Phase-shifted full-bridge model predictive current control method based on predictive error compensation

Abstract

The invention provides a phase-shifted full-bridge model predictive current control method based on predictive error compensation, relates to the technical field of power electronics, and effectively compensates errors in a phase-shifted full-bridge converter system by constructing a predictive error recursive relation from a moment k to a moment k + 1 and utilizing a correction coefficient of a threshold judgment mechanism. On the premise that calculation complexity is not remarkably increased, current prediction deviation caused by factors such as inductance parameter change, sampling errors, discretization errors and inaccurate duty ratio loss compensation can be effectively restrained, and robustness and control precision of a system under parameter disturbance and load change are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power electronics, and more specifically, to a phase-shifted full-bridge model predictive current control method based on prediction error compensation. BACKGROUND

[0002] At present, as a classic DC-DC topology structure, the phase-shifted full-bridge (PSFB) converter is widely used in communication base station power supply, server power supply, industrial drive and new energy power generation system and other fields, because it can realize zero voltage switching (ZVS) of the switching tube, reduce switching loss, improve efficiency, and is suitable for medium and high power application occasions. Its advantages include high voltage conversion ratio, good soft switching characteristics and relatively simple control logic. However, the PSFB converter still faces challenges in control accuracy, dynamic response and robustness in actual operation.

[0003] In terms of control strategy, although the traditional linear control method such as PID control is simple and easy to implement, it has slow dynamic response and limited anti-disturbance ability when dealing with nonlinear and variable parameter systems. In order to improve the system performance, model predictive control (MPC) is gradually applied in power electronic converters. MPC predicts future behavior by using the discrete model of the system, and calculates the control quantity in the form of optimizing the objective function, which has the advantages of fast dynamic response, easy to handle multi-variable and nonlinear constraints. Especially in current control, model predictive current control (MPCC) can realize high-precision current tracking and significantly improve the transient performance of the system.

[0004] Although MPC has many advantages in theory, its actual control performance is highly dependent on the accuracy of the model. The prediction model of the PSFB converter usually depends on the circuit parameters (such as filter inductance, resonance inductance, etc.), and these parameters will drift due to factors such as temperature change, magnetic saturation, aging, etc. in actual work, causing the model to mismatch with the actual object, and then causing prediction error. In addition, the discretization error, sampling delay and quantization error existing in the digital control system, and the duty cycle loss phenomenon unique to PSFB, will further exacerbate the deviation between the predicted value and the actual value. If these errors are not compensated, it will directly lead to the decline of system performance, such as increasing steady-state error, dynamic response oscillation and even system instability.

[0005] In order to improve the robustness of MPC, in recent years, scholars have proposed a variety of methods. For example, some studies use online parameter identification strategies to reduce the mismatch error by updating the model parameters in real time, but this method usually needs to introduce additional observers or identification algorithms, which increases the computational complexity and implementation difficulty of the system. Some studies try to optimize the control law or improve the design of the cost function to enhance the adaptability of the system to parameter changes, such as generalized predictive control, robust predictive control, etc., but these methods often sacrifice response speed or computational efficiency. In addition, some work is committed to improving the sampling and modulation strategy to reduce the impact of discretization and delay, but this method is mostly limited to specific working conditions, and the universality is strong and the comprehensive improvement effect is limited. SUMMARY

[0006] In order to solve the above technical problems, the present application provides a phase-shifted full-bridge model predictive current control method based on prediction error compensation.

[0007] According to one aspect of the present application, a phase-shifted full-bridge model predictive current control method based on prediction error compensation is provided, comprising: Based on the topology structure of the Buck converter, an equivalent circuit model of the phase-shifted full-bridge converter is established, and a prediction control model is established based on the equivalent circuit model; Based on the equivalent circuit model and the prediction control model of the phase-shifted full-bridge converter, the inductor current and the source of system modeling error are analyzed; The prediction error recursive relationship from time k to time k+1 is constructed, and the error in the phase-shifted full-bridge converter system is effectively compensated by using the correction coefficient of the threshold value judgment mechanism; The prediction error compensation correction strategy is introduced into the closed-loop control process of the phase-shifted full-bridge converter, and the duty cycle design is optimized to realize accurate control and dynamic performance improvement.

[0008] Further, the equivalent circuit model of the phase-shifted full-bridge converter is established by regarding the phase-shifted full-bridge converter as a topological derivative structure of the Buck converter, and a second-order equivalent circuit model containing a resonant inductor and a filter inductor is established. By introducing the equivalent input voltage on the secondary side, the MOS tube voltage drop and the resonant inductor voltage drop parameters, a complete equivalent circuit dynamic model is constructed.

[0009] Further, the prediction control model includes a k+1 time inductor current prediction equation based on the current sampling value, a cost function, and an effective duty cycle optimization expression designed according to the inductor current reference value; By establishing the duty loss calculation model when zero voltage switching, combining the period average and forward difference operation of the inductance current expression, the inductance current prediction equation based on the current sampling value is derived, and the corresponding cost function and effective duty expression are designed, thereby the complete prediction control model is constructed.

[0010] Further, the analysis of the inductance current and system modeling error sources is through the discrete period subdivision and the forward Euler method to establish the continuous domain model; the mathematical expressions of various errors are derived, and they are decomposed into two parts of stable error and change error; The stable error includes the inductance parameter mismatch error related to the system output voltage, the fixed deviation between the inductance current prediction value and the true value, the sampling error caused by the steady-state bias of the input and output voltages, and the duty compensation error between the approximate compensation value and the accurate compensation value. The change error includes the inductance parameter mismatch error related to the input voltage and the output duty, the discrete model prediction deviation varying with the system state, and the sampling error varying with the duty.

[0011] Further, the sampling error varying with the duty is expressed as: Wherein, , and represent the sampling errors of the inductance current, the input voltage and the output voltage at k time, respectively. and The prediction error caused by the stable bias deviation of is defined as the stable sampling error .

[0012] Further, the calculation of the correction coefficient includes the following steps: The total prediction error at k time is decomposed into the stable error and the change error . The prediction error transfer equation from k time to k+1 time is established; based on the characteristics of shorter control period, the equivalent relationship of the stable error and in adjacent control periods is determined; the recursive relationship of the change error to is derived; The correction coefficient is introduced to represent the current prediction error; based on the characteristics of shorter control period, the characteristics that the correction coefficient remains constant in adjacent periods is determined ; the recursive expression of the current prediction error at k+1 time is derived; Utilizing k-time prediction error And the duty cycle difference value of adjacent two periods calculates correction coefficient .

[0013] Further, when the duty cycle difference value is greater than a preset threshold, a new correction coefficient is calculated; when the duty cycle difference value is less than the preset threshold, the last period correction coefficient is maintained; the obtained compensation value is added to the prediction current model for correction, so that the system prediction error is compensated.

[0014] According to another aspect of the present application, a phase-shifted full-bridge model prediction current control system based on prediction error compensation is provided, comprising: An equivalent circuit model construction module is configured to establish an equivalent circuit model of the phase-shifted full-bridge converter based on the topology of the Buck converter, and establish a prediction control model based on the equivalent circuit model; An error analysis module is configured to analyze the inductor current and the source of system modeling error based on the equivalent circuit model of the phase-shifted full-bridge converter and the prediction control model; An error compensation module is configured to construct a prediction error recursive relationship from k-time to k+1-time and utilize a correction coefficient of a threshold value judgment mechanism to effectively compensate the error in the phase-shifted full-bridge converter system; A control module is configured to introduce the prediction error compensation correction strategy into the closed-loop control process of the phase-shifted full-bridge converter, and optimize the duty cycle design to achieve accurate control and dynamic performance improvement.

[0015] Compared with the prior art, the phase-shifted full-bridge model prediction current control method based on prediction error compensation provided by the present application can effectively suppress the current prediction deviation caused by factors such as inductor parameter variation, sampling error, discretization error, and inaccurate duty cycle loss compensation, and improve the robustness and control accuracy of the system under parameter disturbance and load variation, without significantly increasing the computational complexity. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor. In the drawings: Figure 1 The equivalent phase-shifted full-bridge circuit model of the phase-shifted full-bridge model prediction current control method based on prediction error compensation according to the embodiments of the present application.

[0017] Figure 2 This is a flowchart illustrating the calculation of prediction error in the phase-shifted full-bridge model prediction current control method based on prediction error compensation according to an embodiment of the present invention.

[0018] Figure 3 The following is a system control block diagram after prediction error compensation correction in the phase-shifted full-bridge model prediction current control method based on prediction error compensation according to an embodiment of the present invention.

[0019] Figure 4 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention.

[0020] Figure 5 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention.

[0021] Figure 6 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention.

[0022] Figure 7 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention.

[0023] Figure 8 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention.

[0024] Figure 9 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention.

[0025] Figure 10 The flowchart shows the improved predictive current control subroutine of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention. Detailed Implementation

[0026] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0027] Figure 1 This is an equivalent phase-shifted full-bridge circuit model of the phase-shifted full-bridge model predictive current control method based on prediction error compensation according to an embodiment of the present invention. For example...Figure 1 As shown, the phase-shifted full-bridge model predictive current control method based on prediction error compensation includes: S1: Establish the equivalent circuit model of the phase-shifted full-bridge converter; derive the inductor current expression based on the Buck converter principle, which considers the equivalent input voltage on the secondary side of the converter, the voltage drop of the MOSFET and the voltage drop of the resonant inductor; establish a duty cycle loss calculation model during zero-voltage switching; obtain the predicted value of the inductor current at time k+1 through periodic averaging and forward differential operation. A design platform is built, the MATLAB / Simulink software platform is configured in the computer system, and subsequent design work is based on the MATLAB / Simulink software platform.

[0028] In terms of system modeling, this invention is based on the Buck converter and deeply analyzes the operating characteristics of phase-shifted full-bridge converters. The phase-shifted full-bridge converter, as a topology derivative of the Buck converter, has the following equivalent circuit model: Figure 1 As shown. Based on this equivalent model, this invention derives a complete expression for the inductor current over one switching cycle. This expression comprehensively considers the influence of multiple key parameters such as the equivalent input voltage on the secondary side of the converter, the MOSFET voltage drop, and the resonant inductor voltage drop.

[0029] Based on the equivalent model of the phase-shifted full-bridge converter, the inductor current over one cycle is expressed as: (1) in: , representing the equivalent input voltage on the secondary side of the converter obtained after calculation; V in , V MOS , V Lr These represent the converter input voltage, MOSFET voltage drop, and resonant inductor Lr voltage drop, respectively. , represents the equivalent inductance of the resonant inductance Lr and the filter inductance Lf referred to the secondary side; V on This represents the equivalent inductance voltage when the MOSFET is turned on; V off This represents the equivalent inductance voltage when the MOSFET is turned off. S2: Design the system control objective and cost function; derive the effective duty cycle expression based on the inductor current reference value; establish a predictive control model to achieve dynamic response control of the system; In the analysis of zero-voltage switching characteristics, this invention establishes an approximate calculation model for duty cycle loss when a phase-shifted full-bridge achieves ZVS. By performing periodic averaging and forward differential operations on the inductor current expression, the predicted value of the inductor current at the next moment is derived. This predicted value is based on the sampled value of the inductor current at the current moment. The PFSB loses duty cycle when achieving ZVS. The approximate value is expressed as: (2) By performing periodic averaging and forward differencing on equation (1), the predicted inductor current at time k+1 can be derived. , is represented as: (3) in, Let be the sampled value of the inductor current at time k, and let the system control objective be... .

[0030] To achieve precise system control, this invention designs corresponding control objectives and cost functions. By optimizing the effective duty cycle expression, it ensures that the system can accurately track the inductor current reference value. This predictive control method effectively improves the system's dynamic response performance and control accuracy. The cost function is designed as follows: (4) To accurately track the inductor current reference value Effective duty cycle Represented as: (5) S3: Perform error analysis steps: Analyze the four sources of system prediction error, including circuit inductance parameter mismatch, discrete model error, sampling error, and duty cycle compensation error; establish the prediction error expression caused by inductance parameter changes; establish a continuous domain model through discrete period subdivision and forward Euler method; derive the mathematical expressions of various errors and decompose them into two parts: stable error and variable error; There are four main factors that cause prediction errors: first, mismatch in circuit inductance parameters; second, error between the discrete model and the actual model of the system; third, sampling error in the system; and fourth, inaccurate compensation for lost duty cycle.

[0031] The first type of error originates from the mismatch of circuit inductance parameters. There is a deviation between the actual inductance value and the set value in the controller; this deviation includes inherent setting errors and dynamic changes during converter operation. Based on the inductor current prediction equation, an expression for the prediction error caused by changes in inductance parameters is derived. This error can be decomposed into two parts: a stable error related only to the system output voltage, and a real-time variation error related to the input voltage and output duty cycle.

[0032] The second type of error arises from the discrepancy between the system's discrete model and the actual model. By subdividing the discrete period into M equal parts and using the forward Euler method for discretization, a transition model from the discrete domain to the continuous domain is established. Through rigorous derivation, an accurate expression for the actual value of the inductor current is obtained, which includes the average value of the equivalent input voltage and output voltage. This yields the prediction error caused by model discretization, which can also be decomposed into two components: a stable error and a variable error.

[0033] The third type of error originates from the system sampling process. This includes sampling errors for inductor current, input voltage, and output voltage. Among these, the steady-state bias of the input and output voltages leads to stable prediction errors, while the portion that varies with the duty cycle generates dynamic prediction bias.

[0034] The fourth type of error arises from the inaccuracy of lost duty cycle compensation. When the control cycle is short, the prediction deviation caused by the error between the approximate compensation value and the accurate compensation value can be regarded as a stable deviation.

[0035] Let the actual value of the inductance L be ,in This is the set value of the inductor in the controller. This represents the change in inductance during converter operation. Combined with equation (3), this represents the inductor current prediction error at time k+1 caused by the change in inductor parameters. It can be represented as: (6) Depending on the nature of the error, it can be... It is divided into two parts, as follows: (7) Given that the circuit parameters are fixed It is only related to the system output voltage and represents the stability error. In addition to being related to the system input voltage, it is also related to the output duty cycle, representing the real-time variation error.

[0036] Discrete period T S The system is divided into M equal parts, and the forward Euler method is used to discretize the system in the form shown in equation (3). When M is large enough, the system is approximately considered to be continuous. The discretization process is shown in equation (8).

[0037] (8) Adding the equations above together, we get: (9) The duty cycle determines the switching time and state of the converter switch within a control cycle, and the duty cycle remains unchanged between time k and k+1. (10) Substituting the above equation into equation (8), the original equation is rewritten as: (11) As M approaches infinity, equation (11) can be solved in the continuous domain, and the actual value of the inductor current at time k+1 can be obtained after solving. It can be represented as: (12) In the formula and These represent the average values ​​of the equivalent input voltage and output voltage at times k to k+1, respectively.

[0038] Let the error between the predicted and actual inductor current at time k+1 caused by fluctuations in the model discretization parameters be denoted as . , is represented as: (13) Will It can also be classified as steady error and variation error Two parts, represented as: (14) (15) The current prediction error caused by sampling error is expressed as: (16) in, , and These represent the sampling errors of the inductor current, input voltage, and output voltage at time k, respectively. and The prediction error caused by the stable bias deviation is defined as the stable sampling error. The prediction bias that varies with duty cycle is defined as... .

[0039] When using an approximate lost duty cycle value for compensation, there is an error between the approximate compensation value and the precise compensation value. , This will generate current prediction error. When the control cycle is short, the prediction deviation caused by this error can be regarded as a steady-state deviation. It is expressed as: (17) S4: Implement the prediction error compensation steps: derive the prediction error at time k+1 based on the prediction error at time k; establish a recursive expression for the prediction error; design a method for calculating the error correction coefficient; introduce a threshold judgment mechanism to solve the problem of calculating the correction coefficient in steady state; integrate the compensation strategy into the system control framework to achieve closed-loop control.

[0040] If we ignore the cross-order errors of the above errors, the total prediction error at time k is... It can be represented as: (18) Based on the different error categories, it will be divided into two parts: (19) in: Represents the stationary error at time k. This represents the error in time k.

[0041] To address the characteristics of the aforementioned prediction errors, this invention proposes a method for prediction error compensation and correction. The main principle is to predict the prediction error at time k+1 based on the prediction error at time k, and then incorporate this error into the system's predicted current model to compensate and correct the predicted current value at time k+1, thereby eliminating current prediction deviations and improving system robustness.

[0042] The prediction error expressions for time k and time k+1 are shown in the following equations: (20) Combining equations (7), (14), (15), (16), and (17), the constant errors of the system at times k and k+1 can be expressed as follows: (twenty one) In the formula , .

[0043] The value of K1 is mainly determined by the real-time inductance parameters in the circuit, while K2 is the initial setting parameter of the inductor, which is a constant. When the control cycle of the system is short, the real-time parameters of the inductor can be considered to remain unchanged over several control cycles.

[0044] In terms of stability error analysis, considering the short control cycle of the system, the inductance parameter can be considered a constant value within adjacent control cycles. Simultaneously, since sampling error and lost duty cycle compensation error occur in the feedback loop of the control channel, these errors can also be considered constant within adjacent cycles. Based on these characteristics, the equivalent relationship of stability errors within two adjacent control cycles is derived.

[0045] In practice, the sampling error and lost duty cycle compensation error of the system occur in the feedback loop of the control channel. The system cannot correct or eliminate them, and they can be considered constant in adjacent control cycles.

[0046] In conclusion, it can be considered that and The values ​​within two adjacent control cycles can be considered equal, that is: (twenty two) Equation (20) is transformed into: (twenty three) From equation (18), the system's variation errors at times k and k+1 are respectively: (twenty four) , representing the correction coefficient for current prediction error.

[0047] When the system control cycle is short The value of is basically constant within adjacent control cycles, and we have: (25) Combining equations (23), (24), and (25), we can obtain: (26) Equation (26) represents the current prediction error at time k+1. The recursive expression.

[0048] The prediction error at time k and the duty cycle of the first two periods. and The difference is derived, and by combining equations (26) and (3), the corrected result can be obtained. for: (27) Equation (27) shows that this prediction deviation compensation method can offset the prediction current deviation caused by the sampling error of the system.

[0049] Equation (26) can be transformed into an expression for M(k) as follows: (28) The results are based on the assumption that the input voltage, inductance changes, and sampling error remain relatively constant within a short control cycle. Furthermore, the parameters required for the next prediction cycle can be calculated based on the state variables from the previous control cycle. value: (29) When the system is in steady state, the duty cycle values ​​of two adjacent cycles may be the same or very close. At this time, the denominator (29) will approach zero, and the value of M(k) will approach infinity, causing the feedback to lose its meaning. Considering the short control cycle of the system, The values ​​used in the previous cycle can be reused. Meanwhile, the expression in the denominator... A threshold is defined; the value is only adjusted when the difference in duty cycles exceeds the threshold. Perform an update calculation; otherwise, continue using the calculations from the previous time step. The control block diagram for calculating the prediction error is shown below. Figure 2 As shown.

[0050] Closed-loop control The prediction error compensation and correction strategy obtained in the above steps is introduced into the control flow of the phase-shifted full-bridge converter system to complete the closed-loop control of the converter.

[0051] It should be noted that this invention uses a simulation platform built in the MATLAB / Simulink environment to simulate and analyze the correction effect of the error correction strategy. Since the value of the resonant inductor after being referred to the secondary side is extremely small compared to the filter inductor, its change is almost negligible. Therefore, this section mainly simulates and analyzes the current and voltage changes when the filter inductor parameters vary within the range of 0.5L0 to 1.5L0, with and without the addition of a compensation correction stage.

[0052] (1) Simulation results when inductance parameters suddenly decrease At t=9ms, the filter inductor parameter suddenly drops to 0.5L0. The output voltage changes with and without the deviation correction circuit are as follows: Figure 5 As shown. Figure 5 (a) and Figure 5 The results in (b) show that when the inductance parameter suddenly decreases to 0.5L0, the output voltage without the deviation correction circuit drops by about 0.2V, and returns to the reference value after 1.4ms of adjustment by the outer voltage loop. After adding the compensation correction circuit, the output voltage fluctuation is very small and can quickly adjust back to the reference value. Due to the decrease in inductance, the voltage ripple increases in both control methods.

[0053] Figure 6 The changes in inductor current under the same conditions described above are listed. From Figure 6 (a) It can be seen that even without inductor parameter mismatch, there is still a certain deviation between the average inductor current and the reference value. This is because the loss of duty cycle compensation uses a steady-state loss of duty cycle approximation compensation, which has a very small deviation and relatively stable error. After adjustment by the voltage outer loop, the inductor current can still output stably. (See Figure 6(a) and...) Figure 6The results in (b) show that when the filter inductor parameter suddenly decreases to 0.5L0, the inductor current without the deviation correction stage stabilizes again after being slowly adjusted by the voltage outer loop, but the deviation from the current reference value further increases. After adding the correction stage, the inductor current can quickly reach a stable output, and the deviation from the reference value is eliminated, achieving accurate tracking.

[0054] Figure 7 A magnified view of the inductor current waveform is provided. From Figure 7 (a) and Figure 7 (b) It can be seen that when the inductance parameter suddenly decreases to 0.5L0, there is a prediction deviation of 0.18A between the average inductance current without the correction circuit and the reference value. After the correction circuit is added, this deviation is eliminated.

[0055] (2) Simulation results when inductance parameters suddenly increase At t=9ms, the filter inductor parameter suddenly increases to 1.5L0. The output voltage changes with and without the deviation correction circuit are as follows: Figure 8 As shown. Figure 8 (a) and Figure 8 The results in (b) show that when the inductance parameter suddenly increases to 1.5L0, the output voltage without the deviation correction stage fluctuates significantly, but returns to the reference value after 2.3ms of adjustment by the outer voltage loop. After adding the compensation correction stage, the output voltage fluctuation is very small and can quickly adjust back to the reference value. Due to the increase in inductance, the voltage ripple is reduced under both control conditions.

[0056] Figure 9 The changes in inductor current under the same conditions described above are listed. Combined with... Figure 9 As can be seen from the results in (a) and Figure 9(b), when the filter inductor parameter suddenly increases to 1.5L0, the inductor current without the deviation correction circuit fluctuates greatly. After the voltage outer loop is slowly adjusted, the output stabilizes again, but a deviation occurs between it and the current reference value. After the correction circuit is added, the inductor current can quickly reach a stable state and the error between it and the reference value is eliminated.

[0057] Figure 10 The magnified details of the inductor current waveform are shown. Combined with... Figure 10 (a) and Figure 10 (b) It can be seen that when the inductance parameter suddenly increases to 1.5L0, there is a prediction deviation of 0.06A between the average inductance current without the correction circuit and the reference value; after the correction circuit is added, the deviation is eliminated.

[0058] In summary, the current prediction control method based on prediction error compensation for a phase-shifted full-bridge model, as described in this invention, is explained. By constructing a recursive relationship of prediction errors from time k to time k+1 and utilizing the correction coefficient of the threshold judgment mechanism, the method effectively compensates for errors in the phase-shifted full-bridge converter system. Without significantly increasing computational complexity, it effectively suppresses current prediction deviations caused by factors such as changes in inductor parameters, sampling errors, discretization errors, and inaccurate compensation for lost duty cycles, thereby improving the robustness and control accuracy of the system under parameter disturbances and load changes.

[0059] Here, those skilled in the art will understand that the specific operations of each step in the above-described phase-shifted full-bridge model predictive current control system based on prediction error compensation have been referenced above. Figures 1 to 10 The predictive current control method based on the phase-shifted full-bridge model with prediction error compensation has been described in detail, and therefore, its repeated description will be omitted.

[0060] In summary, the phase-shifted full-bridge model predictive current control system based on prediction error compensation based on the embodiments of the present invention has been clarified. It effectively compensates for the error in the phase-shifted full-bridge converter system by constructing a recursive relationship of prediction error from time k to time k+1 and using the correction coefficient of the threshold judgment mechanism. Without significantly increasing the computational complexity, it can effectively suppress the current prediction deviation caused by factors such as changes in inductor parameters, sampling errors, discretization errors, and inaccurate compensation for duty cycle loss, thereby improving the robustness and control accuracy of the system under parameter disturbances and load changes.

Claims

1. A phase-shifted full-bridge model predictive current control method based on prediction error compensation, characterized in that, The application relates to a control method and device for a phase-shifted full-bridge converter. The application comprises: establishing an equivalent circuit model of the phase-shifted full-bridge converter based on a topology structure of a Buck converter, and establishing a predictive control model based on the equivalent circuit model; analyzing an inductor current and a system modeling error source based on the equivalent circuit model and the predictive control model of the phase-shifted full-bridge converter; constructing a prediction error recursive relationship from k time to k+1 time and using a correction coefficient of a threshold judgment mechanism to effectively compensate errors in the phase-shifted full-bridge converter system; 2. The phase-shifted full-bridge model predictive current control method based on prediction error compensation according to claim 1, characterized in that, introducing the prediction error compensation correction strategy into a closed-loop control process of the phase-shifted full-bridge converter, optimizing a duty cycle design to realize precise control and dynamic performance improvement.

3. The phase-shifted full-bridge model predictive current control method based on prediction error compensation according to claim 2, characterized in that, The equivalent circuit model of the phase-shifted full-bridge converter is established by regarding the phase-shifted full-bridge converter as a topology derivative structure of the Buck converter, a second-order equivalent circuit model containing a resonant inductor and a filter inductor is established, and a complete equivalent circuit dynamic model is constructed by introducing an equivalent input voltage on the secondary side, a MOS tube voltage drop and a resonant inductor voltage drop parameter. The predictive control model comprises an inductor current prediction equation at k+1 time based on a current sampling value, a cost function and an effective duty cycle optimization expression designed according to an inductor current reference value; 4. The phase-shifted full-bridge model predictive current control method based on prediction error compensation according to claim 3, characterized in that, by establishing a duty cycle loss calculation model when the zero voltage switch is turned on, combining the period average of the inductor current expression and the forward difference operation, the inductor current prediction equation based on the current sampling value is derived, and the corresponding cost function and effective duty cycle expression are designed, thereby constructing the complete predictive control model. The analysis of the inductor current and the system modeling error source is established by discrete period subdivision and forward Euler method; the mathematical expressions of various errors are derived, and they are divided into stable error and change error; The stable error comprises an inductor parameter mismatch error related to the system output voltage, a fixed deviation between the inductor current prediction value and the true value, a sampling error caused by the steady-state bias of the input and output voltages and a duty cycle compensation error between the approximate compensation value and the accurate compensation value; 5. The phase-shifted full-bridge model predictive current control method based on prediction error compensation according to claim 4, characterized in that, The change error comprises an inductor parameter mismatch error related to the input voltage and the output duty cycle, a discrete model prediction deviation changing with the system state and a sampling error changing with the duty cycle. wherein, , and denote the sampling errors of inductor current, input voltage and output voltage at time k, respectively. and The prediction error caused by the stable bias deviation of and is defined as the stable sampling error The prediction error that varies with the duty cycle is defined as .

6. The phase-shifted full-bridge model predictive current control method based on prediction error compensation according to claim 5, characterized in that, The sampling error changing with the duty cycle is expressed as: The total prediction error at time k is decomposed into a stable error and a varying error ;​ Establish the prediction error propagation equation from time k to time k+1; based on the short control period, determine the stability error within adjacent control periods. and The equivalent relationship; deriving the change error arrive The recursive relationship; Introducing correction factor Characterizing current prediction error; based on the feature that control period is short, determining the feature that correction factor keeps constant in adjacent period ; thus deriving the recursive expression of current prediction error at k+1 moment; Utilizing k-time prediction error And adjacent two period duty cycle difference value calculates correction coefficient .

7. The phase-shifted full-bridge model predictive current control method based on prediction error compensation according to claim 6, characterized in that, The calculation of the correction coefficient comprises the following steps:

8. A phase-shifted full-bridge model predictive current control system based on prediction error compensation, characterized in that, When the duty cycle difference is greater than a preset threshold, a new correction coefficient is updated; when the duty cycle difference is less than the preset threshold, the last period correction coefficient is kept; the obtained compensation value is added to the prediction current model for correction, and the system prediction error is compensated. The application relates to a control method and device for a phase-shifted full-bridge converter. The application comprises: an equivalent circuit model construction module for establishing an equivalent circuit model of the phase-shifted full-bridge converter based on a topology structure of a Buck converter, and establishing a predictive control model based on the equivalent circuit model; an error analysis module for analyzing an inductor current and a system modeling error source based on the equivalent circuit model and the predictive control model of the phase-shifted full-bridge converter; an error compensation module for constructing a prediction error recursive relationship from k time to k+1 time and using a correction coefficient of a threshold judgment mechanism to effectively compensate errors in the phase-shifted full-bridge converter system; The control module is used for introducing a prediction error compensation correction strategy into a closed-loop control process of the phase-shifted full-bridge converter, and optimizing a duty cycle design to realize precise control and dynamic performance improvement. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The computer program is executed by the processor to implement the steps of the phase-shifted full-bridge model predictive current control method based on prediction error compensation in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the phase-shifted full-bridge model predictive current control method based on prediction error compensation in any one of claims 1 to 7.

Images