Super capacitor module voltage balancing method and device, terminal and medium

Abstract

The invention discloses a super-capacitor module voltage balancing method and device, a terminal and a medium, and the method comprises the steps: collecting the terminal voltage value of each single capacitor in a super-capacitor module according to a to-be-balanced super-capacitor module; then, according to the terminal voltage value, constructing a time-varying fractional order capacitor model to accurately describe the non-ideal characteristics of the super capacitor, and obtaining an adaptive parameter set comprising a fractional order parameter, an equivalent series resistance parameter, a capacitor capacity parameter and an aging factor parameter in combination with historical equilibrium data; and then a differential game decision-making device is adopted to carry out multi-target collaborative optimization on capacitance balance, power grid reactive power demand and energy consumption limitation, and a current instruction considering capacitance balance, power grid support and energy consumption constraint is dynamically generated by solving Nash equilibrium points of three optimization targets, so that precise balance of the voltage of the super-capacitor module is realized, and the accuracy of the voltage balance of the super-capacitor module is improved. The technical problem of insufficient model precision in a traditional strategy is solved, and the reliability of the system under complex working conditions is improved.

Description

Technical Field

[0001] The present application relates to the field of supercapacitor technology, and in particular to a supercapacitor module voltage balancing method, device, terminal, and medium. Background Art

[0002] In practical applications, supercapacitor modules for energy storage-type reactive power compensation devices typically require the use of several single capacitors in series. Due to differences in parameters (such as capacitance, ESR, and leakage current) between the individual capacitors in series, high-current discharge can easily lead to voltage imbalance across the individual capacitors, causing some capacitors to overcharge or over-discharge, impacting the life of the capacitors and the reliability of the entire circuit.

[0003] In order to reduce the difference in voltage between supercapacitor cells in the system, in large-scale engineering applications where supercapacitors are connected in series and parallel, a management system is usually used to monitor the voltage of each supercapacitor cell and balance the supercapacitor cell voltage. However, the existing supercapacitor module voltage balancing control strategy is based on feedback control implemented between the voltage difference between the two supercapacitor cells with the maximum and minimum voltages in the module. Through charge and discharge control, the voltage difference between the maximum and minimum voltages is made lower than a preset threshold. This control strategy has simple logic and low implementation difficulty, but there is a technical problem of low voltage balancing accuracy. Summary of the Invention

[0004] The present application provides a supercapacitor module voltage balancing method, device, terminal and medium, which are used to solve the technical problem of low balancing accuracy in existing supercapacitor module voltage balancing strategies.

[0005] To solve the above technical problems, the first aspect of the present application provides a supercapacitor module voltage balancing method, comprising:

[0006] According to the supercapacitor module to be balanced, collecting the terminal voltage value of each single capacitor in the supercapacitor module;

[0007] A time-varying fractional-order capacitor model is constructed according to the terminal voltage value, and an adaptive parameter set is obtained based on the time-varying fractional-order capacitor model and historical equalization data, wherein the adaptive parameter set includes: a fractional-order parameter, an equivalent series resistance parameter, a capacitor capacity parameter, and an aging factor parameter;

[0008] The adaptive parameter set and the grid status signal are input into a preset differential game decision maker to determine the balancing current command value of each single capacitor through the game operation of the differential game decision maker, so as to generate the voltage balancing control command of the supercapacitor module according to the balancing current command value, wherein the differential game decision maker includes a preset multi-party differential game model, and the multi-party differential game model includes: a capacitance optimization target for optimizing the balance of the supercapacitor module, a grid optimization target for optimizing the reactive power demand of the grid, and a balancing optimization target for optimizing the balancing loss.

[0009] Preferably, constructing a time-varying fractional-order capacitance model according to the terminal voltage value, and obtaining an adaptive parameter set based on the time-varying fractional-order capacitance model and historical equalization data includes:

[0010] According to the terminal voltage value, a time-varying fractional-order capacitance model is constructed in accordance with a four-element fractional-order equivalent circuit model construction method;

[0011] Based on the time-varying fractional-order capacitor model, combined with the sub-frequency band impedance characteristics, dynamic calculation of model component parameters is performed to determine the values ​​of fractional-order parameters, equivalent series resistance parameters, and capacitor capacity parameters;

[0012] determining, based on the historical equalization data, a deviation between a current terminal voltage change rate and a historical terminal voltage change rate, and determining an aging factor parameter based on the deviation;

[0013] An adaptive parameter set is obtained according to the values ​​of the fractional order parameter, the equivalent series resistance parameter, the capacitance parameter and the aging factor parameter.

[0014] Preferably, constructing a time-varying fractional-order capacitance model according to the terminal voltage value, and obtaining an adaptive parameter set based on the time-varying fractional-order capacitance model and historical equalization data includes:

[0015] Performing high-frequency ripple texture processing according to the terminal voltage value to obtain a fundamental voltage component and a ripple voltage component;

[0016] Calculating the impedance phase shift angle of each single capacitor according to the fundamental voltage component, so as to determine the fractional order parameter according to the impedance phase shift angle;

[0017] Extracting the amplitude attenuation rate and frequency response characteristics of the ripple voltage component to respectively determine equivalent series resistance parameters and capacitance parameters;

[0018] Constructing a time-varying fractional-order capacitance model based on the fractional-order order parameter, the equivalent series resistance parameter, and the capacitance parameter in combination with a preset fractional-order state equation;

[0019] Based on the historical balance data, calculating the energy loss rate increment of the supercapacitor module, and determining an aging factor parameter based on the energy loss rate increment;

[0020] An adaptive parameter set is obtained according to the values ​​of the fractional order parameter, the equivalent series resistance parameter, the capacitance parameter and the aging factor parameter.

[0021] Preferably, the step of inputting the adaptive parameter set and the grid state signal into a preset differential game decision maker to determine the balancing current command value of each single capacitor through a game operation of the differential game decision maker includes:

[0022] Calculating the per-unit voltage difference of the power grid according to the power grid state signal and a preset per-unit voltage difference calculation formula;

[0023] Calculating, based on the adaptive parameter set, a cumulative sum of capacitor-voltage deviations between a voltage parameter of each single capacitor and an average voltage parameter;

[0024] A preset differential game decision maker is inputted according to the accumulated sum of the capacitor voltage deviations and the per-unit voltage difference of the power grid, so as to perform a Nash equilibrium solution operation through the differential game decision maker to determine the balancing current command value of each single capacitor, wherein the optimization objectives of the multi-party differential game model include: a capacitor optimization objective for minimizing the accumulated sum of the capacitor voltage deviations, a power grid optimization objective for closed-loop control of the per-unit voltage difference of the power grid, and a balancing optimization objective for minimizing the total balancing current.

[0025] Preferably, the step of inputting the adaptive parameter set and the grid state signal into a preset differential game decision maker to determine the balancing current command value of each single capacitor through a game operation of the differential game decision maker includes:

[0026] Calculating a reactive power demand intensity coefficient based on the grid status signal and a preset reactive power demand intensity calculation formula;

[0027] Calculating, based on the adaptive parameter set, the product of the deviation between the voltage parameter of each single capacitor and the average voltage parameter and the aging factor parameter to obtain the aging imbalance of each single capacitor;

[0028] A preset differential game decision maker is input according to the aging imbalance and the reactive demand intensity coefficient, so as to perform a Nash equilibrium solution operation through the differential game decision maker to determine the balanced current command value of each single capacitor, wherein the optimization objectives of the multi-party differential game model include: a capacitor optimization objective for minimizing the aging imbalance, a power grid optimization objective for minimizing the reactive demand intensity coefficient, and a balanced optimization objective for minimizing energy transfer loss.

[0029] Preferably, generating the voltage balancing control instruction of the supercapacitor module according to the balancing current instruction value includes:

[0030] The balancing current command value is converted into a corresponding PWM drive signal through a sliding mode controller, and the PWM drive signal is used as a voltage balancing control command for the supercapacitor module.

[0031] Preferably, the objective function in the sliding mode controller is specifically:

[0032] or

[0033] Where, is the sliding mode value of the single capacitor i, is the terminal voltage value of single capacitor i, is the average voltage of each single capacitor, is the balance coefficient between voltage balance and grid demand, is the voltage normalized to the voltage difference, is the grid voltage command value, is the actual value of the grid voltage, is the coupling weight coefficient, is the balancing current command value of the single capacitor, is the actual value of the balancing current of the single capacitor.

[0034] At the same time, the second aspect of the present application provides a supercapacitor module voltage balancing device, comprising:

[0035] A terminal voltage acquisition unit, configured to acquire the terminal voltage value of each single capacitor in the supercapacitor module to be balanced;

[0036] an adaptive parameter set generating unit, configured to construct a time-varying fractional-order capacitor model according to the terminal voltage value, and obtain an adaptive parameter set based on the time-varying fractional-order capacitor model and historical equalization data, wherein the adaptive parameter set includes: a fractional-order parameter, an equivalent series resistance parameter, a capacitance parameter, and an aging factor parameter;

[0037] A game decision processing unit is configured to input the adaptive parameter set and the grid status signal into a preset differential game decider, so as to determine the balancing current command value of each single capacitor through the game operation of the differential game decider, and to generate a voltage balancing control command for the supercapacitor module according to the balancing current command value, wherein the differential game decider includes a preset multi-party differential game model, and the multi-party differential game model includes: a capacitance optimization target for optimizing the balance of the supercapacitor module, a grid optimization target for optimizing the reactive power demand of the grid, and a balancing optimization target for optimizing the balancing loss.

[0038] A third aspect of the present application provides a supercapacitor module voltage balancing terminal, comprising: a memory and a processor;

[0039] The memory is used to store program code, and the program code corresponds to the supercapacitor module voltage balancing method provided in the first aspect of the present application;

[0040] The processor is used to read and execute the program code to implement the above-mentioned supercapacitor module voltage balancing method.

[0041] A fourth aspect of the present application provides a computer-readable storage medium, in which a program code is stored. The program code is used to be read and executed by a processor to implement the supercapacitor module voltage balancing method provided in the first aspect of the present application.

[0042] It can be seen from the above technical solutions that this application has the following advantages:

[0043] The solution provided in this application first collects the terminal voltage values ​​of each single capacitor in the supercapacitor module based on the supercapacitor module to be balanced; then, based on the terminal voltage values, a time-varying fractional-order capacitor model is constructed to accurately describe the non-ideal characteristics of the supercapacitor, and based on the time-varying fractional-order capacitor model and historical balancing data, an adaptive parameter set including fractional-order parameters, equivalent series resistance parameters, capacitor capacity parameters and aging factor parameters is obtained, and then a differential game decision maker is used to perform multi-objective collaborative optimization of capacitor balancing, grid reactive demand and energy consumption restriction. By solving the Nash equilibrium points of the three optimization objectives, a current instruction that takes into account capacitor balancing, grid support and energy consumption constraints is dynamically generated, thereby achieving accurate balancing of the supercapacitor module voltage, solving the technical problem of insufficient model accuracy in traditional strategies, and improving the reliability of the system under complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 This is a flow chart of an embodiment of a supercapacitor module voltage balancing method provided in this application.

[0046] Figure 2 This is a structural diagram of an embodiment of a supercapacitor module voltage balancing device provided in this application.

[0047] Figure 3 This is a structural diagram of an embodiment of a supercapacitor module voltage balancing terminal provided in this application. DETAILED DESCRIPTION

[0048] In the prior art, the supercapacitor module of the energy storage type reactive compensation device is usually composed of multiple single capacitors connected in series. Due to differences in manufacturing processes and material properties, there are inconsistencies in the capacitance, equivalent series resistance and other parameters of the single capacitors, resulting in uneven voltage distribution of each single capacitor during the charging and discharging process. The traditional voltage equalization control strategy is based on feedback regulation of the difference between the highest voltage and the lowest voltage in the module, and the voltage difference is lower than the preset threshold through simple charging and discharging. Although this method is simple to implement, it ignores the coordinated optimization of the reactive power demand of the power grid and the balanced energy consumption, and only focuses on the absolute value control of the voltage difference within the module. In actual operation, this control method will cause the balancing circuit to operate frequently, which not only increases energy loss, but also fails to effectively respond to the reactive compensation needs of the power grid, resulting in insufficient balancing accuracy and difficulty in adapting to performance requirements under complex working conditions.

[0049] In view of this, embodiments of the present application provide a supercapacitor module voltage balancing method, device, terminal, and medium, which are used to solve the technical problem of low balancing accuracy in existing supercapacitor module voltage balancing strategies.

[0050] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0051] First, a detailed description of an embodiment of a supercapacitor module voltage balancing method provided by this application is as follows:

[0052] See also Figure 1 The supercapacitor module voltage balancing method provided in the embodiment of the present application includes:

[0053] Step 101: According to the supercapacitor module to be balanced, the terminal voltage value of each single capacitor in the supercapacitor module is collected.

[0054] Step 102: construct a time-varying fractional-order capacitance model according to the terminal voltage value, and obtain an adaptive parameter set based on the time-varying fractional-order capacitance model and historical equalization data;

[0055] The adaptive parameter set includes: fractional order parameters, equivalent series resistance parameters, capacitance parameters and aging factor parameters.

[0056] Step 103: Input the adaptive parameter set and the grid state signal into a preset differential game decision maker to determine the balancing current command value of each single capacitor through the game operation of the differential game decision maker, and generate the voltage balancing control command of the supercapacitor module according to the balancing current command value;

[0057] Among them, the differential game decision maker includes a preset multi-party differential game model, which includes: a capacitance optimization target for optimizing the balance of the supercapacitor module, a grid optimization target for optimizing the reactive power demand of the grid, and a balance optimization target for optimizing the balanced loss.

[0058] It should be noted that the time-varying fractional-order capacitor model refers to a capacitor dynamic model established using fractional-order calculus theory, which can accurately describe the non-ideal characteristics and memory effects of supercapacitors. It can be specifically implemented through a four-element equivalent circuit combined with a fractional-order differential equation. The adaptive parameter set includes dynamic parameters that reflect the current working state of the capacitor. For example, the aging factor is determined by combining frequency-band impedance characteristic analysis with historical data regression. The differential game decider refers to a multi-objective optimizer based on Nash equilibrium theory. It constructs a game objective function that includes capacitor balance, grid reactive demand matching, and balanced loss, for example, using a dynamic weight allocation algorithm to coordinate conflicts between different optimization objectives.

[0059] Specifically, the method of this embodiment first collects the terminal voltage signal of each single capacitor in real time, and extracts the voltage ripple characteristics through time domain and frequency domain analysis. A dynamic model is constructed based on a four-element fractional-order equivalent circuit, and the aging factor is calculated in combination with the parameter degradation trend in the historical operation data to form an adaptive parameter set that reflects the current state of the capacitor. The parameter set and the grid voltage deviation signal are input into the differential game decision maker, which dynamically generates a current instruction that takes into account capacitor balance, grid support and energy consumption constraints by solving the Nash equilibrium point of the three optimization objectives. Finally, the current instruction is converted into a PWM drive signal through a sliding mode controller. This signal is used to drive a bidirectional DC / DC circuit to achieve directional energy transfer from high-voltage monomers to low-voltage monomers, and reduce transfer losses through aging factor compensation.

[0060] This solution accurately characterizes the dynamic characteristics of capacitors through a time-varying fractional-order model, and realizes adaptive parameter correction by combining the aging factor. The differential game mechanism takes the reactive power demand of the power grid as an independent optimization target, actively responds to changes in the power grid state during the balancing process, and avoids the interference of grid voltage fluctuations on the balancing system in traditional methods. At the same time, by introducing the balanced loss optimization target, the high energy consumption problem of the traditional active balancing circuit is effectively reduced. Through the above technical solution, the precise balancing of the supercapacitor module voltage is achieved, and the balancing accuracy can still be maintained during the aging process of the capacitor, and the energy consumption of the balancing circuit can be reduced while dynamically responding to changes in the reactive power demand of the power grid. This method solves the problems of insufficient model accuracy and single optimization target in traditional strategies, and improves the reliability and economy of the system under complex working conditions.

[0061] On the basis of the above basic embodiment, further, regarding step 102 mentioned in the above basic embodiment, its step process may further include: constructing a time-varying fractional-order capacitor model according to the terminal voltage value in accordance with the four-element fractional-order equivalent circuit model construction method; based on the time-varying fractional-order capacitor model, dynamically calculating the model element parameters in combination with the frequency-dividing impedance characteristics, and determining the numerical values ​​of the fractional-order parameters, equivalent series resistance parameters and capacitor capacitance parameters; based on historical equalization data, determining the deviation between the current terminal voltage change rate and the historical terminal voltage change rate, so as to determine the aging factor parameters based on the deviation; obtaining an adaptive parameter set according to the numerical values ​​of the fractional-order parameters, equivalent series resistance parameters, capacitor capacitance parameters and the aging factor parameters.

[0062] Among them, the four-element fractional-order equivalent circuit model refers to a circuit model composed of fractional-order capacitors, equivalent series resistors, diffuse resistors and diffuse capacitors. Specifically, the impedance spectrum measurement method combined with the nonlinear least squares fitting algorithm can be used to realize parameter identification, which is used to accurately characterize the non-ideal polarization effect and dynamic response characteristics of the supercapacitor. The band-segment impedance characteristics refer to the impedance amplitude and phase change law of the supercapacitor at different frequencies. Specifically, the swept-frequency impedance test combined with the frequency domain decomposition method can be used to realize feature extraction, which is used to dynamically adjust the model parameters for different working conditions. Dynamic calculation of model component parameters refers to iteratively updating the equivalent circuit parameters based on the terminal voltage data collected in real time. Specifically, it can be realized by using the recursive least squares method combined with the forgetting factor optimization algorithm to adapt to changes in the operating state of the supercapacitor. The aging factor parameter refers to a quantitative indicator that characterizes the degree of degradation of capacitor performance. Specifically, it can be obtained by comparing the current voltage change rate with the deviation rate of historical benchmark data, and is used to reflect the impact of capacitor aging on balancing control.

[0063] Specifically, the construction of the four-element fractional-order equivalent circuit model first obtains the complex impedance data of the supercapacitor at different frequencies through impedance spectroscopy testing, and then uses fractional-order calculus equations to establish an equivalent circuit model that includes the polarization diffusion process. The frequency band impedance characteristic analysis divides the operating frequency band into a low-frequency diffusion region, a medium-frequency polarization region, and a high-frequency capacitive region, which correspond to the contributions of diffusion resistance, diffusion capacitance, and equivalent series resistance, respectively. The dominant parameters of each frequency band are extracted through frequency domain decomposition. During the dynamic parameter calculation process, the recursive least squares method processes the real-time voltage data in a sliding window manner, and combines the forgetting factor to reduce the weight coefficient of historical data to achieve online updating of model parameters. The aging factor is mapped to an aging coefficient in the range of 0-1 by establishing a historical database of voltage change rate based on the ratio of the current rate to the historical reference rate.

[0064] For example, fractional orders : In the charge diffusion frequency band (300mHz-100Hz), the imaginary part of the impedance is obtained by constant current charge and discharge test : Apply 50Hz, 2A square wave current to monomer 1 and measure = , substitute into the formula: ; Fitting (A value close to 1 indicates a significant capacitance characteristic).

[0065] Equivalent series resistance :In the resistance characteristic frequency band (>100Hz), the real part of the impedance Approach :Monitor 1 at 100Hz =0.0025 , so =0.0025 .

[0066] Capacitance : In the main capacitor characteristic frequency band (10-100mHz), calculate by the low-frequency impedance imaginary part:

[0067] ,when ;

[0068] Aging factor is generated by comparing the current and historical terminal voltage change rate deviations to quantify the degree of aging: Calculate the voltage change rate: The voltage of monomer 1 rises from 2.65V to 2.71V in 10 seconds, the change rate Compared with historical data: when the cell was new, the change rate was 0.008V / s at the same current.

[0069] Aging Factor : ;

[0070] >0.2 indicates significant aging and the capacity parameter needs to be adjusted downward.

[0071]

[0072] The parameter integration output generates an adaptive parameter set and passes it to the decision maker in step 103. The adaptive parameter set can be found in Table 1:

[0073]

[0074] This solution uses a four-element fractional-order model to accurately describe the polarization diffusion effect, combines it with frequency-band impedance analysis to achieve dynamic parameter matching, and introduces an aging factor to quantify capacitor performance degradation, allowing the model parameters to track the actual working state of the capacitor in real time. Through this technical solution, the problem of voltage deviation accumulation caused by insufficient model accuracy in traditional balancing control is solved. The time-varying fractional-order capacitor model accurately reflects the nonlinear characteristics of the supercapacitor through dynamic parameter calculation. The frequency-band impedance analysis improves the parameter adaptability under different operating conditions. The introduction of the aging factor parameter effectively compensates for the impact of capacitor performance degradation on balancing control, thereby providing a more accurate model foundation for subsequent game decision-making.

[0075] Furthermore, based on the above-mentioned adaptive parameter set generation method, the present application further proposes inputting the adaptive parameter set and the grid status signal into a preset differential game decision maker, and determining the balanced current command value of each single capacitor through the game operation of the differential game decision maker, including: obtaining the grid normalized voltage difference according to the grid status signal combined with the normalized voltage difference calculation formula; calculating the cumulative sum of the capacitor voltage deviations of the voltage parameters of each single capacitor and the average voltage parameter according to the adaptive parameter set; inputting the cumulative sum of the capacitor voltage deviations and the grid normalized voltage difference into the differential game decision maker for Nash equilibrium solution operation to determine the balanced current command value.

[0076] Among them, the grid normalized voltage difference refers to the grid voltage command value With actual value The normalized difference can be calculated as: This parameter can quantify the degree to which the grid voltage deviates from the target value. The cumulative sum of capacitor voltage deviations refers to the sum of the absolute deviations between the voltages of each single capacitor and the average voltage of the module. The calculation formula is: , where N represents the number of single capacitors; this parameter reflects the degree of imbalance in the voltage distribution within the module by collecting voltage data point by point and calculating statistics. The differential game decider refers to an optimization decision module built based on dynamic game theory. Specifically, it can be implemented using a multi-objective dynamic programming algorithm. It makes collaborative decisions by constructing a game model that includes capacitor optimization objectives, power grid optimization objectives, and equilibrium optimization objectives. The Nash equilibrium solution operation refers to finding the optimal balance point of the strategies of all parties in the game. Specifically, it can be implemented by solving the Pareto frontier using an iterative optimization algorithm. This process can coordinate conflicts between different optimization objectives.

[0077] Specifically, the grid's per-unit voltage difference is calculated by real-time monitoring of the dynamic deviation between the grid bus voltage and the command value. Its value reflects the grid's urgency for reactive power compensation. The cumulative sum of capacitor voltage deviations is calculated by collecting real-time voltage data from each individual capacitor, calculating the absolute difference from the module's average voltage, and summing them up. A larger value indicates poorer voltage balance within the module. These two parameters are simultaneously input into the differential game decision maker, which then performs dynamic optimization based on a pre-set three-objective game model: the capacitor optimization objective minimizes the cumulative sum of capacitor voltage deviations to improve balancing accuracy; the grid optimization objective requires closed-loop control of the grid's per-unit voltage difference to meet reactive power requirements; and the balancing optimization objective minimizes the sum of balancing currents to reduce energy consumption. Using a Nash equilibrium solver, the decision maker outputs a balancing current command value that dynamically balances these three objectives. This command value ensures voltage balance within the module while also balancing the grid's reactive power compensation needs and system operating efficiency.

[0078] For example, assume there is a 10kV distribution station with a storage-type reactive power compensation device, where the supercapacitor module consists of 60 cells (3000F / 2.7V) connected in series. The input data includes:

[0079] Adaptive parameter set: Monomer 1: Fractional order =0.92, equivalent resistance =2.5m ,capacity =2239F (aging factor =0.25), ..., monomer 60: =0.89, =2.8m , =2600F( =0.18);

[0080] Grid status signal , where the grid voltage command value is The actual value of the grid voltage can be obtained from the grid dispatching system through the ModbusTCP protocol; The voltage transformer (Rogowski coil, precision 0.2 ), the output is converted to the actual voltage value by the RMS conversion circuit

[0081] Per-unit difference ;

[0082] Modeling of differential game decision makers, which transforms the equilibrium problem into a three-objective optimization problem:

[0083] Goal 1: Minimize the voltage difference between capacitors Minimize voltage variance ;

[0084] Goal 2: Matching grid reactive power demand Tracking per-unit voltage difference , assuming the target value =0.01pu (i.e. 15kVar reactive power is required);

[0085] Goal 3: Limiting the energy consumption of balancing circuits Constrained balanced current sum (avoid overloading of DC / DC circuits), where is the balancing current of single capacitor i;

[0086] Examples of optimization functions could be:

[0087] Among them, the weight coefficient is: =0.6 (voltage balancing priority), =0.3 (reactive matching), =0.1 (energy consumption limit);

[0088] Dynamic game solving process, participants: Grid side: is the state variable, and the goal is to minimize Capacitor bank: Based on single cell voltage is the state variable, and the goal is to minimize .

[0089] Strategy interaction: High-pressure monomer (monomer 1: =2.71V) needs to be discharged, current instruction . Low pressure monomer (monomer 60: =2.58V) needs to be charged, Grid reactive power gap ( =0.013) requires the module's overall output current increment A.

[0090] Nash equilibrium solution: Iterative solution is obtained by gradient descent method, and the convergence condition is voltage variance < 0.01V 2 and ≤0.01pu. After 3 iterations, the following results are obtained: Cell 1 balancing current: -12.5A (discharge); Cell 60 balancing current: +10.2A (charge); The currents of the remaining cells: (fine-tuning);

[0091] Output instructions and performance verification, balanced current instruction set:

[0092] {monomer 1: -12.5A, monomer 2: +0.8A, ..., monomer 60: +10.2A};

[0093] Goal achievement:

[0094] The voltage difference is reduced from 0.13V to 0.05V ( );

[0095] Reduced to 0.009pu (satisfying pu requirements);

[0096] Total balancing current A (energy limited).

[0097] This solution transforms the voltage balancing problem into a dynamic optimization process by constructing a multi-objective game model. While ensuring the module balancing accuracy, it actively responds to the grid's reactive power demand and constrains the energy consumption of the balancing circuit, achieving a coordinated improvement in multi-dimensional performance. Through the above technical solution, it is possible to dynamically balance the relationship between voltage balancing accuracy, grid reactive power compensation demand, and system energy consumption during the discharge process of the supercapacitor module. When the grid voltage fluctuation causes a sudden increase in reactive power demand, the system prioritizes responding to the grid optimization target to adjust the balancing current, avoiding the traditional method of aggravating grid instability due to the simple pursuit of voltage balancing; when the internal voltage difference of the module is large, the system automatically increases the weight of the capacitor optimization target to accelerate the balancing process; when the balancing circuit continues to work, the system limits the total current through the balancing optimization target to reduce energy loss. This dynamic optimization mechanism effectively solves the technical defects of the traditional method of single control dimension and inability to adapt to complex working conditions.

[0098] More specifically, regarding step 102 mentioned in the above basic embodiment, the present application also provides another refined step, which is suitable for higher precision requirements or larger-scale usage scenarios, specifically including: performing high-frequency ripple texture processing according to the terminal voltage value to obtain the fundamental voltage component and the ripple voltage component; calculating the impedance phase offset angle of each monomer capacitor according to the fundamental voltage component to determine the fractional order parameter according to the impedance phase offset angle; extracting the amplitude attenuation rate and frequency response characteristics of the ripple voltage component to determine the equivalent series resistance parameter and the capacitance capacity parameter respectively; constructing a time-varying fractional-order capacitor model based on the fractional order parameter, equivalent series resistance parameter and capacitance capacity parameter, combined with a preset fractional-order state equation; based on historical equilibrium data, statistically calculating the energy loss rate increment of the supercapacitor module to determine the aging factor parameter based on the energy loss rate increment; obtaining an adaptive parameter set based on the numerical values ​​of the fractional order parameter, equivalent series resistance parameter, capacitance capacity parameter and aging factor parameter.

[0099] High-frequency ripple texture processing involves separating the terminal voltage signal into a low-frequency fundamental component and a high-frequency ripple component through a signal decomposition algorithm. This can be achieved using wavelet transforms or Fourier filtering methods to eliminate high-frequency noise interference and extract effective voltage features. The fundamental voltage component refers to the low-frequency voltage component in the terminal voltage signal related to the main charging and discharging process of the capacitor. It can be extracted using a low-pass filter and used to reflect the basic operating state of the capacitor. The ripple voltage component refers to the high-frequency fluctuation component in the terminal voltage signal caused by switching device operation or load fluctuations. It can be separated using a high-pass filter and used to analyze the dynamic response characteristics of the capacitor. The impedance phase offset angle refers to the phase difference between the fundamental voltage and current. It can be calculated using a phase-locked loop or fast Fourier transform and is used to characterize changes in the capacitor's impedance characteristics. The fractional order parameter refers to the order of the fractional differential equation that describes the non-ideal characteristics of the capacitor. It can be determined by fitting the impedance phase angle versus frequency curve to construct a more accurate capacitor model. The amplitude attenuation rate is the degree of attenuation of the fingerprint wave voltage in a specific frequency band. Specifically, the slope of the ripple amplitude changing with frequency can be calculated through frequency domain analysis, which is used to derive the equivalent series resistance parameters. The frequency response characteristic is the corresponding relationship between the fingerprint wave voltage amplitude and frequency. Specifically, it can be obtained through frequency sweep testing or impedance spectrum analysis, and is used to calculate the capacitance parameters. The fractional-order state equation refers to a capacitor dynamic model equation containing a fractional-order differential operator. Specifically, it can be constructed using Caputo fractional-order derivatives to describe the time-varying nonlinear characteristics of the capacitor. The energy loss rate increment refers to the change in capacitor energy loss per unit time. Specifically, it can be calculated by integrating historical balanced current and voltage data to quantify the degree of capacitor aging. The aging factor parameter refers to the proportional coefficient reflecting the degradation of capacitor performance. Specifically, the exponential decay model can be combined with the energy loss increment calculation to correct the model parameters for actual working conditions.

[0100] Specifically, the terminal voltage signal is first processed for high-frequency ripple texture, such as using a second-order Butterworth filter with a cutoff frequency of 1kHz for signal decomposition or using corresponding bandpass filters and high-pass filters for extraction, to obtain the fundamental component and ripple component. For example: Assume that in a certain energy storage type reactive power compensation device, its supercapacitor module consists of 120 BW6101 supercapacitor cells connected in series (rated voltage 2.7V / cell, total voltage 324V). To achieve the first step of data collection for voltage balancing, the specific implementation is as follows:

[0101] Multi-channel synchronous sampling circuit design. Hardware configuration: A parallel acquisition array is constructed using 16-bit precision ADC chips (AD7685), with a total of 15 ADCs (each responsible for 8 cells). FPGA-controlled synchronous triggering of sampling is used, with a sampling rate set to 100kHz. Data generation: The instantaneous terminal voltage values ​​(unit: V) of 120 cells are collected to form a raw voltage matrix (dimension: 120 × 1):

[0102] (Typical value range: 2.5V~2.8V);

[0103] Grid status signal acquisition, voltage command value ( :Get the target voltage of the grid connection point from the grid energy management system (EMS), 10.0kV (50Hz power frequency). Actual voltage value ( ): The actual voltage at the grid connection point is collected through a 0.2-level precision voltage transformer (HVT-200), which is 9.85kV. The grid status is generated: kV

[0104] High-frequency ripple separation and fundamental component extraction: use IIR bandpass filter (center frequency 50Hz, bandwidth 2Hz) process the original matrix and output the fundamental voltage matrix, the fundamental component V of monomer 1 1,基波 =2.70V.

[0105] Ripple component extraction: Use FIR high-pass filter (cut-off frequency 2kHz) to separate the switching frequency ripple (caused by the IGBT switch of the reactive power compensation device, typical frequency 10kHz). The peak-to-peak value of the ripple component of monomer 1 is V 1,纹波 =0.15V.

[0106] Data integration and security monitoring, output data structure:

[0107] Fundamental matrix = [2.70, 2.68, ..., 2.66] T (120×1 vector);

[0108] Ripple matrix = [0.15, 0.14, ..., 0.12]T (120×1 vector, unit: Vpp);

[0109] Grid state pair = (10.0, 9.85) (unit: kV);

[0110] Next, the impedance phase offset angle of each single capacitor is calculated based on the fundamental voltage component to generate a fractional-order parameter set; the amplitude attenuation rate and frequency response characteristics of the ripple voltage component are extracted to generate equivalent series resistance parameters and capacitance parameter sets; the fractional-order parameters and resistance and capacitance parameter sets are integrated to construct a time-varying capacitor model through the fractional-order state equation, and an adaptive parameter set including the aging factor is output; the aging factor is dynamically updated according to the historical energy transfer data to generate a self-correcting parameter set.

[0111] It should be noted that, taking the aforementioned energy storage type reactive power compensation device and its supercapacitor module composed of 120 BW6101 supercapacitor cells (rated voltage 2.7V, nominal capacity 3000F) connected in series as an example, in order to achieve fractional-order modeling and parameter adaptation for voltage balancing, the specific implementation is as follows:

[0112] Extraction of fractional order parameters, fundamental component analysis: Based on the fundamental voltage matrix extracted in step 201 (the fundamental voltage V 3,基波 =2.68V), decouple the impedance phase through the orthogonal mixer. The measured impedance phase deviation angle is 5.2° (normal range: 2°), indicating that the dielectric properties of monomer 3 have changed due to aging. The fractional order is calculated based on the phase shift angle. : (nominal value: 0.95); generate fractional order parameter set: =[0.949,0.953,…,0.942] T (120×1 vector).

[0113] Extraction of equivalent series resistance ESR and capacitance parameter C: ripple voltage matrix (peak-to-peak ripple of monomer 3 V 3,纹波 =0.18V) for spectrum analysis: Resonant frequency shift: Under the excitation of the ripple injection circuit, the resonant frequency of monomer 3 shifted from the nominal value of 10kHz to 10.4kHz, indicating that the equivalent series resistance (ESR) increased. Amplitude attenuation rate: Ripple amplitude attenuation rate (Nominal value: 0.85), reflecting the decrease in capacity.

[0114] Taking cell 3 as an example, the calculation formula for the equivalent series resistance ESR and the capacitance parameter C is as follows:

[0115]

[0116]

[0117] You can get the ESR parameter set [1.7, 1.9, …, 2.3] T and capacity parameter set [3002,2995,…,2980] T (Unit: m / F). Time-varying fractional order model construction and aging factor update, parameter fusion: fractional order , ESR, and capacity parameters are input into the fractional-order state equation: ; Output the adaptive parameter set including the aging factor (monomer 3: =0.942, ESR=2.3m , C=2980F);

[0118] Dynamic update of aging factor: Based on historical energy transfer data (the energy loss rate of monomer 3 increased by 15% in nearly 1,000 equalizations), the aging factor is calculated =1.15 (initial value: 1.0).

[0119] Update formula: . Generate a self-correcting parameter set: =[1.02,1.08,…,1.15] T .

[0120] After model verification, it was found that the updated model predicts that the heat loss of cell 3 under reactive compensation conditions (50Hz fundamental wave + 10kHz ripple) is 21W (nominal value: 15W), which is less than 3% from the actual infrared temperature measurement value of 20.5W.

[0121] Next, by integrating the key data obtained in the previous steps, the adaptive parameter set = [( ,ESR,C, )1=(0.949,1.7m ,3002F,1.02),...,( ,ESR,C, )3=(0.942,2.3m , 2980F, 1.15)],……; it can provide a high-precision dynamic model for subsequent differential game decision-making.

[0122] An example of key parameter update (cell 3) can be found in Table 2 below:

[0123]

[0124] Furthermore, based on the above-mentioned second adaptive parameter set generation method, the present application further proposes another technical means of inputting the adaptive parameter set and the grid status signal into a preset differential game decision maker, and determining the balanced current command value of each single capacitor through game operation, including: calculating the reactive power demand intensity coefficient according to the grid status signal; calculating the aging imbalance of each single capacitor according to the adaptive parameter set; inputting the aging imbalance and the reactive power demand intensity coefficient into the differential game decision maker for Nash equilibrium solution operation to determine the balanced current command value, wherein the optimization objectives of the multi-party differential game model include minimizing the aging imbalance, minimizing the reactive power demand intensity coefficient and minimizing the energy transfer loss.

[0125] Among them, the reactive power demand intensity coefficient refers to the percentage quantitative index of the grid voltage deviation and the command value, which can be specifically determined according to It is calculated by the formula to reflect the degree of reactive power demand of the power grid. The aging imbalance degree refers to the product of the voltage deviation of a single capacitor and the aging factor parameter. Specifically, it can be obtained by multiplying the deviation between the voltage parameter and the average voltage parameter by the aging factor parameter. It is used to characterize the impact of the capacitor aging degree on voltage balance. The differential game decider refers to a dynamic game solver that includes multiple optimization objectives. Specifically, it can be implemented using a Nash equilibrium solution algorithm based on the Hamiltonian function to coordinate conflicts between capacitor balance, grid demand, and energy consumption control. The Nash equilibrium solution operation refers to the stable state solution process for the strategy optimization of multiple game participants. Specifically, it can be implemented using the gradient descent method or the dynamic programming algorithm to find a balance point between multiple optimization objectives.

[0126] Specifically, during the operation of the supercapacitor module, the actual value of the grid voltage is collected in real time and compared with the command value, and the reactive demand intensity coefficient is calculated using a preset formula. At the same time, based on the adaptive parameter set output by the time-varying fractional-order capacitor model, combined with the deviation value between the single capacitor voltage and the average module voltage, and superimposed on the influence of the aging factor parameter, the aging imbalance of each capacitor is calculated. After these two key parameters are input into the differential game decision maker, through the dynamic game solution process, collaborative optimization is performed between the three goals of minimizing the capacitor aging imbalance, reducing the reactive demand intensity of the grid, and reducing energy transfer losses, and finally generating a balanced current command value that meets the multi-objective constraints.

[0127] For example, based on the grid status =10.0kV, =9.85kV), calculate the reactive power demand intensity coefficient, ; When the coefficient is >1%, high reactive power demand mode is triggered.

[0128] Calculation of aging imbalance: Taking monomer 3 in the adaptive parameter set as an example (monomer 3: =0.942, ESR=2.3mΩ, C=2980F, =1.15), calculate the degree to which each cell deviates from the average voltage (2.68V): V; Generate an imbalance set (120-dimensional vector, typical value 0.01V~0.05V);

[0129] Three-party differential game modeling, optimization goal definition: Power grid optimization goal: minimize reactive power demand intensity (objective function ); Capacitor optimization goal: minimize aging weighted imbalance (objective function ); Balanced optimization goal: minimize energy transfer loss (objective function );

[0130] Constraints: Single current command value I b ∈[-30A,30A] (limited by DC / DC circuit capacity)

[0131] By solving the Nash equilibrium, the game equations are combined: the three objective functions are coupled into a multi-objective optimization problem, which is solved iteratively by the gradient projection method:

[0132] Initialize current command ;

[0133] Compute the gradient: (grid sensitivity), (capacitance state sensitivity), (Energy consumption sensitivity); the iteration step size is 0.1, and 10 iterations converge to the equilibrium point.

[0134] Output example: Generate balanced current for cell 3 (high voltage difference + high aging) (Negative value indicates released energy); for monomer 107 (low pressure difference + low aging), I b,107 =+8.2A (positive value indicates absorbed energy).

[0135] Safety verification and output, overcurrent protection: If The current is limited to 25A (H-bridge topology safety threshold). Output data structure: Balanced current instruction set = [-12.5,...,+8.2] T (120×1 vector, unit: A).

[0136] This solution constructs an aging imbalance index by introducing aging factor parameters, which can dynamically reflect the impact of capacitor performance degradation on the balancing process; at the same time, combined with the grid reactive power demand intensity coefficient, the voltage balancing control is coupled with the grid operation status, and the differential game method is used to achieve multi-objective dynamic optimization, which solves the problem of insufficient balancing accuracy caused by single-objective control. Through the above technical solution, the balancing control accuracy of the aging capacitor module can be effectively improved, and the module voltage balance state can still be maintained when the capacitor performance deteriorates. At the same time, by dynamically responding to changes in the grid reactive power demand, the impact on the grid operation is reduced while ensuring the module balancing performance, and the energy loss of the balancing circuit is optimized.

[0137] Furthermore, regarding the method of converting the balancing current command value into a voltage balancing control command, the present application proposes the following steps, including: converting the balancing current command value into a corresponding PWM drive signal through a sliding mode controller, and using the PWM drive signal as the voltage balancing control command of the supercapacitor module.

[0138] A sliding mode controller is a nonlinear controller based on sliding mode variable structure theory. It can be implemented using a preset sliding surface function and convergence law. By adjusting the control signal, the system state trajectory converges to the equilibrium point along the sliding surface. A PWM drive signal is a periodic square wave signal generated through pulse width modulation technology. This is achieved by comparing a fixed-frequency carrier wave with a modulating wave. The on and off times of the switching devices in the balancing circuit are controlled by adjusting the duty cycle.

[0139] Specifically, the sliding mode controller receives the deviation between the balancing current command value and the actual current value as input and calculates a control signal based on a preset sliding mode surface function. This control signal is then linearly transformed and mapped into a PWM duty cycle parameter. A PWM drive signal is generated based on the duty cycle parameter and, through the driver circuit, controls the on and off states of the power switches in the balancing circuit, thereby regulating the charge and discharge current paths of each single capacitor. The nonlinear switching characteristics of the sliding mode controller effectively suppress parameter perturbations and external interference, ensuring rapid tracking of the balancing current command value.

[0140] For example, consider a scenario where a 10kV distribution network energy storage type reactive power compensation device consists of 60 supercapacitor modules (3000F / 2.7V) connected in series, with a total voltage of 162V. Input data includes: balancing current instructions (from step 103): discharge instruction for cell 1: -12.5A; charge instruction for cell 60: +10.2A; instructions for the remaining cells: 0.5A~3A;

[0141] Grid status signal: normalized voltage difference =0.013 (reference voltage =10kV);

[0142] Single cell voltage data: Single cell 1 voltage =2.71V, single cell 60V voltage =2.58V, average voltage =2.65V.

[0143] Sliding mode controller parameter configuration, sliding surface function: ; Coefficient k=0.5 (weighing voltage balance and grid demand, >0.01) with priority given to grid support).

[0144] Control parameters: The reaching law adopts the exponential + constant speed composite form: exponential gain =200 (accelerates approaching the sliding surface); constant velocity gain =100 (suppress steady-state jitter).

[0145] PWM linear relationship: duty cycle (Every 1A of balancing current corresponds to a 5% duty cycle, the maximum current of 50A corresponds to a 250% duty cycle, and the actual limit is 95%).

[0146] The control process is executed, taking cell 1 (high voltage needs to be discharged) as an example: Calculate the sliding surface:

[0147] ;

[0148] Generate control signal: Reaching law:

[0149] ;

[0150] Control output: equivalent current (Combined with instruction value correction);

[0151] Superimposed command value: Final current A (dynamic suppression of voltage fluctuations).

[0152] Converting to PWM: Duty Cycle , and output to the DC / DC switch tube of monomer 1 through the driving circuit.

[0153] Grid feedback function: When Sudden increase to 0.02 (grid voltage drops), make The duty cycle increases by 10% to 15%, accelerating energy transfer to support the power grid.

[0154] Output effect verification, dynamic response: within 200ms, the voltage of single cell 1 dropped from 2.71V to 2.68V, and the voltage of single cell 60 rose from 2.58V to 2.62V, and the range was reduced to 0.06V. At 2A, the sliding mode controller suppresses the voltage tracking error to Within 0.005V (no overshoot).

[0155] PWM output example: Single cell 1: Duty cycle 63% Discharge MOSFET on-time 6.3ms / cycle (10kHz switching frequency); single cell 60: duty cycle 51% The charging MOSFET conduction time is 5.1ms / cycle.

[0156] For example, consider a scenario where 120 BW6101 supercapacitor cells are connected in series in an energy storage-based reactive power compensation device (total voltage 324V). The sliding mode controller converts the balancing current command value (cell 3: -12.5A; cell 107: +8.2A) into a PWM drive signal as follows:

[0157] H-bridge topology control and switch sequence generation, hardware configuration: using a four-channel H-bridge topology (IR2104 driver chip + IRF3710MOSFET), each pair of switch tubes controls an energy transfer path.

[0158] PWM drive signal analysis: Receive the PWM signal set generated in step 204 (cell 3 drive signal duty cycle 32%, cell 107 duty cycle 68%, frequency 20kHz). Generate a complementary high-frequency switching sequence (Q1 / Q2 on, Q3 / Q4 off) using a dead-time controller (approximately 200ns) to prevent shoot-through in the bridge arm.

[0159] Ripple prediction and time window control, ripple component analysis: Based on the ripple matrix extracted in step 201 (peak-to-peak ripple of monomer 3 is 0.18V, frequency is 10kHz), the optimal time window for energy transfer is predicted: the ripple phase starts the transfer in the valley interval (ripple voltage <0.05V) to reduce switching losses. Time window length: 50 (corresponding to half a cycle of 10kHz ripple).

[0160] Directed energy transfer: Within the time window, the H-bridge is controlled to transfer the energy of cell 3 (high voltage 2.71V) to cell 107 (low voltage 2.65V) via the DC / DC circuit. Single transfer energy: mJ.

[0161] Aging factor compensation mechanism, dynamic loss compensation: According to the aging factor of step 202 (cell 3: =1.15), increase the duty cycle by 5% (corrected duty cycle 37%) to compensate for the loss caused by the increase in ESR.

[0162] Real-time monitoring and adjustment: The actual transfer current is detected by a current sensor (ACS712). If the deviation is greater than 10% (theoretical 12.5A vs. measured 11.2A), the duty cycle is dynamically increased to 40%.

[0163] By utilizing a bidirectional DC / DC conversion circuit, energy is transferred directionally between supercapacitor cells according to the PWM drive signal, realizing energy replenishment from high-voltage cells to low-voltage cells and effectively balancing the capacitor voltage. Energy transfer loss is compensated based on the aging factor, and energy loss caused by capacitor aging is reduced by adjusting the duty cycle and other methods, thereby improving energy transfer efficiency. An orthogonal mixer is used to perform amplitude-phase decoupling of the fundamental voltage component, accurately calculating the fractional order and providing key parameters for fractional-order modeling. Capacitor resonance is stimulated by a ripple injection circuit, and the resonance characteristics are measured to calculate the equivalent series resistance and capacitor capacitance parameters. This allows for real-time and accurate acquisition of capacitor parameter changes, providing a timely basis for model updating and balancing control.

[0164] The above is a detailed description of an embodiment of a supercapacitor voltage balancing method provided by the present application. The following is a detailed description of an embodiment of a supercapacitor voltage balancing device provided by the present application.

[0165] See also Figure 2 , an embodiment of the present application provides a supercapacitor module voltage balancing device, comprising:

[0166] The terminal voltage collection unit 201 is used to collect the terminal voltage value of each single capacitor in the supercapacitor module according to the supercapacitor module to be balanced;

[0167] An adaptive parameter set generating unit 202 is configured to construct a time-varying fractional-order capacitor model according to the terminal voltage value, and obtain an adaptive parameter set based on the time-varying fractional-order capacitor model and historical equalization data, wherein the adaptive parameter set includes: fractional-order parameters, equivalent series resistance parameters, capacitance parameters, and aging factor parameters;

[0168] The game decision processing unit 203 is used to input the adaptive parameter set and the grid state signal into a preset differential game decider, so as to determine the balanced current command value of each single capacitor through the game operation of the differential game decider, so as to generate the voltage balanced control command of the supercapacitor module according to the balanced current command value. The differential game decider includes a preset multi-party differential game model, and the multi-party differential game model includes: a capacitor optimization target for optimizing the balance of the supercapacitor module, a grid optimization target for optimizing the reactive power demand of the grid, and a balanced optimization target for optimizing the balanced loss.

[0169] Specifically, the terminal voltage acquisition unit obtains the terminal voltage data of each single capacitor in real time through the voltage sensor and transmits the data to the adaptive parameter set generation unit. The adaptive parameter set generation unit constructs a time-varying fractional-order capacitor model based on the terminal voltage value, and dynamically calculates the fractional order, equivalent series resistance, capacitor capacity and aging factor parameters through frequency band impedance characteristic analysis and historical data statistics. After receiving the grid status signal, the game decision processing unit inputs the adaptive parameter set together with the grid voltage deviation, reactive power demand and other signals into the differential game decision maker, generates the balanced current command value of each single capacitor by solving the Nash equilibrium point, and finally generates the voltage balance control signal through PWM modulation.

[0170] Through the above technical solution, this application solves the problem of insufficient precision in existing voltage-balancing control strategies. Through dynamic modeling and multi-objective game decision-making, it effectively matches the grid's reactive power requirements and reduces energy consumption in the balancing circuit while ensuring voltage balance within the capacitor module. This device can adapt to parameter changes caused by capacitor aging, improving the robustness of balancing control under different operating conditions.

[0171] In addition, the present application also provides a detailed description of an embodiment of a supercapacitor module voltage balancing terminal and a computer-readable storage medium.

[0172] like Figure 3 As shown, the present application provides a supercapacitor module voltage balancing terminal, including: a memory 33 and a processor 31, wherein the memory 33 and the processor 31 can be connected via a communication bus 34; the implementation carriers of the terminal include: a personal computer, an industrial computer, a server and an embedded electronic device.

[0173] The memory 33 is used to store program codes corresponding to the voltage balancing method of the supercapacitor module; the processor 31 is used to read and execute the program codes to implement the voltage balancing method of the supercapacitor module.

[0174] Memory 33 refers to a non-volatile storage medium for storing executable program code, specifically a flash memory chip or EEPROM memory. Its function is to store program instructions and historical balancing data corresponding to the voltage balancing algorithm. Processor 31 refers to an integrated circuit with computing capabilities, specifically an ARM architecture microcontroller or DSP digital signal processor. Its function is to execute the computing logic such as voltage acquisition, model building, and game decision-making in the program code in real time.

[0175] Specifically, the terminal uses a processor to periodically collect the terminal voltage values ​​of each single capacitor in the supercapacitor module. Based on the terminal voltage values, it constructs a time-varying fractional-order capacitor model and calculates an adaptive parameter set including aging factor parameters. The parameter set and grid status signals are then input into a built-in differential game decision maker for multi-objective optimization calculations, ultimately generating voltage balancing control instructions. During operation, the memory continuously stores historical balancing data for adaptive updating of model parameters, and the processor dynamically loads program code to switch balancing strategies under different operating conditions.

[0176] Furthermore, the present application proposes a computer-readable storage medium, in which program code is stored, and the program code is used to be read and executed by a processor to implement a supercapacitor module voltage balancing method for a storage-type reactive compensation device. The method includes collecting the terminal voltage values ​​of each single capacitor, constructing a time-varying fractional-order capacitor model and generating an adaptive parameter set in combination with historical balancing data, inputting the adaptive parameter set and the grid status signal into a differential game decider for game operation to determine the balancing current instruction value, and finally generating a voltage balancing control instruction.

[0177] The term "computer-readable storage medium" refers to a non-volatile storage medium capable of permanently storing program code. Specifically, it can be implemented as a flash memory chip, solid-state drive, or magnetic hard drive. Its function is to provide a reusable code storage foundation for the execution of the voltage balancing method. Program code refers to a set of computer instructions containing executable instructions. Specifically, it can be written in compiled machine code or an interpreted scripting language. Its function is to parse and drive the operation of the voltage balancing control process through a processor. A processor refers to an integrated circuit chip with computing capabilities. Specifically, it can be implemented as a central processing unit, digital signal processor, or microcontroller. Its function is to dynamically execute the acquisition, modeling, game calculation, and instruction generation steps in the voltage balancing method by reading the program code from the storage medium.

[0178] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the terminals, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0179] In the several embodiments provided in this application, it should be understood that the disclosed terminals, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0180] The terms "first," "second," "third," "fourth," and the like (if any) in the specification of the present application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present application described herein, for example, can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or apparatus.

[0181] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0182] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0183] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0184] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. As described above, the above embodiments are merely illustrative of the technical solution of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may be modified or some of the technical features thereof may be replaced by equivalents. Such modifications or replacements do not deviate from the essence of the corresponding technical solution from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A supercapacitor module voltage balancing method, characterized in that: include: According to the supercapacitor module to be balanced, collecting the terminal voltage value of each single capacitor in the supercapacitor module; A time-varying fractional-order capacitor model is constructed according to the terminal voltage value, and an adaptive parameter set is obtained based on the time-varying fractional-order capacitor model and historical equalization data, wherein the adaptive parameter set includes: a fractional-order parameter, an equivalent series resistance parameter, a capacitor capacity parameter, and an aging factor parameter; The adaptive parameter set and the grid status signal are input into a preset differential game decision maker to determine the balancing current command value of each single capacitor through the game operation of the differential game decision maker, so as to generate the voltage balancing control command of the supercapacitor module according to the balancing current command value, wherein the differential game decision maker includes a preset multi-party differential game model, and the multi-party differential game model includes: a capacitance optimization target for optimizing the balance of the supercapacitor module, a grid optimization target for optimizing the reactive power demand of the grid, and a balancing optimization target for optimizing the balancing loss.

2. A supercapacitor module voltage balancing method according to claim 1, characterized in that: The step of constructing a time-varying fractional-order capacitance model according to the terminal voltage value, and obtaining an adaptive parameter set based on the time-varying fractional-order capacitance model and historical balancing data includes: According to the terminal voltage value, a time-varying fractional-order capacitance model is constructed in accordance with a four-element fractional-order equivalent circuit model construction method; Based on the time-varying fractional-order capacitor model, combined with the sub-frequency band impedance characteristics, dynamic calculation of model component parameters is performed to determine the values ​​of fractional-order parameters, equivalent series resistance parameters, and capacitor capacity parameters; determining, based on the historical equalization data, a deviation between a current terminal voltage change rate and a historical terminal voltage change rate, and determining an aging factor parameter based on the deviation; An adaptive parameter set is obtained according to the values ​​of the fractional order parameter, the equivalent series resistance parameter, the capacitance parameter and the aging factor parameter.

3. The supercapacitor module voltage balancing method according to claim 1, characterized in that: The step of constructing a time-varying fractional-order capacitance model according to the terminal voltage value, and obtaining an adaptive parameter set based on the time-varying fractional-order capacitance model and historical balancing data includes: Performing high-frequency ripple texture processing according to the terminal voltage value to obtain a fundamental voltage component and a ripple voltage component; Calculating the impedance phase shift angle of each single capacitor according to the fundamental voltage component, so as to determine the fractional order parameter according to the impedance phase shift angle; Extracting the amplitude attenuation rate and frequency response characteristics of the ripple voltage component to respectively determine equivalent series resistance parameters and capacitance parameters; Constructing a time-varying fractional-order capacitance model based on the fractional-order order parameter, the equivalent series resistance parameter, and the capacitance parameter in combination with a preset fractional-order state equation; Based on the historical balance data, calculating the energy loss rate increment of the supercapacitor module, and determining an aging factor parameter based on the energy loss rate increment; An adaptive parameter set is obtained according to the values ​​of the fractional order parameter, the equivalent series resistance parameter, the capacitance parameter and the aging factor parameter.

4. A supercapacitor module voltage balancing method according to claim 2, characterized in that: Inputting the adaptive parameter set and the grid state signal into a preset differential game decision maker to determine the balancing current command value of each single capacitor through a game operation of the differential game decision maker includes: Calculating the per-unit voltage difference of the power grid according to the power grid state signal and a preset per-unit voltage difference calculation formula; Calculating, based on the adaptive parameter set, a cumulative sum of capacitor-voltage deviations between a voltage parameter of each single capacitor and an average voltage parameter; A preset differential game decision maker is inputted according to the accumulated sum of the capacitor voltage deviations and the per-unit voltage difference of the power grid, so as to perform a Nash equilibrium solution operation through the differential game decision maker to determine the balancing current command value of each single capacitor, wherein the optimization objectives of the multi-party differential game model include: a capacitor optimization objective for minimizing the accumulated sum of the capacitor voltage deviations, a power grid optimization objective for closed-loop control of the per-unit voltage difference of the power grid, and a balancing optimization objective for minimizing the total balancing current.

5. The supercapacitor module voltage balancing method according to claim 3, characterized in that: Inputting the adaptive parameter set and the grid state signal into a preset differential game decision maker to determine the balancing current command value of each single capacitor through a game operation of the differential game decision maker includes: Calculating a reactive power demand intensity coefficient based on the grid status signal and a preset reactive power demand intensity calculation formula; Calculating, based on the adaptive parameter set, the product of the deviation between the voltage parameter of each single capacitor and the average voltage parameter and the aging factor parameter to obtain the aging imbalance of each single capacitor; A preset differential game decision maker is input according to the aging imbalance and the reactive demand intensity coefficient, so as to perform a Nash equilibrium solution operation through the differential game decision maker to determine the balanced current command value of each single capacitor, wherein the optimization objectives of the multi-party differential game model include: a capacitor optimization objective for minimizing the aging imbalance, a power grid optimization objective for minimizing the reactive demand intensity coefficient, and a balanced optimization objective for minimizing energy transfer loss.

6. A supercapacitor module voltage balancing method according to claim 1, characterized in that: Generating the voltage balancing control instruction of the supercapacitor module according to the balancing current instruction value includes: The balancing current command value is converted into a corresponding PWM drive signal through a sliding mode controller, and the PWM drive signal is used as a voltage balancing control command for the supercapacitor module.

7. A supercapacitor module voltage balancing method according to claim 6, characterized in that: The objective function in the sliding mode controller is specifically: or ； Where, is the sliding mode value of the single capacitor i, is the terminal voltage value of single capacitor i, is the average voltage of each single capacitor, is the balance coefficient between voltage balance and grid demand, is the voltage normalized to the voltage difference, is the grid voltage command value, is the actual value of the grid voltage, is the coupling weight coefficient, is the balancing current command value of the single capacitor, is the actual value of the balancing current of the single capacitor.

8. A supercapacitor module voltage balancing device, characterized in that: include: A terminal voltage acquisition unit, configured to acquire the terminal voltage value of each single capacitor in the supercapacitor module to be balanced; an adaptive parameter set generating unit, configured to construct a time-varying fractional-order capacitor model according to the terminal voltage value, and obtain an adaptive parameter set based on the time-varying fractional-order capacitor model and historical equalization data, wherein the adaptive parameter set includes: a fractional-order parameter, an equivalent series resistance parameter, a capacitance parameter, and an aging factor parameter; A game decision processing unit is configured to input the adaptive parameter set and the grid status signal into a preset differential game decider, so as to determine the balancing current command value of each single capacitor through the game operation of the differential game decider, and to generate a voltage balancing control command for the supercapacitor module according to the balancing current command value, wherein the differential game decider includes a preset multi-party differential game model, and the multi-party differential game model includes: a capacitance optimization target for optimizing the balance of the supercapacitor module, a grid optimization target for optimizing the reactive power demand of the grid, and a balancing optimization target for optimizing the balancing loss.

9. A supercapacitor module voltage balancing terminal, characterized in that: include: memory and processor; The memory is used to store program code, and the program code corresponds to the supercapacitor module voltage balancing method according to any one of claims 1 to 7; The processor is used to read and execute the program code to implement the supercapacitor module voltage balancing method.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program code, and the program code is used to be read and executed by a processor to implement the supercapacitor module voltage balancing method according to any one of claims 1 to 7.

Images