Three-phase inverter cascade switch tube fault diagnosis method based on two-stage interval sliding mode observer
By establishing a hybrid logic dynamic model and designing an interval sliding mode observer based on a two-level interval sliding mode observer, the synchronous diagnosis of open-circuit and short-circuit faults of the three-phase inverter switch tubes and the rapid identification of cascade faults are achieved, which improves the comprehensiveness and real-time diagnosis of the three-phase inverter in complex fault scenarios and enhances the reliability and safety of the system operation.
Patent Information
- Application Number
- CN202510953114.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-10
- Publication Date
- 2025-10-17
- Estimated Expiration
- 2045-07-10
AI Technical Summary
Existing technologies are unable to simultaneously handle the mixed diagnosis of both open-circuit and short-circuit faults of switching tubes, and their ability to quickly identify and locate cascading faults is insufficient, impacting the real-time diagnosis of three-phase inverters in complex fault scenarios and the timeliness and effectiveness of system fault handling.
A hybrid logic dynamic model is established based on a two-level interval sliding mode observer. An interval sliding mode observer is designed for fault phase judgment. Synchronous diagnosis and fault isolation of open/short circuit states are achieved through the front-stage and back-stage observers, and cascading faults can be quickly identified.
It achieves synchronous diagnosis of open-circuit faults and short-circuit faults of switching tubes, improves the comprehensiveness and real-time diagnosis of three-phase inverters in complex fault scenarios, and improves the reliability and safety of system operation.
Smart Images

Figure CN120802114A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to a three-phase inverter cascaded switching tube fault diagnosis method based on a two-stage interval sliding mode observer and belongs to the three-phase inverter fault diagnosis technical field. BACKGROUND
[0002] As a core component in power electronic devices, the operation reliability and safety of a three-phase inverter are directly related to the stability of the entire power system. In the actual operation process of the three-phase inverter, switching tubes (such as IGBT, MOSFET, etc.) as key power devices are prone to failure due to overcurrent, overvoltage or aging. Among them, switching tube faults mainly include open circuit fault (switching tube cannot normally turn on) and short circuit fault (switching tube cannot normally turn off).
[0003] At present, the switching tube fault diagnosis method for the three-phase inverter has been widely studied, but the existing technical solutions have the following significant defects:
[0004] Firstly, most fault diagnosis methods can only detect and locate a single type of switching tube fault (such as only for open circuit fault or only for short circuit fault), and cannot realize synchronous diagnosis under complex working conditions with mixed open circuit fault and short circuit fault, which limits the comprehensiveness and practicality of fault diagnosis;
[0005] Secondly, in the actual operation scene of the three-phase inverter, when a switching tube has an open circuit or short circuit fault, the voltage spike generated by the fault may break the switching tube at other positions, causing subsequent cascading faults (i.e. secondary or multiple faults induced by the initial fault).
[0006] However, the existing fault diagnosis method has insufficient ability to quickly identify and locate such cascading faults, which is difficult to meet the real-time diagnosis needs of the three-phase inverter under complex fault working conditions, thereby affecting the timeliness and effectiveness of system fault handling.
[0007] Therefore, it is urgent to propose a three-phase inverter fault diagnosis method that can simultaneously diagnose mixed open circuit fault and short circuit fault of switching tubes and has the ability to quickly identify cascading faults, so as to improve the operation reliability and safety of the three-phase inverter under complex fault scenes. SUMMARY
[0008] To solve the problems in the background art, the application provides a three-phase inverter cascaded switching tube fault diagnosis method based on a two-stage interval sliding mode observer.
[0009] To achieve the above purpose, the application adopts the following technical solution: a three-phase inverter cascaded switching tube fault diagnosis method based on a two-stage interval sliding mode observer, the method comprising the following steps:
[0010] S1: Establish a hybrid logic dynamic model and design an interval sliding mode observer considering the fault state of the switching tube;
[0011] S101: Establish a hybrid logic dynamic model of a three-phase inverter considering open-circuit fault and short-circuit fault of the switching tube:
[0012] (1)
[0013] In formula (1):
[0014] represents the three-phase output current;
[0015] represents the direction of the three-phase output current;
[0016] When , ; when , ;
[0017] represents the resistance value of the load resistance;
[0018] represents the filter inductance value;
[0019] represents the DC bus voltage;
[0020] represents the switching signal of the switching tube, when , it means that the switching tube is on, i.e. the switching tube has a short-circuit fault; when , it means that the switching tube is off, i.e. the switching tube has an open-circuit fault;
[0021] represents time;
[0022] S102: According to formula (1), establish the upper bound and lower bound of the predicted output of the interval sliding mode observer considering the fault state of the switching tube:
[0023] (2)
[0024] In formula (2):
[0025] represents the upper bound of the predicted output;
[0026] represents the lower bound of the predicted output;
[0027] represents the matrix transpose;
[0028] upper bound of the predicted output first derivative;
[0029] lower bound of the predicted output first derivative;
[0030] system matrix, identity matrix;
[0031] control matrix;
[0032] observation matrix;
[0033] direct transmission matrix;
[0034] observer gain matrix;
[0035] input of the observer;
[0036] upper bound of the uncertain parameter;
[0037] lower bound of the uncertain parameter;
[0038] upper bound of the disturbance;
[0039] lower bound of the disturbance;
[0040] sign function;
[0041] sliding mode gain;
[0042] observation value;
[0043] S103: Obtain an interval sliding mode observer:
[0044] (3)
[0045] In formula (3):
[0046] weight factor.
[0047] S2: Fault phase determination;
[0048] S201: Take an interval sliding mode observer as a front-stage observer;
[0049] S202: Set the input switch signal in the pre-stage observer to the original switch signal to realize the fault-free state observation mode;
[0050] S203: Set the abnormal state diagnosis variable:
[0051] (4)
[0052] In formula (4):
[0053] T c represents the fundamental period of the current;
[0054] represents the interval sliding mode observer output residual error vector;
[0055] represents the observed value of the A-phase current;
[0056] represents the observed value of the B-phase current;
[0057] represents the observed value of the C-phase current;
[0058] S204: Set the abnormal state detection threshold:
[0059] (5)
[0060] In formula (5):
[0061] represents a coefficient considering the robustness of the fault diagnosis algorithm;
[0062] represents the discretized system matrix;
[0063] represents the discretized gain matrix;
[0064] represents the discretized sliding mode gain;
[0065] wherein: represents the discretization step, represents a unit matrix of appropriate dimension;
[0066] represents the current time step;
[0067] S205: When , it indicates that a fault has occurred; otherwise, it indicates that no fault has occurred;
[0068] S206: If a fault occurs, set the fault phase detection variable:
[0069] (6)
[0070] In formula (6):
[0071] represents the output current residual of each phase, ;
[0072] S207: Fault phase detection variable The phase corresponding to the maximum of the fault feature variable is the fault phase.
[0073] S3: Fault isolation;
[0074] S301: Define fault feature variable:
[0075] (7)
[0076] In formula (7):
[0077] represents the current residual corresponding to the fault phase in the front-stage observer;
[0078] If , it indicates that there is a probability of upper tube open circuit or lower tube short circuit fault;
[0079] If , it indicates that there is a probability of lower tube open circuit or upper tube short circuit fault,
[0080] Two candidate fault types can be obtained;
[0081] S302: Two interval sliding mode observers are used as the rear-stage observers, and the two rear-stage observers are open circuit observer and short circuit observer respectively;
[0082] S303: In the open circuit observer, the switch signal corresponding to the switch tube with a probability of open circuit fault is , to realize the open circuit state observation mode; in the short circuit observer, the switch signal corresponding to the switch tube with a probability of short circuit fault is , to realize the short circuit state observation mode;
[0083] S304: Calculate the output current residuals of the two rear-stage observers, and select the corresponding fault residual according to the fault phase, respectively, denoted as the open circuit observer fault phase residual and the short circuit observer fault phase residual ;
[0084] S305: Define residual diagnosis variable:
[0085] (8)
[0086] S306: setting residual diagnosis variable threshold , if , it indicates that open circuit fault occurs, if , it indicates that short circuit fault occurs, and the candidate fault type discrimination is completed.
[0087] S4: setting two-stage interval sliding mode observer as observation mode corresponding to the fault to match the actual state;
[0088] S5: repeating S2-S4 to complete the diagnosis of cascaded faults.
[0089] Compared with the prior art, the method has the beneficial effects that:
[0090] The method constructs a hybrid logic dynamic model considering open circuit and short circuit fault states of the switch tube, and designs a two-stage interval sliding mode observer (the front-stage observer is used for fault phase judgment, and the rear-stage observer realizes fault isolation through open circuit / short circuit state observation mode), so that the synchronous diagnosis of complex working conditions with mixed open circuit faults and short circuit faults is realized, the fast identification ability of cascaded faults is simultaneously provided, and the diagnosis comprehensiveness, real-time performance, system operation reliability and safety of the three-phase inverter in the complex fault scene are effectively improved. BRIEF DESCRIPTION OF DRAWINGS
[0091] Figure 1 is a flowchart of the method;
[0092] Figure 2 is a structural schematic diagram of the three-phase inverter;
[0093] Figure 3 is a three-phase current waveform diagram obtained in the experiment in the embodiment;
[0094] Figure 4 is a fault diagnosis variable waveform diagram obtained in the experiment in the embodiment;
[0095] Figure 5 is a fault diagnosis result diagram obtained in the experiment in the embodiment. DETAILED DESCRIPTION
[0096] The technical solutions in the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.
[0097] A three-phase inverter cascaded switch tube fault diagnosis method based on a two-stage interval sliding mode observer, the method comprising the following steps:
[0098] S1: Establish a hybrid logic dynamic model and design an interval sliding mode observer considering the fault state of the switch tube;
[0099] S101: According to the working condition of the three-phase inverter, a hybrid logic dynamic model of the three-phase inverter considering the open-circuit fault and short-circuit fault of the switch tube is established:
[0100] (1)
[0101] In formula (1):
[0102] represents the three-phase output current;
[0103] represents the direction of the three-phase output current;
[0104] , when , the switch tube is on, i.e. the switch tube has a short-circuit fault; when , the switch tube is off, i.e. the switch tube has an open-circuit fault; ;
[0105] represents the resistance value of the load resistor;
[0106] represents the filter inductance value;
[0107] represents the DC bus voltage;
[0108] represents the switching signal of the switch tube, , when , the switch tube is on, i.e. the switch tube has a short-circuit fault; when , the switch tube is off, i.e. the switch tube has an open-circuit fault;
[0109] represents time;
[0110] S102: According to formula (1), the upper bound and lower bound of the output predicted by the interval sliding mode observer considering the fault state of the switch tube are established:
[0111] (2)
[0112] In formula (2):
[0113] represents the upper bound of the predicted output;
[0114] represents the lower bound of the predicted output;
[0115] represents the matrix transpose;
[0116] upper bound of the predicted output first derivative;
[0117] lower bound of the predicted output first derivative;
[0118] system matrix, identity matrix;
[0119] control matrix;
[0120] observation matrix;
[0121] direct transmission matrix;
[0122] observer gain matrix;
[0123] input of the observer;
[0124] upper bound of the uncertain parameter;
[0125] lower bound of the uncertain parameter;
[0126] upper bound of the disturbance;
[0127] lower bound of the disturbance;
[0128] sign function;
[0129] sliding mode gain;
[0130] observation value;
[0131] S103: obtain an interval sliding mode observer:
[0132] (3)
[0133] in equation (3):
[0134] weight factor.
[0135] S2: using an interval sliding mode observer as a front-stage observer, setting it to a fault-free state observation mode, predicting the output current in the fault-free state, calculating the residual error between the predicted output current and the actual output current to determine the fault phase;
[0136] S201: using an interval sliding mode observer as a front-stage observer;
[0137] S202: setting the input switch signal of the front-stage observer to the original switch signal to realize the fault-free state observation mode;
[0138] S203: setting the abnormal state diagnosis variable:
[0139] (4)
[0140] In formula (4):
[0141] T c represents the fundamental period of the current;
[0142] represents the output residual error vector of the interval sliding mode observer;
[0143] represents the observed value of the A-phase current;
[0144] represents the observed value of the B-phase current;
[0145] represents the observed value of the C-phase current;
[0146] S204: setting the abnormal state detection threshold:
[0147] (5)
[0148] In formula (5):
[0149] represents a coefficient considering the robustness of the fault diagnosis algorithm;
[0150] represents the discretized system matrix;
[0151] represents the discretized gain matrix;
[0152] represents the discretized sliding mode gain;
[0153] wherein: represents the discretization step, represents a unit matrix of appropriate dimension;
[0154] denotes the current time step;
[0155] S205: When , it indicates that a fault occurs; otherwise, it indicates that no fault occurs;
[0156] S206: If a fault occurs, set the fault phase detection variable:
[0157] (6)
[0158] In formula (6):
[0159] denotes the output current residual error of each phase, ; and
[0160] S207: The phase corresponding to the maximum value of the fault phase detection variable is the fault phase.
[0161] S3: Fault isolation;
[0162] S301: Define the fault feature variable:
[0163] (7)
[0164] In formula (7):
[0165] denotes the current residual error of the fault phase in the front-stage observer;
[0166] If , it indicates that there is a probability of an upper tube open circuit or a lower tube short circuit fault;
[0167] If , it indicates that there is a probability of a lower tube open circuit or an upper tube short circuit fault,
[0168] Two candidate fault types can be obtained.
[0169] It is indicated that there is no fault, and since a fault has been detected in the foregoing, it will not theoretically occur.
[0170] S302: Use two interval sliding mode observers as the rear-stage observers, and the two rear-stage observers are an open circuit observer and a short circuit observer;
[0171] S303: In the open circuit observer, set the switch signal corresponding to the switch tube with a probability of an open circuit fault to realize the open circuit state observation mode; in the short circuit observer, set the switch signal corresponding to the switch tube with a probability of a short circuit fault to realize the short circuit state observation mode.
[0172] S304: Calculate the output current residual of the two post-stage observers, and select the corresponding fault residual according to the fault phase, denoted as open circuit observer fault phase residual and short circuit observer fault phase residual ;
[0173] S305: Define residual diagnostic variables:
[0174] (8)
[0175] S306: Set the residual diagnostic variable threshold , if , it indicates that an open circuit fault occurs, if , it indicates that a short circuit fault occurs, and the candidate fault type discrimination is completed.
[0176] S4: Set the two-stage interval sliding mode observer (pre-stage observer and post-stage observer) to the observation mode corresponding to the fault to match the actual state;
[0177] S5: Repeat S2-S4 to complete the diagnosis of cascaded faults.
[0178] Embodiment 1:
[0179] The engineering object of the present application is a three-phase inverter. The structure of the three-phase inverter is shown in Figure 2 , which includes a DC power supply, a main inverter circuit, three inductors with the same inductance and three resistive loads with the same resistance, and a control module.
[0180] The main inverter circuit includes three bridge arms, each bridge arm contains two power semiconductor switching devices, a total of six power semiconductor switching devices, denoted as Q1-Q6.
[0181] The input of the control module is a three-phase current measurement value, and the output is six switching signals, wherein the control signal of the lower tube is the non-signal of the control signal of the upper tube.
[0182] The DC bus voltage of the three-phase inverter used in this embodiment is 20V, and the output current amplitude is set to 3.6A.
[0183] The observer gain matrix in S102 , the sliding mode gain .
[0184] In S204 .
[0185] According to the residual of the observer, the in S306 can be set to 10%-20% of the expected current output peak value.
[0186] This embodiment is verified by experiment.
[0187] Figure 3 The three-phase current outputted by the three-phase inverter measured by the current sensor in this embodiment 、 and The Q3 open-circuit fault occurs at 0.015s, and the Q1 short-circuit fault occurs at 0.061s. After the open-circuit fault occurs, the positive half cycle of the corresponding phase current is missing; after the short-circuit fault occurs, the corresponding phase current is positively offset.
[0188] Figure 4 The waveforms of the abnormal state diagnosis variable and the residual diagnosis variable calculated in this embodiment, after the Q3 open-circuit fault occurs, increases, indicating that a fault occurs, and the fault phase is determined to be the B phase. The positive number indicates that an open-circuit fault occurs, and the Q3 open-circuit fault diagnosis is completed. The Q1 short-circuit fault diagnosis is the same, and only M is negative.
[0189] Figure 5 The fault diagnosis result diagram, OC3 is the Q3 open-circuit fault diagnosis result, and SC1 is the Q1 short-circuit fault diagnosis result. As can be seen from the diagram, the completion times of the two fault diagnoses are 0.018s and 0.077s, respectively.
[0190] It is apparent to those skilled in the art that the present application is not limited to the details of the foregoing exemplary embodiments, and that the present application can be implemented in other forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and it is intended that all changes and modifications which come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference signs in the claims should not be construed as limiting the claims to the figures in which the reference signs are used.
[0191] In addition, it should be understood that although the present specification is described in terms of embodiments, not every embodiment contains only one independent technical solution, and the description manner of the specification is only for the sake of clarity, and those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be properly combined to form other embodiments that those skilled in the art can understand.
Claims
1. A three-phase inverter cascade switch fault diagnosis method based on a two-level interval sliding mode observer, characterized by: The method comprises the following steps: S1: Establish a hybrid logic dynamic model and design an interval sliding mode observer considering the fault state of the switch tube; S2: fault phase judgment; S3: Fault isolation; S4: Set the two-level interval sliding mode observer to the observation mode corresponding to the fault to match the actual state; S5: Repeat S2-S4 to complete the diagnosis of the cascading fault.
2. The method for diagnosing faults of cascaded switching tubes of a three-phase inverter based on a two-level interval sliding mode observer according to claim 1, characterized in that: The S1 comprises the following steps: S101: Establish a hybrid logic dynamic model of a three-phase inverter considering open-circuit faults and short-circuit faults of the switch tube: (1) In formula (1): Indicates three-phase output current; Indicates the direction of three-phase output current; ,when hour, ;when hour, ; Indicates the resistance value of the load resistor; Indicates the filter inductance value; Indicates the DC bus voltage; Indicates the switching signal of the switching tube, ,when When , it means the switch tube is turned on, that is, the switch tube has a short circuit fault; when When , it means the switch tube is turned off, that is, the switch tube has an open circuit fault; Indicates time; S102: According to formula (1), the upper and lower bounds of the predicted output of the interval sliding mode observer considering the fault state of the switch tube are established: (2) In formula (2): Represents the upper bound of the predicted output; Represents the lower bound of the predicted output; Represents matrix transpose; Represents the upper bound of the predicted output The first derivative of ; Represents the lower bound of the predicted output The first derivative of ; represents the system matrix, represents the identity matrix; represents the control matrix; represents the observation matrix; represents the direct transfer matrix; represents the observer gain matrix; represents the input of the observer; Indicates the upper bound of the uncertain parameter; Indicates the lower bound of the uncertain parameter; represents the upper bound of the perturbation; represents the lower bound of the perturbation; represents a symbolic function; represents the sliding mode gain; represents the observed value; S103: Get the interval sliding mode observer: (3) In formula (3): Represents the weight factor.
3. The method for diagnosing faults of cascaded switches in a three-phase inverter based on a two-level interval sliding mode observer according to claim 2, characterized in that: The S2 comprises the following steps: S201: Using an interval sliding mode observer as a front-stage observer; S202: setting the switch signal input into the front-stage observer to the original switch signal to achieve a fault-free state observation mode; S203: Setting abnormal status diagnostic variables: (4) In formula (4): T c Indicates the fundamental period of current; represents the residual vector output by the interval sliding mode observer; Represents the observed value of phase A current; represents the observed value of phase B current; represents the observed value of phase C current; S204: Setting the abnormal state detection threshold: (5) In formula (5): represents the coefficient considering the robustness of the fault diagnosis algorithm; represents the discretized system matrix; represents the discretized gain matrix; represents the discretized sliding mode gain; in: represents the discretization step size, represents the identity matrix of appropriate dimensions; Indicates the current time step; S205: When If , it means a fault has occurred; otherwise, it means no fault has occurred; S206: If a fault occurs, set the fault phase detection variable: (6) In formula (6): Represents the output current residual of each phase, ; S207: Fault phase detection variable The phase corresponding to the largest one is the fault phase.
4. The method for diagnosing faults of cascaded switches in a three-phase inverter based on a two-level interval sliding mode observer according to claim 3, characterized in that: The S3 comprises the following steps: S301: Define fault characteristic variables: (7) In formula (7): represents the current residual corresponding to the fault in the previous observer; like , it means there is a probability of upper tube open circuit or lower tube short circuit fault; like , it means there is a probability of the lower tube open circuit or upper tube short circuit fault. So we can get two candidate fault types; S302: Two interval sliding mode observers are used as post-stage observers, where the two post-stage observers are an open-circuit observer and a short-circuit observer respectively; S303: In the open circuit observer, the switch signal corresponding to the switch tube with a probability of open circuit failure is set to , realize the open circuit state observation mode; in the short circuit observer, let the switch signal corresponding to the switch tube with a probability of short circuit fault , realize short-circuit state observation mode; S304: Calculate the output current residuals of the two post-stage observers and select the corresponding fault residual amount according to the fault phase, which is recorded as the open circuit observer fault phase residual and the short-circuit observer fault phase residual ; S305: Define residual diagnostic variables: (8) S306: Setting the residual diagnostic variable threshold ,like , it means an open circuit fault occurs. If , it means a short circuit fault has occurred, completing the classification of candidate fault types.
Citation Information
Patent Citations
Current-magnitude-based open-circuit failure online-diagnosis method for power tube of inverter
CN103701394A
Static fault diagnosis circuit and diagnosis method of three-phase full-bridge inverter
CN105629122A
Inverter open-circuit fault rapid diagnosis system based on ESO-MLD and diagnosis method thereof
CN112363086A
Composite open-circuit fault diagnosis method for NPC three-level inverter
CN113075585A
Fault diagnosis method of single-phase parallel inverter
CN114755610A