Direct current bus voltage control method based on model predictive control and current sharing control

Abstract

The invention discloses a DC bus voltage control method based on model prediction control and current sharing control, and the method comprises the following steps: S1, obtaining a current reference value id through PI control in a voltage outer loop link when a system has a sudden load change and the voltage deviates from a normal operation range; a virtual capacitance current reference value ivir is obtained in a virtual inertia control link, and a reference current iref controlled by an inner loop is obtained by combining the two values; s2, respectively obtaining a current reference value ib-ref of current inner loop control of the energy storage system and a current reference value ig-ref of a G-VSC converter of an alternating current power grid through current sharing control; s3, a multi-step MPC controller is adopted in a current inner loop, and rapid inner loop current tracking of the converter is achieved; and S4, the bidirectional DC / DC converter of the energy storage system and the AC-DC converter of the AC system adjust the power output to the DC bus according to the duty ratio obtained by the current reference value so as to adjust the voltage of the DC bus.

Description

Technical Field

[0001] The present invention relates to the technical field of DC bus voltage control, and in particular to a DC bus voltage control method based on model predictive control and current sharing control. Background Art

[0002] In recent years, the rapid development of clean energy has had a significant impact on the power system landscape. To increase the capacity to absorb renewable energy generation, the traditional power system urgently needs to undergo transformation and upgrading. Consequently, the concept of "PV-storage-direct-flexible" has been proposed. "PV-storage" refers to photovoltaic power generation and energy storage systems, "direct" refers to direct current transmission, and "flexible" refers to the use of flexible loads to achieve flexible power consumption. However, since distributed photovoltaics, as the primary power generation equipment in PV-storage-direct-flexible systems, lack the inertia provided by wind turbines, the system's inertia is relatively low. When the distributed power sources or loads fluctuate, the system is prone to power imbalance, resulting in large fluctuations in the DC bus voltage.

[0003] Energy storage devices in PV-storage-direct-flexible systems are key devices that assist distributed energy resources in generating power. Due to their rapid charging and discharging capabilities, virtual inertia control can be added to the B-DC converter to provide inertial support for the entire system and maintain system voltage stability. However, existing research has significant shortcomings: First, while virtual inertia has a clear mathematical expression, it is difficult to calculate the optimal value of the key parameter, virtual capacitance. Second, energy storage units typically use a dual closed-loop control structure with a voltage outer loop and a current inner loop, requiring the inner loop to have a very fast response speed. However, the PI control used in the inner current loop is a hysteresis control, which to some extent cannot meet the fast response requirements. Therefore, it is urgent to propose a new virtual inertia control method to maintain voltage stability in PV-storage-direct-flexible distribution systems.

[0004] Therefore, there is still room for improvement in the existing technology. Summary of the Invention

[0005] The purpose of the present invention is to provide a DC bus voltage control method based on model predictive control and current sharing control, aiming to solve technical problems such as dynamic instability of DC bus voltage, insufficient performance of traditional control strategies, and lack of system energy efficiency optimization.

[0006] To achieve the above-mentioned purpose, the technical solution of the present invention is: a DC bus voltage control method based on model predictive control and current sharing control, which includes the following steps:

[0007] S1: When the system load suddenly changes and the voltage deviates from the normal operating range, the voltage outer loop obtains the current reference value i through PI control. d The virtual inertia control link calculates the virtual capacitor current reference value i based on the optimal virtual capacitor parameters obtained by model predictive control. virThe combination of the two gives the reference current i for inner loop control. ref ;

[0008] S2: Considering the large inertia of the AC grid, when the energy storage output is insufficient, the AC grid can also provide inertia for the system through additional virtual inertia control. Therefore, the current reference value i of the energy storage system current inner loop control is obtained through current sharing control. b-ref and the current reference value i of the G-VSC converter of the AC grid g-ref ;

[0009] S3: To avoid the hysteresis of the traditional current inner loop control based on PI control, a multi-step MPC controller is used in the current inner loop to achieve fast inner loop current tracking of the converter;

[0010] S4: The bidirectional DC / DC converter of the energy storage system and the AC-DC converter of the AC system adjust the power output to the DC bus according to the duty cycle obtained from the current reference value, thereby regulating the DC bus voltage.

[0011] The DC bus voltage control method based on model predictive control and current sharing control, wherein the voltage outer loop control in step S1 retains the voltage link of the voltage and current dual closed-loop control commonly used in energy storage units, and the size of the virtual capacitor, a key parameter in the additional virtual inertia control link, is obtained through MPC; MPC uses a rolling optimization mechanism to replace global optimization with optimization in a finite time domain; at each sampling moment, MPC first predicts the future trend of the system, then solves an optimization problem in a finite time domain, and finally applies the obtained first group of elements of the control sequence to the system;

[0012] Among them, the optimization problem is the core of MPC, which includes three parts: prediction model, objective function and constraints. The specific form is shown in the following formula:

[0013]

[0014] In the formula, v(k+i) is the control quantity at time k+i; v min (k+i) is the minimum value of the control quantity at time k+i; v max (k+i) is the maximum value of the control quantity at time k+i; Δv(k+i) is the control increment at time k+i; Δv min (k+i) is the minimum value of the control increment at time k+i; Δv max (k+i) is the maximum value of the control increment at time k+i; y(k+i) is the output at time k+i; y min (k+i) is the minimum output value at time k+i; y max(k+i) is the maximum output at time k+i; Δx(k+i|k) is the state increment at time k+i predicted by time k; Δx(k+i+1|k) is the state increment at time k+i+1 predicted by time k; Δd(k+i) is the disturbance increment at time k+i; y(k+i|k) is the output at time k+i predicted by time k; J(x(k,ΔV)) is the objective function, which is a function of the state x(k) at time k and the control sequence ΔV, where ΔV = [Δv(k) Δv(k+1) … Δv(k+m-1)] T ; A, B, C, and D are coefficient matrices. Coefficient matrix A represents the system state transfer matrix, which describes the evolution of the system state variables over time. Coefficient matrix B represents the influence of the control input (virtual inertia coefficient) on the system state variables. Coefficient matrix C represents the influence of the disturbance variable on the system state. Coefficient matrix D links the system state to the output, and represents the relationship between the state variables and the measurable output.

[0015] In the DC bus voltage control method based on model predictive control and current sharing control, the objective function is as follows:

[0016] J=α|U dc (k+1)-U dcn |+(1-α)|εδU dc (k+1)|

[0017] Where U dc (k+1) is the DC bus voltage value at time k+1; U dcn is the rated value of DC bus voltage; δU dc (k+1) is the rate of change of the DC bus voltage at time k+1; ε is the proportional coefficient, which is used to balance the numerical difference between the voltage deviation and the voltage change rate in units; α is the weight coefficient. In the early stage of the load mutation, the virtual capacitor quickly releases the current mainly to suppress the voltage change rate and prevent the voltage from dropping too quickly. Therefore, α should take a small value, that is, α=0.1; during the dynamic voltage fluctuation process, the value of α gradually increases to 1. In this way, in the middle and late stages of the voltage change, the voltage deviation is mainly the main factor, that is, the voltage steady-state error is reduced to prevent the voltage drop from being too large.

[0018] In the DC bus voltage control method based on model predictive control and current sharing control, the specific process of MPC control in step S1 is as follows:

[0019] S11: Obtain real-time operating data of the DC power distribution system and update the prediction model;

[0020] S12: Solve the optimization problem. If the optimization problem has a solution, output the control increment; if the problem has no solution, keep the control amount unchanged and exit MPC;

[0021] S13: Output the control amount and perform virtual inertia control based on the obtained optimal parameters.

[0022] In the DC bus voltage control method based on model predictive control and current sharing control, the specific process of current sharing control in step S2 is as follows:

[0023] S21: First, collect the operating data of the energy storage unit, calculate its state of charge SOC, and establish the reference current i in turn. ref The weight factor γ of , the expression of the weight factor γ is:

[0024]

[0025] Where, SOC max The maximum state of charge allowed by the battery is 90%; SOC min The minimum state of charge allowed for the battery is 10%.

[0026] S22: Set the DC bus current reference value i ref The current reference value i of the energy storage unit is obtained by reasonable allocation through the weight factor γ b-ref and the AC grid current reference value i g-ref , and then the inner loop current control link of the energy storage unit and the G-VSC converter of the AC grid respond to the allocated current reference value.

[0027] In the DC bus voltage control method based on model predictive control and current sharing control, the steps of the current inner loop multi-step predictive control in step S3 are as follows:

[0028] S31: First, the charging and discharging state of the energy storage unit is determined by judging the magnitude of the DC bus voltage. After determining the working state, the i is calculated according to the inner loop current prediction model in the Boost mode and the Buck mode. b (k+n);

[0029] S32: Calculate the output of the objective function based on each predicted value and compare their sizes, output the control variable that minimizes the objective function value, and finally the bidirectional B-DC converter operates according to the output control variable. In the DC bus voltage control method based on model predictive control and current sharing control, the inner loop current prediction model of the energy storage unit in the Boost mode in step S31 is as follows:

[0030]

[0031] Where, U dc (k) is the DC bus voltage at the sampling moment, U bat(k) is the battery voltage at the sampling moment, i bat (k) is the battery current at the sampling moment; i bat (k+1) is the battery current predicted at the next sampling moment; T s is the discrete sampling time;

[0032] In the battery bidirectional DC / DC converter topology, L represents the energy storage inductor, g1 represents the switching state of the switch tube S1, and g2 represents the switching state of the switch tube S2 (g1=1 represents the switch tube is on, g1=0 represents the switch tube is off; the same applies to g2);

[0033] The inner loop current prediction model of the energy storage unit in Buck mode is as follows:

[0034]

[0035] Energy storage unit current inner loop prediction objective function:

[0036] J=|i b (k+n)-i b-ref |

[0037] Where: i b (k+n) is the energy storage unit current when the prediction step size is n; i b-ref It is the reference value of the battery internal loop current.

[0038] Beneficial effects:

[0039] The present invention first adds virtual inertia control to the outer voltage loop. The virtual capacitor in this virtual inertia control is calculated and optimized using an MPC controller based on the current system state. Current sharing control then introduces weighting factors to distribute current reference values to the energy storage unit and the AC grid converter. A multi-step FCS-MPC controller is employed in the energy storage unit's inner current loop to ensure rapid current tracking. Finally, the energy storage unit's B-DC converter and the AC grid's G-VSC adjust their duty cycles based on the current reference values, thereby adjusting the output power.

[0040] Compared to existing technologies, this solution improves the voltage stability and dynamic response performance of DC distribution networks by optimizing virtual inertia through MPC calculations, current sharing control, and multi-step FCS-MPC. First, compared to traditional fixed virtual inertia control, this solution uses MPC to adaptively optimize virtual capacitance based on real-time voltage conditions, making system inertia adjustable and better suppressing voltage fluctuations. This overcomes the problem of fixed inertia parameters being unable to adapt to voltage fluctuations caused by varying loads. Second, it uses weighted factors to optimize current sharing control, rationally distributing the power burden between batteries and grid-connected converters, improving system stability. This also addresses the issue of traditional current sharing control methods that can lead to overloading of energy storage units and low converter utilization. Furthermore, the introduction of FCS-MPC multi-step optimization improves the accuracy and dynamic response speed of the current inner loop control, reduces steady-state errors, and overcomes the current tracking speed limitations of traditional PI control. Finally, power distribution is optimized by adjusting the converter duty cycle, ensuring stable system operation despite load fluctuations. Overall, this solution overcomes the limitations of traditional methods in inertia regulation, current distribution, power optimization, etc., enabling the system to adapt more accurately to different load changes, improve voltage stability, and at the same time improve current tracking accuracy and power distribution balance, thereby enhancing the system's stability and robustness in complex operating environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is the overall control block diagram of the present invention.

[0042] Figure 2 It is the virtual inertia control block diagram of the present invention.

[0043] Figure 3 It is the flow chart of the MPC program in the present invention.

[0044] Figure 4 This is the flow chart of the current sharing control in the present invention.

[0045] Figure 5 This is a flow chart of the current inner loop multi-step predictive control in the present invention.

[0046] Figure 6 This is the topological structure of the battery bidirectional DC / DC converter in the present invention. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clear and explicit, the present invention is further described in detail below with reference to the accompanying drawings and examples.

[0048] like Figure 1-5 As shown, the present invention discloses a DC bus voltage control method based on model predictive control and current sharing control, which includes the following steps:

[0049] S1: When the system load suddenly changes and the voltage deviates from the normal operating range, the voltage outer loop obtains the current reference value i through PI control. d The virtual inertia control link calculates the virtual capacitor current reference value i based on the optimal virtual capacitor parameters obtained by model predictive control. vir The combination of the two gives the reference current i for inner loop control. ref ;

[0050] S2: Considering the large inertia of the AC grid, when the energy storage output is insufficient, the AC grid can also provide inertia for the system through additional virtual inertia control. Therefore, the current reference value i of the energy storage system current inner loop control is obtained through current sharing control. b-ref and the current reference value i of the G-VSC converter of the AC grid g-ref ;

[0051] S3: To avoid the hysteresis of the traditional current inner loop control based on PI control, a multi-step MPC controller is used in the current inner loop to achieve fast inner loop current tracking of the converter;

[0052] S4: The bidirectional DC / DC converter of the energy storage system and the AC-DC converter of the AC system adjust the power output to the DC bus according to the duty cycle obtained from the current reference value, thereby regulating the DC bus voltage.

[0053] The voltage outer loop control in step S1 retains the voltage link of the voltage-current dual closed-loop control commonly used in energy storage units. The key parameter in the additional virtual inertia control link, the size of the virtual capacitor, is obtained through MPC. MPC uses a rolling optimization mechanism to replace global optimization with optimization in a finite time domain. At each sampling moment, MPC first predicts the future trend of the system, then solves an optimization problem in a finite time domain, and finally applies the first set of elements of the obtained control sequence to the system.

[0054] Among them, the optimization problem is the core of MPC, which includes three parts: prediction model, objective function and constraints. The specific form is shown in the following formula:

[0055]

[0056] In the formula, v(k+i) is the control quantity at time k+i; v min (k+i) is the minimum value of the control quantity at time k+i; v max (k+i) is the maximum value of the control quantity at time k+i; Δv(k+i) is the control increment at time k+i; Δv min (k+i) is the minimum value of the control increment at time k+i; v max (k+i) is the maximum value of the control increment at time k+i; y(k+i) is the output at time k+i; ymin (k+i) is the minimum output value at time k+i; y max (k+i) is the maximum output at time k+i; Δx(k+i|k) is the state increment at time k+i predicted by time k; Δx(k+i+1|k) is the state increment at time k+i+1 predicted by time k; Δd(k+i) is the disturbance increment at time k+i; y(k+i|k) is the output at time k+i predicted by time k; J(x(k,ΔV)) is the objective function, which is a function of the state x(k) at time k and the control sequence ΔV, where ΔV = [Δv(k)Δv(k+1)…Δv(k+m-1)] T ; A, B, C, and D are coefficient matrices. Coefficient matrix A represents the system state transfer matrix, which describes the evolution of the system state variables over time. Coefficient matrix B represents the influence of the control input (virtual inertia coefficient) on the system state variables. Coefficient matrix C represents the influence of the disturbance variable on the system state. Coefficient matrix D links the system state to the output, and represents the relationship between the state variables and the measurable output.

[0057] The objective function is as follows:

[0058] J=α|U dc (k+1)-U dcn |+(1-α)|εδU dc (k+1)|

[0059] Where U dc (k+1) is the DC bus voltage value at time k+i; U dcn is the rated value of DC bus voltage; δU dc (k+1) is the rate of change of the DC bus voltage at time k+i; ε is the proportional coefficient, which is used to balance the numerical difference between the voltage deviation and the voltage change rate in units; α is the weight coefficient. In the early stage of the load mutation, the virtual capacitor quickly releases the current mainly to suppress the voltage change rate and prevent the voltage from dropping too quickly. Therefore, α should take a small value, that is, α=0.1; during the dynamic voltage fluctuation process, the value of α gradually increases to 1. In this way, in the middle and late stages of the voltage change, the voltage deviation is mainly the main factor, that is, the voltage steady-state error is reduced to prevent the voltage drop from being too large.

[0060] Figure 3 The MPC program flow for obtaining the optimal virtual capacitance is shown. The specific process of MPC control in step S1 is as follows:

[0061] S11: Obtain real-time operating data of the DC power distribution system and update the prediction model;

[0062] S12: Solve the optimization problem. If the optimization problem has a solution, output the control increment; if the problem has no solution, keep the control amount unchanged and exit MPC;

[0063] S13: Output the control amount and perform virtual inertia control based on the obtained optimal parameters.

[0064] The weight factor γ is set according to the state of charge (SOC) of the energy storage unit, and the reference current is reasonably distributed using the weight factor, thereby determining the reference current of the energy storage unit and the AC grid, thereby achieving the purpose of reasonably distributing the inertial power while providing inertial power to the system. The specific current sharing control flow chart is as follows: Figure 4 As shown; the specific process of the current sharing control in step S2 is as follows:

[0065] S21: First, collect the operating data of the energy storage unit, calculate its state of charge SOC, and establish the reference current i in turn. ref The weight factor γ of , the expression of the weight factor γ is:

[0066]

[0067] Where, SOC max The maximum state of charge allowed by the battery is 90%; SOC min The minimum state of charge allowed for the battery is 10%.

[0068] S22: Set the DC bus current reference value i ref The current reference value i of the energy storage unit is obtained by reasonable allocation through the weight factor γ b-ref and the AC grid current reference value i g-ref , and then the inner loop current control link of the energy storage unit and the G-VSC converter of the AC grid respond to the allocated current reference value.

[0069] like Figure 5 As shown, the steps of the current inner loop multi-step predictive control in step S3 are as follows:

[0070] S31: First, the charging and discharging state of the energy storage unit is determined by judging the magnitude of the DC bus voltage. After determining the working state, the i is calculated according to the inner loop current prediction model in the Boost mode and the Buck mode. b (k+n);

[0071] S32: Calculate the output results of the objective function according to each predicted value and compare their sizes, output the control amount that can minimize the objective function value, and finally the bidirectional B-DC converter operates according to the output control amount.

[0072] like Figure 6 As shown, in the battery bidirectional DC / DC converter topology, U bat Indicates battery voltage, U dc Indicates the DC bus voltage; ibat Indicates the battery current, i dc represents the DC bus current; S1 and S2 represent IGBT switches; D1 and D2 represent anti-parallel diodes; L represents the energy storage inductor, and C represents the energy storage inductor and bus connection capacitor. g1 represents the switching state of switch S1, and g2 represents the switching state of switch S2 (g1=1 means the switch is on, g1=0 means the switch is off; the same applies to g2). Assume i bat 、i dc The positive direction is when the battery is discharging.

[0073] The inner loop current prediction model of the energy storage unit in the Boost mode in step S31 is as follows:

[0074]

[0075] Where U dc (k) is the DC bus voltage at the sampling moment, U bat (k) is the battery voltage at the sampling moment, i bat (k) is the battery current at the sampling moment; i bat (k+1) is the battery current predicted at the next sampling moment; T s is the discrete sampling time L represents the energy storage inductance. ;

[0076] The inner loop current prediction model of the energy storage unit in Buck mode is as follows:

[0077]

[0078] Energy storage unit current inner loop prediction objective function:

[0079] J=|i b (k+n)-i b-ref |,

[0080] Where: i b (k+n) is the energy storage unit current when the prediction step size is n; i b-ref It is the reference value of the battery inner loop current.

[0081] The present invention first adds virtual inertia control on the basis of the voltage outer loop, wherein the virtual capacitor in the virtual inertia control is obtained by optimizing the calculation according to the current system state through the MPC controller; then, through current sharing control, that is, introducing a weight factor to distribute the current reference value to the converter of the energy storage unit and the AC power grid, and at the same time, adopting a multi-step FCS-MPC controller in the current inner loop control of the energy storage unit to meet the rapidity of the inner loop current tracking. Finally, the B-DC converter of the energy storage unit and the G-VSC of the AC power grid adjust the duty cycle according to the current reference value, thereby adjusting the output power. On the one hand, this method can obtain a flexible virtual capacitor value according to the real-time voltage state, thereby providing the system with inertia that meets the needs; on the other hand, it improves the current error tracking rate of the current inner loop, speeds up the time for the voltage to return to steady state, and improves the transient response process of the DC voltage while ensuring the quality of the DC bus voltage.

[0082] The above is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. It should be pointed out that for technicians in this technical field, any modification or equivalent replacement of the technical solution of the present invention without creative work does not depart from the scope of protection of the technical solution of the present invention.

Claims

1. A DC bus voltage control method based on model predictive control and current sharing control, characterized in that: The specific steps include: S1: When the system load suddenly changes and the voltage deviates from the normal operating range, the voltage outer loop obtains the current reference value i through PI control. d The virtual inertia control link calculates the virtual capacitor current reference value i based on the optimal virtual capacitor parameters obtained by model predictive control. vir The combination of the two gives the reference current i for inner loop control. ref ; S2: Considering the large inertia of the AC grid, when the energy storage output is insufficient, the AC grid can also provide inertia for the system through additional virtual inertia control. Therefore, the current reference value i of the energy storage system current inner loop control is obtained through current sharing control. b-ref and the current reference value i of the G-VSC converter of the AC grid g-ref ; S3: To avoid the hysteresis of the traditional current inner loop control based on PI control, a multi-step MPC controller is used in the current inner loop to achieve fast inner loop current tracking of the converter; S4: The bidirectional DC / DC converter of the energy storage system and the AC-DC converter of the AC system adjust the power output to the DC bus according to the duty cycle obtained from the current reference value, thereby regulating the DC bus voltage.

2. The DC bus voltage control method based on model predictive control and current sharing control according to claim 2, characterized in that: The voltage outer loop control in step S1 retains the voltage link of the voltage-current dual closed-loop control commonly used in energy storage units. The key parameter in the additional virtual inertia control link, the size of the virtual capacitor, is obtained through MPC. MPC uses a rolling optimization mechanism to replace global optimization with optimization in a finite time domain. At each sampling moment, MPC first predicts the future trend of the system, then solves an optimization problem in a finite time domain, and finally applies the first set of elements of the obtained control sequence to the system. Among them, the optimization problem is the core of MPC, which includes three parts: prediction model, objective function and constraints. The specific form is shown in the following formula: In the formula, v(k+i) is the control quantity at time k+i; v min (k+i) is the minimum value of the control quantity at time k+i; v max (k+i) is the maximum value of the control quantity at time k+i; Δv(k+i) is the control increment at time k+i; Δv min (k+i) is the minimum value of the control increment at time k+i; Δv max (k+i) is the maximum value of the control increment at time k+i; y(k+i) is the output at time k+i; y min (k+i) is the minimum output value at time k+i; y max (k+i) is the maximum output at time k+i; Δx(k+i|k) is the state increment at time k+i predicted by time k; Δx(k+i+1|k) is the state increment at time k+i+1 predicted by time k; Δd(k+i) is the disturbance increment at time k+i; y(k+i|k) is the output at time k+i predicted by time k; J(x(k,ΔV)) is the objective function, which is a function of the state x(k) at time k and the control sequence ΔV, where ΔV = [Δv(k)Δv(k+1)…Δv(k+m-1)] T ; A, B, C, and D are coefficient matrices. Coefficient matrix A represents the system state transfer matrix, which describes the evolution of the system state variables over time. Coefficient matrix B represents the influence of control input on the system state variables. Coefficient matrix C represents the influence of disturbance variables on the system state. Coefficient matrix D links the system state to the output, and represents the relationship between the state variables and the measurable output.

3. The DC bus voltage control method based on model predictive control and current sharing control according to claim 2, characterized in that: The objective function is as follows: J=α|U dc (k+1)-U dcn |+(1-a)|sedU dc (k+1)| Where U dc (k+1) is the DC bus voltage value at time k+1; U dcn is the rated value of DC bus voltage; δU dc (k+1) is the rate of change of the DC bus voltage at time k+1; ε is the proportional coefficient, which is used to balance the numerical difference between the voltage deviation and the voltage change rate in units; α is the weight coefficient. In the early stage of the load mutation, the virtual capacitor quickly releases the current mainly to suppress the voltage change rate and prevent the voltage from dropping too quickly, so α = 0.1; during the dynamic voltage fluctuation process, the value of α gradually increases to 1. In this way, in the middle and late stages of the voltage change, the voltage deviation is mainly the main factor, that is, the voltage steady-state error is reduced to prevent the voltage drop from being too large.

4. The DC bus voltage control method based on model predictive control and current sharing control according to claim 3, characterized in that: The specific process of MPC control in step S1 is as follows: S11: Obtain real-time operating data of the DC power distribution system and update the prediction model; S12: Solve the optimization problem. If the optimization problem has a solution, output the control increment. If the problem has no solution, keep the control variable unchanged and exit MPC; S13: Output the control amount and perform virtual inertia control based on the obtained optimal parameters.

5. The DC bus voltage control method based on model predictive control and current sharing control according to claim 4, characterized in that: The specific process of current sharing control in step S2 is as follows: S21: First, collect the operating data of the energy storage unit, calculate its state of charge SOC, and establish the reference current i in turn. ref The weight factor γ of , the expression of the weight factor γ is: Where, SOC max The maximum state of charge allowed by the battery is set to 90%; SOC min The minimum state of charge allowed for the battery is set to 10%; S22: Set the DC bus current reference value i ref The current reference value i of the energy storage unit is obtained by reasonable allocation through the weight factor γ b-ref and the AC grid current reference value i g-ref , and then the inner loop current control link of the energy storage unit and the G-VSC converter of the AC grid respond to the allocated current reference value.

6. The DC bus voltage control method based on model predictive control and current sharing control according to claim 5, characterized in that: The steps of the current inner loop multi-step predictive control in step S3 are as follows: S31: First, the charging and discharging state of the energy storage unit is determined by judging the magnitude of the DC bus voltage. After determining the working state, the i is calculated according to the inner loop current prediction model in the Boost mode and the Buck mode. b (k+n); S32: Calculate the output results of the objective function according to each predicted value and compare their sizes, output the control amount that can minimize the objective function value, and finally the bidirectional B-DC converter operates according to the output control amount.

7. The DC bus voltage control method based on model predictive control and current sharing control according to claim 6, characterized in that: In the battery bidirectional DC / DC converter topology, L represents the energy storage inductor, g1 represents the switching state of the switch tube S1, and g2 represents the switching state of the switch tube S2; The inner loop current prediction model of the energy storage unit in the Boost mode in step S31 is as follows: Where U dc (k) is the DC bus voltage at the sampling moment, U bat (k) is the battery voltage at the sampling moment, i bat (k) is the battery current at the sampling moment; i bat (k+1) is the battery current predicted at the next sampling moment; T s is the discrete sampling time; The inner loop current prediction model of the energy storage unit in Buck mode is as follows: Energy storage unit current inner loop prediction objective function: J=|i b (k+n)-i b-ref | Where: i b (k+n) is the energy storage unit current when the prediction step size is n; i b-ref It is the reference value of the battery internal loop current.

Images