A model prediction-based isos system output voltage equalization control method

Abstract

The application discloses a model prediction-based ISOS system output voltage equalization control method, and relates to the field of power electronics, which comprises the following steps: adjusting the total output voltage through a PI regulation system; establishing an input-output mathematical model of a sub-module, which is discretized by a forward Euler formula; establishing a system cost function and selecting an optimal state; establishing a discrete prediction model of each sub-module; sampling the output capacitor voltage of each sub-module, adjusting the inter-bridge phase shift ratio of each sub-module through the discrete prediction model, and performing prediction control on the output capacitor voltage of each sub-module at each sampling time, so as to realize the output voltage equalization control of the ISOS system. The control strategy has the advantages of high reliability, being applicable to a multi-module ISOS system, simple mathematical model, good dynamic response and the like.

Description

TECHNICAL FIELD

[0001] The application relates to the field of power electronics, and particularly relates to an ISOS system output voltage balancing control method based on model prediction. BACKGROUND

[0002] In recent years, with the rapid development of power electronics technology, modular DC converters have been widely researched and applied. Standardized module combination modes include series connection, parallel connection and cascade connection, and according to different input and output modes, series-parallel DC converters can be divided into input parallel output parallel (IPOP) systems, input parallel output series (IPOS) systems, input series output parallel (ISOP) systems and input series output series (ISOS) systems, wherein the DC converter of the ISOS system is suitable for input low-voltage output medium-high-voltage occasions, such as high-voltage DC power supply, high-speed electric railway and ship power supply systems, due to its low production cost and the characteristic that the total power is evenly distributed among the sub-modules. To ensure stable system operation, the most critical problem is to solve the input and output voltage balancing problem.

[0003] The existing schemes are all realized based on PI controllers, including three-loop control strategies: an output voltage loop ensures stable system output voltage, an input voltage balancing loop ensures input voltage balancing through a current inner loop, and simultaneously realizes output voltage balancing; and a switching duty ratio strategy: the output voltages of two sub-modules are respectively taken as the feedback of the input of the other, to track the same given signal, so as to ensure equal output voltages. These voltage balancing control strategies are only applicable to two sub-modules, and each sub-module lacks independence, resulting in low modularity of the ISOS system, and affecting the system scalability and reliability. SUMMARY

[0004] The present application provides an ISOS system output voltage balancing control method based on model prediction, which adjusts the phase-shift ratio between the bridges of each sub-module to realize sub-module output capacitor voltage balancing, reduces system complexity, and improves system reliability and overall dynamic performance.

[0005] The technical scheme of the present application is as follows:

[0006] The ISOS system output voltage balancing control method based on model prediction comprises the following steps: first, adjusting the total output voltage of the system through a PI regulator; second, establishing a discrete-time model of the sub-module based on the input-output relationship of the sub-module; when the cost function is zero, selecting the optimal output capacitor voltage of the sub-module according to the total output voltage, and establishing a discrete prediction model of the sub-module; finally, calculating the difference between the output balancing voltage and the output capacitor voltage of each sub-module at the current sampling time, and substituting the difference into the discrete prediction model, and adopting the strategy of optimizing the phase-shift ratio between the bridges of each sub-module to realize sub-module capacitor voltage balancing.

[0007] Further technical solutions are that the control method specifically comprises the following steps:

[0008] First step: constructing a system output voltage closed loop. The system output voltage closed loop comprises an output voltage reference of the system, a total output voltage feedback value of the system and a PI regulator, and the output of the PI regulator is a common bridge interphase shift ratio of the system.

[0009] Second step: establishing a mathematical model of the input-output relationship of a single sub-module, and establishing a discrete-time model of the single sub-module through a forward Euler formula.

[0010] Third step: establishing a cost function of the system, and selecting optimal output capacitor voltages of each sub-module when the value of the cost function is zero.

[0011] Fourth step: deriving a discrete prediction model of each sub-module according to the discrete-time model of each sub-module and the optimal output capacitor voltages.

[0012] Fifth step: calculating a difference between an output balanced voltage and an output instantaneous voltage of each sub-module at a current sampling moment, predicting a current gain of each sub-module about the bridge interphase shift ratio at a next sampling moment according to the discrete prediction model, and adjusting the bridge interphase shift ratio of each sub-module to make the output capacitor voltages of each sub-module tend to be balanced at the next sampling moment.

[0013] The beneficial technical effects of the present application are:

[0014] 1. No weighting factor design. The present application mainly studies the output voltage balance of the ISOS system, the total output voltage is adjusted by a conventional PI regulator, and the output voltage balance of each sub-module is achieved by a model prediction control strategy based on the bridge interphase shift ratio, so that no weighting factor is needed.

[0015] 2. The proposed control method optimizes the bridge interphase shift ratio under constant switching frequency, and only involves the shift between two full bridges, without bridge internal shift, effectively avoiding the problems of direct current bias of the switching tube of the conventional MPC and uneven power distribution.

[0016] 3. The model is simple to construct and only involves the output capacitor.

[0017] 4. The system has high scalability, and each sub-module has high independence.

[0018] 5. The system has good dynamic response. The conventional MPC selects the switching state that minimizes the cost function, while the present method calculates the optimal bridge interphase shift ratio with the cost function being zero. DETAILED DESCRIPTION

[0019] Figure 1 is a structural diagram of the ISOS system provided by the present application.

[0020] Figure 2 is a system submodule DC-DC converter structure provided by the present application.

[0021] Figure 3 is an ISOS system output voltage equalization control method flowchart provided by the present application.

[0022] Figure 4 is a model prediction-based overall control structure diagram provided by the present application. DETAILED DESCRIPTION

[0023] The specific embodiments of the present application will be further described below in conjunction with the accompanying drawings.

[0024] As shown in Figure 1 , the ISOS system includes N submodules, a DC voltage source V in is connected in series to the input ends of the N submodules, and the output ends of the N submodules are respectively connected in parallel to output capacitors C oN , the output capacitors C oN are connected in series across the output load R L , and the submodules are DC-DC converters. Optionally, the submodule DC-DC converter used in this example is a CLLLC (resonant type) DC-DC converter, the structure of which is shown in Figure 2 . Since this structure is not the focus of the present application, its specific connection relationship will not be described in detail here.

[0025] Based on the above structure, the present application provides an ISOS system output voltage equalization control method based on model prediction, which specifically includes the following steps in conjunction with Figure 3 , Figure 4 .

[0026] Step 1: Build an output voltage closed loop of the ISOS system to obtain the common bridge interphase shift ratio of the ISOS system.

[0027] The output voltage closed loop includes an output voltage reference value V o_ref of the system, a total output voltage feedback value V o of the system, and a PI regulator. Among them, the output voltage reference value V o_ref and the total output voltage feedback value V o are inputs of the PI regulator, and the output of the PI regulator is the common bridge interphase shift ratio d of the system, which is input into the discrete prediction model of each submodule.

[0028] Step 2: Establish a discrete-time model of each submodule based on the input-output relationship of a single submodule.

[0029] 1) Establish a mathematical model of the input-output relationship of a single submodule, the expression of which is:

[0030] C oN dv oN / dt=f(d N )i in -i o (1)

[0031] where C oN is the output capacitance of the Nth sub-module, v oN is the output capacitance voltage of the Nth sub-module; f(d N ) is the current gain of the Nth sub-module, and the current gain is positively correlated with the inter-bridge phase shift of the sub-module; i in is the system input current, and i o is the system output current.

[0032] 2) The mathematical model (1) is discretized by using the forward Euler formula to obtain the discrete-time model of a single sub-module, which is expressed as:

[0033] v oN (k+1)-v oN (k)=T s {f(d N (k+1))i in (k)-i o (k)} / C oN (2)

[0034] where k is the current sampling time, k+1 is the next sampling time, and T s is the sampling period.

[0035] where the expressions of the input and output currents of the ISOS system are established based on the common inter-bridge phase shift d, which is obtained by a PI regulator; V o (k) is the total output voltage of the system at the k sampling time, and R L is the output load.

[0036]

[0037] where d(k) is the common inter-bridge phase shift of the system at the k sampling time, which is obtained by a PI regulator; V o (k) is the total output voltage of the system at the k sampling time, and R L is the output load.

[0038] Third step: Establish the cost function of the ISOS system, and select the optimal output capacitance voltage of each sub-module when the cost function is zero.

[0039] 1) The cost function is established by the sum of the absolute values of the output capacitance voltage difference between adjacent sub-modules and the output capacitance voltage difference between the first and last sub-modules, which is expressed as:

[0040] g=|v o1 (k+1)-v o2 (k+1)|+…+|voN (k+1)-v o1 (k+1)| (3)

[0041] where v oN (k+1) is the output capacitor voltage of the Nth submodule at k+1 sampling time.

[0042] 2) When the output voltage of the ISOS system is balanced, g = 0, the optimal output capacitor voltage of a single submodule is obtained as the total output voltage of the system, and the cost function of equation (3) is further expressed as:

[0043]

[0044] where V o (k+1) is the total output voltage of the system at k+1 sampling time.

[0045] 3) Neglecting the deviation between adjacent sampling times (i.e., k and k+1), i.e., the system output does not change significantly within a sampling interval, so:

[0046]

[0047] Fourth step: Establish a discrete prediction model for each submodule by combining the discrete-time model and the optimal output capacitor voltage.

[0048] Substitute equation (5), the input and output current relationship i in (k) and i o (k) into equation (2), and the expression of the discrete prediction model for each submodule is obtained as:

[0049]

[0050] where v is the output balanced voltage of each submodule, i.e., the optimal output capacitor voltage of a single submodule, v oN (k) is the output capacitor voltage (i.e., instantaneous voltage) of the Nth submodule at k sampling time; f(d(k)) is the current gain of the system at k sampling time; f(d N (k+1)) is the predicted current gain of the Nth submodule at k+1 sampling time, so as to obtain the prediction value of the inter-bridge phase shift ratio of each submodule at the next sampling time.

[0051] Fifth step: At the current sampling time k, obtain the total output voltage V o (k) of the system, the output capacitor voltage v oN (k) of each submodule, and the common inter-bridge phase shift ratio d(k) of the system, and substitute them into the corresponding discrete prediction model (6) to predict the inter-bridge phase shift ratio d N(k+1). At constant switching frequency, the controller adjusts the phase shift between the bridges of each sub-module to the predicted value, achieving the equalization of the output voltage of the sub-modules.

[0052] The specific implementation of the above method is that at the k-1 sampling moment, the ISOS system output voltage is equalized and in a stable state, at this time v o1 (k-1) =... = v oN (k-1), when the output capacitor is disturbed at the k sampling moment, the output capacitor voltage of one or more sub-modules in the system will change. Suppose N = 2, and the output capacitor voltage relationship of the two sub-modules is v o1 (k) < (v o1 (k) + v o2 (k)) / 2 < v o2 (k). The control strategy is set to adjust from the k sampling moment, the PI regulator is used to adjust the total output voltage of the system to the preset value, the difference between the output balance voltage and the output capacitor voltage of each sub-module is calculated and substituted into the corresponding discrete prediction model (6). Since the difference calculated by sub-module 1 is greater than zero, the predicted phase shift between the bridges d1(k+1) is increased relative to d1(k) based on formula (6), the controller sets the phase shift between the bridges of sub-module 1 to the predicted value, so that the output capacitor voltage of sub-module 1 is increased. Similarly, the difference calculated by sub-module 2 is less than zero, so the predicted phase shift between the bridges d2(k+1) is reduced relative to d2(k) based on formula (6), the controller sets the phase shift between the bridges of sub-module 2 to the predicted value, so that the output capacitor voltage of sub-module 2 is reduced, and the system can track the balance point voltage of each sub-module output voltage when the output is disturbed, thereby achieving the equalization control of the output voltage of the two modules. Optionally, if the difference between the output balance voltage and the output capacitor voltage of the sub-module is equal to zero, the predicted phase shift between the bridges d N (k+1) is equal to d N (k).

[0053] The above only describes the preferred embodiments of the present application, and the present application is not limited to the above embodiments. It can be understood that other improvements and changes directly derived or thought by those skilled in the art without departing from the spirit and concept of the present application should be considered within the protection scope of the present application.

Claims

1. A model prediction based ISOS system output voltage equalization control method, characterized by, The ISOS system includes N Sub-modules N Each submodule has an output capacitor connected in parallel at its output terminal. The output capacitor is connected in series across the output load. The submodule is a DC-DC converter. The control method comprises: a discrete-time model of each sub-module is established based on an input-output relationship of the single sub-module; a cost function of the ISOS system is established, and optimal output capacitor voltages of each sub-module when the cost function is zero are selected; a discrete prediction model of each sub-module is established in combination with the discrete-time model and the optimal output capacitor voltages; at a current sampling time, a total output voltage of the system and an output capacitor voltage of each sub-module are obtained, and are substituted into the corresponding discrete prediction model to predict an inter-bridge phase shift ratio of each sub-module at a next sampling time; the inter-bridge phase shift ratio of each sub-module is adjusted to the predicted value respectively, so that the output voltage of the sub-module is balanced and adjusted; the control method further comprises: an inter-bridge phase shift ratio of the ISOS system is obtained; an expression of input and output currents of the ISOS system is established based on the inter-bridge phase shift ratio, and is as follows: ； wherein, i in ( k ) is k the system input current at the sampling instant, i o ( k ) is k the system output current at the sampling instant, d ( k ) is k the common inter-bridge phase shift of the system at the sampling instant, f ( d ( k )) is k the current gain of the system at the sampling instant, V o ( k ) is k the total output voltage of the system at the sampling instant, R L is the output load.

2. The model prediction based ISOS system output voltage equalization control method of claim 1, wherein, the discrete-time model of each sub-module is established based on the input-output relationship of the single sub-module, and comprises: a mathematical model of the input-output relationship of the single sub-module is established, and an expression is as follows: （1） wherein, C oN is the output capacitance of the first N submodule, v oN is the output capacitance voltage of the first N submodule; f ( d N ) is the current gain of the first N submodule, the current gain being positively correlated with the inter-bridge phase shift of the submodule; i in is the system input current, i o is the system output current; the mathematical model is discretized by using a forward Euler formula to obtain the discrete-time model of the single sub-module, and an expression is as follows: （2） wherein, k is the current sampling time, T s is the sampling period, f d N k is the predicted current gain of the N th sub-module at the k +1 sampling time.​​ 3. The model prediction based ISOS system output voltage equalization control method of claim 2, wherein, the cost function of the ISOS system is established, and comprises: a cost function is established by using an absolute value of a difference between output capacitor voltages of adjacent sub-modules and an absolute value of a difference between output capacitor voltages of the first and last sub-modules, and a sum of the two, and an expression is as follows: （3） wherein, v oN ( k +1) is the output capacitor voltage of the first N submodule at the sampling time point of k k +1.

4. The model prediction based ISOS system output voltage equalization control method of claim 3, wherein, the optimal output capacitor voltages of each sub-module when the cost function is zero are selected, and comprise: When the output voltage of the ISOS system is balanced, g =0, it is derived that the optimal output capacitor voltage of a single sub-module is equally divided into the total output voltage of the system, and the cost function of equation (3) is further expressed as: （4） in, V o ( k +1) is k The total output voltage of the system at sampling time +1; a deviation at an adjacent sampling time is ignored, so that the following is obtained: （5） wherein V o ( k ) is k the total output voltage of the system at the sampling instant.

5. The model prediction based ISOS system output voltage equalization control method of claim 4, wherein, the discrete prediction model of each sub-module is established in combination with the discrete-time model and the optimal output capacitor voltages, and comprises: the expression (5) and the input and output current relationship of the system are substituted into the expression (2) to obtain an expression of the discrete prediction model of each sub-module, and is as follows: （6） in, To balance the output voltage of each submodule, v oN ( k ) is the first N Each submodule in k The output capacitor voltage at the sampling moment; f ( d ( k ))for k The current gain of the system at the sampling time; f ( d N ( k +1)) is the predicted number. N Each submodule in k The current gain at the +1 sampling time is used to obtain the predicted value of the inter-bridge shift ratio of each submodule at the next sampling time.

6. The model prediction based ISOS system output voltage equalization control method of claim 1, wherein, the inter-bridge phase shift ratio of the ISOS system is obtained, and comprises: an output voltage closed loop of the ISOS system is constructed, and the output voltage closed loop comprises an output voltage reference value of the system, a total output voltage feedback value of the system and a PI regulator, wherein the output voltage reference value and the total output voltage feedback value are input to the PI regulator, and an output of the PI regulator is the inter-bridge phase shift ratio of the system and is input to the discrete prediction model of each sub-module.

7. The model prediction based ISOS system output voltage equalization control method of claim 1, wherein, the inter-bridge phase shift ratio of each sub-module is adjusted to the predicted value respectively, and comprises: the inter-bridge phase shift ratio of each sub-module is adjusted to the predicted value respectively under a constant switching frequency.

Images