A V2G grid-side converter control method and electric vehicle charging system

Abstract

The present invention provides a V2G grid-side converter control method and an electric vehicle charging system, wherein the V2G grid-side converter includes a DC-side LC filter and an AC-side LC filter; the method includes: generating a state space model of the V2G grid-side converter based on a d-q coordinate system, and obtaining a first discrete time model of the AC-side LC filter and a second discrete time model of the DC-side LC filter based on the state space model; constructing a multi-objective cost function to track the AC-side current, the AC-side capacitor voltage and the DC-side voltage in the first discrete time model, and setting a third weight through an adaptive weight adjustment module; finally solving the multi-objective cost function to obtain an initial stabilization vector; performing secondary prediction based on the initial stabilization vector to obtain an optimal stabilization vector; and modulating the optimal stabilization vector through a support vector machine. The present invention can improve the voltage stability of an electric vehicle charging system.

Description

Technical Field

[0001] The present invention relates to the technical field of V2G charging pile control, and in particular to a V2G grid-side converter control method and an electric vehicle charging system. Background Art

[0002] V2G (Vehicle-to-grid) technology is an important solution for decarbonizing transportation and energy systems. It provides auxiliary services to the grid, such as voltage and frequency control, rotating and non-rotating reserves, and reducing peak load fluctuations, by controlling the flow of energy between electric vehicles and the grid, thereby enhancing the stability, reliability and efficiency of the grid. In the actual scenario of V2G technology, the grid-side converter converts DC power into AC power and acts as an interface between the battery of the plug-in electric vehicle and the grid. Therefore, the power electronic inverter is a key supporting technology for V2G applications.

[0003] At present, one of the challenges of grid-side converters is how to simultaneously solve the synchronization problem with the grid, track active and reactive power references, attenuate harmonic distortion, and suppress internal and external interference. The more common solution is to use direct power control methods to achieve power tracking, using active and reactive power as state variables, and implementing the direct power link with proportional integral, proportional resonance, sliding film control, fuzzy logic, state feedback, and model predictive control, such as MPC (Model Predictive Control, MATLAB Model Predictive Control) controller, which uses the mathematical model of the system to predict its future behavior, and then determines the optimal control parameters to be applied based on the objective function, which is suitable for dealing with multivariable problems. However, the performance of the above-mentioned types of converters will be degraded due to reasons such as stability problems caused by system parameter mismatch or huge fluctuations in grid voltage. To meet this challenge, the known solutions include: 1. Designing a super-torsion integral sliding mode control strategy for three-phase voltage source converters, which shows robustness and maintains low harmonic current in the case of filter parameter drift. 2. A state feedback controller with integral action and based on linear matrix inequality for three-phase bidirectional chargers for electric vehicles. 3. Tracking the current reference signal through a repetitive controller. 4. Model predictive control method using permanent magnet synchronous motor drive.

[0004] However, the above scheme, on the one hand, only considers the AC side performance of the grid-side converter, and has not yet considered the stability problem caused by the multi-level connection of the electric vehicle charging system; on the other hand, the traditional model predicts variable frequency control, which makes the filter design of the V2G grid-side converter difficult. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a V2G grid-side converter control method and an electric vehicle charging system, which solve the problems in the prior art that the grid-side converter ignores the AC side performance when predicting and selecting control parameters, cannot ensure the voltage stability of the electric vehicle charging system, and the filter design of the V2G grid-side converter is difficult to cooperate with the model prediction method it adopts.

[0006] According to a first aspect of an embodiment of the present invention, there is provided a V2G grid-side converter control method, wherein the V2G grid-side converter includes a DC-side LC filter and an AC-side LC filter;

[0007] The V2G grid-side converter control method comprises:

[0008] Generate a state space model of the V2G grid-side converter based on a dq coordinate system, and obtain a first discrete time model of an AC-side LC filter and a second discrete time model of a DC-side LC filter based on the state space model;

[0009] Constructing a multi-objective cost function to track the AC side current, the AC side capacitor voltage in the first discrete time model, and the DC side voltage in the second discrete time model; wherein the multi-objective cost function includes a first weight of the AC side current, a second weight of the AC side capacitor voltage, and a third weight of the DC side voltage, and the third weight is set by an adaptive weight adjustment module;

[0010] Solving the multi-objective cost function to obtain an initial stabilization vector;

[0011] Performing secondary prediction based on the initial stabilization vector to obtain an optimal stabilization vector;

[0012] The optimal stabilization vector is modulated by a support vector machine to obtain N switch control signals, and the N switch control signals are used to control the switch states of N switches in the V2G grid-side converter, where N is a positive integer.

[0013] Optionally, based on the state space model, obtaining a first discrete time model of the AC side LC filter and a second discrete time model of the DC side LC filter includes:

[0014] Discretizing the state space model by a zero-order hold method to obtain the first discrete time model;

[0015] The second discrete-time model is obtained based on the capacitor in the DC-side LC filter.

[0016] Optionally, the state space model of the V2G grid-side converter is formulated as follows:

[0017] ;

[0018] in, i L,dq is the inductor current of the DC side filter, i L,d and i L,q are respectively the projection of the inductor current of the DC side filter on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; K abc→dq It represents the transformation process from the three-phase stationary coordinate system to the dq coordinate system; i L,a 、i L,b and i L,c are respectively the projection of the inductor current of the DC side filter on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; v C,dq is the capacitor voltage of the LC filter on the AC side, v C,d and v C,q are the projection of the capacitor voltage of the LC filter on the AC side on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system, respectively. v C,a 、v C,b and v C,c are respectively the projection of the capacitor voltage of the LC filter on the AC side on the a-axis of the three-phase stationary coordinate system, the projection on the b-axis of the three-phase stationary coordinate system, and the projection on the c-axis of the three-phase stationary coordinate system; i o,dq represents the load current of the load connected to the V2G grid-side converter, i o,d and i o,q are the projection of the load current on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system respectively; i o,a , i o,b and i o,c They are respectively the projection of the load current on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; v i,dq represents the load voltage of the load connected to the V2G grid-side converter, v i,d and v i,qare the projection of the load voltage on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system respectively; v i,a , v i,b and v i,c They are respectively the projection of the load voltage on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system;

[0019] The first discrete time model is formulated as follows:

[0020] ;

[0021] Where, k represents the kth moment;

[0022] in, , , , ;

[0023] Among them, a 11 ,a 12 ,a 21 ,a 22 Indicates that the value of the two-dimensional matrix is ​​equal to , b 11 ,b 12 ,b 21 ,b 22 Indicates that the value of the two-dimensional matrix is ​​equal to , L represents the inductance parameter of the LC filter on the AC side, and C represents the capacitance parameter of the LC filter on the AC side;

[0024] Furthermore, the discretization form based on the first discrete-time model is:

[0025] ;

[0026] Wherein, Vref represents the reference voltage, ωref represents the reference angular frequency, and Ts represents a sampling period;

[0027] The second discrete time model is formulated as follows:

[0028] ;

[0029] in, C dc Indicates the capacitance parameters of the DC side filter, v dc Indicates the DC voltage of the DC side filter.

[0030] Optionally, a multi-objective cost function is constructed, the formula is:

[0031] ;

[0032] ;

[0033] in, ; ;

[0034] in, G v represents the target cost function for tracking the capacitor voltage of the AC side LC filter, G i represents the target cost function for tracking the inductor current of the DC side filter, G dc represents the target cost function for tracking the DC voltage of the DC side filter; the first weight is 1 and the second weight is W i , the third weight is W dc ; i L,dq is the inductor current of the DC side filter, i L,d and i L,q are respectively the projection of the inductor current of the DC side filter on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; v C,dq is the capacitor voltage of the LC filter on the AC side, v C,d and v C,q are respectively the projection of the capacitor voltage of the LC filter on the AC side on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; v dc It represents the DC voltage of the DC side filter, and Ts represents the sampling period.

[0035] Optionally, the adaptive weight adjustment module adjusts the third weight by outputting an adaptive weight;

[0036] The calculation formula of the adaptive weight is:

[0037] ;

[0038] in, W adc represents the adaptive weight, h = Δ v C / v Callow , v CallowIndicates the preset capacitor voltage error of the AC side LC filter, v dcallow Indicates the preset DC voltage error of the DC side filter, Δ v dc Indicates the oscillation ripple of the DC voltage of the AC side LC filter, Δ v C Indicates the amplitude error of the DC voltage of the DC side filter.

[0039] Optionally, before solving the multi-objective cost function and obtaining the initial stabilization vector, the method includes:

[0040] The multi-objective cost function is represented by the sum of a first function and a second function, the independent variable of the first function is a d-axis, and the independent variable of the second function is a q-axis, wherein the d-axis and the q-axis are coordinate axes of a dq coordinate system;

[0041] The second-order partial derivative of the first function and the second-order partial derivative of the second function are analyzed to verify the existence of the minimum multi-objective cost function.

[0042] Optionally, solving the multi-objective cost function to obtain an initial stabilization vector includes:

[0043] When the minimum multi-objective cost function exists, an initial stabilization vector is calculated based on the multi-objective cost function represented by the first discrete-time model, the second discrete-time model, and the sum of the first function and the second function.

[0044] Optionally, the multi-objective cost function represented by the sum of the first function and the second function is expressed as follows:

[0045] ;

[0046] in, , ;

[0047] in, G d is the first function, G is the second function, G v,d express G v The d-axis component of G i,d express G i The d-axis component of G dc,d express G dc The d-axis component of G v,q express G vThe q-axis component of G i,q express G i The q-axis component of G dc,q express G dc The q-axis component of W i is the second weight, W dc The third weight.

[0048] Optionally, the multi-objective cost function is solved to obtain an initial stabilization vector, which is formulated as follows:

[0049] ;

[0050] in, , , , , , , ;

[0051] in, , , , ; a 11 ,a 12 ,a 21 ,a 22 Indicates that the value of the two-dimensional matrix is ​​equal to , b 11 ,b 12 ,b 21 ,b 22 Indicates that the value of the two-dimensional matrix is ​​equal to , L represents the inductance parameter of the LC filter on the AC side, and C represents the capacitance parameter of the LC filter on the AC side; W i for The second weight, W dc is the third weight, i L,dq is the inductor current of the DC side filter, v C,dq is the capacitor voltage of the LC filter on the AC side, v dc represents the DC voltage of the DC side filter, i o,dq It represents the load current of the load connected to the V2G grid-side converter, and Ts represents a sampling period.

[0052] Optionally, performing secondary prediction based on the initial stabilization vector to obtain an optimal stabilization vector includes:

[0053] The reference value and predicted value of the initial stabilization vector at time k+2 are used to replace the reference value and predicted value of the initial stabilization vector at time k+1 to obtain the optimal stabilization vector.

[0054] A second aspect provides an electric vehicle charging system, comprising a converter control device and a V2G grid-side converter; the V2G grid-side converter comprises a DC side LC filter and an AC side LC filter, and the converter control device is connected between the DC side LC filter and the AC side LC filter;

[0055] The converter control device executes any one of the above-mentioned V2G grid-side converter control methods.

[0056] The present invention has the following beneficial effects: constructing a multi-cost objective function to track the unstable factors that may be caused by the multi-level connection of the electric vehicle charging system, adjusting the DC voltage and the AC voltage, and flexibly adjusting various weights at the same time, automatically adjusting the DC side and AC side performance of the system under disturbance to obtain the optimal stabilization vector. Finally, the optimal stabilization vector is modulated by SVM to obtain the optimal control method, that is, the control method of the N switches in the V2G grid-side converter. In addition, in the above process, variable frequency control is not used for the DC side LC filter and the AC side LC filter, thereby reducing the difficulty of filter design of the V2G grid-side converter. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Schematic diagram of the implementation flow of the V2G grid-side converter control method according to an embodiment of the present invention;

[0058] Figure 2 This is a schematic diagram of the structure of an electric vehicle charging system according to an embodiment of the present invention;

[0059] Figure 3 A simulation result diagram of a traditional model predictive control according to an embodiment of the present invention;

[0060] Figure 4 It is a simulation result diagram of a V2G grid-side converter control method and an electric vehicle charging system according to an embodiment of the present invention;

[0061] Figure 5 This is another simulation result diagram of the V2G grid-side converter control method and electric vehicle charging system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0063] like Figure 1 As shown, an embodiment of the present invention proposes a V2G grid-side converter control method, wherein the V2G grid-side converter includes a DC side LC filter and an AC side LC filter. The V2G grid-side converter control method includes but is not limited to the following steps:

[0064] S1. Generate a state space model of the V2G grid-side converter based on a dq coordinate system, and obtain a first discrete time model of the AC-side LC filter and a second discrete time model of the DC-side LC filter based on the state space model;

[0065] S2. Construct a multi-objective cost function to track the AC side current, the AC side capacitor voltage in the first discrete time model, and the DC side voltage in the second discrete time model; wherein the multi-objective cost function includes a first weight of the AC side current, a second weight of the AC side capacitor voltage, and a third weight of the DC side voltage, and the third weight is set by an adaptive weight adjustment module;

[0066] S3, solving the multi-objective cost function to obtain an initial stabilization vector;

[0067] S4, performing secondary prediction based on the initial stabilization vector to obtain an optimal stabilization vector;

[0068] S5. Modulate the optimal stabilization vector through a support vector machine to obtain N switch control signals, where the N switch control signals are used to control the switch states of the N switches in the V2G grid-side converter.

[0069] In the above step S1, the power part of the V2G grid-side converter is model analyzed to obtain a state space model. In a specific application, the dynamic equation of the AC side LC filter of the V2G grid-side converter is represented in the three-phase stationary coordinate system abc as follows:

[0070] (1)

[0071] Among them, the AC side LC filter is an LC filter, L and C represent the inductance parameters and capacitance parameters of the AC side LC filter, iL, vC, vi, and io represent the inductance current, the capacitance voltage, the voltage vector of the V2G grid-side converter, and the load current, respectively.

[0072] In a specific application, the dynamic equation of the DC side filter of the V2G grid-side converter is expressed in the dq coordinate system, and the power loss of the switch of the V2G grid-side converter is ignored. The dynamic equation is:

[0073] (2)

[0074] Wherein, vs is the voltage from the passive front-end interface, Ldc and Cdc are filters on the DC side, vdc and idc are DC voltage and DC current respectively, iL,dq is the inductor current on the DC side, and vi,dq is the voltage vector on the AC side.

[0075] It should be noted that, since the DC inductance is relatively large and the DC source voltage and the DC link voltage are continuous, the DC current is regarded as a disturbance. On this basis, the embodiment of the present invention assumes that the DC current remains constant between two sampling periods.

[0076] In the embodiment of the present invention, in order to simplify the calculation, the dynamic equation of the AC side LC filter of the V2G grid-side converter represented in the three-phase stationary coordinate system is converted to the dq coordinate system, thereby obtaining the state space model of the V2G grid-side converter in the above step S1. It should be noted that the dq coordinate system in the embodiment of the present invention is a coordinate system based on the dq axis of the motor, wherein the d axis represents the DC component and the q axis represents the AC component.

[0077] First, the dynamic equation of the V2G grid-side converter in the dq coordinate system is:

[0078] (3)

[0079] Then, based on the above dynamic equation, the state space model of the V2G grid-side converter is as follows:

[0080] (4)

[0081] in, i L,dq is the inductor current of the DC side filter, i L,d and i L,q are respectively the projection of the inductor current of the DC side filter on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; K abc→dq It represents the transformation process from the three-phase stationary coordinate system to the dq coordinate system; i L,a 、i L,b and i L,c are respectively the projection of the inductor current of the DC side filter on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; v C,dq is the capacitor voltage of the LC filter on the AC side, v C,d and v C,qare the projection of the capacitor voltage of the LC filter on the AC side on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system, respectively. v C,a 、v C,b and v C,c are respectively the projection of the capacitor voltage of the LC filter on the AC side on the a-axis of the three-phase stationary coordinate system, the projection on the b-axis of the three-phase stationary coordinate system, and the projection on the c-axis of the three-phase stationary coordinate system; i o,dq represents the load current of the load connected to the V2G grid-side converter, i o,d and i o,q are the projection of the load current on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system respectively; i o,a , i o,b and i o,c They are respectively the projection of the load current on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; v i,dq represents the load voltage of the load connected to the V2G grid-side converter, v i,d and v i,q are the projection of the load voltage on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system respectively; v i,a , v i,b and v i,c They are respectively the projection of the load voltage on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system;

[0082] In a preferred implementation, the state space model is discretized by a zero-order hold method to obtain the first discrete time model. Exemplarily, the first discrete time model is formulated as follows:

[0083] (5)

[0084] Where, k represents the kth moment;

[0085] in, , , , ;

[0086] Among them, a11 ,a 12 ,a 21 ,a 22 Indicates that the value of the two-dimensional matrix is ​​equal to , b 11 ,b 12 ,b 21 ,b 22 Indicates that the value of the two-dimensional matrix is ​​equal to , L represents the inductance parameter of the LC filter on the AC side, and C represents the capacitance parameter of the LC filter on the AC side;

[0087] Furthermore, the discretization form based on the first discrete-time model is:

[0088] (6)

[0089] Among them, V ref represents the reference voltage, ω ref represents the reference angular frequency, Ts represents a sampling period;

[0090] Since the DC current remains constant between two sampling periods, the second discrete time model of the DC side LC filter can be directly expressed as:

[0091] (7)

[0092] in, C dc Indicates the capacitance parameters of the DC side filter, v dc Indicates the DC voltage of the DC side filter.

[0093] It should be noted that the fixed frequency control technology used in the above process has lower requirements on the filter design of the V2G grid-side converter compared to variable frequency control.

[0094] For the above step S2, in a specific application, for an electric vehicle charging system based on a V2G grid-side converter, it is very important to introduce a cost function to simultaneously adjust the DC voltage and the AC voltage, thereby improving the output voltage quality of the V2G grid-side converter in the electric vehicle charging system. The multi-objective cost function constructed in the embodiment of the present invention also considers the error of the DC voltage relative to the steady-state value, and tracks the AC side capacitor voltage, the inductor current, and the DC side voltage.

[0095] Exemplarily, the constructed multi-objective cost function is formulated as follows:

[0096] (8)

[0097] (9)

[0098] in, ; ;

[0099] in, G v represents the target cost function for tracking the capacitor voltage of the AC side LC filter, G i represents the target cost function for tracking the inductor current of the DC side filter, G dc represents the target cost function for tracking the DC voltage of the DC side filter; the first weight is 1 and the second weight is W i , the third weight is W dc ; i L,dq is the inductor current of the DC side filter, i L,d and i L,q are respectively the projection of the inductor current of the DC side filter on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; v C,dq is the capacitor voltage of the LC filter on the AC side, v C,d and v C,q are respectively the projection of the capacitor voltage of the LC filter on the AC side on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; v dc It represents the DC voltage of the DC side filter, and Ts represents the sampling period.

[0100] The multi-objective cost function constructed in the embodiment of the present invention can track the AC side current, the AC side band voltage and the DC side voltage, which not only ensures the AC side performance of the V2G grid-side converter, but also takes into account the stable operation of the DC side of the V2G system, thereby improving the reliability of battery charging and discharging in the electric vehicle charging system and avoiding stability problems caused by multi-stage connection of the electric vehicle charging system.

[0101] In specific applications, adjusting the first weight, the second weight and the third weight can effectively improve the model predictive control effect of the V2G grid-side converter. An embodiment of the present invention provides a flexible and stable adjustment method, that is, outputting an adaptive weight to adjust the third weight through an adaptive weight adjustment module.

[0102] Exemplarily, the calculation formula of the adaptive weight is:

[0103] (10)

[0104] in, W adc represents the adaptive weight, h = Δ v C / v Callow , v Callow Indicates the preset capacitor voltage error of the AC side LC filter, v dcallow Indicates the preset DC voltage error of the DC side filter, Δ v dc Indicates the oscillation ripple of the DC voltage of the AC side LC filter, Δ v C Indicates the amplitude error of the DC voltage of the DC side filter.

[0105] The above adaptive weight adjustment module automatically adjusts W according to the oscillation dc The size of the control unit can reduce the impact on communication while maintaining stability.

[0106] For the above step S3, the embodiment of the present invention solves the multi-objective cost function and obtains the minimum value of the multi-objective cost function. By minimizing the multi-objective cost function, the optimal control parameters can be found, thereby achieving simultaneous regulation of the DC voltage and the AC voltage. Therefore, before solving, that is, before the above step S3, it is also necessary to verify the existence of the minimum multi-objective cost function.

[0107] In a preferred implementation, the embodiment of the present invention verifies the existence of the minimum multi-objective cost function through the zero value of the second-order partial derivative, including:

[0108] The multi-objective cost function is represented by the sum of a first function and a second function, the independent variable of the first function is a d-axis, and the independent variable of the second function is a q-axis, wherein the d-axis and the q-axis are coordinate axes of a dq coordinate system;

[0109] The second-order partial derivative of the first function and the second-order partial derivative of the second function are analyzed to verify the existence of the minimum multi-objective cost function.

[0110] Exemplarily, the multi-objective cost function represented by the sum of the first function and the second function is expressed as follows:

[0111] (11)

[0112] in, , ;

[0113] in, G d is the first function, G is the second function,G v,d express G v The d-axis component of G i,d express G i The d-axis component of G dc,d express G dc The d-axis component of G v,q express G v The q-axis component of G i,q express G i The q-axis component of G dc,q express G dc The q-axis component of W i is the second weight, W dc The third weight.

[0114] For example, the second-order partial derivative of the above formula can be expressed as:

[0115] (12)

[0116] (13)

[0117] According to the above equations (9) and (11), the above equation (13) holds.

[0118] According to the above steps, when the minimum multi-objective cost function exists, the implementation of the above step S3 is:

[0119] An initial stabilization vector is calculated based on the first discrete-time model, the second discrete-time model, and a multi-objective cost function represented by a sum of a first function and a second function.

[0120] In the embodiment of the present invention, the calculation of the above steps is: Substitute (5) to (6) into (11) to obtain the expression of the initial stabilization vector OSV. Exemplarily, the initial stabilization vector is obtained as follows:

[0121] (14)

[0122] in, , , , , , , ;

[0123] in, , , , ; a 11 ,a 12 ,a 21 ,a 22 Indicates that the value of the two-dimensional matrix is ​​equal to , b 11 ,b 12 ,b 21 ,b 22 Indicates that the value of the two-dimensional matrix is ​​equal to , L represents the inductance parameter of the LC filter on the AC side, and C represents the capacitance parameter of the LC filter on the AC side; W i for The second weight, W dc is the third weight, i L,dq is the inductor current of the DC side filter, v C,dq is the capacitor voltage of the LC filter on the AC side, v dc represents the DC voltage of the DC side filter, i o,dq It represents the load current of the load connected to the V2G grid-side converter, and Ts represents a sampling period.

[0124] The above step S4 processes the initial stabilization vector obtained in the above step S3 to avoid control delay in practical applications, including:

[0125] The reference value and predicted value of the initial stabilization vector at time k+2 are used to replace the reference value and predicted value of the initial stabilization vector at time k+1 to obtain the optimal stabilization vector.

[0126] According to the initial stabilization vector shown in the above formula (14), the embodiment of the present invention also provides a formula for the optimal stabilization vector:

[0127] (15)

[0128] Among them, the reference value v C,dq *(k+2),i L,dq *(k+2),v dc *(k+2) can be obtained by a third-order Lagrangian extrapolation strategy and a one-step delay.

[0129] For the above step S5, the embodiment of the present invention uses support vector machines (SVM) to modulate the optimal stabilization vector, so as to obtain the optimal control method according to the optimal stabilization vector, that is, the control method of the N switches in the V2G grid-side converter.

[0130] Through the above steps S1 to S5, the embodiment of the present invention constructs a multi-cost objective function to track the unstable factors that may be caused by the multi-level connection of the electric vehicle charging system, adjust the DC voltage and the AC voltage, and flexibly adjust the weights at the same time, automatically adjust the DC side and AC side performance of the system under disturbance to obtain the optimal stabilization vector. Finally, the optimal stabilization vector is modulated by SVM to obtain the optimal control method, that is, the control method of the N switches in the V2G grid-side converter. In addition, in the above process, variable frequency control is not used for the DC side LC filter and the AC side LC filter, thereby reducing the difficulty of filter design of the V2G grid-side converter.

[0131] like Figure 2 As shown, another embodiment of the present invention provides an electric vehicle charging system, including a converter control device 10 and a V2G grid-side converter 20; the V2G grid-side converter 20 includes a DC side LC filter 21 and an AC side LC filter 22, and the converter control device 10 is connected between the DC side LC filter 21 and the AC side LC filter 22; the converter control device 10 executes the V2G grid-side converter control method in the above embodiment.

[0132] Figure 2 In the embodiment, the converter control device 10 obtains the current and voltage data of the AC side LC filter 22, including i based on the three-term stationary coordinate system. La ,i Lb ,i Lc , v Ca , v Cb , v Cc ,i oa ,i ob ,i oc , and the above parameters are transformed from the three-phase stationary coordinate system abc to the dq coordinate system, and finally i is obtained L,dq 、v C,dq and i o,dq The converter control device 10 obtains the current and voltage data of the DC side LC filter 21, including i dc and v dc In the converter control device 10, i L,dq 、v C,dq and v dcThe first discrete time model and the second discrete time model are obtained by the discrete method shown in (5) and (7). Then, the multi-objective cost function is constructed using (9), and the minimum multi-objective cost function is solved by (11). The initial stabilization vector is obtained according to the minimized cost function and the optimal stabilization vector OSV obtained after delay processing is represented by (15). It should be noted that the solution of (8) and (9) in the above embodiment does not involve iteration. According to (15), 11 operations are required to obtain OSV, which reduces the number of operations and simplifies the implementation process. In addition, since the modulator SVM modulates the optimal stabilization vector OSV and outputs the switch control signal Sa, the switch control signal Sb and the switch control signal Sc, the V2G grid-side converter control method proposed in the embodiment of the present invention is essentially a constant switching frequency control method.

[0133] The embodiment of the present invention also provides an actual simulation case, combining the traditional model predictive control to illustrate the application effect of the V2G grid-side converter control method and electric vehicle charging system provided by the embodiment of the present invention. First, the simulation parameters are as follows: Input voltage v s = 300V, DC side LC filter: L dc = 5mH, C dc = 30µF, AC side LC filter: L =2.4mH, C = 100µF, Linear Load R = 40Ω, switching frequency f = 10kHz, sampling period Ts = 25µs, reference voltage amplitude V ref = 110V, reference angular frequency ω ref = 120π rad / s. The simulation results are as follows Figure 3 and Figure 4 As shown, Figure 3 To use the simulation results of traditional model predictive control (MPC), Figure 4 The simulation results of the V2G grid-side converter control method and electric vehicle charging system provided by the embodiment of the present invention, the horizontal axis is time, the unit is s, and the vertical axis is voltage v C and v dc , the unit is V. According to Figure 3 and Figure 4 , it can be seen that Figure 3 It shows that there is DC link oscillation, which threatens the normal operation of the electric vehicle charging system. Figure 4In addition, the embodiment of the present invention also simulates the V2G grid-side converter control method and electric vehicle charging system provided in the case of a single-phase fault on the AC side at 0.01 seconds. The simulation results are as follows: Figure 5 As shown, the horizontal axis is time t, in seconds, and the vertical axis is voltage v C and v dc , the unit is V. According to Figure 5 It can be seen that, for the DC side, the voltage of the grid-side converter stabilizes quickly due to the control action after the fault; for the grid side, fast dynamic response is also achieved, and the deviation between the voltage and the reference voltage is small.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A V2G grid-side converter control method, characterized in that: The V2G grid-side converter includes a DC-side LC filter and an AC-side LC filter; The V2G grid-side converter control method comprises: Generate a state space model of the V2G grid-side converter based on a dq coordinate system, and obtain a first discrete time model of an AC-side LC filter and a second discrete time model of a DC-side LC filter based on the state space model; Constructing a multi-objective cost function to track the AC side current, the AC side capacitor voltage in the first discrete time model, and the DC side voltage in the second discrete time model; wherein the multi-objective cost function includes a first weight of the AC side current, a second weight of the AC side capacitor voltage, and a third weight of the DC side voltage, and the third weight is set by an adaptive weight adjustment module; Solving the multi-objective cost function to obtain an initial stabilization vector; Performing secondary prediction based on the initial stabilization vector to obtain an optimal stabilization vector; Modulating the optimal stabilization vector by a support vector machine to obtain N switch control signals, where the N switch control signals are used to control the switch states of N switches in the V2G grid-side converter, where N is a positive integer; The dynamic equation of the AC side LC filter of the V2G grid-side converter is represented in the three-phase stationary coordinate system abc, and the dynamic equation of the DC side filter of the V2G grid-side converter is represented in the dq coordinate system. The dynamic equation of the AC side LC filter of the V2G grid-side converter represented in the three-phase stationary coordinate system is converted to the dq coordinate system to obtain the state space model; Discretizing the state space model by a zero-order hold method to obtain the first discrete time model; The second discrete-time model is obtained based on the capacitor in the DC-side LC filter.

2. The V2G grid-side converter control method according to claim 1, characterized in that: The state space model of the V2G grid-side converter is as follows: in, i L,dq is the inductor current of the DC side filter, i L,d and i L,q are respectively the projection of the inductor current of the DC side filter on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; K abc→dq It represents the transformation process from the three-phase stationary coordinate system to the dq coordinate system; i L,a 、i L,b and i L,c are respectively the projection of the inductor current of the DC side filter on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; v C,dq is the capacitor voltage of the LC filter on the AC side, v C,d and v C,q are the projection of the capacitor voltage of the LC filter on the AC side on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system, respectively. v C,a 、v C,b and v C,c are respectively the projection of the capacitor voltage of the LC filter on the AC side on the a-axis of the three-phase stationary coordinate system, the projection on the b-axis of the three-phase stationary coordinate system, and the projection on the c-axis of the three-phase stationary coordinate system; i o,dq represents the load current of the load connected to the V2G grid-side converter, i o,d and i o,q are the projection of the load current on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system respectively; i o,a , i o,b and i o,c They are respectively the projection of the load current on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; v i,dq Indicates the load voltage of the load connected to the LC filter on the AC side, v i,d and v i,q are the projection of the load voltage on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system respectively; v i,a , v i,b and v i,c They are respectively the projection of the load voltage on the a-axis of the three-phase stationary coordinate system, the projection of the b-axis of the three-phase stationary coordinate system, and the projection of the c-axis of the three-phase stationary coordinate system; The first discrete time model is formulated as follows: ； Where, k represents the kth moment; in, , , , ; Among them, a 11 ,a 12 ,a 21 ,a 22 Indicates that the value of the two-dimensional matrix is ​​equal to , b 11 ,b 12 ,b 21 ,b 22 Indicates that the value of the two-dimensional matrix is ​​equal to , L represents the inductance parameter of the LC filter on the AC side, and C represents the capacitance parameter of the LC filter on the AC side; Furthermore, the discretization form based on the first discrete-time model is: ； Among them, V ref represents the reference voltage, ω ref represents the reference angular frequency, Ts represents a sampling period; The second discrete time model is formulated as follows: ； in, C dc Indicates the capacitance parameters of the DC side filter, v dc represents the DC voltage of the DC side filter, i dc is a direct current.

3. The V2G grid-side converter control method according to claim 1, characterized in that: The constructed multi-objective cost function is formulated as follows: ； ； in, ; ; in, G v represents the target cost function for tracking the capacitor voltage of the AC side LC filter, G i represents the target cost function for tracking the inductor current of the DC side filter, G dc represents the target cost function for tracking the DC voltage of the DC side filter; the first weight is 1 and the second weight is W i , the third weight is W dc ; i L,dq is the inductor current of the DC side filter, i L,d and i L,q are respectively the projection of the inductor current of the DC side filter on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; v C,dq is the capacitor voltage of the LC filter on the AC side, v C,d and v C,q are respectively the projection of the capacitor voltage of the LC filter on the AC side on the d-axis of the dq coordinate system and the projection of the q-axis of the dq coordinate system; v dc represents the DC voltage of the DC side filter, Ts represents the sampling period, and k represents the kth moment.

4. The V2G grid-side converter control method according to claim 3, characterized in that: The adaptive weight adjustment module adjusts the third weight by outputting an adaptive weight; The calculation formula of the adaptive weight is: ； in, W adc represents the adaptive weight, h = Δ v C / v Callow , v Callow Indicates the preset capacitor voltage error of the AC side LC filter, v dcallow Indicates the preset DC voltage error of the DC side filter, Δ v dc Indicates the oscillation ripple of the DC voltage of the AC side LC filter, Δ v C Represents the amplitude error of the DC voltage of the DC side filter, V ref Indicates the reference voltage.

5. The V2G grid-side converter control method according to claim 1, characterized in that: Solving the multi-objective cost function and obtaining the initial stabilization vector includes: The multi-objective cost function is represented by the sum of a first function and a second function, the independent variable of the first function is a d-axis, and the independent variable of the second function is a q-axis, wherein the d-axis and the q-axis are coordinate axes of a dq coordinate system; Analyze the second-order partial derivative of the first function and the second-order partial derivative of the second function to verify the existence of the minimum multi-objective cost function; The formula of the multi-objective cost function represented by the sum of the first function and the second function is: ； in, , ; in, G d is the first function, G is the second function, G v,d express G v The d-axis component of G i,d express G i The d-axis component of G dc,d express G dc The d-axis component of G v,q express G v The q-axis component of G i,q express G i The q-axis component of G dc,q express G dc The q-axis component of W i is the second weight, W dc The third weight.

6. The V2G grid-side converter control method according to claim 5, characterized in that: Solving the multi-objective cost function to obtain an initial stabilization vector includes: When the minimum multi-objective cost function exists, an initial stabilization vector is calculated based on the multi-objective cost function represented by the first discrete-time model, the second discrete-time model, and the sum of the first function and the second function.

7. The V2G grid-side converter control method according to claim 1, characterized in that: Solving the multi-objective cost function, the initial stabilization vector is obtained, and the formula is: ； in, , , , , , , ; in, , , , ; a 11 ,a 12 ,a 21 ,a 22 Indicates that the value of the two-dimensional matrix is ​​equal to , b 11 ,b 12 ,b 21 ,b 22 Indicates that the value of the two-dimensional matrix is ​​equal to , L represents the inductance parameter of the LC filter on the AC side, and C represents the capacitance parameter of the LC filter on the AC side; W i for The second weight, W dc is the third weight, i L,dq is the inductor current of the DC side filter, v C,dq is the capacitor voltage of the LC filter on the AC side, v dc represents the DC voltage of the DC side filter, i o,dq represents the load current of the load connected to the V2G grid-side converter, Ts represents a sampling period, k represents the kth moment, V dc Indicates the DC voltage of the DC side filter, C dc Indicates the capacitance parameter of the DC side filter, i dc Indicates direct current.

8. The V2G grid-side converter control method according to claim 1, characterized in that: Performing secondary prediction based on the initial stabilization vector to obtain an optimal stabilization vector includes: The reference value and predicted value of the initial stabilization vector at time k+2 are used to replace the reference value and predicted value of the initial stabilization vector at time k+1 to obtain the optimal stabilization vector.

9. An electric vehicle charging system, characterized in that: It includes a converter control device and a V2G grid-side converter; the V2G grid-side converter includes a DC side LC filter and an AC side LC filter, and the converter control device is connected between the DC side LC filter and the AC side LC filter; The converter control device executes the V2G grid-side converter control method according to any one of claims 1 to 8.

Images