A method for fault diagnosis of pulse power device

Abstract

The present invention specifically relates to a pulse power device fault diagnosis method, which solves the technical problem that when a component of a Marx fails during the charging process in an existing large-scale pulse power device, the entire primary energy storage unit will fail. The large-scale pulse power device has a large number of Marxs, and the traditional method of checking one by one is extremely labor-intensive and difficult to quickly and accurately determine the specific location of the fault point on the Marx. The pulse power device fault diagnosis method includes the following steps: 1) setting a Marx fault detection device on the charging cable between the Marx and an external high-voltage constant-current power supply (8); determining the Marx fault type, which includes a charging cable fault and a Marx fault; 2) establishing a Marx circuit model and a Marx fault model; 3) using the Marx fault detection device to detect the charging current on each Marx charging cable in real time, and when the peak current I of the charging current is Peak When the preset current is exceeded, a fault alarm is given, the Marx fault type is determined, and the specific location of the Marx fault is obtained.

Description

Technical Field

[0001] The present invention particularly relates to a fault diagnosis method for a pulse power device. Background Art

[0002] With the rapid development of pulsed power technology in recent years, pulsed power devices have demonstrated the characteristics of high power, large size, and multiple devices operating in parallel. Large-scale pulsed power devices based on linear transformer drivers (LTDs) used in Z-pinch fusion research typically require hundreds of LTD induction cavities connected in series and tens of thousands of high-voltage gas switches operating in parallel. The Z-pinch Fusion Experimental System (PTS) built by the China Academy of Engineering Physics utilizes multiple Marx generators. During the parallel discharge process of multiple Marx generators, components within the multiple Marx generators may experience faults such as open circuits and life failures. Furthermore, components may experience poor contact and high-voltage breakdown, which can cause failures in large-scale pulsed power devices. Implementing fault diagnosis for multiple Marx generators is a key research topic for improving the operating efficiency of large-scale pulsed power devices.

[0003] Marx is an important component of large-scale pulse power devices and is their primary energy storage unit. When the voltage level remains unchanged, multiple Marx generators are usually connected in parallel, and one or more power supplies are used to charge the Marx generators. During the charging process of large-scale pulse power devices, the most common faults are: (1) abnormal operation of individual switches causing self-discharge; (2) capacitor breakdown; (3) high-voltage breakdown of the charging cable. When the above faults occur, not only the energy on the capacitor closest to the fault on the same Marx is released through the faulty unit, but also the energy on other capacitors on the Marx and other parallel Marxs will be released through the faulty unit, causing the performance of the faulty unit to deteriorate seriously, causing the stability of the large-scale pulse power device to decay rapidly. Therefore, in the case of Marx charging failure, it is necessary to complete the maintenance of the Marx immediately to prevent the fault from expanding.

[0004] The Marx in existing large-scale pulse power devices all use transformer oil insulation and are large in size. To carry out maintenance, the transformer oil must be drained first, or the Marx must be lifted out for inspection, which results in a huge maintenance workload. If the specific location of the fault is unknown, the maintenance work will be aimless, which will greatly extend the maintenance time and make it difficult to achieve accurate maintenance. Therefore, accurate positioning of the fault before maintenance is of great significance for rapid maintenance and ensuring the stable operation of large-scale pulse power devices. Chinese patent CN109659920B "A fault detection and grounding protection system for large-capacity pulse power devices" adopts the method of light-emitting diode + optical fiber + photosensitive diode. By observing the serial number of the light-emitting diode, it can determine which Marx has failed and achieve rapid grounding. This method can quickly detect faults, but it cannot determine the type of fault, nor can it determine the specific location of the fault point on the Marx.

[0005] In summary, during the charging process of Marx in existing large-scale pulse power devices, if a component of a Marx fails, the entire primary energy storage unit will fail. Large-scale pulse power devices have a large number of Marxs, and the traditional one-by-one troubleshooting method is extremely labor-intensive and difficult to quickly and accurately determine the specific location of the fault point on the Marx. Summary of the Invention

[0006] The purpose of the present invention is to solve the technical problem that when a component of a Marx in an existing large-scale pulse power device fails during the charging process, it will cause the failure of the entire primary energy storage unit. Since a large-scale pulse power device has a large number of Marxs, the traditional method of checking one by one is extremely labor-intensive and difficult to quickly and accurately determine the specific location of the fault point on the Marx. A pulse power device fault diagnosis method is provided to achieve rapid and accurate determination of the fault type and fault location on the Marx.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0008] A method for diagnosing a fault of a pulse power device is characterized in that it comprises the following steps:

[0009] 1) A Marx fault detection device is installed on the charging cable between Marx and the external high-voltage constant-current power supply; the Marx fault type is determined, which includes charging cable fault and Marx fault;

[0010] 2) Establish a Marx circuit model and calculate the peak current I corresponding to each fault type according to the Marx fault type Peak and half-peak width time t FWHM , establish the Marx fault model;

[0011] 3) Use the Marx fault detection device in step 1) to detect the charging current on each Marx charging cable in real time. When the peak current I Peak When the current exceeds the preset value, the fault diagnosis is started and a fault alarm is given. Peak and half-peak width time t FWHM , corresponding to the Marx fault model obtained in step 2), determine the Marx fault type and obtain the specific location of the Marx fault.

[0012] Furthermore, in step 1), the Marx fault detection device includes a plurality of current sensors and a data acquisition and processing unit;

[0013] The multiple current sensors are respectively mounted on the outside of the semiconductor layers of the multiple Marx charging cables, and a shielding shell is provided on the outside of the current sensors; the shielding shell is connected to the metal shielding mesh of the charging cable;

[0014] The multiple current sensors are all electrically connected to the data acquisition and processing unit.

[0015] Furthermore, in step 1), the charging cable fault includes a charging cable circuit breaker fault and a charging cable high voltage breakdown fault;

[0016] Marx faults include switch self-discharge faults and capacitor partial breakdown faults.

[0017] Furthermore, step 2) is specifically as follows:

[0018] Perform Marx circuit modeling, simulate each Marx fault, obtain the current-time waveform corresponding to all Marx faults, and calculate the peak current I of each current-time waveform Peak and half-peak width time t FWHM , get "Fault-I Peak -t FWHM " list to complete the establishment of the Marx fault model.

[0019] Furthermore, step 3) is specifically as follows:

[0020] 3.1. Use the current sensor in step 1) to collect the charging current on the charging cables of all Marx in real time. When the charging current on the charging cable corresponding to any Marx is 0, it is determined that the Marx has a charging cable circuit breaker fault, and the data acquisition and processing unit gives a fault alarm; when the peak current I Peak When the current exceeds the preset value, the data acquisition and processing unit will give a fault alarm and execute step 3.2;

[0021] 3.2. Detect the charging current on all Marx charging cables. When the charging current on any charging cable has the same amplitude as the charging current on other charging cables but flows in the opposite direction, other Marx discharges reversely to the breakdown point of the Marx charging cable through the corresponding charging cable, then it is determined that the Marx has a charging cable high-voltage breakdown fault; otherwise, a Marx fault occurs, and go to step 3.3;

[0022] 3.3. According to the charging current on all Marx charging cables, the corresponding "Fault-I Peak -t FWHM "List, determine the fault Marx, and obtain the specific location of the Marx fault.

[0023] Further, in step 2), obtain "Fault-I Peak -t FWHM The specific list is:

[0024] 2.1. Use Pspice to model the Marx circuit, simulate the self-discharge fault of each switch in Marx, obtain the current-time waveform corresponding to all switch self-discharge faults, and obtain the peak current I of each switch current-time waveform. Peak and half-peak width t FWHM , get "Switch Self-Discharge Fault-I Peak -t FWHM "list;

[0025] 2.2. Use Pspice to model the Marx circuit, simulate the capacitor breakdown fault in each Marx, obtain the current-time waveform corresponding to all capacitor breakdown faults, and obtain the peak current I of each capacitor current-time waveform. Peak and half-peak width t FWHM , get "capacitor partial breakdown fault-I Peak -t FWHM ” list.

[0026] Furthermore, in step 3.3, the fault Marx is determined to be:

[0027] At the same time, the maximum peak current I among all Marx charging currents Peak The corresponding Marx is the fault Marx, and the fault type is the switch self-discharge fault;

[0028] Among all Marx charging currents, the Marx corresponding to the negative charging current is the fault Marx, and the fault type is the capacitor partial breakdown fault;

[0029] In step 3.3, the specific location of the Marx fault is obtained as follows:

[0030] When the Marx fault is a switch self-discharge fault, according to the "Switch Self-Discharge Fault-I Peak -t FWHM "List, determine the switch self-discharge fault, and determine the specific location of the faulty switch;

[0031] When the Marx fault is a capacitor breakdown fault, according to the "capacitor breakdown fault-I Peak -t FWHM "List, determine the capacitor partial breakdown fault, and determine the specific location of the faulty capacitor.

[0032] Furthermore, in step 3.1, before Marx starts charging, the current sensor is cleared.

[0033] Furthermore, in step 1), the input range of the data acquisition and processing unit is greater than the output range of the current sensor.

[0034] Furthermore, in step 1), the current sensor is a ring structure;

[0035] In step 1), the sampling rate of the data acquisition and processing unit is greater than 5 kHz.

[0036] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0037] 1. The pulse power device fault diagnosis method of the present invention sets a non-contact current sensor in each Marx charging circuit to detect the charging current value of each Marx, determine whether a large-scale pulse power device with multiple Marxes connected in parallel has a fault, and quickly and accurately determine the fault type and fault location. This reduces the fault location time from several days to a few minutes, reduces the maintenance workload, and improves the operating efficiency of the large-scale pulse power device.

[0038] 2. The pulse power device fault diagnosis method of the present invention detects the current change of each Marx in real time during the charging process. By comparing multiple Marxs horizontally, comparing the historical data of a single Marx, and analyzing the charging waveform of the Marx itself, it can accurately judge the working status of each Marx, capture minor faults of a single capacitor or switch, and terminate the pressurization program of a large pulse power device in time, thereby preventing the occurrence of larger-scale failures of the entire large pulse power device.

[0039] 3. The present invention's pulse power device fault diagnosis method utilizes a non-contact current sensor and mounting method, enabling measurement of tiny currents at high voltages (hundreds of kV) in large pulse power devices. The shielded outer shell of the current sensor reduces damage to the sensor caused by high-voltage discharge, solving the problem of measuring small currents in high-voltage environments and demonstrating its potential for widespread adoption. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a cross-sectional diagram of the installation of a current sensor in an embodiment of the pulse power device fault diagnosis method of the present invention;

[0041] Figure 2 This is a schematic diagram of a parallel structure of three Marx units in an embodiment of a pulse power device fault diagnosis method of the present invention;

[0042] Figure 3 This is a schematic diagram of a Marx3 front-end charging cable disconnection fault in an embodiment of the pulse power device fault diagnosis method of the present invention;

[0043] Figure 4 This is a schematic diagram of a high-voltage breakdown fault in a Marx3 front-end charging cable in an embodiment of a pulse power device fault diagnosis method of the present invention;

[0044] Figure 5 This is a schematic diagram of a self-discharge fault of the switch S1 in Marx3 in an embodiment of the fault diagnosis method for a pulse power device of the present invention;

[0045] Figure 6 This is a schematic diagram of a self-discharge fault of the switch S2 in Marx3 in an embodiment of the fault diagnosis method for a pulse power device of the present invention;

[0046] Figure 7 This is a schematic diagram of a self-discharge fault of switch S3 in Marx3 in an embodiment of a fault diagnosis method for a pulse power device of the present invention;

[0047] Figure 8 This is a waveform diagram of the current flowing through the current sensor H1, the current sensor H2, and the current sensor H3 when the switch S1 in Marx3 self-discharges in the embodiment of the pulse power device fault diagnosis method of the present invention;

[0048] Figure 9 This is a waveform diagram of the current flowing through the coil of the current sensor H3 when the switches S1, S2, and S3 in Marx3 are broken down in the embodiment of the pulse power device fault diagnosis method of the present invention;

[0049] Figure 10 This is a schematic diagram of a partial breakdown fault of capacitor C1 in Marx3 in an embodiment of a fault diagnosis method for a pulse power device of the present invention;

[0050] Figure 11 This is a waveform diagram of the current flowing through the current sensor H1, the current sensor H2, and the current sensor H3 when the capacitor C11 in the capacitor C1 of Marx3 breaks down in the embodiment of the pulse power device fault diagnosis method of the present invention;

[0051] Figure 121 is a current waveform diagram when capacitors C1 to C4 in Marx3 are partially broken down in an embodiment of the pulse power device fault diagnosis method of the present invention.

[0052] The reference numerals are as follows:

[0053] 1-cable core wire, 2-insulation layer, 3-semiconductor layer, 4-metal shielding mesh, 5-outer sheath, 6-current sensor, 7-shielded shell, 8-external high-voltage constant current power supply, 9-data acquisition and processing unit. DETAILED DESCRIPTION

[0054] The pulse power device fault diagnosis method of the present invention detects the charging current on the parallel Marx charging cable and the fault current under fault conditions, compares them according to the method of the present invention, and determines the fault location and cause. Specifically, it includes the following steps:

[0055] 1) If Figure 1 As shown, a Marx fault detection device is provided;

[0056] The Marx fault detection device includes multiple current sensors 6 and a data acquisition and processing unit 9. The charging cable includes an insulating layer 2, a semiconducting layer 3, a metal shielding mesh 4, and an outer sheath 5, which are sequentially wrapped around the cable core 1. The current sensor 6 is sheathed outside the semiconducting layer 3, and a shielding shell 7 is provided outside the current sensor 6. The shielding shell 7 is connected to the metal shielding mesh 4. The multiple current sensors 6 are electrically connected to the data acquisition and processing unit 9, completing the setup of the Marx fault detection device. A current sensor 6 is installed on each Marx charging cable and all current sensors 6 are connected to the data acquisition and processing unit 9. The input range of the data acquisition and processing unit 9 is greater than the output range of the current sensor 6. Ensure that the shielding shell 7 is well connected to the metal shielding mesh 4, that the metal shielding mesh 4 is well grounded, and that no current flows through the shielding shell 7. The output cable of the current sensor 6 should use a twisted-pair shielded cable.

[0057] The cable core 1 conducts high voltage, while the insulation layer 2 provides insulation between the high voltage and ground potential. The semiconducting layer 3, located outside the insulation layer 2, equalizes the electric field. The metal shielding mesh 4 provides a stable ground potential for the high-voltage cable, shielding the high-voltage electric field from the outside. The outer sheath 5 protects the internal structure from damage. The current sensor 6 is a ring-shaped structure that fits over the exterior of the semiconducting layer 3 and measures the current passing through its inner hole. During high-voltage charging and discharging, a weak current flows through the metal shielding mesh 4, preventing it from passing through the current sensor 6. High-voltage environments are subject to significant electromagnetic interference and discharge, so a shielding housing 7 protects the current sensor 6.

[0058] The input range of the data acquisition and processing unit 9 is greater than the output range of the current sensor 6. The data acquisition and processing unit 9 employs isolation and shielding measures (for example, when away from the Marx charging environment, it is installed in an external electromagnetic shielding box) to ensure the authenticity and reliability of the sampled current signal. The data acquisition and processing unit 9 has a sampling rate greater than 5kHz, which meets the requirements of continuous waveform acquisition, analysis, high resolution, and real-time detection to meet the measurement requirements of charging current and abnormal discharge current. In this embodiment, the data acquisition and processing unit 9 uses an oscilloscope (RIGOLMSO8104) for real-time current detection.

[0059] 2) Establish a Marx circuit model and calculate the peak current I corresponding to each fault type according to the Marx fault type Peak and half-peak width time t FWHM , establish the Marx fault model;

[0060] Model the actual Marx parallel structure. Using an RCS (R - resistor, C - capacitor, S - switch) structure, modeling is performed based on the Marx series, charging polarity, switch position, and resistor position. In this embodiment, the switch breakdown process is considered to be fully conductive, and the capacitor breakdown is modeled as a partial breakdown of the high-voltage capacitor. Other distributed parameters such as Marx inductance, switch on-resistance, and switch on-inductance are ignored.

[0061] Perform Marx circuit modeling, simulate each Marx fault, obtain the current-time waveform corresponding to all Marx faults, and calculate the peak current I of each current-time waveform Peak and half-peak width time t FWHM , get "Fault-I Peak -t FWHM " list to complete the establishment of the Marx fault model.

[0062] like Figure 2-Figure 7 As shown, the input terminals U1 of the three Marxs are all connected to the external high-voltage constant-current power supply 8, and the output terminals Uo of the three Marxs are connected; in step 1), the current sensor 6 is set between the input terminal U1 of the Marx and the external high-voltage constant-current power supply 8.

[0063] Get "Fault-I Peak -t FWHM The specific list is:

[0064] 2.1. Use Pspice to model the Marx circuit, simulate the self-discharge fault of each switch in Marx, obtain the current-time waveform corresponding to all switch self-discharge faults, and obtain the peak current I of each switch current-time waveform. Peak and half-peak width t FWHM, get "Switch Self-Discharge Fault-I Peak -t FWHM "list;

[0065] 2.2. Use Pspice to model the Marx circuit, simulate the capacitor breakdown fault in each Marx, obtain the current-time waveform corresponding to all capacitor breakdown faults, and obtain the peak current I of each capacitor current-time waveform. Peak and half-peak width t FWHM , get "capacitor partial breakdown fault-I Peak -t FWHM ” list.

[0066] In this embodiment, the Marx circuit takes the first single Marx as an example, including a resistor Ro, a resistor Rc1, a resistor Rc2, a resistor Rc3, a resistor Rg1, a resistor Rg2, a resistor Rg3, a capacitor C1, a capacitor C2, a capacitor C3, a capacitor C4, a switch S1, a switch S2 and a switch S3; one end of the resistor Ro is connected to the charging cable, the other end of the resistor Ro is connected to one end of the resistor Rc1 and one end of the capacitor C1, and the other end of the capacitor C1 is grounded; the other end of the resistor Rc1 is respectively connected to one end of the switch S1, one end of the capacitor C2 and one end of the resistor Rc2, and the other end of the switch S1 is grounded; the other end of the capacitor C2 is respectively connected to the charging cable, the other end of the resistor Rc1 is respectively connected to the charging cable, the other end of the resistor Rc1 is connected to the charging cable, the other end of the resistor Rc1 is connected to the charging cable, the other end of the resistor Rc1 is connected to the charging cable, the other end of the capacitor C2 is connected to the charging cable, the other end of the capacitor C4 is connected to the charging cable, the other end of the resistor Rc1 ... capacitor C4 is connected to the charging cable, the other end of the capacitor One end of the resistor Rg1, one end of the switch S2 and one end of the resistor Rg2, the other end of the resistor Rg1 is grounded; the other end of the switch S2 is respectively connected to the other end of the resistor Rc2, one end of the capacitor C3 and one end of the resistor Rc3; the other end of the capacitor C3 is respectively connected to the other end of the resistor Rg2, one end of the switch S3 and one end of the resistor Rg3; the other end of the resistor Rc3 is respectively connected to the other end of the switch S3 and one end of the capacitor C4; the other end of the capacitor C4 is connected to the other end of the resistor Rg3; wherein the resistance value of the resistor Rc1, the resistor Rc2 and the resistor Rc3 is Rc; the resistance value of the resistor Rg1, the resistor Rg2 and the resistor Rg3 is Rg;

[0067] Then, multiple Marxes are connected in parallel and charged simultaneously using the same external high-voltage constant-current power supply 8. The charging terminals of multiple Marxes are connected in parallel, and the high-voltage terminals are connected in parallel. The external high-voltage constant-current power supply 8 is used to charge the capacitors in all Marxes. When the charging voltage reaches the specified voltage, the external high-voltage constant-current power supply 8 is turned off, all switches in the Marx are turned on, and the capacitors are discharged in series, forming a high-voltage pulse discharge.

[0068] 3) Use the Marx fault detection device in step 1) to detect the charging current on each Marx charging cable in real time. When the peak current I Peak When the current exceeds the preset value, the fault diagnosis is started and a fault alarm is given.Peak and half-peak width time t FWHM , corresponding to the Marx fault model obtained in step 2), determine the Marx fault type and obtain the specific location of the Marx fault;

[0069] Marx fault types include charging cable faults and Marx faults. Charging cable faults include charging cable break faults and charging cable high-voltage breakdown faults. When a cable break occurs, the Marx charging current is zero. When a high-voltage breakdown occurs, all Marx capacitors discharge in the opposite direction toward the breakdown point through their respective charging cables. The current on the charging cables is characterized by the same amplitude and reverse flow. Marx faults include switch self-discharge faults and partial capacitor breakdown faults. A switch self-discharge fault manifests itself when a switch shorts during charging. All capacitors on the Marx connected to that switch discharge through that switch; the capacitors on other Marx capacitors discharge through the parallel charging cables. The current on the charging cable is characterized by the current on the faulty Marx equal to the sum of the charging currents on the other Marx charging cables, and the current flows in the opposite direction. A partial capacitor breakdown fault manifests itself as an incomplete breakdown, resulting in a change in the capacitance of the new capacitor, causing current fluctuations on the charging cable. In this case, the current on the charging cable fluctuates slightly and for a short period of time.

[0070] The above four types of faults can all be evaluated by the current characteristics on the charging cable. In order to obtain accurate charging current data, the current signal of the current sensor 6 is cleared before each use. When charging multiple Marx units, the current value flowing through each charging cable is detected in real time to ensure that the charging current is within a reasonable range (i.e., the rated charging current). When the charging current exceeds the reasonable range, the data acquisition and processing unit 9 issues a fault alarm. At the same time, the fault diagnosis program is started to analyze the time-waveform within a period of time before and after this moment, and the fault diagnosis results are given according to the criteria, as follows:

[0071] 3.1. Before Marx starts charging, the current sensor 6 is cleared; the current sensor 6 in step 1) is used to detect the charging current on the charging cables of all Marx in real time. When the charging current on the charging cable corresponding to any Marx is 0, it is determined that the Marx has a charging cable circuit breaker fault, and the data acquisition and processing unit 9 gives a fault alarm; when the peak current I Peak When the current exceeds the preset value, the data acquisition and processing unit 9 generates a fault alarm and executes step 3.2;

[0072] 3.2. Detect the charging current on all Marx charging cables. When the charging current on any charging cable has the same amplitude as the charging current on other charging cables but flows in the opposite direction, the other Marx reversely discharges to the breakdown point of the Marx charging cable through the corresponding charging cable. It is determined that the Marx has a high-voltage breakdown fault in the charging cable, and the data acquisition and processing unit 9 issues a fault alarm; otherwise, a Marx fault occurs and step 3.3 is executed;

[0073] 3.3. According to the charging current on all Marx charging cables, the corresponding "Fault-I Peak -t FWHM "List, judge the fault Marx, and obtain the specific location of the Marx fault. Among them, judging the fault Marx is specifically:

[0074] At the same time, the maximum peak current I among all Marx charging currents Peak The corresponding Marx is the fault Marx, and the fault type is the switch self-discharge fault; the Marx corresponding to the negative charging current among all Marx charging currents is the fault Marx, and the fault type is the capacitor partial breakdown fault;

[0075] The specific location of the Marx fault is:

[0076] When the Marx fault is a switch self-discharge fault, according to the "Switch Self-Discharge Fault-I Peak -t FWHM "List, judge the switch self-discharge fault, determine the specific location of the faulty switch; when the Marx fault is a capacitor breakdown fault, according to the "capacitor breakdown fault-I Peak -t FWHM "List, determine the capacitor partial breakdown fault, and determine the specific location of the faulty capacitor.

[0077] In order to further illustrate the technical solution of the present invention, three Marx primary energy storage units connected in parallel are taken as an example.

[0078] The current sensor 6 used in the present invention is different from the conventional current sensor 6 in that the current sensor 6 in the present invention is installed in a shielding shell 7, which can effectively reduce spatial interference in complex electromagnetic environments such as high-voltage discharge and arc.

[0079] like Figure 2 As shown in the figure, the primary energy storage unit is composed of three Marxs connected in parallel, namely Marx1, Marx2 and Marx3. Each Marx is composed of three switches, four capacitors and multiple resistors. Among them, the capacitor is 40nF and the resistor is 1.4kΩ. The primary energy storage unit adopts single-end charging. Figure 2 The Marx input terminal U1 is connected to the output terminal of the external high-voltage constant-current power supply 8. The Marx output terminal Uo is the high-voltage output of the Marx, and the three Marx output terminals Uo are connected in parallel. Each current sensor 6 is installed at the input end of the Marx charging cable. During the charging process, it can detect the charging current value of the corresponding Marx in real time; in the event of a Marx fault, it can detect the fault current to provide fault judgment criteria.

[0080] The following describes different faults:

[0081] 1. Charging cable circuit breaker failure

[0082] like Figure 3 As shown, the charging cable to Marx3 is broken. At this point, the external high-voltage constant-current power supply 8 can only charge Marx1 and Marx2. The current flowing through the coils of current sensors H1 and H2 is i1 + i2 = I / 2 (I is the output current of the external high-voltage constant-current power supply 8). The current measured by current sensor H3 is 0. Therefore, it can be determined that the charging cable to Marx3 is broken.

[0083] 2. A high-voltage breakdown fault occurs in the charging cable

[0084] like Figure 4 As shown in the figure, during the charging process, the high-voltage cable is damaged and cannot withstand the high voltage, causing discharge between cable core 1 and metal shield 4, which appears as a short circuit. At this time, Marx1, Marx2, and Marx3 discharge in the opposite direction through their respective charging cables toward the breakdown point. The outputs of current sensors H1, H2, and H3 exhibit the same amplitude and negative polarity.

[0085] 3. Switch self-discharge failure

[0086] like Figure 5 、 Figure 6 、 Figure 7 As shown in the figure, the switch self-discharge fault model of Marx3 switches S1, S2, and S3 occurs respectively. When a switch in Marx3 self-discharges, all capacitors in Marx3 will discharge through this switch. At the same time, all capacitors in Marx1 and Marx2 will also discharge through this switch. Pspice is used to model the circuit and simulate each switch self-discharge fault. The current-time waveforms under all switch self-discharge fault conditions are obtained, and the peak current I of each switch waveform is obtained. Peak and half-peak width time t FWHM According to the waveform, the "switch self-discharge fault-I Peak -t FWHM ” list, see Table 1.

[0087] Table 1 "Switch Self-Discharge Fault-I Peak -t FWHM List

[0088]

[0089] The Marx criterion corresponding to the switch self-discharge fault. Figure 8 As shown in the figure, when switch S1 of Marx3 experiences a self-discharge fault, the current flowing through current sensors H1, H2, and H3 versus time. Since Marx1 and Marx2 are identical, the currents flowing through current sensors H1 and H2 are also identical, but in opposite directions. Their sum equals the current flowing through current sensor H3. Therefore, it can be determined that Marx3, corresponding to current sensor 6 (i.e., current sensor H3) with the largest absolute current value, has failed at the same moment. Therefore, the criterion for determining a faulty Marx is: the Marx corresponding to the maximum peak current is the faulty Marx.

[0090] Fault switch judgment in Marx. Figure 9 As shown in Figure 1, the current-time waveform of the current sensor H3 is shown when switches S1, S2, and S3 in Marx3 are broken down. As the fault Marx moves away from the charging current inlet, the current peak value flowing through the current sensor H3 becomes smaller and the half-peak width time t FWHM The switch that keeps getting longer is a faulty switch.

[0091] Check the waveform of current sensor H3, the highest peak current I Peak is 310A, half-peak width time t FWHM The data in Table 1 is consistent with the typical data of the first switch breakdown, so it is determined that the fault occurred at the second switch of Marx.

[0092] 4. Partial breakdown of capacitor

[0093] Conventional high-voltage capacitors are composed of multiple small capacitors connected in series. When one capacitor fails, the others won't immediately break down. Only after multiple breakdowns occur will the entire capacitor breakdown. Therefore, in this embodiment, the model uses two capacitors to simulate the use of a single high-voltage capacitor and simply simulates the current change when one capacitor breaks down.

[0094] like Figure 10As shown in Figure 1, the Marx criterion corresponding to the partial breakdown fault of the capacitor is that in Marx3, the capacitor C1 partially breaks down. The value of capacitor C1 is 40nF. Capacitor C1 is composed of two capacitors in series, capacitor C11 and capacitor C12. Each capacitor is 80nF. Capacitors C11 and C12 form a capacitor voltage divider. The voltage of capacitors C11 and C12 is 0.5U1. When partial breakdown occurs, capacitor C11 is short-circuited, and the voltage of capacitor C1 is equal to the voltage of capacitor C12, causing other capacitors to charge capacitor C12. Therefore, the current on current sensor H1 and current sensor H2 is positive (i.e., outflow), and the current on current sensor H3 is negative (i.e., inflow), so the fault Marx can be judged as Marx3.

[0095] like Figure 11 As shown in the current waveform diagram when the capacitor C1 is partially broken down in Marx3, the current of the current sensor H3 is negative, and the currents of the current sensors H1 and H2 are positive. Therefore, it can be determined that the capacitor breakdown occurs in Marx3.

[0096] Faulty capacitor determination in Marx. The current flowing through current sensor H3 is mainly the charging current of Marx1 and Marx2 to Marx3, which can be equivalent to Cmarx1 and Cmarx2. The faulty capacitor is charged through multiple resistors Rc in Marx3. According to the RC charging model, the charging time t = RC. For the same capacitance C, the larger the resistance R value, the longer the charging time, that is, t FWHM The larger the half-peak width time t FWHM The corresponding capacitor is the faulty capacitor.

[0097] Use Pspice to model the circuit and simulate the partial breakdown fault of each capacitor in Marx to obtain the current-time waveform under the condition of partial breakdown fault of all capacitors and the peak current I of each capacitor waveform. Peak and half-peak width time t FWHM According to the waveform, the "capacitor partial breakdown fault-I Peak -t FWHM ” list, see Table 2.

[0098] Table 2 "Capacitor Partial Breakdown Fault -I Peak -t FWHM List

[0099]

[0100] Complete the above modeling and "Capacitor Partial Breakdown Fault-I Peak -t FWHM"After the list is compiled, use the external high-voltage constant-current power supply 8 to charge Marx1, Marx2, and Marx3, and the charging current is set to 30mA. At this time, the charging current on the current sensor H1, current sensor H2, and current sensor H3 should all be 10mA. Set the charging current deviation to 10%. When the current sensor H3 reading exceeds 11mA, the data acquisition and processing unit 9 will issue a fault alarm.

[0101] The data acquisition and processing unit 9 collects and plots the data for a certain period of time before and after the fault alarm. The user compares the data to obtain the peak current I of the current sensor H3. Peak The peak current I of current sensor H1 and current sensor H2 is larger and negative. Peak They are equal and the polarity is positive, so it is determined that Marx3 is faulty.

[0102] like Figure 12 As shown in the figure, the waveform of the current sensor H3 is as follows: the highest peak current IPeak is 630A, and the half-peak width time t FWHM The current waveforms for the partial breakdown of capacitors C1 to C4 in Marx3 are shown in Table 2.

Claims

1. A pulse power device fault diagnosis method, characterized in that: The following steps are involved: 1) A Marx fault detection device is provided on a charging cable between Marx and an external high-voltage constant-current power supply (8); a Marx fault type is determined, wherein the Marx fault type includes a charging cable fault and a Marx fault; the Marx fault detection device includes a plurality of current sensors (6) and a data acquisition and processing unit (9); 2) Establish a Marx circuit model and calculate the peak current I corresponding to each fault type according to the Marx fault type Peak and half-peak width time t FWHM , establish the Marx fault model: Perform Marx circuit modeling, simulate each Marx fault, obtain the current-time waveform corresponding to all Marx faults, and calculate the peak current I of each current-time waveform Peak and half-peak width time t FWHM , get "fault-I Peak -t FWHM "List, complete the establishment of Marx fault model; 3) Use the Marx fault detection device in step 1) to detect the charging current on each Marx charging cable in real time. When the peak current I Peak When the current exceeds the preset value, the fault diagnosis is started and a fault alarm is given. Peak and half-peak width time t FWHM , corresponding to the Marx fault model obtained in step 2), determine the Marx fault type and obtain the specific location of the Marx fault: 3.

1. Using the current sensor (6) in step 1) to collect the charging current on the charging cables of all Marx in real time, when the charging current on the charging cable corresponding to any Marx is 0, it is determined that the Marx has a charging cable circuit breaker fault, and the data acquisition and processing unit (9) issues a fault alarm; When any Marx corresponds to the peak current I of the charging current on the charging cable Peak When the current exceeds the preset value, the data acquisition and processing unit (9) generates a fault alarm and executes step 3.2; 3.

2. Detect the charging current on all Marx charging cables. When the charging current on any charging cable has the same amplitude as the charging current on other charging cables but flows in the opposite direction, other Marx discharges reversely to the breakdown point of the Marx charging cable through the corresponding charging cable, then it is determined that the Marx has a charging cable high-voltage breakdown fault; otherwise, a Marx fault occurs, and go to step 3.3; 3.3、According to the charging current on all Marx charging cables, the corresponding "Fault-I Peak -t FWHM "List, determine the fault Marx, and obtain the specific location of the Marx fault.

2. A pulse power device fault diagnosis method according to claim 1, characterized in that: In step 1), the plurality of current sensors (6) are respectively sleeved on the outside of the semiconductor layers (3) of the plurality of Marx charging cables, and a shielding shell (7) is provided on the outside of the current sensor (6); the shielding shell (7) is connected to the metal shielding net (4) of the charging cable; The plurality of current sensors (6) are all electrically connected to the data acquisition and processing unit (9).

3. A pulse power device fault diagnosis method according to claim 2, characterized in that: In step 1), the charging cable fault includes a charging cable circuit breaker fault and a charging cable high voltage breakdown fault; The Marx fault includes a switch self-discharge fault and a capacitor partial breakdown fault.

4. A pulse power device fault diagnosis method according to claim 3, characterized in that: In step 2), the acquisition of "Fault-I Peak -t FWHM The specific list is: 2.

1. Use Pspice to model the Marx circuit, simulate the self-discharge fault of each switch in Marx, obtain the current-time waveform corresponding to all switch self-discharge faults, and obtain the peak current I of each switch current-time waveform. Peak and half-peak width t FWHM , get "Switch Self-Discharge Fault-I Peak -t FWHM "list; 2.

2. Use Pspice to model the Marx circuit, simulate the capacitor breakdown fault in each Marx, obtain the current-time waveform corresponding to all capacitor breakdown faults, and obtain the peak current I of each capacitor current-time waveform. Peak and half-peak width t FWHM , get "capacitor partial breakdown fault-I Peak -t FWHM ” list.

5. A pulse power device fault diagnosis method according to claim 4, characterized in that: In step 3.3, the fault determination Marx is specifically: At the same time, the maximum peak current I among all Marx charging currents Peak The corresponding Marx is the fault Marx, and the fault type is the switch self-discharge fault; Among all Marx charging currents, the Marx corresponding to the negative charging current is the fault Marx, and the fault type is the capacitor partial breakdown fault; In step 3.3, the specific location of the Marx fault is obtained as follows: When the Marx fault is a switch self-discharge fault, according to the "Switch Self-Discharge Fault-I Peak -t FWHM "List, determine the switch self-discharge fault, and determine the specific location of the faulty switch; When the Marx fault is a capacitor breakdown fault, according to the "capacitor breakdown fault-I Peak -t FWHM "List, determine the capacitor partial breakdown fault, and determine the specific location of the faulty capacitor.

6. A pulse power device fault diagnosis method according to claim 5, characterized in that: In step 3.1, before Marx starts charging, the current sensor (6) is cleared.

7. A pulse power device fault diagnosis method according to claim 6, characterized in that: In step 1), the input range of the data acquisition and processing unit (9) is greater than the output range of the current sensor (6).

8. A pulse power device fault diagnosis method according to claim 7, characterized in that: In step 1), the current sensor (6) is a ring structure; In step 1), the sampling rate of the data acquisition and processing unit (9) is greater than 5 kHz.

Images