Open-circuit fault diagnosis method for switching tube of dual-active bridge converter

Abstract

The invention discloses an open-circuit fault diagnosis method for a switching tube of a dual-active bridge converter. The open-circuit fault diagnosis method comprises the following specific steps: modeling the dual-active bridge converter; analyzing voltage and inductance waveform characteristics of the dual-active bridge converter in a fault state of each switching tube on the primary side and the secondary side; comparing and analyzing fault features of each switching tube, and selecting a difference value between an inductive current average value and a bridge arm midpoint voltage average value in a fault state and a normal state as a fault feature; and designing a fault diagnosis algorithm according to the selected fault features. Compared with a traditional fault diagnosis method, only the voltage of one bridge arm of the primary side and the voltage of one bridge arm of the secondary side are selected as fault features, fewer sensors are needed, and real-time online fault diagnosis is more accurate and rapid.

Description

Technical Field

[0001] The present invention relates to the fault diagnosis technology of power electronic converters, and particularly to an open-circuit fault diagnosis method for switching tubes of a dual-active-bridge converter. Background Art

[0002] Due to advantages such as high power density, high efficiency, and bidirectional power flow ability, the dual-active-bridge converter is widely used in fields such as electric vehicle charging, renewable energy energy storage, data center power supply, and smart grid. However, the dual-active-bridge converter usually consists of multiple power switching tubes, and during long-term operation, open-circuit faults of the switching tubes may occur due to overcurrent, overheating, or abnormal drive signals. If not detected in time, it may lead to out-of-control output voltage, increased current stress, and even system collapse, seriously affecting the safety and stability of the equipment.

[0003] Currently, three main types of methods are mainly used for fault diagnosis of DAB converters: The first type is the direct detection method, which judges faults by monitoring the current / voltage changes of the switching tubes. Although the principle is simple, it requires additional installation of sensors, increasing the system cost and complexity; the second type is the method based on a model observer, which predicts electrical parameters by establishing a mathematical model and compares them with the measured values, but has extremely high requirements for model accuracy and a large amount of calculation; the third type is the artificial intelligence method, which has a relatively high recognition accuracy, but requires a large amount of training data and poor real-time performance. These existing methods generally have three main defects: one is the relatively high hardware cost, and traditional solutions need to monitor the electrical signals of all switching tubes; the second is the slow diagnostic response speed, especially complex algorithms are difficult to meet the real-time requirements; the third is the weak anti-interference ability, and misjudgment is likely to occur under dynamic load or input voltage fluctuations. These limitations severely restrict the practical application effect of the DAB converter fault diagnosis technology.

[0004] In view of the above technical problems, the present invention innovatively proposes a fault diagnosis method based on the waveform characteristics of the midpoint voltage of the bridge arm and the inductor current. By deeply analyzing the distortion characteristics of the midpoint voltage of the bridge arm and the offset law of the inductor current under different fault states of the switching tubes, an accurate fault feature database is established. During specific implementation, only the voltage signals of one bridge arm on the primary side and the secondary side need to be monitored, and combined with the change trend of the average value of the inductor current, the faulty switching tube can be quickly and accurately located. Compared with the prior art, the advantages of the present invention are that the number of sensors used is significantly reduced, the system cost and complexity are greatly reduced; the diagnostic speed is improved, and fast diagnosis is achieved through an optimized feature extraction algorithm and a lightweight diagnostic logic; the anti-interference ability is enhanced, and a multi-feature fusion diagnostic strategy is adopted to effectively avoid misjudgment problems that may be caused by single-signal detection. This method is not only applicable to the standard DAB topology, but also can be extended to derivative converters such as CLLC, and has broad application prospects and important engineering value. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for diagnosing open-circuit faults of switching tubes in a dual-active-bridge converter based on multi-feature fusion, aiming at the problems of strong sensor dependence, poor diagnostic real-time performance, and weak anti-interference ability in the existing fault diagnosis methods for dual-active-bridge converters. This method extracts the time-domain features of the midpoint voltage of the bridge arm and the inductor current, establishes a three-level diagnostic logic, and effectively fuses the feature information of different fault signals. It not only retains the timing characteristics of the original signal but also synthesizes the correlation features of multiple parameters, facilitating fast and accurate fault location and significantly improving the real-time performance and reliability of fault diagnosis.

[0006] The technical solution adopted by the present invention to solve its technical problems is as follows:

[0007] The present invention provides a method for diagnosing open-circuit faults in a dual-active-bridge converter, which includes the following steps:

[0008] Step 1: Model the dual-active-bridge converter;

[0009] Step 2: Analyze the waveform characteristics of the midpoint voltage of the bridge arm and the inductor current under the fault states of each switching tube on the primary and secondary sides of the dual-active-bridge converter, and summarize the waveform characteristics of the inductor current and the midpoint voltage of the bridge arm under the faults of each switching tube;

[0010] Step 3: Through comparative analysis of the fault characteristics of each switching tube, select the difference between the average value of the inductor current and the average value of the midpoint voltage of the bridge arm in the fault state and the normal state as the fault feature;

[0011] Step 4: Design a fault diagnosis algorithm based on the selected fault feature.

[0012] Preferably, the dual-active-bridge converter in Step 1 is:

[0013] The circuit structure of the dual-active-bridge converter includes two full-bridge modules H1 and H2 and a high-frequency isolation transformer T. This topology uses eight MOSFET power tubes T1-T8 as switching devices, and each switching tube is shunted with a freewheeling diode D1--D8. The circuit is provided with an input port U1 and an output port U2, corresponding to an input current I1 and an output current I2. In the circuit topology, A, B and C, D respectively represent the midpoints of the primary and secondary full-bridge arms, and O, P are the reference ground nodes on both sides. The AC voltages generated between the midpoints of the bridge arms are respectively denoted as u AB and u CD , to suppress voltage spikes, the circuit is configured with absorption capacitors C1 and C2 on the input and output sides. The inductor L is composed of the leakage inductance of the transformer and an externally added phase-shifting inductor, and the voltage and current across its two ends are respectively represented as U L and i L, the turns ratio of the transformer is \(n = N1 / N2\), and the circuit operating parameters include: the switching frequency \(f\) S 、the switching period \(T\) and its half period \(T\) h , the voltage transfer ratio \(K\) is defined as the ratio of the input voltage to the converted output voltage, that is, \(K = U1 / (nU2)\), and usually \(K\geq1\) is required to ensure normal operation.

[0014] Preferably, the content of the analysis of the waveform characteristics of the midpoint voltage of the bridge arm in the fault state in step 2 is:

[0015] When the switching tubes T1 or T4 have an open - circuit fault, the waveform characteristics of the full - bridge output voltage \(u\) AB 、\(u\) CD and the inductor current \(i\) L show consistency; however, in the case of a single fault of T1, the midpoint voltage \(u\) of the bridge arm AO shows waveform distortion characteristics during the fault transient period. When T2 and T3 are faulty, only the fault of T2 will cause \(u\) AO to be distorted during the transient stage. Similarly, in the fault condition on the secondary side, only the faults of T5 and T6 will cause distortion of the midpoint voltage \(u\) of the bridge arm CP during the transient process. Based on this characteristic difference, the midpoint voltages \(u\) of the bridge arm AO and \(u\) CP can be used as the discriminant basis for distinguishing the faults of the switching tubes at the diagonal positions. Specifically, when implementing, the average values of \(u\) AO and \(u\) CP , \(u\) AO_mean and \(u\) CP_mean are selected as the key fault characteristic quantities. In the normal operation state of the dual - active converter, since the midpoint voltages \(u\) of the bridge arm AO and \(u\) CP are both standard square - wave signals with a duty cycle of 50%, their average values can be calculated as follows:

[0016]

[0017] It can be seen from this that only when the switching tubes T1, T2, T5, and T6 have open - circuit faults, the average values of the midpoint voltages \(u\) of the bridge arm AO_mean and \(u\) CP_mean will change during the fault transient stage, and their change characteristics can be summarized as:

[0018] (1) When an open - circuit fault occurs in the primary - side switching tube:

[0019] When the T1 tube is faulty, \(u\) AO_mean \(<1 / 2U1\), \(u\) CP_mean \( = 1 / 2U2\);

[0020] When the T2 tube is faulty, \(u\) AO_mean \(>1 / 2U1\), \(u\)CP_mean = 1 / 2U2;

[0021] When the T3 or T4 tube fails, u AO_mean = 1 / 2U1, u CP_mean = 1 / 2U2;

[0022] (2) When an open - circuit fault occurs in the secondary - side switching tube:

[0023] When the T5 tube fails, u AO_mean = 1 / 2U1, u CP_mean <1 / 2U2;

[0024] When the T6 tube fails, u AO_mean = 1 / 2U1, u CP_mean > 1 / 2U2;

[0025] When the T7 or T8 tube fails, u AO_mean = 1 / 2U1, u CP_mean = 1 / 2U2;

[0026] (3) The fault diagnosis rule is:

[0027] When u AO_mean ≠ 1 / 2U1 and u CP_mean = 1 / 2U2, it is determined that the primary - side T1 or T2 tube fails;

[0028] When u AO_mean = 1 / 2U1 and u CP_mean ≠ 1 / 2U2, it is determined that the secondary - side T5 or T6 tube fails;

[0029] When u AO_mean = 1 / 2U1 and u CP_mean = 1 / 2U2, it is necessary to make a comprehensive judgment by combining the bias characteristics of the inductor current.

[0030] Preferably, the content summary of the bias characteristics of the inductor current i L in the fault state in step 2 is:

[0031] (1) When the system is operating normally, the inductor current i L shows no bias state, and its average value i L_mean = 0;

[0032] (2) When a single - tube open - circuit fault occurs:

[0033] ① When the primary - side switching tube fails, the inductor current forms a significant unipolar bias within 1 - 2 switching cycles, where:

[0034] When the T1 or T4 tube fails, it shows a negative bias: i L_mean <0;

[0035] When the T2 or T3 tube fails, it shows a forward bias: i L_mean > 0;

[0036] Through the positive and negative relationship of i L_mean When a primary-side fault occurs, the T1 / T4 and T2 / T3 switching tubes can be distinguished;

[0037] ② When a secondary-side switching tube fails, the inductor current needs 3 - 5 switching cycles to form a bipolar bias, where:

[0038] When the T5 or T8 tube fails, it shows a forward bias: i L_mean > 0;

[0039] When the T6 or T7 tube fails, it shows a negative bias: i L_mean < 0;

[0040] Through the positive and negative relationship of i L_mean When a secondary-side fault occurs, the T5 / T8 and T6 / T7 switching tubes can be distinguished;

[0041] (3) The bias amplitude is characterized by the absolute value of i L_mean .

[0042] Preferably, the method for processing the fault characteristics in step 3 is:

[0043] When an open-circuit fault occurs in the switching tubes on both the primary and secondary sides, it will cause a DC bias in the inductor current, but the degree of bias of the inductor current is different. Through experiments, it is known that the DC bias of the inductor current caused by an open-circuit fault on the primary side is greater than that caused by an open-circuit fault on the secondary side. Therefore, a threshold θ is set, and by setting the threshold, it is judged which side has an open-circuit fault. Specifically, it is manifested as: First, calculate the average deviation of the inductor current Δi L , that is, the difference between the fault state and the i L_mean in the normal state is used to characterize the current bias characteristics. When Δi L is zero, it indicates no bias state; when Δi L is greater than zero, it characterizes a forward bias; when Δi L is less than zero, it characterizes a negative bias. When the absolute value of Δi L exceeds the preset threshold θ, it is determined as a significant bias caused by a primary-side fault; when the absolute value of Δi L is less than θ and not zero, it is determined as a slight bias caused by a secondary-side fault, which is used to distinguish the bias degrees of different fault types. Further, since an open-circuit fault in T1 / T4 and T2 / T3 will cause the inductor current waveform to have different directions of bias, therefore, through Δi LT1 / T4 can be distinguished from T2 / T3. When open-circuit faults occur in T1 and T4, it will cause a negative bias in the inductor current waveform. Therefore, when the value of Δi L is less than -θ, it can be determined that an open-circuit fault has occurred in the primary-side switch tube T1 or T4; when open-circuit faults occur in T2 and T3, it will cause a positive bias in the inductor current waveform. Therefore, when the value of Δi L is greater than θ, it can be determined that an open-circuit fault has occurred in the primary-side T2 or T3; the same principle applies to the secondary-side push process. At the same time, by calculating the average value deviation Δu AO and Δu CP , which are the differences between the fault state and the normal state of u AO_mean and u CP_mean respectively, the specific faulty switch tube can be further located.

[0044] Preferably, the specific content of introducing a threshold value to the processed fault feature quantity in step 3 is as follows:

[0045] By setting threshold parameters to eliminate the influence of measurement errors on the diagnosis result. When the measured values of Δi L , Δu AO or Δu CP are near zero, it indicates that the corresponding waveform has not changed significantly. Considering that factors such as switching noise, parasitic parameters, and sensor errors in actual operation will cause inherent fluctuations in i L , u AO and u CP under normal operating conditions, three positive threshold parameters α, β, and γ are specifically introduced to perform tolerance processing on the measurement results of Δi L , Δu AO and Δu CP respectively, to avoid misdiagnosis caused by system inherent interference.

[0046] Preferably, the specific content of designing a fault diagnosis algorithm value according to the selected fault feature in step 4 is as follows:

[0047] By introducing the variables α, β, and γ, the specific faulty switch tube is further located. Specifically, it is manifested as: Since only when open-circuit faults occur in the switch tubes T1, T2, T5, and T6, the average values of the bridge-arm midpoint voltages u AO_mean and u CP_mean will change during the fault transient stage, and T1 / T4 and T2 / T3 will cause different-direction DC biases in the inductor current waveform. Therefore, when it is determined that an open-circuit fault has occurred in the primary-side switch tube T1 or T4 through Δi L , the specific faulty switch tube can be located through Δu AO , and when an open-circuit fault occurs on the primary side, Δu CPThe absolute values of all are less than γ, so when Δu AO is less than -β, it is determined that the T1 switch has an open - circuit fault. When Δu AO is less than β, it can be determined that the T4 switch has an open - circuit fault; when it is determined through Δi L that the open - circuit fault occurs in the primary - side T2 or T3 switch, if the value of Δu AO is greater than β, it is determined that the T2 switch has an open - circuit fault. If the value of Δu AO is less than β, it is determined that the T3 switch has an open - circuit fault; when an open - circuit fault occurs on the secondary - side, the absolute values of Δu AO are all less than β. After it is determined through Δi L that the open - circuit fault occurs in the secondary - side T5 or T8 switch, if the value of Δu CP is less than -γ, it is determined that the T5 switch has an open - circuit fault. If the value of Δu CP is less than γ, it is determined that the T8 switch has an open - circuit fault; when it is determined through Δi L that the open - circuit fault occurs in the secondary - side T6 or T7 switch, if the value of Δu CP is greater than γ, it is determined that the T5 switch has an open - circuit fault. If the value of Δu CP is less than γ, it is determined that the T8 switch has an open - circuit fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is the main circuit diagram of the dual - active - bridge converter of the present invention;

[0049] Figure 2 is the algorithm flow chart of the fault diagnosis of the present invention;

[0050] Figure 3 is the waveform comparison and diagnosis location diagram before and after the T1 switch fault of the present invention;

[0051] Figure 4 is the waveform comparison and diagnosis location diagram before and after the T4 switch fault of the present invention;

[0052] Figure 5 is the waveform comparison and diagnosis location diagram before and after the T5 switch fault of the present invention;

[0053] Figure 6 is the waveform comparison and diagnosis location diagram before and after the T8 switch fault of the present invention; DETAILED DESCRIPTION OF THE INVENTION

[0054] The following further describes an embodiment of the open - circuit fault diagnosis of the dual - active - bridge converter of the present invention with reference to the drawings:

[0055] Figure 1is the main circuit diagram of a dual-active-bridge converter. Its circuit structure includes two full-bridge modules H1 and H2 and a high-frequency isolation transformer T. This topology uses eight MOSFET power transistors T1 - T8 as switching devices, and each switching transistor is paralleled with a freewheeling diode D1 - D8. The circuit is provided with an input port U1 and an output port U2, corresponding to an input current I1 and an output current I2. In the circuit topology, A, B and C, D respectively represent the midpoints of the primary and secondary full-bridge arms, and O, P are the reference ground nodes on both sides. The AC voltages generated between the midpoints of the arms are respectively denoted as u AB and u CD . To suppress voltage spikes, the circuit is configured with snubber capacitors C1 and C2 on the input and output sides. The inductor L is composed of the leakage inductance of the transformer and an externally added phase-shifting inductor. The voltage and current across its two ends are respectively expressed as U L and i L . The turns ratio of the transformer is n = N1 / N2. The circuit operating parameters include: the switching frequency f S , the switching period T and its half-period T h . The voltage transfer ratio K is defined as the ratio of the input voltage to the converted output voltage, that is, K = U1 / (nU2), and usually K≥1 is required to ensure normal operation.

[0056] Figure 2 is the flowchart of the fault diagnosis algorithm of the present invention. Its steps are as follows: First, collect the original data i L、 u AO and u CP . By calculation, obtain the fault index quantities Δi L , Δu AO and Δu CP . Subsequently, compare the fault index quantities with the selected threshold variables θ, α, β and γ. Finally, the position of the faulty switching transistor can be located. The fault diagnosis algorithm of the present invention is verified through an example below.

[0057] Build a simulation model of a dual-active-bridge converter in the MATLAB / Simulink 2023b environment to simulate the open-circuit fault conditions of each power switching transistor. Among them, the main circuit parameters: the input voltage U1 is 200V, the output voltage U2 is 145V, the rated power P n is 400W, and the switching frequency fs is 20 kHz. The parameters of the fault diagnosis module: the sampling frequency is 5 times the switching frequency, 100 kHz. The sliding window width is 4 switching periods in the case of an open-circuit fault of the primary switching transistor and 5 switching periods in the case of an open-circuit fault of the secondary switching transistor. It can be known from experimental summaries that the diagnostic thresholds θ, α, β and γ are taken as 5A, 1A, 0.5V and 0.5V respectively.

[0058] Figure 3 and 4respectively show the three fault characteristic signals when the primary side switching transistors T1 and T4 have open - circuit faults, the mid - point voltage u AO 、u CP and the inductor current i L simulation waveforms, as well as the results given by the execution of the fault diagnosis program. The solid lines in the figure are the simulation waveforms after the fault, while the dashed lines are the simulation waveforms under non - fault conditions. An open - circuit fault occurs at time t1. From the Figure 3 simulation result figure, it can be seen that within one switching period after the fault occurs, the waveform of the mid - point voltage u AO of the primary side bridge arm has obvious distortion; after a very short transient of the inductor current i L , a complete negative bias is generated, and its peak value reaches - 10.1 A, about twice that of normal operation. From the Figure 4 simulation result figure, it can be seen that although the inductor current generates the same negative bias after the fault occurs, the waveform of the mid - point voltage u AO of the primary side bridge arm does not distort, so the faulty switching transistor of T1 or T4 can be accurately located. At the same time, after the faults of T1 and T4, since the inductor current will quickly generate an obvious DC bias, the fault can be quickly located after one - half cycle. It can be seen from the figure that the faulty switching transistor is located at time t2. In summary, the simulation results initially show that the processing method and diagnosis process of the three fault characteristic signals proposed in the present invention can quickly and accurately judge and locate the primary side open - circuit fault transistors T1 and T4, and issue a fault indication signal. Although T1 or T4 open - circuit has the same inductor current waveform characteristics, the diagnosis method can still correctly distinguish these two faults by introducing different mid - point voltage u AO characteristic criteria. The diagnosis method is equally effective for distinguishing the open - circuit faults of T2 or T3.

[0059] Figure 5 And 6 also give the relevant simulation waveforms and fault diagnosis results when the secondary side switching transistors T5 and T8 have open - circuit faults. The solid lines in the figure are the simulation waveforms after the fault, while the dashed lines are the simulation waveforms under non - fault conditions. An open - circuit fault occurs at time t1. From the Figure 5 simulation result figure, it can be seen that it takes about 3 switching periods of transient from the occurrence of the fault to entering the fault steady state. During the fault transient process, the width of the square wave of the mid - point voltage u CP of the secondary side bridge arm changes and then gradually recovers; the waveform of the inductor current i L also gradually biases positively and finally remains in a stable partial bias situation, and its peak value is about 8 A, close to but less than twice that of normal operation. From the Figure 6 simulation result figure, it can also be seen that the inductor current i LThe waveform also gradually biases positively and finally remains in a stable partial bias condition, but the midpoint voltage u of the secondary side bridge arm CP of the square wave width does not change significantly, so the faulty switch tubes of T5 or T8 can be accurately located. At the same time, after the faults of T5 and T8, since the DC bias of the inductor current is generated slowly, the fault location can be obtained after two half-cycles. It can be seen from the figure that the faulty switch tubes are located at the moment t2. Similar to the case of the primary side fault, combining Figure 5 and Figure 6 the waveforms in the case of the open circuit fault shown, it can be seen that the proposed fault diagnosis scheme can also effectively identify the open circuit faults of the secondary side switch tubes T6 and T7.

Claims

1. A method for diagnosing an open circuit fault of a switch tube of a dual active bridge converter, characterized in that: The following steps are involved: Step 1: Modeling the dual active bridge converter; Step 2, analyzing the waveform characteristics of the bridge arm midpoint voltage and the inductor current under the fault state of each switch tube on the primary side and the secondary side of the dual active bridge converter, and summarizing the waveform characteristics of the inductor current and the bridge arm midpoint voltage under the fault state of each switch tube; Step 3: By comparing and analyzing the fault characteristics of each switch tube, the difference between the average value of the inductor current and the average value of the bridge arm midpoint voltage in the fault state and the normal state is selected as the fault characteristic; Step 4: Design a fault diagnosis algorithm based on the selected fault characteristics.

2. The open circuit fault diagnosis method of the switch tube of the dual active bridge converter according to claim 1, characterized in that: The dual active bridge converter is modeled as: The circuit structure of the dual-active bridge converter includes two full-bridge modules H1 and H2 and a high-frequency isolation transformer T. The topology uses eight MOSFET power tubes T1-T8 as switching devices. Each switch tube is connected in parallel with a freewheeling diode D1--D8. The circuit has an input port U1 and an output port U2, corresponding to the input current I1 and the output current I2. In the circuit topology, A, B and C, D represent the midpoints of the bridge arms of the primary and secondary full bridges respectively, O and P are the reference ground nodes on both sides, and the AC voltage generated between the midpoints of the bridge arms is recorded as u AB and u CD To suppress voltage spikes, the circuit is equipped with absorption capacitors C1 and C2 on the input and output sides. The inductor L is composed of the transformer leakage inductance and the external phase-shifting inductor. The voltage and current at both ends are expressed as U L and i L , the transformer turns ratio is n = N1 / N2, the circuit operating parameters include: switching frequency f S , switching period T and its half period T h The voltage transfer ratio K is defined as the ratio of the input voltage to the converted output voltage, that is, K = U1 / (nU2), and K ≥ 1 is usually required to ensure normal operation.

3. The open circuit fault diagnosis method of the switch tube of the dual active bridge converter according to claim 1, characterized in that: The content of the step 2 analyzing the waveform characteristics of the midpoint voltage of the bridge arm of the dual active bridge converter under the fault state of each switch tube on the primary side and the secondary side includes: When the T1 or T4 switch fails, the full-bridge output voltage u AB 、u CD and the inductor current i L The waveform characteristics of the bridge arm are consistent; however, in the case of a single fault of T1, the bridge arm midpoint voltage u AO The waveform distortion characteristics appear during the fault transient period, and when T2 and T3 fail, only T2 failure will cause u AO In the transient stage, distortion occurs. Similarly, under the fault condition on the secondary side, only T5 and T6 faults will cause the bridge arm midpoint voltage u in the transient process. CP Based on this characteristic difference, the bridge arm midpoint voltage u AO and u CP As the basis for distinguishing the fault of the diagonal switch tube, in the specific implementation, u is selected AO and u CP The average value u AO_mean and u CP_mean As a key fault characteristic quantity, under the normal operation of the dual active converter, due to the bridge arm midpoint voltage u AO and u CP They are all standard square wave signals with a duty cycle of 50%. Their average values ​​can be calculated as follows: It can be seen from this that when only the switch tubes T1, T2, T5 and T6 have an open circuit fault, the average voltage u AO_mean and u CP_mean Changes will occur during the transient stage of the fault, and their changing characteristics can be summarized as follows: (1) When an open circuit fault occurs on the primary side switch: When T1 fails, u AO_mean <1 / 2U1,u CP_mean =1 / 2U2; When T2 tube fails, u AO_mean >1 / 2U1,u CP_mean =1 / 2U2; When T3 or T4 tube fails, u AO_mean =1 / 2U1,u CP_mean =1 / 2U2; (2) When an open circuit fault occurs on the secondary switch: When the T5 tube fails, u AO_mean =1 / 2U1,u CP_mean <1 / 2U2; When the T6 tube fails, u AO_mean =1 / 2U1,u CP_mean >1 / 2U2; When T7 or T8 tube fails, u AO_mean =1 / 2U1,u CP_mean =1 / 2U 2。 4. The method for diagnosing open circuit fault of a switch tube of a dual active bridge converter according to claim 1, characterized in that: The step 2 analyzes the inductor current i of the dual active bridge converter when each switch tube on the primary side and the secondary side is faulty. L The contents of the bias characteristics include: (1) When the system is operating normally, the inductor current i L In an unbiased state, its average value i L_mean =0; (2) When a single-tube open circuit fault occurs: ① The fault of the primary side switch tube causes the inductor current to form a significant unipolar bias within 1-2 switching cycles, where: When T1 or T4 fails, it will be negatively biased: i L_mean <0; When T2 or T3 fails, it will be forward biased: i L_mean >0; ② When the secondary switch fails, the inductor current needs 3-5 switching cycles to form a bipolar bias, where: When the T5 or T8 tube fails, it will be forward biased: i L_mean >0; When the T6 or T7 tube fails, it will be negatively biased: i L_mean <0; (3) The bias amplitude is determined by i L_mean The absolute value representation of .

5. The open circuit fault diagnosis method of the switch tube of the dual active bridge converter according to claim 1, characterized in that: The step 3 selects the difference between the average value of the inductor current and the average value of the bridge arm midpoint voltage in the fault state and the normal state as the fault feature, and its content is: When the switch tube open circuit fault occurs on the primary and secondary sides, a DC bias will occur in the inductor current, but the bias degree of the inductor current is different. Through experiments, it is known that the DC bias of the inductor current caused by the switch tube open circuit fault on the primary side is greater than the DC bias caused by the switch tube open circuit fault on the secondary side. Therefore, a threshold value θ is set to determine which side has an open circuit fault. Specifically, the threshold value is set as follows: First, the inductor current average value deviation Δi is calculated L , that is, the fault state and the normal state i L_mean The difference between the current bias characteristics, Δi L A value of zero indicates no bias; Δi L Greater than zero indicates forward bias; Δi L Less than zero indicates a negative bias. L When the absolute value exceeds the preset threshold value θ, it is determined to be a significant bias caused by the primary side fault; when Δi L When the absolute value is less than θ and not zero, it is judged as a slight bias caused by the secondary side fault, which is used to distinguish the bias degree of different fault types. At the same time, by calculating the average voltage deviation Δu of the bridge arm midpoint AO and Δu CP , which are respectively u in fault state and u in normal state AO_mean and u CP_mean The difference between and can effectively characterize the distortion characteristics of the voltage waveform at the midpoint of the corresponding bridge arm.

6. The open circuit fault diagnosis method of the switch tube of the dual active bridge converter according to claim 5, characterized in that: In actual application, the processed fault feature quantity also needs to introduce a threshold to avoid errors in the diagnosis process, and its content is: The influence of measurement error on the diagnosis result is eliminated by setting the threshold parameter. The specific implementation method is: when Δi L , Δu AO or Δu CP When the measured value is near zero, it indicates that the corresponding waveform has not changed significantly. Considering the switching noise, parasitic parameters and sensor errors in actual operation, the measured value of i under normal working conditions will be L 、u AO and u CP There are inherent fluctuations, so three positive threshold parameters α, β and γ are introduced to adjust Δi L , Δu AO and Δu CP The measurement results are processed with tolerance to avoid misdiagnosis caused by inherent interference in the system.

7. The open circuit fault diagnosis method of the switch tube of the dual active bridge converter according to claim 6, characterized in that: The method can realize real-time online fault diagnosis in a short time by using only one current sensor and two voltage sensors, and can quickly and accurately identify the open circuit fault of the switch tube, thereby improving the reliability and safety of the system.

Images