A model predictive control method and system for energy storage bidirectional converter

Abstract

The present invention proposes a model predictive control method and system for energy storage bidirectional converters, which uses Kalman filtering to identify linear parameters in the model predictive control algorithm online to avoid the influence of the noise tolerance strategy on the control target accuracy; at the same time, two-step predictive control is introduced to reduce the output voltage fluctuation caused by calculation delay; the noise tolerance coefficient is introduced to reconstruct the model prediction, which greatly improves the robustness of the converter at the cost of slightly sacrificing dynamic performance. Compared with the existing noise tolerance improvement method, the control strategy proposed by the present invention has lower system hardware cost, lower processor computing performance requirements, and almost unchanged dynamic performance.

Description

Technical Field

[0001] The present invention belongs to the technical field of model predictive control of converters, and in particular relates to a model predictive control method and system for an energy storage bidirectional converter. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The increase of distributed generation equipment in new power systems has put forward higher requirements for emergency rescue of power systems. When distributed generation equipment is disconnected from the main grid due to a fault, energy storage devices can quickly make up for the power shortage. The series resonant dual-bridge-active (DAB) converter has the advantages of bidirectional power flow, no DC bias, and smaller switch-off current. It has broad application prospects in energy storage systems. In order to improve the dynamic performance of energy storage converters and achieve rapid response to power demand, model predictive control has been gradually applied to the field of power electronics, but its inherent problem of high sensitivity to sampling noise limits its application in practical engineering.

[0004] In order to solve the noise sensitivity problem of model predictive control, the existing solutions mainly solve it by improving processor performance, adding peripheral hardware filtering circuits, performing software filtering on measured values, and reconstructing state space models. For example, the paper "Multirate digital signal processing and noise suppression for dual activebridge DC–DC converters in a power electronic traction transformer" analyzes the generation mechanism of switching noise, quantization noise and measurement noise and their impact on dual active full-bridge converters, and suppresses them by increasing sampling frequency, reducing signal transmission delay, and increasing controller main frequency. The paper "Sensorless current predictive control of active full-bridge converter" introduces a Kalman filter scheme to improve the robustness of continuous set model predictive control, and uses the optimal estimated value of state variables instead of the actual measured value, which effectively suppresses the influence of measurement noise. The paper "Robust predictive control of grid-connected converters: sensor noise suppression with parallel-cascade extended state observer" combines the hybrid cascade extended state observer with model-free predictive control to enhance the interference suppression and noise suppression capabilities of the system.

[0005] In summary, the existing solutions to solve the noise sensitivity problem in model predictive control have the following problems:

[0006] (1) Improving processor performance and increasing sampling frequency will increase system costs, which is not conducive to engineering applications.

[0007] (2) When the system is in a steady state, the software filtering algorithm still relies on the filter to estimate the state variables, which occupies more storage space and computing power of the digital chip. Summary of the invention

[0008] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a model predictive control method and system for a bidirectional energy storage converter. Compared with the existing noise tolerance improvement method, the control strategy proposed in the present invention has lower system hardware cost, lower requirements on processor computing performance, and almost unchanged dynamic performance.

[0009] To achieve the above object, a first aspect of the present invention provides a model predictive control method for an energy storage bidirectional converter, comprising:

[0010] Obtaining output voltage, output current and target voltage given values ​​of the energy storage bidirectional converter;

[0011] A two-step model predictive control is adopted. At the current moment, the optimal phase shift control quantity is calculated based on the model predictive control algorithm according to the output voltage, output current and target voltage given values ​​at the previous moment. The Kalman filter is used to perform online identification of the linear parameters in the model predictive control algorithm based on the optimal phase shift control quantity at the current moment, and the optimal phase shift control quantity at the next moment is obtained by combining the voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient.

[0012] Based on the optimal phase shift control amount at the next moment, the control signal of each switch tube of the next control cycle of the energy storage bidirectional converter is obtained to control the energy storage bidirectional converter.

[0013] A second aspect of the present invention provides a model predictive control system for an energy storage bidirectional converter, comprising:

[0014] An acquisition module is used to obtain the output voltage, output current and target voltage given value of the energy storage bidirectional converter;

[0015] The prediction module is used to adopt a two-step model predictive control to calculate the optimal phase shift control amount at the current moment based on the output voltage, output current and target voltage given value at the previous moment based on the model predictive control algorithm; the Kalman filter is used to perform online identification of the linear parameters in the model predictive control algorithm based on the optimal phase shift control amount at the current moment, and the voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient is combined to obtain the optimal phase shift control amount at the next moment;

[0016] The control module is used to obtain the control signal of each switch tube in the next control cycle of the energy storage bidirectional converter based on the optimal phase shift control amount at the next moment, and control the energy storage bidirectional converter.

[0017] The third aspect of the present invention provides a computer device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, a model predictive control method for an energy storage bidirectional converter is performed.

[0018] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, executes a model predictive control method for an energy storage bidirectional converter.

[0019] One or more of the above technical solutions have the following beneficial effects:

[0020] In the present invention, Kalman filtering is used to identify the linear parameters in the model predictive control algorithm online to avoid the influence of the noise tolerance strategy on the control target accuracy; at the same time, two-step predictive control is introduced to reduce the output voltage fluctuation caused by calculation delay; the noise tolerance coefficient is introduced to reconstruct the model prediction, and the robustness of the converter is greatly improved at the cost of slightly sacrificing dynamic performance. Compared with the existing noise tolerance improvement method, the control strategy proposed by the present invention has lower system hardware cost, lower processor computing performance requirements, and almost unchanged dynamic performance.

[0021] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0023] Figure 1 This is a topological structure diagram of a series resonant DAB converter in Embodiment 1 of the present invention;

[0024] Figure 2 This is a working waveform diagram of the series resonant DAB converter in the first embodiment of the present invention;

[0025] FIG3( a ) is a diagram of the system operation process under single-step predictive control in Embodiment 1 of the present invention;

[0026] FIG3( b ) is a diagram of the system operation process under two-step predictive control in Embodiment 1 of the present invention;

[0027] Figure 4 : is the system amplitude-frequency characteristic curve under different noise tolerance coefficients in the first embodiment of the present invention;

[0028] FIG5( a ) is a system operating waveform when the noise tolerance factor Y=0.03 in the first embodiment of the present invention;

[0029] FIG5( b ) is a system operating waveform when the noise tolerance factor Y=0.12 in the first embodiment of the present invention;

[0030] Figure 6 This is a flowchart of the model predictive control method for the energy storage bidirectional converter in the first embodiment of the present invention. DETAILED DESCRIPTION

[0031] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0032] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0033] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0034] Embodiment 1

[0035] This embodiment discloses a model predictive control method for an energy storage bidirectional converter, including:

[0036] Obtaining output voltage, output current and target voltage given values ​​of the energy storage bidirectional converter;

[0037] A two-step model predictive control is adopted to calculate the optimal phase shift control amount at the current moment based on the model predictive control algorithm according to the output voltage, output current and target voltage given values ​​at the previous moment. A Kalman filter is used to perform online identification of the linear parameters in the model predictive control algorithm based on the optimal phase shift control amount at the current moment, and the optimal phase shift control amount at the next moment is obtained by combining the voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient.

[0038] Based on the optimal phase shift control amount at the next moment, the control signal of each switch tube of the next control cycle of the energy storage bidirectional converter is obtained to control the energy storage bidirectional converter.

[0039] The solution of this embodiment adopts two-step continuous set model predictive control, online identification of prediction model related parameters, introduces noise tolerance coefficient, slightly sacrifices dynamic performance to achieve a significant improvement in system stability. Compared with the existing noise tolerance improvement method, the solution of this embodiment does not increase hardware cost, and the implementation process is simple and reliable, which effectively reduces the computing burden of the processor.

[0040] This embodiment is not limited to the application of series resonant DAB converters. Figure 6 As shown, this embodiment takes a series resonant DAB converter as an example for detailed description, and specifically includes:

[0041] 1. Construction of prediction model for series resonant DAB converter.

[0042] The series resonant DAB converter topology is as follows: Figure 1 The main circuit includes input side support capacitor C1, output side support capacitor C2, resonant capacitor L r , resonant inductor C r , high frequency transformer T, switch tubes S1~S8 and their anti-parallel diodes. The dual active bridge series resonant converter includes four bridge arms, and the upper and lower switch tubes of each bridge arm are complementary. The driving signal of the complementary switch tube is a square wave signal with a duty cycle of 50%, that is, each switch tube is turned on for half a switching cycle.

[0043] Figure 2 The working waveform of the series resonant DAB converter under single phase shift modulation. The diagonal switch tubes of the primary or secondary full bridge are considered as a group (for example, switch tubes S1 and S4 are considered as a group). During the dead time before any group of switch tubes are turned on, the parallel parasitic capacitors of the switch tubes and the complementary switch tubes on the same side complete the charging and discharging process, so that the terminal voltage of the switch tube group before turning on is zero, realizing zero voltage switching. Figure 2 Medium T hs is a half switching period, i L is the resonant inductor current, D is the phase shift duty cycle, and the phase difference between the inductor current and the bridge voltage can be controlled by adjusting the phase shift duty cycle to achieve free switching of the power flow size and direction.

[0044] The deviation between the actual value of the output voltage and the given value depends on the phase shift duty cycle D and the output current I o When the average current I2 of the full bridge DC side determined by D is equal to the output current I o When they are equal, the output voltage reaches a given value, that is, ΔV2 is zero. The output voltage deviation can be obtained by the following formula:

[0045]

[0046] Among them, X LC is the resonant cavity impedance, f c To control the frequency.

[0047] In order to avoid the impact of the subsequent noise tolerance strategy on the control target accuracy, the phase shift duty cycle related terms can be linearized, and the Kalman filter algorithm can be used to estimate the linearization coefficient online. The simplified output voltage deviation model is:

[0048]

[0049] In the formula, G is the control gain. As the circuit parameters change, the control gain also changes. Write the above formula in matrix form:

[0050]

[0051] in,

[0052]

[0053] In order to ensure accurate control of the output voltage, this embodiment uses the Kalman filter algorithm to identify the control gain G(k) in equation (3) online. The implementation process is as follows:

[0054]

[0055] In the formula, is the state vector, P(k) represents the covariance matrix of the Kalman estimation error at the kth moment, K is the Kalman gain, R and Q represent the covariance matrices of the measurement noise and process noise respectively, and H(k) is the measurement system parameter.

[0056] 2. Two-step predictive control.

[0057] In the single-step model predictive control, the controller aims to ensure that the output voltage at the next moment reaches its given value as soon as possible, and solves the optimal phase shift duty cycle within the control cycle. The system operation process is shown in Figure 3(a). Before the kth moment, the output voltage deviates greatly from the given value, and the system outputs at maximum power. At this time, the phase shift duty cycle is 0.5. The output voltage sampling value at the kth moment is V 2.cal (k), in order to make the output voltage reach V2 at the k+1th moment * , the controller calculates the optimal operating trajectory of the system and obtains the optimal phase shift duty cycle as D1. Considering the controller calculation time, the output voltage has risen to V at the time of phase shift duty cycle update. 2.rea (k), so the output voltage cannot reach V2 at time k+1 * Under single-step predictive control, as time goes by, the actual value of the output voltage gradually approaches its given value. However, since the calculation time of the controller cannot be ignored, the output voltage cannot move along the predetermined trajectory, the dynamic response of the system becomes slower, and the output voltage sampling value fluctuates greatly during the transient process, resulting in poor system stability.

[0058] In order to solve the problem of controller calculation time delay, this embodiment adopts a two-step predictive control model to solve the optimal phase-shift duty cycle.

[0059] According to formula (2), the output voltage deviation at time k+1 is:

[0060]

[0061] Combining equations (2) and (6), the output voltage prediction value at the k+2th moment is:

[0062]

[0063] The implementation process of the continuous set simplified model predictive control proposed in this embodiment is as follows: the controller calculates the optimal phase-shift duty ratio D1 at the k+1 moment at the kth moment, and first updates the phase-shift duty ratio at the k+1 moment, and then calculates the optimal phase-shift duty ratio D2 at the k+2 moment, and so on. The calculation process is shown in Figure 3(b). The two-step predictive control predicts the optimal phase-shift duty ratio at the next moment at this moment, and updates the phase-shift duty ratio in time at the beginning of the next control cycle to ensure that the output voltage at the next moment reaches the reference value as quickly as possible.

[0064] In formula (6), V2(k) and I o (k) can be obtained through measurement, D(k) is the optimal phase shift duty cycle required by the predictive control algorithm at the previous moment, G(k) can be iterated by formula (5), and the load current does not change much within a control cycle, so it can be considered that: I o (k+1)=I o (k). The variable G(k+1) cannot be measured directly, but its size can be obtained by fitting historical data.

[0065] 3. Strategy for improving noise tolerance of model predictive control.

[0066] When the output voltage measurement value fluctuates due to sampling noise or low bit number of the controller AD converter, the measurement noise will be fed back to the prediction model, and the phase shift duty cycle will change accordingly. Assuming that the output voltage measurement noise is δV2 when the system is in steady state, the phase shift duty cycle change caused by it is δD. According to formula (2), we can get:

[0067]

[0068] It can be seen from formula (8) that a small voltage fluctuation will cause D to jump between 0 and 0.5. When D jumps to a larger value, the switch tube turn-off current increases significantly, the turn-off loss increases, and the energy storage unit needs to withstand higher voltage and current stress, which accelerates the aging of the device; when D jumps to a smaller value, the primary and secondary side switches cannot all achieve zero voltage turn-on under non-unit voltage gain, and the system efficiency is reduced; when the phase shift duty cycle fluctuates between 0 and 0.5, the output voltage noise increases accordingly, which is not conducive to the stable operation of voltage-sensitive loads.

[0069] Now define the noise tolerance factor Y:

[0070]

[0071] Substituting equation (9) into equation (2), the output voltage deviation can be expressed as:

[0072] ΔV2=G·DY·I o (10)

[0073] When the system reaches steady state, the phase shift duty cycle is:

[0074]

[0075] The correction of the control gain G makes the output voltage deviation zero. Therefore, for different values ​​of Y, the Kalman filter algorithm can always obtain the optimal estimate of G to satisfy V2(k+2)=V2 * .

[0076] The operating conditions of the converter system remain unchanged under the same transmission power and different noise margin coefficients. Therefore, when Y increases or decreases, the steady-state control gain increases or decreases proportionally with the change of the noise margin coefficient. It can be seen from formula (8) that when G increases, the phase shift duty cycle can tolerate larger output voltage fluctuations in the linear region.

[0077] Set the system steady-state operating point to V1 = 100V, V2 = 100V, P = 250W. Apply a small disturbance with a frequency range of 10Hz-10kHz and an amplitude of 0.01V on the output voltage side, which can be equivalent to the reference voltage side. The closed-loop transfer function of the system is:

[0078]

[0079] Among them, v2 * is a small perturbation applied to the reference voltage.

[0080] The frequency sweep results of the closed-loop transfer function under different Y values ​​are as follows Figure 4As shown. The noise tolerance coefficient Y has a negative impact on the system bandwidth. As the noise tolerance coefficient Y increases, the control system bandwidth decreases and the theoretical dynamic response time of the system slows down. To further quantify the impact of the increase in Y on the dynamic performance of the system, Figure 5(a) and Figure 5(b) are the converter operating waveforms during load switching under different noise tolerance coefficients. The experimental results show that as the noise tolerance coefficient increases, the sensitivity of the phase shift duty cycle to sampling noise is significantly reduced, and the dynamic performance is almost unaffected.

[0081] 4. The specific implementation method of the noise tolerance improvement strategy in the series resonant DAB converter model prediction control is as follows:

[0082] Step 1: The sensor first outputs the voltage V2 and the current I o Sampling, setting the output voltage given value V2 * ;

[0083] Step 2: Combine the optimal phase shift duty ratio D at the current moment obtained in step 3 of the previous moment opt (k), the control gain G(k) is calculated by formula (5);

[0084] Step 3: Substitute G(k) into the unified simplified model shown in equation (10) to calculate the optimal phase shift duty cycle D at the next moment: opt (k+1), for use in step 2 at the next moment;

[0085] Step 4: Input the obtained phase-shift duty cycle into the phase-shift modulator for phase-shift modulation to obtain the control signal of each switch tube and control the converter to achieve fast dynamic response.

[0086] By looping to step 2 above, the model predictive control of the series resonant DAB converter can be realized. This control strategy effectively improves the dynamic performance of the system and effectively reduces the sensitivity of the phase shift duty cycle to the output voltage sampling noise. Compared with the existing noise tolerance strategy, this embodiment significantly reduces the cost.

[0087] Embodiment 2

[0088] The purpose of this embodiment is to provide a model predictive control system for an energy storage bidirectional converter, including:

[0089] An acquisition module is used to obtain the output voltage, output current and target voltage given value of the energy storage bidirectional converter;

[0090] A prediction module is used to adopt a two-step model predictive control, and at the current moment, according to the output voltage, output current and target voltage given value at the previous moment, the optimal phase shift control amount at the current moment is calculated based on the model predictive control algorithm, and a Kalman filter is used to perform online identification of linear parameters in the model predictive control algorithm based on the optimal phase shift control amount at the current moment, and the optimal phase shift control amount at the next moment is obtained by combining the voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient;

[0091] The control module is used to obtain the control signal of each switch tube in the next control cycle of the energy storage bidirectional converter based on the optimal phase shift control amount at the next moment, and control the energy storage bidirectional converter.

[0092] Embodiment 3

[0093] The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the program.

[0094] Embodiment 4

[0095] The purpose of this embodiment is to provide a computer-readable storage medium.

[0096] A computer-readable storage medium stores a computer program, which executes the steps of the above method when executed by a processor.

[0097] The steps involved in the apparatuses of the above embodiments 2, 3 and 4 correspond to the method embodiment 1, and the specific implementation methods can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0098] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0099] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A model predictive control method for an energy storage bidirectional converter, characterized in that: include: Obtaining output voltage, output current and target voltage given values ​​of the energy storage bidirectional converter; A two-step model predictive control is adopted. At the current moment, the optimal phase shift control quantity is calculated based on the model predictive control algorithm according to the output voltage, output current and target voltage given values ​​at the previous moment. The Kalman filter is used to perform online identification of the linear parameters in the model predictive control algorithm based on the optimal phase shift control quantity at the current moment, and the optimal phase shift control quantity at the next moment is obtained by combining the voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient. Based on the optimal phase shift control amount at the next moment, the control signal of each switch tube of the next control cycle of the energy storage bidirectional converter is obtained to control the energy storage bidirectional converter; The voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient is: the product of the control gain and the phase shift control amount, and the difference between the product of the noise tolerance coefficient and the output current of the energy storage bidirectional converter; wherein the noise tolerance coefficient is determined based on the ratio of the output side support capacitance of the energy storage bidirectional converter and the product of the switching frequency.

2. The model predictive control method for an energy storage bidirectional converter according to claim 1, characterized in that: Energy storage bidirectional converter The voltage deviation model at the moment is: ， in, for Output voltage deviation at all times, for Output voltage at all times, for The control gain at the moment, for The phase shift control amount at the moment, C 2 is the output side support capacitor, fc is the control frequency, for The output current at the moment.

3. The model predictive control method for an energy storage bidirectional converter according to claim 1, characterized in that: In the two-step model predictive control, Time to The optimal phase shift control quantity at the moment is calculated and First update the phase shift control value at any time, and then The optimal phase shift duty ratio is calculated at the moment to make the The output voltage reaches the target reference value at that moment.

4. The model predictive control method for an energy storage bidirectional converter according to claim 1, characterized in that: Kalman filtering is used to identify the control gain of the optimal phase shift control amount at the current moment online, specifically: Update the Kalman gain based on the covariance matrix of the previous moment; The control gain is estimated based on the updated Kalman gain, the state vector at the previous moment, the voltage target control quantity at the current moment, and the measurement system parameters.

5. The model predictive control method for an energy storage bidirectional converter according to claim 4, characterized in that: The state vector is composed of the output side support capacitance and control frequency of the energy storage bidirectional converter, and the control gain of the phase shift control amount; the measurement system parameters are composed of the output current and phase shift control amount of the energy storage bidirectional converter.

6. The model predictive control method for an energy storage bidirectional converter according to claim 4, characterized in that: The optimal control gain estimated by the Kalman filter algorithm satisfies the The output voltage at that moment is equal to the target voltage value.

7. A model predictive control system for an energy storage bidirectional converter, characterized in that: A method for implementing a model predictive control of an energy storage bidirectional converter as claimed in any one of claims 1 to 6 comprises: An acquisition module is used to obtain the output voltage, output current and target voltage given value of the energy storage bidirectional converter; The prediction module is used to adopt a two-step model predictive control to calculate the optimal phase shift control amount at the current moment based on the output voltage, output current and target voltage given value at the previous moment based on the model predictive control algorithm; the Kalman filter is used to perform online identification of the linear parameters in the model predictive control algorithm based on the optimal phase shift control amount at the current moment, and the optimal phase shift control amount at the next moment is obtained by combining the voltage deviation model of the energy storage bidirectional converter based on the noise tolerance coefficient; The control module is used to obtain the control signal of each switch tube in the next control cycle of the energy storage bidirectional converter based on the optimal phase shift control amount at the next moment, and control the energy storage bidirectional converter.

8. A computer device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, a model predictive control method for an energy storage bidirectional converter as described in any one of claims 1 to 6 is executed.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the model predictive control method for an energy storage bidirectional converter according to any one of claims 1 to 6 is executed.

Images