A dynamic management and control system of a lithium battery energy storage system

Abstract

The application relates to the technical field of energy storage system management and control, and discloses a dynamic management and control system of a lithium battery energy storage system, which comprises a sensing layer used for collecting voltage data, current data, temperature data and impedance data of a battery array during operation; a processing layer used for dividing the battery array into a plurality of logical management clusters according to the voltage data, the current data, the temperature data and the impedance data, and calculating a bus current change rate in real time, wherein the current change rate is the absolute value of the first derivative of the current with respect to time; and the processing layer is further used for triggering high-frequency sampling and electrochemical impedance online inversion when the current change rate exceeds a preset threshold value. The high-frequency sampling triggered by the current change rate and the electrochemical impedance online inversion can capture the battery polarization characteristics in real time and correct an SOC calculation model, and the SOC estimation error is reduced by about 10% compared with a traditional fixed-frequency BMS under a dynamic stress test working condition, and is stabilized within 5%.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy storage system management and control, in particular to a dynamic management and control system of a lithium battery energy storage system. BACKGROUND

[0002] In the current energy storage industry, the battery management system BMS generally adopts fixed frequency sampling and passive balancing technology. However, with the increasing requirements of the power grid on the response speed of the energy storage system such as primary frequency modulation and secondary frequency modulation, the existing technology has the following technical defects:

[0003] The traditional BMS adopts fixed to Low frequency sampling cannot capture the transient polarization characteristics when the charging and discharging current suddenly changes. In the dynamic stress test DST and other fast variable power conditions, the SOC state of charge error can accumulate to above in minutes, which seriously affects the accuracy of power grid scheduling.

[0004] In the existing battery management architecture, the physical battery cluster is strongly coupled with the control logic, and the charging and discharging depth of each cluster is fixedly allocated by the hardware circuit. When a certain battery cluster appears performance degradation, it cannot dynamically adjust its power quota, which leads to single point failure easily causing the overall system to run at reduced capacity or even shut down.

[0005] The traditional active balancing technology adopts resistance discharging method, which converts the excess charge of the high-voltage monomer in the cluster into heat energy through resistance and dissipates it. This method not only has low energy utilization rate with a loss of to , but also generates additional heat, which increases the burden of the thermal management system and cannot fundamentally correct the consistency degradation caused by the internal resistance difference. SUMMARY

[0006] In view of the deficiencies of the prior art, the present application provides a dynamic management and control system of a lithium battery energy storage system, which solves the problems raised in the above background art.

[0007] The present application provides the following technical solution: a dynamic management and control system of a lithium battery energy storage system, comprising:

[0008] A perception layer for collecting voltage data, current data, temperature data and impedance data of the battery array during operation;

[0009] A processing layer for dividing the battery array into a plurality of logical management clusters according to the voltage data, current data, temperature data and impedance data, and calculating the bus current rate of change in real time, wherein the current rate of change is the absolute value of the first derivative of the current with respect to time;

[0010] The processing layer is further configured to increase the sampling frequency of each logical management cluster when the current rate of change exceeds a preset threshold, and correct the battery state parameter based on the extracted impedance parameter;

[0011] The execution layer includes a power distribution unit and a fluid control unit;

[0012] The power distribution unit includes a plurality of bidirectional converters, each corresponding to a logical management cluster, and adjusts the PWM duty cycle of each bidirectional converter to distribute the charging and discharging power according to the scheduling instructions output by the processing layer;

[0013] The fluid control unit includes a plurality of proportional regulating valves, which adjust the flow rate of the cooling fluid at each battery monomer according to the temperature control instructions output by the processing layer, adjust the local temperature by changing the convective heat transfer coefficient of the monomer surface, and correct the internal resistance difference between the battery monomers by using the negative correlation between internal resistance and temperature.

[0014] Preferably, the processing layer performs multi-scale sampling when:

[0015] When the current rate of change exceeds a preset threshold, the sampling frequency is switched from the normal sampling frequency to the high-frequency sampling frequency, and the power exchange module is instructed to inject a weak sinusoidal excitation signal to the battery array to extract the electrochemical impedance characteristics;

[0016] Wherein, the high-frequency sampling frequency is times to times of the normal sampling frequency.

[0017] Preferably, the processing layer corrects the SOC calculation model using the extracted dynamic impedance:

[0018] The dynamic impedance is obtained by inverse calculation of the relationship between the terminal voltage, open circuit voltage and current, and includes Ohmic resistance, polarization resistance and diffusion resistance obtained by high-frequency sampling data inversion;

[0019] The processing layer corrects the ampere-hour integral result in real time by pre-stored compensation parameter table, and the compensation parameter table records the corresponding relationship between Ohmic resistance deviation, temperature deviation and SOC correction value.

[0020] Preferably, the power distribution logic of the power distribution unit includes:

[0021] Collect the health, real-time impedance and temperature rise rate of each logical management cluster;

[0022] Based on the correlation between the health and the real-time impedance, the PWM duty cycle of the corresponding bidirectional converter is adjusted, so that the logical management cluster with higher health and lower real-time impedance carries a higher proportion of charging and discharging current.

[0023] Preferably, the processing layer is further configured to, when the temperature rise rate of a specific logical management cluster reaches or exceeds a preset safety threshold, adjust the power distribution unit by a smooth transition function to transfer the current load of the logical management cluster to the remaining logical management clusters;

[0024] wherein the smooth transition function is a logistic function , is the migration start time, is the steepness coefficient.

[0025] Preferably, the fluid control unit adjusts the opening of the proportional regulating valve to change the cooling liquid flow rate, thereby changing the convective heat transfer coefficient of the monomer surface, and using the negative correlation between the battery internal resistance and the temperature to correct the deviation.

[0026] The present application has the following beneficial effects:

[0027] High-frequency sampling triggered by current change rate Combined with online electrochemical impedance inversion, the battery polarization characteristics can be captured in real time and the SOC calculation model can be corrected, and under dynamic stress test conditions, the SOC estimation error is reduced by about compared with the traditional fixed frequency BMS, and is stably maintained within .

[0028] Through the modular power scheduling architecture, the physical battery cluster is taken as a management unit that can dynamically allocate load, when a management cluster appears to be in a sub-healthy state such as rapid temperature rise or internal resistance increase, the system can smoothly transfer its load to other healthy clusters within milliseconds, avoiding the risk of electric arc and power drop caused by the disconnection of contactors in the traditional scheme.

[0029] The temperature field active intervention technology is used to replace the traditional resistance discharge balancing, by adjusting the monomer level cooling flow rate, using the temperature negative correlation characteristic Arrhenius effect of the battery internal resistance to realize current redistribution, this method does not need to convert electrical energy into heat energy dissipation, the energy loss of the balancing process is reduced by compared with the traditional active balancing, and effectively delays the aging of the thermal management system, in charge-discharge rate, times of cycle test, the total energy throughput efficiency of the system is improved by about .

[0030] The control actions are all realized by the physical components of the hardware execution layer 4, bidirectional converter and proportional regulating valve, without relying on complex software logic, the system has strong robustness. The load migration process uses a smooth transition function to control the PWM duty cycle, ensuring that the bus current fluctuation rate is less than , meeting the strict requirements of power quality of the power grid access standard. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 The overall architecture diagram of the dynamic management and control system 1 of the application;

[0032] Figure 2 The flow chart of multi-scale adaptive sampling trigger logic;

[0033] Figure 3 The block diagram of electrochemical impedance inversion and SOC correction algorithm;

[0034] Figure 4 The schematic diagram of modular power scheduling and load smoothing migration;

[0035] Figure 5 The control closed-loop principle diagram of temperature field active balancing.

[0036] 1, dynamic management and control system; 2, perception layer; 3, processing layer; 4, execution layer; 5, battery array; 6, DC bus;

[0037] 101, temperature sensor; 102, current sensor; 103, voltage acquisition unit; 104, communication bus; 105, data acquisition card;

[0038] 201, heterogeneous computing platform; 202, field programmable gate array; 203, microprocessor; 204, memory sharing bus; 205, scheduling control module; 206, state estimation module;

[0039] 301, power distribution unit; 302, fluid control unit; 303, bidirectional converter; 304, proportional regulating valve; 305, step driver; 306, cooling liquid pipeline;

[0040] 401, battery monomer; 402, battery cluster; 403, temperature detection point; 404, current detection point; 405, cooling liquid inlet; 406, cooling liquid outlet;

[0041] 501, power exchange module; 502, isolation drive circuit; 503, control signal interface;

[0042] 601, cooling pump; 602, liquid storage tank; 603, flow meter; 604, pressure sensor. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the application.

[0044] Referring to Figure 1 - Figure 5 A dynamic management system of a lithium battery energy storage system, comprising:

[0045] A perception layer 2 for collecting voltage data, current data, temperature data and impedance data of the battery array 5 during operation;

[0046] A processing layer 3 for dividing the battery array 5 into a plurality of logical management clusters according to the voltage data, current data, temperature data and impedance data, and calculating a bus current rate of change in real time, wherein the current rate of change is the absolute value of the first derivative of the current with respect to time;

[0047] The processing layer 3 is further configured to increase the sampling frequency of each logical management cluster when the current rate of change exceeds a preset threshold, and correct the battery state parameters based on the extracted impedance parameters;

[0048] An execution layer 4 comprising a power distribution unit 301 and a fluid control unit 302;

[0049] The power distribution unit 301 comprises a plurality of bidirectional converters 303, each corresponding to a logical management cluster, and adjusts the PWM (Pulse Width Modulation, a technique for using the digital output of a microprocessor to control an analog circuit) duty cycle of each bidirectional converter 303 to distribute charge and discharge power according to the scheduling instructions output by the processing layer 3;

[0050] The fluid control unit 302 comprises a plurality of proportional regulating valves 304, which adjust the flow rate of the cooling fluid at each battery monomer 401 according to the temperature control instructions output by the processing layer 3, adjust the local temperature by changing the convective heat transfer coefficient of the monomer surface, and correct the internal resistance difference between the battery monomers 401 by using the negative correlation between internal resistance and temperature.

[0051] In a preferred embodiment, when the processing layer 3 performs multi-scale sampling:

[0052] When the current rate of change exceeds the preset threshold, the sampling frequency is switched from the normal sampling frequency to the high-frequency sampling frequency, and the power exchange module 501 is instructed to inject a weak sinusoidal excitation signal into the battery array 5 to extract the electrochemical impedance characteristics;

[0053] Wherein, the high-frequency sampling frequency is 100 to 1000 times the normal sampling frequency.

[0054] In a preferred embodiment, the processing layer 3 corrects the SOC (State of Charge, representing the percentage of the remaining battery capacity relative to the rated capacity) calculation model using the extracted dynamic impedance:

[0055] The dynamic impedance is obtained by back calculation of the relationship between the terminal voltage, open circuit voltage and current, including Ohmic impedance, polarization impedance and diffusion impedance obtained by high frequency sampling data inversion;

[0056] The processing layer 3 corrects the ampere-hour integral result in real time by pre-stored compensation parameter table, and the compensation parameter table records the corresponding relationship of Ohmic internal resistance deviation, temperature deviation and SOC correction value.

[0057] In a preferred embodiment: the power distribution logic of the power distribution unit 301 includes:

[0058] Collect the health degree, real-time impedance and temperature rise rate of each logical management cluster;

[0059] Based on the correlation between the health degree and the real-time impedance, adjust the PWM (Pulse Width Modulation, a technology that uses the digital output of a microprocessor to control an analog circuit) duty cycle of the corresponding bidirectional converter 303, so that the logical management cluster with higher health degree and lower real-time impedance carries a higher proportion of charging and discharging current.

[0060] In a preferred embodiment: the processing layer 3 is further configured to adjust the power distribution unit 301 through a smooth transition function when the temperature rise rate of a specific logical management cluster reaches or exceeds a preset safety threshold, and transfer the current load of the logical management cluster to the remaining logical management clusters;

[0061] The smooth transition function is a logistic function , is the migration start time, is the steepness coefficient.

[0062] In a preferred embodiment: the fluid control unit 302 implements the balancing logic including:

[0063] Based on the Arrhenius equation, obtain the sensitivity factor of the internal resistance of the target monomer to temperature;

[0064] According to the deviation of the target monomer and the average current in the cluster, calculate the target temperature difference compensation amount;

[0065] Adjust the opening of the proportional regulating valve 304 through the stepper driver 305 to change the cooling liquid flow rate, so as to change the convective heat transfer coefficient of the monomer surface, and correct the deviation by using the negative correlation between the internal resistance and the temperature of the battery.

[0066] A dynamic management and control method of a dynamic management and control system of a lithium battery energy storage system, comprising the following steps:

[0067] S1: Real-time monitoring of the first derivative of the bus current absolute value, and triggering high-frequency sampling mode when it exceeds the preset threshold, injecting excitation signal and extracting battery impedance parameters by inversion;

[0068] S2: Constructing the modular power scheduling architecture of battery array 5, adjusting the duty cycle of bidirectional converter 303 according to battery health characteristics to allocate power quota;

[0069] S3: When detecting the inconsistency of inter-cell current, adjust the cooling fluid flow rate to actively correct the local temperature field of the battery, and then realize lossless equalization through internal resistance compensation.

[0070] In step S2, the droop control logic is used to adjust the output voltage of each logical management cluster to ensure that the fluctuation rate of the total output power of the system during load migration is less than 1%.

[0071] Embodiment 1: System hardware architecture and data interaction

[0072] As shown in Figure 1 The dynamic control system 1 is composed of three parts: the perception layer 2, the processing layer 3 and the execution layer 4. Each layer realizes data interaction through a high-speed communication bus 104, and the core management object is the battery array 5.

[0073] The perception layer 2 uses flexible thin film platinum resistance temperature sensor 101 and Hall current sensor 102. The voltage acquisition unit 103 uses an isolated 24-bit Σ-Δ ADC (Analog-to-Digital Converter, which converts continuous variable analog signals into discrete digital signals) chip, which has high common mode rejection ratio;

[0074] The voltage acquisition unit 103 contains a high-precision A / D (Analog-to-Digital) sampling circuit, and its signal input end is connected in parallel to the positive and negative poles of each battery monomer 401 through metal leads. The temperature sensor 101 is physically attached to the temperature detection point 403 of each monomer. The Hall current sensor 102 is set on the power output bus of each battery cluster 402 to capture cluster-level current.

[0075] After the collected analog signals are converted into digital signals by the data acquisition card 105, they are transmitted to the processing layer 3 through the dual-redundancy communication bus 104.

[0076] Figure 1 CNA-FD 5Mbps in the table indicates that the Controller Area Network Flexible Data-rate (Controller Area Network Flexible Data-rate) protocol is used between the data acquisition card 105 and the field programmable gate array 202, and the transmission efficiency is 5Mbps;

[0077] FPGA, Field Programmable Gate Array, is an integrated circuit that defines functions through programming at the hardware level. FPGA processes tasks simultaneously through physical connection of internal logic gates, and is used for high-speed signal processing at the bottom layer;

[0078] AXI4, Advanced eXtensible Interface 4, is a high-performance, high-bandwidth, and low-latency on-chip bus protocol. FPGA 202 and microprocessor 203 use AXI4 bus protocol.

[0079] The processing layer 3 adopts a heterogeneous computing platform 201, including FPGA 202 and microprocessor 203. FPGA 202 is responsible for processing high-frequency real-time tasks at the millisecond level, and microprocessor 203 runs a Linux (open source, Unix-like operating system) real-time operating system, carrying a scheduling control module 205 and a state estimation module 206. FPGA 202 and microprocessor 203 exchange data through a memory sharing bus 204, with a data throughput of .

[0080] The execution layer 4 includes a power distribution unit 301 and a fluid control unit 302. The power distribution unit 301 is composed of a plurality of bidirectional converters 303. The bidirectional converter 303 adopts a CLLC resonant topology, and uses a symmetric LC resonant network on the primary and secondary sides to realize bidirectional high-efficiency exchange of electric energy;

[0081] and the input end is connected to the corresponding battery cluster 402, and the output end is uniformly collected and accessed to the DC bus 6. Each bidirectional converter 303 corresponds to a battery cluster 402. The fluid control unit 302 adopts a proportional regulating valve 304 driven by a step driver 305 to control the flow rate in the cooling liquid pipeline 306;

[0082] The control signal sent by the scheduling control module 205 controls the on-off of the power exchange module 501 through the isolation driving circuit 502, so as to adjust the energy flow of the battery cluster into the DC bus 6.

[0083] The battery array 5 is divided into N battery clusters 402, and each battery cluster 402 is composed of a plurality of battery monomers 401 in series and parallel connection. Temperature detection points 403 and current detection points 404 are arranged on the surface of each battery monomer 401. Cooling liquid enters through the cooling liquid inlet 405, flows out from the cooling liquid outlet 406 after passing through the surface of the battery monomer 401.

[0084] The fluid control unit 302 forms a closed cooling cycle. The coolant reservoir 602 stores the cooling medium, and its outlet is pressurized by the cooling pump 601 and pumped into the main coolant line 306. A flow meter 603 and a pressure sensor 604 are installed on the main cooling cycle line to monitor cycle stability. Before entering the battery array 5, the coolant line 306 branches into several branches, each equipped with a proportional regulating valve 304. The proportional regulating valve 304 is controlled by a stepper driver 305. The regulated fluid flows over the surface of the battery cells 401 for heat exchange and finally converges and flows back to the reservoir 602 through the outlet.

[0085] Example 2: Reference Figure 2 The triggering and execution process of multi-scale adaptive sampling includes the following steps:

[0086] Step 1: Real-time calculation of current change rate

[0087] The field-programmable gate array 202 of processing layer 3 uses a differential algorithm to calculate the rate of change of current on DC bus 6 in real time. Current sensor 102 collects current data at current detection point 404 and uploads it to field-programmable gate array 202 via data acquisition card 105. Let the current sampling time be... The sampling interval is Baseline mode Then the first derivative current rate of change The calculation formula is:

[0088]

[0089] in, for The bus current at any given time is expressed in amperes (A). To suppress measurement noise, the actual implementation uses... The difference is calculated after point median filtering;

[0090] Step 2: Dynamic switching of sampling frequency

[0091] The system presets the current change rate threshold. This threshold is determined based on typical energy storage application scenarios, and the primary frequency regulation response time. seconds, when calculated Exceed At this time, the field-programmable gate array 202 sends a frequency switching command to the data acquisition card 105 via the communication bus 104, switching the sampling rate of the voltage acquisition unit 103 from the normal sampling frequency to the high-frequency sampling frequency, and maintaining the high-frequency sampling window for 5 seconds. If within this window... Persistently below It will automatically restore to Reference sampling, settings To minimize hysteresis and avoid frequent switching;

[0092] The high-frequency sampling frequency ranges from 100 times to 1000 times of the normal sampling frequency. In extreme dynamic conditions such as power grid short circuit protection test, the sampling frequency can be increased to 1000 times, but considering the data processing load and actual application requirements, 100 times is sufficient to capture the battery polarization characteristics in regular applications;

[0093] Step 3: Excitation signal injection and impedance spectrum measurement

[0094] In the high-frequency sampling mode, the scheduling control module 205 of the processing layer 3 instructs the power exchange module 501 to inject the excitation signal through the control signal interface 503. The power exchange module 501 controls the bidirectional converter 303 to superimpose a sinusoidal disturbance current with an amplitude of 1A to 5A and a frequency of to swept in logarithmic scale on the basis of the charging and discharging main current. The excitation duration is an integer multiple of the period of each frequency point, typically 20 to 50 periods, to ensure the accuracy of the correlation detection;

[0095] The field programmable gate array 202 uses a digital lock-in amplification algorithm to extract the electrochemical impedance.

[0096] The algorithm principle is: multiply and integrate the high-frequency voltage response collected by the voltage acquisition unit 103 with the excitation current i(t), extract the real part and the imaginary part of the same frequency component.

[0097] In implementation, the cosine reference signal and the sine reference signal are multiplied by respectively and integrated for N periods, and the calculation formula is:

[0098]

[0099]

[0100] where is the excitation angular frequency, and N is the number of integration periods;

[0101] Step 4: Equivalent circuit parameter extraction

[0102] Based on the measured complex impedance spectrum data, the state estimation module 206 uses the least squares fitting algorithm to extract the fractional order equivalent circuit model ECM (Equivalent Circuit Model) parameters. The model impedance expression is:

[0103]

[0104] where, is the ohmic internal resistance, which is obtained by The above high-frequency point The average value extraction, The charge transfer resistance, The time constant, The fractional order exponent , The Warburg diffusion coefficient;

[0105] The fitting algorithm uses a nonlinear least squares method, and the objective function is the complex residual square sum minimization of the measured value and the model value;

[0106] Step 5: SOC (State of Charge, representing the percentage of the remaining capacity of the battery to the rated capacity) correction model

[0107] The extracted ohmic resistance For correcting the SOC (State of Charge, representing the percentage of the remaining capacity of the battery to the rated capacity) calculation of the traditional ampere-hour integral method;

[0108] The correction logic is: first, calculate the polarization voltage drop according to the difference between the open circuit voltage and the terminal voltage, the formula is:

[0109]

[0110] Wherein The measured terminal voltage, The open circuit voltage-SOC compensation lookup table is pre-calibrated through a low-rate static experiment;

[0111] Then, based on the ohmic resistance deviation, the SOC drift is corrected, and the correction formula is:

[0112]

[0113] Wherein, The ampere-hour integral preliminary calculation value, The reference resistance, the nominal value of the new battery at 25°C, The reference temperature 25°C, And The linear compensation coefficient, the linear compensation coefficient is obtained by dynamic cycle experiment fitting, the typical value , , stored in the 256 byte lookup table of EEPROM (Electrically Erasable Programmable Read-Only Memory, a non-volatile storage warehouse used by the processing layer 3 to store key calibration data);

[0114] The SOC compensation lookup table is And For two-dimensional index, each unit stores the corresponding SOC correction value;

[0115] Under dynamic stress test conditions, the average absolute error MAE (Mean Absolute Error) of SOC (State of Charge, representing the percentage of the remaining battery capacity to the rated capacity) estimation using the multi-scale sampling and impedance inversion technology of the embodiment is 0.7%, while the traditional fixed The error of the sampled BMS (Battery Management System, an electronic system for monitoring and managing the state of the battery) reaches ± 1.5% after 500 consecutive charge and discharge cycles.

[0116] Embodiment 3: The physical implementation process of modular power scheduling and load migration is as follows:

[0117] The scheduling control module 205 of the processing layer 3 periodically collects three key parameters of each battery cluster 402: the health SOH (State of Health, representing the degree of attenuation of the current performance of the battery relative to the new battery) calculated by the state estimation module 206, the real-time impedance obtained by electrochemical impedance inversion, and the temperature rise rate collected by the temperature sensor 101 at the temperature detection point 403.

[0118] The scheduling control module 205 calculates the comprehensive score of each battery cluster 402:

[0119]

[0120] Among them , , are weight coefficients, , , are weight coefficients adjusted according to the application scenario, and the typical value is , , , which satisfies the normalization condition ;

[0121] Then according to the total power demand and the cluster score, the target power quota is calculated:

[0122]

[0123] Finally, the The PWM (Pulse Width Modulation, a technique of using digital output of a microprocessor to control an analog circuit) duty cycle of the corresponding bidirectional converter 303 is converted For a step-down converter, the conversion relationship is:

[0124]

[0125] wherein Vbus is the voltage of the DC bus 6, Ib is the rated current of the th bidirectional converter 303, the PWM control word is issued to the isolation drive circuit 502 through the control signal interface 503, and the isolation drive circuit 502 drives the power switch tube of the bidirectional converter 303;

[0126] When the temperature sensor 101 of a certain battery cluster 402 (denoted as cluster A) detects that the temperature rise rate reaches or exceeds the set safety threshold, the dispatch control module 205 determines that it is in a thermal stress risk state, and starts the load migration program;

[0127] The dispatch control module 205 adopts a logistic function as a transition curve (Logistic smooth migration function):

[0128]

[0129] wherein t0 is the migration start time, k is the steepness coefficient, and the typical value is , corresponding to The transition is completed within t, and during the transition period, the PWM duty cycle of cluster A is smoothly lowered according to : t

[0130]

[0131] At the same time, the duty cycles of the remaining battery clusters 402 are proportionally increased:

[0132]

[0133] The process is realized by the real-time task period of the microprocessor 203, and the PWM (Pulse Width Modulation, a technique of using digital output of a microprocessor to control an analog circuit) control word is updated step by step. Since the transition time is much greater than the response time of the current control loop, the bus current remains smooth during the entire migration process, and the actual fluctuation rate satisfies the power quality standard for grid access;​

[0134] During load migration, no contactor or circuit breaker disconnection is involved, avoiding the risk of contact erosion caused by electric arcing in traditional solutions. All control is achieved through PWM (Pulse Width Modulation, a technique that uses the digital output of a microprocessor to control analog circuits), significantly improving safety. Once the temperature of cluster A returns to the normal range, the system automatically initiates the reverse migration procedure to restore power balance.

[0135] Example 4: Closed-loop control based on internal resistance compensation and equalization technology under temperature field regulation;

[0136] The internal resistance of the lithium battery cell 401 is negatively correlated with temperature, following the Arrhenius equation:

[0137]

[0138] Where A is the pre-exponential factor. Activation energy, typical value for lithium-ion batteries to R is the gas constant. Where T is the absolute temperature in Kelvin (K). Differentiating this equation yields the temperature sensitivity factor of the internal resistance:

[0139]

[0140] Typical lithium batteries in hour, That is, for every 1°C increase in temperature, the internal resistance decreases by about 5%. By utilizing this characteristic, internal resistance compensation can be achieved by actively adjusting the local temperature of the individual unit, thereby correcting the uneven current distribution in the parallel circuit.

[0141] The system uses current sensor 102 to measure the current of each battery cell 401 within the cluster in real time at current detection point 404. When a battery cell 401 is detected, it is recorded as... , current with average current within the cluster deviation When the threshold is reached or exceeded, the balancing process is initiated.

[0142] First, the electrochemical impedance spectroscopy from Example 2 was used to obtain... Real-time internal resistance , with average internal resistance within the cluster Comparison. If ,illustrate The excessive internal resistance leads to insufficient current distribution, requiring a temperature increase to reduce the internal resistance. The target temperature difference compensation amount is... The calculation formula is derived as follows:

[0143] According to Ohm's law, in a parallel circuit, the current is inversely proportional to the internal resistance; in order to make The current increases Its internal resistance needs to be reduced to Change in internal resistance Then, based on the internal resistance-temperature sensitivity factor... The required temperature change is:

[0144]

[0145] Substitute typical values If it is necessary to reduce internal resistance Then the temperature needs to be increased by about 4°C.

[0146] The fluid control unit 302 of the execution layer 4 calculates... ,adjust The corresponding proportional control valve 304 opening degree, the specific control logic is as follows:

[0147] If heating is required This reduces the opening of the proportional control valve 304, decreasing the flow rate in the coolant line 306 from the normal flow rate. Down to The convective heat transfer coefficient of the surface of the 401 battery cell From the agreement dropped to about ;

[0148] According to the heat balance equation, the temperature of the monomer will rise slowly until a new steady state is reached;

[0149] If cooling is required Then increase the opening of the proportional control valve 304 to increase the coolant flow rate to The convective heat transfer coefficient is increased to approximately This accelerates heat dissipation from individual units;

[0150] After receiving a control command, the stepper driver 305 drives the stepper motor of the proportional control valve 304 to rotate a specific angle, typically 15 degrees. Changes in flow rate alter the valve opening.

[0151] Temperature sensor 101 continuously collects the real-time temperature of battery cell 401 at temperature detection point 403 and feeds it back to scheduling control module 205. Scheduling control module 205 uses a PID (Proportional-Integral-Derivative, a commonly used closed-loop feedback control algorithm) controller with a Smith predictor (a predictive compensation control strategy designed for systems with large time lag) to calculate the control quantity. :

[0152]

[0153] in The temperature error is the difference between the target temperature and the measured temperature. , , These are the PID parameters, which are tuned using the Ziegler-Nichols method (a heuristic method for rapidly obtaining PID (proportional-integral-derivative) controller parameters through experimentation). Typical values ​​are... , , , Smith estimates the compensation term based on the fluid heat transfer time lag model, with a typical lag time of 30 to 60 seconds, predicts the future temperature response and adjusts it in advance.

[0154] In a real-world test case, the internal resistance of a certain cell 401 within a cluster decreased due to aging. Increase to The current measured at current detection point 404 was 12% lower than the average. After detecting the deviation, the system adjusted the corresponding proportional control valve 304 via stepper driver 305, reducing the cooling flow rate from... Down to After 8 minutes of adjustment, temperature sensor 101 measured the temperature of the monomer at temperature detection point 403, which showed an increase from 28°C to 32°C, and the internal resistance decreased. The current deviation measured at current detection point 404 was reduced to 3%. The entire process requires no resistor discharge, and the energy loss is only a slight increase in the power of the cooling pump 601. Compared to traditional balancing methods, the loss Reduced by more than 90%.

[0155] Example 5: Building a system with a capacity of The lithium battery energy storage system, battery array 5 includes Each battery cluster 402 consists of 60 lithium iron phosphate battery cells 401 connected in series and 100 parallel. The sensing layer 2 is equipped with 240 temperature sensors 101, 4 current sensors 102, and 1 voltage acquisition unit 103. The processing layer 3 uses a Xilinx Zynq-7020 (specific product model of the programmable logic device) heterogeneous computing platform 201. The execution layer 4 contains 4... A bidirectional converter 303 and 60 proportional control valves 304. The cooling system includes a cooling pump 601 with power... 602, liquid storage tank, 603, flow meter and pressure sensor 604.

[0156] Benchmark System: A 2023 standard liquid-cooled BMS from a certain international brand was selected as the benchmark. This system uses a fixed... Sampling, resistive active equalization, each string is configured Equalized resistance, overall liquid cooling control, and fixed flow rate .

[0157] Test conditions: Conducted in a professional battery testing laboratory, with ambient temperatures set as follows: and .

[0158] The test items include:

[0159] 1. Dynamic Stress Test (DST) (Dynamic Stress Test, a standard test procedure that simulates the complex driving conditions of electric vehicles): Simulates the variable power charging and discharging conditions of electric vehicles.

[0160] 2. Primary frequency modulation simulation: using The frequency fluctuates within ±10% of the power range;

[0161] 3. Long cycle life test: Charge and discharge cycles, 1000 times in total;

[0162] Comparison of results:

[0163] Experimental data demonstrate that the dynamic control system proposed in this application significantly outperforms existing technologies in key indicators such as state estimation accuracy, system availability, and energy efficiency, and has good engineering application value.

[0164] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0165] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. A dynamic control system for a lithium battery energy storage system, characterized in that, include: The sensing layer (2) is used to collect voltage data, current data, temperature data and impedance data of the battery array (5) during operation; The processing layer (3) is used to divide the battery array (5) into several logic management clusters based on voltage data, current data, temperature data and impedance data, and to calculate the bus current change rate in real time, wherein the current change rate is the absolute value of the first derivative of the current with respect to time. The processing layer (3) is also used to increase the sampling frequency of each logic management cluster when the current change rate exceeds the preset threshold, and to correct the battery state parameters based on the extracted impedance parameters. The execution layer (4) includes a power distribution unit (301) and a fluid control unit (302). The power distribution unit (301) includes several bidirectional converters (303), each bidirectional converter (303) corresponds to a logic management cluster, and adjusts the PWM duty cycle of each bidirectional converter (303) according to the scheduling instructions output by the processing layer (3) to distribute charging and discharging power. The fluid control unit (302) includes several proportional regulating valves (304), which adjust the flow rate of cooling fluid at each battery cell (401) according to the temperature control command output by the processing layer (3), adjust the local temperature by changing the convective heat transfer coefficient of the cell surface, and correct the internal resistance difference between battery cells (401) by utilizing the negative correlation between internal resistance and temperature.

2. The dynamic control system for a lithium battery energy storage system according to claim 1, characterized in that, When processing layer (3) performs multi-scale sampling: When the rate of change of current exceeds the preset threshold, the sampling frequency is switched from the normal sampling frequency to the high-frequency sampling frequency, and the power exchange module (501) is instructed to inject a weak sinusoidal excitation signal into the battery array (5) to extract the electrochemical impedance characteristics. Among them, the high-frequency sampling frequency is the normal sampling frequency. Doubled times.

3. The dynamic control system for a lithium battery energy storage system according to claim 2, characterized in that, The processing layer (3) uses the extracted dynamic impedance to correct the SOC calculation model: Dynamic impedance is obtained by inversely calculating the relationship between terminal voltage, open-circuit voltage and current, including ohmic impedance, polarization impedance and diffusion impedance obtained by inverse high-frequency sampling data; The processing layer (3) performs real-time drift correction on the ampere-hour integration result through a pre-stored compensation parameter table. The compensation parameter table records the correspondence between the ohmic internal resistance deviation, temperature deviation and SOC correction value.

4. The dynamic control system for a lithium battery energy storage system according to claim 1, characterized in that, The power distribution logic of the power distribution unit (301) includes: Collect the health status, real-time impedance, and temperature rise rate of each logic management cluster; Based on the correlation between health status and real-time impedance, the PWM duty cycle of the corresponding bidirectional converter (303) is adjusted so that the logic management cluster with higher health status and lower real-time impedance can carry a higher proportion of charging and discharging current.

5. A dynamic control system for a lithium battery energy storage system according to claim 1 or 4, characterized in that, The processing layer (3) is also used to adjust the power distribution unit (301) through a smooth transition function when the temperature rise rate of a specific logic management cluster reaches or exceeds a preset safety threshold, so as to transfer the current load of the logic management cluster to the other logic management clusters. Wherein, the smooth transition function is the logistic function. , For the migration start time, This is the steepness coefficient.

6. The dynamic control system for a lithium battery energy storage system according to claim 1, characterized in that, The logic for implementing balancing in the fluid control unit (302) includes: The temperature sensitivity factor of the internal resistance of the target monomer is obtained based on the Arrhenius equation. The target temperature difference compensation is calculated based on the deviation between the target single cell and the average current within the cluster. The opening of the proportional control valve (304) is adjusted by the stepper driver (305) to change the coolant flow rate, thereby changing the convective heat transfer coefficient of the cell surface and using the negative correlation between the battery internal resistance and temperature to correct the deviation.

Images