy storage needs to be configured and obtain the configuration result. S22. Based on the configuration result, the grid-connected power of the new energy and the expected grid-connected power instruction are obtained by using the optimization target of the grid-connected power fluctuation rate of the new energy to meet the fluctuation rate constraint and the adaptive wavelet packet decomposition algorithm.

4. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 3, characterized in that: The configuration results include: if the new energy output power of the typical new energy scenario meets the grid-connected volatility constraint, there is no need to configure energy storage and the system can be directly connected to the grid; if the grid-connected volatility constraint is not met, energy storage is configured.

5. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 4, characterized in that: In the S22, a spectrum analysis is performed on the output power that needs to be smoothed by power fluctuation through discrete Fourier transform to obtain the amplitude-frequency characteristic results under the output scenario. At a fixed scale, a wavelet packet function is defined to obtain a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm. The renewable energy output power is decomposed into high-frequency components and low-frequency components. The low-frequency component is judged using the grid-connected fluctuation rate constraint. When the amplitude fluctuation of the low-frequency component meets the grid-connected fluctuation constraint for the first time, the optimal number of decomposition layers is determined. At the same time, the low-frequency component is used as the expected grid-connected power, and the difference between the expected grid-connected power and the original output power of the renewable energy is used as the expected grid-connected power instruction.

6. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 1, characterized in that: In said S3, solving the energy storage configuration model includes solving the energy storage system rated power and the energy storage system rated capacity; Solving the energy storage system rated power includes solving the energy storage system expected power based on the expected grid-connected power instruction, taking into account the energy storage comprehensive cycle efficiency and energy storage charging and discharging efficiency, obtaining the primary power instruction of the energy storage system, and then considering the satisfaction of the power balance constraint to obtain the energy storage system rated power; Solving the rated capacity of the energy storage system includes solving the energy fluctuation of the energy storage system during its operation cycle compared to its initial state based on the primary power command, obtaining a calculation formula for the energy storage system's stored energy at any time, and then obtaining a state of charge expression while satisfying the upper and lower limit constraints of the energy storage system's charge and discharge. The energy storage system's rated capacity is obtained by calculating the ratio of the difference between the maximum energy fluctuation of the energy storage system during its entire operation cycle and the difference between its upper and lower limits of charge and discharge.

7. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out the fluctuation of new energy power according to claim 1, characterized in that: The initial charge state in S4 is calculated specifically by the following expression: ； Where, represents the initial charge state, is the rated capacity of the energy storage system, Indicates the upper limit of energy storage charging, Indicates the lower limit of energy storage discharge, It is the storage energy of the energy storage system at any time.

8. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 1, characterized in that: In S4, the specific expression of DC resistance is: ； Where, are the polynomial coefficients; is the DC resistance, is an exponential function of the state of charge.

9. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 1, characterized in that: In S4, the specific expression of DC resistance is: ； Where, is the rated capacity of the single cell; when t=1, ; The sampling interval for renewable energy output data; This is the data of parallel connection of battery cells; is the time corresponding to the sampling point; is the initial state of charge; When calculating the DC resistance, the series and parallel connection of the energy storage cells is taken into account, and it is assumed that the charge state of each cell in the energy storage system remains consistent; it is assumed that the charging and discharging power of the lithium battery is constant during the sampling period.

10. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 8, characterized in that: The calculation formula of the state of charge at each moment in the operation cycle is specifically expressed as follows: ； Where, is the rated capacity of the energy storage system; is the time point corresponding to the end of the sampling period; is the grid-connected power corresponding to the energy storage at time t.

11. The method for calculating the dynamic heat generation power of a lithium battery for smoothing out fluctuations in new energy power according to claim 1, characterized in that: The dynamic heat generation formula of lithium batteries under the energy storage to smooth fluctuations is specifically expressed as follows: ； Where X(·), Y(·), and Z(·) are the abstract functions of dynamic heating power, operating current, and DC internal resistance, respectively; F(·) is the abstract function of dynamic heating power obtained after solution.

Images