Multi-module adaptive current sharing control method and system for household energy storage stack system

Abstract

The present application relates to the technical field of energy storage system parallel current sharing control, and discloses a kind of multi-module adaptive current sharing control method and system of household energy storage stacking system.The method includes obtaining bus voltage, bus current, bus side end voltage, branch current and capacity parameter;According to the bus side end voltage, the bus voltage and the branch current, determine the line impedance, construct impedance state characteristics and input the pre-trained support vector machine, obtain impedance state grade;According to the capacity parameter, determine the capacity shunt ratio, target branch current and current distribution deviation;According to the impedance state grade, update effective line impedance, and determine target current sharing impedance, basic virtual impedance and target virtual impedance;According to the target virtual impedance and the branch current, determine the output reference voltage, and adjust the output of the two-way DC / DC converter corresponding to each energy storage module.This method can realize adaptive distribution of multi-module branch current.

Description

Technical Field

[0001] This invention relates to the field of parallel current sharing control technology for energy storage systems, and particularly to a multi-module adaptive current sharing control method and system for residential energy storage stacking systems. Background Technology

[0002] Currently, residential energy storage stacking systems typically employ multiple energy storage modules connected in parallel to a DC bus, with different energy storage capacity configurations achieved by adding or removing energy storage modules. With the development of power conversion, operational status monitoring, and industrial big data technologies, the application of multi-module stacking structures in energy storage systems is gradually increasing. Each energy storage module shares the bus current during charging and discharging; therefore, the current distribution among the parallel branches directly affects the operational consistency of the energy storage modules and the overall lifespan of the system.

[0003] In a prior art, fixed current sharing parameters or droop control parameters are typically set based on the rated parameters of the energy storage module. The output of the corresponding converter for each energy storage module is then adjusted according to these parameters to distribute the bus current to each parallel branch. Under this control method, the control parameters remain relatively fixed during operation. However, in actual operation, differences exist in the length of connecting cables, the contact state of connecting terminals, and the temperature rise of connection points for different energy storage modules. This causes differences in the line impedance of each parallel branch, which changes with the operating state. Simultaneously, load switching causes fluctuations in the line impedance determined based on real-time voltage and current within a short period. This makes it difficult to adjust the fixed current sharing parameters according to changes in line conditions. Directly adjusting based on real-time line impedance simultaneously responds to both continuous line condition changes and short-term operational disturbances, causing the branch current of some energy storage modules to deviate from the corresponding current distribution relationship.

[0004] In summary, existing technologies have the problem that when multiple modules of a residential energy storage stacking system are running in parallel, the dynamic changes in branch impedance and the inability of fixed control parameters to adapt to the branch status in real time lead to a continuous deviation of the current distribution of each branch from the expected distribution relationship. Summary of the Invention

[0005] This invention provides a multi-module adaptive current sharing control method and system for a residential energy storage stacking system to solve the problem in the prior art where the current distribution of each branch continuously deviates from the expected distribution relationship.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a multi-module adaptive current sharing control method for a residential energy storage stacking system, comprising:

[0007] Obtain the bus voltage, bus current, bus-side voltage of each energy storage module, branch current and capacity parameters of the residential energy storage stacking system;

[0008] The line impedance is determined based on the bus side voltage, the bus voltage, and the branch current. Impedance state features are constructed based on the line impedance of multiple consecutive current sharing control cycles. The impedance state features are then input into a pre-trained support vector machine to obtain the impedance state level.

[0009] The capacity shunting ratio is determined based on the capacity parameters, the target branch current is determined based on the capacity shunting ratio and the bus current, and the current distribution deviation is determined based on the branch current and the target branch current.

[0010] The effective line impedance is updated according to the impedance status level, the target current sharing impedance is determined according to the capacity shunting ratio and the effective line impedance, the basic virtual impedance is determined according to the difference between the target current sharing impedance and the effective line impedance, and the basic virtual impedance is corrected according to the current distribution deviation to obtain the target virtual impedance.

[0011] The output reference voltage is determined based on the target virtual impedance and the branch current, and the output of the bidirectional DC / DC converter corresponding to each energy storage module is adjusted based on the output reference voltage.

[0012] Secondly, the present invention provides a multi-module adaptive current sharing control system for a residential energy storage stacking system, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] (1) This invention determines the line impedance based on the bus side voltage, bus voltage and branch current, and forms the impedance state characteristics based on the line impedance of multiple consecutive current sharing control cycles. The impedance state level is determined by the support vector machine, and the line impedance update method is selected according to the impedance state level. This allows the continuously changing line impedance to enter the basic virtual impedance update process, and maintains the most recent effective line impedance under transient disturbance state, thereby distinguishing the impact of continuous changes in line state and short-term operating disturbance on the impedance identification results.

[0015] (2) The present invention determines the target current sharing impedance based on the capacity current sharing ratio and the effective line impedance, and determines the basic virtual impedance based on the difference between the target current sharing impedance and the effective line impedance. Then, the basic virtual impedance is corrected based on the current distribution deviation between the actual branch current and the target branch current to obtain the target virtual impedance. This allows the target equivalent impedance of each energy storage module to be adjusted according to the effective line impedance, the capacity current sharing ratio and the actual current distribution state, so that the branch current of different capacity energy storage modules approaches the corresponding target branch current.

[0016] (3) The present invention determines the output reference voltage of the bidirectional DC / DC converter based on the target virtual impedance and the branch current, and adjusts the bidirectional DC / DC converter corresponding to each energy storage module according to the output reference voltage. After adjustment, the branch current is obtained again and the current sharing deviation between the branch current and the target branch current is judged, forming a closed-loop control process from branch state identification, virtual impedance correction to converter output adjustment. This enables the residential energy storage stacking system to continuously adjust the current distribution when the branch state of each energy storage module changes, and maintain the current sharing state between each energy storage module corresponding to the capacity current sharing ratio. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart of the multi-module adaptive current sharing control method for a residential energy storage stacking system provided in the first embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram of the impedance state level identification result provided in the first embodiment of the present invention;

[0019] Figure 3 This is a schematic diagram of the virtual impedance adaptive correction process provided in the first embodiment of the present invention;

[0020] Figure 4 This is a schematic diagram of the consistency of current sharing in the current sharing branch provided in the first embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Reference Figure 1 The first embodiment of the present invention provides a multi-module adaptive current sharing control method for a residential energy storage stacking system, comprising the following steps:

[0023] S1, obtain the bus voltage, bus current, bus-side voltage of each energy storage module, branch current and capacity parameters of the residential energy storage stacking system;

[0024] S2, determine the line impedance based on the bus side voltage, the bus voltage and the branch current, construct impedance state features based on the line impedance of multiple consecutive current sharing control cycles, and input the impedance state features into a pre-trained support vector machine to obtain the impedance state level.

[0025] S3, determine the capacity shunting ratio based on the capacity parameters, determine the target branch current based on the capacity shunting ratio and the bus current, and determine the current distribution deviation based on the branch current and the target branch current;

[0026] S4, update the effective line impedance according to the impedance status level, determine the target current sharing impedance according to the capacity shunting ratio and the effective line impedance, determine the basic virtual impedance according to the difference between the target current sharing impedance and the effective line impedance, and correct the basic virtual impedance according to the current distribution deviation to obtain the target virtual impedance;

[0027] S5. Determine the output reference voltage based on the target virtual impedance and the branch current, and adjust the output of the bidirectional DC / DC converter corresponding to each energy storage module based on the output reference voltage.

[0028] In step S1, the bus voltage, bus current, bus-side voltage of each energy storage module, branch current, and capacity parameters of the residential energy storage stacking system are obtained, including:

[0029] Obtain the bus voltage and bus current of the common DC bus;

[0030] Obtain the port voltage of the bidirectional DC / DC converter connected to the common DC bus for each energy storage module, determine the port voltage as the bus-side terminal voltage of the corresponding energy storage module, and collect the branch current between the bidirectional DC / DC converter and the common DC bus for each energy storage module.

[0031] Read the current available capacity output by the battery management system of each energy storage module, and determine the current available capacity as the capacity parameter.

[0032] In one implementation, the common DC bus is a DC bus loop in which the bidirectional DC / DC converters corresponding to each energy storage module are connected in parallel. The bus voltage is the real-time voltage at the junction of the branches of each energy storage module, and the bus current is the real-time total current formed after the branches of each energy storage module are combined.

[0033] The bus voltage and the bus-side voltage of each energy storage module are acquired through an isolated voltage sampling circuit. The isolated voltage sampling circuit includes a high-impedance voltage divider network, an isolation amplifier circuit, and an analog-to-digital conversion (ADC) channel. The high-impedance voltage divider network converts the measured DC voltage to the input range of the ADC channel, the isolation amplifier circuit provides electrical isolation between the power circuit and the controller sampling circuit, and the ADC channel converts the analog voltage into a digital sample value.

[0034] During the system commissioning phase, each voltage sampling channel undergoes zero-point calibration and rate conversion calibration. Zero-point calibration involves inputting 0 volts to the sampling channel and recording the corresponding analog-to-digital conversion value. Rate conversion calibration involves inputting a known voltage provided by a calibration DC source to the sampling channel, and calculating the voltage conversion factor based on the known voltage, the zero-point analog-to-digital conversion value, and the corresponding sampling value. The zero-point value and voltage conversion factor of each channel are written into the controller parameter area. During operation, the analog-to-digital conversion value is converted to the actual voltage based on the zero-point value and voltage conversion factor.

[0035] It should be noted that the bus-side voltage is taken from the port of the corresponding bidirectional DC / DC converter closest to the common DC bus. The bus-side voltage and the bus voltage correspond to the voltages at the bus-side port of the bidirectional DC / DC converter and the common DC bus junction, respectively, and the connection branch between the two voltage sampling locations is the connection branch of the corresponding energy storage module.

[0036] The bus current and the branch current of each energy storage module are acquired using closed-loop Hall current sensors. The bus current sensor is installed on the total current path after the branches of each energy storage module converge, and the branch current sensor is installed on the connection branch between the corresponding bidirectional DC / DC converter bus-side port and the common DC bus. The closed-loop Hall current sensor outputs an analog measurement signal corresponding to the actual current, and the controller obtains the corresponding current sampling value through an analog-to-digital conversion channel.

[0037] In this embodiment, each voltage sampling channel and each current sampling channel are synchronously triggered by the same controller time base. The bidirectional DC / DC converter generates PWM control signals using a center-aligned method, and the analog-to-digital conversion is triggered at the center of the carrier wave, away from the switching edge, in each PWM carrier cycle. The synchronous voltage and synchronous current sampling values ​​obtained within a current sharing control cycle are arithmetically averaged, and the average result is used as the bus voltage, bus current, bus-side voltage, and branch current for the current current sharing control cycle.

[0038] In this embodiment, the current sharing control period is set to 20 milliseconds, corresponding to a current sharing parameter update frequency of 50 Hz. The outer virtual impedance adjustment time constant (described later) is 400 milliseconds; therefore, 20 current sharing parameter updates are performed within one outer virtual impedance adjustment time constant. The bidirectional DC / DC converter's own PWM control, current loop control, and voltage loop control operate with a control cycle of less than 20 milliseconds, while the current sharing control, as an outer control process, updates with a 20-millisecond cycle.

[0039] Within each 20-millisecond current sharing control cycle, the bus voltage, bus current, bus-side voltage of each energy storage module, and branch current are associated with the same control cycle time marker. When calculating the branch voltage drop and line impedance of the same energy storage module in step S2, the bus-side voltage, bus voltage, and branch current corresponding to the same time marker are retrieved.

[0040] The capacity parameter represents the current available capacity of the corresponding energy storage module in its current operating state. The controller receives status data frames sent by the battery management system of each energy storage module via the CAN communication interface. It extracts the raw capacity value according to the current available capacity field position specified in the corresponding battery management system's communication protocol, converts it to an actual capacity value according to the data ratio recorded in the communication protocol, and determines this actual capacity value as the capacity parameter. The capacity parameter of each energy storage module uniformly uses ampere-hours (AH) as the unit of capacity.

[0041] When the battery management system's capacity data update cycle is longer than 20 milliseconds, the most recent valid capacity parameter with a timestamp no later than the current control time is used in the current current sharing control cycle; after receiving the new current available capacity, the updated capacity parameter is used starting from the next current sharing control cycle.

[0042] The energy storage modules participating in the current current sharing control must have their corresponding bidirectional DC / DC converters in operation and their capacity parameters greater than 0. After an energy storage module exits operation, it is removed from the set of energy storage modules participating in the current sharing calculation in the next current sharing control cycle. Once a new energy storage module enters operation and obtains valid capacity parameters, it is added to the set of energy storage modules in the next current sharing control cycle. When the set of energy storage modules participating in the current sharing calculation is empty, subsequent capacity ratio and virtual impedance updates are paused, and step S1 is re-executed once an energy storage module that meets the participation conditions appears.

[0043] For example, the residential energy storage stack system includes four energy storage modules in a discharged state. During a 20-millisecond current sharing control cycle, the bus voltage of the common DC bus is measured to be 399.868 volts, and the bus current is 40.0 amps. The bus-side voltages of the four energy storage modules are 399.985 volts, 399.993 volts, 400.004 volts, and 399.998 volts, respectively, corresponding to branch currents of 11.8 amps, 10.4 amps, 9.7 amps, and 8.1 amps. The current available capacities output by the battery management system of the four energy storage modules are read as 100 amp-hours, 90 amp-hours, 80 amp-hours, and 70 amp-hours, respectively, and these values ​​are determined as the capacity parameters of the corresponding energy storage modules.

[0044] In step S2, the line impedance is determined based on the bus-side voltage, the bus voltage, and the branch current. Impedance state features are constructed based on the line impedance over multiple consecutive current-sharing control cycles. These impedance state features are then input into a pre-trained support vector machine to obtain the impedance state level, including:

[0045] The absolute value of the voltage difference between the bus side voltage and the bus voltage of each energy storage module is calculated according to the same sampling time, and the absolute value of the corresponding branch current is extracted. The absolute value of the voltage difference and the absolute value of the corresponding branch current are combined to obtain impedance sampling data.

[0046] The impedance sampling data is fitted using least squares, and the slope of the fitted line is determined as the line impedance of the corresponding energy storage module.

[0047] Arrange multiple consecutive line impedances according to the current sharing control cycle to obtain a line impedance sequence;

[0048] The line impedance change amplitude and slope are determined based on the line impedance sequence. The impedance fitting deviation is determined based on the deviation between the impedance sampling data and the corresponding fitted line. The line impedance change amplitude, the line impedance change slope, and the impedance fitting deviation are combined to obtain the impedance state characteristics.

[0049] The impedance state characteristics are input into the support vector machine to obtain the impedance state level.

[0050] In one implementation, the line impedance corresponds to the equivalent DC resistance formed by the connecting conductors, connecting terminals, and contact points between the bus-side port of the bidirectional DC / DC converter and the common DC bus junction location.

[0051] For each energy storage module, according to the control cycle time marker established in step S1, the bus-side voltage, bus voltage, and branch current corresponding to the same time marker are extracted. The absolute value of the difference between the bus-side voltage and the bus voltage is calculated, and the obtained value is determined as the branch voltage drop; at the same time, the absolute value of the branch current is calculated, and the absolute value of one branch voltage drop and one branch current corresponding to the same control cycle are combined into a set of impedance sampling data.

[0052] Both charging and discharging states generate impedance sampling data according to the correspondence between the absolute value of branch voltage drop and the absolute value of branch current.

[0053] In one implementation, impedance sampling data is extracted from the stable current operating point formed during the operation of the bidirectional DC / DC converter. When the difference in the absolute values ​​of the branch currents corresponding to two consecutive 20-millisecond current-sharing control cycles is not greater than the maximum absolute measurement error of the corresponding branch current sensor, the data corresponding to the subsequent current-sharing control cycle is recorded as the stable operating point. The 40-millisecond observation time corresponding to two consecutive current-sharing control cycles corresponds to the inner current loop stabilization time obtained after debugging the bidirectional DC / DC converter in this embodiment.

[0054] The maximum absolute measurement error of the branch current sensor reads the technical parameters of the corresponding closed-loop Hall current sensor within the current range and operating temperature range. In this embodiment, the maximum absolute measurement error of the branch current sensor is 0.2 amps. Therefore, when the difference in the absolute value of the branch current for two consecutive control cycles is not greater than 0.2 amps, the subsequent control cycle is recorded as the stable operating point.

[0055] A single least-squares fit uses at least three sets of stable operating point data. Least-squares fitting simultaneously determines two parameters of the fitted line: the slope and the intercept. Therefore, the three sets of data create a data constraint exceeding the number of parameters to be determined. The difference between the maximum and minimum absolute values ​​of all branch currents participating in the single fit is greater than twice the maximum absolute measurement error of the corresponding branch current sensor. This twice-maximum absolute measurement error corresponds to the maximum difference that may occur when two current sampling values ​​are at their maximum positive and negative measurement deviations, respectively. For example, when the maximum absolute measurement error of the corresponding branch current sensor is 0.2 amps, the span of the branch currents participating in the fit is greater than 0.4 amps.

[0056] If the number of stable operating points is less than 3 or the branch current span does not meet the above conditions, continue to acquire subsequent stable operating points; if line impedance identification results have been formed, retain the most recent line impedance identification results until new valid fitting results are formed. When the system is running for the first time and line impedance identification results have not yet been formed, form no less than 3 sets of impedance sampling data according to the first line impedance fitting method described later.

[0057] In one implementation, when the system is running for the first time and the currently acquired impedance sampling data has not yet met the initial line impedance fitting condition, at least three current reference values ​​are sequentially set through the current inner loop of the corresponding bidirectional DC / DC converter. Each current reference value is within the rated branch current range of the corresponding bidirectional DC / DC converter, and the difference between two adjacent current reference values ​​is greater than twice the maximum absolute measurement error of the corresponding branch current sensor. After each current reference value is set, the branch current is judged to have entered a steady state according to the aforementioned stable operating point determination method. After the branch current enters a steady state, the bus-side voltage, bus voltage, and branch current corresponding to the same sampling time are collected, and a set of impedance sampling data is obtained according to the aforementioned impedance sampling data formation method.

[0058] After obtaining at least three sets of impedance sampling data that meet the aforementioned current span condition, the line impedance of the corresponding energy storage module is calculated using the aforementioned least squares fitting method. The calculated line impedance is then determined as the initial line impedance identification result. After completing the initial line impedance fitting, the corresponding bidirectional DC / DC converter is restored to the current reference value corresponding to the current operating state. During subsequent operation, the impedance sampling data is updated based on the stable operating point formed during normal charging and discharging, and the line impedance is recalculated.

[0059] When performing least-squares fitting on the impedance sampling data, the absolute value of the branch current is used as the independent variable, and the corresponding branch voltage drop is used as the dependent variable. The line impedance of the i-th energy storage module is determined according to the following formula:

[0060] ;

[0061] in, Let represent the line impedance of the branch corresponding to the i-th energy storage module, and n represent the number of stable operating points participating in this least squares fitting. This represents the absolute value of the branch current corresponding to the i-th energy storage module at the k-th stable operating point. i represents the arithmetic mean of the absolute values ​​of all branch currents participating in this fitting. This represents the branch voltage drop corresponding to the i-th energy storage module at the k-th stable operating point. This represents the arithmetic mean of the voltage drops of all branches participating in this fitting.

[0062] The slope calculated above has the dimension of volts divided by amperes, corresponding to ohms. This slope is then used to determine the line impedance of the corresponding energy storage module. The intercept of the fitted straight line is recorded as a fixed voltage bias between the bus-side voltage sampling channel and the bus voltage sampling channel. The line impedance is represented by the slope corresponding to the branch voltage drop as a function of the branch current.

[0063] This embodiment updates the impedance sampling data after obtaining a new stable operating point and refits the data according to the most recently obtained stable operating point, so that the line impedance is updated as the state of the connecting conductors and terminals changes. For example, the fit uses the five most recent stable operating points. The fitted line contains two parameters to be determined, and in addition to the two parameters, three sets of additional observation data are retained from the five stable operating points. Simultaneously, data corresponding to historical operating states are restricted from continuously being included in the current line impedance calculation.

[0064] For example, the absolute values ​​of the branch currents corresponding to the five most recently obtained stable operating points of the first energy storage module are 8 A, 12 A, 16 A, 20 A, and 24 A, respectively, and the corresponding branch voltage drops are 0.083 V, 0.121 V, 0.163 V, 0.201 V, and 0.242 V, respectively. The arithmetic mean of the five branch current absolute values ​​is 16 A, and the arithmetic mean of the five branch voltage drops is 0.162 V. The slope of the fitted straight line calculated using the above least squares relationship is approximately 0.00995 ohms. This 0.00995 ohms is determined as the line impedance of the first energy storage module, i.e., 9.95 milliohms.

[0065] The same method was used to fit three other energy storage modules, and the corresponding line impedances were found to be 12.00 milliohms, 14.00 milliohms and 16.00 milliohms, respectively.

[0066] Furthermore, after obtaining the line impedance of each energy storage module, the line impedance is continuously recorded according to the current sharing control cycle. In this embodiment, the line impedance of the most recent 20 current sharing control cycles is selected to form a line impedance sequence. Each current sharing control cycle is 20 milliseconds, so 20 current sharing control cycles correspond to 400 milliseconds, which corresponds to the outer virtual impedance adjustment time constant of this embodiment, so that the impedance state characteristics reflect the line impedance changes within an outer virtual impedance adjustment time scale.

[0067] In one implementation, the impedance state characteristics include the line impedance change amplitude, the line impedance change slope, and the impedance fitting deviation. The absolute value of the difference between the line impedance in the current current sharing control cycle and the line impedance in the previous current sharing control cycle is calculated, and this value is determined as the line impedance change amplitude. A linear fit is performed using the cycle number of the 20 current sharing control cycles as the independent variable and the corresponding line impedance as the dependent variable, and the slope of the fitted line is determined as the line impedance change slope. For each impedance sampling data used in the current line impedance fitting, the difference between the actual branch voltage drop and the corresponding branch voltage drop of the fitted line is calculated. The square root of the average of the squares of each difference is determined as the impedance fitting deviation. The impedance state characteristics are obtained by combining these three values ​​in the order of line impedance change amplitude, line impedance change slope, and impedance fitting deviation.

[0068] The support vector machine employs a radial basis function kernel and a soft-margin classification method. During the system debugging phase, training samples are established by connecting adjustable DC resistors to the connection branches of the corresponding energy storage modules, forming stable, continuously drifting, and transient disturbance states. Keeping the DC resistor value constant and the load stable, the corresponding samples are marked as stable states. The DC resistor value is gradually changed along the same direction, marking the corresponding samples as continuously drifting states. Keeping the DC resistor value constant and switching the load current, the data generated over 20 consecutive current sharing control cycles after the load switch are marked as transient disturbance states.

[0069] The operating range from 0 to the rated branch current is divided into equal intervals of 25% of the rated branch current. Three internal points of 25%, 50%, and 75% of the rated branch current are selected to form low, medium, and high current-carrying training conditions, avoiding the 0 current position and the rated branch current boundary. Data for no less than 100 current-sharing control cycles are continuously collected under each impedance state and each current-carrying condition. 20 consecutive current-sharing control cycles are used as one sample window, and the sample windows do not overlap. In this way, no less than 5 training samples are formed for each impedance state under each current-carrying condition, so that each current-carrying condition can enter the training and verification process during 5-fold cross-validation.

[0070] For each training sample, the line impedance change amplitude, line impedance change slope, and impedance fitting deviation are calculated using the same method as in the running phase. The arithmetic mean and standard deviation of the three types of features in the training set are calculated separately. Each feature is subtracted from its corresponding arithmetic mean and then divided by its corresponding standard deviation to obtain the standardized training features. The corresponding arithmetic mean and standard deviation are then saved.

[0071] This embodiment employs a pairwise alternative to establish a three-radial basis function kernel support vector machine. The training process aims to increase the classification margin between samples with different impedance states and to penalize misclassified samples, using a sequential minimum optimization method. Candidate values ​​for the penalty coefficient are selected from 2^−5 to 2^15 exponents of 2, and candidate values ​​for the radial basis function kernel parameters are selected from 2^−15 to 2^3 exponents of 2, with an interval of 2 between adjacent candidate exponents to cover different regularization strengths and kernel widths on an order of magnitude.

[0072] For each set of candidate parameters, 5-fold hierarchical cross-validation was performed. Each impedance state sample was divided into 5 subsets in equal proportions. In each round, 4 subsets were used for training, and the remaining subset was used for validation. This process was repeated for 5 rounds, and the average classification error rate was calculated. The parameter combination with the lowest average classification error rate was selected as the penalty coefficient and radial basis function kernel parameters for the corresponding support vector machine. When multiple parameter combinations with the same average classification error rate existed, the parameter combination with the smaller penalty coefficient was selected. After training, the support vectors, dual coefficients, bias terms, radial basis function kernel parameters, and the aforementioned arithmetic mean and standard deviation of the features from the three support vector machines were written into the controller parameter area.

[0073] During operation, the impedance state features are calculated and standardized based on the data from the most recent 20 current sharing control cycles using the same method as in the training phase. The standardized impedance state features are then input into three trained support vector machines to obtain corresponding decision values. The category with the largest decision value is determined as the current impedance state level.

[0074] For example, Figure 2 This is a schematic diagram of the impedance state level identification results. The diagram shows the slope of line impedance change and the distribution of impedance fitting deviation for independent verification samples. The stable state is concentrated in the region where the slope of line impedance change is close to zero and the impedance fitting deviation is small. The continuous drift state is characterized by obvious positive or negative line impedance change slopes, while the transient disturbance state has a large impedance fitting deviation. The support vector machine can effectively distinguish the three impedance state levels.

[0075] In step S3, the capacity shunting ratio is determined according to the capacity parameters, the target branch current is determined according to the capacity shunting ratio and the bus current, and the current distribution deviation is determined according to the branch current and the target branch current.

[0076] Determining the capacity offloading ratio based on the capacity parameters includes:

[0077] The total capacity of the modules is obtained by summing the capacity parameters of each energy storage module.

[0078] Calculate the ratio between the capacity parameter of each energy storage module and the total capacity of the module to obtain the capacity diversion ratio of each energy storage module.

[0079] In one implementation, the capacity parameters of each energy storage module participating in the current current sharing control in step S1 are extracted, and the capacity parameters are summed to obtain the total capacity of the modules. The capacity parameter of each energy storage module is divided by the total capacity of the modules, and the resulting ratio is determined as the capacity sharing ratio of the corresponding energy storage module.

[0080] The capacity diversion ratio is a dimensionless value, and the sum of the capacity diversion ratios corresponding to all energy storage modules participating in the current current sharing control is 1.

[0081] When the capacity parameters of any energy storage module are updated, the total module capacity and the total capacity distribution ratio are recalculated from the next current sharing control cycle. When an energy storage module is taken out of operation, the total module capacity and the total capacity distribution ratio are recalculated based on the capacity parameters of the remaining energy storage modules. When a new energy storage module is put into operation, the capacity parameters of the energy storage module are added to the total module capacity, and the total capacity distribution ratio is recalculated.

[0082] When the total capacity of the energy storage modules participating in the current current sharing control is 0, the current capacity ratio calculation ends, and the set of energy storage modules participating in the current sharing control is re-evaluated according to step S1.

[0083] For example, the capacity parameters of the four energy storage modules are 100 Ah, 90 Ah, 80 Ah, and 70 Ah, respectively. Summing these four capacity parameters yields a total module capacity of 340 Ah. Dividing 100 Ah, 90 Ah, 80 Ah, and 70 Ah by 340 Ah yields corresponding capacity shunting ratios of approximately 0.2941, 0.2647, 0.2353, and 0.2059. These four values ​​are displayed with four decimal places; subsequent calculations use the unrounded capacity shunting ratios stored internally by the controller.

[0084] The determination of the target branch current based on the capacity shunting ratio and the bus current, and the determination of the current distribution deviation based on the branch current and the target branch current, include:

[0085] Calculate the absolute value of the bus current to obtain the bus current amplitude, and calculate the product of the capacity shunting ratio of each energy storage module and the bus current amplitude to obtain the corresponding target branch current;

[0086] Calculate the absolute value of the branch current of each energy storage module to obtain the branch current amplitude. Calculate the difference between the branch current amplitude and the corresponding target branch current to obtain the current distribution deviation. A positive current distribution deviation indicates that the branch current amplitude of the corresponding energy storage module is higher than the target branch current.

[0087] In one implementation, the bus current corresponding to the current sharing control cycle in step S1 is extracted, and the absolute value of the bus current is taken to obtain the bus current amplitude. The capacity shunting ratio corresponding to each energy storage module is multiplied by the bus current amplitude, and the resulting product is determined as the target branch current of the corresponding energy storage module. The sum of all target branch currents equals the current bus current amplitude.

[0088] Furthermore, the absolute value of the branch current of each energy storage module within the current current sharing control cycle is taken to obtain the corresponding branch current amplitude. The target branch current is subtracted from each branch current amplitude, and the resulting difference is determined as the current distribution deviation.

[0089] When the current distribution deviation is greater than 0, the amplitude of the branch current of the corresponding energy storage module is higher than the target branch current; when the current distribution deviation is less than 0, the amplitude of the branch current of the corresponding energy storage module is lower than the target branch current; when the current distribution deviation is equal to 0, the amplitude of the branch current of the corresponding energy storage module is the same as the target branch current.

[0090] The capacity ratio shunt for both charging and discharging states is calculated using current amplitude, and the current direction is represented by the signed branch current in step S5.

[0091] For example, the current bus current is 40.0 amps, and the absolute value of the bus current is taken to obtain the bus current amplitude of 40.0 amps. Calculated according to the unrounded capacity shunt ratio, the target branch currents of the four energy storage modules are approximately 11.76 amps, 10.59 amps, 9.41 amps, and 8.24 amps, respectively.

[0092] The currents of the four branches within the same control cycle are 11.8 A, 10.4 A, 9.7 A, and 8.1 A, respectively. The differences between the branch current amplitude and the corresponding target branch current are calculated, and the current distribution deviations of the four energy storage modules are approximately 0.04 A, -0.19 A, 0.29 A, and -0.14 A, respectively.

[0093] In step S4, the effective line impedance is updated according to the impedance status level, the target current sharing impedance is determined according to the capacity shunting ratio and the effective line impedance, the basic virtual impedance is determined according to the difference between the target current sharing impedance and the effective line impedance, and the basic virtual impedance is corrected according to the current distribution deviation to obtain the target virtual impedance.

[0094] Specifically, updating the effective line impedance based on the impedance status level, determining the target current-sharing impedance based on the capacity shunt ratio and the effective line impedance, and determining the basic virtual impedance based on the difference between the target current-sharing impedance and the effective line impedance include:

[0095] Update the corresponding effective line impedance according to the impedance status level of each energy storage module;

[0096] Specifically, when the impedance state level is in a continuous drift state, the current line impedance is determined as the effective line impedance; when the impedance state level is in a stable state, the arithmetic mean of the current line impedance and the most recent effective line impedance is updated as the effective line impedance; and when the impedance state level is in a transient disturbance state, the most recent effective line impedance is maintained.

[0097] The impedance weight value is obtained by multiplying the effective line impedance of each energy storage module by the corresponding capacity shunt ratio.

[0098] The maximum value among the impedance weights is selected as the reference value for current sharing impedance.

[0099] Calculate the ratio between the current sharing impedance benchmark value and the capacity current sharing ratio corresponding to each energy storage module to obtain the target current sharing impedance of each energy storage module.

[0100] The difference between the target current-sharing impedance of each energy storage module and the corresponding effective line impedance is calculated to obtain the corresponding basic virtual impedance.

[0101] In one implementation, the effective line impedance of the corresponding energy storage module is updated according to the impedance state level determined in step S2. The effective line impedance represents the line impedance involved in the calculation of the target current sharing impedance and the basic virtual impedance during the current current sharing control process. When no historical effective line impedance exists, the most recent line impedance identification result is determined as the current effective line impedance. When the impedance state level is in a stable state, the arithmetic mean of the current line impedance and the most recent effective line impedance is calculated, and the resulting average value is updated as the current effective line impedance. When the impedance state level is in a continuously drifting state, the current line impedance is updated as the effective line impedance. When the impedance state level is in a transient disturbance state, the most recent effective line impedance is maintained, and the impedance correction amount is continued to be calculated based on the current current distribution deviation.

[0102] The effective line impedance of each energy storage module is multiplied by its corresponding capacity shunt ratio, and the resulting product is determined as the impedance weight value. The impedance weight values ​​of all energy storage modules participating in the current current sharing control are compared, and the maximum value is determined as the current sharing impedance reference value.

[0103] For each energy storage module, the current sharing impedance benchmark value is divided by the corresponding capacity shunting ratio, and the result is determined as the target current sharing impedance of the energy storage module. Therefore, the product of the target current sharing impedance and the corresponding capacity shunting ratio for each energy storage module is equal to the same current sharing impedance benchmark value. Energy storage modules with larger capacity shunting ratios correspond to smaller target current sharing impedances, and energy storage modules with smaller capacity shunting ratios correspond to larger target current sharing impedances.

[0104] The target current-sharing impedance of each energy storage module is subtracted from the corresponding effective line impedance, and the difference is determined as the basic virtual impedance. The target current-sharing impedance represents the target equivalent impedance formed by the effective line impedance and the basic virtual impedance.

[0105] Since the current-sharing impedance benchmark value is the maximum value among all impedance weights, for any energy storage module participating in the current current-sharing control, the current-sharing impedance benchmark value is not less than the product of the effective line impedance and the capacity current sharing ratio of the energy storage module. Since each capacity current sharing ratio is greater than 0, the target current-sharing impedance obtained by dividing the current-sharing impedance benchmark value by the capacity current sharing ratio is not less than the corresponding effective line impedance. Therefore, the basic virtual impedance obtained by subtracting the effective line impedance from the target current-sharing impedance is not less than 0.

[0106] Each energy storage module uses the same rated bus voltage as the reference for subsequent virtual impedance adjustment. Under the condition that the common DC bus voltage is the same, the target equivalent impedance corresponding to each energy storage module is configured according to the reciprocal of the capacity shunt ratio, so that different energy storage modules form a branch current distribution relationship corresponding to the capacity shunt ratio.

[0107] For example, the capacity shunting ratios of the four energy storage modules are approximately 0.2941, 0.2647, 0.2353 and 0.2059, respectively, corresponding to effective line impedances of 9.95 milliohms, 12.00 milliohms, 14.00 milliohms and 16.00 milliohms, respectively.

[0108] By multiplying the effective line impedance by the corresponding capacity shunting ratio, four impedance weight values ​​are obtained: approximately 2.926 milliohms, 3.176 milliohms, 3.294 milliohms, and 3.294 milliohms. Therefore, 3.294 milliohms is determined as the reference value for current sharing impedance.

[0109] Based on the unrounded capacity shunting ratio, dividing 3.294 milliohms by the corresponding capacity shunting ratio yields target current-sharing impedances of approximately 11.20 milliohms, 12.44 milliohms, 14.00 milliohms, and 16.00 milliohms for the four energy storage modules. Subtracting the corresponding effective line impedance from each target current-sharing impedance yields basic virtual impedances of approximately 1.25 milliohms, 0.44 milliohms, 0 milliohms, and 0 milliohms.

[0110] The process of correcting the basic virtual impedance based on the current distribution deviation to obtain the target virtual impedance includes:

[0111] The proportional-integral calculation is performed on the current distribution deviation of each energy storage module to obtain the corresponding impedance correction amount;

[0112] The corrected virtual impedance is obtained by summing the basic virtual impedance and the corresponding impedance correction amount.

[0113] The modified virtual impedance is limited according to the allowable virtual impedance range of the corresponding bidirectional DC / DC converter to obtain the target virtual impedance.

[0114] In one implementation, the impedance correction amount for the i-th energy storage module in the k-th current sharing control cycle is determined according to the following formula:

[0115] ;

[0116] in, This represents the impedance correction amount of the i-th energy storage module in the k-th current sharing control cycle. This represents the current distribution deviation of the i-th energy storage module in the k-th current sharing control cycle. This represents the corresponding proportionality coefficient. This represents the corresponding integral coefficient, and q represents the sequence number of the flow-sharing control cycle during integral accumulation. This represents the actual time interval between two adjacent current sharing control cycles. In this embodiment, a fixed current sharing control cycle of 20 milliseconds is used. Take 0.02 seconds.

[0117] When the current distribution deviation is positive, the proportional-integral operation generates a positive impedance correction, which increases the virtual impedance relative to the base virtual impedance; when the current distribution deviation is negative, the proportional-integral operation generates a negative impedance correction, which decreases the virtual impedance relative to the base virtual impedance.

[0118] The permissible virtual impedance range is determined based on the permissible adjustment range of the output reference voltage, the rated bus voltage, and the rated branch current of the corresponding bidirectional DC / DC converter. The permissible adjustment range of the output reference voltage represents the voltage range within which the output reference voltage can be adjusted relative to the rated bus voltage when the bidirectional DC / DC converter performs current sharing control. During system configuration, the permissible adjustment range of the output reference voltage is determined based on the lower and upper limits of the output reference voltage recorded in the corresponding bidirectional DC / DC converter controller, and the lower limit, upper limit, and rated branch current are written into the controller parameter area.

[0119] Calculate the voltage difference between the rated bus voltage and the lower limit of the allowable adjustment range of the output reference voltage, and the voltage difference between the upper limit of the allowable adjustment range of the output reference voltage and the rated bus voltage. Select the smaller of the two voltage differences as the allowable voltage offset. Divide the allowable voltage offset by the rated branch current of the corresponding bidirectional DC / DC converter, and determine the result as the upper limit of the allowable virtual impedance range. The lower limit of the allowable virtual impedance range is set to 0 ohms.

[0120] For example, the rated bus voltage of a residential energy storage stack system is 400 volts. The allowable adjustment range of the output reference voltage when the bidirectional DC / DC converter performs current sharing control is 399–401 volts. The voltage difference between the rated bus voltage and both the lower and upper limits of the allowable adjustment range of the output reference voltage is 1 volt. Therefore, 1 volt is determined as the allowable voltage offset. The rated branch current of the bidirectional DC / DC converter is 25 amps. Dividing 1 volt by 25 amps yields an upper limit of the allowable virtual impedance range of 0.04 ohms, or 40 milliohms. Therefore, the allowable virtual impedance range is 0–40 milliohms.

[0121] The proportional and integral coefficients are determined based on the upper limit of the allowable virtual impedance range, the rated branch current, and the settling time of the inner current loop of the corresponding bidirectional DC / DC converter.

[0122] The stabilization time of the inner current loop was measured during the commissioning phase of the bidirectional DC / DC converter. Maintaining the common DC bus voltage and the energy storage module terminal voltage stable, the current reference value was changed within the rated branch current range. When the absolute difference between the actual branch current and the new current reference value was not greater than the maximum absolute measurement error of the corresponding branch current sensor, and this was maintained for one 20-millisecond current sharing control cycle, the time interval between the moment when the actual branch current first entered the error range and the moment when the current reference value was changed was determined as the stabilization time of that current loop.

[0123] For example, the internal current loop stabilization time measured in the discharge direction is 36 milliseconds, and the internal current loop stabilization time measured in the charging direction is 40 milliseconds. Therefore, 40 milliseconds is written into the controller parameter area as the internal current loop stabilization time in this embodiment.

[0124] Divide the upper limit of the allowable virtual impedance range by the rated branch current, and determine the result as a proportionality coefficient. This proportionality coefficient ensures that when the current distribution deviation reaches the level of the rated branch current, the impedance correction corresponding to the proportional term is within the allowable virtual impedance range.

[0125] The outer virtual impedance adjustment time constant is determined to be 10 times the settling time of the inner current loop. This 10-fold time relationship makes the timescale of the outer virtual impedance adjustment one order of magnitude slower than the settling process of the inner current loop. In this embodiment, the settling time of the inner current loop is 40 milliseconds, therefore the outer virtual impedance adjustment time constant is taken as 400 milliseconds.

[0126] Divide the proportional coefficient by the outer virtual impedance adjustment time constant, and determine the result as the integral coefficient.

[0127] For example, the upper limit of the allowed virtual impedance range is 0.04 ohms, and the rated branch current is 25 amps. Dividing 0.04 ohms by 25 amps yields a proportionality coefficient of 0.0016 ohms per amp. The outer virtual impedance adjustment time constant is 0.4 seconds. Dividing 0.0016 ohms per amp by 0.4 seconds yields an integral coefficient of 0.004 ohms per amp-second.

[0128] When an energy storage module first joins the current current sharing control, the corresponding integral accumulation term is initialized to 0; when the energy storage module leaves the current current sharing control, the corresponding integral accumulation term is cleared, and integral accumulation is performed again when it joins the current sharing control again.

[0129] The corrected virtual impedance is obtained by summing the basic virtual impedance and the corresponding impedance correction. When the corrected virtual impedance is within the allowable virtual impedance range, it is determined as the target virtual impedance; when the corrected virtual impedance is less than the lower limit of the allowable virtual impedance range, the lower limit is determined as the target virtual impedance; when the corrected virtual impedance is greater than the upper limit of the allowable virtual impedance range, the upper limit is determined as the target virtual impedance.

[0130] When the target virtual impedance reaches the upper limit of the allowable virtual impedance range and the current current distribution deviation continues to generate positive impedance correction, the increase of the corresponding integral term in the current control cycle is paused; when the target virtual impedance reaches the lower limit of the allowable virtual impedance range and the current current distribution deviation continues to generate negative impedance correction, the decrease of the corresponding integral term in the current control cycle is paused. Integral accumulation resumes after the current distribution deviation changes direction and the corrected virtual impedance re-enters the allowable virtual impedance range.

[0131] For example, the base virtual impedance of the first energy storage module is 1.25 milliohms, the current current distribution deviation is 0.04 amps, and the initial value of the integral accumulation is 0. The impedance correction amount corresponding to the proportional term is 0.064 milliohms. Adding 0.064 milliohms to 1.25 milliohms yields a corrected virtual impedance of 1.314 milliohms. 1.314 milliohms falls within the allowable virtual impedance range of 0 to 40 milliohms; therefore, 1.314 milliohms is determined as the target virtual impedance for the current current sharing control cycle.

[0132] For example, Figure 3 This is a schematic diagram of the virtual impedance adaptive correction process. The diagram shows the changes of the basic virtual impedance and the target virtual impedance of the energy storage module 1 over time. The target virtual impedance is corrected by proportional-integral correction based on the current distribution deviation on the basis of the basic virtual impedance, and is continuously adjusted with the change of effective line impedance. The difference between the two curves is naturally formed by the calculation process, and the target virtual impedance is always within the allowable virtual impedance range.

[0133] In step S5, the output reference voltage is determined based on the target virtual impedance and the branch current, and the output of the bidirectional DC / DC converter corresponding to each energy storage module is adjusted based on the output reference voltage.

[0134] The process of determining the output reference voltage based on the target virtual impedance and the branch current includes:

[0135] The sign of the branch current of each energy storage module is determined by taking the discharge direction as the positive direction and the charging direction as the negative direction, thus obtaining the signed branch current.

[0136] The virtual impedance voltage drop is obtained by multiplying the target virtual impedance by the corresponding signed branch current.

[0137] The difference between the rated bus voltage of the residential energy storage stack system and the virtual impedance voltage drop is calculated to obtain the corresponding output reference voltage of the bidirectional DC / DC converter.

[0138] In one implementation, the current flowing from the bus-side port of the bidirectional DC / DC converter to the common DC bus is defined as positive, and the current flowing from the common DC bus to the bus-side port of the bidirectional DC / DC converter is defined as negative. When the energy storage module is in a discharging state, the corresponding signed branch current is positive; when the energy storage module is in a charging state, the corresponding signed branch current is negative.

[0139] The rated bus voltage is the common DC bus rated operating voltage recorded in the configuration parameters of the residential energy storage stacking system. Each energy storage module uses the same rated bus voltage under the same operating configuration. In this embodiment, the rated bus voltage is 400 volts.

[0140] The output reference voltage of the bidirectional DC / DC converter corresponding to the i-th energy storage module is determined according to the following formula:

[0141] ;

[0142] in, This represents the output reference voltage of the bidirectional DC / DC converter corresponding to the i-th energy storage module. Indicates the rated bus voltage. Let represent the target virtual impedance of the i-th energy storage module. This represents the signed branch current of the i-th energy storage module.

[0143] In the discharge state, the signed branch current is positive, and the product of the target virtual impedance and the signed branch current forms a positive virtual impedance voltage drop. In the charging state, the signed branch current is negative, and the product forms a negative virtual impedance voltage drop. The output reference voltage is determined according to the above-mentioned relationship of the signed branch current in both the charging and discharging states.

[0144] The target virtual impedance obtained in step S4 is a virtual impedance parameter formed after the basic virtual impedance has been corrected for current distribution deviation. The basic virtual impedance is determined based on the difference between the target current-sharing impedance and the effective line impedance. Therefore, this step directly uses the target virtual impedance and the signed branch current to calculate the virtual impedance voltage drop.

[0145] For example, the rated bus voltage is 400 volts, the current target virtual impedance of the first energy storage module is 1.314 milliohms, and the signed branch current is 11.8 amps. Multiplying 1.314 milliohms by 11.8 amps gives a virtual impedance voltage drop of approximately 0.0155 volts; subtracting 0.0155 volts from 400 volts gives an output reference voltage of approximately 399.9845 volts for the corresponding bidirectional DC / DC converter.

[0146] When the first energy storage module is in the charging state and the signed branch current is −11.8 amps, the virtual impedance voltage drop is approximately −0.0155 volts, and the output reference voltage is approximately 400.0155 volts according to the same calculation relationship.

[0147] The adjustment of the output of the bidirectional DC / DC converter corresponding to each energy storage module according to the output reference voltage includes:

[0148] The voltage regulation deviation is obtained by calculating the difference between the output reference voltage of the bidirectional DC / DC converter corresponding to each energy storage module and the corresponding bus-side voltage.

[0149] Adjust the pulse width modulation duty cycle of the corresponding bidirectional DC / DC converter according to the voltage regulation deviation, re-collect the branch current of each energy storage module, and obtain the feedback branch current;

[0150] The absolute difference between the absolute value of the feedback branch current of each energy storage module and the corresponding target branch current is calculated to obtain the current sharing deviation.

[0151] If the current sharing deviation of each energy storage module is not greater than the preset allowable current sharing deviation, then the feedback branch current is determined as the current sharing branch current corresponding to the capacity current sharing ratio.

[0152] In one implementation, the bidirectional DC / DC converter employs a bidirectional Buck-Boost converter structure capable of adjusting bidirectional power transfer via PWM duty cycle. The bidirectional DC / DC converter uses a cascaded closed-loop control structure consisting of an outer voltage loop and an inner current loop. The outer voltage loop receives the output reference voltage, while the inner current loop adjusts the PWM duty cycle of the corresponding power switching devices.

[0153] The voltage regulation deviation is determined by subtracting the current bus-side voltage from the output reference voltage of the corresponding bidirectional DC / DC converter. The outer voltage loop performs proportional-integral calculations on the voltage regulation deviation and outputs the corresponding signed current reference value.

[0154] The difference between the signed current reference value and the current signed branch current is calculated to obtain the current regulation deviation. The inner current loop performs proportional-integral calculation on the current regulation deviation to obtain the PWM duty cycle correction amount. The current PWM duty cycle is updated according to the PWM duty cycle correction amount, and the corresponding power switch drive signal is output by the PWM modulation module.

[0155] When the bidirectional DC / DC converter is in the discharging state, the PWM modulation module updates the duty cycle of the corresponding power switching device according to the power flow direction of energy transfer from the energy storage module to the common DC bus; when the bidirectional DC / DC converter is in the charging state, the PWM modulation module updates the duty cycle of the corresponding power switching device according to the power flow direction of energy transfer from the common DC bus to the energy storage module.

[0156] The voltage outer loop and current inner loop are determined and written into the controller parameter area using proportional-integral (PI) parameters during the commissioning phase of the bidirectional DC / DC converter. During commissioning, the current inner loop is tuned first. Within the rated branch current range, the current reference value is changed, and the PI parameters of the current inner loop are adjusted. The stability determination method used to determine the settling time of the inner current loop is then used to determine whether the actual branch current has entered a stable state. Subsequently, while keeping the PI parameters of the current inner loop unchanged, the PI parameters of the voltage outer loop are adjusted so that the bus-side voltage follows the given voltage reference change, and the dynamic change rate of the voltage outer loop is lower than that of the current inner loop. After tuning, the corresponding PI parameters of the voltage outer loop and the current inner loop are written into the controller parameter area. During current sharing control, these parameters are called to execute the voltage and current regulation of the bidirectional DC / DC converter.

[0157] After updating the output reference voltage for the current current sharing control cycle, at the next synchronous sampling moment, the branch current of each energy storage module is re-acquired according to the branch current acquisition process described in step S1, and the re-acquired branch current is determined as the feedback branch current.

[0158] Calculate the absolute value of the current in each feedback branch, and then calculate the absolute difference between the absolute value of the current in each feedback branch and the corresponding target branch current. The obtained value is determined as the current sharing deviation of the corresponding energy storage module.

[0159] It should be noted that the preset current sharing allowable deviation is determined during the system configuration phase based on the maximum absolute measurement error of the bus current measurement link and the branch current measurement link.

[0160] Read the maximum absolute measurement error of the bus current sensor within the current range and operating temperature range, and then read the maximum absolute measurement error of each branch current sensor within the corresponding range and operating temperature range. Select the maximum value from the maximum absolute measurement errors of all branch current sensors, add the maximum value to the maximum absolute measurement error of the bus current sensor, and determine the result as the preset current sharing allowable deviation. Write the current sharing allowable deviation into the current sharing parameter area of ​​the controller.

[0161] The target branch current is obtained by multiplying the capacity shunting ratio by the bus current amplitude. The capacity shunting ratio ranges from 0 to 1. Therefore, the absolute error of the bus current measurement error after being transmitted to the individual target branch current will not exceed the maximum absolute measurement error of the bus current sensor. The bus current measurement error and the branch current measurement error are added together at their maximum values ​​to form a unified current sharing judgment boundary.

[0162] For example, the maximum absolute measurement error of the bus current sensor in this embodiment is 0.2 A within the corresponding range and operating temperature range, and the maximum absolute measurement error of the four branch current sensors is also 0.2 A. Adding 0.2 A to 0.2 A yields the allowable deviation of 0.4 A for current sharing, and this 0.4 A is written into the current sharing parameter area of ​​the controller.

[0163] When the current sharing deviation of all energy storage modules is no greater than 0.4 amps, the current feedback branch current is determined as the current sharing branch current corresponding to the capacity current sharing ratio.

[0164] When the current sharing deviation of any energy storage module is greater than 0.4 amps, the current feedback branch current is used as the branch current corresponding to the next current sharing control cycle, the current distribution deviation is recalculated, the target virtual impedance and output reference voltage are re-determined based on the updated current distribution deviation, and the corresponding bidirectional DC / DC converter is adjusted based on the updated output reference voltage.

[0165] After the allowable deviation for current sharing is reached, subsequent current sharing control cycles continue to acquire bus current, capacity parameters, and branch current. When the capacity parameters or the energy storage modules participating in current sharing control change, the capacity shunting ratio and target branch current are redefined. When a new line impedance and impedance status level are formed in step S2, the effective line impedance is updated according to the impedance status level described in step S4. When the effective line impedance is updated, the target current sharing impedance and basic virtual impedance are redefined based on the updated effective line impedance. When the feedback branch current change causes the current sharing deviation to exceed the allowable deviation for current sharing, the impedance correction amount and target virtual impedance are updated according to the new current distribution deviation, and the corresponding bidirectional DC / DC converter continues to be adjusted.

[0166] It should be noted that in the actual use of residential energy storage stacking systems, users can add or remove energy storage modules through the stacking structure to adjust the total system capacity. When a new energy storage module is put into operation and completes the communication connection with the battery management system, the controller will add the energy storage module to the current sharing calculation from the next current sharing control cycle and re-execute the process of determining the capacity current sharing ratio and the target branch current. When the energy storage module is taken out of operation, it will be removed from the next current sharing control cycle and the capacity current sharing ratio of the remaining modules will be recalculated.

[0167] For example, the target branch currents of the four energy storage modules are approximately 11.76 A, 10.59 A, 9.41 A, and 8.24 A, respectively. After the adjustment of the current sharing control cycle is completed, the feedback branch currents of the four modules are re-acquired and found to be 11.77 A, 10.55 A, 9.45 A, and 8.23 ​​A, respectively.

[0168] The absolute difference between the absolute value of the feedback branch current and the corresponding target branch current was calculated, resulting in current sharing deviations of 0.01 A, 0.04 A, 0.04 A, and 0.01 A for the four energy storage modules. Since none of the four deviations exceeded 0.4 A, 11.77 A, 10.55 A, 9.45 A, and 8.23 ​​A were determined as the current sharing branch currents corresponding to the respective capacity current sharing ratios.

[0169] For example, Figure 4 This is a schematic diagram of the consistency of current sharing in the branch circuit. The diagram shows the correspondence between the feedback branch current and the target branch current of the four energy storage modules. The data points of each energy storage module are closely distributed near the current sharing branch current. The maximum current sharing deviation in the data is 0.159 amps, which is less than the allowable deviation of 0.4 amps, indicating that the feedback branch current can maintain a current sharing state corresponding to the capacity current sharing ratio.

[0170] In summary, this invention obtains the voltage, current, and capacity parameters of the common DC bus and the corresponding branches of each energy storage module to determine the line impedance corresponding to each energy storage module. It then constructs impedance state characteristics based on the line impedance over multiple consecutive current sharing control cycles and determines the impedance state level using a support vector machine. The effective line impedance is updated based on the impedance state level, and the target current sharing impedance and basic virtual impedance are determined by combining the capacity current sharing ratio. The basic virtual impedance is corrected based on the current distribution deviation to obtain the target virtual impedance. Finally, the output reference voltage of the bidirectional DC / DC converter is determined based on the target virtual impedance and the branch current, and the current distribution state is continuously updated based on the feedback branch current, thus achieving multi-module adaptive current sharing control.

[0171] The second embodiment of the present invention provides a multi-module adaptive current sharing control system for a residential energy storage stacking system, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.

[0172] It should be noted that the system embodiments described above are merely illustrative. 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 the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0173] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A multi-module adaptive current sharing control method for a residential energy storage stacking system, characterized in that, include: Obtain the bus voltage, bus current, bus-side voltage of each energy storage module, branch current and capacity parameters of the residential energy storage stacking system; The line impedance is determined based on the bus side voltage, the bus voltage, and the branch current. Impedance state features are constructed based on the line impedance of multiple consecutive current sharing control cycles. The impedance state features are then input into a pre-trained support vector machine to obtain the impedance state level. The capacity shunting ratio is determined based on the capacity parameters, the target branch current is determined based on the capacity shunting ratio and the bus current, and the current distribution deviation is determined based on the branch current and the target branch current. The effective line impedance is updated according to the impedance status level, the target current sharing impedance is determined according to the capacity shunt ratio and the effective line impedance, the basic virtual impedance is determined according to the difference between the target current sharing impedance and the effective line impedance, and the basic virtual impedance is corrected according to the current distribution deviation to obtain the target virtual impedance. The output reference voltage is determined based on the target virtual impedance and the branch current, and the output of the bidirectional DC / DC converter corresponding to each energy storage module is adjusted based on the output reference voltage.

2. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The acquisition of bus voltage, bus current, bus-side voltage of each energy storage module, branch current, and capacity parameters of the residential energy storage stacking system includes: Obtain the bus voltage and bus current of the common DC bus; Obtain the port voltage of the bidirectional DC / DC converter connected to the common DC bus for each energy storage module, determine the port voltage as the bus-side terminal voltage of the corresponding energy storage module, and collect the branch current between the bidirectional DC / DC converter and the common DC bus for each energy storage module. Read the current available capacity output by the battery management system of each energy storage module, and determine the current available capacity as the capacity parameter.

3. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The process of determining the line impedance based on the bus-side voltage, the bus voltage, and the branch current, and constructing impedance state characteristics based on the line impedance over multiple consecutive current-sharing control cycles, includes: The absolute value of the voltage difference between the bus side voltage and the bus voltage of each energy storage module is calculated according to the same sampling time, and the absolute value of the corresponding branch current is extracted. The absolute value of the voltage difference and the absolute value of the corresponding branch current are combined to obtain impedance sampling data. The impedance sampling data is fitted using least squares, and the slope of the fitted line is determined as the line impedance of the corresponding energy storage module. Arrange multiple consecutive line impedances according to the current sharing control cycle to obtain a line impedance sequence; The impedance change amplitude and slope are determined based on the impedance sequence. The impedance fitting deviation is determined based on the deviation between the impedance sampling data and the corresponding fitted line. The impedance state characteristics are obtained by combining the impedance change amplitude, the impedance change slope, and the impedance fitting deviation.

4. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The step of determining the capacity offloading ratio based on the capacity parameters includes: The total capacity of the modules is obtained by summing the capacity parameters of each energy storage module. Calculate the ratio between the capacity parameter of each energy storage module and the total capacity of the module to obtain the capacity diversion ratio of each energy storage module.

5. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The step of determining the target branch current based on the capacity shunting ratio and the bus current, and determining the current distribution deviation based on the branch current and the target branch current, includes: Calculate the absolute value of the bus current to obtain the bus current amplitude, and calculate the product of the capacity shunting ratio of each energy storage module and the bus current amplitude to obtain the corresponding target branch current; Calculate the absolute value of the branch current of each energy storage module to obtain the branch current amplitude. Calculate the difference between the branch current amplitude and the corresponding target branch current to obtain the current distribution deviation. A positive current distribution deviation indicates that the branch current amplitude of the corresponding energy storage module is higher than the target branch current.

6. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The steps of updating the effective line impedance according to the impedance status level, determining the target current-sharing impedance according to the capacity shunt ratio and the effective line impedance, and determining the basic virtual impedance according to the difference between the target current-sharing impedance and the effective line impedance include: Update the corresponding effective line impedance according to the impedance status level of each energy storage module; Specifically, when the impedance state level is in a continuous drift state, the current line impedance is determined as the effective line impedance; when the impedance state level is in a stable state, the arithmetic mean of the current line impedance and the most recent effective line impedance is updated as the effective line impedance; and when the impedance state level is in a transient disturbance state, the most recent effective line impedance is maintained. The impedance weight value is obtained by multiplying the effective line impedance of each energy storage module by the corresponding capacity shunt ratio. The maximum value among the impedance weights is selected as the reference value for current sharing impedance. Calculate the ratio between the current sharing impedance benchmark value and the capacity current sharing ratio corresponding to each energy storage module to obtain the target current sharing impedance of each energy storage module. The difference between the target current-sharing impedance of each energy storage module and the corresponding effective line impedance is calculated to obtain the corresponding basic virtual impedance.

7. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The step of correcting the basic virtual impedance based on the current distribution deviation to obtain the target virtual impedance includes: The proportional-integral calculation is performed on the current distribution deviation of each energy storage module to obtain the corresponding impedance correction amount; The corrected virtual impedance is obtained by summing the basic virtual impedance and the corresponding impedance correction amount. The modified virtual impedance is limited according to the allowable virtual impedance range of the corresponding bidirectional DC / DC converter to obtain the target virtual impedance.

8. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The step of determining the output reference voltage based on the target virtual impedance and the branch current includes: The sign of the branch current of each energy storage module is determined by taking the discharge direction as the positive direction and the charging direction as the negative direction, thus obtaining the signed branch current. The virtual impedance voltage drop is obtained by multiplying the target virtual impedance by the corresponding signed branch current. The difference between the rated bus voltage of the residential energy storage stack system and the virtual impedance voltage drop is calculated to obtain the corresponding output reference voltage of the bidirectional DC / DC converter.

9. The multi-module adaptive current sharing control method for a residential energy storage stacking system according to claim 1, characterized in that, The step of adjusting the output of the bidirectional DC / DC converter corresponding to each energy storage module according to the output reference voltage includes: The voltage regulation deviation is obtained by calculating the difference between the output reference voltage of the bidirectional DC / DC converter corresponding to each energy storage module and the corresponding bus-side voltage. Adjust the pulse width modulation duty cycle of the corresponding bidirectional DC / DC converter according to the voltage regulation deviation, re-collect the branch current of each energy storage module, and obtain the feedback branch current; The absolute difference between the absolute value of the feedback branch current of each energy storage module and the corresponding target branch current is calculated to obtain the current sharing deviation. If the current sharing deviation of each energy storage module is not greater than the preset allowable current sharing deviation, then the feedback branch current is determined as the current sharing branch current corresponding to the capacity current sharing ratio.

10. A multi-module adaptive current sharing control system for a residential energy storage stacking system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method according to any one of claims 1 to 9.

Images