Control method based on three-level cascade Buck-Boost converter and energy storage system

Abstract

The invention provides a control method and an energy storage system based on a three-level cascade Buck-Boost converter, and the method comprises the steps: employing a mode switching point threshold value calculation method based on the flying capacitor voltage and inductive current dynamic characteristics, and achieving the switching between different working modes of the Buck-Boost converter, in combination with real-time data such as input and output voltage, inductance equivalent parameters and a control period, an inductive current switching point is dynamically determined, smooth transition among different modes is realized, inductive current abrupt change caused by mode switching is effectively inhibited, impact on direct current bus voltage is reduced, and the reliability of the system is improved. And the stability and the dynamic response speed of voltage stabilization control are improved.

Description

Technical Field

[0001] The present application relates to the field of power electronics technology, and in particular to a control method and energy storage system based on a three-level cascaded Buck-Boost converter. Background Art

[0002] Three-level cascaded Buck-Boost converters have broad application prospects in energy storage systems, renewable energy generation, and DC microgrids. These converters typically require stable control of the DC bus voltage and active balancing of the flying capacitor voltage. The performance of their control strategy directly impacts system reliability, dynamic response, and overall efficiency. Traditional control architectures typically employ a fixed-threshold mode switching method combined with a voltage and current dual closed-loop proportional-integral (PI) controller to achieve DC voltage regulation. The mode switching thresholds are often calculated and set offline based on nominal circuit parameters, making them incapable of real-time adjustment to actual operating conditions such as load power variations, device parameter drift, or external disturbances. Furthermore, balancing the flying capacitor voltage is typically achieved by introducing an additional voltage balancing loop. The parameters of this control loop are typically fixed, failing to account for the impact of load power on the dynamic balancing process. This results in insufficient response under high-power conditions and prone to overshoot or oscillation under low-power conditions.

[0003] Existing control methods still have the following defects: fixed switching thresholds are prone to cause sudden changes in inductor current, which in turn leads to DC bus voltage drops or overshoots; non-adaptive capacitor voltage balancing strategies are difficult to balance dynamic response and steady-state stability within the full power range, which may cause long-term capacitor voltage imbalance; and predictive control methods based on static models have model mismatch when parameters change, resulting in insufficient system robustness. Summary of the Invention

[0004] In view of this, the embodiments of the present application provide a control method and energy storage system based on a three-level cascade Buck-Boost converter, aiming to solve the technical problem that the existing converter using a fixed switching threshold is prone to causing sudden changes in the inductor current.

[0005] A first aspect of an embodiment of the present application provides a control method based on a three-level cascaded Buck-Boost converter, wherein the converter includes an input side and an output side, an inductor is located between the input side and the output side, and the inductor is used as a multiplexed inductor to store or release energy in different operating states of the converter; The control method includes: collecting the current of the inductor, the voltage of the flying capacitor on the input side, and the voltage of the flying capacitor on the output side in real time, and calculating the threshold of the inductor current modal switching point in combination with the input voltage, output voltage, control period, and parameters of the inductor of the converter; Based on the real-time voltage deviations of the flying capacitor on the input side and the flying capacitor on the output side, respectively, the capacitor voltage adjustment signals on the input side and the output side are calculated by an adaptive P controller to dynamically correct the inductor current reference signal; Constructing a predictive control model including load parameters, the predictive control model is used to estimate the equivalent parameter changes of the inductor in real time, and dynamically adjust the prediction weight coefficients of the inductor current and capacitor voltage in the predictive control model according to the estimation results; Constructing a prediction cost function including the prediction weight coefficient, determining a target mode based on the prediction cost function, and calculating a target duty cycle of the switching tube of the converter based on the switching tube action logic of the target mode and the load parameter; The target duty cycle is input to the carrier stacking and carrier phase shift modulation unit to generate control signals for each switch tube, and the control signals are input to the converter. In one embodiment, the input side includes a flying capacitor and , switch tube 、 、 and The input side bridge arm, switch tube and For the main control switch tube, switch tube and For auxiliary switch tube; flying capacitor and The series midpoint of the circuit serves as the midpoint potential of the input side; The output side includes a flying capacitor and , switch tube 、 、 and The output side bridge arm, switch tube and For the main control switch tube, switch tube and For auxiliary switch tube; flying capacitor and The series midpoint of the circuit serves as the midpoint potential of the output side; The inductor L1 is connected between the input side and the output side.

[0006] In one embodiment, the step of calculating the threshold of the inductor current mode switching point includes: Determining a mode type of the current converter based on a voltage gain formula; the mode types include Buck1, Buck2, Boost1, Boost2, Buck-Boost1, and Buck-Boost2; If the current mode is Buck1, Buck2, Boost1 or Boost2, the threshold of the switching point is determined by an integral calculation formula; If the current mode is Buck-Boost 1 or Buck-Boost 2, the threshold of the switching point is determined by a piecewise linear interpolation formula.

[0007] In one embodiment, the integral calculation formula is: ;in is the instantaneous value of the voltage across the inductor L1 at time t in the current mode; Indicates the inductor current threshold of the jth mode switching point in the current cycle k, represents the inductor current threshold of the previous mode switching point j-1 in the current cycle k, Indicates the real-time equivalent inductance value of inductor L1 at time t; The piecewise linear interpolation formula: .

[0008] In one embodiment, the calculating the capacitor voltage regulation signals at the input side and the output side respectively by the adaptive P controller includes: Dynamically adjusting the capacitor voltage regulation signals on the input side and the output side by dynamically adjusting the proportional coefficient of the adaptive P controller according to the power demand of the converter at the current moment; The capacitor voltage regulation signal on the input side is calculated by the following formula: ;in 、 is the proportional coefficient of the input-side P controller, Indicates the reference command value of the inductor current in the current cycle k; The capacitor voltage regulation signal on the output side is calculated by the following formula: ; in 、 is the proportional coefficient of the output side P controller, Indicates the reference value of the converter output voltage, What is calculated is the voltage regulation deviation of the converter output DC bus.

[0009] In one embodiment, dynamically adjusting the prediction weight coefficients of the inductor current and capacitor voltage in the prediction control model based on the estimation results includes: calculating the equivalent inductance value and the external disturbance value of the inductor; calculating the load resistance; and dynamically adjusting the prediction weight coefficient based on the equivalent inductance value, the external disturbance value and the load resistance.

[0010] In one embodiment, constructing a prediction cost function including the prediction weight coefficients includes: The inductor current deviation under the target mode , the flying capacitor voltage deviation on the input side , the flying capacitor voltage deviation on the output side Perform weighted summation; The calculation formula of the prediction cost function is: ; in, 、 、 is the prediction weight coefficient.

[0011] In one embodiment, after determining the target mode based on the prediction cost function, it also includes: if the target mode is Buck-Boost1 or Buck-Boost2, further determining the symmetry deviation of the flying capacitor voltage on the input side and the output side; if the symmetry deviation exceeds a preset threshold, introducing symmetry compensation to calculate the target duty cycle.

[0012] In one embodiment, the target duty cycle is calculated by the following formula: ; in, is the duty cycle without introducing symmetry compensation, 、 is the capacitor voltage symmetry compensation coefficient, 、 is the symmetry deviation of the flying capacitor voltages at the input side and the output side.

[0013] A second aspect of the embodiments of the present application provides an energy storage system, including a three-level cascade Buck-Boost converter and a control unit, wherein the control unit is configured to implement the control method mentioned in any one of the above embodiments.

[0014] Beneficial effects of the embodiments of the present application: This technical solution breaks through the existing practice of using fixed thresholds or offline setting of switching points based on simplified circuit models in determining the switching point of the inductor current. It proposes a dynamic calculation method based on real-time parameters and modal judgment, thereby maintaining the continuity and smoothness of the inductor current during the mode switching process.

[0015] By introducing a power demand-driven adaptive regulation strategy, the problem of fixed proportional controller parameters and difficulty in balancing voltage balance speed and current loop stability under high and low power conditions in the existing technology is effectively solved.

[0016] A combined extended state observer and online parameter identification method was introduced to achieve real-time estimation of load parameters and dynamic adjustment of prediction weights. The estimated parameters are then incorporated into the MPC model, dynamically adjusting the weights of the inductor current deviation and capacitor voltage deviation terms based on the deviation of the inductor or load resistance from the nominal value. This allows the MPC to maintain high prediction accuracy and optimal control despite device parameter drift, load variations, or external disturbances.

[0017] When the target mode is Buck-Boost, a capacitor voltage symmetry judgment and compensation mechanism is introduced to ensure the voltage symmetry of the flying capacitor and prevent the midpoint voltage offset from adversely affecting the DC bus voltage and output waveform quality, thereby further improving the system's voltage regulation accuracy and output quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] Figure 1 A schematic diagram of a three-level cascaded Buck-Boost converter provided in an embodiment of the present application; Figure 2 A flowchart of a control method provided in one embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0021] It should be noted that when an element is referred to as being “fixed on” or “disposed on” another element, it may be directly on the other element or indirectly on the other element. When an element is referred to as being “connected to” another element, it may be directly connected to the other element or indirectly connected to the other element.

[0022] It should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application.

[0023] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0024] See also Figure 1 An embodiment of the present application provides a control method based on a three-level cascaded Buck-Boost converter. The converter includes an input side and an output side. An inductor L1 is located between the input side and the output side. The inductor L1 serves as a multiplexed inductor for storing or releasing energy in different working states of the converter.

[0025] The input side includes a DC input power supply , flying capacitor and , switch tube 、 、 and The input side bridge arm, switch tube and For the main control switch tube, switch tube and For auxiliary switch tube; flying capacitor and The midpoint of the series connection serves as the midpoint potential of the input side and is connected to the input-side main switch bridge arm. Buck or Boost energy conversion on the input side is achieved by turning on different switch combinations. The circuit's energy transmission element is a reusable inductor L1, located between the input and output sides, used to store and release energy in different operating modes. Between the input and output bridge arms, inductor L1 serves as both an output filter element in Buck mode and an energy storage element in Boost mode, supporting bidirectional power flow.

[0026] The output side structure is similar to the input side, and the output side includes a flying capacitor and , switch tube 、 、 and The output side bridge arm, switch tube and For the main control switch tube, switch tube and For auxiliary switch tube, flying capacitor and The midpoint of the series connection serves as the midpoint potential of the output side. The output side can be operated in a boost, buck, or Buck-Boost mixed mode by different conduction combinations of the switch tubes. The output side is connected to a load or energy storage unit for bidirectional energy exchange.

[0027] See also Figure 2 The control method based on the three-level cascade Buck-Boost converter provided in the embodiment of the present application includes the following steps: S100 , collecting the inductor current, the flying capacitor voltage on the input side, and the flying capacitor voltage on the output side in real time, and calculating the threshold of the inductor current modal switching point by combining the input voltage, output voltage, control period, and inductor parameters of the converter.

[0028] In this embodiment, the inductance of the three-level cascade Buck-Boost converter is collected in real time. Real-time current , Input side flying capacitor and Voltage 、 , output side flying capacitor and Voltage 、 , and based on the inductor current at the current moment , the voltage value of the flying capacitor on the input and output sides, combined with the input voltage of the converter , output voltage , control cycle and inductance Parameter characteristics, calculate the threshold of the inductor current mode switching point .

[0029] S200, based on the real-time voltage deviation of the flying capacitor on the input side and the flying capacitor on the output side, the capacitor voltage adjustment signal on the input side and the output side is calculated by the adaptive P controller respectively, which is used to dynamically correct the inductor current reference signal. In this embodiment, according to the real-time voltage deviation of the flying capacitor on the input side and the output side, the input side capacitor voltage adjustment signal is input to the adaptive P controller on the input side and the output side respectively, and the input side capacitor voltage adjustment signal is generated. and output side capacitor voltage regulation signal , used to dynamically correct the inductor current reference signal .

[0030] S300, constructing a prediction control model including load parameters, the prediction control model is used to estimate the equivalent parameter changes of the inductor in real time, and dynamically adjust the prediction weight coefficients of the inductor current and capacitor voltage in the prediction control model according to the estimation results.

[0031] In this embodiment, based on the inductor current collected in step S100 , the adjustment signal generated in step S200 、 , and the flying capacitor voltages on the input and output sides 、 、 、 , construct a predictive control model (MPC) including a load parameter estimation module, which estimates the inductance in real time through an extended state observer The equivalent parameter changes are detected and the prediction weight coefficients of the inductor current and capacitor voltage in the MPC model are dynamically adjusted based on the estimation results.

[0032] S400 , constructing a prediction cost function including a prediction weight coefficient, determining a target mode based on the prediction cost function, and calculating a target duty cycle of the switching tube of the converter based on the switching tube action logic and load parameters of the target mode.

[0033] In this embodiment, by constructing a prediction cost function including a prediction weight coefficient, the control effect of each preset mode under the current working conditions is quantitatively evaluated, providing a quantitative basis for the selection of the optimal target mode, so that the prediction cost function can accurately reflect the control priority of the system under different working conditions, thereby screening out the optimal target mode that can minimize the state variable deviation and adapt to the changes in load parameters, providing a reliable target mode benchmark for subsequent duty cycle calculations, and ensuring that the control strategy can still achieve stable control of the DC bus voltage under parameter fluctuations or disturbances.

[0034] S500. Input the target duty cycle into the carrier stacking and carrier phase-shift modulation unit to generate the control signals of each switching tube, and input the control signals into the converter.

[0035] In the embodiments of the present application, there are significant improvements in the DC voltage regulation accuracy and system robustness, mainly due to the introduction of a method for calculating the threshold of the mode switching point based on the dynamic characteristics of the flying capacitor voltage and inductor current. When switching between different operating modes (Buck, Boost, Buck-Boost) in the embodiments of the present application, instead of using traditional fixed thresholds or simplified circuit models for calculation, real-time data such as input and output voltages, inductor equivalent parameters, and control period are combined to dynamically determine the inductor current switching point, achieving smooth transitions between different modes, effectively suppressing the sudden change of the inductor current caused by mode switching, thereby reducing the impact on the DC bus voltage and improving the stability and dynamic response speed of the voltage regulation control.

[0036] In one embodiment, the specific steps for calculating the threshold of the inductor current mode switching point in step S100 include: Determine the current mode type of the converter based on the voltage gain formula; the mode types include Buck1, Buck2, Boost1, Boost2, Buck-Boost1, and Buck-Boost2.

[0037] If the current mode is Buck1, Buck2, Boost1, or Boost2, determine the threshold of the switching point through the integral calculation formula.

[0038] If the current mode is Buck-Boost1 or Buck-Boost2, determine the threshold of the switching point through the piecewise linear interpolation formula.

[0039] Specifically, based on the input voltage , output voltage , inductor equivalent parameters and control period , through the voltage gain formula: [[ID=三十]]Determine the current mode type (Buck1, Buck2, Boost1, Boost2, Buck-Boost1, Buck-Boost2). If the current mode is Buck2 (0.5 < m < 1) or Boost2 (1 < m < 2), then through the integral calculation formula: ; Determine the switching point current, where is the instantaneous value of the voltage across inductor L1 at time t within the current mode; Indicates the inductor current threshold of the jth mode switching point in the current cycle k, represents the inductor current threshold of the previous mode switching point j-1 in the current cycle k, Indicates the real-time equivalent inductance value of inductor L1 at time t. If the current mode is Buck-Boost1 or Buck-Boost2 (corresponding Dead zone), then the piecewise linear interpolation formula is: ; Determine the switching point current, where is the dead zone transition coefficient; through this dynamic switching point calculation method, a smooth transition between different modes is achieved, avoiding the impact of the inductor current mutation caused by mode switching on the DC bus voltage.

[0040] In one embodiment, in step S200, the capacitor voltage regulation signals of the input side and the output side are calculated by the adaptive P controller respectively, including: inputting the real-time deviation of the flying capacitor voltage of the input side and the output side into the adaptive P controller of the input side and the output side respectively to generate the input side capacitor voltage regulation signal and output side capacitor voltage regulation signal , used to dynamically correct the inductor current reference instruction , the flying capacitor voltage on the input and output sides regulates the signal 、 The generation of is combined with the load power requirements, specifically including: Input side flying capacitor voltage regulation signal: ;in 、 is the proportional coefficient of the input-side P controller, Indicates the reference command value of the inductor current in the current cycle k.

[0041] Output side flying capacitor voltage regulation signal: ;in 、 is the proportional coefficient of the output side P controller, Indicates the reference value of the converter output voltage, What is calculated is the voltage regulation deviation of the converter output DC bus.

[0042] When the converter is in constant power charge and discharge mode (such as bidirectional energy storage scenario), it dynamically adjusts the power consumption according to the current power demand. 、 、 、 The coefficient value, for example, increases during high power charging and discharging. and The coefficient value is used to prioritize the balance of the flying capacitor voltage. In low power mode, the value is reduced. and The coefficient value is used to reduce the current loop regulation intensity. Through this power demand-driven regulation signal generation strategy, the flying capacitor voltage can respond quickly and the DC bus voltage can be stabilized under the charging and discharging power fluctuation scenario. The calculation formula of the flying capacitor voltage regulation signal on the input and output sides has the core function of dynamically maintaining the stability and balance of the voltage across the flying capacitor by accurately calculating and generating the corresponding regulation signal. Specifically, the input side formula targets the flying capacitor in the input link of the circuit, and the output side formula focuses on the flying capacitor in the output link. The two work together to ensure that under different working conditions, the flying capacitor voltage is always within the set target range, thereby providing stable voltage support for the entire circuit system, ensuring power conversion efficiency, suppressing harmonic interference, and avoiding device damage or circuit performance degradation due to capacitor voltage imbalance, ultimately achieving safe and efficient circuit operation.

[0043] In one embodiment, in step S300 , dynamically adjusting the prediction weight coefficients of the inductor current and the capacitor voltage in the prediction control model according to the estimation result includes: Calculate the equivalent inductance of the inductor and the external disturbance.

[0044] Calculate the load resistance.

[0045] Based on the equivalent inductance value, external disturbance value and load resistance, the prediction weight coefficient is dynamically adjusted. In this embodiment, based on the inductor current collected in step S100 , the adjustment signal generated in step S200 、 , and the flying capacitor voltages on the input and output sides 、 、 、 , build a predictive control model including a load parameter estimation module. The specific implementation of the load parameter estimation module includes: The inductor is analyzed by the extended state observer (ESO). Equivalent parameters of Perform real-time estimation, ESO uses inductor current , input voltage , output voltage As input, the equivalent inductance of inductor L1 and external disturbance are calculated by state equation At the same time, the voltage deviation of the flying capacitor and the inductor current The real-time relationship is combined with the least squares method or Kalman filter algorithm to estimate the load resistance. , estimated 、 and Used to dynamically adjust the weight coefficients of the inductor current deviation term and the capacitor voltage deviation term in the MPC model 、 , where the weight coefficient adjustment formula is: ;in 、 is the nominal value, 、 The dynamic weight adjustment is used to preset the adjustment coefficient to solve the prediction deviation problem caused by load parameter changes in the traditional MPC model and improve the adaptability of voltage stabilization control.

[0046] After completing the generation of the flying capacitor voltage regulation signals on the input side and the output side and the dynamic correction of the inductor current reference instruction in step S200, step S300 constructs an MPC model including a load parameter estimation module based on the corrected inductor current reference signal. At this time, the regulation signal generated in step S200 not only serves as the basis for correcting the current instruction, but the flying capacitor voltage deviation information and current deviation characteristics contained therein will also serve as an auxiliary input of the load parameter estimation module, providing a richer dynamic characteristic reference for the extended state observer, so that the estimation of the inductor equivalent parameters and the load resistance value is more in line with the actual working conditions, thereby ensuring that the MPC model can achieve reliable prediction of the future state of the system based on the real-time corrected reference instructions and the accurately estimated load parameters. In one embodiment, a prediction cost function including a prediction weight coefficient is constructed in step S400, and the calculation of the prediction cost function specifically includes: the inductor current deviation under the target mode , Input side flying capacitor voltage deviation , output side flying capacitor voltage deviation Perform weighted summation to form the total cost function: ; Input side flying capacitor voltage deviation: ; Output side flying capacitor voltage deviation: ;in, 、 、 The dynamic prediction weight coefficient is based on the load parameter estimated in step S300. 、 and external disturbance Make real-time adjustments.

[0047] For example, when the inductor Compared with nominal value When it is high, Increase to give priority to suppressing current deviation; when the load resistance When the fluctuation is large, or Increase to prioritize stabilizing the capacitor voltage; through the design of this multi-dimensional dynamic cost function, the MPC model can still prioritize responding to key state variables when parameters change, thereby reducing the sensitivity of voltage regulation control.

[0048] By constructing a prediction cost function containing multi-dimensional dynamic weights, the control effect of each preset mode under the current working conditions is quantitatively evaluated, providing a quantitative basis for the selection of the optimal mode. By dynamically adjusting the weight coefficient, the cost function can accurately reflect the control priority of the system under different working conditions, thereby screening out the optimal mode that can minimize the state variable deviation and adapt to the changes in load parameters, providing a reliable target mode benchmark for subsequent duty cycle calculations, and ensuring that the control strategy can still achieve stable control of the DC bus voltage under parameter fluctuations or disturbances.

[0049] In one embodiment, after determining the target modality based on the prediction cost function, the method further includes: If the target mode is Buck-Boost 1 or Buck-Boost 2, the symmetry deviation of the flying capacitor voltages on the input side and the output side is further determined.

[0050] If the symmetry deviation exceeds a preset threshold, symmetry compensation is introduced to calculate the target duty cycle. The calculation incorporates the flying capacitor voltage balance constraints, specifically: After selecting the target mode with the lowest cost at the current moment, if the target mode is Buck-Boost 1 or Buck-Boost 2, the symmetry deviation of the flying capacitor voltages on the input and output sides is further determined. The symmetry deviation of the input side transcapacitor voltage is: ; Symmetry deviation of the output side transcapacitor voltage: ; Determine whether the symmetry deviation exceeds the preset threshold If it exceeds, adjust the duty cycle calculation formula of the target mode and introduce the capacitor voltage symmetry compensation term: ; in is the calculated value of duty cycle without compensation, 、 Capacitor voltage symmetry compensation coefficient, if not exceeded, then output directly .

[0051] By using the above constraints, the midpoint voltage offset problem caused by capacitor voltage asymmetry is avoided, thereby indirectly improving the DC bus voltage. In order to ensure the stability and output waveform quality of the converter, the optimal switching mode is dynamically determined and the target duty cycle of the main control switch is accurately calculated, taking into account both the efficient operation and operational stability of the converter: on the one hand, by screening the mode with the lowest cost, it is ensured that the switching action logic meets the optimization target under the current working conditions, providing a basis for efficient energy conversion; on the other hand, the flying capacitor voltage balance constraint is introduced when calculating the duty cycle. By judging the voltage symmetry deviation and dynamically adjusting the calculation formula, the midpoint voltage offset caused by capacitor voltage asymmetry is effectively avoided, thereby indirectly ensuring the stability of the DC bus voltage and the output waveform quality, and ultimately achieving stable and reliable operation of the converter based on efficient operation.

[0052] In one embodiment, in step S500, the target duty cycle is input to the carrier stacking and carrier phase shift modulation unit to generate a control signal for each switch tube, and the control signal is input to the converter, which specifically includes: The target duty cycle will be calculated The input is sent to the carrier stacking and carrier phase shift modulation unit to generate control signals for each switch tube. The control signals are output to the input and output main control switches of the three-level cascade Buck-Boost converter through a fixed switching frequency modulation strategy (such as SVPWM) to achieve coordinated control of the inductor current and the flying capacitor voltage. At the same time, the model deviation caused by load parameter changes or component aging is compensated by dynamically adjusted prediction weight coefficients, thereby improving the DC bus voltage. voltage regulation accuracy and system robustness.

[0053] Furthermore, the specific implementation of the carrier stacking and carrier phase shift modulation unit includes: The target duty cycle will be dynamically calculated Decomposed into the input side Buck converter main control switch tube ( and ) duty cycle and the output side Boost converter main control tube ( and ) duty cycle ,in, ; The input side Buck converter and the output side Boost converter are in cascade relationship. When decomposing, the Buck stage on the input side can be made to work first to stabilize the voltage at an intermediate value, and then the Boost stage on the output side adjusts this intermediate value to the final target voltage, and presets a fixed input side Buck duty cycle, for example , and then according to the target duty cycle Calculated based on the above formula (1) Similarly, you can also preset a fixed output side Boost duty cycle, and then adjust the duty cycle according to the target duty cycle. Calculated based on the above formula (2) .

[0054] Through the cascade modulation strategy, and The phase difference between the input and output carrier signals is adjusted by phase shift modulation strategy. , in order to suppress the inductor current ripple caused by the dynamic change of duty cycle, the phase difference The adjustment formula is: ;

[0055] in is the ripple rejection coefficient, is the inductor current deviation at the current moment. This modulation strategy achieves low-loss control at a fixed switching frequency while reducing the impact of inductor current ripple on DC bus voltage regulation.

[0056] The working principle of this scheme is to achieve the stability of the DC bus voltage of the energy storage inverter through a closed-loop control logic with multiple links coordinated: first, the core state quantities such as the inductor current and the input / output side flying capacitor voltage are collected in real time, and the dynamic mode switching threshold is calculated to avoid the current mutation caused by mode switching; then, based on the flying capacitor voltage deviation and power demand, the adaptive P controller generates a regulation signal to correct the inductor current reference instruction and balance the capacitor voltage; then, an MPC model with a load parameter estimation module is constructed to estimate the inductor, load parameters and disturbances in real time and dynamically adjust the prediction weight to improve the model's adaptability to parameter changes; then, the prediction cost function of each preset mode is calculated, the optimal mode is screened, and the target duty cycle is dynamically calculated by combining hardware limitations and capacitor voltage balance constraints; finally, the switching signal is generated through carrier stacking and phase shift modulation to achieve coordinated control of the inductor current and flying capacitor voltage, compensate for parameter deviations, and ultimately ensure the stability of the DC bus voltage and the output waveform quality in scenarios such as charging and discharging power fluctuations and parameter changes. The present invention provides an energy storage system comprising a three-level cascaded Buck-Boost converter and a control unit configured to implement the control method provided in the present invention. In a specific embodiment, the system implements bidirectional energy conversion and voltage regulation between a battery and a DC bus for a bidirectional energy storage inverter within the energy storage system.

[0057] The specific requirements are as follows: Input: Battery pack (voltage range 400-600 VDC); Output: DC bus (target voltage 500 VDC, allowable ripple ≤ ±10 V); Load power: 0-5 kW (supports charge and discharge mode switching, power fluctuation rate ≤ 2 kW / 10 ms); Environment: Operating temperature -20°C to 60°C, with tolerance to device parameter drift (e.g., ±10% change in inductor value due to aging).

[0058] Based on the control method provided in this application, a verification test is carried out. The process is as follows: First, the dynamic mode switching point is calculated, and the state quantity is collected (at time k, i.e., the kth control cycle). The collected values ​​are: input voltage , output voltage , inductor current .

[0059] Input side capacitor voltage , ;Output side capacitor voltage , .

[0060] Performing modal judgment: voltage gain , it is determined to be Buck-Boost mode (corresponding to dead zone m≈1).

[0061] Switching point threshold calculation: Estimation of inductor equivalent value via extended state observer , dead zone transition coefficient .

[0062] Using piecewise linear interpolation formula: Substitute the data: .

[0063] Secondly, the flying capacitor voltage balance adjustment signal is generated. Power demand judgment: current load power (High power mode), the flying capacitor voltage must be balanced first.

[0064] Input side regulation signal calculation: input side voltage difference .

[0065] Inductor current reference command , current deviation .

[0066] In high power mode, the proportional coefficient (increase to quickly balance the capacitance), , reduce to reduce the current loop fluctuation: .

[0067] Output side regulation signal calculation: output side voltage difference .

[0068] Output voltage deviation (Target 500V).

[0069] In high power mode, , . . Corrected inductor current reference instruction: After correction (After unit conversion) Secondly, load parameters are estimated and MPC weights are adjusted.

[0070] Parameter Estimation: Extending State Observer Inputs 、 、 , the estimated external disturbance .

[0071] Least Squares Method for Estimating Load Resistance (Nominal value , ).

[0072] Weight coefficient calculation: inductance deviation , preset adjustment coefficient .

[0073] .

[0074] Load resistance has no deviation, preset , ; Output side capacitor voltage weight (fixed).

[0075] Secondly, the prediction cost function is calculated. For the six preset modes (Buck1 / 2, Boost1 / 2, Buck-Boost1 / 2), the cost function is calculated: .

[0076] After comparison, Buck-Boost 1 mode was found to have the lowest cost and was selected as the target mode. In the DC voltage regulation control method for energy storage inverters based on a three-level Buck-Boost system with flying capacitors and inductor reuse, selecting the lowest-cost mode as the target mode is a key step in achieving precise regulation within the entire control logic. This process is based on real-time data collected by the system, including the inductor current, input / output flying capacitor voltages, and input / output voltages. The system first dynamically calculates the modal switching point threshold to determine the potential operating mode range under the current operating conditions. Then, combining the flying capacitor voltage deviation with the power demand, it generates a regulation signal to correct the inductor current reference command, ensuring that the capacitor voltage balance requirement is factored into the control. Subsequently, using an extended state observer and online parameter identification technology, key parameters such as the inductor and load, as well as external disturbances, are estimated in real time. The weight coefficients of the inductor current deviation and capacitor voltage deviation terms in the model predictive control are dynamically adjusted, ensuring that the cost function accurately reflects the control priority under different operating conditions.

[0077] Finally, by calculating the prediction cost function of each preset mode (Buck1, Buck2, Boost1, Boost2, Buck-Boost1, Buck-Boost2), the mode with the lowest cost is selected as the target mode.

[0078] This technical solution significantly improves DC voltage regulation accuracy and system robustness. When switching between different operating modes, rather than using traditional fixed thresholds or simplified circuit model calculations, this solution combines real-time data such as input and output voltages, inductor equivalent parameters, and control cycles, dynamically determining the inductor current switching point using integral calculations or piecewise linear interpolation. This method achieves smooth transitions between different modes, effectively suppressing sudden changes in inductor current caused by mode switching, thereby reducing the impact on the DC bus voltage and improving the smoothness and dynamic response speed of voltage regulation control.

[0079] This technical solution is superior to existing technologies in terms of flying capacitor voltage balance and load fluctuation adaptability. The key lies in the use of a P controller regulation signal generation strategy based on power demand adaptive adjustment. Unlike the traditional fixed proportional coefficient capacitor voltage balance control, this solution dynamically adjusts the proportional coefficient of the P controller according to the real-time changes in charge and discharge power, giving priority to enhancing the capacitor voltage balancing ability in high power mode, and reducing the current loop regulation strength in low power mode to reduce fluctuations. This power-driven coefficient adjustment mechanism enables the system to maintain rapid response and balance of the flying capacitor voltage in scenarios with rapid charge and discharge power fluctuations, while maintaining the stability of the DC bus voltage, taking into account both steady-state performance and dynamic adaptability.

[0080] In the model predictive control section, this solution introduces a real-time load parameter estimation mechanism using an extended state observer combined with the least squares method or Kalman filter. This not only estimates the inductor equivalent value and external disturbances, but also obtains the load resistance in real time and dynamically adjusts the MPC weight coefficient based on parameter changes. Unlike existing technologies that assume constant circuit parameters and reduce prediction accuracy due to parameter drift, this solution uses online adjustment of weights such as w1, w2, and w3 to enable the MPC to maintain the accuracy of the prediction model when the inductance, load resistance, and external disturbances change. This technical approach significantly improves the system's adaptability and control robustness in the face of component aging, sudden load changes, and other conditions, thereby achieving higher-precision DC voltage regulation control.

[0081] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A control method based on a three-level cascaded Buck-Boost converter, wherein the converter includes an input side and an output side, an inductor is located between the input side and the output side, and the inductor serves as a multiplexed inductor for storing or releasing energy in different operating states of the converter; It is characterized by: The control method includes: collecting the current of the inductor, the voltage of the flying capacitor on the input side, and the voltage of the flying capacitor on the output side in real time, and calculating the threshold of the inductor current modal switching point in combination with the input voltage, output voltage, control period, and parameters of the inductor of the converter; Based on the real-time voltage deviations of the flying capacitor on the input side and the flying capacitor on the output side, respectively, the capacitor voltage adjustment signals on the input side and the output side are calculated by an adaptive P controller to dynamically correct the inductor current reference signal; Constructing a predictive control model including load parameters, the predictive control model is used to estimate the equivalent parameter changes of the inductor in real time, and dynamically adjust the prediction weight coefficients of the inductor current and capacitor voltage in the predictive control model according to the estimation results; Constructing a prediction cost function including the prediction weight coefficient, determining a target mode based on the prediction cost function, and calculating a target duty cycle of the switching tube of the converter based on the switching tube action logic of the target mode and the load parameter; The target duty cycle is input into the carrier stacking and carrier phase shift modulation unit to generate a control signal for each switch tube, and the control signal is input into the converter.

2. The control method according to claim 1, wherein: The input side includes a flying capacitor and , switch tube 、 、 and The input side bridge arm, switch tube and For the main control switch tube, switch tube and For auxiliary switch tube; flying capacitor and The series midpoint of the circuit serves as the midpoint potential of the input side; The output side includes a flying capacitor and , switch tube 、 、 and The output side bridge arm, switch tube and For the main control switch tube, switch tube and For auxiliary switch tube; flying capacitor and The series midpoint of the circuit serves as the midpoint potential of the output side; The inductor L1 is connected between the input side and the output side.

3. The control method according to claim 1, wherein: Calculating the threshold of the inductor current mode switching point includes: Determining a mode type of the current converter based on a voltage gain formula; the mode types include Buck1, Buck2, Boost1, Boost2, Buck-Boost1, and Buck-Boost2; If the current mode is Buck1, Buck2, Boost1 or Boost2, the threshold of the switching point is determined by an integral calculation formula; If the current mode is Buck-Boost 1 or Buck-Boost 2, the threshold of the switching point is determined by a piecewise linear interpolation formula.

4. The control method according to claim 3, wherein: The integral calculation formula is: ;in is the instantaneous value of the voltage across the inductor L1 at time t in the current mode; Indicates the inductor current threshold of the jth mode switching point in the current cycle k, represents the inductor current threshold of the previous mode switching point j-1 in the current cycle k, Indicates the real-time equivalent inductance value of inductor L1 at time t; The piecewise linear interpolation formula: 。 5. The control method according to claim 1, wherein: The calculating the capacitor voltage regulation signals of the input side and the output side respectively by the adaptive P controller includes: Dynamically adjusting the capacitor voltage regulation signals on the input side and the output side by dynamically adjusting the proportional coefficient of the adaptive P controller according to the power demand of the converter at the current moment; The capacitor voltage regulation signal on the input side is calculated by the following formula: ;in 、 is the proportional coefficient of the input-side P controller, Indicates the reference command value of the inductor current in the current cycle k; The capacitor voltage regulation signal on the output side is calculated by the following formula: ； in 、 is the proportional coefficient of the output side P controller, The reference value of the converter output voltage is shown. What is calculated is the voltage regulation deviation of the converter output DC bus.

6. The control method according to claim 1, wherein: The dynamically adjusting the prediction weight coefficients of the inductor current and the capacitor voltage in the prediction control model according to the estimation result includes: Calculating the equivalent inductance value and the external disturbance amount of the inductor; calculating the load resistance; and dynamically adjusting the prediction weight coefficient based on the equivalent inductance value, the external disturbance amount, and the load resistance.

7. The control method according to claim 1, wherein: The construction of the prediction cost function including the prediction weight coefficient includes: calculating the inductor current deviation under the target mode , the flying capacitor voltage deviation on the input side , the flying capacitor voltage deviation on the output side Perform weighted summation; the calculation formula of the prediction cost function is: ; in, 、 、 is the prediction weight coefficient.

8. The control method according to claim 1, wherein: After determining the target modality based on the prediction cost function, the method further includes: If the target mode is Buck-Boost 1 or Buck-Boost 2, further determining the symmetry deviation of the flying capacitor voltages on the input side and the output side; If the symmetry deviation exceeds a preset threshold, symmetry compensation is introduced to calculate the target duty cycle.

9. The control method according to claim 8, wherein: The target duty cycle is calculated by the following formula: ; in, is the duty cycle without introducing symmetry compensation, 、 is the capacitor voltage symmetry compensation coefficient, 、 is the symmetry deviation of the flying capacitor voltages at the input side and the output side.

10. An energy storage system, characterized in that: The invention comprises a three-level cascade Buck-Boost converter and a control unit, wherein the control unit is used to implement the control method according to any one of claims 1 to 9.

Images