GIS isolation switch mechanical fault detection method and system

Abstract

The invention provides a GIS disconnecting switch mechanical fault detection method and system, and belongs to the technical field of high-voltage electrical equipment fault monitoring, and the detection method comprises the following steps: collecting the winding current data of a drive motor and the rotation angular displacement data of an outer crank arm in the switching-on and switching-off processes when a disconnecting switch is in a normal operation state and in a mechanical fault; analyzing to obtain feature data; calculating the weight of each feature by using an MIC algorithm; selecting several-dimensional feature data with the maximum weight value, inputting the selected feature data into the decision tree model, carrying out classification by using a random forest algorithm, and training the decision tree model; and using the trained decision tree model to carry out GIS isolation switch mechanical fault detection. According to the method, the MIC and the random forest are adopted for feature selection and data classification, GIS disconnecting switch operation data monitoring system development and mechanical fault diagnosis result analysis are achieved, the fault diagnosis time is shortened, and the nonobjective defect inspection frequency is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage electrical equipment fault monitoring, and particularly relates to a mechanical fault detection method and system for a GIS disconnector. Background Art

[0002] In the power system, as one of the key switch devices for switching operations and changing the operation mode of electrical equipment, the disconnector is mainly used to form a reliable disconnection point between the equipment and lines to be repaired and the power source, preventing breakdown accidents caused by overvoltage and current flashover. The reliable operation of the disconnector is of great significance for ensuring the personal safety of operation and maintenance personnel and the stability of the power system. As a combined electrical equipment with the characteristics of small floor area, low failure rate, and long operation and maintenance cycle, GIS (Gas Insulated Metal Enclosed Switchgear) seals circuit breakers, disconnectors, earthing switches, busbars, lightning arresters, voltage transformers, current transformers, etc. except transformers in a substation in a metal shell filled with SF6 insulating gas at a certain pressure. In recent years, the occupancy rate of GIS in substations with voltage levels of 110 kV and above has been increasing.

[0003] The existing disadvantages of mechanical fault detection for GIS disconnectors include:

[0004] (1) Fine scratches and tiny metal particles inside the GIS are difficult to detect during factory production and acceptance. After long-term operation, the electric field distribution around the scratches and metal particles becomes uneven and may eventually develop into insulation faults such as partial discharge.

[0005] (2) Due to poor production processes, the static contact cavity of the disconnector is relatively long. The ordinary machining method of drilling holes from both sides of the cavity easily causes axial misalignment of the static contact cavity. When the disconnector passes through this position, the abnormally increased frictional force causes a sharp increase in the output torque of the driving motor. When the torque exceeds the shear force of metal components such as the connecting pins of the transmission rod, the pins break, ultimately resulting in the disconnector failing to close or open in place.

[0006] (3) The enclosed and invisible mechanical structure makes it impossible for on-site operation and maintenance personnel to judge the internal operation status of the disconnector in the GIS. In addition, the SF6 gas leakage and large-area power outages caused by disassembly and repair result in the disconnector in the GIS often being in a state of untimely maintenance.

[0007] (4) When operating outdoors in a harsh environment for a long time, the outer surfaces of components such as the outer crank arm and transmission rod of the operating mechanism of the disconnector in the GIS are prone to rust, resulting in an increase in the frictional resistance between the operating mechanisms of the disconnector. The transmission gears, transmission rods, connecting pins, and metal connecting keys are severely worn after multiple closing and opening operations, and their mechanical strength drops significantly.

[0008] During the operation of a GIS disconnector, rich equipment status information exists in its operating mechanism and drive motor and other parts, mainly including the winding current of the drive motor, the rotational angular displacement of the outer toggle arm, etc. Compared with vibration signals and sound signals, current signals and angular displacement signals have the advantages of being easy to collect and not being interfered by environmental noise, which provides a good theoretical basis for mechanical fault diagnosis based on the winding current of the drive motor and the rotational angular displacement of the outer toggle arm.

[0009] The Chinese invention patent with the application publication number CN112557895A discloses a fault diagnosis method for GIS disconnectors based on multiple characteristic quantities. This method needs to collect the motor current, spindle rotation angle and image information of the GIS disconnector, and two diagnoses are required: the first diagnosis is carried out simultaneously with the operation of the GIS disconnector, and it is judged whether the motor is burned due to refusal to open / close during the opening and closing of the disconnector by detecting the signals of the motor current and the spindle rotation angle; the second diagnosis is carried out after the opening and closing of the GIS disconnector is completed, and it is judged whether the disconnector has faults such as jamming, loose fasteners, and incomplete opening and closing by processing, analyzing and comparing the collected current, rotation angle and image signals, and the fault area is identified. This patent needs to judge faults based on the absolute value of the rotation angle signal and the corresponding current amplitude, and diagnose the fault type based on the opening and closing action time, motor current and image. The diagnosis process is cumbersome, inefficient, and the number of fault types that can be diagnosed is small, which is difficult to meet the requirements of the actual scenario. Summary of the Invention

[0010] Aiming at the technical problems in the prior art that it is difficult to quickly and accurately detect the mechanical faults of GIS disconnectors and judge the fault types, the present invention provides a mechanical fault detection method and system for GIS disconnectors, which uses MIC and random forest for feature selection and data classification, realizes the development of the operation data monitoring system of GIS disconnectors and the analysis of mechanical fault diagnosis results, shortens the fault diagnosis time, and reduces the number of aimless defect inspections.

[0011] The technical solution adopted by the present invention is: a mechanical fault detection method for GIS disconnectors, including the following steps:

[0012] Collect the winding current data of the drive motor and the rotational angular displacement data of the outer toggle arm during the opening and closing processes of the disconnector in the normal operation state.

[0013] Collect the winding current data of the drive motor and the rotational angular displacement data of the outer toggle arm during the opening and closing processes of the disconnector when the disconnector has mechanical faults.

[0014] According to the drive motor winding current data and the external link arm rotation angular displacement data collected during the opening and closing processes of the disconnector under normal operating conditions and mechanical failures, the characteristic data of the disconnector under normal operating conditions and the characteristic data of the disconnector mechanical failures are analyzed; the weight of each feature of the characteristic data is calculated using the MIC (Max Information Coefficient) algorithm;

[0015] Select several dimensions of feature data with the largest weight values from the calculation results of the MIC algorithm and input them into the decision tree model, and use the random forest algorithm for classification to train the decision tree model;

[0016] Use the trained decision tree model to detect the mechanical failures of the GIS disconnector, and judge whether there are mechanical failures in the disconnector and the types of mechanical failures.

[0017] Furthermore, use AC and DC current sensors to collect the drive motor winding current of the mechanism phase to obtain the drive motor winding current data; measure the rotation angle of the drive motor main shaft through an angular displacement sensor, and convert the rotation angle of the drive motor main shaft into the rotation angle of the external link arm according to the mechanical structure relationship between the drive motor main shaft and the external link arm to obtain the external link arm rotation angular displacement data.

[0018] Furthermore, the types of mechanical failures of the disconnector include various ones such as rust on the housing end cover, misalignment of the static and moving contact axes, pin detachment between phases AB, pin detachment between phases BC, too low power supply voltage, and too high power supply voltage.

[0019] Furthermore, the mechanical failures of the disconnector are realized through simulation, and only one type of failure is simulated each time.

[0020] Furthermore, the simulation method for rust on the housing end cover is to increase the torque of the drive motor during the entire opening and closing process of the disconnector;

[0021] The simulation method for misalignment of the static and moving contact axes is to keep the torque of the drive motor unchanged during the opening distance stage and increase the over-travel torque;

[0022] The simulation method for pin detachment between phases AB is that the pin detachment causes the drive part of the drive motor to operate with load. By removing the pin of the transmission tie rod between phase A and phase B, the drive motor drives phase B and phase C to operate;

[0023] The simulation method for pin detachment between phases BC is to remove the pin of the transmission tie rod between phase C and phase B, so that the drive motor only drives phase C for opening and closing operations;

[0024] The simulation method for too low power supply voltage is to lower the power supply voltage;

[0025] The analog method for too high power supply voltage is to increase the power supply voltage.

[0026] Furthermore, the characteristic data includes:

[0027] The moment of maximum value of the current in the opening starting stage of the moving contact, the moment of minimum value of the current in the overtravel stage of opening, the moment of maximum value of the current in the opening distance stage of opening, the moment of maximum value of the current in the ending stage of opening and the ending moment of opening, the maximum value of the current in the opening starting stage of the moving contact, the minimum value of the current in the overtravel stage of opening, the maximum value of the current in the opening distance stage of opening and the maximum value of the current in the ending stage of opening;

[0028] The moment of maximum value of the current in the closing starting stage of the moving contact, the moment of maximum value of the current in the just - closed stage of the moving and static contacts, the moment of minimum value of the current in the overtravel stage of closing and the ending moment of closing, the maximum value of the current in the closing starting stage of the moving contact, the maximum value of the current in the just - closed stage of the moving and static contacts and the minimum value of the current in the overtravel stage of closing;

[0029] The starting moment of opening of the moving contact, the moment of just - separated of the moving and static contacts and the moment of in - place of opening of the moving contact;

[0030] The starting moment of closing of the moving contact, the moment of just - closed of the moving and static contacts and the moment of in - place of closing of the moving contact.

[0031] Furthermore, when using the random forest algorithm for classification, the Bootstrap algorithm is adopted to extract multiple sample sets with the number of N in a sampling - with - replacement manner to train the decision tree model; the selection principle of the feature nodes is also to extract in a random sampling - with - replacement manner.

[0032] The technical solution adopted by the present invention is also: a GIS disconnector mechanical fault detection system, including

[0033] AC - DC current sensors, which are used to collect the drive motor winding current data during the opening and closing processes of the disconnector;

[0034] Angular displacement sensors, which are used to collect the rotation angle of the drive motor spindle during the opening and closing processes of the disconnector;

[0035] A storage module, which is used to store the characteristic data of the normal operation state of the disconnector and the characteristic data of the mechanical faults of the disconnector,

[0036] A characteristic data weight calculation module, which is used to calculate the weight of each feature of the characteristic data by using the MIC algorithm according to the data collected by the AC - DC current sensors and angular displacement sensors and the characteristic data stored in the storage module;

[0037] A classification module, which is used to calculate the calculation result of the feature data weight calculation module, and use the trained decision tree model to adopt the random forest algorithm to detect the mechanical fault of the GIS disconnector, determine whether there is a mechanical fault in the disconnector, and the fault type of the mechanical fault.

[0038] Further, the feature data weight calculation module uses the rotation angle of the driving motor main shaft during the opening and closing processes of the disconnector, and according to the mechanical structure relationship between the driving motor main shaft and the outer crank arm, converts the rotation angle of the driving motor main shaft into the rotation angle of the outer crank arm to obtain the rotation angular displacement data of the outer crank arm.

[0039] The technical solution adopted by the present invention is also: a computer-readable storage medium, including a program, and the program can be executed by a processor to implement a method for detecting mechanical faults of a GIS disconnector as described in any one of claims 1-7.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] (1) By analyzing the force characteristics of the operating mechanism of the GIS, the output of the operating mechanism can be indirectly judged by the winding current of the driving motor. By collecting and analyzing the winding current of the driving motor, the change of the load carried by the operating mechanism can be judged; and according to the mechanical structure relationship between the driving motor main shaft and the outer crank arm, the rotation angle of the driving motor main shaft is converted into the rotation angle of the outer crank arm to obtain the rotation angular displacement data of the outer crank arm. The present invention selects feature data from the winding current of the driving motor and the rotation angular displacement of the outer crank arm, and has the advantages of easy collection and no environmental noise interference compared with vibration signals and sound signals.

[0042] (2) The present invention selects 22 feature data from the winding current data of the driving motor and the rotation angular displacement data of the outer crank arm during the opening and closing processes, covering all stages of the opening and closing processes; and uses the MIC algorithm to screen out the positively correlated features beneficial to data classification, and then uses the decision tree as the base learner, and creates multiple tree models at the same time. The decision results of all tree models are integrated together, which is the final classification result. The present invention adopts the Bootstrap algorithm in the random forest algorithm, which has randomness and diversity, and effectively overcomes the defect of overfitting of the decision tree.

[0043] (3) The present invention analyzes the generation principle of common mechanical faults of the disconnector in the GIS, and combines the force characteristics analysis of the operating mechanism of the GIS to simulate the mechanical faults respectively, which can conveniently and accurately collect fault data, and the fault data is more typical, improving the accuracy of classification.

[0044] (4) The present invention plays an important guiding role in improving the stability of the power system and enhancing the equipment maintenance strategy. The present invention can shorten the fault diagnosis time, reduce the number of aimless defect inspections, provide certain guidance for on-site maintenance personnel, and improve the equipment maintenance strategy, which has a positive significance for enhancing the reliability and operation and maintenance level of the disconnector in GIS. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flowchart of an embodiment of the present invention;

[0046] Figure 2 is a force analysis diagram of the operating mechanism of an embodiment of the present invention;

[0047] Figure 3 is an equivalent velocity component diagram of the operating mechanism of an embodiment of the present invention;

[0048] Figure 4 is a current curve and angular displacement curve diagram in the normal operating state of an embodiment of the present invention;

[0049] Figure 5 is a current curve and angular displacement curve diagram when the power supply voltage is too low in an embodiment of the present invention;

[0050] Figure 6 is a current curve and angular displacement curve diagram when the power supply voltage is too high in an embodiment of the present invention;

[0051] Figure 7 is a current curve and angular displacement curve diagram when the pin between phases A and B falls off in an embodiment of the present invention;

[0052] Figure 8 is a current curve and angular displacement curve diagram when the pin between phases B and C falls off in an embodiment of the present invention;

[0053] Figure 9 is a current curve and angular displacement curve diagram when the outer shell end cover is rusted in an embodiment of the present invention;

[0054] Figure 10 is a current curve and angular displacement curve diagram when the axes of the moving and static contacts are misaligned in an embodiment of the present invention;

[0055] Figure 11 is a characteristic number - classification error schematic diagram of an embodiment of the present invention;

[0056] Figure 12 is a fault diagnosis result diagram of an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be described in detail below with reference to the drawings and specific embodiments.

[0058] Embodiments of the present invention analyze the mechanical characteristics of the ZF16-252 angular disconnect switch and the force conditions of each component of the operating mechanism, and provide a method for detecting mechanical faults of GIS disconnect switches, such as Figure 1 shown, which includes the following steps:

[0059] Step 1: In this embodiment, the force characteristics of the operating mechanism of GIS are analyzed, as Figure 2 shown, and the equivalent velocity component diagram is as Figure 3 shown. When the disconnect switch is in normal operation, the reaction force of the contact operating chamber is in balance with the output force of the operating mechanism. The output force of the operating mechanism can be indirectly judged by the winding current of the driving motor. By collecting and analyzing the winding current of the driving motor, the change of the load carried by the operating mechanism can be judged.

[0060] The reaction force of the contact operating chamber of the disconnect switch in GIS includes the friction force at each contact point of the operating mechanism, the friction force between the moving and static contacts, and the gravity of each component. During the operation of the disconnect switch, the equivalent mass of each component in the contact operating chamber reduced to point D is M D . The initial velocity of the moving contact is zero. According to the law of conservation of energy, the mathematical expression between the output force of the motor operating mechanism and the velocity of the moving contact is:

[0061]

[0062] In the formula, F is the component of the output force of the operating mechanism at point D, F D is the reaction force of the contact operating chamber, and due to the rotation of the toggle arm, both F and F D are variables; v d is the moving velocity of the moving contact, and l is the stroke of the intermediate guide rod.

[0063] The mechanical characteristics of the disconnect switch operating system are as follows:

[0064]

[0065] In the formula, F i is the output force of the operating mechanism, F j is the load reaction force of the operating mechanism, m i is the mass of each component of the operating mechanism, F lin is the reaction force at point A of the transmission tie rod between the mechanism phase and the non-mechanism phase, M D is the equivalent mass, l i is the initial stroke, l j is the final stroke, l m is the overtravel, η is the friction coefficient between each contact point of the operating mechanism, and g is the acceleration due to gravity.

[0066] During the opening distance stage, the motor operating mechanism is only affected by F D action, FD The reaction force at point A is:

[0067]

[0068] In the formula, θ0, θ1, and θ2 are the angular values set for calculating the components of F A , F B and F D respectively, as shown in Figure 2 .

[0069] During the overtravel stage, the operating mechanism is under the combined action of F D and F f , and the reaction force of F D at point A is equal to that in the opening stroke stage. The reaction force of the reaction force F f acting on point A is:

[0070]

[0071] In the formula, F f is the frictional resistance between the components of the arc extinguishing chamber, l1 is the length of the insulating pull rod CD, and l2 is the length of the toggle arm BC.

[0072] During the opening stroke stage, the rotation angle of the insulating pull rod is 0 to 46.6°, and the equivalent load torque of the disconnector in this stage is relatively small. When the disconnector operates to the just-separated and just-closed positions, due to the action of the spring fingers in the static contact cavity, the frictional force between the moving and static contacts changes suddenly. At this time, the reaction force of the operating mechanism reaches the maximum. Under normal operating conditions, the maximum torque of the driving motor is 25 N·m. As the driving motor continues to rotate, both the frictional resistance between the moving and static contacts and the force arm of the insulating pull rod decrease, so the equivalent load torque decreases accordingly. Under the action of the spring fingers, the torque of the driving motor in the overtravel stage is greater than that in the opening stroke stage.

[0073] The moment of inertia, as an important parameter of the driving motor, changes with the rotation of the motor spindle during the operation of the disconnector. According to the law of conservation of energy, the calculation formula of the moment of inertia during the opening and closing operations of the disconnector is as follows:

[0074]

[0075] v a = v a1 + v a2 (6)

[0076] In the formula, J o is the total moment of inertia, the angular velocity of the toggle arm OA is ω, the linear velocity of the toggle arm OA is v a , the mass of the transmission pull rod is m ln , the mass of the intermediate guide rod is m md , and the moment of inertia of the toggle arm OA is JOA The moment of inertia of the drive pull rod is J ln The moment of inertia of the crank arm BC is J BC The moment of inertia of the insulating pull rod is J ins The length of the drive pull rod is a. v a1 、v a2 and v b are all set speeds.

[0077] Collect the drive motor winding current data and the external crank arm rotational angular displacement data during the opening and closing processes of the disconnector in the normal operating state of the disconnector, including the opening current curve, closing current curve, opening angular displacement curve, and closing angular displacement curve, as Figure 4 shown.

[0078] Use an AC / DC current sensor to collect the drive motor winding current of the mechanism phase to obtain the drive motor winding current data; measure the rotational angle of the drive motor main shaft through an angular displacement sensor, and convert the rotational angle of the drive motor main shaft into the rotational angle of the external crank arm according to the mechanical structure relationship between the drive motor main shaft and the external crank arm to obtain the external crank arm rotational angular displacement data.

[0079] Step 2: The mechanical faults of the disconnector in the GIS can be simulated or the data during actual faults can be collected. Relatively speaking, simulating mechanical faults can collect fault data more conveniently and accurately, and the fault data is also more typical.

[0080] Simulate the mechanical faults of the disconnector in the GIS. The fault types of the mechanical faults of the disconnector mainly include various types such as rust on the shell end cover, misalignment of the static and moving contact axes, pin drop between phases A and B, pin drop between phases B and C, too low power supply voltage, and too high power supply voltage.

[0081] The simulation methods for various mechanical faults of the disconnector are as follows:

[0082] (1) The simulation method for rust on the shell end cover is to increase the torque of the drive motor during the entire opening and closing process of the disconnector; it is known that when the disconnector is operating without faults, the peak torque of the drive motor is about 25 N·m.

[0083] (2) The simulation method for misalignment of the static and moving contact axes is to keep the torque of the drive motor unchanged during the opening distance stage and increase the overtravel torque; to simulate the jamming fault caused by misalignment of the static and moving contact axes.

[0084] (3) The simulation method for pin drop between phases A and B is that pin drop causes part of the load of the drive motor to operate. By removing the pin of the drive pull rod between phases A and B, the drive motor drives phases B and C (mechanism phase) to operate.

[0085] (4) The simulation method for the pin detachment between phases B and C is to remove the drive rod pin between phase C and phase B, so that the drive motor only drives phase C for opening and closing operations.

[0086] (5) The simulation method for too low power supply voltage is to lower the power supply voltage. The rated voltage of the drive motor of the ZF16-252 angular disconnector is 220V at power frequency. By adjusting the power supply voltage to 180V, it is used to simulate that the power supply voltage of the drive motor is too low.

[0087] (6) The simulation method for too high power supply voltage is to raise the power supply voltage. Adjust the power supply voltage of the drive motor to 240V to simulate that the power supply voltage of the drive motor is too high.

[0088] Among them, the influence of power supply voltage fluctuation on the contact situation of moving and static contacts is very small, but it may affect the motor current and the opening and closing speed of the disconnector. When the load carried by the drive motor is certain, the opening and closing speed may increase with the increase of voltage, thus affecting the mechanical fault diagnosis result of the disconnector. For the drive motor of the disconnector in the GIS operating in the substation, its working voltage is not a stable value. Therefore, simulate the power supply voltage fluctuation of the drive motor and detect the winding current of the drive motor and the rotational angular displacement of the outer cranked arm when the non-fault disconnector operates at this voltage.

[0089] When simulating the mechanical faults of the disconnector, only one type of fault is simulated each time, and it is ensured that other types of faults do not occur, so as to ensure the accuracy of fault diagnosis. Collect the winding current data of the drive motor and the rotational angular displacement data of the outer cranked arm during the opening and closing processes of the disconnector when simulating the mechanical faults of the disconnector. The corresponding data collected for each mechanical fault is the same as the data types when the disconnector is operating normally, and also includes the opening current curve, closing current curve, opening angular displacement curve, and closing angular displacement curve, as Figures 5 - 10 shown.

[0090] Step 3: According to the winding current data of the drive motor and the rotational angular displacement data of the outer cranked arm collected in Step 1 and Step 2, analyze and obtain the characteristic data of the normal operating state of the disconnector and the characteristic data of the mechanical faults of the disconnector; use the MIC algorithm to calculate the weight of each feature of the characteristic data.

[0091] The opening operation of the disconnector in the GIS mainly includes four stages:

[0092] The first part: The starting stage of the moving contact opening. The drive motor drives the operating mechanism and the moving contact to do work against the static friction force. Since the starting point of each current signal acquisition is inconsistent, in order to facilitate its processing and analysis, the starting moment of the disconnector opening and closing is taken as the time zero point. When the output torque of the drive motor reaches equilibrium with the static friction force between the moving and static contacts and each operating component, the moving contact starts to move.

[0093] Second part: Tripping overtravel stage (moving and static contacts contacting stage). When the moving contact starts to move, the winding current rapidly decreases and then transitions to a steady operation stage. The running time and current value of the driving motor in this stage are mainly affected by the dynamic friction forces between the moving and static contacts and various transmission components. The current variation can be used to analyze the resistance situation between the moving and static contacts and various transmission components. An increase in resistance will cause the current passing through this position by the driving motor to increase and the running time to be prolonged.

[0094] Third part: Tripping opening distance stage. After the moving and static contacts separate, the driving motor continues to drive the moving contact to operate to form a reliable insulation gap. The winding current in this stage is mainly related to the damping torques of various transmission components of the three-phase operating mechanism. When the friction force increases due to rust between the transmission components, the winding current increases and the running time is prolonged. When the pin between the transmission tie rods breaks, causing the driving motor to only drive the mechanism phase and part of the non-mechanism phase to perform the tripping operation, the winding current and running time decrease accordingly.

[0095] Fourth part: Tripping end stage. At the end stage of the disconnector tripping, when the rotating main shaft of the driving motor collides with the limit device, the winding current rises slightly. At the same time, the control circuit of the driving motor is disconnected, and the winding current rapidly drops to zero, and the disconnector tripping ends.

[0096] Taking the extreme current values and their occurrence times in each stage as characteristic values, the tripping current characteristic values are shown in Table 1.

[0097] Table 1 Tripping current characteristic value table

[0098]

[0099]

[0100] t1 - t5 respectively represent the moment of the maximum current value in the moving contact tripping start stage, the moment of the minimum current value in the tripping overtravel stage, the moment of the maximum current value in the tripping opening distance stage, the moment of the maximum current value in the tripping end stage, and the tripping end moment. i1 - i4 respectively represent the maximum current value in the moving contact tripping start stage, the minimum current value in the tripping overtravel stage, the maximum current value in the tripping opening distance stage, and the maximum current value in the tripping end stage.

[0101] The closing operation of the disconnector in the GIS consists of four stages:

[0102] First part: Moving contact closing start stage. The driving motor drives the operating mechanism to start moving from a standstill. Through the winding current and running time in this stage, it is possible to diagnose whether there are any jams or other mechanical faults in each component.

[0103] Part II: Closing stroke stage. This is the stage where the moving contact runs from the open position to the position where the moving and static contacts just close. The frictional resistance between the transmission components of the operating mechanism can be judged during this stage.

[0104] Part III: Stage when the moving and static contacts just close. Under the action of the spring finger pressure in the static contact cavity, the resistance at the position where the moving and static contacts just close increases significantly, and the winding current increases accordingly.

[0105] Part IV: Closing overtravel stage. This is the contact stage of the moving and static contacts. The winding current and operating time during this stage can reflect the comprehensive force conditions between the moving and static contacts and the operating mechanism.

[0106] The characteristic values of the winding current during the closing process of the disconnector in GIS are shown in Table 2. Among them, t6 - t9 respectively represent the moments of the maximum value of the current during the starting stage of the moving contact closing, the moment of the maximum value of the current during the stage when the moving and static contacts just close, the moment of the minimum value of the current during the closing overtravel stage, and the closing end moment. i6 - i8 respectively represent the maximum value of the current during the starting stage of the moving contact closing, the maximum value of the current during the stage when the moving and static contacts just close, and the minimum value of the current during the closing overtravel stage.

[0107] Table 2 Characteristic values of closing current

[0108]

[0109]

[0110] In the angular displacement curve, based on the mathematical relationship between the stroke of the moving contact of the disconnector and the rotation angle of the crank arm, the contact situation of the disconnector during operation can be judged, and the starting moment of the moving contact, the moments of just opening and just closing, and the moments of opening and closing in place can be extracted and used as the characteristic values of the angular displacement signal, as shown in Tables 3 and 4. The moment of just opening is the moment when the moving and static contacts just separate during the opening process.

[0111] Table 3 Characteristic values of opening angular displacement

[0112] Operating status <![CDATA[t 11 > <![CDATA[t 12 > <![CDATA[t 13 > Normal operating status 0.12 1.71 4.64 Power supply voltage too low 0.14 1.73 4.71 Power supply voltage too high 0.09 1.7 4.65 Pin between phases A and B dropped off 0.13 1.71 4.65 Pin between phases B and C dropped off 0.09 1.7 4.65 Outer shell end cover rusted 0.3 2.22 5.38 Moving and static contact axes not aligned 0.25 2.38 5.14

[0113] t 11 -t 13 respectively represent the starting moment of the moving contact opening, the moment when the moving and static contacts just separate, and the moment when the moving contact opening is in place.

[0114] Table 4 Characteristic values of closing angular displacement

[0115] Operating status <![CDATA[t 14 > <![CDATA[t 15 > <![CDATA[t 16 > Normal operating status 0.12 3.07 4.91 Power supply voltage too low 0.14 3.07 4.98 Power supply voltage too high 0.08 3.01 4.84 Pin between phases A and B dropped off 0.11 3.07 4.92 Pin between phases B and C dropped off 0.08 3.03 4.88 Outer shell end cover rusted 0.11 3.34 5.42 Moving and static contact axes not aligned 0.11 3.07 5.6

[0116] t 14 -t 16 respectively represent the starting moment of the moving contact closing, the moment when the moving and static contacts just close, and the moment when the moving contact closing is in place.

[0117] From the drive motor winding current data and the outer link rotation angular displacement data in Steps 1 and 2, a total of 22 types of characteristic data listed in Tables 1 - 4 are analyzed and obtained, forming the characteristic data of the normal operating state of the disconnector and the characteristic data of the mechanical faults of the disconnector respectively.

[0118] The essence of the mechanical fault diagnosis of the disconnector in GIS is a problem of data classification based on the current characteristic values and angular displacement characteristic values under different operating states of the disconnector. In the original feature set, there are not only positive correlation features that are beneficial to data classification, but also negative correlation features that affect the classification accuracy rate, as well as irrelevant features that affect the operation efficiency. Therefore, it is necessary to screen the high-dimensional data features.

[0119] This embodiment uses the MIC algorithm to calculate the weight of each feature in the characteristic data. MIC is a feature selection method based on information criteria, used to measure the correlation degree between two variables and realize the calculation of the sample feature weights. The accuracy of MIC is better than that of mutual information, and at the same time, MIC has the characteristics of universality, fairness, and symmetry. The calculation steps of the MIC algorithm are as follows:

[0120] Calculate the maximum mutual information value:

[0121]

[0122] In the formula, p(X; Y) is the joint probability of variables X and Y; p(X) is the probability of variable X; p(Y) is the probability of variable Y;

[0123] Normalize the maximum mutual information value:

[0124]

[0125]

[0126] In the formula, B is a variable.

[0127] Step 4: Select several dimensions of characteristic data with the largest weight values from the calculation results of the MIC algorithm and input them into the decision tree model, use the random forest algorithm for classification, train the decision tree model, determine the most suitable number of dimensions of the characteristic data, and set the optimal number of decision trees. The classification results include the normal operating state and 6 types of fault types.

[0128] Random forest is an ensemble learning algorithm based on decision trees. Taking decision trees as the base learners, multiple tree models are created simultaneously, and the decision results of all tree models are integrated together, which is the final classification result. Compared with a single decision tree, random forest has randomness and diversity, and effectively overcomes the defect of overfitting of decision trees. There are mainly two methods, Bootstrap and out-of-bag estimate, in the construction process of random forest. In random forest, if all decision trees have the same model, their sample processing results are the same, and at this time, the classification result of random forest is the same as that of a single decision tree. To improve the difference between decision trees, the Bootstrap algorithm is used to draw multiple sample sets with a size of N in a with-replacement manner for training the decision tree model. The random with-replacement drawing method can avoid the overfitting phenomenon caused by completely different sample sets. There may be a phenomenon of data aliasing of certain categories in the sample set, resulting in a small difference between decision trees. To address this problem, the present invention improves its diversity by selecting different features in the tree model. The selection principle of feature nodes is the same as the sampling method of the decision tree training set, and it is also drawn in a random with-replacement manner. Suppose the dimension of the sample feature set is K, and a feature subset with a dimension of k is selected from K, where k << K. The optimal feature is selected from this feature subset for training the decision tree, and this process is repeated n times to complete the selection of feature nodes of n different decision trees.

[0129] After mathematical morphology preprocessing, the high-frequency burrs and abnormal spikes in the measured current curve are basically eliminated. By analyzing the winding current and angular displacement signals, the current extreme values and their occurrence times at different stages are used as feature values, and the starting moment of the moving contact, the moment of just opening and closing, and the moment of ending opening and closing are used as angular displacement feature values.

[0130] When diagnosing the mechanical faults of disconnectors in GIS based on the random forest method, for the convenience of data processing, the normal and various simulated mechanical fault operating states are respectively represented by data labels from 1 to 7. The above-mentioned 5-dimensional features with higher weights are imported into the program to construct a random forest classification model. Since the classification error decreases with the increase in the number of decision trees, when the number increases to a certain extent, the out-of-bag error rate no longer decreases and fluctuates within a certain range. Therefore, the number of decision trees when the out-of-bag error rate is first reduced to the lowest is taken as the optimal number. If the number of decision trees in the random forest is greater than this value, its classification accuracy may decrease, and the classification efficiency of the overall sample may decrease due to the increase in computational complexity. When diagnosing mechanical faults through the disconnector opening current, the number of decision trees when the out-of-bag error rate is first reduced to the lowest is 41. When the number of features used to divide internal nodes is different, the classification errors of the constructed decision trees are different. Figure 11 It is the classification error graph of decision trees under different numbers of feature nodes.

[0131] Step 5: Use the trained decision tree model to perform mechanical fault detection on the GIS disconnector, determine whether there is a mechanical fault in the disconnector, and the type of mechanical fault.

[0132] After the data samples traverse all the feature nodes of the decision tree, each leaf node should only contain samples of the same class, and all leaf nodes should cover all the operating states of the disconnectors in the GIS. At this time, the accuracy rate of the random forest is 100%. However, in the actual classification process, due to parameter settings such as the number of decision trees and feature selection, some leaf nodes contain two or more types of data, thus affecting the accuracy rate of the random forest.

[0133] To verify the accuracy of this embodiment, 5 groups of data are included for each operating state of the disconnector. The fault diagnosis results are as Figure 12 shown, Figure 12 In the figure, 1-7 respectively represent the normal operating state, too low supply voltage, too high supply voltage, pin drop between phases A and B, pin drop between phases B and C, corrosion of the outer shell end cover, and misalignment of the axes of the moving and static contacts. In the classification results, one group of normal data is misclassified as too high voltage, one group of too low voltage is misclassified as normal, and one group of too high voltage is misclassified as normal. The accuracy rate of the mechanical fault diagnosis based on the opening current is 91.42%.

[0134] This embodiment also includes a GIS disconnector mechanical fault detection system, including

[0135] AC and DC current sensors, which are used to collect the drive motor winding current data during the opening and closing processes of the disconnector.

[0136] Angular displacement sensors, which are used to collect the rotation angle of the drive motor spindle during the opening and closing processes of the disconnector.

[0137] A storage module, which is used to store the characteristic data of the normal operating state of the disconnector and the characteristic data of the mechanical faults of the disconnector.

[0138] The characteristic data weight calculation module uses the rotation angle of the drive motor spindle during the opening and closing processes of the disconnector, and according to the mechanical structure relationship between the drive motor spindle and the outer toggle arm, converts the rotation angle of the drive motor spindle into the rotation angle of the outer toggle arm to obtain the rotation angular displacement data of the outer toggle arm.

[0139] The characteristic data weight calculation module is used to analyze and obtain the characteristic data to be judged of the disconnector based on the rotation angular displacement data of the outer toggle arm and the drive motor winding current data, combine the characteristic data of the normal operating state of the disconnector and the characteristic data of the mechanical faults of the disconnector stored in the storage module, and use the MIC algorithm to calculate the weight of each feature of the characteristic data.

[0140] A classification module, which is used to input several dimensions of features with higher weights into a trained decision tree model according to the calculation result of the feature data weight calculation module, and use the trained decision tree model to perform GIS disconnector mechanical fault detection by using the random forest algorithm, determine whether there is a mechanical fault in the disconnector, and directly determine the fault type of the mechanical fault if there is a mechanical fault.

[0141] For disconnectors in GIS of different models, the weights of each feature of the feature data calculated by the MIC algorithm may be different; the number of dimensions of the features input into the decision tree model may also be different.

[0142] This embodiment also includes a computer-readable storage medium, including a program that can be executed by a processor to implement the above-mentioned method for detecting mechanical faults of GIS disconnectors.

[0143] The above has described the present invention in detail through embodiments, but the above content is only an exemplary embodiment of the present invention and cannot be considered as limiting the implementation scope of the present invention. The protection scope of the present invention is defined by the claims. Those who use the technical solutions described in the present invention, or those skilled in the art inspired by the technical solutions of the present invention, within the essence and protection scope of the present invention, design similar technical solutions to achieve the above technical effects, or make equivalent changes and improvements to the scope of the application, etc., should still fall within the patent coverage protection scope of the present invention.

Claims

1. A method for detecting mechanical faults of GIS disconnect switches, characterized by: It includes the following steps: Collect the drive motor winding current data and the external link arm rotation angular displacement data during the opening and closing processes of the disconnector in its normal operating state. Collect the drive motor winding current data and the external link arm rotation angular displacement data during the opening and closing processes of the disconnector in case of mechanical failure. Based on the drive motor winding current data and the external link arm rotation angular displacement data collected during the opening and closing processes of the disconnector in its normal operating state and in case of mechanical failure, analyze to obtain the characteristic data of the disconnector in its normal operating state and the characteristic data of the disconnector mechanical failure. Use the MIC (Maximum Information Coefficient) algorithm to calculate the weight of each feature of the characteristic data. Select several dimensions of characteristic data with the largest weight values from the calculation results of the MIC algorithm and input them into the decision tree model, and use the random forest algorithm for classification to train the decision tree model. Use the trained decision tree model to detect the mechanical failure of the GIS disconnector, and judge whether there is a mechanical failure of the disconnector and the type of mechanical failure.

2. The mechanical fault detection method of a GIS disconnector according to claim 1, characterized in that: Use an AC / DC current sensor to collect the drive motor winding current of the mechanism phase to obtain the drive motor winding current data; measure the rotation angle of the drive motor main shaft through an angular displacement sensor, and convert the rotation angle of the drive motor main shaft into the rotation angle of the external link arm according to the mechanical structure relationship between the drive motor main shaft and the external link arm to obtain the external link arm rotation angular displacement data.

3. A mechanical fault detection method for a GIS disconnector according to claim 1, characterized in that: The types of mechanical failures of the disconnector include various ones such as rust on the shell end cover, misalignment of the moving and static contact axes, pin detachment between phases A and B, pin detachment between phases B and C, too low power supply voltage, and too high power supply voltage.

4. The mechanical fault detection method of a GIS disconnector according to claim 3, characterized in that: The mechanical failure of the disconnector is realized through simulation, and only one type of failure is simulated each time.

5. A method for detecting the mechanical failure of a GIS disconnector according to claim 4, characterized in that: The simulation method for rust on the shell end cover is to increase the torque of the drive motor during the entire opening and closing process of the disconnector. The simulation method for misalignment of the moving and static contact axes is to keep the torque of the drive motor unchanged during the opening distance stage and increase the overtravel torque. The simulation method for pin detachment between phases A and B is that the pin detachment causes the drive part of the drive motor to operate with load. By removing the pin of the transmission tie rod between phase A and phase B, the drive motor drives phase B and phase C to operate. ​ ​ ​ 6. The mechanical fault detection method of a GIS disconnector according to claim 1, wherein: ​ ​ The moment of maximum value of the current during the closing start stage of the moving contact, the moment of maximum value of the current during the just-closed stage of the moving and static contacts, the moment of minimum value of the current during the closing overtravel stage, and the closing end moment; the maximum value of the current during the closing start stage of the moving contact, the maximum value of the current during the just-closed stage of the moving and static contacts, and the minimum value of the current during the closing overtravel stage; The opening start moment of the moving contact, the just-separated moment of the moving and static contacts, and the opening-in-place moment of the moving contact; The closing start moment of the moving contact, the just-closed moment of the moving and static contacts, and the closing-in-place moment of the moving contact.

7. The mechanical fault detection method for a GIS disconnector according to claim 1, characterized in that: When using the random forest algorithm for classification, the Bootstrap algorithm is adopted to extract multiple sample sets with a quantity of N in a with-replacement manner to train the decision tree model; the selection principle of the feature nodes is also to extract in a random with-replacement manner.

8. A mechanical fault detection system for GIS disconnectors, characterized in that: including An AC / DC current sensor for collecting the drive motor winding current data during the opening and closing processes of the disconnector; An angular displacement sensor for collecting the rotation angle of the drive motor main shaft during the opening and closing processes of the disconnector; A storage module for storing the characteristic data of the normal operating state of the disconnector and the characteristic data of the mechanical faults of the disconnector, A characteristic data weight calculation module for calculating the weight of each feature of the characteristic data by using the MIC algorithm according to the data collected by the AC / DC current sensor and the angular displacement sensor, and the characteristic data stored in the storage module; A classification module for performing GIS disconnector mechanical fault detection by using the trained decision tree model with the random forest algorithm according to the calculation result of the characteristic data weight calculation module, and judging whether there are mechanical faults in the disconnector and the fault types of the mechanical faults.

9. The mechanical fault detection system for a GIS disconnector according to claim 8, wherein: The characteristic data weight calculation module uses the rotation angle of the drive motor main shaft during the opening and closing processes of the disconnector, and according to the mechanical structure relationship between the drive motor main shaft and the outer toggle arm, converts the rotation angle of the drive motor main shaft into the rotation angle of the outer toggle arm to obtain the rotation angular displacement data of the outer toggle arm.

10. A computer-readable storage medium, characterized in that: Including a program that can be executed by a processor to implement a GIS disconnector mechanical fault detection method according to any one of claims 1-7.

Images