A neutral point potential control method applied to a neutral point clamped three-level inverter

Abstract

The present invention discloses a method for controlling the neutral point potential of a neutral point clamped three-level inverter. Based on the space vector modulation technology with one-phase clamping, the switching vectors and the action time are determined according to different clamping modes and the polarity of the neutral point current. The corresponding positive and negative clamping modes are selected according to the magnitude relationship of the DC side bus voltage. The normalized neutral point current under the corresponding clamping mode is calculated, and whether the triple-frequency AC fluctuation of the neutral point potential is controllable is judged by its polarity, and then the proportionality factor k is obtained. According to the clamping mode and the proportionality factor k, the final small sector division is determined, the switching sequence and the action time of each switching vector are determined, and the switching signals of each bridge arm are obtained by using a symmetric sequence. The present invention combines the neutral point potential control and the discontinuous modulation technology with one-phase clamping, enables the inverter to effectively control the neutral point potential balance and eliminate the triple-frequency AC fluctuation of the neutral point potential under full modulation index and full power factor conditions, and minimizes the switching loss to the greatest extent to improve the efficiency.

Description

Technical Field

[0001] The present invention belongs to the field of neutral-point clamped three-level inverters, and particularly relates to a high-efficiency neutral-point potential control method for effectively eliminating the triple-frequency fluctuation of the neutral-point potential and reducing the switching loss under full power factor and modulation index conditions. Background Art

[0002] Compared with two-level inverters, neutral-point clamped three-level inverters have smaller dv / dt, lower current harmonics and higher efficiency. Therefore, as an AC-DC energy conversion unit, they are widely used in medium-voltage high-power grid-connected systems, motor traction systems, RLC load island systems, etc. However, regardless of whether the switching network is of T-type or I-type structure, due to the charging and discharging processes of the positive and negative bus capacitors on the DC side, the neutral-point potential of the neutral-point clamped three-level inverter will generate triple-frequency AC fluctuations or even DC offsets, resulting in output voltage distortion, increased device voltage stress, and reduced capacitor life. Therefore, how to effectively control the neutral-point potential is one of the key issues.

[0003] There are many existing neutral-point potential control methods based on pulse width modulation strategies, including five-segment discontinuous modulation strategy (Discontinuous PWM, DPWM), seven-segment space vector modulation strategy (Space Vector PWM, SVPWM), and nine-segment virtual space vector modulation strategy (Virtual Space Vector PWM, VSVPWM) and corresponding carrier generation methods. Among them, for the neutral-point potential control methods based on DPWM and SVPWM, most of the triple-frequency fluctuations of the neutral-point potential under various conditions can be eliminated by switching or canceling positive and negative small vectors. However, under high modulation index and low power factor conditions, the triple-frequency fluctuations of the neutral-point potential still exist. Although the VSVPWM method can effectively control the neutral-point potential and eliminate the triple-frequency fluctuations under all conditions, there are 8 switching actions in each switching cycle, resulting in a large switching loss and reducing the efficiency of the inverter. Research scholars have also proposed some neutral-point potential control methods based on the hybrid of DPWM and VSVPWM, which can reduce the switching loss while eliminating the triple-frequency fluctuations of the neutral-point potential. However, in some methods, under extreme conditions, such as when the power factor angle is ±90°, the modulation strategy fails and the neutral-point potential is even uncontrollable; some methods use a look-up table screening method, which has a large amount of calculation, is complex to implement, and the switching loss has not been reduced to the optimal value. Summary of the Invention

[0004] The object of the present invention is to solve the above deficiencies and provide a neutral-point potential control method applied to a neutral-point clamped three-level inverter, which can effectively control the neutral-point potential and eliminate the triple-frequency fluctuations under full modulation index and full power factor conditions, and at the same time ensure that there is no switching action in one phase clamping in each switching cycle to reduce the switching loss.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A method for controlling the neutral point potential of a neutral point clamped three-level inverter, comprising the following steps:

[0007] Step 1: Obtain the clamping mode factor k according to the relative magnitude relationship between the positive and negative bus voltages on the DC side cp ;

[0008] Step 2: Let the proportionality factor k be 1, and calculate the normalized neutral point current i refα,β clamped to the positive level and the normalized neutral point current i La,b,c clamped to the negative level within the current switching period according to the reference voltage vector v op and the three-phase inductor current i on ;

[0009] Step 3: Determine the polarity of the corresponding neutral point current i cp obtained in Step 2 according to the clamping mode factor k op or i on obtained in Step 2 to obtain the proportionality factor k;

[0010] Step 4: Determine the corresponding switching sequence according to the reference voltage vector v refα,β , the clamping mode factor k cp obtained in Step 1 and Step 3, and the proportionality factor k, calculate the action time of each switching vector and obtain the switching signals of each phase.

[0011] In Step 1, the clamping mode factor k cp is selected as follows:

[0012]

[0013] where u c1 is the capacitor voltage of the positive bus on the DC side, and u c2 is the capacitor voltage of the negative bus on the DC side.

[0014] In Step 2, the normalized neutral point currents i op and i on are calculated as follows:

[0015] First, determine the large sector BS where the current voltage reference vector v refα,β is located:

[0016]

[0017] where the fix function represents truncating and taking the integer.

[0018] To simplify the implementation of the program, the voltage reference vector is transformed to the gh coordinate system for calculation, and the voltage reference vector v in the gh coordinate system is obtained. g ′ and v h ′:

[0019]

[0020] The voltage reference vectors located in other large sectors are all rotated to the first large sector to obtain the new voltage reference vectors v g and v h :

[0021]

[0022] Let k = 1 and determine the small sector SS where the current voltage reference vector is located:

[0023]

[0024] When k = 1, there is no 5th small sector. According to the switching action vector and the volt-second balance principle at k = 1, calculate the normalized midpoint current i with positive clamping level in the current switching period op and the normalized midpoint current i with negative clamping level on , and obtain the expressions of i op and i on as follows:

[0025]

[0026]

[0027] The i in the above formula x,y,z is determined by the large sector and the three-phase inductor current i La,b,c :

[0028]

[0029] In step three, the proportionality factor k is selected as follows:

[0030]

[0031] Finally, in step four, according to the voltage reference vector, the obtained clamping mode factor k cp and the proportionality factor k, determine the new small sector division, obtain the corresponding switching vectors and action times, and adopt the symmetric sequence arrangement V1-V2-V3-V4-V3-V2-V1, and the action time arrangement is T1 / 2-T2 / 2-T3 / 2-T4-T3 / 2-T2 / 2-T1 / 2, and then obtain the drive signals of the three-phase bridge arms.

[0032] The switching vectors are arranged in a symmetric sequence: V1 - V2 - V3 - V4 - V3 - V2 - V1. The switching vectors of the first major sector are shown in Table 1, and those of other major sectors are similar.

[0033] Table 1 Switching sequence of the first major sector

[0034]

[0035] Calculate the action time of each switching vector according to the volt - second balance principle:

[0036] V1T1 + V2T2 + V3T3 + V4T4 = V ref T s

[0037]

[0038] where T s is the switching period, and T1 * , T2 * , T3 * , T4 * are the per - unit action times of each switching vector.

[0039] The time sequence arrangement is T1 / 2 - T2 / 2 - T3 / 2 - T4 - T3 / 2 - T2 / 2 - T1 / 2. Taking the first major sector as an example, the action time of each switching vector is calculated as shown in Table 2, and those of other major sectors are similar.

[0040] Table 2 Action time of each switching vector in the first major sector

[0041]

[0042] Advantages of the present invention:

[0043] Compared with the prior art, the present invention is based on the space vector modulation technology with one - phase clamping. The switching vectors and action times are determined according to different clamping modes and the polarity of the mid - point current. The corresponding positive and negative clamping modes are selected according to the magnitude relationship of the DC - side bus voltage. Then, the normalized mid - point current under the corresponding clamping mode is calculated, and whether the triple - frequency AC fluctuation of the mid - point potential is controllable is judged by its polarity, and then the proportionality factor k is obtained. According to the clamping mode and the proportionality factor k, the final small - sector division is determined, the switching sequence and the action time of each switching vector are determined, and the switching signals of each bridge arm are obtained by using a symmetric sequence. The present invention effectively combines the mid - point potential control and the discontinuous modulation technology of one - phase clamping, enabling the neutral - point - clamped three - level inverter to effectively control the mid - point potential balance and eliminate the triple - frequency AC fluctuation of the mid - point potential under full modulation index and full power - factor conditions, while minimizing the switching loss to improve the efficiency to the greatest extent. Description of the drawings

[0044] Figure 1 This is the topology and system structure diagram of the neutral point clamped three-level inverter in the present invention;

[0045] Figure 2 These are the implementation steps of the present invention;

[0046] Figure 3 This is the specific method flow block diagram of the present invention;

[0047] Figure 4 This is the conventional voltage space vector diagram of the neutral point clamped three-level inverter;

[0048] Figure 5 This is the clamping mode factor k cp = 1, the available switching vectors and sector division when the proportionality factor k = 1;

[0049] Figure 6 This is the clamping mode factor k cp = 0, the available switching vectors and sector division when the proportionality factor k = 1;

[0050] Figure 7 This is the clamping mode factor k cp = 1, the available switching vectors and sector division when the proportionality factor k = 0;

[0051] Figure 8 This is the clamping mode factor k cp = 0, the available switching vectors and sector division when the proportionality factor k = 0;

[0052] Figure 9 This is the clamping mode factor k cp = 1, the switching sequence and action time allocation diagram when the voltage reference vector is in the 5th small sector of the first large sector with the proportionality factor k = 0;

[0053] Figure 10 This is the simulation comparison diagram of the neutral point potential fluctuation control effect under the condition that the power factor angle is -45°;

[0054] Figure 11 This is the simulation comparison diagram of the neutral point potential fluctuation control effect under the condition that the power factor angle is 45°;

[0055] Figure 12 This is the simulation comparison diagram of the neutral point potential fluctuation control effect under the condition that the power factor angle is -90°;

[0056] Figure 13 This is the simulation comparison diagram of the neutral point potential fluctuation control effect under the condition that the power factor angle is 90°;

[0057] Figure 14 This is the simulation comparison diagram of the device switching losses under various power factor conditions. Detailed implementation manners

[0058] The midpoint-clamped three-level inverter topology of the present invention is as Figure 1 shown, and the implementation steps are as Figure 2 shown. The specific method flow block diagram is as Figure 3 shown. The conventional voltage space vector diagram of the midpoint-clamped three-level inverter is as Figure 4 shown. The specific steps are as follows:

[0059] Step 1: Determine the value of the clamping mode factor k cp .

[0060] Compare the magnitudes of the positive and negative bus voltages u c1 and u c2 obtained by sampling to get k cp :

[0061]

[0062] where k cp = 1 indicates that one phase is clamped to the positive level, and k cp = 0 indicates that one phase is clamped to the negative level.

[0063] Step 2: Calculate the normalized midpoint currents i op and i on .

[0064] First, determine the large sector BS where the current voltage reference vector v refα,β is located:

[0065]

[0066] where the fix function represents truncating and taking the integer.

[0067] Convert the voltage reference vector v refα,β to the gh coordinate system for calculation to obtain the voltage reference vectors v g ' and v h ':

[0068]

[0069] From equations (2) and (3), rotate the voltage reference vectors of other large sectors to the first large sector to obtain the new voltage reference vectors v g and v h :

[0070]

[0071] Let k = 1, and in combination with equation (4), determine the small sector SS where the current voltage reference vector is located:

[0072]

[0073] Depend on Figure 5 and Figure 6 It can be seen that there is no fifth small sector when k = 1. According to the switching action vector and the volt-second balance principle when k = 1, the normalized midpoint current i clamped to a positive level in the current switching cycle is calculated. op and the normalized midpoint current i clamped to a negative level on , we can get i from equation (4) and equation (5) op and i on The expression is as follows:

[0074]

[0075]

[0076] where i x,y,z The large sector BS and the three-phase inductor current i La,b,c Determine, as shown in formula (8):

[0077]

[0078] Step 3: Determine the value of the scaling factor k.

[0079] When k cp =1 when comparing i op The polarity of k cp =0 when comparing i on The polarity of , the proportional factor k can be obtained from equations (1), (6) and (7):

[0080]

[0081] Step 4: Determine the corresponding switching sequence and calculate the action time of each switching vector.

[0082] See also Figures 5 - 8 , according to the clamping mode factor k cp The new size sector division and the available switch vector are determined by the scale factor k. The size sector where the current voltage reference vector is located can be obtained from equations (2) and (5).

[0083] According to the large and small sectors combined with Table 1, the switching sequence of the first large sector can be obtained, and the switching sequences of other large sectors are similar.

[0084] The action time of each switch vector is calculated according to the volt-second balance principle, as shown in equations (10) and (11):

[0085] V1T1+V2T2+V3T3+V4T4=V ref T s (10)

[0086]

[0087] Among them, T s is the switching period, T1 * , T2 * , T3 * , T4 * are the per-unit action times of each switching vector.

[0088] From Equation (1), Equation (9), Equation (10), Equation (11) and the large and small sectors, the action times of each switching vector in the first large sector can be obtained, as shown in Table 2. The action times of the switching vectors in other large sectors are similar.

[0089] Finally, by adopting the method of generating a symmetric switching sequence, within one switching period, the action sequence of the switching vectors is V1 - V2 - V3 - V4 - V3 - V2 - V1, and the corresponding action time arrangement is T1 / 2 - T2 / 2 - T3 / 2 - T4 - T3 / 2 - T2 / 2 - T1 / 2. Taking the fifth small sector of the first large sector when k cp = 1 and k = 0 as an example, as Figure 9 shown, the action sequence of the switching vectors is PNN - PON - POO - PPO - POO - PON - PNN, and the driving signal of phase A is clamped at the positive level throughout the switching period without switching action. It can be seen from Table 2 that at this time T2 = 0, so PON actually does not act.

[0090] To test the present invention, a neutral-point-clamped three-level inverter grid-connected system was built in MATLAB / Simulink. The parameters are shown in Table 3. During operation, the modulation index is about 1. The amplitude and phase of the inductor current are controlled by single current closed-loop control to achieve operation under low power factor conditions. At the same time, the loss model of the switching device is introduced to test the switching losses under different modulation strategies.

[0091] To verify that the present invention can effectively control the neutral-point potential and eliminate the third-harmonic fluctuation while reducing the switching loss under all working conditions, the present invention (abbreviated as the new clamping method), the neutral-point potential control method based on DPWM (abbreviated as the DPWM method), and the neutral-point potential control method based on VSVM (abbreviated as the VSVM method) are compared. The simulation results are shown in Figures 10 - 14 .

[0092] Table 3 Simulation system parameters

[0093]

[0094]

[0095] See Figures 10 - 13, the power factor angles of the operating conditions are -45°, 45°, -90°, and 90° respectively. It can be seen that there is a triple-frequency fluctuation in the midpoint potential under the condition of high modulation index and low power factor in the DPWM method, while the new clamping method and the VSVM method can eliminate the triple-frequency fluctuation of the midpoint potential.

[0096] See Figure 14 , it can be seen that under the condition of high power factor, the switching loss of the new clamping method is basically close to that of the DPWM method, while under the condition of full power factor, the switching loss of the new clamping method is greatly reduced compared with the VSVM method.

[0097] In summary, compared with the DPWM method, the present invention enables the neutral-point clamped three-level inverter to effectively control the neutral-point potential and eliminate the triple-frequency AC fluctuation of the neutral-point potential under the conditions of full modulation index and full power factor. At the same time, compared with the VSVM method, the switching loss is greatly reduced to improve the efficiency. The simulation results verify the correctness and effectiveness of the present invention.

Claims

1. A method for controlling the neutral point potential applied to a neutral point clamped three-level inverter, characterized in that It includes the following steps: Step 1: Obtain the clamping mode factor k according to the relative magnitude relationship of the positive and negative bus voltages on the DC side cp ; Step 2: Let the scaling factor k be 1, and calculate, according to the reference voltage vector v refα,β and the three-phase inductor current i La,b,c , the normalized midpoint current i op clamped to the positive level and the normalized midpoint current i on clamped to the negative level within the current switching period; Step 3: According to the clamping mode factor k obtained in Step 1 cp , determine the polarity of the corresponding midpoint current i op or i on obtained in Step 2, and obtain the proportionality factor k; Step 4: According to the reference voltage vector v refα,β , the clamping mode factor k cp obtained in Step 1 and Step 3 and the scaling factor k, determine the corresponding switching sequence, calculate the action time of each switching vector, and obtain the switching signals of each phase; In step one, the selection method of the clamping mode factor k cp is as follows: where, u c1 is the positive bus voltage on the DC side, and u c2 is the negative bus voltage on the DC side; In step two, the calculation methods of the normalized midpoint currents i op and i on are as follows: First, determine the large sector BS where the current voltage reference vector v is located: refα,β ​ where the fix function represents truncating and rounding; To simplify the program implementation, the voltage reference vector is transformed to the gh coordinate system for calculation, and the voltage reference vectors v g ′ and v h ′ are obtained: Rotate the voltage reference vectors located in other large sectors to the first large sector to obtain new voltage reference vectors v g and v h : Let k = 1 and determine the small sector SS where the current voltage reference vector is located: When k = 1, the fifth smallest sector does not exist. According to the switching action vector and volt-second balance principle at k = 1, calculate the normalized midpoint current i clamped to a positive level within the current switching period op and the normalized midpoint current i clamped to a negative level on , and obtain the expressions of i op and i on as follows: The i in the above formula x,y,z is determined by the large sector BS and the three-phase inductor current i La,b,c as follows:

2. The midpoint potential control method for a neutral point clamped three-level inverter according to claim 1, characterized in that, In step three, the selection method of the scaling factor k is as follows:

3. A neutral point potential control method applied to a neutral point clamped three-level inverter according to claim 1, characterized in that, In step 4, according to the voltage reference vector, the obtained clamping mode factor k cp and the scaling factor k, determine a new small sector division, obtain the corresponding switching vectors and action times, and adopt a symmetric sequence arrangement of V1-V2-V3-V4-V3-V2-V1, with the action time arrangement of T1 / 2-T2 / 2-T3 / 2-T4-T3 / 2-T2 / 2-T1 / 2, so as to obtain the driving signals of the three-phase bridge arms; Calculate the action time of each switching vector according to the volt-second balance principle: V1T1 + V2T2 + V3T3 + V4T4 = V ref T s where T s is the switching period, T1 * , T2 * , T3 * , T4 * are the per-unit action times of each switching vector.

Images