阅读说明：本技术 适用于多电平变换器的热应力平衡预测控制方法及系统 (Thermal stress balance prediction control method and system suitable for multi-level converter ) 是由 罗安 韩蓉 徐千鸣 丁红旗 唐成 汪亮 于 2019-10-23 设计创作，主要内容包括：本发明公开了一种适用于多电平变换器的热应力平衡预测控制方法及系统,涉及电力电子变换器控制技术。所述热应力平衡预测控制方法及系统,先根据控制变量参考值和控制变量输出预测值来构建代价函数,确定最优的输出总电平和电平增量,再根据电平增量确定能动作的模块,最后根据结温值设计开关分配函数,确定需要动作的模块,将电平增量分配到需要动作的模块,该控制方法中代价函数根据输出预测值来设计,且无修正,保证了输出精度；开关分配函数以用于反应各模块热应力的结温值为设计依据,通过开关分配函数确定模块的开关状态,达到了多模块热应力平衡的效果。(The invention discloses a thermal stress balance prediction control method and system suitable for a multilevel converter, and relates to the control technology of a power electronic converter. The thermal stress balance prediction control method and the system construct a cost function according to a control variable reference value and a control variable output predicted value, determine an optimal output total level and a level increment, then determine an operable module according to the level increment, finally design a switch distribution function according to a junction temperature value, determine a module required to be operated, and distribute the level increment to the module required to be operated, wherein the cost function is designed according to the output predicted value in the control method without correction, so that the output precision is ensured; the switch distribution function takes the junction temperature value for reflecting the thermal stress of each module as a design basis, and determines the switch state of the module through the switch distribution function, so that the effect of multi-module thermal stress balance is achieved.)

1. A thermal stress balance prediction control method suitable for a multilevel converter is characterized by comprising the following steps:

step 1: collecting control variable output instantaneous values of the multilevel converter at the moment k, and obtaining control variable output predicted values under different switch state combinations at the moment k +1 by the control variable output instantaneous values;

step 2: setting a control variable reference value at the moment k +1, constructing a cost function according to the control variable reference value at the moment k +1 and the control variable output predicted value in the step 1, taking the total level corresponding to the minimum cost function as the optimal output total level at the moment k +1 under the constraint of a limited control set, and obtaining the level increment at the optimal state at the moment k + 1;

and step 3: selecting a module capable of acting at the moment k +1 according to the level increment in the step 2;

and 4, step 4: collecting the conduction voltage and current of each power device, and estimating the junction temperature value of each module power device according to the conduction voltage and current;

and 5: building a switch distribution function of the modules according to the junction temperature value in the step 4, and selecting the modules needing to be operated at the moment k +1 from the modules capable of being operated in the step 3 according to the switch distribution function of each module;

step 6: and (4) allocating the level increment in the step (2) to a specific bridge arm of a module needing action so as to realize the balance of the thermal stress of the multiple modules.

2. The thermal stress balance predictive control method of claim 1, wherein in step 2, the specific expression of the cost function is:

wherein g is a cost function of the multilevel converter,

3. The thermal stress balance predictive control method according to claim 1, wherein in step 2, Z level states adjacent to the output total level H (k) of the multilevel converter at the time k are used as a finite control set, and the constraint of the output total level H (k +1) at the time k +1 in the finite control set is:

H(k+1)∈{H(k)-Z,H(k)-Z+1,…,H(k),…,H(k)+Z-1,H(k)+Z}

wherein H (k) ═ Q₁(k)+…+Q_i(k)+…+Q_m(k) Qi (k) is the output level of the ith module at time k, m is the total number of modules in the multilevel converter, Q_i(k)＝1,0,-1。

4. The method for predictive control of thermal stress balance as set forth in claim 1, wherein in step 3, the selection of the modules that can be activated at time k +1 is:

when level is increased by Δ H>At 0, the level Q is output from the current_i(k) Selecting an action module from the modules with the value of-1 or 0; when level is increased by Δ H<At 0, the level Q is output from the current_i(k) Selecting an action module from the modules of 1 or 0; when the level increment Δ H is 0, the next time the level Q is output_i(k +1) and the current output level Q_i(k) And keeping consistent, wherein the state of the power device in each module at the next moment is kept consistent with the state at the current moment.

5. The thermal stress balance predictive control method of claim 1, wherein in step 5, the specific expression of the switching distribution function of the ith module is

Wherein, C_i(k) The function is assigned to the switch of the ith module at time k,

6. The method for predictive control of thermal stress balance as set forth in claim 1, wherein in step 5, the selection of the module requiring action is:

step 5.1: in each control period, calculating switch distribution function values of all modules capable of acting;

step 5.2: the calculated switch allocation function values are sorted,

step 5.3: selecting | Δ H | modules corresponding to the switch distribution function values from the maximum value of the switch distribution function value sequence as modules needing action at the moment k +1, wherein Δ H is a level increment at the moment k + 1;

step 5.4: and clearing the switch action frequency accumulated value corresponding to the module needing action, and continuously accumulating and counting the switch action frequency accumulated value corresponding to the module not needing action.

7. The thermal stress balance predictive control method of claim 1, wherein in step 6, the specific method of assigning a level increment to a specific leg of a module requiring action is: the bridge arm of each module comprises a left bridge arm and a right bridge arm, and a status flag bit flagi is introduced into the ith module;

if flagi is equal to 1, preferentially judging the switching state of the left bridge arm or the right bridge arm of the ith module, if the switches of the left bridge arm or the right bridge arm can act, operating the left bridge arm or the right bridge arm according to the level increment, and if the switches of the left bridge arm or the right bridge arm cannot act, operating the right bridge arm or the left bridge arm according to the level increment;

if flagi is-1, preferentially judging the switching state of the right bridge arm or the left bridge arm of the ith module, if the switch of the right bridge arm or the left bridge arm can act, operating the right bridge arm or the left bridge arm according to the level increment, and if the switch of the right bridge arm or the left bridge arm can not act, operating the left bridge arm or the right bridge arm according to the level increment;

and after the switch of the right bridge arm or the left bridge arm acts each time, flagi is inverted so as to balance the action times of the switches of the left bridge arm and the right bridge arm of each module.

8. A thermal stress balance predictive control system for a multilevel converter, for implementing the thermal stress balance predictive control method according to any one of claims 1 to 7, comprising:

the output prediction module is used for acquiring a control variable output instantaneous value of the multilevel converter at the current moment and obtaining a control variable output prediction value under different switch state combinations at the next moment according to the control variable output instantaneous value;

the finite set switch state optimizing module is used for obtaining the output total level at the next moment and the level increment at the next moment according to the given control variable reference value at the next moment and the control variable output predicted value under the constraint of a finite control set;

the voltage and current acquisition module is used for acquiring the conduction voltage and current of each module power device;

the junction temperature estimation module is used for estimating a junction temperature value of each module power device according to the conduction voltage and the current of each module power device;

the switch distribution function calculation module is used for acquiring the switch action time accumulated value of each module at the current moment and obtaining a switch distribution function value of each module according to the power device temperature junction value and the switch action time accumulated value of each module;

the switch state distribution module is used for sequencing the switch distribution function values of all the modules and selecting the module needing action at the next moment from the maximum value of the switch distribution function values according to the level increment at the next moment;

and the switching action bridge arm distribution module is used for distributing the level increment at the next moment to the bridge arm needing action of the module needing action.

Technical Field

The invention belongs to the control technology of power electronic converters, and particularly relates to a thermal stress balance prediction control method and system suitable for a multilevel converter.

Background

With the gradual development of economy in China, the demand of the medium-high voltage and high-power converters is gradually increased, and the demand of the capacity of the converters even reaches hundreds of MVA. However, due to the limitations of the voltage class and capacity of the power electronic devices themselves, the conventional two-level converter cannot meet the existing industrial requirements in many cases. In order to improve the capability of power electronic devices to handle larger power, the multi-level converter has received extensive attention from academia, and a lot of research work has been carried out by various national scholars. The multilevel converter is a novel high-voltage high-power converter starting from a topological structure, can obtain high-quality output waveforms, can overcome the defects of a plurality of traditional two-level converters, and does not need a dynamic voltage-sharing circuit and an output end transformer. The multilevel converter connects a plurality of power devices by changing the topological structure of the multilevel converter, outputs more step waves, outputs a waveform closer to a sine wave, and can meet the requirements of medium and high voltage, large capacity and high performance.

The multilevel converter has good output performance, and besides the output performance, the reliability is usually a very important design standard of the converter, especially for special application occasions with limited access environment, such as high-temperature and high-pressure environment, warships, closed cabins and the like. With respect to reliability, power semiconductors are one of the most sensitive components in the converter, whose life cycle is mainly affected by the magnitude of the thermal cycling and the average junction temperature. However, there is a problem of thermal imbalance between modules due to imbalance of electrical parameters and control variables of the multilevel converter under complex output tasks. The higher the likelihood of damage to the module, which is at high junction temperatures for long periods of time, may cause the entire device and even the system to crash, severely degrading the reliability of the system operation. The method poses a great threat to the normal operation of the equipment, so that the reasonable thermal stress management among the modules is of great significance.

Currently, the control strategy considering thermal stress can be classified into four types, namely a power device layer, a converter modulation layer, a converter control layer and a system control layer according to the operation level. The thermal stress control of the power device layer is mainly to reduce the thermal stress by adjusting the switching track of the power device and controlling the conduction voltage drop, such as changing the driving voltage, the driving resistance, the on/off delay time and the like. The method is applied to occasions of reducing uneven thermal stress of parallel devices, restraining junction temperature fluctuation of the wind power converter and the like at present, although the method is high in universality, extra hardware circuits are generally required to be added. The thermal stress control of the modulation layer of the converter is a method for changing the switching sequence of a power device by changing a modulation method so as to improve the loss of the device. The thermal stress control of the converter control layer is to optimize the thermal stress by adjusting state variables related to junction temperature, such as switching frequency, output power, direct-current side voltage and the like, but the thermal control target is difficult to be compatible with other control targets, and the thermal stress control usually affects the output effect. The method utilizes the complementary characteristics of different devices among the systems, has little influence on the overall cost and the output performance of the system, is an ideal heat control strategy, but has poor universality and is difficult to be applied to a converter which is used independently.

The proposed thermal stress control strategy is mostly based on a loss optimization or junction temperature smoothing method, and relatively few researches are made on the thermal stress balance problem between multi-level converter modules by reducing the thermal stress of the power device or reducing the thermal swing to improve the thermal performance.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides a thermal stress balance prediction control method and a system suitable for a multilevel converter, which can balance the thermal stress among modules and balance the aging speed of the modules while ensuring the output precision, thereby maximizing the service life of the whole device.

The invention solves the technical problems through the following technical scheme: a thermal stress balance prediction control method suitable for a multilevel converter comprises the following steps:

step 1: collecting control variable output instantaneous values of the multilevel converter at the moment k, and obtaining control variable output predicted values under different switch state combinations at the moment k +1 by the control variable output instantaneous values;

step 2: setting a control variable reference value at the moment k +1, constructing a cost function according to the control variable reference value at the moment k +1 and the control variable output predicted value in the step 1, taking the total level corresponding to the minimum cost function as the optimal output total level at the moment k +1 under the constraint of a limited control set, and obtaining the level increment at the optimal state at the moment k + 1;

and step 3: selecting a module capable of acting at the moment k +1 according to the level increment in the step 2;

and 4, step 4: collecting the conduction voltage and current of each power device, and estimating the junction temperature value of each module power device according to the conduction voltage and current;

and 5: building a switch distribution function of the modules according to the junction temperature value in the step 4, and selecting the modules needing to be operated at the moment k +1 from the modules capable of being operated in the step 3 according to the switch distribution function of each module;

step 6: and (4) allocating the level increment in the step (2) to a specific bridge arm of a module needing action so as to realize the balance of the thermal stress of the multiple modules.

The thermal stress balance prediction control method comprises the steps of firstly constructing a cost function according to a control variable reference value and a control variable output predicted value, selecting an optimal output total level through a limited control set, determining a level increment, then determining a module capable of acting according to the level increment, realizing on-off state distribution of the module, and finally constructing a switch distribution function according to a junction temperature value and determining the module needing to act; in the control method, the cost function is designed according to the output predicted value without correction, so that the output precision is ensured; the switch distribution function takes the junction temperature value for reflecting the thermal stress of each module as a design basis, and determines the switch state of the module through the switch distribution function, so that the effect of balancing the thermal stress of multiple modules is achieved, the loss of the multi-level converter is reduced, the reliability is improved, the service life is prolonged, and the difficulty in heat dissipation design is simplified.

Further, in step 2, a specific expression of the cost function is as follows:

wherein g is a cost function of the multilevel converter,

for the jth control variable reference value at time k +1, x_j ^p(k +1) is the output predicted value of the jth control variable under different switch state combinations at the moment of k +1, and lambda_jIs the weight coefficient corresponding to the jth control variable, and n is the number of control variables.

The cost function is directly obtained according to the control variable reference value and the control variable output predicted value without correction, and the output precision of the multilevel converter is guaranteed.

Further, in step 2, the Z level states adjacent to the front and back of the output total level H (k) of the multi-level converter at time k are used as a finite control set, and the constraint that the output total level H (k +1) at time k +1 in the finite control set is as follows:

H(k+1)∈{H(k)-Z,H(k)-Z+1,…,H(k),…,H(k)+Z-1,H(k)+Z}

wherein H (k) ═ Q₁(k)+…+Q_i(k)+…+Q_m(k)，Q_i(k) Is the output level of the ith module at the moment of k, m is the total number of modules of the multilevel converter, Q_i(k)＝1,0,-1。

Further, in step 3, a method for selecting modules capable of operating at the time k +1 is as follows:

when level is increased by Δ H>At 0, the level Q is output from the current_i(k) Selecting an action module from the modules with the value of-1 or 0; when level is increased by Δ H<At 0, the level Q is output from the current_i(k) Selecting an action module from the modules of 1 or 0; when the level increment Δ H is 0, the next time the level Q is output_i(k +1) and the current output level Q_i(k) And keeping consistent, wherein the state of the power device in each module at the next moment is kept consistent with the state at the current moment.

Further, in step 5, the specific expression of the switch distribution function of the ith module is

Wherein, C_i(k) The function is assigned to the switch of the ith module at time k,

the cumulative value of the action times of the switch at the time 0 to k of the ith module,

and (4) the temperature value of the power device of the ith module at the moment k, wherein alpha is a distribution factor.

The switch distribution function of the module comprises two parts, wherein one part is an accumulated value of the action times of the module switch, reflects the switch action frequency of the module and ensures the uniform switch action frequency of each module, and the other part is a junction temperature value related to the thermal stress; increasing the partition factor alpha so thatAnd

kept in the same order of magnitude to better reflect

And

the balance of thermal stress among the modules is achieved.

Further, in step 5, a method for selecting a module requiring an action includes:

step 5.1: in each control period, calculating switch distribution function values of all modules capable of acting;

step 5.2: the calculated switch allocation function values are sorted,

step 5.3: selecting | Δ H | modules corresponding to the switch distribution function values from the maximum value of the switch distribution function value sequence as modules needing action at the moment k +1, wherein Δ H is a level increment at the moment k + 1;

step 5.4: and clearing the switch action frequency accumulated value corresponding to the module needing action, and continuously accumulating and counting the switch action frequency accumulated value corresponding to the module not needing action.

Further, in step 6, a specific method for allocating the level increment to a specific bridge arm of the module that needs to be operated includes: the bridge arm of each module comprises a left bridge arm and a right bridge arm, and a status flag bit flagi is introduced into the ith module;

if flagi is equal to 1, preferentially judging the switching state of the left bridge arm or the right bridge arm of the ith module, if the switches of the left bridge arm or the right bridge arm can act, operating the left bridge arm or the right bridge arm according to the level increment, and if the switches of the left bridge arm or the right bridge arm cannot act, operating the right bridge arm or the left bridge arm according to the level increment;

if flagi is-1, preferentially judging the switching state of the right bridge arm or the left bridge arm of the ith module, if the switch of the right bridge arm or the left bridge arm can act, operating the right bridge arm or the left bridge arm according to the level increment, and if the switch of the right bridge arm or the left bridge arm can not act, operating the left bridge arm or the right bridge arm according to the level increment;

and after the switch of the right bridge arm or the left bridge arm acts each time, flagi is inverted so as to balance the action times of the switches of the left bridge arm and the right bridge arm of each module.

Correspondingly, a thermal stress balance predictive control system suitable for a multilevel converter is used for implementing the thermal stress balance predictive control method, and comprises the following steps:

the output prediction module is used for acquiring a control variable output instantaneous value of the multilevel converter at the current moment and obtaining a control variable output prediction value under different switch state combinations at the next moment according to the control variable output instantaneous value;

the finite set switch state optimizing module is used for obtaining the output total level at the next moment and the level increment at the next moment according to the given control variable reference value at the next moment and the control variable output predicted value under the constraint of a finite control set;

the voltage and current acquisition module is used for acquiring the conduction voltage and current of each module power device;

the junction temperature estimation module is used for estimating a junction temperature value of each module power device according to the conduction voltage and the current of each module power device;

the switch distribution function calculation module is used for acquiring the switch action time accumulated value of each module at the current moment and obtaining a switch distribution function value of each module according to the power device temperature junction value and the switch action time accumulated value of each module;

the switch state distribution module is used for sequencing the switch distribution function values of all the modules and selecting the module needing action at the next moment from the maximum value of the switch distribution function values according to the level increment at the next moment;

and the switching action bridge arm distribution module is used for distributing the level increment at the next moment to the bridge arm needing action of the module needing action.

Advantageous effects

Compared with the prior art, the thermal stress balance prediction control method suitable for the multilevel converter, provided by the invention, comprises the steps of firstly constructing a cost function according to a control variable reference value and a control variable output predicted value, selecting an optimal output total level through a limited control set, determining a level increment, then determining an operable module according to the level increment, realizing the state allocation of the module, finally designing a switch allocation function according to a junction temperature value, determining the module required to be operated, and allocating the level increment to the module required to be operated; in the control method, the cost function is designed according to the output predicted value without correction, so that the output precision is ensured; the switch distribution function takes the junction temperature value for reflecting the thermal stress of each module as a design basis, and determines the switch state of the module through the switch distribution function, so that the effect of balancing the thermal stress of multiple modules is achieved, the aging speed of each module is balanced, the loss of each module of the multilevel converter is reduced, the reliability is improved, the service life is prolonged, and the difficulty in heat dissipation design is simplified.

Drawings

In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only one embodiment of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.

FIG. 1 is a flow chart of a thermal stress balance predictive control method in an embodiment of the invention;

FIG. 2 illustrates four output level states of the multilevel converter module in accordance with an embodiment of the present invention;

fig. 3 is a control schematic diagram of a thermal stress balance predictive control system in an embodiment of the invention.

Detailed Description

The technical solutions in the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

As shown in fig. 1, the method for predictive control of thermal stress balance for a multilevel converter provided by the present invention includes the following steps:

1. acquiring the jth of a multilevel converter at time kControl variable output instantaneous value x_j(k) From the mathematical model of the multilevel converter, by x_j(k) Obtaining the output predicted value of the jth control variable under different switch state combinations at the moment of k +1

As shown in fig. 2, each module has four different switch states, and only one of the upper and lower power devices of a single bridge arm can be turned on, where an upper tube is turned on and a lower tube is turned off to indicate that the bridge arm state is 1, and an upper tube is turned on and a lower tube is turned off to indicate that the bridge arm state is 0. The multi-level converter is composed of a plurality of modules, and corresponds to a plurality of different switch state combinations, and the total output level of the multi-level converter is different under different switch state combinations.

Let the left half-bridge state of the ith module be B_2i-1And the right half-bridge state is marked as B_2iThe output level of the ith module is Q_iThen the total output level H of the multilevel converter with n blocks is:

H＝Q₁+Q₂+…+Q_i+…+Q_n(1)

wherein the content of the first and second substances,

the mathematical models of different types of multilevel converters are different, and the construction method of the mathematical model of the multilevel converter is the prior art, for example, the construction method of the mathematical model of the cascaded H-bridge type multilevel converter can refer to: dinghongqi, research on modular multilevel power amplifiers and device development [ D ]. Hunan university, 2018.

2. Let the jth control variable reference value at time k +1 be x_j ^*(k +1) from

And x_j ^p(k +1) constructing a cost function of the multilevel converter, and optimizing the cost function under the constraint of a limited control setAnd taking the small corresponding total level as the optimal output total level at the moment k +1, and obtaining the level increment at the optimal state at the moment k + 1.

The specific expression of the cost function is as follows:

wherein g is a cost function of the multilevel converter,

reference value, x, for the j-th control variable at time k +1_j ^p(k +1) is the output predicted value of the jth control variable under different switch state combinations at the moment of k +1, and lambda_jWeight coefficient (lambda) corresponding to jth control variable_jEmpirically set), n is the number of control variables. The cost function is directly obtained according to the control variable reference value and the control variable output predicted value without correction, and the output precision of the multilevel converter is guaranteed.

Taking Z level states adjacent to the front and back of the total level H (k) output by the multi-level converter at the time k as a finite control set, and the constraint of the total level H (k +1) output at the time k +1 in the finite control set is as follows:

H(k+1)∈{H(k)-Z,H(k)-Z+1,…,H(k),…,H(k)+Z-1,H(k)+Z} (4)

wherein H (k) ═ Q₁(k)+…+Q_i(k)+…+Q_m(k)，Q_i(k) Is the output level of the ith module at the moment of k, m is the total number of modules of the multilevel converter, Q_i(k)＝1,0,-1。

Under the constraint of a limited control set, the cost function g is minimum as an optimization target, the total level corresponding to the minimum value of the cost function is taken as the output total level H (k +1) at the moment k +1, the level increment delta H at the moment k +1 is obtained according to H (k +1) and H (k), the delta H is H (k +1) -H (k), and the delta H belongs to (-Z, Z).

3. And selecting a module capable of acting at the moment k +1 according to the level increment delta H.

When level is increased by Δ H>At 0, the current output level Q_i(k) Module of 1Can increase the level again, from the current output level Q_i(k) Selecting an action module from the modules with the value of-1 or 0; when level is increased by Δ H<At 0, the current output level Q_i(k) The-1 module can no longer reduce the level, from the current output level Q_i(k) Selecting an action module from the modules of 1 or 0; when the level increment Δ H is 0, the next time the level Q is output_i(k +1) and the current output level Q_i(k) And keeping consistent, wherein the state of the power device in each module at the next moment is kept consistent with the state at the current moment.

4. Collecting the conduction voltage and current of each module power device, and estimating the junction temperature value of each module power device according to the conduction voltage and current.

The junction temperature value of the power device is related to the thermal stress, the average junction temperature of each module is reflected, a switch distribution function is constructed according to the junction temperature value of the power device, and the switch state of each module is distributed according to the switch distribution function, so that the thermal stress balance among the modules is ensured. The estimation of the junction temperature value of the power device is the prior art, and can refer to Ahmed M R, Putrus G A. the method for predicting IGBT junction temperature under conversion condition [ C ]// reference of the IEEE industry Electronics society. IEEE, 2008.

5. Acquiring the accumulated value of the switching action times of each module at the current moment, constructing a switching distribution function of each module according to the accumulated value of the switching action times and the junction temperature value of the power device, and selecting modules needing to be operated at the moment of k +1 from the modules which can be operated according to the switching distribution function of each module, wherein the number of the modules needing to be operated is equal to the absolute value of the level increment. The concrete expression of the switch distribution function of the ith module is

Wherein, C_i(k) The function is assigned to the switch of the ith module at time k,is the ithThe accumulated value of the action times of the module switch from 0 to k,

and (4) the temperature value of the power device of the ith module at the moment k, wherein alpha is a distribution factor. The distribution factor alpha is set by adopting a trial-and-error method or according to experience, and the specific process determined by adopting the trial-and-error method is as follows: let the initial value of alpha be

And observing the effect of thermal stress balance, and if the effect is poor, replacing the effect with a value adjacent to the initial value of alpha until the position with good thermal stress balance effect is reached.

The switch distribution function of each module comprises two parts, one part is an accumulated value of the action times of the module switch

Reflects the switching action frequency of the modules, ensures the uniform switching action frequency of each module, and has the other part of junction temperature values related to thermal stressThe switch distribution of the multiple modules is carried out through the temperature value, so that the thermal stress balance of each module is ensured; increased partition factor alpha, such thatAnd

kept in the same order of magnitude to better reflectAnd

the balance of thermal stress among the modules is achieved.

The selection method of the module needing action comprises the following steps:

5.1: in each control period, calculating switch distribution function values of all modules capable of acting;

5.2: the calculated switch allocation function values are sorted from large to small,

5.3: selecting | Δ H | modules corresponding to the switch distribution function values from the maximum value of the switch distribution function value sequence as modules needing action at the moment k +1, wherein Δ H is a level increment at the moment k + 1;

5.4: integrating the switch operation times corresponding to the module to be operatedAnd clearing, and continuously accumulating and counting the switch action time accumulated value corresponding to the module which does not need to be acted.

6. And distributing the level increment delta H to a specific bridge arm of a module needing to act so as to realize the balance of the thermal stress of the multiple modules. The bridge arm of each module comprises a left bridge arm and a right bridge arm, a state flag bit flagi is introduced into the ith module, and the switching state of which bridge arm of the module is preferentially judged according to the state flag bit:

if flagi is 1, preferentially judging the switching state of the left bridge arm of the ith module, if the switch of the left bridge arm can act, operating the left bridge arm according to the level increment, and if the switch of the left bridge arm can not act, operating the right bridge arm according to the level increment; and if flagi is equal to-1, preferentially judging the switching state of the right arm of the ith module, if the switch of the right arm can be operated, operating the right arm according to the level increment, and if the switch of the right arm can not be operated, operating the left arm according to the level increment.

Or: if flagi is 1, preferentially judging the switching state of the right bridge arm of the ith module, if the switch of the right bridge arm can be operated, operating the right bridge arm according to the level increment, and if the switch of the right bridge arm can not be operated, operating the left bridge arm according to the level increment; and if flagi is equal to-1, preferentially judging the switching state of the left arm of the ith module, if the switch of the left arm can be operated, operating the left arm according to the level increment, and if the switch of the left arm can not be operated, operating the right arm according to the level increment.

And after the switch of the right bridge arm or the left bridge arm acts each time, flagi is inverted so as to balance the action times of the switches of the left bridge arm and the right bridge arm of each module.

As shown in fig. 3, a thermal stress balance prediction control system suitable for a multilevel converter is used to implement the thermal stress balance prediction control method, and includes:

an output prediction module 1 for obtaining the jth control variable output instantaneous value x of the multilevel converter 6 at the current time_j(k) And outputs the instantaneous value x according to the jth control variable_j(k) Obtaining the output predicted value x of the jth control variable under different switch state combinations at the next moment_j ^p(k+1)；

A finite set switch state optimizing module 2, which is used for setting the jth control variable reference value at the given next moment as x under the constraint of the finite control set_j ^*(k +1) th and jth control variable output prediction valuesObtaining the output total level H (k +1) at the next moment and the level increment delta H at the next moment;

the voltage and current acquisition module 3 is used for acquiring the conduction voltage and current of each module power device;

a junction temperature estimation module 4 for estimating the junction temperature value of each module power device according to the conduction voltage and current of each module power device

A switch distribution function calculation module 5 for obtaining the switch action times accumulated value of each module at the current time

And according to the junction temperature value of the power device of each module

And number of switching operationsCumulative value

Obtaining switch distribution function value C of each module_i(k)；

A switch state allocation module 6 for allocating the switch of all modules to a function value C_i(k) Sorting, and selecting a module needing action at the next moment from the maximum value of the switch distribution function value according to the level increment delta H at the next moment;

and the switching action bridge arm distribution module 7 is used for distributing the level increment delta H at the next moment to the action required bridge arm of the action required module.

The multilevel converter 8 controls the action of the power device of each module according to the action condition of the left and right bridge arms of each module at the next moment to achieve the control effect, the whole control system is closed-loop control, the effect of multi-module thermal stress balance can be achieved, R is load, x is_jAre control variables.

The above disclosure is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of changes or modifications within the technical scope of the present invention, and shall be covered by the scope of the present invention.