Battery cell rotation discharging control method of semi-solid-state ups power supply

Abstract

This invention relates to the field of power electronics and electrochemical energy storage control technology, and discloses a method for rotating discharge control of individual battery cells in a semi-solid-state UPS power supply. The method includes: collecting individual battery cell data to analyze the state of charge and health status; dynamically calculating the duration of the control cycle based on load fluctuation characteristics; mapping individual cells to a two-dimensional priority interval based on state values; calculating the marginal lifetime cost index in conjunction with expected physical losses; performing a two-dimensional sorting based on discharge priority and marginal lifetime cost to generate an ordered discharge sequence; accumulating and truncating the sequence according to the DC bus reference voltage, determining the target working group and controlling its connection to the main circuit, while the remaining cells are bypassed and isolated. This invention decouples the physical series connection limitation of batteries through an intelligent scheduling strategy, overcomes the battery pack's bottleneck effect, and achieves flexible isolation of faulty cells, extending the overall service life of the system while ensuring accurate and stable output voltage.

Description

Technical Field

[0001] This invention relates to the field of power electronics and electrochemical energy storage control technology, specifically to a method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply. Background Technology

[0002] Semi-solid-state lithium-ion batteries, with their high energy density, high safety, and wide temperature range adaptability, are gradually replacing traditional lead-acid batteries and liquid lithium batteries, becoming the core energy storage device for uninterruptible power supply (UPS) systems in data centers and critical infrastructure. In megawatt-level UPS applications, to meet the high voltage requirements of the DC bus, energy storage systems typically consist of hundreds of individual battery cells connected in series to form a high-voltage battery cluster.

[0003] In traditional fixed series topology, the overall performance of a battery cluster is limited by the worst-performing individual cell, exhibiting a "weakest link" effect. With increasing usage time and charge / discharge cycles, the dispersion of parameters such as internal resistance, capacity retention, and self-discharge rate among individual cells inevitably increases due to manufacturing process tolerances and environmental temperature gradients. When individual cells in the series circuit experience increased internal resistance or capacity decay due to aging, they will reach the cutoff voltage first during discharge, forcing the battery management system to disconnect the main circuit of the entire battery cluster to prevent over-discharge. At this point, most of the remaining cells still retain a significant amount of usable charge, resulting in low overall system capacity utilization. Furthermore, if an individual cell experiences an internal short circuit or thermal runaway risk, traditional hard-connected structures cannot physically isolate it in time, potentially triggering a chain reaction and reducing the overall safety of the system.

[0004] Existing battery balancing technologies are mainly divided into passive balancing and active balancing. Passive balancing dissipates energy through resistors, resulting in low efficiency and high heat generation. While active balancing achieves energy transfer, its circuit topology is complex and costly, and the balancing current is typically small, making it difficult to meet the real-time differential voltage adjustment requirements under high-current discharge conditions in UPS systems. Although modular multilevel converters (MMCs) or cascaded H-bridge structures that have emerged in recent years possess a certain degree of fault tolerance, their control dimensions are mostly focused on the synthesis of output voltage waveforms, lacking in-depth optimization for the electrochemical characteristics of semi-solid-state batteries.

[0005] In particular, existing control strategies often employ fixed-cycle switching or simple voltage threshold determination, failing to fully consider the impact of battery state of health (SOH) and state of charge (SOC) on marginal lifetime cost, and lacking a dynamic response mechanism to load power fluctuations. During periods of severe load fluctuation, fixed-cycle scheduling may lead to response lag; while during periods of stable load, unnecessary frequent switching introduces additional switching losses. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply. This method solves the problem that the overall lifespan of a traditional series-connected battery pack is limited by the worst-performing cell due to the "weakest link" effect caused by inconsistent aging of individual cells, and the inability to flexibly isolate faults online.

[0007] To achieve the above objectives, the present invention provides a method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply, the method comprising:

[0008] Data Acquisition and Status Analysis Phase: Real-time acquisition of terminal voltage and temperature data for individual battery cells, as well as load power data at the inverter output. Based on the acquired terminal voltage and temperature data, the current state of charge and health status of each battery cell are analyzed, providing a data foundation for subsequent hierarchical scheduling.

[0009] Dynamic Time Window Construction Phase: Fluctuation characteristics of the collected load power data are analyzed, and the control cycle duration for the current rotational discharge process is dynamically calculated based on the analysis results. The specific dynamic construction logic follows the inverse correlation mapping principle: the standard deviation of the load power sample points is calculated within a preset sampling period to characterize the load volatility; when the load volatility is higher than a preset high threshold, the control cycle duration is set to a preset minimum value to respond to high-frequency load fluctuations; when the load volatility is lower than a preset low threshold, the control cycle duration is set to a preset maximum value to maintain discharge topology stability; when the load volatility is between the high and low thresholds, the control cycle duration decreases linearly with the increase of load volatility.

[0010] Priority interval mapping stage: Based on the analyzed state of charge (SOC) and state of health (SOC) values, individual battery cells are mapped to a preset two-dimensional state interval table, thereby determining the discharge priority category of each battery cell. The discharge priority categories include main discharge, secondary discharge, standby discharge, and endangered state. The main discharge category corresponds to battery cells with both high SOC and SOC; the endangered category corresponds to battery cells with both low SOC and SOC; and the secondary discharge and standby discharge categories represent transitional states between these two categories.

[0011] Marginal life cost quantification stage: The determined duration of the control cycle is used as the prediction time benchmark. A life decay model is used to estimate the expected physical losses of individual battery cells within this duration. These expected physical losses encompass cycle aging losses based on the predicted depth of discharge, calendar aging losses based on the time dimension, and accelerated aging losses based on temperature-induced thermal effects. Based on this, a quantified marginal life cost index is calculated using health status values. The calculation logic for this index is as follows: a ratio relationship is established, with the depreciation amount per discharge as the numerator and the product of the expected total released energy and the future health status value at the end of the control cycle as the denominator. This calculation logic ensures that, under the same physical losses and total energy output, battery cells with lower health status values ​​will be assigned a higher marginal life cost index, thus being ranked lower in subsequent sorting.

[0012] The two-dimensional sorting and truncation execution phase involves sorting individual battery cells in two dimensions to generate an ordered discharge sequence. The first sorting dimension is based on discharge priority category, with higher priority cells appearing at the top of the sequence. The second sorting dimension is based on marginal lifetime cost; when priority categories are the same, cells with lower cost values ​​appear at the top of the sequence. After generating the sequence, the system truncates the target workgroup based on the inverter's voltage requirements to the DC bus. Specifically, following the ordered discharge sequence, the real-time terminal voltage of each battery cell is accumulated sequentially from front to back. Accumulation stops when the accumulated voltage reaches or exceeds the DC bus reference voltage, thus determining the sequence cutoff point.

[0013] Power switching execution phase: Battery cells at or before the cutoff position are assigned to the target working group, and an access command is sent; battery cells after the cutoff position are assigned to the bypass group, and a bypass command is sent. This command is executed by the power switching execution unit connected to each battery cell. The power switching execution unit includes a first power switch connected in series in the main discharge path and a second power switch connected in parallel in the bypass shunt path. During state switching, a dead-time control strategy is adopted, ensuring that one switch is completely turned off and a preset delay time is elapsed before the other switch is turned on to prevent short circuits. Furthermore, the system monitors time-triggered, undervoltage-triggered, and over-temperature-triggered conditions in real time during discharge; once these conditions are met, the current process is terminated and a new control cycle is initiated.

[0014] This invention provides a method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply. It offers the following advantages:

[0015] 1. This invention establishes a two-dimensional priority interval based on health status and state of charge, and introduces a marginal lifetime cost index for two-dimensional sorting, enabling intelligent identification and protection of battery cells in poor condition. This strategy breaks the limitations of the traditional battery pack's bottleneck effect, allowing cells with high health to undertake more discharge tasks, while cells with low health are allowed to rest, thereby maximizing the total discharge volume of the semi-solid-state battery pack throughout its entire life cycle and reducing long-term operation and maintenance costs.

[0016] 2. This invention employs a dynamic time window construction technique based on load power fluctuation characteristics. When the load fluctuates drastically, the system automatically shortens the control cycle to quickly respond to voltage demands; when the load is stable, the system extends the control cycle to reduce switching losses and heat generation. This adaptive mechanism effectively solves the contradiction of low efficiency caused by response lag or frequent switching in traditional fixed-cycle switching control, thus improving the dynamic performance of the UPS power supply.

[0017] 3. This invention utilizes a voltage accumulation and cutoff strategy combined with the bypass isolation mechanism of the power switching execution unit to flexibly select the optimal battery combination for circuit connection based on real-time DC bus voltage requirements. This approach not only achieves precise output voltage regulation but also possesses strong fault tolerance: once a battery cell enters a critical state or fails, the system can immediately bypass and isolate it without affecting the normal power supply of the remaining battery cells, ensuring uninterrupted and highly reliable operation of the UPS system. Attached Figure Description

[0018] Figure 1 This is the main flowchart of the battery cell rotation discharge control method for a semi-solid UPS power supply provided in an embodiment of the present invention;

[0019] Figure 2 A schematic diagram illustrating the principle of the cascaded sorting and voltage accumulation cutoff strategy provided in an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of the circuit topology of the power switching execution unit provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions in 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] See attached document Figure 1The control process described in this embodiment uses a centralized cluster control module as the main logic operator, which coordinates with distributed unit control modules for execution. The control logic is a periodic closed-loop process, comprising four consecutive stages: state awareness, parameter calculation, decision sequencing, and execution scheduling.

[0023] During the state-aware phase, data acquisition circuits distributed at each battery cell acquire real-time data on cell terminal voltage and surface temperature, and calculate the current state of charge and health status by integrating the current. Simultaneously, monitoring devices located at the inverter output side collect load power data. All acquired analog quantities are converted into digital signals and aggregated via a communication bus to a centralized cluster control module, forming a comprehensive system state dataset.

[0024] Upon entering the parameter calculation phase, the logic processing unit first processes the load power data. By calculating the standard deviation or volatility of the load power within a preset time period, the stability of the current load condition is quantified. Based on the load stability index, the duration of the current control cycle is dynamically defined in the time domain. For cases with high load volatility, a shorter control cycle is set to match power changes, ensuring that the release of stored energy can respond promptly to load impacts; for cases with low load volatility, a longer control cycle is set to reduce the frequency of switching actions, minimizing mechanical contact wear and switching losses.

[0025] After defining the time parameters, the control logic performs a tiered screening of individual battery cells. Based on the health status and state of charge (SOC) values ​​of each cell, it maps them to a pre-defined two-dimensional state interval table. The state interval table categorizes cells into different discharge priority classes, including primary discharge, secondary discharge, standby discharge, and critical discharge. Cells falling into the critical discharge class or below the safety threshold are logically marked as unusable, excluded from subsequent discharge sequences, and placed into maintenance or restoring recovery mode.

[0026] Subsequently, for battery cells classified into the same discharge priority category, a refined ranking based on marginal lifetime cost is performed. The main logic operation calls the lifetime degradation prediction model, inputting the current depth of discharge, temperature, and time parameters, to estimate the expected physical loss of the cell in the next control cycle. Then, the expected physical loss is compared with the cell's current health state to calculate a quantified marginal lifetime cost index. This calculation process ensures that, under the same loss conditions, cells with poorer health or higher temperatures receive higher cost values, thus placing them in a lower position in the ranking algorithm.

[0027] During the decision-making and execution scheduling phase, the system generates the final sequence of switching instructions. The initial sorting is based on discharge priority categories, prioritizing main discharge cells. Within each category, a secondary sort is performed based on marginal lifetime cost, prioritizing cells with lower cost values. Based on the total voltage or power required by the current load, several cells at the top of the sorted sequence are selected as working groups. The centralized cluster control module sends a closing instruction to the corresponding distributed unit control module, connecting it to the main circuit for discharge; and sends bypass instructions to the remaining cells, isolating them.

[0028] The system maintains the above discharge topology. During operation, if the control cycle time is exhausted or the state value of any working cell falls below the preset safety trigger threshold, the control logic will immediately trigger a new round of state perception and decision-making process, thereby achieving dynamic tracking of the battery pack state and load conditions.

[0029] This embodiment discloses a semi-solid-state uninterruptible power supply (UPS) system architecture that supports independent control at the unit level. At the physical layer, the system adopts a two-level control topology combining centralized decision-making and distributed execution, and mainly consists of a semi-solid-state battery pack, a distributed unit control module, a centralized cluster control module, and load-side monitoring devices.

[0030] A semi-solid-state battery pack consists of multiple semi-solid-state lithium-ion battery cells connected in series via a power bus. Each semi-solid-state battery cell is an independent physical replacement unit, containing a positive electrode, a negative electrode, and a semi-solid electrolyte with a specific viscosity. Temperature sensing probes are attached to the tabs or surfaces of the battery cells to capture thermal reaction data during charging and discharging. This data provides a physical basis for subsequent calculations of temperature-accelerated aging losses.

[0031] The distributed unit control module is configured in a one-to-one correspondence with the semi-solid-state battery cell. Each unit control module is physically fixed to the corresponding battery cell or integrated into the battery module's bracket. The hardware circuitry of the unit control module includes the following functional units:

[0032] High-precision data acquisition front end: includes a differential voltage amplifier circuit and an analog-to-digital converter, connected to the positive and negative terminals of the battery cells for real-time acquisition of the cell terminal voltage; includes a temperature signal conditioning circuit, connected to the aforementioned temperature sensing probe, for real-time acquisition of the cell operating temperature.

[0033] Power switching execution unit: A switching matrix composed of power semiconductor devices (such as MOSFETs or IGBTs) or high-power relays. This switching matrix is ​​connected in series in the DC main circuit of the UPS and has two circuit states: connected and bypassed. In the connected state, the current path of the battery cell is open, participating in external discharge; in the bypass state, the battery cell is short-circuited through a parallel branch, allowing the main circuit current to bypass the cell, thereby physically isolating the specific cell from the discharge sequence.

[0034] The underlying communication interface is equipped with an isolated CAN bus transceiver or RS485 transceiver, which is used to package and send the collected voltage and temperature data, and to receive the switching control commands sent from the upper layer.

[0035] The centralized cluster control module serves as the core of the system's computation and scheduling, establishing electrical connections with all distributed unit control modules via an industrial fieldbus. The module's hardware platform is built on an embedded microprocessor (such as a DSP or ARM architecture processor) and equipped with non-volatile memory. The memory contains a pre-built SOH-SOC state priority mapping table and a semi-solid-state battery life degradation prediction model. The cluster control module periodically polls each unit control module via the bus to obtain the real-time state vectors of all battery cells and executes sorting algorithms and sequence generation logic.

[0036] The load-side monitoring device is physically installed on the output circuit of the UPS inverter or on the downstream power distribution bus. This device mainly includes current transformers, voltage transformers, and a high-frequency power metering chip. Its sampling frequency is set to meet the standard for capturing instantaneous load fluctuations (e.g., a power frequency multiple higher than 50Hz) to calculate the effective value and instantaneous rate of change of load power in real time. The load-side monitoring device transmits real-time load power data to the centralized cluster control module via an independent communication line or shared bus, providing input variables for the dynamic time window algorithm and establishing a coupling relationship between load conditions and battery replacement cycles.

[0037] In terms of electrical connections, the power output terminals of each semi-solid-state battery cell are connected in series via copper busbars or cables to form a high-voltage DC input, which is then connected to the DC side of the UPS inverter. The communication ports of each distributed unit control module are connected to the main communication port of the centralized cluster control module via twisted-pair cables in a daisy-chain or star topology. The signal output terminal of the load-side monitoring device is connected to the analog input terminal or digital communication terminal of the centralized cluster control module. The entire system forms a closed-loop control physical environment that includes source-side status perception, load-side operating condition monitoring, central strategy calculation, and end-point power execution.

[0038] During the operation of a semi-solid-state UPS power system, the dynamic characteristics of the load conditions directly affect the effectiveness of the battery swapping strategy. To quantify the stability of the load conditions, the system introduces a load volatility index. The calculation of the load volatility index is performed by the digital signal processing unit in the centralized cluster control module.

[0039] Define a discrete time sampling sequence with a sampling period of 1. The system continuously collects data within each sampling period. One load power sample point.

[0040] First, based on the collected data For each load power sample point, calculate the arithmetic mean load power over the sampling period of that sample point. The calculation formula is as follows:

[0041] ;

[0042] in, This represents the average load power within the current sampling period; This indicates the total number of sampling points included within the sampling period; Indicates the first The instantaneous values ​​of load power collected at each sampling time, where Values ​​range from 1 to Integers.

[0043] Based on the obtained average load power, the standard deviation of the load power is further calculated as the load volatility. The quantitative representation of . The calculation formula is as follows:

[0044] ;

[0045] in, It represents the load volatility, and its physical dimensions are the same as power. Its value reflects the degree of dispersion of the load power from the average value. , and The definition is consistent with the one mentioned above.

[0046] This calculation process occurs in each sampling period. Triggered once at the end, and the calculated result The value is stored in the register of the control module, serving as an input variable for subsequently determining the length of the dynamic time window. This is achieved through real-time updates. The system can capture the time-varying characteristics of the load at the UPS output end and distinguish between stable load conditions and impulsive load conditions.

[0047] In obtaining load volatility Then, the system dynamically adjusts the current control cycle, i.e., the time window, based on this indicator. This adjustment mechanism establishes a piecewise linear mapping function that converts load fluctuation characteristics into time parameters of the control system to match response requirements under different operating conditions.

[0048] Set two critical thresholds for load volatility: low volatility threshold and high volatility threshold These two thresholds are preset and stored in the control module based on the UPS system's rated capacity and load type, and are used to define the stable range, transition range, and drastic fluctuation range of the load condition. Simultaneously, boundary values ​​for the control cycle are set, namely the minimum control cycle. and maximum control cycle .

[0049] Based on the above parameters, the time window The calculation formula is defined as follows:

[0050] ;

[0051] Adjustment coefficient in the formula The linear slope used to determine the change in volatility within the transition interval of the time window is calculated as follows:

[0052] ;

[0053] in, Indicates the duration of the current rotation control cycle, in minutes or seconds; This represents the current load volatility calculated from the preceding steps; This indicates the high threshold of load volatility. When the volatility exceeds the high threshold, it is determined to be a severe volatility condition. This represents the lower threshold of load volatility. When the volatility is below the upper threshold, it is considered a stable operating condition. This represents the lower limit of the control cycle, corresponding to the rapid response mode under drastic fluctuation conditions. This indicates the upper limit of the control cycle, corresponding to the low-frequency switching mode under stable operating conditions.

[0054] Under this mechanism, when Greater than or equal to At that time, the control cycle is locked at the lower limit value. This ensures the system refreshes the discharge sequence at the highest frequency to adapt to rapid load changes; when Less than or equal to At that time, the control cycle is locked at the upper limit value. Maintain the current discharge topology and reduce the number of switching device operations; when In and In between, control cycle Follow The value decreases linearly with the increase of [the value], achieving a smooth transition between response speed and switching frequency. The calculated [response rate]... The value is loaded into the system's timer as a countdown reference for the current round of discharge.

[0055] The multidimensional state evaluation system constructed in the primary screening strategy based on state intervals is described in detail.

[0056] In the control logic of a semi-solid-state UPS power system, the primary screening strategy aims to establish the safety boundaries and basic availability of individual battery cells. To this end, the system constructs a health status (…) ) and state of charge ( A two-dimensional evaluation system is constructed. This two-dimensional evaluation system abandons the traditional UPS evaluation method that only considers individual unit voltage or a single... Instead of using the logic of switching indicators, it introduces a system that reflects the degradation characteristics throughout the battery's entire lifespan. As the primary weighted dimension, the indicator will reflect the energy reserves of the current cycle. The metric is used as the second weighting dimension.

[0057] Definition of the first The health status of each battery cell is It represents the retention rate of the maximum storable capacity of a single battery cell at the current moment relative to its factory rated capacity, and also implicitly represents the growth of the battery's internal resistance and the stability of the semi-solid electrolyte interface. The values ​​are obtained through an estimation algorithm within the unit control module. This algorithm is based on the voltage-current response characteristics during the discharge process or the cumulative capacity method. In the evaluation system, Primarily used to define the physical safety level and long-term service value of individual units, low This value corresponds to high internal resistance and high thermal runaway risk, and therefore has the highest veto power (circuit breaker) in logical judgment.

[0058] Definition of the first The state of charge of each battery cell is It represents the percentage of the current remaining active charge within a single battery cell relative to its current maximum usable capacity. The value is calculated in real time by the unit control module using the ampere-hour integration method combined with the open-circuit voltage correction method. In the evaluation system, This is used to measure the continuous power supply capability of a single unit within the current control window. The centralized cluster control module collects data from each single unit. and The data is mapped to a two-dimensional orthogonal coordinate system to form state coordinate points. The core technical feature of the multidimensional state evaluation system lies in the establishment of a nonlinear state evaluation logic: the system not only focuses on whether a single unit has electricity ( More attention is paid to whether the individual cells are healthy. Through this multidimensional evaluation, the system can physically distinguish between healthy but low-charge cells and high-charge but aged cells, providing basic data support for subsequent priority interval division and life cost calculation, and ensuring that the scheduling strategy is based on the actual physical state of the battery.

[0059] Based on the construction of a multi-dimensional state evaluation system, the system uses a lookup table method to map all battery cells to four discrete discharge priority levels according to preset logical thresholds. ( ) and a maintenance mode range.

[0060] The three threshold values ​​for the health status dimension are set as follows: , , And satisfy ; Set two threshold values ​​for the state of charge dimension as follows , And satisfy Specifically, in this embodiment, the values ​​of each threshold parameter are: .

[0061] The system iterates through the state vectors of all individual units. The priority of its affiliation is determined according to the following set of logical rules. :

[0062] Priority 1 (Main Zone):

[0063] When the monomer simultaneously satisfies and At that time, the judgment =1. Individual units in this range possess excellent health and sufficient power, and are marked by the system as core power sources for priority use.

[0064] Priority 2 (second-best zone):

[0065] The system implements an equivalent substitution strategy at this level, defining the following two states as having equal priority for judgment. =2:

[0066] Status A: (High health) and (Medium power);

[0067] State B: (China Health) and (High charge). This logic indicates that the system considers healthy low-charge cells and slightly aged fully charged cells to have the same discharge utility value and are given equal priority in the scheduling sequence.

[0068] Priority 3 (Spare Zone):

[0069] Similarly, the system categorizes the following two states into the third priority for judgment. =3:

[0070] State C: (China Health) and (Medium power);

[0071] State D: (Low health) and (High battery level)

[0072] Priority 4 (Endangered Area):

[0073] When the monomer satisfies and At that time, the judgment =4. The cells in this range are on the verge of aging and have moderate charge. The system places them at the end of the discharge sequence and only activates them when the load demand exceeds the total capacity of the first three levels.

[0074] Maintenance mode (circuit breaker mechanism):

[0075] If the monomer satisfies or If any of the conditions is met, the system immediately triggers the circuit breaker logic. The affected cell is forcibly removed from the discharge candidate pool, and its control commands are locked in bypass mode until it is manually replaced or recharged to above the threshold. This mechanism constitutes the system's absolute safety baseline. Through the above logical division, the system completes the preliminary qualitative classification of all battery cells, generating a primary sequence containing priority tags, providing a hierarchical basis for subsequent quantitative cost optimization.

[0076] This embodiment focuses on the core evaluation metric in the secondary optimization algorithm: marginal lifetime cost. The quantitative formula and its physical meaning of ) are explained in detail. This is done in obtaining semi-solid-state battery cells. After predicting the degradation value, the system needs to transform the physical-level lifespan loss into an economic indicator that can be used for ranking and comparison. To this end, the system constructs a marginal lifespan cost calculation model. This model not only calculates simple depreciation costs, but also introduces a health status weighting factor to construct a mathematical penalty mechanism for batteries with low health.

[0077] Based on the prediction obtained from the aforementioned steps Health status value at all times Calculate the first The marginal lifetime cost of each individual battery cell during this discharge process. The calculation formula is defined as follows:

[0078] ;

[0079] Among them, output indicators Indicates the first The economic cost of lifespan loss corresponding to a unit of effective energy output by a single battery cell within the current control cycle. The smaller the value of this indicator, the higher the discharge economy of that single cell;

[0080] Asset value parameters Indicates the first The total lifecycle replacement cost or initial purchase price of each battery cell. This parameter is stored in the system's non-volatile memory and serves as a benchmark constant for calculating the monetary value of lifespan depreciation.

[0081] Lifetime depreciation value item (Molecular part): This represents the physical lifetime decay of a single cell during this discharge process (including the sum of cycle, calendar, and temperature aging). Compare this with... Multiplying these amounts gives the depreciation amount of the asset resulting from this discharge.

[0082] Expected energy output (Denominator): indicates the first... Individual battery cells within the time window The total amount of electrical energy expected to be released within the cell. This value is calculated by integrating the current terminal voltage of the cell, the expected discharge current, and the time window length. Health penalty factor. (Denominator): This is the key mathematical construct for building the passive life management characteristics of this algorithm. The system directly uses the predicted health status value as the multiplier factor in the denominator.

[0083] This formula establishes the following nonlinear cost mapping relationship: in the numerator (i.e., the value of lifetime loss) and Under the same conditions of (energy output) The lower the value of a single battery cell, the lower its calculated value. The numerical value is actually larger. This inverse relationship mathematically creates a price lever for aging batteries. When the system performs an ascending sort (lower cost first) in subsequent steps, severely aged battery cells, due to their lower cost, are prioritized. Artificially amplified, they naturally settle to the end of the sequence. This mechanism requires no additional conditional logic to automatically achieve adaptive lifespan extension control by using older batteries less and healthier batteries more.

[0084] This embodiment elaborates on the dual sorting logic in the cascading sorting and dynamic execution strategy. After completing the priority division of battery cells and the calculation of marginal lifetime cost, the system establishes a system containing cell number and priority identifier. and marginal life cost The centralized cluster control module processes this dataset using a cascaded comparison algorithm to generate an ordered queue for guiding power switching. .

[0085] The cascaded comparison algorithm defines a two-dimensional sort key value. The first dimension, the sort key value, is a priority identifier. The first principle is that the smaller the value, the higher the priority; the second dimension, the sorting key value, is the marginal lifetime cost. The smaller the value, the earlier it appears in the sorting.

[0086] For any two battery cells and In an ordered queue The relative positions within are uniquely determined by the following set of logical rules:

[0087] Rule 1: Hierarchical Priority Logic

[0088] like Then, they are sorted only according to priority indicators. If Determine the monomer Superior to monomers ,monomer In an ordered queue The position in the middle is in the single unit Previously, this rule ensured that units in the primary discharge zone were always called upon before units in the secondary or backup zones, even if the latter were less costly. This layer of logic established the system's safety and functional boundaries, preventing safety hazards caused by cross-level scheduling.

[0089] Rule 2: Cost Optimization Logic

[0090] like Then they are arranged according to marginal life cost. If Determine the monomer Superior to monomers ,monomer Ranked in single Previously; if ,monomer Ranked in single Previously, this rule ensured that within the same security level, the system always prioritized calling the unit with the lowest current cost, achieving local economic optimization.

[0091] Based on the above rules, the system constructs the final discharge control sequence. The mathematical expression of this sequence is as follows:

[0092] ;

[0093] in, This represents the final, strictly ordered queue of candidate battery cells for discharge. Indicates priority identifier equal A subset of singletons, i.e. ,in ; Represents a set The individual elements within are arranged according to The ordered subsequence generated after the indicators are sorted in ascending order; This represents a sequence concatenation operator that directly joins the right subsequence to the end of the left subsequence to form a long sequence.

[0094] The generated discharge candidate queue is achieved through this dual sorting logic. The sequence exhibits a strict hierarchical structure: the head of the sequence contains high-quality monomers with high health and low loss cost; the middle of the sequence contains monomers with acceptable health but currently high loss under operating conditions; and the tail of the sequence contains monomers with low health or extremely high loss. This logic ensures that the control strategy strictly adheres to safety classification at the macro level and fully explores economic benefits at the micro level, avoiding safety risks or economic waste that may result from ranking based on a single indicator.

[0095] See attached document Figure 2 This embodiment elaborates on the dynamic scheduling and execution logic in the cascading sorting and dynamic execution strategy. It generates a strictly ordered discharge candidate queue. Afterward, the system enters the physical execution phase. The core task of this phase is to transform the logical sequence into physical switching matrix actions and establish a real-time circuit breaker mechanism to ensure that the system dynamically follows environmental changes.

[0096] The centralized cluster control module reads the reference voltage requirement of the current inverter DC bus. Based on the principle of series superposition, the system follows... The battery cells in the queue are sorted sequentially from front to back, and their real-time terminal voltages are accumulated one by one until the accumulated value meets the reference voltage requirement. This determines the cutoff point for the battery cells that need to be connected to the main circuit. Cut-off position The calculation formula is as follows:

[0097] ;

[0098] in, This indicates the total number of battery cells that need to be put into operation during this scheduling cycle, i.e., the cutoff index value; Indicates an ordered queue Middle row in Real-time terminal voltage of a single battery cell; This indicates the lower limit of the DC bus voltage required for the UPS inverter to operate normally. Represents the set of positive integers.

[0099] Determine the deadline Then, the system generates a global switch instruction vector. For an ordered queue Mid-ranking index For each battery cell, the corresponding instruction status is set to "accessed" (logic 1); for the ranking index... For each individual battery cell, the corresponding command status is set to bypass (logic 0). The cluster control module unpacks the command vector and distributes it to each distributed unit control module via the fieldbus. The unit control module that receives the command drives the power switching devices to operate, completing the physical circuit reconfiguration.

[0100] During the discharge process, the system starts a parallel monitoring thread to calculate the cyclic reset trigger conditions in real time. .once When the value becomes True, the system immediately terminates the current scheduling cycle, ignores the remaining time window countdown, and jumps directly back to the state awareness phase to start a new round of calculation. This interruption mechanism ensures the real-time security of the control strategy.

[0101] Reset trigger conditions in a loop Defined as the result of the logical OR operation of time-triggered, safety-triggered, and load-triggered events, the calculation formula is as follows:

[0102] ;

[0103] in, This indicates a Boolean reset trigger signal; a value of 1 triggers a system recalculation. This indicates the current time recorded by the system's internal clock. Indicates the end time of the current dynamic time window, by Sure; This represents the set of battery cells currently in a state of discharge. Represents a set The Middle Real-time terminal voltage of each individual cell; This indicates the preset individual unit undervoltage protection threshold; Represents a set The real-time temperature of the i-th monomer; This indicates the preset over-temperature protection threshold for a single unit.

[0104] Through the above logic, the system implements a dynamic execution strategy combining quota throttling and conditional circuit breaking. Quota throttling ensures the highest quality... Each battery carries the load, achieving economical implementation; conditional fuses ensure that the system can respond in milliseconds and switch to a new optimal topology when any single cell experiences a voltage drop or abnormal temperature rise, or when the time window expires, thereby maintaining the continuity and safety of UPS power supply.

[0105] See attached document Figure 3 This embodiment details the internal hardware circuit structure of the power switching execution unit installed on each semi-solid-state battery cell. This circuit topology aims to physically achieve seamless serial insertion and isolation of battery cells, and provides hardware-level short-circuit protection. This circuit topology includes a first power switch transistor. Second power switching transistor Hardware interlock logic circuit and freewheeling diode.

[0106] The positive terminal of the semi-solid-state battery cell is connected to the positive power output terminal of the module. The negative terminal of the semi-solid-state battery cell is connected to the first power switch. The drain of the first power switch. The source is connected to the negative power output terminal of the module. This branch forms the main discharge path.

[0107] Second power switch The drain is connected to the positive power output terminal of the module. Its source is connected to the negative power output terminal of the module. This branch road forms a bypass and diversion route.

[0108] First power switch transistor With the second power switch All devices are enhanced N-channel power MOSFETs or IGBTs. A fast recovery diode (FRD) is connected in reverse parallel between the drain and source of each switch. and This diode is used to provide a freewheeling path for inductive current during the dead time of switching operation, preventing voltage spikes on the DC bus from damaging the switching device.

[0109] To prevent and Simultaneous conduction causes a direct short circuit in a single battery cell, and the execution unit is equipped with a hardware interlock logic circuit. This circuit consists of NOT gates and delay buffers. The raw control signals come from the microprocessor of the distributed unit control module. Enter directly to The gate drive circuit; simultaneously, the original control signal After being inverted by a logic NOT gate, the input is... The gate drive circuit.

[0110] The operating logic of this circuit is as follows:

[0111] Access status: When the original control signal is input When it is low (logic 0), Cut-off. After inversion by the NOT gate, The gate receives a high-level signal. The circuit is on. At this point, the current path passes through the following points in sequence: The battery's negative terminal, the battery's positive terminal, and finally... The individual battery cells are connected in series to the main circuit.

[0112] Bypass state: When the original control signal is input When it is high (logic 1), The circuit is turned on. After being inverted by the NOT gate, the circuit is activated. Cut off. At this point, the main circuit current path passes through the following sequentially: , It is best to arrive Current is transmitted directly, bypassing individual battery cells, thus achieving physical isolation.

[0113] In hardware interlocked logic circuits, asymmetric delay circuits are connected in series on the signal paths leading to the gates of the two switching transistors. Opening delay time Greater than shutdown delay time And set Opening delay time Greater than shutdown delay time This time difference setting ensures that during state switching, the previously conducting switch is disconnected first, and after a preset microsecond dead time, the other switch is closed, thereby eliminating the risk of thermal runaway due to a short circuit in a single battery cell at the physical level.

Claims

1. A method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply, characterized in that, include: Real-time acquisition of terminal voltage data, temperature data, and load power data at the inverter output terminal of each battery cell; and parsing the state of charge and health status values ​​of each battery cell based on the terminal voltage data and temperature data. The collected load power data is subjected to fluctuation characteristic analysis to obtain fluctuation characteristic analysis results. The duration of the control cycle of the current rotation discharge process is dynamically calculated based on the fluctuation characteristic analysis results. Based on the parsed state of charge and health values, individual battery cells are mapped to a preset two-dimensional state interval table to determine the discharge priority category of each battery cell. The specific logic is as follows: A battery cell whose health status value is higher than the first health threshold and whose state of charge value is higher than the first charge threshold is classified as a main discharge type. If the health status value of a battery cell is higher than the first health threshold but the state of charge value is between the first energy threshold and the second energy threshold, or if the health status value of a battery cell is between the first health threshold and the second health threshold and the state of charge value is higher than the first energy threshold, it is classified as a secondary discharge. A battery cell whose health status value is between the first health threshold and the second health threshold and whose state of charge value is between the first energy threshold and the second energy threshold, or whose health status value is between the second health threshold and the third health threshold and whose state of charge value is higher than the first energy threshold, is classified as a standby discharge cell. A battery cell whose health status value is between the second health threshold and the third health threshold, and whose state of charge value is between the first charge threshold and the second charge threshold, is classified as endangered. The calculated duration of the control cycle is used as the prediction time benchmark. The expected physical loss of a single battery cell during the duration is estimated using the life decay model. The quantitative marginal life cost index is calculated by combining the health status value. The battery cells are sorted according to the first dimension based on the discharge priority category and the second dimension based on the marginal lifetime cost index to generate an ordered discharge sequence. According to the load voltage requirement, a target working group is selected from the ordered discharge sequence, and the battery cells in the target working group are controlled to be connected to the main circuit for discharge, while the remaining battery cells are bypassed and isolated.

2. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The step of dynamically calculating the duration of the control cycle for the current rotational discharge process based on the fluctuation characteristic analysis results includes: Multiple load power sample points are collected within a preset sampling period, and the standard deviation of the multiple load power sample points is calculated as the load volatility. Establish an inverse correlation between load volatility and duration: When the load volatility exceeds the preset high threshold, the duration will be set to the preset minimum value to respond to high-frequency load fluctuations. When the load volatility is lower than the preset low threshold, the duration is set to the preset maximum value to maintain the stability of the discharge topology; When the load volatility is between the high threshold and the low threshold, the duration decreases linearly as the load volatility increases.

3. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The estimation of the expected physical loss of a single battery cell during the specified duration using the lifetime degradation model involves calculating the sum of the following three degradation components: The loss of electrochemical active material is calculated using a power function model based on the expected depth of discharge of the battery cell during the duration of the control cycle. The duration of the control cycle is used as a time variable, and the static capacity loss is calculated using the exponential decay model in combination with the current state of charge of the battery cell. The additional losses due to thermal effects are calculated using the Arrhenius equation, which includes the reaction activation energy parameter, by combining real-time temperature data of individual battery cells.

4. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The steps for calculating the quantified marginal life cost index include: Obtain the full lifecycle asset value parameters of a single battery cell; The depreciation amount per discharge is obtained by multiplying the expected physical loss of a single battery cell during the duration of the control cycle by the asset value parameter. Calculate the total electrical energy expected to be released by a single battery cell during the duration of the control cycle; Predict the future health status of individual cells at the end of the control cycle based on expected physical losses. A ratio formula is constructed, with the depreciation amount per discharge as the numerator and the product of the expected total amount of electrical energy released and the future health status value as the denominator, to calculate the marginal life cost index. The future health status value is inversely proportional to the marginal lifetime cost index, so that under the same physical loss and total electrical energy conditions, the battery cell with a lower health status value can be calculated to have a higher marginal lifetime cost index.

5. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The generation of the ordered discharge sequence includes: First, compare the discharge priority categories of individual battery cells and rank the battery cells with higher priority categories at the top of the sequence. Among them, the main discharge category is higher than the secondary discharge category, the secondary discharge category is higher than the backup discharge category, and the backup discharge category is higher than the endangered category. If the discharge priority categories are the same, the marginal lifetime cost index of the individual battery cells is compared, and the battery cells with the smaller marginal lifetime cost index value are ranked at the top of the sequence.

6. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The step of selecting a target work group from the ordered discharge sequence based on the load voltage requirement and controlling the individual battery cells within the target work group to connect to the main circuit for discharge includes: Obtain the DC bus reference voltage required for the inverter to operate normally; According to the order of the ordered discharge sequence, the real-time terminal voltage of each battery cell is accumulated from front to back. When the accumulated voltage value reaches or exceeds the DC bus reference voltage, the accumulation stops, and the cutoff position of the currently accumulated battery cell in the ordered discharge sequence is marked. The battery cells at or before the cutoff position in the ordered discharge sequence are assigned to the target working group, and an access command is sent. The battery cells that are after the cutoff position in the ordered discharge sequence are assigned to the bypass group, and a bypass command is sent.

7. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The control of individual battery cells within the target working group being connected to the main circuit for discharge also includes real-time monitoring of the following triggering conditions: The execution time of the current discharge process reaches the calculated duration of the control cycle; The terminal voltage of any battery cell in the discharge state is lower than the undervoltage protection threshold. The temperature of any single battery cell in the discharge state exceeds the over-temperature protection threshold. When any of the above triggering conditions are detected, the current discharge process is immediately terminated and the process returns to the steps of real-time acquisition of the terminal voltage data, temperature data and load power data of the inverter output, and a new round of control process is started.

8. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 1, characterized in that, The step of connecting individual battery cells in the control target working group to the main circuit for discharge and bypassing and isolating the remaining individual battery cells is performed based on the power switching execution unit connected to each individual battery cell. The power switching execution unit includes a first power switch connected in series in the main discharge path and a second power switch connected in parallel in the bypass shunt path. When the access command is executed, the second power switch is turned off and the first power switch is turned on, so that current flows through the battery cell. When the bypass command is executed, the first power switch is turned off and the second power switch is turned on, so that the current bypasses the battery cell.

9. The method for controlling the alternating discharge of individual battery cells in a semi-solid-state UPS power supply according to claim 8, characterized in that, When controlling the state switching between the first power switch and the second power switch, a dead-time control strategy is executed: When switching from the access state to the bypass state, the first power switch is first turned off, and after a first preset delay time, the second power switch is turned on. When switching from bypass state to access state, the second power switch is first turned off, and after a second preset delay time, the first power switch is turned on. The dead time control strategy is used to prevent the first power switch and the second power switch from being turned on at the same time, which would cause a short circuit in a single battery cell.

Images