Fault-tolerant control method for t-type three-level rectifier based on dpwm modulation

Abstract

The application discloses a T-type three-level rectifier fault-tolerant control method based on DPWM modulation, establishes a mathematical model of a midpoint current of a T-type three-level rectifier, divides one fundamental wave period into four regions, and analyzes fault regions when inner and outer switch tubes are faulty; the midpoint current corresponding to each clamping state is calculated; a prediction model of the midpoint current is obtained by state space modeling of the T-type three-level rectifier, the square of the difference between the midpoint current in the prediction model and an actual midpoint current reference value is taken as an objective function, the clamping state is selected by the next moment predicted current, and the T-type three-level rectifier is subjected to fault-tolerant control by selecting a suitable carrier type and direction according to the clamping state of the switch tube open circuit fault. The application improves the working efficiency of the converter, reduces the switching times, reduces the switching loss, effectively reduces the calculation amount, improves the dynamic response speed of the system, and has good practical significance.

Description

Technical Field

[0001] This invention belongs to the field of fault control technology for power electronic converters, specifically relating to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation. Background Technology

[0002] The development and utilization of new energy sources has become a hot industry, and power converters, as key components of new energy power generation systems, have developed rapidly. T-type three-level converters, with their advantages of low device stress, low output voltage harmonic content, and high efficiency, are widely used in the field of new energy power generation. When power devices in a rectifier operate for extended periods, they may experience short-circuit or open-circuit faults. Open-circuit faults, unlike short-circuit faults which produce a sudden surge of large current, cause distortion of the output current waveform, affecting the safe and stable operation of the converter.

[0003] However, current fault-tolerant control technology for T-type three-level rectifiers still faces shortcomings and challenges: exacerbated midpoint voltage imbalance, including DC deviation and oscillation, leads to increased stress on switching devices and increased harmonic distortion of the current waveform; power losses in switching devices result in low efficiency, increasing the size and cost of heat sinks. Therefore, research on fault-tolerant control strategies that balance reducing switching losses with achieving midpoint voltage balance is particularly important. Summary of the Invention

[0004] The purpose of this invention is to provide a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, which solves the problem of reduced efficiency of the T-type three-level converter due to DC deviation and oscillation of the midpoint voltage in the current fault-tolerant control technology for T-type three-level rectifiers.

[0005] The technical solution adopted in this invention is:

[0006] A fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation is proposed. S1: Establish a mathematical model of the midpoint current of the T-type three-level rectifier. Under the DPWM modulation strategy, based on the reference voltage and current amplitude of the faulty phase, divide one fundamental cycle into four regions and analyze the fault regions when the internal and external switches fail respectively. S2: Calculate the midpoint current corresponding to each clamping state. S3: Obtain a prediction model of the midpoint current by performing state-space modeling on the T-type three-level rectifier. Input the sampled midpoint current reference value into the prediction model. Use the square of the difference between the midpoint current in the prediction model and the actual midpoint current reference value as the objective function. Select the clamping state based on the predicted current at the next moment. S4: Select an appropriate carrier type and direction to generate PWM for fault-tolerant control of the T-type three-level rectifier based on the clamping state of the open-circuit fault of the switch.

[0007] The invention is further characterized by:

[0008] S1 specifically refers to:

[0009] Based on the input and output of the rectifier, establish the relevant model of the rectifier system;

[0010] In a T-type three-level rectifier, with V dc When / 2 is the reference voltage, the three-phase sinusoidal modulation wave is:

[0011]

[0012] Where m is the modulation index, f s The fundamental frequency of the AC side voltage;

[0013] To improve the utilization rate of DC bus voltage, the three-phase reference voltage in equation (1) is adjusted as follows, and the correction value is shown in the following equation:

[0014]

[0015] Among them, V zvs For the added zero-sequence voltage component, u min u max These are the maximum and minimum values ​​of the three-phase reference voltage in equation (1), respectively;

[0016] Based on the correction value of the fault phase reference voltage and the current amplitude, a DPWM fundamental cycle is divided into four regions: A, B, C, and D. The size of regions B and D depends on the filter impedance and the rectifier power factor. When the external switching transistor S... X1 When an open-circuit fault occurs, since region B contains the affected current, the fault region is region B. When the external switch S... X2 When an open-circuit fault occurs, region D contains fault current, so region D is a fault region at this time. When the fault current is less than zero and the rectifier is operating at the midpoint, if the switching transistor S... X2 When an open-circuit fault occurs, regions B and C containing the fault current are considered the fault regions; similarly, when the switching transistor S... X3 When an open circuit fault occurs, areas A and D are the fault areas.

[0017] S2 specifically refers to:

[0018] To handle open-circuit faults in the switching transistor, fault-tolerant control is first required within the normal time period. Within this period, open-circuit faults do not affect the normal operation of the rectifier. Therefore, clamping states need to be selected to achieve midpoint voltage balance and reduce switching losses. This requires calculating the midpoint current corresponding to each clamping state, which can be expressed as:

[0019]

[0020] Among them, u mo Indicates the reference voltage corresponding to different clamping states; i mo For u mo The corresponding phase current;

[0021] When the internal switching transistor fails within the abnormal time period, the corresponding clamping state and its corresponding midpoint current will vary depending on the operating conditions.

[0022] When V a =V max hour;

[0023] i NP =[1-sign(u mid )u mid i mid +[1-sign(u min )u min i min (4);

[0024] When V a =V mid hour;

[0025] i NP =[1-sign(u max )u max i max +[1-sign(u min )u min i min (5);

[0026] When V a =V mid hour;

[0027] i NP =[1-sign(u max )u max i max +[1-sign(u mid )u mid i mid (6);

[0028] Among them, V a For phase A sinusoidal modulation wave, V max V mid V min These correspond to the maximum, median, and minimum values ​​of the three-phase reference voltage after the injection of the zero-sequence voltage component, i.e., the corrected values ​​of the three-phase reference voltage. max i midi min Then they are respectively V max V mid V min The corresponding current.

[0029] S3 specifically refers to:

[0030] The reference voltages obtained after zero-sequence voltage injection are sorted, and then the midpoint current corresponding to the clamping state is calculated.

[0031] A prediction model for the midpoint current is obtained by performing state-space modeling on the three-level rectifier, and the sampled reference value of the midpoint voltage is input into the prediction model.

[0032] The objective function is the square of the difference between the midpoint current in the prediction model and the actual reference value of the midpoint current.

[0033] The clamping state is selected by predicting the current at the next moment, and the appropriate carrier type and direction are selected according to the open circuit fault of the switching transistor to generate PWM for fault-tolerant control of the T-type three-level rectifier.

[0034] We first need to obtain the discrete-time mathematical model of the two capacitor voltages on the DC side of the T-type three-level rectifier, that is, the predicted values ​​of the DC side capacitor voltages at time k+1 are as follows:

[0035]

[0036]

[0037] Among them, U dc1 (k+1), U dc2 (k+1) represent the predicted k+1 voltage values ​​for capacitors C1 and C2, respectively. dc1 (k), U dc2 (k) represents the predicted voltage value at time k, i dc (k) represents the predicted bus current at time k, T S This refers to the switching frequency.

[0038] When the error range is small, fuzzy control is used to compensate for the small threshold error caused by the model's inability to accurately track the reference voltage. Based on the purpose of eliminating errors in fuzzy input variables and fuzzy outputs, fuzzy control rules need to be constructed to meet the effective control of the fuzzy controller. After obtaining the fuzzy output, the fuzzy control output needs to be defuzzified. The area centroid method shown in the following formula (9) is selected for defuzzification to obtain U. dc1 with U dc2 The precise value of the fuzzy control output is then input into the objective function for compensation.

[0039]

[0040] After compensation, in order to achieve grid-side current tracking control and DC-side capacitor voltage balance, the objective function is defined as follows:

[0041]

[0042] Where i max (k+1),i mid (k+1),i min (k+1) represents the actual value of the current sample. To predict the estimated value, λ dc The coefficient corresponding to the voltage deviation between the upper and lower capacitors is used to ensure that the voltages of the upper and lower capacitors on the DC side remain balanced. Δv(k+1) is the error between the predicted values ​​of the DC side capacitor voltage at time k+1. o This is the compensation value;

[0043] After the prediction model is established, the clamping state is selected by calculating the predicted current value.

[0044] S4 specifically refers to:

[0045] The carrier type and direction are selected based on the predicted clamping state;

[0046] Within the normal time range;

[0047] When V max >0.5, at this time u max The corresponding phase is achieved by using a convex in-phase carrier.

[0048] When u mid When u > 0, the carrier type adopts a concave in-phase carrier. mid When ≤0, the carrier type adopts a convex in-phase carrier;

[0049] When V max When the reference voltage amplitude is less than 0.5, the carrier type adopts a convex in-phase carrier; while when the reference voltage amplitude is greater than 0, the carrier type adopts a concave in-phase carrier.

[0050] Within the abnormal time period;

[0051] When the external switch fails, the selection of the carrier type and direction corresponds to that in the normal time region.

[0052] When the internal switching transistor fails, the carrier type and direction are as follows:

[0053] The selection rules for phases A, B, and C are the same; the following uses phase A as an example.

[0054] When V max>0.5;

[0055] V a,ref =V max At that time, phase A uses a convex single carrier.

[0056] V a,ref =V mid At that time, phase A uses a concave single carrier.

[0057] V a,ref =V min At that time, phase A uses a concave single carrier.

[0058] When V max When ≤0.5;

[0059] V a,ref =V max At that time, phase A uses a convex single carrier.

[0060] V a,ref =V mid At that time, phase A uses a concave single carrier.

[0061] V a,ref =V min At that time, phase A uses a concave single carrier.

[0062] The beneficial effects of this invention are as follows: This invention provides a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation. This method employs a DPWM (Discontinuous Pulse Width Modulation) strategy, superimposing a specific zero-sequence voltage component into the three-phase reference voltage, so that the adjusted reference voltage is located at the carrier boundary value, thereby reducing the number of switching operations and lowering switching losses. By combining model predictive control and fault-tolerant control strategies, the midpoint current value at the next moment can be predicted, and the objective function can be used for optimal solution, effectively reducing the amount of computation and improving the dynamic response speed of the system; it has good practical significance. Attached Figure Description

[0063] Figure 1 This is a topology diagram of the T-type three-level rectifier in the fault-tolerant control method of the T-type three-level rectifier based on DPWM modulation of the present invention;

[0064] Figure 2 This invention relates to the fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, which uses a DPWM modulation strategy to divide the fault region under an open-circuit fault of the switching transistor.

[0065] Figure 3 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, within the normal time region V. max When >0.5, u max Corresponding phase carrier type and direction;

[0066] Figure 4 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, within the normal time region V. max When ≤0.5, u max Corresponding phase carrier type and direction;

[0067] Figure 5 In the fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation of this invention, u mid Corresponding phase carrier type and direction;

[0068] Figure 6 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, within the normal time region V. max When >0.5, u min Corresponding phase carrier type and direction;

[0069] Figure 7 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, within the normal time region V. max When ≤0.5, u min Corresponding phase carrier type and direction;

[0070] Figure 8 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, which addresses the issue of u within abnormal time regions. max Corresponding phase carrier type and direction;

[0071] Figure 9 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, which addresses the issue of u within abnormal time regions. mid Corresponding phase carrier type and direction;

[0072] Figure 10 This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, which addresses the issue of u within abnormal time regions. min Corresponding phase carrier type and direction;

[0073] Figure 11 This is a block diagram of the model prediction fault-tolerant control strategy based on DPWM modulation in the fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation in this invention.

[0074] Figure 12 This is a schematic diagram of the overall process of the fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation according to the present invention;

[0075] Figure 13 This is an improved MPC control flowchart in the fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, as described in this invention. Detailed Implementation

[0076] The fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation according to the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0077] The following detailed description of the fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation according to the present invention will be provided through specific embodiments.

[0078] This invention proposes a fault-tolerant control method for a T-type three-level rectifier that simultaneously considers midpoint voltage balance and reduces switching losses. This method employs a DPWM modulation strategy, which superimposes a specific zero-sequence voltage component into the three-phase reference voltage, so that the adjusted reference voltage is located at the carrier boundary value, thereby reducing the number of switching operations and lowering switching losses.

[0079] In a T-type three-level rectifier, the switching action of the switching transistor is determined by the magnitude relationship between the modulating wave and the carrier wave. When the modulating wave is greater than the carrier wave, the switching transistor S... X1 and S X2 On, switch S X3 and S X4 When the phase is turned off, the corresponding output state is P; when the modulation wave is less than the upcarrier wave and greater than the downcarrier wave, the switch S... X2 and S X3 On, switch S X1 and S X4 When the circuit is turned off, the corresponding phase output state is 0: when the modulation wave is smaller than the download wave, the switching transistor S... X3 and S X4 On, switch S X1 and S X2 When the circuit is turned off, the corresponding phase output state is N. The generation of 0.5V is represented by P, O, and N respectively. dc 0, -0.5V dc The output voltage is referenced to zero at the connection point between the negative terminal of DC bus capacitor C1 and the positive terminal of capacitor C2. The fault-tolerant control method is as follows: First, the fundamental frequency period is divided into a normal time region and an abnormal time region. In the normal time region, a zero-sequence voltage component injection method is used to select an appropriate clamping state to adjust the midpoint voltage and reduce the number of switching operations. Simultaneously, the carrier direction is selected to reduce switching actions within one cycle. In the abnormal time region, fault-tolerant operation is first achieved using single / dual carriers, and then an appropriate zero-sequence voltage and carrier direction are selected based on the situation to achieve efficient operation. Furthermore, by combining model predictive control and fault-tolerant control strategies, the midpoint current value at the next moment can be predicted, and the objective function can be used for optimal solution, effectively reducing computational load and improving the system's dynamic response speed.

[0080] This invention relates to a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation;

[0081] Step 1: Based on the input and output of the rectifier, establish a mathematical model of the midpoint voltage of the T-type three-level rectifier. According to the required DPWM wave and the reference voltage and current direction of the faulty phase, divide one fundamental cycle into four regions and analyze the fault regions when the internal and external switching transistors fail respectively.

[0082] Step 2: Calculate the midpoint current corresponding to each clamping state;

[0083] Step 3: Obtain the prediction model of the midpoint current by performing state-space modeling on the T-type three-level rectifier, and input the sampled midpoint current reference value into the prediction model; use the square of the difference between the midpoint current in the prediction model and the actual midpoint current reference value as the objective function, and select the clamping state by predicting the current at the next moment.

[0084] Step 4: Select the appropriate carrier type and direction based on the clamping state of the open-circuit fault of the switching transistor to generate PWM for fault-tolerant control of the T-type three-level rectifier.

[0085] Example 1;

[0086] Step one specifically includes:

[0087] Based on the input and output of the rectifier, establish the relevant model of the rectifier system;

[0088] In a T-type three-level rectifier, with V dc When / 2 is the reference voltage, the three-phase sinusoidal modulation wave is:

[0089]

[0090] Where m is the modulation index, f s This is the fundamental frequency of the AC side voltage.

[0091] To improve the utilization rate of DC bus voltage, the three-phase reference voltage in equation (1) is adjusted as follows, and the correction value is shown in the following equation:

[0092]

[0093] In the formula, V zvs For the added zero-sequence voltage component, u min u max These are the maximum and minimum values ​​of the three-phase reference voltage in equation (1), respectively.

[0094] Based on the correction value of the fault phase reference voltage and the current amplitude, a DPWM fundamental cycle is divided into four regions: A, B, C, and D. The size of regions B and D depends on the filter impedance and the rectifier power factor. When the external switching transistor S... X1When an open-circuit fault occurs, since region B contains the affected current, the fault region is region B. When the external switch S... X2 When an open-circuit fault occurs, region D contains fault current, so region D is a fault region at this time. When the fault current is less than zero and the rectifier is operating at the midpoint, if the switching transistor S... X2 When an open-circuit fault occurs, regions B and C containing the fault current are considered the fault regions. Similarly, when the switching transistor S... X3 When an open circuit fault occurs, areas A and D are the fault areas.

[0095] Example 2;

[0096] In step two, to handle open-circuit faults in the switching transistor, fault-tolerant control is first required within the normal time period. Within this period, open-circuit faults do not affect the normal operation of the rectifier. Therefore, clamping states need to be selected to achieve midpoint voltage balance and reduce switching losses. This requires calculating the midpoint current corresponding to each clamping state, which can be expressed as:

[0097]

[0098] Among them, u mo Indicates the reference voltage corresponding to different clamping states; i mo For u mo The corresponding phase current.

[0099] When the internal switching transistor fails within the abnormal time period, the corresponding clamping state and its corresponding midpoint current will vary depending on the operating conditions.

[0100] When V a =V max hour;

[0101] i NP =[1-sign(u mid )u mid i mid +[1-sign(u min )u min i min (4);

[0102] When V a =V mid hour;

[0103] i NP =[1-sign(u max )u max i max +[1-sign(u min )u min imin (5);

[0104] When V a =V mid hour;

[0105] i NP =[1-sign(u max )u max i max +[1-sign(u mid )u mid i mid (6);

[0106] Among them, V a For phase A sinusoidal modulation wave, V max V mid V min These correspond to the maximum, median, and minimum values ​​of the three-phase reference voltage after the injection of the zero-sequence voltage component, i.e., the corrected values ​​of the three-phase reference voltage. max i mid i min Then they are respectively V max V mid V min The corresponding current.

[0107] Example 3;

[0108] Further, step three is specifically described as follows: (1) Sort the reference voltages obtained after zero-sequence voltage injection, and then calculate the midpoint current corresponding to the clamping state; (2) Obtain the prediction model of the midpoint current by performing state space modeling on the three-level rectifier, and input the sampled midpoint voltage reference value into the prediction model; (3) Use the square of the difference between the midpoint current in the prediction model and the actual midpoint current reference value as the objective function; (4) Select the clamping state by predicting the current at the next moment, and then select the appropriate carrier type and direction according to the open circuit fault of the switching tube to generate PWM for fault-tolerant control of the T-type three-level rectifier.

[0109] Therefore, we first need to obtain the discrete-time mathematical model of the voltages of the two capacitors on the DC side of the T-type three-level rectifier, that is, the predicted values ​​of the DC side capacitor voltages at time k+1 are as follows:

[0110]

[0111]

[0112] Among them, U dc1 (k+1), U dc2 (k+1) represent the predicted k+1 voltage values ​​for capacitors C1 and C2, respectively. dc1 (k), Udc2 (k) represents the predicted voltage value at time k, i dc (k) represents the predicted bus current at time k, T S This refers to the switching frequency.

[0113] When the error range is small, fuzzy control is used to compensate for the small threshold error caused by the model's inability to accurately track the reference voltage. Based on the purpose of eliminating errors in fuzzy input variables and fuzzy outputs, fuzzy control rules need to be constructed to meet the effective control of the fuzzy controller. After obtaining the fuzzy output, the fuzzy control output needs to be defuzzified. This invention selects the area centroid method as shown in equation (9) for defuzzification to obtain U. dc1 with U dc2 The precise value of the fuzzy control output is then input into the objective function for compensation.

[0114]

[0115] After compensation, in order to achieve the above two control objectives: grid-side current tracking control and DC-side capacitor voltage balance, the objective function is defined as follows:

[0116]

[0117] Among them, i max (k+1),i mid (k+1),i min (k+1) represents the actual value of the current sample. To predict the estimated value, λ dc The coefficient corresponding to the voltage deviation between the upper and lower capacitors is used to ensure that the voltages of the upper and lower capacitors on the DC side remain balanced. Δv(k+1) is the error between the predicted values ​​of the DC side capacitor voltage at time k+1. o This is the compensation value. After establishing the prediction model, the calculated predicted current value can be used to select a suitable clamping state, which can reduce switching losses and provide a basis for selecting the carrier type and direction.

[0118] Example 4;

[0119] Step four can be further explained as selecting the carrier type and direction based on the predicted clamping state. Specifically, within the normal time range, when V... max >0.5, at this time u max The corresponding phase will contain two states, P and O. Therefore, when the switching sequence in the corresponding phase begins and ends in the P state, no additional switching action will occur during the adjacent switching cycles of that phase. This control objective can be achieved by using a convex in-phase carrier. Similarly, when u midWhen the corresponding phase is in different clamping states, it contains a zero state. Therefore, if the phase switching sequence begins and ends with a zero state, switching actions between adjacent switching cycles will be effectively avoided. Specifically, when u mid When u > 0, a concave in-phase carrier is used. mid When u ≤ 0, a convex in-phase carrier is used; similarly, regardless of the clamping mode used, u min Each phase contains an N state. If the phase switching sequence begins and ends with an N state, no additional switching action will occur during adjacent switching cycles. To achieve this control objective, this phase must use a concave in-phase carrier. When V max When u ≤ 0.5, there are three clamping states, and regardless of the clamping state used, u max u mid and u min Each phase contains an 0 state. Therefore, if the phase switching sequence starts and ends with an 0 state, no additional switching action will occur during adjacent cycles. Specifically, when the reference voltage amplitude is less than 0, a convex in-phase carrier is used; and when the reference voltage amplitude is greater than 0, a concave in-phase carrier is used.

[0120] Within the abnormal time period, when the external switch fails, the selection of the carrier type and direction is similar to that within the normal time period. However, when the internal switch fails, the carrier type and direction are specifically determined as follows: when V... max >0.5 and V a,ref =V max When the A-phase switching sequence begins and ends in state P, additional switching actions during the switching cycle can be eliminated. Therefore, A-phase uses a convex single-carrier configuration. a,ref =V mid Regardless of the clamping state used, phase A will always exist in two states: P and N. In this case, to eliminate switching actions between adjacent switching cycles, phase A can use a convex single-carrier configuration; V a,ref =V min During the abnormal time period, there are three potential clamping states. If phase A uses a concave single carrier, the switching sequence of phase A will start and end in state N. In this case, no additional switching action will occur during adjacent switching cycles in phase A. When V max ≤0.5 and V a,ref =V max If phase A uses a convex single-carrier configuration, the phase A switching sequence will begin and end in state P, and no additional switching action will occur during adjacent switching cycles of phase A. When any of these four potential clamping states are used in the phase corresponding to the reference voltage, state O will appear in all of these phases. Therefore, if the switching sequence of the phase corresponding to the reference voltage begins and ends in state O, no additional switching action will occur during adjacent switching cycles of this phase; V a,ref =V midAt that time, phase A uses a concave single carrier; V a,ref =V min At the same time, using a concave single carrier can effectively reduce the on-off loss.

[0121] This invention addresses the fault-tolerant control of open-circuit faults in power devices of a T-type three-level rectifier during prolonged operation. It ensures undistorted output current and reduced harmonic content under open-circuit fault conditions, thereby improving the converter's efficiency. The circuit topology is shown below. Figure 1 As shown.

[0122] In performing fault-tolerant control, this invention first divides the fault area, such as... Figure 2 As shown, V x,ref,off For the corrected voltage reference value, i x Based on the reference voltage and current direction, a fundamental frequency cycle is divided into four regions, named regions A, B, C, and D. Additionally, depending on whether the output voltage is affected, each fundamental frequency cycle can be divided into two parts: a normal time region and an abnormal time region. In the normal time region, a switching transistor fault does not affect the normal operation of the rectifier, while in the abnormal time region, a switching transistor fault will affect the normal operation of the rectifier. Based on the impact of the faulty device on the current and the fact that regions A, B, C, and D are included in the fault region... Figure 2 It is shown in the middle.

[0123] Example 5;

[0124] For the selection of carrier direction and type in fault-tolerant control within the normal time range, when V max >0.5, at this time u max When the corresponding phase switching sequence begins and ends in state P, no additional switching action will occur during adjacent switching cycles of that phase. This control objective can be achieved by using a convex in-phase carrier, such as... Figure 3 As shown; similarly, when u mid If the phase switching sequence begins and ends with a 0 state, then switching actions between adjacent switching cycles will be effectively avoided. In this case, the principle for selecting the carrier type is as follows: Figure 5 As shown. Similarly, u min Each phase contains an N-state, and the phase switching sequence begins and ends with an N-state. Therefore, no additional switching action will occur during adjacent switching cycles. To achieve this control objective, this phase must use a concave in-phase carrier, such as... Figure 6 As shown. When V max When u ≤ 0.5, there are three clamping states, and regardless of the clamping state used, u max u mid and u minIf each phase contains a zero state, and the switching sequence begins and ends with a zero state, then no additional switching action will occur between adjacent cycles. Figure 4 , 5 7 and 8 are u max u mid and u min A schematic diagram illustrating the principle of selecting the corresponding phase carrier type.

[0125] In the abnormal time region, when the external switch fails, its carrier direction and type are selected the same as in the normal time region; when the internal switch fails, the faulty phase first adopts two-level carrier modulation technology, which effectively avoids the use of fault state O, and first realizes the fault-tolerant operation of the converter. Then, according to the comparison between the A-phase reference voltage and the single carrier, when the modulated wave is greater than the carrier, the output state P is output, and when the modulated wave is less than the carrier, the output state N is output. Figure 8 , 9 10 and 10 are u max u mid and u min A schematic diagram showing the selection of corresponding phase carrier type and direction.

[0126] Based on the above, the overall control flow diagram of this invention is as follows: Figure 11 As shown, its main content is as follows: the obtained DC bus voltage value is transformed to obtain the current value under the dq coordinate axis, then PI-regulated and inversely transformed to obtain the three-phase reference voltage. After injecting zero-sequence voltage components into the three-phase reference voltage, the reference voltage is sorted. A prediction model of the midpoint current is obtained by performing state-space modeling on the three-level rectifier. The sampled midpoint voltage reference value and current value are input into the prediction model. The square of the difference between the midpoint current in the prediction model and the actual midpoint current reference value is used as the cost function. The clamping state is selected by predicting the current at the next moment. Then, according to the open circuit fault of the switching transistor, an appropriate carrier type and direction are selected to generate PWM for fault-tolerant control of the T-type three-level rectifier.

[0127] The fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, as described in the present invention, has the following overall flowchart: Figure 12 As shown, upon receiving a PWM interrupt signal, the ADC is first triggered to sample the voltage and current, obtaining the three-phase voltage and current feedback data of the power grid. The calculated zero-sequence voltage component is then injected into the three-phase reference voltage, which is then sorted. Subsequently, deadbeat control is performed on the inner current loop based on the sampled current values. The obtained values ​​are input into the MPC control program for midpoint current calculation. Finally, the calculated current value is output to the PWM module to control the switching transistors. After this cycle is completed, the system waits for the next cycle interrupt signal to arrive, thus achieving continuous control. Figure 13To improve the MPC control flowchart, the ADC sampling results are used to enter fuzzy control, calculate the AC / DC deviation compensation value, and add it to the cost function of the improved MPC. The improved MPC uses the cost function to traverse and optimize to find the clamping state corresponding to the optimal midpoint current. Finally, the value is output to the PWM module to control the switching transistor.

[0128] This invention presents a fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation. This method ensures that the output current of the T-type three-level rectifier remains undistorted and harmonic content is reduced under open-circuit fault conditions, thereby improving the converter's operating efficiency under fault conditions. It can also reduce the number of switching operations and lower switching losses. By combining model predictive control strategies and fault-tolerant control methods, and using the objective function for optimal solution, it effectively reduces computational load and improves the system's dynamic response speed. This method has significant practical value.

Claims

1. A fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation, characterized in that, S1: Establish a mathematical model of the midpoint current of the T-type three-level rectifier. Under the DPWM modulation strategy, divide one fundamental cycle into four regions based on the reference voltage and current amplitude of the faulty phase, and analyze the fault regions when the internal and external switches fail respectively. S2: Calculate the midpoint current corresponding to each clamping state. S3: Obtain a prediction model of the midpoint current by performing state-space modeling on the T-type three-level rectifier. Input the sampled midpoint current reference value into the prediction model. Use the square of the difference between the midpoint current in the prediction model and the actual midpoint current reference value as the objective function. Select the clamping state based on the predicted current at the next moment. S4: Select an appropriate carrier type and direction to generate PWM for fault-tolerant control of the T-type three-level rectifier based on the clamping state of the open-circuit fault of the switch. S2 specifically refers to: To handle open-circuit faults in the switching transistor, fault-tolerant control is first required within the normal time period, during which an open-circuit fault will not affect the normal operation of the rectifier. Selecting clamping states to achieve midpoint voltage balance and reduce switching losses requires calculating the midpoint current corresponding to each clamping state. The midpoint current corresponding to each clamping state can be expressed as: （3）； in, u mo This indicates the reference voltage corresponding to different clamping states; i mo for u mo The corresponding phase current; When the internal switching transistor fails within the abnormal time period, the corresponding clamping state and its corresponding midpoint current will vary depending on the operating conditions. when hour; （4）； when hour; （5）； when hour; （6）； in, V a This is a sinusoidal modulation wave of phase A. V max , V mid , V min These correspond to the maximum, median, and minimum values ​​of the three-phase reference voltage after the injection of the zero-sequence voltage component, i.e., the corrected values ​​of the three-phase reference voltage. i max , i mid , i min They are respectively V max , V mid , V min The corresponding current; Step 4 specifically involves: The carrier type and direction are selected based on the predicted clamping state; Within the normal time range; when V max >0.5, at this time u max The corresponding phase is achieved by using a convex in-phase carrier. when u mid When >0, the carrier type adopts a concave in-phase carrier. u mid When ≤0, the carrier type adopts a convex in-phase carrier; when V max When the reference voltage amplitude is less than 0.5, the carrier type adopts a convex in-phase carrier; while when the reference voltage amplitude is greater than 0, the carrier type adopts a concave in-phase carrier. Within the abnormal time period; When the external switch fails, the selection of the carrier type and direction corresponds to that in the normal time region. When the internal switching transistor fails, the carrier type and direction are as follows: The selection rules for phases A, B, and C are the same; the following uses phase A as an example. when V max >0.5; V a,ref = V max At that time, phase A uses a convex single carrier. V a,ref = V mid At that time, phase A uses a concave single carrier. V a,ref = V min At that time, phase A uses a concave single carrier. when V max When ≤0.5; V a,ref = V max At that time, phase A uses a convex single carrier. V a,ref = V mid At that time, phase A uses a concave single carrier. V a,ref = V min At that time, phase A uses a concave single carrier.

2. The fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation according to claim 1, characterized in that, S1 specifically refers to: Based on the input and output of the rectifier, establish the relevant model of the rectifier system; In a T-type three-level rectifier, with V dc When / 2 is the reference voltage, the three-phase sinusoidal modulation wave is: （1）； in, m In order to adjust the system, f s The fundamental frequency of the AC side voltage; To improve the utilization rate of DC bus voltage, the three-phase reference voltage in equation (1) is adjusted as follows, and the correction value is shown in the following equation: （2）； in, V zvs For the added zero-sequence voltage component, u min , u max These are the maximum and minimum values ​​of the three-phase reference voltage in equation (1), respectively; Based on the correction value of the fault phase reference voltage and the current amplitude, a DPWM fundamental cycle is divided into four regions: A, B, C, and D. The size of regions B and D depends on the filter impedance and the rectifier power factor. When the external switching transistor... S X1 When an open-circuit fault occurs, since region B contains the affected current, the fault region is region B. When the external switching transistor... S X2 When an open-circuit fault occurs, region D contains fault current, so region D is a fault region at this time. When the fault current is less than zero and the rectifier is operating at the midpoint, if the switching transistor... S X2 When an open-circuit fault occurs, regions B and C containing the fault current are considered the fault regions; similarly, when the switching transistor... S X3 When an open circuit fault occurs, areas A and D are the fault areas.

3. The fault-tolerant control method for a T-type three-level rectifier based on DPWM modulation according to claim 1, characterized in that, S3 specifically refers to: The reference voltages obtained after zero-sequence voltage injection are sorted, and then the midpoint current corresponding to the clamping state is calculated. A prediction model for the midpoint current is obtained by performing state-space modeling on the three-level rectifier, and the sampled reference value of the midpoint voltage is input into the prediction model. The objective function is the square of the difference between the midpoint current in the prediction model and the actual reference value of the midpoint current. The clamping state is selected by predicting the current at the next moment, and the appropriate carrier type and direction are selected according to the open circuit fault of the switching transistor to generate PWM for fault-tolerant control of the T-type three-level rectifier. We first need to obtain the discrete-time mathematical model of the two capacitor voltages on the DC side of the T-type three-level rectifier, that is, the predicted values ​​of the DC side capacitor voltages at time k+1 are as follows: （7）； （8）； in, U dc1 (k+1), U dc2 (k+1) represents the predicted voltage values ​​at k+1 for capacitors C1 and C2, respectively. U dc1 (k), U dc2 (k) represents the predicted voltage value at time k. i dc1 (k) represents the predicted value of the bus current at time k. T S The switching frequency; When the error range is small, fuzzy control is used to compensate for the small range error caused by the model's inability to accurately track the reference voltage. Based on the purpose of eliminating errors in fuzzy input variables and fuzzy outputs, fuzzy control rules need to be constructed to meet the effective control of the fuzzy controller. After obtaining the fuzzy output, the fuzzy control output needs to be defuzzified. The area centroid method shown in the following formula (9) is selected for defuzzification to obtain the result. U dc1 and U dc2 The precise value of the fuzzy control output is then input into the objective function for compensation. （9）； After compensation, in order to achieve grid-side current tracking control and DC-side capacitor voltage balance, the objective function is defined as follows: （10）； in , , This is the actual value of the current sample. , , To predict the estimated value, This is a coefficient corresponding to the voltage deviation between the upper and lower capacitors, used to ensure that the voltages of the upper and lower capacitors on the DC side remain balanced. The error between the predicted DC-side capacitor voltage at time k+1 is... This is the compensation value; After the prediction model is established, the clamping state is selected by calculating the predicted current value.

Images