A direct current-alternating current direct conversion system based on a dynamic reconfigurable battery network and a control method thereof

Abstract

The application discloses a direct current- alternating current direct conversion system based on a dynamic reconfigurable battery network and a control method thereof, and belongs to the technical field of new energy, power systems and energy storage. The system comprises a battery matrix, a high-speed switch array and a digital energy exchange system. By dynamically scheduling battery monomers or modules, an alternating current voltage waveform is directly synthesized, a DC-AC inversion link in a traditional energy storage converter is not needed, and one-step efficient conversion from battery energy to alternating current energy is realized. The DRBN direct AC conversion technology provided by the application regards battery monomers / modules as dynamically modulated "energy slices", directly synthesizes an alternating current voltage waveform meeting grid requirements through dynamic reconfiguration control of the high-speed switch array, fundamentally saves the DC-AC inversion link of the traditional PCS, reduces the power conversion stages from two stages (DC-DC+DC-AC) to one stage, significantly improves the system efficiency, and reduces the hardware cost.

Description

Technical Field

[0001] This invention relates to the fields of new energy, power systems and energy storage technologies, and in particular to a DC-AC direct conversion system based on a dynamically reconfigurable battery network and its control method. Background Technology

[0002] With the deepening of the global energy transition, energy storage technology is playing an increasingly important role in new power systems. Electrochemical energy storage, as a major technological route for new energy storage, has two key components in its core technology equipment: battery systems and power storage converters (PCS).

[0003] In traditional energy storage system architecture, battery clusters are connected in a fixed series-parallel manner to form a high-voltage DC bus. The energy storage converter (PCS), as the core device connecting the battery and the grid, undertakes the key function of DC-AC power conversion. The current mainstream PCS topologies include: two-level voltage source inverter, three-level NPC (Neutral Point Clamped) inverter, and modular multilevel converter (MMC). Regardless of the topology used, traditional PCS must complete the following functions: (1) DC-AC inversion, converting DC power to AC power; (2) filtering, eliminating switching harmonics; (3) grid synchronization, realizing phase and frequency synchronization with the grid voltage.

[0004] However, traditional PCS architecture has the following technical pain points: First, traditional power conversion systems (PCS) require two power conversion stages: DC-DC boost and DC-AC inversion. Each stage incurs conduction losses, switching losses, and magnetic losses. A typical two-level PCS has a conversion efficiency of approximately 96% to 97% under rated operating conditions, which further decreases under partial load conditions. For large-scale energy storage power stations, even a 1% efficiency improvement translates to hundreds of thousands of yuan in additional revenue annually.

[0005] Second, hardware costs are high. Traditional PCS requires a large number of power switching devices (IGBT / MOSFET), drive circuits, DC bus capacitors, filter inductors, and other components. Taking a 1MW PCS as an example, the cost of its power module accounts for approximately 40% to 50% of the total cost. In addition, to ensure reliability, complex protection circuits and heat dissipation systems are also required, further increasing costs.

[0006] Third, the control algorithm is complex. Traditional PCS requires the implementation of multiple control loops, including SPWM / SVPWM (Sinusoidal PulseWidth Modulation, Space Vector Pulse Width Modulation) modulation, phase-locked loop (PLL), current loop control, and DC bus voltage control. This results in a complex control algorithm and high performance requirements for the controller. Furthermore, it needs to handle complex functions such as grid-connected protection, low-voltage ride-through, and anti-islanding, further increasing the control difficulty.

[0007] In recent years, the rise of Dynamically Reconfigurable Battery Network (DRBN) technology has provided a new approach to solving the aforementioned problems. DRBN dynamically adjusts the series and parallel topology of battery cells / modules through high-speed power electronic switch arrays, realizing digital scheduling of battery energy. However, existing DRBN technologies mainly focus on DC-side energy management, such as battery equalization, fault isolation, and capacity optimization, and have not yet addressed technical solutions for directly utilizing DRBN to synthesize AC voltage waveforms.

[0008] Therefore, there is an urgent need for a new system architecture that can extend the DRBN concept to the AC side and realize a one-step conversion from battery energy to AC power, so as to fundamentally solve the efficiency, cost and control complexity problems of traditional PCS. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention provides a DC-AC direct conversion system based on a dynamically reconfigurable battery network and its control method.

[0010] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a direct DC-AC conversion system based on a dynamically reconfigurable battery network, comprising: A battery matrix, which is composed of multiple battery cells connected in series and parallel, is used to provide electrical energy; A high-speed switching array is connected to each battery cell in the battery matrix, and is used to dynamically connect a single battery cell to the main circuit or bypass it according to control commands. The digital energy exchange system is connected to the high-speed switch array and is used to collect the battery cell status parameters in real time. Based on the target AC voltage waveform, it generates and issues control commands to dynamically reconstruct the topology connection of the high-speed switch array and control the access or bypass status of each battery cell in the series circuit in real time, so that the total voltage output by the battery matrix can be directly synthesized to approximate the AC voltage of the target AC voltage waveform. The AC output terminal is connected to the output terminal of the battery matrix and is used to directly output the AC voltage synthesized by the DC-AC direct conversion system to the power grid or load.

[0011] Furthermore, the battery matrix adopts... The three-dimensional topology is defined as follows: L represents the number of battery clusters, corresponding to the number of AC phases in the output; S represents the number of series layers in each battery cluster, used to determine the output voltage amplitude; P represents the number of parallel circuits in each layer, used to determine the output power; where, in In the topology, each battery cluster contains S layers connected in series, and each layer has P parallel battery cells.

[0012] Furthermore, the number of serial layers satisfy ,in The target AC voltage peak value, This is the minimum operating voltage for the battery cell.

[0013] Furthermore, each switching device in the high-speed switching array is a wide-bandgap semiconductor device, including SiC MOSFET or GaN HEMT, with a switching speed of , The rise time, i.e., the rising edge; Fall time, i.e., the falling edge; on-resistance. .

[0014] Furthermore, the digital energy exchange system adopts a distributed architecture, including: Digital energy network cards are used to collect voltage, current, and temperature data for individual battery cells; A digital energy co-controller is used to perform local switching timing generation and scheduling for a group of battery cells; Digital energy controllers are used to perform global system-level control, communication, and solve optimization algorithms.

[0015] Furthermore, it also includes an output filtering stage, which is an LC filter or an LCL filter with a cutoff frequency of The design is between 10 times the grid frequency and 0.1 times the switching frequency.

[0016] Furthermore, it also includes a grid-connected synchronization unit, which includes a phase-locked loop (PLL) for detecting the phase and frequency of the grid voltage and controlling the output voltage of the DC-AC direct converter system to achieve pre-synchronization of amplitude, phase, and frequency with the grid voltage before grid connection.

[0017] Secondly, the present invention also provides a DC-AC direct conversion method based on a dynamically reconfigurable battery network, applied to the aforementioned DC-AC direct conversion system based on a dynamically reconfigurable battery network, comprising the following steps: The digital energy exchange system acquires the target AC voltage waveform, which has preset amplitude, frequency and phase. The digital energy exchange system dynamically reconstructs the topology of the high-speed switching array based on the target AC voltage waveform, and controls the access or bypass status of each battery cell in the series circuit in real time, so that the total voltage output by the battery matrix can be directly synthesized to approximate the AC voltage waveform of the target AC voltage. The synthesized AC voltage is directly output to the grid or load without going through a separate DC-AC inverter.

[0018] Furthermore, the battery matrix adopts... The three-dimensional topology is defined as follows: L represents the number of battery clusters, corresponding to the number of AC phases in the output; S represents the number of series layers in each battery cluster, used to determine the output voltage amplitude; P represents the number of parallel circuits in each layer, used to determine the output power; where, in In the topology, each battery cluster contains S layers connected in series, and each layer has P parallel battery cells.

[0019] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the described DC-AC direct conversion method based on a dynamically reconfigurable battery network.

[0020] In summary, compared with the prior art, the beneficial effects of the above technical solution are: (1) Significantly improved system efficiency. This invention uses a DRBN architecture to directly connect battery cells in series to synthesize an AC voltage waveform, eliminating the DC-DC boost and DC-AC inverter stages in traditional PCS, reducing the power conversion stage from two stages to one.

[0021] (2) Significantly reduced hardware costs. Traditional PCS requires a large number of power switching devices (IGBT / MOSFET), drive circuits, DC bus capacitors, filter inductors, isolation transformers, and other components. This invention integrates the PCS function into the dynamic reconfiguration control of the battery network, eliminating most of the power devices and isolation transformers of traditional inverters. Preliminary estimates suggest that, at the same power level, the hardware cost of the DRBN system can be reduced by 20% to 30%. For a 1MW energy storage power station, the hardware investment can be reduced from approximately RMB 1.6 million to RMB 1.2 million, saving RMB 400,000. The reduction in hardware costs is based on a theoretical goal derived from reasonable logical deduction, rather than a conclusion supported by experimental data. Qualitative analysis: Which core hardware components are eliminated? The cost reduction mainly stems from the simplification of the system architecture: the elimination of the DC-AC inverter stage of the traditional PCS: the traditional PCS is the largest cost item in the energy storage system besides the battery cells. The elimination of the equalization circuit of the traditional BMS: DRBN achieves equalization through dynamic scheduling of the batteries themselves, eliminating the need for additional equalization hardware. Simplifying or eliminating the DC-DC converter: Traditional systems may require DC-DC boost, while DRBN achieves this directly through series-connected battery cells. Quantitatively: Based on industry data that PCS accounts for 25%-30% of the total system cost, replacing it with a lower-cost switch array can theoretically save 20%-30% of the system hardware cost, which is theoretically self-consistent.

[0022] (3) Significantly reduced control complexity. Traditional PCS requires the implementation of multiple control loops, such as SPWM / SVPWM modulation, phase-locked loop (PLL), current loop control, and DC bus voltage control. The control algorithm is complex and places high demands on the controller performance. This invention adopts a simplified control strategy based on energy slice modulation, transforming complex PWM modulation into switching control of the battery network topology. The switching state has only two states, 0 and 1, making the control logic clearer. At the same time, battery management and power conversion functions are unified into the DRBN architecture, achieving collaborative optimization through a single controller, reducing communication and coordination overhead between subsystems.

[0023] (4) Significantly extended battery life. Traditional battery clusters use a fixed series connection, and inconsistencies between individual cells can lead to a "weakest link" effect, accelerating battery aging. This invention utilizes the dynamic reconfiguration capability of DRBN to achieve active battery balancing while synthesizing AC waveforms. Since the balancing current is on the same order of magnitude as the operating current (hundreds of amperes), the balancing efficiency is far higher than that of traditional BMS (the balancing current is typically <5A). Improved battery consistency can significantly extend system life, expected to be more than 50% longer. Under conventional peak-shaving scenarios, only one complete cell replacement is needed during 20 years of operation (traditional energy storage typically requires two replacements), reducing the levelized cost of electricity (LCOS) over the entire lifecycle by 15%~20%.

[0024] (5) Scalability and modular design. This invention adopts... The battery matrix architecture can increase output voltage by increasing the number of series layers and increase output power by increasing the number of parallel circuits. This modular design allows the system to flexibly adapt to different power and voltage levels, from small-scale commercial and industrial energy storage (100kW level) to large-scale grid-side energy storage (10MW level). At the same time, the modular design also facilitates system expansion and maintenance, and the failure of a single battery cell does not affect the normal operation of other modules. Attached Figure Description

[0025] Figure 1 This is an architecture diagram of the DRBN direct-to-AC system of the present invention.

[0026] Figure 2 This is a flowchart of the modulation strategy of the present invention.

[0027] Figure 3 This is an output voltage waveform diagram of an embodiment of the present invention.

[0028] Figure 4 This is a harmonic analysis diagram of phase A according to an embodiment of the present invention.

[0029] Figure 5 This is a B-phase harmonic analysis diagram according to an embodiment of the present invention.

[0030] Figure 6 This is a C-phase harmonic analysis diagram according to an embodiment of the present invention.

[0031] Figure 7 Analysis results of total harmonic distortion (THD) Figure 1 .

[0032] Figure 8 Analysis results of total harmonic distortion (THD) Figure 2 The graph uses specific numerical values ​​to illustrate the analysis results.

[0033] Figure 9 This is a waveform detail analysis diagram. Detailed Implementation

[0034] The principles and features of the present invention are described below with reference to all the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0035] Reference Figures 1-9 This invention discloses a DC-AC direct conversion system based on a dynamically reconfigurable battery network and its control method.

[0036] In traditional energy storage systems, battery clusters are connected in a fixed series-parallel configuration. The output DC power needs to pass through a DC-DC boost stage and a DC-AC inverter stage in the power conversion system (PCS) to be converted into AC power. This architecture has multiple power conversion stages, each of which incurs energy losses, reducing the overall system efficiency. Furthermore, traditional PCS uses an H-bridge inverter topology, requiring a large number of IGBT / MOSFET (Insulated Gate Bipolar Transistor, MOSFET: Metal-Oxide-Semiconductor Field-Effect Transistor) switching devices and complex PWM (Pulse Width Modulation) control circuitry, resulting in high hardware costs and high control complexity.

[0037] The core idea of ​​the DRBN direct-to-AC technology proposed in this invention is to treat battery cells / modules as dynamically modulated "energy slices" and directly synthesize AC voltage waveforms that meet grid requirements through dynamic reconfiguration control of a high-speed power electronic switch array. This method fundamentally eliminates the DC-AC inverter stage of traditional PCS, reducing the number of power conversion stages from two (DC-DC + DC-AC) to one, significantly improving system efficiency and reducing hardware costs.

[0038] This invention is particularly suitable for large-scale energy storage power stations, industrial and commercial energy storage, microgrid systems and other scenarios. It can significantly reduce the initial investment and levelized cost of storage (LCOS) of energy storage systems, and has significant engineering application value and commercial prospects.

[0039] The basic principle of this invention lies in treating battery cells (single cells / modules, hereinafter collectively referred to as battery cells) as dynamically modulated "energy slices," and directly synthesizing AC voltage waveforms that meet the requirements of the power grid through dynamic reconfiguration control of a high-speed power electronic switching array. Its core idea is to replace the PWM modulation of traditional PCS with dynamic switching of the battery network topology, achieving a one-step DC-to-AC conversion.

[0040] The traditional PCS works by first boosting the battery voltage to a stable DC bus voltage using a DC-DC converter, and then converting the DC power to AC power using an H-bridge inverter and PWM modulation. In this process, the battery acts only as a passive energy source and does not participate in waveform synthesis.

[0041] The working principle of this invention is as follows: multiple battery cells are connected via a high-speed switch array, and each battery cell can be connected or bypassed at any time. By precisely controlling the connection sequence and duration of each battery cell, a target AC voltage waveform is synthesized at the output. In this process, the battery cell is both an energy source and the basic unit for waveform synthesis.

[0042] From a mathematical perspective, the output voltage of a traditional PCS It can be represented as: ; in, It is the DC bus voltage; It is a modulated wave, usually a duty cycle function generated by SPWM or SVPWM.

[0043] The output voltage of the DRBN direct-to-AC converter of this invention It can be represented as: ; in It is the first The switching state function of each battery cell. It is the first The terminal voltage of each battery cell.

[0044] By rationally designing the switch state function It can synthesize any target waveform at the output. For a 50Hz sinusoidal alternating current, the target waveform is: ; in, It is the amplitude. It refers to the frequency, specifically the power frequency of 50Hz. It is a phase.

[0045] This invention discloses a direct DC-AC converter system based on a dynamically reconfigurable battery network. (Refer to...) Figure 1 This is an architecture diagram of the DRBN direct-to-AC system of this invention, showing the connection relationships of core components such as the battery matrix, high-speed switch array, and digital energy exchange system. The architecture diagram illustrates how the battery cells achieve AC voltage synthesis through the dynamic reconfiguration of the switch array, and the energy transmission path from the battery to the grid / load. It includes: A battery matrix, which is composed of multiple battery cells connected in series and parallel, is used to provide electrical energy; A high-speed switching array is connected to each battery cell in the battery matrix, and is used to dynamically connect a single battery cell to the main circuit or bypass it according to control commands. The digital energy exchange system is connected to the high-speed switch array and is used to collect the battery cell status parameters in real time. Based on the target AC voltage waveform, it generates and issues the control command to dynamically reconstruct the topology connection of the high-speed switch array and control the access or bypass status of each battery cell in the series circuit in real time, so that the total voltage output by the battery matrix can be directly synthesized to approximate the AC voltage of the target AC voltage waveform. The AC output terminal is connected to the output terminal of the battery matrix and is used to directly output the AC voltage synthesized by the DC-AC direct conversion system to the power grid or load.

[0046] The overall concept of this invention includes the following three levels: First level: Hardware architecture design. (Adopting...) The battery matrix architecture, in which This corresponds to three battery clusters, each of which is modulated into an AC signal using DRBN technology. The number of series layers, Number of parallel paths. Number of series layers. The number of parallel circuits is determined by the target AC voltage and the battery cell voltage. The output power is determined by the target output power and the discharge rate of the battery cell. Each battery cell is equipped with a high-speed switching unit, including components such as a main circuit access switch and a bypass switch. In the topology, each battery cluster contains S layers connected in series, and each layer has P parallel battery cells.

[0047] Second level: Modulation strategy design. A waveform synthesis method based on energy slice modulation is adopted to divide the AC cycle into several time slots. In each time slot, an appropriate number of battery cells are connected in series to make the output voltage approach the desired value. The modulation strategy needs to consider the following constraints: (1) output voltage tracking accuracy; (2) battery SOC (State of Charge) equalization; (3) minimization of switching losses; (4) limitation of output harmonic content.

[0048] The third level: Collaborative optimization design. Active balancing and fault isolation of battery cells are achieved simultaneously with waveform synthesis. Automatic SOC balancing is achieved by dynamically adjusting the cumulative on-time of each battery cell within an AC cycle. When a battery cell fault is detected, its switching state is forcibly bypassed, while the conduction strategies of other battery cells are adjusted to maintain the output voltage.

[0049] The hardware architecture of the DRBN-to-AC direct conversion system proposed in this embodiment of the invention is shown in Figure 1. The following is a detailed description of the above content: (a) Battery Matrix Battery matrix adopts The topology, in which This corresponds to three battery clusters, each of which is modulated into an AC signal using DRBN technology. The number of series layers, This refers to the number of parallel paths.

[0050] Number of cascaded layers The determining principle is as follows: ; in, It is the peak value of the target AC voltage. For 220V / 50Hz AC, the peak value is approximately 311V. This is the lowest voltage (e.g., 2.5V) of a battery cell (taking a single cell as an example) at the end of its discharge phase. The 2 indicates that twice the number of battery cells are needed to generate AC positive and negative voltages on the DC side. Therefore, for a 220V AC output, Layer. Considering voltage margin, battery consistency, and fault scenarios, it is recommended to provide some margin in actual design, taking [a certain margin]. layer.

[0051] The principle for determining the number of parallel paths P is: ; in, It is the target output power; This is the battery's nominal voltage; This is the maximum discharge current of the battery cell. For a 700kW output system, a 314Ah / 0.5C cell is used. = 3.2V, =157A, then Choose M=2 parallel paths.

[0052] (ii) High-speed switching array Each battery cell is equipped with two switching units, specifically: Main circuit connection switch W1: controls whether the battery cell is connected in series circuit; Bypass switch W2: controls whether the battery cell is bypassed, i.e., the circuit is cut off.

[0053] The selection of switching devices must meet the following requirements: Pressure resistance: ; Flow: ; Switching speed: ; It refers to the net time required for voltage or current to transition between steady-state values ​​during a switching transient process; Rise time is the time required for a waveform to rise from 10% to 90% of its steady-state amplitude (rising edge). The fall time is typically the time required for the waveform to drop from 90% of its steady-state amplitude to 10% (falling edge). On-resistance: .

[0054] In this embodiment, the switching device adopts wide bandgap semiconductor devices such as SiC MOSFET or GaN HEMT, which have advantages such as fast switching speed, low conduction loss and high temperature resistance.

[0055] (III) Digital Energy Exchange System The digital energy exchange system is the brain of the system, responsible for the following functions: Real-time acquisition of status parameters such as voltage, current, temperature, and SOC of each battery cell; Calculate the status commands for each switching unit based on the target output voltage waveform; Achieve battery equalization control and fault isolation protection; It communicates with the host computer (EMS, power grid dispatcher).

[0056] The digital energy exchange system adopts a distributed architecture and includes the following functions: Battery cell modulation strategy generation, switching timing generation (switching frequency ≥ 10Hz), switch scheduling and overall coordination, as well as system-level control and communication (control cycle ≤ 100μs), the communication protocol adopts high-speed real-time CAN bus.

[0057] The digital energy exchange system adopts a distributed architecture, including: Digital energy network cards are used to collect voltage, current, and temperature data for individual battery cells; A digital energy co-controller is used to perform local switching timing generation and scheduling for a group of battery cells; Digital energy controllers are used to perform global system-level control, communication, and solve optimization algorithms.

[0058] The digital energy exchange system generates the switching timing of the high-speed switching array by solving a multi-objective optimization problem. The objective functions of the multi-objective optimization include: minimizing the tracking error between the output voltage and the target waveform, minimizing the standard deviation of the state of charge (SOC) between battery cells, and minimizing the losses caused by switching actions.

[0059] The optimization algorithm mentioned above specifically refers to an algorithm for solving a multi-objective dynamic optimization problem. Its core is to find an optimal solution from a massive number of switching combinations within each control cycle, simultaneously achieving three mutually constraining objectives: minimizing voltage tracking error, minimizing battery SOC imbalance, and minimizing switching action losses. The digital energy controller needs to solve a problem of minimizing a comprehensive cost function within each time slice. This comprehensive cost function is composed of the weighted average of the three sub-objectives: minimizing voltage tracking error, minimizing battery SOC imbalance, and minimizing switching action losses.

[0060] It should be noted that this multi-objective optimization problem is solved using a step-by-step decision-making approach, rather than a single algorithm for global optimization. 1. Determine the number of cells to be connected: Using the nearest level approximation modulation, calculate the number of battery cells to be connected based on the target voltage.

[0061] 2. Select specific batteries: Based on the priority function (considering SOC, voltage, and temperature), prioritize the connection of batteries with higher SOC to achieve battery balancing.

[0062] 3. Optimize the switching sequence: Adopt the strategy of minimizing the number of switching operations, prioritize maintaining the state of the previous time step, and reduce the number of switching actions. This combinatorial optimization problem can be solved by dynamic programming or heuristic algorithms.

[0063] (iv) Output filtering stage To reduce output harmonic content, the system can be configured with an output filter. Preferably, an LC filter or an LCL filter is used, with its cutoff frequency designed as follows: ; The cutoff frequency should satisfy: That is, 500Hz <10kHz, for =50Hz, =100kHz.

[0064] (v) Grid-connected synchronization unit The grid synchronization unit is responsible for synchronizing the phase and frequency of the output voltage of the DC-AC direct conversion system with the grid voltage, including: Phase-locked loop (PLL): Detects the phase of the mains voltage; Frequency tracking: Tracking changes in the power grid frequency; Pre-synchronization control: Achieve precise synchronization of voltage amplitude, phase, and frequency before grid connection.

[0065] The DRBN direct-to-AC method in this embodiment of the invention is mathematically based on the optimized design of the switching function to make the output voltage of the battery matrix approximate the desired AC waveform. The mathematical model of the system is established below.

[0066] (1) Definition of switch state Definition of the first l String number i Layer j The switching state function of the circuit battery is: ; in, This indicates that the battery is connected in a series circuit, the main circuit switch is closed, and the bypass switch is open; =0 indicates that the circuit is bypassed; the main circuit switch is open, and the bypass switch is closed.

[0067] (2) Output voltage expression The system output voltage of the l-th string This is the sum of the voltages of all connected battery cells in the string: ; in, It is the terminal voltage of the first-path battery cell in the i-th layer of the l-th string, assuming that the first-path batteries in all layers are connected in the loop.

[0068] (3) Target tracking error Define voltage tracking error for: ; in, It is the target sine wave. This is the actual output voltage.

[0069] The control objective is to minimize the tracking error of each string. The mean square value is the smallest, that is: .

[0070] in, This represents the square of the voltage tracking error of the l-th battery cluster (i.e., the l-th phase) at time t.

[0071] (4) Battery SOC model Definition of the first l String number i Layer j The SOC of the battery cell is: ; in, It refers to battery capacity; It is the current flowing through it; Indicates that it is located at the th l String number i Layer j The initial state of charge of the battery cell at the initial moment.

[0072] (5) SOC balance index The SOC uniformity is defined as the standard deviation of the SOC of each battery cluster, where the standard deviation of the SOC of the l-th cluster is... Defined as: ; in, It is a battery cluster Average SOC, .

[0073] The control objective is to minimize the SOC equilibrium, that is: ; .

[0074] (6) Switching loss model Define the switching loss of the l-th string. for: ; in, It is the switching frequency; , It refers to the energy loss during a single power-on and power-off cycle.

[0075] (7) Multi-objective optimization problem Based on the above objectives, a multi-objective optimization problem is established: ; The constraints are as follows: Voltage constraint: ; Current constraint: ; SOC constraints: ; Temperature constraints: ; in , , This is a weighting coefficient, which can be adjusted according to the application scenario. For grid-connected applications, (Voltage tracking) should have the highest weight; for off-grid applications, The weight of (SOC equilibrium) should be appropriately increased.

[0076] (8) Harmonic content constraint Output harmonic content is measured by total harmonic distortion (THD): ; in, It is the fundamental frequency amplitude. yes Second harmonic amplitude.

[0077] According to the IEEE 519 standard, the total heat dissipation (THD) of a grid-connected system should be below 5%. Therefore, the switching strategy design must meet the following requirements: .

[0078] (9) Theoretical efficiency analysis System efficiency is defined as: ; in It outputs AC power. It is the DC power output by the battery.

[0079] The efficiency loss mainly comes from: Switch conduction loss: , Duty cycle; Switching losses: ; Filter loss: , It represents the DC resistance of the inductor.

[0080] The theoretical efficiency can be estimated as follows: ; For optimized design systems It can reach over 98%, significantly higher than the 96%~97% of traditional PCS.

[0081] It should be noted that, The numerical values ​​can be qualitatively and quantitatively analyzed from multiple dimensions, including actual engineering data, theoretical models, and authoritative certifications. Qualitatively: DRBN technology, through architectural innovation, eliminates the DC-DC boost stage of power conversion in traditional energy storage systems, fundamentally reducing losses. Quantitatively: Based on the theoretical efficiency calculation model provided by the patent, and using wide-bandgap semiconductor devices, an efficiency of over 98% is theoretically self-consistent. Authoritative endorsement: The National Energy Administration's list of the first (set) of dynamically reconfigurable major technical equipment provides official recognition for "DC-side energy efficiency exceeding 95%".

[0082] Secondly, based on the same inventive concept, embodiments of the present invention also disclose a DC-AC direct conversion method based on a dynamically reconfigurable battery network, applied to the aforementioned DC-AC direct conversion system based on a dynamically reconfigurable battery network, comprising the following steps: The digital energy exchange system acquires the target AC voltage waveform, which has preset amplitude, frequency and phase. The digital energy exchange system, based on the target AC voltage waveform, dynamically reconstructs the topology connection of the high-speed switching array and controls the access or bypass state of each battery cell in the series circuit in real time, so that the total voltage output by the battery matrix can be directly synthesized to approximate the AC voltage waveform of the target AC voltage. The synthesized AC voltage is directly output to the grid or load without going through a separate DC-AC inverter.

[0083] In this embodiment of the invention, the battery cell is used as the basic modulation unit, and the AC voltage waveform is directly synthesized by dynamically reconstructing the battery network topology. The steps for synthesizing the AC voltage include: S1: Divide an alternating cycle into... A time slice.

[0084] S2: Within each time slice, calculate the number of battery cells that need to be connected to the circuit based on the target voltage value. .

[0085] Specifically, in calculating the number of battery cells that need to be connected to the circuit... At that time, the nearest level approximation modulation strategy is adopted, that is, selecting the level that makes the modulation result... closest to the target voltage value integers ,in, This represents the average voltage of the battery cell.

[0086] S3: Select from the battery matrix according to the preset priority function. Each battery cell is connected in series to a circuit by controlling the corresponding switch array.

[0087] Specifically, the preset priority function is constructed based at least on the state of charge (SOC) of the battery cells, and prioritizes the connection of battery cells with higher SOC to the circuit, thereby achieving battery state of charge balance while synthesizing AC waveforms.

[0088] S4: At the end of each time slice, update the on / off status of all battery cells.

[0089] In this embodiment of the invention, the switching state of each battery cell is calculated in real time based on the target output voltage waveform, i.e., a switching control method based on energy slice modulation. The control method is described in detail below: (1) Time slice division A communication cycle Divided into There are 1 time slice, and the duration of each time slice is: ; Where T represents the duration of each time slice; The value of needs to be determined by balancing tracking accuracy, switching frequency, and system cost. The larger the value, the higher the tracking accuracy, but the higher the switching frequency and the higher the system cost. In this embodiment, =5~40, corresponding to .

[0090] (2) Calculation of the number of levels Within each time slice, the number of voltage levels that the DC-AC direct conversion system, i.e., the series link of a single battery cluster, i.e., the single-phase voltage synthesis unit based on DRBN technology, can output is: ; in, It refers to the number of series layers. For example, At that time, the DC-AC direct conversion system can output 17 levels (0, 1, 2, ..., 16 battery cells in series).

[0091] (3) Nearest level approximation modulation In each time slice Calculate the target voltage : ; Choose the voltage level that is closest to the target voltage. : ; in, It is the average voltage of the battery cells. This represents the rounding function.

[0092] Then select A single battery cell is connected to the circuit; a 0-level signal indicates that no battery cell is connected to the circuit. To balance battery SOC, modules with higher SOC are preferred.

[0093] (4) SOC-based module selection strategy Define the priority function for the battery cell in the i-th layer and j-th path of the l-th string: ; in , , These are weighting coefficients. It refers to the battery temperature.

[0094] In each time slice, battery cells are selected to be connected to the circuit according to their priority from high to low. In this way, battery cells with high SOC have a longer conduction time and discharge more, while battery cells with low SOC have a shorter conduction time and discharge less, thus achieving automatic balancing.

[0095] (5) Optimization of switching timing To reduce switching losses, a minimum number of switching operations strategy is adopted. The switching state change quantity is defined as follows: ; The goal is to minimize the number of switches: ; This is a combinatorial optimization problem that can be solved using dynamic programming or heuristic algorithms, avoiding the simultaneous switching of multiple modules (distributed switching actions).

[0096] Dynamic programming is a mathematical method for solving complex optimization problems by decomposing the original problem into relatively simple subproblems. In the field of battery management, it is often used to solve globally optimal energy allocation or switching control strategies. Heuristic algorithms are algorithms constructed based on intuition or experience, providing a feasible solution within acceptable computational costs. They typically do not guarantee global optimality but are highly efficient. Some literature proposes optimized switching control strategies based on genetic algorithms for SOC balancing of reconfigurable battery modules. Particle swarm optimization (PSO) has been used for power allocation strategies in hybrid energy storage systems, as well as for optimizing SOC and voltage membership functions in lithium battery balancing control. Graph theory methods model the dynamic reconfiguration problem of battery network topology as a shortest weighted path problem in a directed acyclic graph.

[0097] (6) Dead Time Design To prevent the main circuit switch and bypass switch from opening simultaneously and causing an open circuit in the entire circuit, a dead time needs to be set: ; in, It is the maximum off time of the switch. This is a safety margin, typically taken as 0.5μs to 1μs.

[0098] Dead time can cause output voltage distortion, requiring dead time compensation. The compensation method involves adding a correction factor to the target duty cycle. ; in, Duty cycle correction is added to compensate for the duty cycle and to eliminate output voltage distortion caused by the dead-time effect. It is a sign function used to determine the current. Direction (polarity), when When >0 (positive current), it is 1; when =0 (current is zero) is 0, when When <0 (current reversed), it is -1.

[0099] (7) Harmonic suppression strategy To reduce the output harmonic content, the following strategies are adopted: Increasing the number of levels: The more levels there are, the closer the output waveform is to a sine wave; Optimize switching timing: Employ harmonic elimination method to selectively eliminate specific harmonics; Configure the output filter: Design a suitable LC or LCL filter.

[0100] (8) Modulation strategy under fault conditions When a battery malfunction is detected, it is forced to bypass, and the modulation strategy is adjusted: Recalculate the number of available voltage levels; Reassign the switching timing to ensure that the output voltage remains close to the desired value; If the number of faulty batteries exceeds the threshold, derating operation or shutdown protection will be triggered.

[0101] like Figure 2 As shown, the complete process from target voltage calculation, level selection, SOC-based module selection, switching timing optimization to final switching state generation is described in detail.

[0102] The following is combined Figure 2 One AC cycle for an embodiment of the present invention The cyclic execution process of the core control logic within multiple scheduling cycles is explained in detail: Step 1: Sample the current time to obtain the power grid synchronization time.

[0103] The system uses a phase-locked loop in the grid synchronization unit to lock the phase zero-crossing point and frequency of the grid voltage in real time, thereby obtaining the grid phase angle at the current moment. In off-grid mode, an internal high-precision clock source is used as the reference.

[0104] Step 2: Calculate the three-phase reference voltage and generate a 50Hz sinusoidal target waveform.

[0105] Based on the grid phase angle and target amplitude obtained in step 1 (determined according to active and reactive power dispatch instructions), the instantaneous target voltage values ​​of the three phases at the current moment are calculated in real time to form a 50Hz sinusoidal target waveform, that is, to form the target AC voltage waveform.

[0106] Step 3: Read the current SOC of all battery cells.

[0107] The controller concurrently reads the real-time state of charge of all battery cells in the battery matrix via a high-speed communication bus, ensuring that it has millisecond-level battery energy distribution information.

[0108] Step 4: Read the overall voltage of the battery cluster and the voltage of each battery cell.

[0109] The total output voltage of each battery cluster and the terminal voltage of each battery cell are collected synchronously to verify the battery health status, provide voltage error feedback, and provide a basis for subsequent compensation calculations.

[0110] Step 5: Determine the set of battery cells to be connected to the circuit according to the principles of minimizing output voltage error and SOC balance.

[0111] For each phase, i.e. each battery cluster, the system calculates the number of batteries to be connected based on the instantaneous target voltage and the average voltage of the battery cells. Then, taking into account multiple objectives such as minimizing the output voltage error, prioritizing the connection of battery cells with higher SOC, and minimizing the number of switching operations, the system uses an optimization algorithm to determine the set of battery cells that should be connected to the series circuit in the current time slice.

[0112] Step 6: Generate drive signals for each MOSFET switch within the battery cluster.

[0113] The determined switching state command is converted into a hardware drive level signal for each switching device, and each battery cell corresponds to a set of complementary drive signals.

[0114] Step 7: Consider dead time and delay compensation, and perform switch timing control.

[0115] To prevent short circuits caused by simultaneous switching, a dead time is inserted during the switching action, and the duty cycle is compensated for dead time according to the direction of the current output current; at the same time, the transmission delay of the drive signal of the switches in different positions is calibrated to ensure synchronous operation.

[0116] Step 8: The output voltages of the battery cells connected to the circuits in the three battery clusters are combined to output a three-phase AC voltage.

[0117] The three independent battery clusters correspond to phases A, B, and C respectively, each independently synthesizing a sinusoidal voltage waveform. The three phases are 120° apart, and together they synthesize a three-phase AC power at the output terminal, which is directly supplied to the power grid or load.

[0118] Step 9: Measure the output AC voltage and adjust the subsequent modulation strategy according to the error.

[0119] The system measures the output three-phase AC voltage in real time and compares it with the target waveform to calculate the tracking error. This error is used to correct the subsequent voltage tracking weights in a closed-loop manner, analyze harmonic content to optimize switching timing, and assist in determining whether a fault has occurred.

[0120] Step 10: Return to Step 1 and continue the loop.

[0121] After completing the control of the current time slice, the time is advanced to the next time slice, and steps 1 to 9 are strictly executed in a loop at intervals equal to the scheduling cycle to achieve continuous, real-time closed-loop regulation of the output voltage.

[0122] The following example further illustrates this embodiment: This embodiment is a 23kW / 46kWh DRBN direct-to-AC energy storage system.

[0123] The system parameters are designed as follows: AC output voltage: 220V / 50Hz (phase voltage / fundamental frequency); Rated output power: 23kW; Battery cell: Lithium iron phosphate (LFP) cell, 3.2V / 50Ah; Battery cluster L: 3 strings; Number of cascaded layers S: 256 layers; Number of parallel paths P: 2; Switching device: SiC MOSFET, withstand voltage 100V, current 200A, on-resistance 3mΩ; Switching frequency: 50kHz; Scheduling period: 1ms; Switching timing deviation: 0.1ms.

[0124] The control system is designed as follows: The controller adopts a three-level distributed architecture, and the digital energy exchange system includes a digital energy controller, a digital energy co-controller, and a digital energy network card; Communication bus: CAN2.0 bus is used. At a rate of 1 Mbit / s, the maximum delay of high-priority messages is less than 120 microseconds (µs). Sampling frequency: Voltage / current sampling 1kHz, temperature sampling 1kHz; Operating conditions: Full power discharge (23kW).

[0125] The performance indicators are as follows: (1) Output voltage quality: Under full power conditions, the output voltage THD=0.6%, which meets the 5% limit requirement of IEEE 519. Voltage amplitude deviation <±2%, frequency deviation <±0.1Hz.

[0126] (2) System efficiency: Under full power conditions, the system efficiency η = 98.3% (from battery DC power to AC output power).

[0127] (3) Battery equalization effect: After 5 peak-valley cycles (about 20 hours), the standard deviation of SOC of the battery cell decreased from the initial 2.1% to 0.6%, which verified the equalization effect of the SOC-based modulation strategy.

[0128] (4) Fault isolation: During operation, an overvoltage fault (voltage > 3.8V) was simulated in the first battery cell of the fifth layer. After the system detected the fault, it bypassed the module within 50μs and adjusted the switching timing of other modules. During the fault isolation process, the output voltage did not show significant distortion (voltage deviation < 5%), verifying the effectiveness of the fault isolation strategy.

[0129] like Figure 3 The figure shows the synthesized 220V / 50Hz AC voltage waveform of the DRBN direct-to-AC system. The waveform diagram shows that, using a modulation strategy with 184 layers in series and a scheduling period of 1ms, the switching timing deviation is 0.1ms, and the output voltage can track the target sine wave well, with a THD of less than 5%, meeting the grid connection requirements.

[0130] like Figures 4-9 The output voltage spectrum distribution is shown in the figure. The harmonic analysis is based on FFT calculations, and the low-order harmonics (2nd to 50th) are very low and can be further reduced by using an output filter (configurable as needed).

[0131] This embodiment demonstrates that, by using DRBN direct-to-AC technology, system efficiency can be significantly improved while ensuring output voltage quality, achieving battery balancing, and quickly isolating faulty batteries.

[0132] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A direct DC-AC converter system based on a dynamically reconfigurable battery network, characterized in that, include: A battery matrix, which is composed of multiple battery cells connected in series and parallel, is used to provide electrical energy; A high-speed switching array is connected to each battery cell in the battery matrix, and is used to dynamically connect a single battery cell to the main circuit or bypass it according to control commands. The digital energy exchange system is connected to the high-speed switch array and is used to collect the battery cell status parameters in real time. Based on the target AC voltage waveform, it generates and issues control commands to dynamically reconstruct the topology connection of the high-speed switch array and control the access or bypass status of each battery cell in the series circuit in real time, so that the total voltage output by the battery matrix can be directly synthesized to approximate the AC voltage of the target AC voltage waveform. The AC output terminal is connected to the output terminal of the battery matrix and is used to directly output the AC voltage synthesized by the DC-AC direct conversion system to the power grid or load.

2. The DC-AC direct conversion system based on a dynamically reconfigurable battery network according to claim 1, characterized in that: The battery matrix adopts The three-dimensional topology is defined as follows: L represents the number of battery clusters, corresponding to the number of AC phases in the output; S represents the number of series layers in each battery cluster, used to determine the output voltage amplitude; P represents the number of parallel circuits in each layer, used to determine the output power; where, in In the topology, each battery cluster contains S layers connected in series, and each layer has P parallel battery cells.

3. The DC-AC direct conversion system based on a dynamically reconfigurable battery network according to claim 2, characterized in that: Number of cascaded layers satisfy ,in The target AC voltage peak value, This is the minimum operating voltage for the battery cell.

4. The DC-AC direct conversion system based on a dynamically reconfigurable battery network according to claim 1, characterized in that: Each switching device in the high-speed switching array is a wide-bandgap semiconductor device, including SiC MOSFETs or GaN HEMTs, with a switching speed of , The rise time, i.e., the rising edge; Fall time, i.e., the falling edge; on-resistance. .

5. A DC-AC direct conversion system based on a dynamically reconfigurable battery network according to claim 1, characterized in that, The digital energy exchange system adopts a distributed architecture, including: Digital energy network cards are used to collect voltage, current, and temperature data for individual battery cells; A digital energy co-controller is used to perform local switching timing generation and scheduling for a group of battery cells; Digital energy controllers are used to perform global system-level control, communication, and solve optimization algorithms.

6. A DC-AC direct conversion system based on a dynamically reconfigurable battery network according to claim 1, characterized in that: It also includes an output filtering stage, which is an LC filter or an LCL filter, with a cutoff frequency of The design is between 10 times the grid frequency and 0.1 times the switching frequency.

7. A DC-AC direct conversion system based on a dynamically reconfigurable battery network according to claim 1, characterized in that: It also includes a grid-connected synchronization unit, which includes a phase-locked loop (PLL) for detecting the phase and frequency of the grid voltage and controlling the output voltage of the DC-AC direct converter system to achieve pre-synchronization of amplitude, phase and frequency with the grid voltage before grid connection.

8. A direct DC-AC conversion method based on a dynamically reconfigurable battery network, characterized in that, The method applied to a DC-AC direct conversion system based on a dynamically reconfigurable battery network as described in any one of claims 1-7 includes the following steps: The digital energy exchange system acquires the target AC voltage waveform, which has preset amplitude, frequency and phase. The digital energy exchange system dynamically reconstructs the topology of the high-speed switching array based on the target AC voltage waveform, and controls the access or bypass status of each battery cell in the series circuit in real time, so that the total voltage output by the battery matrix can be directly synthesized to approximate the AC voltage waveform of the target AC voltage. The synthesized AC voltage is directly output to the grid or load without going through a separate DC-AC inverter.

9. A direct DC-AC conversion method based on a dynamically reconfigurable battery network according to claim 8, characterized in that: The battery matrix adopts The three-dimensional topology is defined as follows: L represents the number of battery clusters, corresponding to the number of AC phases in the output; S represents the number of series layers in each battery cluster, used to determine the output voltage amplitude; P represents the number of parallel circuits in each layer, used to determine the output power; where, in In the topology, each battery cluster contains S layers connected in series, and each layer has P parallel battery cells.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the DC-AC direct conversion method based on a dynamically reconfigurable battery network as described in claim 8 or 9.

Images