Optimization control method of three-phase dual-active bridge isolation DC-DC converter system

Abstract

The invention discloses an optimal control method for a three-phase dual-active bridge isolated DC-DC converter system, which comprises the following steps of: constructing a steady-state model of a circuit, constructing an optimization problem cost function, inputting an optimization working condition parameter, selecting a particle swarm optimization parameter, initializing a particle swarm, and calculating a case circuit running state and a target cost function, storing the current-generation optimal cost function, sequentially judging whether particle swarm optimization quitting conditions and optimization indexes are met or not, and finally performing particle swarm optimization in a full control interval on the model to obtain optimal control parameters under a specific working condition; according to the invention, the circuit operation efficiency or the voltage and current stress of the element can be effectively optimized, the power density of the converter is improved, and then the optimal control parameters under each working condition are found in the three-degree-of-freedom space through the particle swarm optimization. And finally, controller parameters generated through optimization can cover the whole required working condition in a segmented one-time piecewise linear interpolation mode.

Description

Technical Field

[0001] The invention belongs to the technical field of power electronics, and in particular relates to an optimization control parameter calculation method suitable for a three-phase dual active bridge (3p-DAB) isolated DC-DC converter system. Background Art

[0002] The 3p-DAB converter has attracted widespread attention due to its advantages of galvanic isolation, bidirectional power conversion flow and easy implementation of zero voltage switching (ZVS). This technology can be applied to DC conversion devices, photovoltaic power generation systems, electric vehicle charging piles, etc.

[0003] The traditional three-phase DAB circuit uses a single phase shift control method (SPS) to control the output voltage and output power of the converter. This control method can meet the zero voltage switching of the input and output dual bridge arms under the condition of input and output voltage matching, so that the conversion efficiency of the converter can reach more than 95%.

[0004] However, when the input and output voltages do not match, this control method can only meet the full-bridge ZVS in a very small range where the power is close to the converter output limit. At the same time, since the switching frequency of the three-phase DAB is generally high and there are more switches in the circuit topology, the switching loss of the SPS is high and the efficiency is low under non-ideal conditions.

[0005] In order to solve the above problems and improve the working efficiency of the 3p-DAB converter under voltage mismatch conditions, choosing a control scheme with greater control freedom is a feasible option. DCC control adds two new control freedoms, namely the duty cycle of the front bridge arm (input bridge arm) and the rear bridge arm (output bridge arm), on the basis of single phase-shift control, so as to effectively control the internal operating state of the circuit while controlling the output power.

[0006] However, due to the difficulty in analyzing 3p-DAB under DCC control and the high degree of control freedom of the control method, it is difficult to implement an analytical calculation formula for the optimal control parameters. It becomes very complicated to analytically analyze the circuit and obtain the optimal control parameters. Summary of the invention

[0007] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to propose an optimization control method for a three-phase dual active bridge isolated DC-DC converter system.

[0008] The technical solution adopted by the present invention to solve the technical problem is: an optimization control method for a three-phase dual active bridge isolated DC-DC converter system, wherein the three-phase dual active bridge isolated DC-DC converter is composed of a front bridge arm and a rear bridge arm through a transformer and an inductor L sa , L sb , L scThe front bridge arm is connected by the switch tube S 14 ~S 16 The rear bridge arm is composed of switch tube S 21 ~S 26 The six switch tubes constitute the upper and lower bridge arms respectively; the steps are as follows:

[0009] S1, construct the cost function of the optimization problem and input the optimization condition parameters: use the piecewise linearization analysis method to analyze the inductance in steady-state operation in sections, calculate the inductor current in the entire interval through the switch timestamp, the front and rear bridge arm switch functions and the input and output voltages, and then calculate the output power. Through the above piecewise linear method, construct a three-phase DAB steady-state circuit model under DCC control;

[0010] S2, constructing a cost function composed of circuit parameters according to the optimization target: repackaging the steady-state circuit model to construct an optimizable function model whose input is a variable control parameter and whose output is a cost function corresponding to the optimization target;

[0011] S3, select particle swarm optimization parameters: select appropriate particle swarm optimization parameters according to the type of problem to be optimized;

[0012] S4, initializing the particle swarm: performing particle swarm optimization on the optimizable function model under the set working conditions. If the cost function corresponding to the optimized control parameter is higher than the cost function corresponding to the SPS control parameter under the same working condition, the optimization is considered invalid and the particle swarm optimization needs to be performed again.

[0013] S5, calculating the circuit model operation state and the target cost function;

[0014] S6, store the current optimal cost function, determine whether the conditions for exiting particle swarm optimization are met, otherwise perform a particle swarm iteration and return to S5;

[0015] S7, if yes, further determine whether the optimization index is met, otherwise return to S4;

[0016] S8, if yes, output the optimal control parameters and end the operation.

[0017] Further, the three-phase DAB steady-state circuit model is constructed by the following steps:

[0018] (1) Generate a three-phase front and rear bridge arm switch action timestamp sequence within a single cycle based on the three-degree-of-freedom control input. The timestamp element consists of an action time value and an action type label. Then, the timestamp sequence is sorted according to the action time. The A-phase switch action time is shown in the following formula:

[0019]

[0020] Where t Ta1 Indicates the time between the upper bridge arm switch tube of input A phase being turned on and the lower bridge arm switch tube being turned off, t Ta2 Indicates the time between the lower bridge arm switch tube of input A phase being turned on and the upper bridge arm switch tube being turned off, t Sa1 Indicates the time when the upper bridge arm switch of output phase A is turned on and the lower bridge arm switch is turned off, t Sa2 Indicates the time when the lower bridge arm switch of output phase A is turned on and the upper bridge arm switch is turned off; T s Represents the switching cycle, D1 and D2 are the switching duty ratios of the front and rear bridge arms respectively, and D f D is the ratio of the front bridge arm lagging behind the rear bridge arm. f =1 means that the front bridge arm lags behind the rear bridge arm by half a switching cycle, and n is the number of cycles from time zero to now; the switching action time of phase B lags behind phase A by 1 / 3 switching cycle, and the switching action time of phase C is ahead of phase A by 1 / 3 switching cycle. The expression is:

[0021] (2) According to the switching action timestamp, the entire switching cycle is partitioned, and the segmented switching functions of the front and rear bridge arms are drawn according to the sorted switching action timestamp labels:

[0022]

[0023] Where δ(t) is the impulse function. When t0∈[t1,t2], for any f(x) we have At the same time if but

[0024] (3) Calculate the inductance L based on the segmented switching function sa , L sb , L sc The voltage at both ends is: Where U dc Represents the DC side voltage of the three-phase bridge arm, S a , S b , S c They represent the switching functions of the front and rear bridge arms of the three-phases, respectively, and their values ​​are the switching functions of the front bridge arm of the corresponding phase minus the switching functions of the rear bridge arm of the corresponding phase, wherein for the switching functions of the front and rear bridge arms, when the corresponding upper bridge arm is turned on, it is 1, and when the corresponding lower bridge arm is turned on, it is -1;

[0025] (4) According to the inductance L sa , L sb , L sc The voltage at both ends and the duration of each small interval are used to calculate the change in inductor current in each small interval. The change in inductor current in the nth small interval formed by the A phase timestamp is: Where L is the equivalent inductance in the three-phase circuit, tn-1 and t n are the start time and end time of the cell respectively;

[0026] (5) Based on the inductor current change calculated in the previous step, the inductor current offset at the beginning of the switching cycle is calculated. The inductor current at the beginning of the cycle is Where t0 represents the start time of the cycle, I′(t) represents that the inductor current is 0 at the start of the cycle, and therefore the inductor current in the entire switching interval is I(t)=I0+I′(t).

[0027] Furthermore, the output charge in a single switching cycle is calculated based on the inductor current and the switching function of the rear bridge arm. The output power expression is as follows:

[0028]

[0029] The above model parameters can be further transformed into an optimized target cost function.

[0030] Furthermore, the mathematical expression of the cost function optimization problem in step S5 is:

[0031] minimize f cost

[0032] subject to P o =P ref

[0033] others,

[0034] Where f cost Represents the cost function; the cost function constructed by integrating the constraints into the cost function is:

[0035] CF=target+M1×ε(P o -P ref |-P ref ×0.1%)+M2×ε[f cost ],

[0036] In the formula, target represents a circuit indicator that needs to be optimized to the minimum value, that is, the original cost function f described in the above formula cost , M i It represents a very large value, which is absolutely greater than any possible target value, so that the cost function of the control target cost function will jump to a very large value when the output power does not meet the requirements. f(x) represents other possible constraint expressions except output power.

[0037] The beneficial effects of the present invention are:

[0038] 1. The present invention is based on a 3p-DAB converter circuit under three-degree-of-freedom duty cycle control, which avoids the complicated analytical process and complicated partitioning conclusions and can quickly obtain its steady-state operation state.

[0039] 2. The present invention realizes a fully automated scheme for generating optimal control parameters of a 3p-DAB circuit based on a simulation model. The generated control parameters can effectively improve the working efficiency of the circuit under the condition that the input and output voltages do not match. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flow chart of the optimization control method of the present invention;

[0041] Figure 2 This is a topological structure diagram of a three-phase dual active bridge isolated DC-DC converter system of the present invention;

[0042] Figure 3 It is a steady-state model of the three-phase dual active bridge isolated DC-DC converter system of the present invention;

[0043] Figure 4 This is a simulation diagram of a three-phase dual active bridge isolated DC-DC converter system of the present invention;

[0044] Figure 5 A schematic diagram of the switching function of the front and rear bridge arms of a three-phase DAB circuit under three-degree-of-freedom control used in the steady-state model of the present invention: when the upper bridge arm is turned on, the switching function is 1, and when the lower bridge arm is turned on, the switching function is 0;

[0045] Figure 6 This is a diagram showing the effect of 300V input / 600V output simulation of inductor current reduction under full working conditions of the present invention;

[0046] Figure 7 This is a diagram of the simulated inductor current reduction effect of 600V input / 300V output under full working conditions of the present invention. DETAILED DESCRIPTION

[0047] The present invention will be further described in detail with reference to the accompanying drawings.

[0048] When the input and output voltages do not match, it is difficult for the three-phase dual active bridge converter system using the traditional SPS control method to maintain zero voltage turn-on of all bridge arm switches, and the current reflux is large and there are many topological switches, which makes the circuit efficiency low under this condition. Therefore, it is necessary to adopt a steady-state digital modeling method and a three-degree-of-freedom duty cycle control (DCC) optimization control parameter calculation method.

[0049] Reference Figure 1 As shown, the present invention discloses a method for calculating optimal control parameters of a three-phase dual active bridge isolated DC-DC converter system under DCC control, and the steps are as follows.

[0050] S1, start running, construct the cost function of the optimization problem, and input the optimization operating parameters: use the piecewise linearization analysis method to analyze the inductance in steady-state operation in segments, calculate the inductor current in the entire interval through the switching timestamp, the front and rear bridge arm switching functions and the input and output voltages, and then calculate the output power. Through the above piecewise linear method, construct a three-phase DAB steady-state circuit model under DCC control.

[0051] Description of steady-state computer modeling of three-phase DAB circuit under DCC control: Figure 2 What is shown is the final form of the three-phase DAB steady-state circuit model. The input parameters are composed of circuit element parameters, input and output voltages, and control inputs. The basic output is output power, but other circuit operation parameters in this class can also be called. The detailed construction process is as follows.

[0052] 1) Generate a three-phase front and rear bridge arm switch action timestamp sequence within a single cycle based on the three-degree-of-freedom control input, where the timestamp element consists of an action time value and an action type label. Then sort the timestamp sequence according to the action time. Taking phase A as an example, its switch action time is shown as follows:

[0053]

[0054] Among them, t Ta1 Indicates the time between the upper bridge arm switch tube of input A phase being turned on and the lower bridge arm switch tube being turned off, t Ta2 Indicates the time between the lower bridge arm switch tube of input A phase being turned on and the upper bridge arm switch tube being turned off, t Sa1 Indicates the time when the upper bridge arm switch of output phase A is turned on and the lower bridge arm switch is turned off, t Sa2 Indicates the time when the lower bridge arm switch of output phase A is turned on and the upper bridge arm switch is turned off; T s Represents the switching cycle, D1 and D2 are the switching duty ratios of the front and rear bridge arms respectively, and D f D is the ratio of the front bridge arm lagging behind the rear bridge arm. f =1 means that the front bridge arm lags behind the rear bridge arm by half a switching cycle, and n is the number of switching cycles from time zero to now. The switching action time of phase B and phase C lags behind and leads the switching cycle of phase A by 1 / 3 of the switching cycle, respectively. The expression is:

[0055] (2) According to the switching action timestamp, the entire switching cycle is partitioned, and the segmented switching functions of the front and rear bridge arms are drawn according to the sorted switching action timestamp labels. When the corresponding upper bridge arm is turned on, the switching function is 1, and when the corresponding lower bridge arm is turned on, the switching function is -1; its expression is:

[0056]

[0057] Where δ(t) is the impulse function. When t0∈[t1,t2], for any f(x) we have At the same time if but

[0058] (3) Calculate the inductance L based on the segmented switching function sa , L sb , L sc The voltage across the two ends, Figure 4 The figure shows the equivalent circuit of the three-phase DAB circuit. The three-phase bridge arm on the input and output sides and the input and output voltages are equivalent to forming two controllable voltage sources at both ends of the inductor. When analyzing the circuit, the virtual center potentials at both ends of the transformer are connected. Through the superposition principle, it can be concluded that in any switching state, the voltage across the equivalent three-phase inductor is controlled by the switching function of the three-phase bridge arm. Taking the equivalent voltage source of the front bridge arm of phase A as an example, the expression of its controlled voltage is Where U dc Represents the DC side voltage of the three-phase bridge arm, S a , S b , S c They represent the switching functions of the three-phase bridge arms respectively, and their values ​​are the switching functions of the corresponding front bridge arm minus the switching functions of the corresponding rear bridge arm. For the front and rear bridge arm switching functions, when the corresponding upper bridge arm is turned on, it is 1, and when the corresponding lower bridge arm is turned on, it is -1, such as Figure 5 shown.

[0059] (4) According to the inductance L sa , L sb , L sc The voltage at both ends and the duration of each small interval are used to calculate the change in inductor current in each small interval. Taking phase A as an example, the change in inductor current in the nth small interval formed by the timestamp is: Where L is the equivalent inductance in the three-phase circuit, t n-1 and t n They are the start time and end time of the cell respectively.

[0060] (5) The inductor current offset at the beginning of the switching cycle is calculated based on the inductor current change calculated in the previous step. The calculation basis is that the inductor current should not have a DC bias in a single switching cycle, that is, the integral of the inductor current in the entire cycle should be 0. Because the three phases of this circuit should be symmetrical, if the three-phase inductors have the same DC bias, it violates Kirchhoff's current law. Therefore, the inductor current at the beginning of the cycle is Where t0 represents the start time of the cycle, I′(t) represents the inductor current value in the entire switching cycle calculated based on the above linear change of the inductor current when the inductor current is 0 at the start of the cycle, and the inductor current in the entire switching interval is I(t)=I0+I′(t).

[0061] S2, construct a cost function composed of circuit parameters according to the optimization target: repackage the steady-state circuit model to construct an optimizable function model whose input is a variable control parameter and whose output is a cost function corresponding to the optimization target.

[0062] The output charge in a single switching cycle is calculated based on the inductor current and the switching function of the rear bridge arm. Because the inductor charge will only be transferred to the output filter capacitor and the output load when the inductor current is positive and the output corresponding upper bridge arm is turned on, the output power expression is as follows:

[0063]

[0064] In the above calculation process, a sequence of inductor current magnitudes at each switching action moment is generated. The inductor current in this sequence can be used to determine whether the switching action is in a soft switching state and calculate the effective value of the inductor current. The on-off and conduction losses of the switching element can also be calculated based on this. The above model parameters can be further transformed into an optimization target cost function.

[0065] S3, select particle swarm optimization parameters: select appropriate particle swarm optimization parameters according to the type of problem to be optimized. In this step, a more general particle swarm optimization parameter can also be used.

[0066] This patent uses a particle swarm optimization scheme to intelligently optimize the optimization target cost function. The particle swarm optimization (PSO) algorithm is a swarm-based search algorithm based on an intelligent optimization algorithm that imitates swarm intelligence behavior. In PSO, individuals called particles "fly" in a hyperdimensional search space, and each particle represents a potential solution.

[0067] The pattern of particle position change in the search space is based on the social psychological tendency of individuals to imitate the successful experience of other individuals. Therefore, the change of particles in the particle swarm is affected by the experience or knowledge of all particles or their neighbors. The result of modeling this social behavior is that particles tend to approach the known optimal point at a random speed during the iteration process. The particle swarm algorithm distributes the initialized particle swarm to the entire multidimensional control space, and has a certain random optimization method during the optimization process, so that it has better global optimization ability and has better optimization effect on the piecewise nonlinear optimization model described in this patent. If the cost function corresponding to the control parameters optimized this time is not lower than the corresponding cost function of the SPS control, it means that the optimization is invalid and needs to be re-optimized. In this way, the optimal control parameters that meet the optimization objectives can be obtained.

[0068] S4, initialize particle swarm: perform particle swarm optimization on the optimizable function model under the set working conditions. If the cost function corresponding to the optimized control parameters is higher than the cost function corresponding to the SPS control parameters under the same working conditions, the optimization is considered invalid and particle swarm optimization needs to be performed again.

[0069] The particle position is changed by adding the velocity v to the current position. i (t) to achieve, that is:

[0070] x i (t+1)=x i (t)+v i (t+1),

[0071] Where t represents the discrete time step, x i (t) represents the position of particle i in the search space at time step t, v i (t) is the velocity vector of the optimization process, which is determined by the particle's own empirical knowledge and the exchange information from the particle's neighborhood (or all) particles. The particle's empirical knowledge is generally called the cognitive component, which is proportional to the distance from the particle's historical best position found since the first time step (called the particle's personal best position). The social exchange information is called the social component in the velocity equation, which is proportional to the overall best component of the field particles or all particles. Its calculation formula is:

[0072]

[0073] Where V ij (t) represents the velocity of particle i in dimension j at time step t, y ij (t) and Respectively represent the best historical position of the particle itself and the field or the whole up to the tth step, r 1j (t) and r 2j(t) is a random value uniformly distributed between [0, 1] to enhance the randomness of the particle swarm algorithm, and ω is the inertia parameter.

[0074] The particle swarm algorithm has two basic strategies, namely the global best particle swarm algorithm and the neighborhood best particle swarm algorithm. The difference is It represents the best historical position of n particles in the field or the best historical position of the particle swarm as a whole. In contrast, the local best particle swarm algorithm has stronger exploration ability and can better avoid falling into the local optimal solution. Since the effective interval involved in this paper accounts for a small proportion of the overall space, in order to avoid the particle swarm being attracted to the local optimal point outside the effective interval too early, this paper uses the neighborhood best particle swarm algorithm.

[0075] S5, calculate the case circuit operation status and the target cost function.

[0076] Explanation of cost function construction and optimized parameter processing. The mathematical expression of the optimization problem of the patent method is:

[0077] minimize f cost

[0078] subject to P o =P ref

[0079] others,

[0080] Where f cost Represents the cost function. Since the digital optimization algorithm may cause excessive calculation or difficulty in convergence by using full-precision constraints, this patent method retains a certain error range when setting constraints and integrates the constraints into the cost function. The constructed cost function is as follows:

[0081] CF=target+M1×ε(P o -P ref |-P ref ×0.1%)+M2×ε[f cost ],

[0082] Where target represents a circuit indicator that needs to be optimized to the minimum value, that is, the original cost function f described in the above formula cost , M i represents a very large value, which is absolutely greater than any possible target value, so that the cost function of the control target will jump to a very large value when the output power does not meet the requirements. f(x) represents other possible constraint expressions except output power.

[0083] S6, store the contemporary optimal cost function, determine whether the conditions for exiting particle swarm optimization are met, otherwise perform a particle swarm iteration and return to S5.

[0084] S7, if yes, further determine whether the optimization index is met, otherwise return to S4.

[0085] S8, if yes, output the optimal control parameters and end the operation.

[0086] If you need to optimize the control parameters for a certain range of working conditions, you need to perform the above optimization process by discretely collecting some working points within the working condition. The more points you collect, the better the optimization effect will be, but the optimization time required in the early stage will be longer. If there are more collection points, you can consider performing a certain low-frequency filtering on the generated control parameters. Finally, through a piecewise linear interpolation method, the optimized controller parameters can cover the entire required working condition.

[0087] Figure 2 The topological structure diagram of the three-phase DAB circuit is shown in FIG. 1 , which consists of the front bridge arm and the rear bridge arm through the transformer and the inductor L sa , L sb , L sc The front bridge arm is connected by the switch tube S 14 ~S 16 The rear bridge arm is composed of switch tube S 21 ~S 26 The six switch tubes constitute the upper and lower bridge arms respectively. The left input is the control parameter under the model, where the switching frequency is a fixed control parameter. The upper bridge duty cycle of the front and rear bridge arms and the shift ratio of the front and rear bridge arms are the three controllable degrees of freedom in DCC control. The controllable range of the bridge arm duty cycle is [0, 1], and the controllable range of the shift ratio is [0, 0.5], which corresponds to the phase shift from no phase shift to the rear bridge arm lagging the front bridge arm by 90°; the right output, that is, the output power, is the constraint target of the optimization control to ensure that the selected control parameters must meet certain output power conditions; the upper side represents the fixed component parameters and the current operating parameters in the circuit. The circuit component parameters that are more important to this model are the transformer ratio and the equivalent leakage inductance. The current working state is determined by the three parameters of input and output voltage and output power. The input and output voltages are the input parameters of the model, and the output power is the output parameter controlled by the control parameters shown on the left side of the model.

[0088] Figure 3The figure shows the steady-state model of the three-phase DAB circuit. The circuit operation state obtained in the intermediate process of calculating the output is centered on the three-phase inductor operation state in the topological center, because in this converter, the three-phase inductor is the core element for storing and transferring energy. At the same time, the switch current at the switching moment can be obtained through the inductor current and the switch function, and the circuit magnetic loss can also be measured through the effective value of the inductor current. The various circuit operation state parameters obtained above can be used to form our optimization target and construct a cost function with minimization as the target. Therefore, the above model can be further processed into: an optimization target cost function with three-degree-of-freedom control parameters as input and cost function as output.

[0089] Explanation of the parameter processing after optimization: Under a certain range of operating conditions, the parameters obtained after optimization may fluctuate slightly. In order to reduce the adverse effects of such fluctuations on control, it is possible to consider using low-frequency filtering to trim a few offset control parameters within the entire waveform.

[0090] After low-frequency filtering, a series of control parameter sequences in the entire operating range can be obtained. It is possible to consider performing a piecewise linear interpolation on the control parameters generated after optimization within the operating range so that the obtained control parameters can cover the entire operating range.

[0091] Figure 6 and Figure 7 The optimization results of optimizing the effective value of inductor current by using the patented method obtained through simulation are as follows: the control parameter calculation method of the patented optimization is used to optimize the control parameters of the three-phase DAB circuit with input voltage of 300V, output voltage of 600V, equivalent leakage inductance of 6.5uH, transformer ratio of 1:1, switching frequency of 14kHz, and maximum output power of 130kW under all working conditions of forward energy flow and reverse energy flow. The optimization target is the effective value of inductor current. Studies have shown that the smaller the effective value of inductor current, the smaller the overall conduction loss of the circuit. The conduction loss consists of two parts: magnetic loss of magnetic components and conduction loss of switches.

[0092] After obtaining the optimized parameters, simulation verification was carried out. The circuit operation under various working conditions of 300V input and 600V output and 600V input and 300V output, and the output power was set from no-load to 130kW output upper limit, and the optimization effect of the optimized DCC controller on the effective value of the inductor current under various working conditions was recorded. The results are shown in the attached figure. Figure 4 shown.

[0093] From the simulation results, it can be seen that the optimization parameters calculated using the optimization method proposed in this patent can effectively reduce the effective value of the inductor current within the power output range of the sample converter, achieving the predetermined optimization goal.

[0094] The above embodiments are only illustrative of the principles and effects of the present invention, as well as some embodiments of its application. For those skilled in the art, several modifications and improvements may be made without departing from the creative concept of the present invention, and all of these belong to the protection scope of the present invention.

Claims

1. An optimization control method for a three-phase dual active bridge isolated DC-DC converter system, wherein the three-phase dual active bridge isolated DC-DC converter system comprises a front bridge arm and a rear bridge arm connected by a transformer and an inductor L sa , L sb , L sc The front bridge arm is connected by the switch tube S 14 ~S 16 The rear bridge arm is composed of switch tube S 21 ~S 26 Composition; it is characterized by: The steps are as follows S1, adopts the piecewise linearization analysis method to analyze the inductor in steady-state operation in sections, calculates the inductor current in the entire interval through the switch timestamp, the front and rear bridge arm switch functions and the input and output voltages, and then calculates the output power to construct a three-phase DAB steady-state circuit model; S2, repackage the steady-state circuit model to construct an optimizable function model whose input is a variable control parameter and whose output is a cost function corresponding to the optimization target; S3, select appropriate particle swarm optimization parameters; S4, particle swarm optimization is performed on the optimizable function model. If the cost function corresponding to the optimized control parameter is higher than the cost function corresponding to the SPS control parameter under the same working condition, the optimization is considered invalid and the particle swarm optimization is performed again. S5, calculating the circuit model operation state and the target cost function; S6, store the current optimal cost function, determine whether the conditions for exiting particle swarm optimization are met, otherwise perform a particle swarm iteration and return to S5; S7, if yes, further determine whether the optimization index is met, otherwise return to S4; S8, if yes, output the optimal control parameters and end the operation.

2. The optimization control method of a three-phase dual active bridge isolated DC-DC converter system according to claim 1, characterized in that: The three-phase DAB steady-state circuit model in step S1 is constructed by the following steps: (1) Generate a three-phase front and rear bridge arm switch action timestamp sequence within a single cycle based on the three-degree-of-freedom control input, where the timestamp element consists of an action time value and an action type label. Then, the timestamp sequence is sorted according to the action time. The A-phase switch action time is shown in the following formula: Where t Ta1 Indicates the time between the upper bridge arm switch tube of input A phase being turned on and the lower bridge arm switch tube being turned off, t Ta2 Indicates the time between the lower bridge arm switch tube of input A phase being turned on and the upper bridge arm switch tube being turned off, t Sa1 Indicates the time when the upper bridge arm switch of output phase A is turned on and the lower bridge arm switch is turned off, t Sa2 Indicates the time when the lower bridge arm switch of output phase A is turned on and the upper bridge arm switch is turned off; T s Represents the switching cycle, D1 and D2 are the switching duty ratios of the front and rear bridge arms respectively, and D f is the shift ratio of the front bridge arm lagging behind the rear bridge arm, n is the number of cycles from time zero to now; the switching action time of phase B lags behind phase A by 1 / 3 switching cycle, and the switching action time of phase C is ahead of phase A by 1 / 3 switching cycle. Its expression is: (2) Partition the entire switching cycle according to the switching action timestamp, and draw the segmented switching function of the front and rear bridge arms according to the sorted switching action timestamp labels. Where δ(t) is the impulse function. When t0∈[t1,t2], for any f(x) we have At the same time if but (3) Calculate the inductance L based on the segmented switching function sa , L sb , L sc The voltage at both ends is: Where U dc Represents the DC side voltage of the three-phase bridge arm, S a , S b , S c They represent the switching functions of the three-phase bridge arms respectively, and their values ​​are the switching functions of the front bridge arm of the corresponding phase minus the switching functions of the rear bridge arm of the corresponding phase; (4) According to the inductance L sa , L sb , L sc The voltage at both ends and the duration of each small interval are used to calculate the change in inductor current in each small interval. The change in inductor current in the nth small interval formed by the A phase timestamp is: Where L is the equivalent inductance in the three-phase circuit, t n-1 and t n are the start time and end time of the cell respectively; (5) The inductor current offset at the beginning of the switching cycle is calculated based on the change in the inductor current. The inductor current at the beginning of the cycle is Where t0 represents the start time of the cycle, I′(t) represents that the inductor current is 0 at the start of the cycle, and the inductor current in the entire switching interval is I(t)=I0+I′(t).

3. The optimization control method of a three-phase dual active bridge isolated DC-DC converter system according to claim 2, characterized in that: In step S2, the output charge in a single switching cycle is calculated based on the inductor current and the switching function of the rear bridge arm. The output power expression is: Further transformation is performed to obtain the optimized objective cost function.

4. The optimization control method of a three-phase dual active bridge isolated DC-DC converter system according to claim 3, characterized in that: The mathematical expression of the cost function optimization problem in step S5 is: minimize f cost subject to P o =P ref others, Where f cost Represents the cost function; the cost function constructed by integrating the constraints into the cost function is: CF=target+M1×ε(|P o -P ref |-P ref ×0.1%)+M2×ε[f cost ], Where target represents the cost function f that needs to be optimized to the minimum value cost , M i represents a target value, which makes the cost function of the control target cost function jump to a very large value when the output power does not meet the requirements. f(x) represents other possible constraint expressions except output power.

Images