A multi-level inverter predictive control method and system without weight coefficient
By dividing the level combinations of multilevel inverters into redundant groups and calculating the cost function of current and capacitor voltage deviations, the problems of large computational load and complex weight adjustment in existing technologies are solved, and efficient and accurate level prediction is achieved.
Patent Information
- Application Number
- CN202210395935.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-15
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-04-15
AI Technical Summary
Existing predictive control methods for multilevel inverters involve large computational loads and require weight adjustments under different operating conditions, affecting the rationality of the output results.
A multi-level inverter predictive control method with unweighted coefficients is adopted. By dividing the level combination into redundant groups, calculating the cost function of current prediction deviation and capacitor voltage deviation, and selecting the optimal level combination, the amount of calculation is reduced and it is applicable to different operating conditions.
It achieves low computational load, high speed, high accuracy, wide applicability, no need to adjust weights according to operating conditions, and more accurate output level prediction.
Smart Images

Figure CN115021598B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of current conversion control, and in particular relates to a multi-level inverter predictive control method and system without weight coefficients. Background Art
[0002] In the decision-making of current conversion control, it is necessary to select a level combination whose current prediction result is close to the expected one, while trying to ensure the stability of the capacitor voltage. In order to comprehensively consider the above two factors, the MPCC algorithm designs a cost function and corresponding process. The most common MPCC algorithm adds weights to the cost function so that the calculation result of the cost function includes the result of current prediction and the impact of capacitor voltage changes. The algorithm only needs to select a set of level combinations to output based on the calculated value of the cost function. The cost function and process of the algorithm are as follows: Figure 1 As shown. The cost function used is g=Dis+λ*Dif, where Dis is the "prediction-reference" deviation of the current vector, Dif is the total deviation of the capacitor voltage, and λ is the weight in the cost function, which determines the influence coefficient of the current prediction deviation and the capacitor voltage deviation. After setting a reasonable λ value, the smaller g is, the better the selected level combination is. This algorithm ensures the quality of the output current and the stability of the capacitor voltage, but the amount of calculation is too large, and different cost function weights need to be set according to different working conditions. The rationality of the weight has a great influence on the circuit output result. Summary of the invention
[0003] The present invention provides a multi-level inverter prediction control method and system without weight coefficients, thereby solving the technical problems of large amount of calculation in the inverter output level prediction method in the prior art and the need to change the weights when applied to different working conditions.
[0004] The first aspect of the present invention discloses a multi-level inverter predictive control method without weight coefficients, comprising:
[0005] Acquire all level combinations of the multilevel inverter, calculate the voltage vector of each level combination, divide the level combinations with the same voltage vector into one group, and obtain N redundant groups;
[0006] Calculating a first cost function of each of the redundant groups, and obtaining M redundant groups whose first cost functions satisfy a preset condition, wherein the first cost function is a current prediction deviation function;
[0007] A second cost function of each level combination in the M redundant groups is calculated, and the level combination with the smallest second cost function is used as the output level, where the second cost function is a capacitor voltage deviation function.
[0008] Preferably, calculating the first cost function of each of the redundancy groups specifically includes:
[0009] Obtaining parameters of the multilevel inverter at time k under a level combination in each redundant group, the parameters including a current vector, a voltage vector and a load back electromotive force vector;
[0010] Determine, according to the parameters at the time k, a predicted current value of the multilevel inverter corresponding to each of the redundant groups at the time k+1;
[0011] A first cost function of each of the redundant groups is determined according to the current prediction value at the k+1 moment and a preset reference current.
[0012] Preferably, determining the current prediction value of the multilevel inverter corresponding to each redundant group at time k+1 according to the parameter at time k specifically includes:
[0013] The current prediction value of the multilevel inverter corresponding to each of the redundancy groups at time k+1 is determined according to a first formula, wherein the first formula is:
[0014]
[0015] In the formula, I P (k+1) is the current prediction value of the multilevel inverter at time k+1 under a level combination in each redundant group, R and L are the load resistance and load inductance of the multilevel inverter respectively, Ts is the sampling time, I(k) is the current vector of the multilevel inverter at time k, V(k) is the voltage vector of the multilevel inverter at time k, and e(k) is the load back electromotive force vector of the multilevel inverter at time k.
[0016] Preferably, the load back electromotive force vector of the multilevel inverter at time k is determined according to a second formula, and the second formula is:
[0017]
[0018] Wherein, e(k-1) is the load back electromotive force vector of the multilevel inverter at the k-1 moment, V(k-1) is the voltage vector of the multilevel inverter at the k-1 moment, R and L are the load resistance and load inductance of the multilevel inverter respectively, Ts is the sampling time, I(k) is the current vector of the multilevel inverter at the k moment, and I(k-1) is the current vector of the multilevel inverter at the k-1 moment.
[0019] Preferably, determining the first cost function of each of the redundancy groups according to the current prediction value at the k+1 time and the preset reference current specifically includes:
[0020] The first cost function of each of the redundancy groups is determined according to a third formula, where the third formula is:
[0021]
[0022] Where g1 is the first cost function, is the current prediction value I of the multilevel inverter at time k+1 P The real part of (k+1), is the current prediction value I of the multilevel inverter at time k+1 P The imaginary part of (k+1), is the preset reference current I at time k+1 r The real part of (k+1), is the preset reference current I at time k+1 r The imaginary part of (k+1).
[0023] Preferably, obtaining M redundancy groups whose first cost function satisfies a preset condition specifically includes:
[0024] Sorting the first cost function of each of the redundant groups from small to large to obtain a sequence;
[0025] Get the first M redundancy groups in the sequence.
[0026] Preferably, calculating the second cost function of each level combination in the M redundancy groups specifically includes:
[0027] Obtaining a voltage value of each capacitor and a current value flowing through the capacitor of the multi-level inverter at time k under each level combination in the M redundancy groups;
[0028] Determine a predicted voltage value of each capacitor at time k+1 according to the voltage value of each capacitor and the current value flowing through the capacitor of the multilevel inverter at time k;
[0029] The second cost function of the corresponding level combination is determined according to the voltage prediction value of each capacitor at the k+1 time.
[0030] Preferably, determining the predicted voltage value of each capacitor at time k+1 according to the voltage value of each capacitor and the current value flowing through the capacitor at time k of the multilevel inverter specifically includes:
[0031] The voltage prediction value of each capacitor at time k+1 is determined according to the fourth formula, which is:
[0032]
[0033] Where V C (k+1) is the predicted value of the capacitor voltage at time k+1, V C(k) is the voltage value of the capacitor at time k, Ts is the sampling time, C is the capacitance, I C (k) is the current value flowing through the capacitor at time k.
[0034] Preferably, determining the second cost function of the corresponding level combination according to the voltage prediction value of each capacitor at the k+1 time specifically includes:
[0035] The second cost function of the corresponding level combination is determined according to the fifth formula, and the fifth formula is:
[0036]
[0037] Where g2 is the second cost function, V Cj (k+1) is the voltage prediction value of the jth capacitor in a level combination at time k+1, and j=1,2,…,M.
[0038] A second aspect of the present invention discloses a multi-level inverter predictive control system without weight coefficients, comprising:
[0039] A redundancy group determination module, the redundancy group determination module is used to obtain all level combinations of the multilevel inverter, calculate the voltage vector of each level combination, and divide the level combinations with the same voltage vector into one group to obtain N redundancy groups;
[0040] A screening module, the screening module is used to calculate a first cost function of each of the redundant groups, and obtain M redundant groups whose first cost functions meet preset conditions, wherein the first cost function is a current prediction deviation function;
[0041] An output level prediction module is used to calculate a second cost function for each level combination in the M redundant groups, and take the level combination with the smallest second cost function as the output level, wherein the second cost function is a capacitor voltage deviation function.
[0042] Compared with the prior art, the present invention has the following beneficial effects:
[0043] The inverter circuit output level prediction method and system of the present invention have small calculation amount, fast calculation speed, high accuracy and can be applied to different working conditions without modification, and have a wide range of applications. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 It is a flow chart of the method for predicting the output level of an inverter circuit in the prior art;
[0045] Figure 2 A schematic flow chart of a multi-level inverter predictive control method without weight coefficients provided in an embodiment of the present invention;
[0046] Figure 3 A detailed flow chart of a multi-level inverter predictive control method without weight coefficients provided in an embodiment of the present invention;
[0047] Figure 4 A circuit diagram of a four-level three-phase inverter provided by an embodiment of the present invention;
[0048] Figure 5 A vector radar diagram of a four-level three-phase inverter provided by an embodiment of the present invention;
[0049] Figure 6 A schematic diagram of the structure of a multi-level inverter predictive control system without weight coefficients provided in an embodiment of the present invention.
[0050] In the figure, 101 is a redundancy group determination module, 102 is a screening module, and 103 is an output level prediction module. DETAILED DESCRIPTION
[0051] The technical scheme of the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation cases. It should be understood that the following embodiments are only exemplary descriptions and explanations of the present invention and should not be construed as limiting the scope of protection of the present invention. All technologies implemented based on the above content of the present invention are included in the scope that the present invention is intended to protect.
[0052] A first aspect of the present invention provides a multi-level inverter predictive control method without weight coefficients, such as Figure 2 and Figure 3 As shown, the following steps are included:
[0053] Step 1: Obtain all level combinations of the multilevel inverter, calculate the voltage vector of each level combination, and divide the level combinations with the same voltage vector into one group to obtain N redundant groups.
[0054] The multi-level inverter in the embodiment of the present invention is specifically a four-level three-phase inverter.
[0055] The following will be Figure 4 The four-level three-phase inverter shown is taken as an example to illustrate the method and system of the present invention.
[0056] For a four-level three-phase inverter, there are 4×4×4=64 level combinations for the three phases. Then, the voltage vector of each level combination is calculated using formula (1):
[0057]
[0058] Where V is a voltage vector of a level combination, V dc is the DC voltage of the four-level three-phase inverter, L a , L b and Lc is the output level of the three-phase circuit, i is the output current of the four-level three-phase inverter.
[0059] When the three-phase levels are 1, 2, and 3 respectively, the voltage vector calculated by formula (1) is:
[0060]
[0061] When the three-phase levels are 0, 1, and 2 respectively, the voltage vector calculated by formula (1) is:
[0062]
[0063] It can be seen that the voltage vectors of the level combination 1, 2, 3 and the level combination 0, 1, 2 are equal.
[0064] The embodiment of the present invention divides the level combinations with equal voltage vectors into one group, and can divide the 64 level combinations into 37 redundant groups. According to the values of the real and imaginary parts of the voltage vector, each group is marked at the corresponding position in the complex coordinate system, and the following can be obtained: Figure 5 The vector radar diagram shown in the figure. The level combinations at the same point in the figure have equal voltage vectors, and they are redundant combinations of each other.
[0065] Step 2: Calculate the first cost function of each redundant group, and obtain M redundant groups whose first cost functions meet preset conditions. The first cost function is a current prediction deviation function, which specifically includes:
[0066] Step 21: Obtain parameters of the multilevel inverter at time k in a level combination in each redundant group, the parameters including a current vector, a voltage vector, and a load back electromotive force vector.
[0067] Step 22: Determine the current prediction value of the multilevel inverter corresponding to each redundancy group at time k+1 according to the parameters at time k.
[0068] The load currents of the three phases of the four-level three-phase inverter in the embodiment of the present invention are I a ,I b ,I c , then the current vector I is as follows:
[0069]
[0070] For a circuit with resistive-inductive load, assuming that the load resistance and inductance of each phase are R and L, and the load back electromotive force vector is e, the relationship between the voltage vector and the current vector of the circuit is as shown in formula (3):
[0071]
[0072] Using a discrete time model with sampling time Ts, replacing dIdt in formula (3) by Euler approximation, we obtain the discrete relationship described in formula (4):
[0073]
[0074] If the k moment is regarded as the current moment, I(k) is regarded as the current current vector, and V(k) is regarded as the voltage vector calculated by the currently selected level combination, then I(k+1) can be regarded as the predicted value of the current vector at the next moment, and the superscript P is used to represent the predicted value. The relationship between the current vector and the voltage vector can be organized as follows:
[0075]
[0076] In the formula, I P (k+1) is the current prediction value of the multilevel inverter at time k+1 under a level combination in each redundant group, R and L are the load resistance and load inductance of the multilevel inverter respectively, Ts is the sampling time, I(k) is the current vector of the multilevel inverter at time k, V(k) is the voltage vector of the multilevel inverter at time k, and e(k) is the load back electromotive force vector of the multilevel inverter at time k.
[0077] Since e(k) cannot be calculated from the observed value at the current moment, and the value of e(k) does not change much within a sampling time, the embodiment of the present invention uses e(k-1) to approximate e(k) when predicting current. The calculation formula of e(k) is as shown in formula (6):
[0078]
[0079] Wherein, e(k-1) is the load back electromotive force vector of the multilevel inverter at time k-1, V(k-1) is the voltage vector of the multilevel inverter at time k-1, R and L are the load resistance and load inductance of the multilevel inverter respectively, Ts is the sampling time, I(k) is the current vector of the multilevel inverter at time k, and I(k-1) is the current vector of the multilevel inverter at time k-1.
[0080] Among them, I(k) and I(k-1) are calculated by the observed values of the load current at the current moment and the previous moment respectively through formula (2), and V(k) and V(k-1) are calculated by the level combination selected at the current moment and the previous moment respectively through formula (1).
[0081] Since the voltage vectors calculated by the redundant groups are equal, their corresponding current prediction results are also equal. Therefore, the embodiment of the present invention uses the current prediction value of a level combination in the redundant group to represent the current prediction value of the entire redundant group.
[0082] Step 23: Determine the first cost function of each redundancy group according to the current prediction value at time k+1 and the preset reference current.
[0083] The embodiment of the present invention determines the first cost function of each redundancy group according to formula (7):
[0084]
[0085] Where g1 is the first cost function, which is the current prediction deviation function, is the current prediction value I of the multilevel inverter at time k+1 P The real part of (k+1), is the current prediction value I of the multilevel inverter at time k+1 P The imaginary part of (k+1), is the preset reference current I at time k+1 r The real part of (k+1), is the preset reference current I at time k+1 r The imaginary part of (k+1).
[0086] Step 24: sort the first cost function of each redundant group from small to large to obtain a sequence; obtain the first M redundant groups in the sequence.
[0087] In the embodiment of the present invention, the smaller the first cost function is, the closer the result of the predicted current is to the reference value. Therefore, the present invention selects the smaller M redundant groups in the sequence.
[0088] Step 3: Calculate the second cost function of each level combination in the M redundant groups, and use the level combination with the smallest second cost function as the output level. The second cost function is a capacitor voltage deviation function, which specifically includes:
[0089] Step 31, obtaining the voltage value of each capacitor and the current value flowing through the capacitor at time k of the multi-level inverter under each level combination in the M redundancy groups;
[0090] Step 32: Determine a predicted voltage value of each capacitor at time k+1 according to the voltage value of each capacitor and the current value flowing through the capacitor of the multi-level inverter at time k.
[0091] In an ideal four-level three-phase inverter, the three capacitors on the DC bus are identical, and the voltage across each capacitor is constant. However, in reality, the capacitor voltage changes as the charge on the capacitor changes. As long as there is current passing through the capacitor, the capacitor voltage cannot be constant. The change in capacitor voltage will affect the voltage value of the output level, destroy the waveform of the output current, and may cause overvoltage in capacitors with larger voltages and their parallel circuit components, or even damage the components. Therefore, when designing the model prediction current control (MPCC) algorithm, the change in capacitor voltage must be predicted.
[0092] For a capacitor with a capacity of C, if the voltage across it is V, the charge is Q, and the current flowing through it is I, then the following relationship exists:
[0093]
[0094] By replacing dV in formula (8) with the Euler approximation C dt, we can get:
[0095]
[0096] Where V C (k+1) is the predicted voltage value of the capacitor at time k+1, Ts is the sampling time, C is the capacitance, and V C (k+1), we need to get the capacitor voltage value V at time k C (k) and the current flowing through the capacitor I C (k). Where V C (k) can be directly obtained by observing the capacitance, I C (k) needs to be calculated by combining the observed results of the circuit's three-phase load current with formulas (10) and (11).
[0097] Assume that the three-phase load currents of the current circuit are I a (k) I b (k) I c (k) Definition Figure 4 The current flowing out of NP1 and NP2 on the DC bus is I NP1 and I NP2 , then there exists a relationship as shown in formula (10):
[0098]
[0099] L in formula (10) a V C (k+1) corresponding to the a-phase switch state, L b VC (k+1) corresponding b-phase switch state, L c V C (k+1) corresponds to the c-phase switching state.
[0100] The current I passing through capacitors C1, C2, and C3 C1 (k) I C2 (k) and I C3 (k) can be calculated by formula (11):
[0101]
[0102] Substituting formula (11) into formula (9), the predicted voltage value of each capacitor at time k+1 can be obtained.
[0103] Step 33: Determine the second cost function of the corresponding level combination according to the voltage prediction value of each capacitor at time k+1.
[0104] The embodiment of the present invention determines the second cost function of the corresponding level combination according to formula (12):
[0105]
[0106] Where g2 is the second cost function, V Cj (k+1) is the voltage prediction value of the jth capacitor in a level combination at time k+1, and j=1, 2, ..., M, V dc is the DC link voltage.
[0107] Taking a four-level three-phase inverter as an example, step 33 is described in detail.
[0108] In a four-level three-phase inverter, the second cost function of the level combination is:
[0109]
[0110] Step 34: Use the level combination with the minimum second cost function as the output level.
[0111] The smaller the second cost function is, the more balanced the capacitor voltage is. Therefore, the embodiment of the present invention uses the level combination with the smallest second cost function as the output level.
[0112] The unweighted algorithm proposed in the present invention first divides the 64 level combinations into 37 redundant groups, then calculates the cost function g1 of the current 37 redundant groups, selects M (3 in the embodiment of the present invention) redundant groups with the smallest g1 result, and then calculates the cost function g2 of all level combinations in the 3 redundant groups, selects the level combination with the smallest g2 result, and uses the selected level combination as the output of the circuit at the next moment.
[0113] For example, among the 37 redundant groups calculated first, the g1 results of the three groups "110 / 221 / 332", "220 / 331", and "210 / 321" are the smallest. Then the g2 of the seven level combinations "110", "221", "332", "220", "331", "210", and "321" are calculated respectively. It is calculated that the g2 result of "331" is the smallest, so "331" is the level combination output by the circuit at the next moment.
[0114] In the above example, the algorithm calculates g1 37 times and g2 7 times, and the total number of cost function calculations is 44 times. For the existing weighted MPCC algorithm, if the cost function g is regarded as g1+λ*g2, the algorithm will calculate the cost function 128 times in one sampling cycle, which is much more than the algorithm proposed in the present invention. In addition, the prior art has to calculate V separately each time g2 is calculated. C1 (k+1), V C2 (k+1) and V C3 (k+1), so the amount of calculation of g2 is much larger than that of g1, so the amount of calculation of the algorithm proposed in the present invention is much smaller than that of the weighted algorithm, the amount of calculation is greatly reduced, and the calculation speed is improved. In addition, the algorithm proposed in the present invention does not use weights in calculations. When applied to different working conditions, there is no need to change the weights given by the algorithm, which reduces the complexity and does not cause the problem of inaccurate results caused by unreasonable weight settings.
[0115] The embodiment of the present invention takes into account that the priority of current prediction deviation is higher than that of capacitor voltage deviation, so the first cost function uses the current prediction deviation function; the second cost function uses the capacitor voltage deviation function, so that the prediction result is more accurate.
[0116] A second aspect of the present invention provides a multi-level inverter predictive control system without weight coefficients, such as Figure 6 As shown, it includes a redundancy group determination module 101, a screening module 102 and an output level prediction module 103;
[0117] The redundancy group determination module 101 is used to obtain all level combinations of the multi-level inverter, calculate the voltage vector of each level combination, and divide the level combinations with the same voltage vector into one group to obtain N redundancy groups;
[0118] The screening module 102 is used to calculate the first cost function of each redundant group, and obtain M redundant groups whose first cost functions meet preset conditions, where the first cost function is a current prediction deviation function;
[0119] The output level prediction module 103 is used to calculate the second cost function of each level combination in the M redundant groups, and take the level combination with the minimum second cost function as the output level, where the second cost function is a capacitor voltage deviation function.
[0120] The above is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with the art can make many possible changes and modifications to the technical solution of the present invention by using the technical content disclosed above, or modify it into an equivalent embodiment of equivalent changes without departing from the scope of the technical solution of the present invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention should fall within the scope of protection of the technical solution of the present invention.
Claims
1. A multi-level inverter predictive control method without weight coefficients, characterized in that: include: Obtain all level combinations of the multilevel inverter, calculate the voltage vector of each level combination, and group the level combinations with the same voltage vector into one group to obtain N redundant groups; Calculating a first cost function of each of the redundancy groups to obtain M redundancy groups whose first cost functions satisfy a preset condition, where the first cost function is a current prediction deviation function; Calculating a second cost function for each level combination in the M redundancy groups, and taking the level combination with the minimum second cost function as the output level, where the second cost function is a capacitance-voltage deviation function; Calculating a first cost function of each of the redundancy groups specifically includes: Obtaining parameters of the multilevel inverter at time k in a level combination in each redundant group, the parameters including a current vector, a voltage vector, and a load back electromotive force vector; Determining a predicted current value of the multilevel inverter at time k+1 corresponding to each of the redundant groups according to the parameters at time k; Determining a first cost function of each of the redundant groups according to the current prediction value at the k+1 moment and a preset reference current; Calculating a second cost function for each level combination in the M redundancy groups specifically includes: Obtaining a voltage value of each capacitor and a current value flowing through the capacitor of the multi-level inverter at time k under each level combination in the M redundancy groups; Determine a predicted voltage value of each capacitor at time k+1 based on the voltage value of each capacitor and the current value flowing through the capacitor at time k of the multilevel inverter; The second cost function of the corresponding level combination is determined according to the voltage prediction value of each capacitor at the k+1 time.
2. The method according to claim 1, wherein: Determining, according to the parameters at time k, a predicted current value of the multilevel inverter corresponding to each redundancy group at time k+1, specifically comprising: The current prediction value of the multilevel inverter corresponding to each redundancy group at time k+1 is determined according to a first formula, where the first formula is: Where, I P (k+1) is the current prediction value of the multilevel inverter at time k+1 under a level combination in each redundant group, R and L are the load resistance and load inductance of the multilevel inverter, respectively, Ts is the sampling time, I(k) is the current vector of the multilevel inverter at time k, V(k) is the voltage vector of the multilevel inverter at time k, and e(k) is the load back electromotive force vector of the multilevel inverter at time k.
3. The method according to claim 2, wherein: The load back electromotive force vector of the multilevel inverter at time k is determined according to a second formula, which is: Wherein, e(k-1) is the load back electromotive force vector of the multilevel inverter at time k-1, V(k-1) is the voltage vector of the multilevel inverter at time k-1, R and L are the load resistance and load inductance of the multilevel inverter respectively, Ts is the sampling time, I(k) is the current vector of the multilevel inverter at time k, and I(k-1) is the current vector of the multilevel inverter at time k-1.
4. The method according to claim 1, wherein: Determining a first cost function of each of the redundancy groups according to the current prediction value at the k+1 time and a preset reference current specifically includes: The first cost function of each redundancy group is determined according to a third formula, wherein the third formula is: Where g1 is the first cost function, is the current prediction value I of the multilevel inverter at time k+1 P The real part of (k+1), is the current prediction value I of the multilevel inverter at time k+1 P The imaginary part of (k+1), The reference current I at the preset time k+1 r The real part of (k+1), The reference current I at the preset time k+1 r The imaginary part of (k+1).
5. The method according to claim 1, wherein: Obtaining M redundancy groups whose first cost function satisfies a preset condition, specifically including: sorting the first cost function of each of the redundant groups from small to large to obtain a sequence; Get the first M redundant groups in the sequence.
6. The method according to claim 1, wherein: Determining a predicted voltage value of each capacitor at time k+1 based on the voltage value of each capacitor and the current value flowing through the capacitor of the multi-level inverter at time k specifically includes: The predicted voltage value of each capacitor at time k+1 is determined according to the fourth formula, which is: Where V C (k+1) is the predicted voltage value of the capacitor at time k+1, V C (k) is the voltage value of the capacitor at time k, Ts is the sampling time, C is the capacitance, I C (k) is the current value flowing through the capacitor at time k.
7. The method according to claim 1, wherein: Determining the second cost function of the corresponding level combination according to the voltage prediction value of each capacitor at the k+1 time specifically includes: The second cost function of the corresponding level combination is determined according to the fifth formula, and the fifth formula is: Where g2 is the second cost function, V Cj (k+1) is the predicted voltage value of the j-th capacitor in a level combination at time k+1, and j=1, 2, …, M.
8. A multi-level inverter predictive control system using the method according to any one of claims 1 to 7 without weight coefficients, characterized in that: include: a redundancy group determination module, the redundancy group determination module being configured to obtain all level combinations of the multilevel inverter, calculate a voltage vector for each level combination, and group the level combinations having the same voltage vector into one group to obtain N redundancy groups; a screening module, the screening module being configured to calculate a first cost function of each of the redundant groups, and obtain M redundant groups whose first cost functions satisfy a preset condition, wherein the first cost function is a current prediction deviation function; An output level prediction module is used to calculate a second cost function for each level combination in the M redundant groups, and take the level combination with the smallest second cost function as the output level, where the second cost function is a capacitor voltage deviation function.
