High-low voltage switch cabinet state evaluation method and system based on multi-dimensional features and AI recognition

Abstract

This application provides a method and system for assessing the condition of high and low voltage switchgear based on multi-dimensional features and AI recognition. The method acquires mechanical characteristic parameter data, calculates the degree of parameter deviation, the completeness of actions, and the degree of coordination of actions of the switchgear during opening and closing operations, and determines the mechanical health score of the switchgear. Based on the mechanical characteristic parameter data, it calculates the shock resistance, fluctuation resistance, and self-recovery degree of the switchgear during opening and closing operations, and determines the toughness score of the switchgear. Based on the mechanical health score and the toughness score, the mechanical state of the switchgear is determined, and this mechanical state is used as the assessment result of the mechanical actions of the switchgear in high and low voltage switchgear. This application effectively overcomes the shortcomings of single-parameter assessment methods, which can only reflect a certain local state information of the switchgear and have insufficient ability to identify early deterioration characteristics.

Description

Technical Field

[0001] This application relates to the field of intelligent manufacturing technology, and in particular to a method and system for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition. Background Technology

[0002] High and low voltage switchgear is the core set of equipment in the power system for receiving, distributing, controlling and protecting electrical energy. The circuit breakers and other switching equipment configured inside it undertake the key task of normal switching and connection between different power supply circuits. Whether the above-mentioned switching equipment can operate reliably directly determines the continuity and safety of the power supply system.

[0003] The reliable operation of a circuit breaker fundamentally depends on the health of its mechanical condition. In actual long-term operation, due to the combined effects of accumulated operation frequency, changes in ambient temperature and humidity, vibration, and material aging, the mechanical components of the circuit breaker will inevitably experience performance degradation. For example, the energy storage mechanism may jam, resulting in insufficient energy storage or prolonged energy storage time; the transmission linkage may loosen, causing energy loss and travel deviation during the action transmission process; the contact spring may fatigue and the elastic coefficient may decrease, resulting in insufficient contact pressure or uneven local stress between the moving and stationary contacts. The above-mentioned mechanical performance degradation will not only directly affect the opening and closing characteristics of the circuit breaker, leading to switching and connection failures, but will also cause a series of chain problems such as increased contact resistance, abnormal local heating, and accelerated insulation aging. In severe cases, it may even lead to major accidents such as contact burnout, phase-to-phase short circuits, and electrical fires, threatening the safe and stable operation of the power grid.

[0004] Currently, methods for assessing the mechanical condition of switchgear include vibration signal analysis, coil current waveform analysis, integrated data-driven analysis (IDDA) models, and tensor fusion methods. Among these, vibration signal analysis and coil current waveform analysis assess the mechanical condition of switchgear using a single type of parameter. However, a single type of monitoring parameter can only reflect the state information of a certain part or aspect of the mechanical system, making it difficult to accurately reflect the mechanical health status of the circuit breaker. This results in insufficient ability to identify early deterioration characteristics, and the accuracy and reliability of fault diagnosis cannot meet the needs of actual engineering. Integrated data-driven analysis models and tensor fusion methods are health assessment models built based on multi-dimensional monitoring data. Relying solely on the health assessment model to assess the mechanical condition of the circuit breaker lacks physical mechanism support, has poor model interpretability, and insufficient dynamic adaptability.

[0005] Therefore, it is crucial to achieve comprehensive perception and evaluation of the mechanical condition of circuit breakers through various types of monitoring parameters, thereby improving the accuracy of mechanical condition assessment. Summary of the Invention

[0006] In view of the shortcomings of the prior art described above, this application provides a method and system for evaluating the status of high and low voltage switchgear based on multi-dimensional features and AI recognition to solve the above-mentioned technical problems.

[0007] According to one aspect of the embodiments of this application, a method for evaluating the status of high- and low-voltage switchgear based on multi-dimensional features and AI recognition is provided. The method includes: acquiring mechanical characteristic parameter data generated by switchgear in the high- and low-voltage switchgear during the execution of opening and closing actions; the mechanical characteristic parameter data includes: stroke curve, opening and closing time, bounce, synchronization difference, and overtravel; the switchgear includes: a circuit breaker; calculating, based on the mechanical characteristic parameter data, the parameter deviation degree, action integrity degree, and action coordination degree of the switchgear during the execution of opening and closing actions; and based on the parameter deviation degree, the action integrity degree, and the action coordination degree... The mechanical health score of the switchgear is determined based on the mechanical characteristic parameter data. The shock resistance, fluctuation resistance, and self-recovery of the switchgear during opening and closing operations are calculated. Based on these parameters, a resilience score is determined. This resilience score characterizes the dynamic response capability of the switchgear during opening and closing operations. The mechanical state of the switchgear is determined based on the mechanical health score and the resilience score, and this mechanical state is used as the evaluation result of the switchgear's mechanical operation in the high and low voltage switchgear.

[0008] According to another aspect of the embodiments of this application, a high- and low-voltage switchgear status assessment system based on multi-dimensional features and AI recognition is also provided, comprising: a data acquisition module, used to acquire mechanical characteristic parameter data generated by the switchgear in the high- and low-voltage switchgear during the execution of opening and closing actions; the mechanical characteristic parameter data includes: stroke curve, opening and closing time, bounce, synchronization difference, and overtravel; the switchgear includes: a circuit breaker; and a health score calculation module, used to calculate, based on the mechanical characteristic parameter data, the parameter deviation degree, action integrity degree, and action coordination degree of the switchgear during the execution of opening and closing actions; and based on the parameter deviation degree, the action integrity degree, and the action coordination degree... The system comprises a mechanical health rating module for the switchgear, a resilience rating calculation module for calculating the shock resistance, fluctuation resistance, and self-recovery of the switchgear during opening and closing operations based on the mechanical characteristic parameter data, and determining the resilience rating of the switchgear based on the shock resistance, fluctuation resistance, and self-recovery. The resilience rating characterizes the dynamic response capability of the switchgear during opening and closing operations. A state determination module determines the mechanical state of the switchgear based on the mechanical health rating and the resilience rating, and uses the mechanical state as the evaluation result of the switchgear's mechanical operation in the high and low voltage switchgear.

[0009] The beneficial effects of this application are as follows: This application obtains mechanical characteristic parameter data of switchgear in high and low voltage switchgear during the execution of opening and closing operations. Based on the mechanical characteristic parameter data, it calculates the parameter deviation, operational integrity, and operational coordination of the switchgear during the execution of opening and closing operations. Based on these parameters, it determines the mechanical health score of the switchgear. Based on the mechanical characteristic parameter data, it calculates the shock resistance, fluctuation resistance, and self-recovery of the switchgear during the execution of opening and closing operations. Based on these parameters, it determines the toughness score of the switchgear. Based on the mechanical health score and toughness score, it determines the mechanical state of the switchgear. The mechanical state is used as the evaluation result of the mechanical operation of the switchgear in high and low voltage switchgear. The above process obtains the mechanical characteristic parameter data of switchgear during the execution of opening and closing operations. The mechanical characteristic parameter data generated in the process comprehensively assesses the mechanical health of switchgear from three dimensions: parameter deviation, operational integrity, and operational coordination. It also assesses the dynamic response resilience of switchgear from three dimensions: shock resistance, fluctuation resistance, and self-recovery. The mechanical health score and resilience score are combined to determine the mechanical condition of the switchgear. This achieves comprehensive perception and evaluation of multiple types of monitoring parameters, effectively overcoming the shortcomings of single-parameter evaluation methods that can only reflect a local state of the switchgear and lack the ability to identify early deterioration characteristics. Simultaneously, it avoids the problems of pure data-driven models lacking physical mechanism support, poor interpretability, and insufficient dynamic adaptability. This improves the accuracy, reliability, and interpretability of switchgear mechanical condition assessment, providing a scientific basis for preventative maintenance and fault early warning of high and low voltage switchgear, and effectively ensuring the continuity and safety of power system operation.

[0010] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0011] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:

[0012] Figure 1 This is a flowchart illustrating an exemplary embodiment of the present application of a method for assessing the status of high and low voltage switchgear based on multidimensional features and AI recognition;

[0013] Figure 2This is a flowchart illustrating an exemplary embodiment of the present application, showing how to determine the toughness score of a switchgear based on its shock resistance, fluctuation resistance, and self-recovery.

[0014] Figure 3 This is an exemplary embodiment of the present application illustrating a flowchart of determining the mechanical condition of a switchgear based on a mechanical health score and a toughness score;

[0015] Figure 4 This is an exemplary embodiment of the present application illustrating the operation interface diagram of high and low voltage switchgear status assessment based on multi-dimensional features and AI recognition;

[0016] Figure 5 This is a block diagram illustrating a high- and low-voltage switchgear condition assessment system based on multi-dimensional features and AI recognition, as shown in an exemplary embodiment of this application. Detailed Implementation

[0017] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0018] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0019] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the present application. However, it will be apparent to those skilled in the art that embodiments of the present application may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the present application.

[0020] The implementation details of the technical solutions in the embodiments of this application are described in detail below:

[0021] Figure 1 This is a flowchart illustrating an exemplary embodiment of the high- and low-voltage switchgear condition assessment method based on multi-dimensional features and AI recognition. (Refer to...) Figure 1As shown, the high- and low-voltage switchgear status assessment method based on multi-dimensional features and AI recognition includes at least steps S110 to S140, which are described in detail below:

[0022] In step S110, mechanical characteristic parameter data generated by the switching equipment in the high and low voltage switchgear during the opening and closing operations are acquired. In one embodiment of this application, the switching equipment includes a circuit breaker; the mechanical characteristic parameter data includes: stroke curve, opening and closing time, bounce, synchronization difference, overtravel, etc. The stroke curve is the relationship curve of the displacement of the moving contact of the switch changing with time during the entire opening and closing movement (the horizontal axis is time, and the vertical axis is the stroke displacement). The slope of the relationship curve reflects the movement speed of the moving contact of the switch, the sudden change in slope reflects the jamming or buffering abnormality of the mechanism, the plateau segment of the relationship curve reflects the overtravel compression process after the moving contact of the switch, and the oscillation segment of the relationship curve reflects the closing bounce or opening rebound. The real-time displacement data of the moving contact of the switch is directly acquired by a displacement sensor, and the corresponding timestamp is recorded synchronously to obtain timestamp-displacement data. The stroke curve is plotted using the timestamp-displacement data. The displacement data is acquired synchronously at a sufficiently high sampling rate (e.g., a sampling rate of not less than 10kHz) to ensure that short-term dynamic processes such as closing bounce, opening rebound, and buffering abnormality can be distinguished. Displacement measurement can be performed using linear displacement sensors, rotation angle sensors, etc.; alternatively, acceleration signals can be collected by accelerometers and the displacement can be indirectly obtained through integration. The opening and closing times include: opening time and closing time. The opening time is the time interval from the moment the circuit breaker receives the opening command (the opening coil is energized) to the moment all pole contacts separate. The opening time reflects the speed at which the circuit breaker clears the fault current; the shorter the time, the better the protection effect. The closing time is the time interval from the moment the circuit breaker receives the closing command (the closing coil is energized) to the moment all pole contacts make contact. The closing time reflects the response speed of the circuit breaker upon activation. The acquisition process for the opening (or closing) time includes: capturing the energization moment of the opening coil (or closing coil) through Hall sensor or coil voltage detection; starting a high-precision timer (usually based on a DSP (Digital Signal Processor) or microcontroller, with a time resolution of up to 0.1ms); after the timer starts, synchronously sampling the break signal (or transient signal, displacement signal, etc.) at high speed; identifying the end point of the contact action through signal processing algorithms (e.g., short-time energy method, change-point detection method, high-frequency signal acquisition method, etc.); and calculating the time difference between the end point and the start point to obtain the opening (or closing) time. The break signal can be used to identify the moment of change in the on / off state of the moving and stationary contacts; the transient signal can be used to identify the transient characteristics generated at the moment of contact or separation of the moving and stationary contacts; and the displacement signal can be used to identify the key position changes during the movement of the moving contact.Bounce characterizes the phenomenon of repeated separation and contact between the moving and stationary contacts of a switch during the closing operation, caused by mechanical impact and incomplete dissipation of kinetic energy. Bounce includes: the number of bounces, bounce duration, and bounce amplitude. The number of bounces represents the number of times the moving and stationary contacts repeatedly separate (close) during the closing process. The bounce duration represents the time from the first closing of the moving and stationary contacts to the end of the last bounce (i.e., the contacts are stably closed and no longer separate). The bounce amplitude represents the maximum displacement value of the moving contact rebounding from the closed position during the closing operation. The calculation and acquisition of the number of bounces are achieved through methods such as the break-point signal method, displacement sensor method, and vibration sensor method. The calculation and acquisition of the bounce duration are achieved through methods such as the break-point signal method and displacement sensor method. The bounce amplitude is extracted from the stroke curve. The process involves differentiating the stroke curve to obtain the velocity curve. The moment the velocity curve abruptly changes from a positive value to zero (or the difference from zero is less than a difference threshold) is the moment the moving contact first reaches the closing position. The displacement value at this moment is the reference position for the bounce amplitude. Within the bounce interval (from the first closing to the last closing), the minimum displacement value in the stroke curve is found and taken as the farthest position the moving contact bounces back to during the bounce process. The difference between the reference position and the minimum displacement value is recorded as the bounce amplitude. Synchronization difference characterizes the synchronicity of the three-phase contact actions, including: inter-phase synchronization difference and intra-phase synchronization difference. Inter-phase synchronization difference represents the difference between the maximum and minimum closing times or the maximum and minimum opening times among the three phases. Intra-phase synchronization difference refers to the difference between the maximum and minimum closing times or the maximum and minimum opening times among multiple breaks within the same phase. Intra-phase synchronization difference is applicable to circuit breaker structures with multiple breaks within the same phase. Overtravel (also known as contact travel or overtravel) refers to the extra distance that the operating mechanism continues to push the moving contact forward after the moving and stationary contacts have made contact during the closing operation (at which time the contact pressure spring is compressed). Overtravel is extracted from the travel curve. The process of extracting overtravel includes: determining the point where the moving and stationary contacts just close by using the abrupt change point of the slope of the travel curve, the characteristic point of speed change, or the moment of on / off change of the break signal; determining the final travel of the moving contact at its final stopping position; and the displacement difference between the final travel and the travel at the point of just close is taken as overtravel.

[0023] In step S120, based on the mechanical characteristic parameter data, the degree of parameter deviation, the degree of action integrity, and the degree of action coordination of the switchgear during the opening and closing operation are calculated; and based on the degree of parameter deviation, the degree of action integrity, and the degree of action coordination, the mechanical health score of the switchgear is determined. In one embodiment of this application, the degree of parameter deviation is calculated based on the stroke curve, opening and closing time, bounce, and overtravel in the mechanical characteristic parameter data. The calculation formula for the degree of parameter deviation is as follows:

[0024] Equation (1)

[0025] in, Indicates the degree of parameter deviation. Indicates the first The weights of different types of mechanical characteristic parameter data Indicates the first The degree of deviation of the mechanical characteristic parameter data of each type .

[0026] No. The formula for calculating the deviation of the mechanical characteristic parameter data for each type is as follows:

[0027] Equation (2)

[0028] in, Indicates the first The degree of deviation of the mechanical characteristic parameter data of each type Indicates the first Types of mechanical characteristic parameter data, Indicates the first Reference values ​​for various types of mechanical characteristic parameters; the deviation of opening and closing time, the deviation of moving contact speed during opening and closing, the deviation of overtravel, and the deviation of bounce duration can all be calculated using formula (2). The moving contact speed during opening and closing is obtained by differentiating the stroke curve, or by using differential calculation for the corresponding discrete data. The reference values ​​for opening and closing time, moving contact speed during opening and closing, and overtravel are all rated values ​​(not zero), and the reference value for bounce duration is a recommended value (e.g., 1ms~2ms).

[0029] The formula for calculating the completeness of a movement is as follows:

[0030] Equation (3)

[0031] in, Indicates the degree of completeness of the action. This indicates the number of parameter items in the mechanical characteristic parameter data that meet the judgment criteria. The judgment criteria are whether the mechanical characteristic parameter data is within a specified range, or whether the mechanical characteristic parameter data is within a range determined by a limit value. This indicates the total number of parameters that need to be checked. During the opening process, the parameters that need to be checked include: opening time, initial opening speed, stroke, overtravel, and synchronization difference. During the closing process, the parameters that need to be checked include: closing time, initial closing speed, stroke, overtravel, bounce time, and synchronization difference of the three phases. The instantaneous opening and closing speeds can be extracted from the travel curves. The process involves: finding the maximum travel value after the travel stabilizes on the closing travel curve, taking it as the closing endpoint; measuring a distance equal to the overtravel value from the closing endpoint in the direction of decreasing travel, obtaining the travel point as the instantaneous closing point (the position where the moving and stationary contacts just touch during closing); and taking the average speed within a preset time period or preset distance before the instantaneous closing point as the instantaneous closing speed. Similarly, finding the minimum travel value after the travel stabilizes on the opening travel curve, taking it as the opening endpoint; measuring a distance equal to the overtravel value from the opening endpoint in the direction of increasing travel, obtaining the travel point as the instantaneous opening point (the position where the moving and stationary contacts just separate during opening); and taking the average speed within a preset time period or preset distance after the instantaneous opening point as the instantaneous opening speed. Travel refers to the total distance traveled by the moving contact from the fully open position to the fully closed position, including the minimum safe distance when the moving and stationary contacts separate and the overtravel. Its main function is to ensure that a sufficiently long insulation gap can be formed between the moving and stationary contacts in the open state to reliably extinguish the arc.

[0032] The formula for calculating the degree of coordination of movements is as follows:

[0033] Equation (4)

[0034] in, Indicates the degree of coordination of movements. Indicates the reference time base (e.g., the manufacturer's specified rated operating time or synchronization difference limit). This represents the maximum value of the difference between the same period and the same period.

[0035] In one embodiment of this application, when the synchronization difference is the phase-to-phase synchronization difference, formula (4) calculates the phase-to-phase coordination degree, and the average synchronization difference is the average of the phase-to-phase synchronization differences; when the synchronization difference is the same-phase synchronization difference, formula (4) calculates the same-phase coordination degree, and the average synchronization difference is the average of the same-phase synchronization differences. For a three-phase circuit breaker, the same-phase synchronization differences of phases A, B, and C can be calculated separately, resulting in three values.

[0036] In one embodiment of this application, after obtaining the coordination degree of interphase actions and the coordination degree of three in-phase actions, the minimum coordination degree among the interphase actions and the three in-phase actions is selected as the final coordination degree; or, the average value of the three in-phase actions is calculated, and the average value is weighted and summed with the interphase actions to obtain the final coordination degree; or, the minimum value among the three in-phase actions is selected, and the minimum value is weighted and summed with the interphase actions to obtain the final coordination degree.

[0037] In step S130, based on the mechanical characteristic parameter data, the shock resistance, fluctuation resistance, and self-recovery degree of the switchgear during the opening and closing operation are calculated; and based on the shock resistance, fluctuation resistance, and self-recovery degree, the toughness score of the switchgear is determined. In one embodiment of this application, the toughness score is used to characterize the dynamic response capability of the circuit breaker during the opening and closing operation. The formula for calculating the shock resistance is as follows:

[0038] Equation (5)

[0039] in, Indicates the degree of impact resistance. Indicates speed weight, Represents the first velocity factor. Indicates the bounce weight. Indicates the bounce factor. Indicates the overrange weight. The overtravel factor is represented by the sum of the speed weight, bounce weight, and overtravel weight, which equals 1. Formula (5) is used to measure the circuit breaker's ability to withstand mechanical and electrodynamic shocks at the moment of closing, and is calculated based on the mechanical characteristic parameter data during the closing process.

[0040] In one embodiment of this application, the formula for calculating the first velocity factor is as follows:

[0041] Equation (6)

[0042] in, Represents the first velocity factor. Indicates the net velocity. The reference value for the engagement speed is set by the manufacturer. The first speed factor characterizes the degree of matching between the engagement speed and the reference value. The larger the first speed factor, the higher the degree of matching between the engagement speed and the reference value. The value range of the first speed factor is [0-1]. When the first speed factor exceeds [0-1], truncation is required. That is, if the first speed factor is greater than the upper limit of the value range, the upper limit of the value range is used as the first speed factor; if the first speed factor is less than the lower limit of the value range, the lower limit of the value range is used as the first speed factor.

[0043] The formula for calculating the bounce factor is as follows:

[0044] Equation (7)

[0045] in, Indicates the bounce factor. Indicates the duration of the bounce. This represents the maximum bounce duration. The bounce factor is inversely proportional to the bounce duration; a larger bounce factor results in a smaller bounce duration, indicating better performance. The bounce factor ranges from [0-1]. When the bounce factor exceeds [0-1], it needs to be truncated.

[0046] The formula for calculating the overtravel factor is as follows:

[0047] Equation (8)

[0048] in, Indicates the overrange factor. Indicates exceeding the travel limit. The reference value for overtravel is set by the manufacturer. The overtravel factor characterizes the degree of matching between the overtravel and the reference value. The larger the overtravel factor, the higher the degree of matching between the overtravel and the reference value. The overtravel factor ranges from [0-1]. When the overtravel factor exceeds [0-1], truncation is required.

[0049] The formula for calculating the degree of resistance to fluctuations is as follows:

[0050] Equation (9)

[0051] in, Indicates the degree of resistance to fluctuations. Indicates the second velocity weight. Indicates the second velocity factor. Indicates the velocity of the first component. Indicates time weighting, Indicates the time factor. Indicates the weight of consistency. It represents the synchronization factor (i.e., the degree of coordination and consistency of the final action). The sum of the second speed weight, time weight, and consistency weight is equal to 1. Formula (9) is used to measure the circuit breaker's ability to resist arc energy fluctuations, mechanical vibrations, and system disturbances during the opening process. It is calculated based on the mechanical characteristic parameter data during the opening process. The second speed factor is calculated by inputting the reference values ​​of the opening speed and the opening speed into formula (6).

[0052] The formula for calculating the time factor is as follows:

[0053] Equation (10)

[0054] in, Indicates the time factor. Indicates the opening time. The reference value for the opening time is set by the manufacturer. The time factor is used to characterize the ratio between the actual opening speed and the standard opening speed. The larger the time factor, the faster the actual opening speed and the better the performance. The value range of the time factor is [0-1]. When the time factor exceeds [0-1], it needs to be truncated.

[0055] The formula for calculating the degree of self-recovery is as follows:

[0056] Equation (11)

[0057] in, Indicates the degree of self-recovery. Indicates the time weighting without current. Indicates no current time factor. Indicates the short-term weight of gold. Indicates the short-term factor of gold. Indicates the weight of the secondary closing operation. Indicates the secondary closing factor. Indicates the secondary closing bounce weight. The secondary closing bounce factor is represented by the weight of no-current time, the weight of short-time, the weight of secondary closing, and the weight of secondary closing bounce. The sum of these weights equals 1. Formula (11) is used to measure the ability of a circuit breaker to reliably recover to a reoperable state after opening. It is calculated based on the mechanical characteristic parameter data in the reclosing cycle. The no-current time is used to evaluate the insulation recovery conditions at the fault point, the short-time is used to evaluate the ability to withstand the impact of permanent faults, and the secondary closing time and the secondary closing bounce time are combined to comprehensively characterize the mechanical performance and contact reliability of the operating mechanism under continuous operation.

[0058] The formula for calculating the current-free time factor is as follows:

[0059] Equation (12)

[0060] in, Indicates no current time factor. Indicates the time without current. This refers to the reference value for the no-current time. The no-current time characterizes the time interval during automatic reclosing, from the moment the arc in the arc-extinguishing chamber recovers to a level sufficient to withstand the system voltage without re-breakdown, thus creating safe conditions for re-conducting current. It is the time interval from the moment the arc in all poles is finally extinguished after the previous opening operation to the moment current first flows through any pole in the next closing operation. The no-current time should ideally be greater than or equal to this reference value to ensure the safety of the circuit breaker during opening and closing. When the mechanical characteristic parameter data is formed by an opening-closing-opening sequence, then both reclosing and secondary closing refer to the closing operation. When the mechanical characteristic parameter data is formed by a closing-opening-closing sequence, then both reclosing and secondary closing refer to the last closing operation.

[0061] The formula for calculating the gold short-term time factor is as follows:

[0062] Equation (13)

[0063] in, Indicates the short-term factor of gold. It means that gold is in a short time. This refers to the short-circuit time reference value. The short-circuit time characterizes the circuit breaker's ability to withstand the thermal and electrodynamic effects of short-circuit current in the closed state. During automatic reclosing, when the circuit breaker recloses to the faulty line, it is the time interval from the instant of contact (or pre-breakdown) until all pole arcs are finally extinguished (contacts completely separated). It is preferable that the short-circuit time be less than or equal to the short-circuit time reference value to ensure the safety of the circuit breaker during opening and closing. When the mechanical characteristic parameter data is formed by the opening-closing-opening action sequence, both the no-current time and the short-circuit time can be measured using a circuit breaker dynamic characteristic tester.

[0064] The formula for calculating the secondary closing factor is as follows:

[0065] Equation (14)

[0066] in, Indicates the secondary closing factor. Indicates the time of the second closing. This indicates the rated closing time. The shorter the secondary closing time, the faster the circuit breaker executes the secondary closing command or the faster its response speed. In this case, the secondary closing factor is larger.

[0067] The formula for calculating the secondary closing bounce factor is as follows:

[0068] Equation (15)

[0069] in, This represents the secondary closing bounce factor. Indicates the duration of the second closing bounce. This indicates the limit value for the secondary closing bounce time. The smaller the secondary closing bounce time, the larger the secondary closing bounce factor, indicating better bounce performance.

[0070] In one embodiment of this application, when the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the opening-closing-opening action sequence, or when the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the closing-opening-closing action sequence, the degree of self-recovery can be calculated by formula (11). When the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the closing-opening-closing action sequence, the degree of self-recovery is calculated using the no-current time, the secondary closing time, the secondary closing bounce time, and the short time obtained from the last measurement.

[0071] In another embodiment of this application, the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the opening-closing-opening action sequence, and the self-recovery degree is calculated by formula (11). When the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the closing-opening-closing action sequence, the self-recovery degree is obtained by weighted summation of the no-current time, the secondary closing time, and the secondary closing bounce time.

[0072] In another embodiment of this application, the mechanical characteristic parameter data includes at least mechanical characteristic parameter data formed by the opening-closing-opening action sequence, or at least includes mechanical characteristic parameter data formed by the closing-opening-closing action sequence. When calculating the self-recovery level, the calculation is performed based on the mechanical characteristic parameter data formed by the opening-closing-opening action sequence most recently at the current calculation time, or the calculation is performed based on the mechanical characteristic parameter data formed by the closing-opening-closing action sequence most recently at the current calculation time. When calculating the mechanical health level score, the calculation is performed based on the mechanical characteristic parameter data formed by the opening action most recently at the current calculation time, or the calculation is performed based on the mechanical characteristic parameter data formed by the closing action most recently at the current calculation time.

[0073] In step S140, the mechanical state of the switchgear is determined based on the mechanical health score and resilience score, and this mechanical state is used as the evaluation result of the switchgear's mechanical operation in high and low voltage switchgear. In one embodiment of this application, by acquiring the mechanical characteristic parameter data generated by the switchgear during the execution of opening and closing operations, the mechanical health of the switchgear is comprehensively evaluated from three dimensions: parameter deviation, operation integrity, and operation coordination. The dynamic response resilience of the switchgear is evaluated from three dimensions: shock resistance, fluctuation resistance, and self-recovery. The mechanical health score and resilience score are combined to determine the mechanical state of the switchgear, realizing comprehensive perception and comprehensive evaluation of multiple types of monitoring parameters. This effectively overcomes the shortcomings of single-parameter evaluation methods, which can only reflect a certain local state information of the switchgear and have insufficient ability to identify early deterioration characteristics. At the same time, it avoids the problems of pure data-driven models lacking physical mechanism support, poor interpretability, and insufficient dynamic adaptability. This improves the accuracy, reliability, and interpretability of switchgear mechanical state evaluation, provides a scientific basis for preventive maintenance and fault early warning of high and low voltage switchgear, and effectively ensures the continuity and safety of power system operation.

[0074] In one embodiment of this application, the process of determining the mechanical health score of a switching device based on the degree of parameter deviation, the degree of action integrity, and the degree of action coordination includes:

[0075] The current health score of the switching equipment is obtained by weighted summation of the parameter deviation, the integrity of the action, and the coordination of the action. In one embodiment of this application, the weights of the parameter deviation, the integrity of the action, and the coordination of the action during the weighted summation can be preset fixed weights or weights set by expert experience, and are not specifically limited here.

[0076] The current health status score is fuzzified to obtain the membership degree of each health status level. In one embodiment of this application, the process of fuzzifying the current health status score to obtain the membership degree of each health status level includes: pre-constructing the membership function of each health status level; inputting the current health status score into each membership function to obtain the initial membership degree of each health status level; and normalizing the initial membership degree of each health status level according to the sum of the initial membership degrees of each health status level to obtain the membership degree of each health status level. Each health status level includes: healthy state, good state, usable state, and fault state. Healthy state indicates that all indicators are normal and the equipment performance is good; good state indicates that a few indicators deviate slightly but do not affect safe operation; usable state indicates that some indicators are significantly deteriorated and there is a potential risk; and fault state indicates that key indicators are seriously exceeded and have affected equipment function. The membership function of the health status is shown below:

[0077] Equation (16)

[0078] in, Membership functions representing health status. This indicates the current health status score. This indicates the second scoring threshold. This indicates the third scoring threshold.

[0079] The membership function for a good state is shown below:

[0080] Equation (17)

[0081] in, Membership functions representing a good state. This indicates the current health status score. This indicates the first scoring threshold. This indicates the second scoring threshold. This indicates the third scoring threshold. This indicates the fourth scoring threshold.

[0082] The membership function of the available state is shown below:

[0083] Equation (18)

[0084] in, Membership functions representing the availability state, This indicates the current health status score. This indicates the first scoring threshold. This indicates the second scoring threshold. This indicates the third scoring threshold.

[0085] The membership function for the fault state is shown below:

[0086] Equation (19)

[0087] in, Membership functions representing fault states. This indicates the current health status score. This indicates the first scoring threshold. This represents the second scoring threshold. The relationship between the first, second, third, fourth, and fifth scoring thresholds is as follows: .

[0088] The degree of parameter deviation, the degree of action completeness, and the degree of action coordination are fuzzified to obtain the membership degree of each dimension parameter. Then, according to preset fuzzy rules, fuzzy inference is performed on the membership degree of each health status level and the membership degree of each dimension parameter to obtain the activation intensity of each inferred health status level. Finally, the activation intensity is normalized to obtain the probability distribution of the inferred health status level. In one embodiment of this application, the processes of fuzzifying the degree of parameter deviation, the degree of action completeness, and the degree of action coordination are all consistent with the process of fuzzifying the current health status score. The degree of parameter deviation, the degree of motion integrity, and the degree of motion coordination are all divided into low, medium, and high levels. The membership degree of parameter deviation includes: the membership degree of parameter deviation to the low level, the membership degree of parameter deviation to the medium level, and the membership degree of parameter deviation to the high level. The membership degree of motion integrity includes: the membership degree of motion integrity to the low level, the membership degree of motion integrity to the medium level, and the membership degree of motion coordination includes: the membership degree of motion coordination to the low level, the membership degree of motion coordination to the medium level, and the membership degree of motion coordination to the high level.

[0089] In one embodiment of this application, the preset fuzzy rules include: R1: IF parameter deviation is low level AND action integrity is high level AND action coordination is high level AND health status level is healthy state THEN inference health status level is healthy state; R2: IF parameter deviation is medium level AND action integrity is high level AND action coordination is medium level AND health status level is usable state THEN inference health status level is usable state; R3: IF parameter deviation is high level AND action integrity is low level AND action coordination is low level AND health status level is faulty state THEN inference health status level is faulty state, etc. Not all rule cases are listed here. For each fuzzy rule, the activation strength is the minimum value of all membership degrees in the antecedent; if there are multiple fuzzy rules pointing to the same inference health status level in the consequent, the maximum value among the multiple activation strengths is determined as the activation strength of the corresponding inference health status level, thereby obtaining the activation strength of each inference health status level.

[0090] The mechanical health status score is obtained by weighting and summing the representative scores of each inference health status level using the probability distribution of each inference health status level as the weight. In one embodiment of this application, taking the inference health status levels as including: healthy state, good state, usable state, and faulty state as an example, the probability distribution of assignment includes: the probability of assignment for healthy state, the probability of assignment for good state, the probability of assignment for usable state, and the probability of assignment for faulty state. The representative score of each inference health status level is determined according to the scoring interval of the corresponding inference health status level. For example, if the scoring interval for healthy state is set to [80, 100], the scoring interval for good state is set to [60, 80], the scoring interval for usable state is set to [30, 60], and the scoring interval for faulty state is set to [0, 30], the representative score of each inference health status level is the midpoint value of the scoring interval of the corresponding inference health status level. The calculation formula for the mechanical health status score is as follows:

[0091] Equation (20)

[0092] in, Indicates the health status score of the machinery. Indicates the first The probability of belonging to each health status level. Indicates the first Representative scores for each health status level The total number of health status levels is represented by formula (20). Formula (20) is used to characterize the mechanical health of the switching equipment at the current moment.

[0093] In one embodiment of this application, fuzzy inference is performed on the membership degrees of each health status level and each dimension parameter to obtain the activation intensity of each inferred health status level. The activation intensity is then normalized and transformed into a probability distribution of the inferred health status level. On the one hand, fuzzy rules encode domain expert experience into computable inference rules, achieving deep fusion of the membership degrees of each health status level and each dimension parameter, avoiding the problem that a single dimension score may mask local anomalies. On the other hand, the probability distribution of the inferred health status retains the uncertainty information of the mechanical health level, and the probability distribution of the inferred health status is used as a weight to assign values ​​to each inferred health status. The representative scores of the levels are weighted and summed so that the final machine health score reflects both the expected level of machine health and the confidence level of the machine health judgment. This provides a richer basis for subsequent fusion decision-making and makes up for the shortcomings of the current health score, which is that parameter deviation, action integrity and action coordination are independent of each other and cannot reflect the nonlinear interaction between parameter deviation, action integrity and action coordination. This makes it easy for local anomalies in one dimension to be masked by normal performance in other dimensions, and the current health score cannot reflect the confidence level.

[0094] In one embodiment of this application, the process of determining the toughness score of a switchgear based on its shock resistance, fluctuation resistance, and self-recovery capability includes:

[0095] The current resilience score of the switchgear is obtained by weighted summation of the impact resistance, fluctuation resistance, and self-recovery level. In one embodiment of this application, the weights of impact resistance, fluctuation resistance, and self-recovery level during the weighted summation can be preset fixed weights or weights set by expert experience; no specific limitation is made here. The sum of the weights of impact resistance, fluctuation resistance, and self-recovery level equals 1.

[0096] The historical resilience scores of the switchgear are obtained, and the current resilience score and the historical resilience scores are concatenated in chronological order to obtain a resilience score sequence. In one embodiment of this application, the historical resilience score includes resilience scores from multiple historical moments.

[0097] A first score segment is obtained by filtering from the resilience score sequence according to a first time span; and a second score segment is obtained by filtering from the resilience score sequence according to a second time span. In one embodiment of this application, the first time span is longer than the second time span, and the last resilience score of the first score segment and the last resilience score of the second score segment are both the current resilience score; the first time span is set according to the actual situation, the second time span is set according to the actual situation, the first score segment includes at least 5 resilience scores, and the second score segment includes at least 10 resilience scores.

[0098] Linear regression analysis is performed on the resilience score in the first scoring segment to obtain a first resilience trend quantification index; and linear regression fitting is performed on the resilience score in the second scoring segment to obtain a second regression slope. In one embodiment of this application, the process of performing linear regression fitting on the resilience score in the second scoring segment includes: fitting the resilience score in the second scoring segment using the least squares method to obtain a second fitting equation; and using the slope of the second fitting equation as the second regression slope. The first quantitative indicators of resilience trend include: the first regression slope, goodness of fit, and significance probability. The process of performing linear regression analysis on the resilience scores in the first scoring segment includes: using different time points as independent variables and the resilience scores in the first scoring segment as dependent variables, using the least squares method to fit the resilience scores in the first scoring segment to obtain the first fitting equation, and the slope of the first fitting equation as the first regression slope; calculating the coefficient of determination as the goodness of fit based on the resilience scores in the first scoring segment, the fitted resilience scores, and the mean of the resilience scores; calculating the standard error of the residuals and the standard error of the slope based on the resilience scores in the first scoring segment and the fitted resilience scores; using the ratio of the first regression slope to the standard error of the slope as the t-statistic; and using the t-statistic and degrees of freedom as a benchmark, looking up the t-distribution table to obtain the significance probability.

[0099] The formula for calculating goodness of fit is shown below:

[0100] Equation (21)

[0101] in, Indicates the goodness of fit. This indicates the total number of resilience scores in the first scoring segment. Indicates the first scoring segment A resilience score, Indicates the first scoring segment The fitted resilience score is obtained through the first fitted equation. This represents the mean resilience score, which is the average resilience score in the first score segment. Goodness of fit measures how well the first fitting equation fits the resilience scores in the first score segment. A higher goodness of fit indicates a better fit and a clearer linear trend. When the difference between the goodness of fit and 1 is less than a preset difference threshold, it indicates that the change in resilience scores over time is mainly dominated by a linear trend with small random fluctuations. If the goodness of fit is low (the difference between the goodness of fit and 0 is less than the preset difference threshold), it indicates large data fluctuations and an unreliable linear trend.

[0102] The formula for calculating the standard error of the residuals is as follows:

[0103] Equation (22)

[0104] in, Indicates the standard error of the residual. This indicates the total number of resilience scores in the first scoring segment. Indicates the first scoring segment A resilience score, Indicates the first scoring segment A fitted resilience score is obtained through the first fitted equation. The standard error of the residuals represents the degrees of freedom. It measures the average deviation between the estimated value output by the first fitted equation and the actual observed value, reflecting the typical amplitude of the random fluctuation of the dependent variable (resilience score) after excluding the linear effects of the independent variables (at different times). A smaller standard error indicates that the resilience score closely surrounds the first fitted equation, resulting in a better fit and more accurate prediction.

[0105] The formula for calculating the standard error of the slope is as follows:

[0106] Equation (23)

[0107] in, The standard error representing the slope, Indicates the standard error of the residual. This indicates the total number of resilience scores in the first scoring segment. Indicates the first scoring segment The time sequence number of each resilience score. This represents the average of all resilience scores corresponding to the time sequence in the first scoring segment. The standard error of the slope is used to measure the accuracy and stability of the first regression slope; the smaller the standard error of the slope, the closer the first regression slope is to the true value.

[0108] Based on the first resilience trend quantification index and the second regression slope, the current resilience score is corrected to obtain the resilience score. In one embodiment of this application, the changing trend of the resilience score is incorporated into the current resilience score, which can effectively suppress score jumps caused by instantaneous interference, smooth out short-term random fluctuations, realize the fusion of historical patterns and real-time status, and significantly improve the stability, continuity, and noise robustness of the resilience score calculation.

[0109] In one embodiment of this application, the process of correcting the current resilience score based on a first resilience trend quantification index and a second regression slope to obtain the resilience score includes:

[0110] A first correction factor is determined based on a first resilience trend quantification index. In one embodiment of this application, the calculation formula for the first correction factor is as follows:

[0111] Equation (24)

[0112] in, Indicates the first correction factor. This represents the normalized value of the first regression slope, which is the ratio of the first regression slope to the theoretical upper limit of the resilience score (e.g., 100 points). It characterizes the relative rate of change of the resilience score per unit time (i.e., the proportion of the change in the resilience score per unit time to the theoretical upper limit), and its dimension is the reciprocal of the unit time. This represents the sensitivity adjustment factor, used to balance the dimensions of the normalized value of the first regression slope and control the correction magnitude of the normalized value of the first regression slope. The dimension is unit time (e.g., hour or day). Indicates credibility. Formula (24) is used to characterize the credibility level. Based on this benchmark, and according to confidence level and a sensitivity adjustment factor, the normalized value of the first regression slope is converted into a reward or penalty for the current resilience score. When Less than 0 (indicating a tendency to deteriorate). Less than 1 (first correction factor downgraded, score reduced), when Greater than 0 (indicating an increasing trend). Greater than 1 (first correction factor increased, improving score), when The difference between 0 and 0 is within a preset range (representing a stable trend). It equals 1 (the first correction factor remains unchanged).

[0113] The formula for calculating credibility is as follows:

[0114] Equation (25)

[0115] in, Indicates credibility. Indicates the goodness of fit. This represents the significance probability value. Equation (25) in A higher slope and a lower P value indicate that the trend of the first regression slope is reliable. In this case, the correction to the current resilience score is relatively large. Otherwise, the correction to the current resilience score is relatively small.

[0116] The current resilience score is multiplied by a first correction factor to obtain a first resilience score. In one embodiment of this application, the current resilience score is corrected by the first correction factor, enabling differentiated evaluation of the current resilience score through trend-based rewards and penalties. This helps alleviate the lag defect of instantaneous scoring and improves the robustness of the resilience score.

[0117] A second correction factor is determined based on the comparison between the first and second regression slopes. In one embodiment of this application, comparing the first and second regression slopes can identify trend abrupt changes and trend divergences, which helps to enable the second correction factor to have trend consistency verification capabilities.

[0118] The first resilience score is multiplied by the second correction factor to obtain the resilience score. In one embodiment of this application, a multiplication gating mechanism is used to apply the result of trend consistency verification as an adjustment weight to the first resilience score. This upgrade from "single trend tracking" to "dual verification evaluation" is achieved while maintaining the stability of the resilience score magnitude, thus improving the reliability of the resilience score. After obtaining the resilience score, if the resilience score exceeds the upper limit, it is set as the upper limit; if the resilience score is lower than the lower limit, it is set as the lower limit.

[0119] In one embodiment of this application, the process of determining the second correction factor based on the comparison result of the first regression slope and the second regression slope includes:

[0120] If the second regression slope is less than the first regression slope, and the first regression slope is less than the first preset slope threshold, then the first correction factor is used as the second correction factor. In one embodiment of this application, the first preset slope threshold can be 0 or other thresholds, and the first correction factor is less than the base factor; the base factor can be set to 1 or other values, and the range of the first correction factor is (0.95-0.98), with a value of 0.96. When the second regression slope is less than the first regression slope, and the first regression slope is less than the first preset slope threshold, it indicates that the switching equipment is in an accelerated degradation stage, and the first correction factor is less than the base factor, indicating a trend penalty on the first resilience score.

[0121] If the first regression slope is less than the second regression slope, and the second regression slope is less than the first preset slope threshold, then the second correction factor is used as the second correction factor. In one embodiment of this application, the second correction factor is greater than the benchmark factor; the value range of the second correction factor is (1.02-1.05), and the value can be 1.03. If the first regression slope is less than the second regression slope, and the second regression slope is less than the first preset slope threshold, it indicates that the switching equipment is in a degradation deceleration stage, and the second correction factor is greater than the benchmark factor, indicating that there is a trend reward for the first resilience score.

[0122] If the second regression slope is greater than the first preset slope threshold and the first regression slope is less than the first preset slope threshold, then the third correction factor is used as the second correction factor. In one embodiment of this application, the second correction factor is less than the third correction factor; the value range of the third correction factor is (1.03-1.06), and the value can be 1.05. If the second regression slope is greater than the first preset slope threshold and the first regression slope is less than the first preset slope threshold, it indicates that the switching equipment is in the recent recovery phase, and the second correction factor is less than the third correction factor, indicating that there is a trend reward for the first resilience score.

[0123] If the absolute value of the difference between the second regression slope and the first preset slope threshold is less than the preset absolute value threshold, and the absolute value of the difference between the first regression slope and the first preset slope threshold is also less than the preset absolute value threshold, then the benchmark multiplier is used as the second correction factor. In one embodiment of this application, the preset absolute value threshold is set according to the actual situation. If the absolute value of the difference between the second regression slope and the first preset slope threshold is less than the preset absolute value threshold, and the absolute value of the difference between the first regression slope and the first preset slope threshold is also less than the preset absolute value threshold, then it indicates that the switching equipment is in a trend-stable stage, the second correction factor is the benchmark multiplier, and no trend reward or trend penalty is applied to the first resilience score.

[0124] If the absolute value of the difference between the first regression slope and the first preset slope threshold is less than the preset absolute value threshold, and the absolute value of the second regression slope is greater than the second preset slope threshold, then the fourth correction factor is used as the second correction factor. In one embodiment of this application, the second preset slope threshold is greater than the first preset slope threshold; the second preset slope threshold is set according to actual conditions (e.g., 2), and the fourth correction factor is less than the first correction factor; the value range of the fourth correction factor is (0.90-0.95), and the value can be 0.94. If the absolute value of the difference between the first regression slope and the first preset slope threshold is less than the preset absolute value threshold, and the absolute value of the second regression slope is greater than the second preset slope threshold, it indicates that the switching equipment is in a sudden disturbance stage, and the fourth correction factor is less than the first correction factor, indicating that there is a trend penalty for the first resilience score.

[0125] If the second regression slope is greater than the first regression slope, and the first regression slope is greater than the first preset slope threshold, then the fifth correction factor is used as the second correction factor. In one embodiment of this application, the fifth correction factor is greater than the third correction factor; the value range of the fifth correction factor is (1.05-1.1), and the value can be 1.06. If the second regression slope is greater than the first regression slope, and the first regression slope is greater than the first preset slope threshold, it indicates that the switching equipment is in a continuously excellent stage, and the fifth correction factor is greater than the third correction factor, indicating that the first resilience score has a trend reward.

[0126] If the second regression slope is less than the first regression slope, and the first regression slope is greater than the first preset slope threshold, then the sixth correction factor is used as the second correction factor. In one embodiment of this application, the sixth correction factor is greater than the first correction factor, and the sixth correction factor is less than the benchmark factor. The value range of the sixth correction factor is (0.96-0.99), and it can be 0.97. If the second regression slope is less than the first regression slope, and the first regression slope is greater than the first preset slope threshold, it indicates that the switching equipment is in the recovery obstruction stage. The sixth correction factor being less than the benchmark factor indicates that there is a trend penalty on the first resilience score.

[0127] In one embodiment of this application, there are three cases where the second regression slope is less than the first regression slope and the first regression slope is greater than the first preset slope threshold: the second regression slope is less than the first preset slope threshold, the second regression slope is less than the first regression slope, and the first regression slope is greater than the first preset slope threshold (referred to as the first case); the second regression slope is equal to the first preset slope threshold, the second regression slope is less than the first regression slope, and the first regression slope is greater than the first preset slope threshold (referred to as the second case); and the second regression slope is greater than the first preset slope threshold, the second regression slope is less than the first regression slope, and the first regression slope is greater than the first preset slope threshold (referred to as the third case). The first case indicates that the switching equipment is in a recovery deceleration phase, that is, both the long-term and short-term trends are recovering, but the short-term recovery speed is lower than the long-term recovery speed, and the recovery momentum is slowing down; the second case indicates that the switching equipment is in a short-term stagnation phase, that is, the long-term trend is positive, but the short-term recovery is temporarily stagnant; and the third case indicates that the switching equipment is in a short-term regression phase, that is, the long-term trend is positive, but there are signs of regression in the short term.

[0128] In one embodiment of this application, adaptive hierarchical control of the correction intensity is achieved by comparing the first regression slope and the second regression slope. This means that the stage of the switching equipment can be distinguished through threshold gating, and the correction intensity can be dynamically adjusted according to the different stages of the switching equipment, significantly improving the flexibility, robustness, and reliability of the resilience score. If the conditions for determining the second correction factor listed above are not met, the benchmark ratio is used as the second correction factor.

[0129] In one embodiment of this application, the process of determining the mechanical condition of a switchgear based on a mechanical health score and a toughness score includes:

[0130] The contact resistance value, cumulative number of operations, and environmental parameter data of the switching equipment are acquired. In one embodiment of this application, the contact resistance value is obtained by testing with a loop resistance tester, the cumulative number of operations is determined based on usage records, and the environmental parameter data includes ambient temperature and humidity, which are collected by temperature and humidity sensors.

[0131] Based on the contact resistance value, the resistance change rate and the three-phase imbalance rate are calculated. In one embodiment of this application, the contact resistance value includes: three-phase contact resistance, and the resistance change rate includes: three-phase resistance change rate. The calculation formula for the resistance change rate of each phase is as follows:

[0132] Equation (26)

[0133] in, Indicates the first Phase resistance change rate, Indicates the first Phase contact resistance value, Indicates the first Phase reference resistance value. The phase can be phase A, phase B, or phase C.

[0134] The formula for calculating the three-phase unbalance rate is as follows:

[0135] Equation (27)

[0136] in, This represents the three-phase imbalance rate. This represents the maximum value among the three-phase contact resistances. This represents the minimum value of the three-phase contact resistance. This represents the average value of the three-phase contact resistance.

[0137] Based on the comparison results of the mechanical health score and the preset health score, the toughness score and the preset toughness score, the resistance change rate and the resistance change rate threshold, and the three-phase imbalance rate and the three-phase imbalance rate threshold, the first mechanical state and confidence level of the switchgear are determined. In one embodiment of this application, the first mechanical state includes: normal state, warning state, defective state, and pending confirmation state; the resistance change rate threshold is determined based on the toughness score; the three-phase imbalance rate threshold is determined based on the toughness score; the process of determining the resistance change rate threshold based on the toughness score is the same as the process of determining the three-phase imbalance rate threshold based on the toughness score.

[0138] The contact resistance value, resistance change rate, three-phase imbalance rate, mechanical health score, toughness score, environmental parameter data, and cumulative number of operations are input into the AI ​​model to obtain the second mechanical state and state probability of the switchgear. In one embodiment of this application, the second mechanical state includes: normal state, warning state, defective state, and pending confirmation state. The preset model can be an XGBoost (eXtreme Gradient Boosting) model or a LightGBM (Light Gradient Boosting Machine) model, etc. The AI ​​(Artificial Intelligence Model) model is trained on a pre-defined model based on sample data. The sample data includes: historical contact resistance time series, historical resistance change rate time series, historical three-phase imbalance rate time series, historical mechanical health score time series, historical resilience score time series, historical environmental parameter data time series, historical cumulative operation count time series, and historical mechanical status labels. The historical resistance change rate is calculated based on the contact resistance at the corresponding historical moment, the historical three-phase imbalance rate is calculated based on the contact resistance at the corresponding historical moment, the historical mechanical health score is calculated based on the mechanical characteristic parameter data within the corresponding historical time period, and the historical resilience score is calculated based on the historical mechanical characteristic parameter data within the corresponding historical time period. After obtaining the sample data, the historical contact resistance time series, historical resistance change rate time series, historical three-phase imbalance rate time series, historical mechanical health score time series, historical resilience score time series, historical environmental parameter data time series, and historical cumulative operation count time series are time-aligned to obtain feature data. The feature data is input into a preset model, which outputs mechanical state prediction results. The loss value between the historical mechanical state label (represented by one-hot encoding) and the label corresponding to the mechanical state prediction result (represented by one-hot encoding) is calculated. An optimizer (e.g., Adam (Adaptive Moment Estimation) or SGD (Stochastic Gradient Descent)) is used to adjust the parameters in the preset model based on the loss value. The above process is repeated until the number of training iterations reaches a preset threshold, or the loss value is less than or equal to the preset loss threshold, and the AI ​​model is output.

[0139] The mechanical state of the switchgear is determined based on the second mechanical state, the state probability, the first mechanical state, and the confidence level. In one embodiment of this application, both the second mechanical state and the first mechanical state are the mechanical states at the current moment. The state probability is used to characterize the likelihood of the second mechanical state occurring, and the confidence level is used to characterize the degree of consistency between the first mechanical state and the multidimensional input data. The process of determining the mechanical state of the switchgear based on the second mechanical state, the state probability, the first mechanical state, and the confidence level improves the accuracy and robustness of the mechanical state determination by utilizing the cross-validation mechanism of multi-source evaluation results. Furthermore, by introducing the state probability and the confidence level, a quantitative credibility representation of the evaluation results is achieved, effectively avoiding misjudgments and omissions caused by a single evaluation path.

[0140] In one embodiment of this application, the process of determining the mechanical state of a switching device based on a second mechanical state, a state probability, a first mechanical state, and a confidence level includes:

[0141] If the mechanical state with the highest confidence level in the first mechanical state is consistent with the mechanical state with the highest state probability in the second mechanical state, then the mechanical state with the highest confidence level or the mechanical state with the highest state probability is taken as the mechanical state. In one embodiment of this application, the accuracy of mechanical state determination is improved by mutually verifying the confidence level and the state probability.

[0142] If the mechanical state with the highest confidence in the first mechanical state is inconsistent with the mechanical state with the highest state probability in the second mechanical state, then the historical mechanical states of the switching equipment and the state transition probability vector under the historical mechanical states are obtained. The historical mechanical states and the state transition probability vector are multiplied to obtain the probability distribution under the predicted mechanical state. The probability value of the mechanical state with the highest confidence in the probability distribution is used as the matching degree between the predicted mechanical state and the mechanical state with the highest confidence, denoted as the first matching degree. The probability value of the mechanical state with the highest state probability in the probability distribution is used as the matching degree between the predicted mechanical state and the mechanical state with the highest state probability, denoted as the second matching degree. In one embodiment of this application, the historical mechanical state includes: the mechanical state of the previous evaluation period. Taking the mechanical state of the previous evaluation period as the normal state as an example, the state transition probability vector includes: the probability of transitioning from the normal state to the normal state, the probability of transitioning from the normal state to the warning state, the probability of transitioning from the normal state to the defect state, and the probability of transitioning from the normal state to the pending confirmation state. The state transition probability can be set based on expert experience or obtained from historical data statistics. For example, the total number of state transitions in historical mechanical states can be obtained from historical data, and the number of times the historical mechanical state transitioned to the normal state, the number of times the warning state, the number of times the defect state, and the number of times the pending confirmation state were counted separately can be combined as the state transition probability vector.

[0143] In one embodiment of this application, the predicted mechanical state is taken as the third mechanical state at the current moment. Taking the mechanical state with the highest confidence as the normal state as an example, the first matching degree is the probability value of the normal state in the probability distribution. Taking the mechanical state with the highest state probability as the warning state as an example, the second matching degree is the probability value of the warning state in the probability distribution.

[0144] If the difference between the first matching degree and the second matching degree is greater than a preset difference threshold, the mechanical state corresponding to the larger matching degree between the first and second matching degrees is taken as the mechanical state. In one embodiment of this application, when the difference between the first matching degree and the second matching degree is greater than the preset difference threshold, the mechanical state corresponding to the larger matching degree is selected as the final mechanical state, ensuring the reliability and accuracy of the final mechanical state.

[0145] If the difference between the first matching degree and the second matching degree is less than or equal to a preset difference threshold, then the mechanical state with the higher anomaly level between the mechanical state with the highest confidence and the mechanical state with the highest state probability is taken as the mechanical state. In one embodiment of this application, when the difference between the first matching degree and the second matching degree is less than or equal to the preset difference threshold, the mechanical state with the higher anomaly level is taken as the mechanical state, reducing unnecessary computational overhead and decision fluctuations, and achieving an effective balance between decision stability and accuracy.

[0146] In one embodiment of this application, a third mechanical state is determined by introducing a Markov state transition model, and the mechanical state with the highest confidence and the mechanical state with the highest state probability are verified and selected based on the third mechanical state, thereby ensuring the stability, reliability and accuracy of the determination of the mechanical state.

[0147] In one embodiment of this application, the process of determining the first mechanical state and confidence level of the switchgear based on the comparison results of the mechanical health score and the preset health score, the toughness score and the preset toughness score, the resistance change rate and the preset resistance change rate, and the three-phase imbalance rate and the preset three-phase imbalance rate includes:

[0148] The comparison result between the resistance change rate and the resistance change rate threshold is taken as the first comparison result; and the comparison result between the three-phase imbalance rate and the three-phase imbalance rate threshold is taken as the second comparison result. In one embodiment of this application, the resistance change rate includes: phase A resistance change rate, phase B resistance change rate, and phase C resistance change rate; the resistance change rate threshold includes: phase A resistance change rate threshold, phase B resistance change rate threshold, and phase C resistance change rate threshold; the first comparison result includes: phase A resistance change rate greater than phase A resistance change rate threshold (i.e., phase A exceeds the limit), phase A resistance change rate less than or equal to phase A resistance change rate threshold (i.e., phase A does not exceed the limit); phase B resistance change rate greater than phase B resistance change rate threshold (i.e., phase B exceeds the limit), phase B resistance change rate less than or equal to phase B resistance change rate threshold (i.e., phase B does not exceed the limit); phase C resistance change rate greater than phase C resistance change rate threshold (i.e., phase C exceeds the limit), phase C resistance change rate less than or equal to phase C resistance change rate threshold (i.e., phase C does not exceed the limit); the second comparison result includes: three-phase imbalance rate greater than three-phase imbalance rate threshold (i.e., three-phase imbalance rate exceeds the limit), three-phase imbalance rate less than or equal to three-phase imbalance rate threshold (i.e., three-phase imbalance rate does not exceed the limit).

[0149] The loop resistance state is determined based on the first and second comparison results. In one embodiment of this application, the loop resistance state includes: normal state, three-phase common degradation state, single-phase local defect state, and data anomaly state. The loop resistance state also includes: two-phase common degradation state, severely deteriorated state, and poor contact state. The mapping relationship between the first and second comparison results and the loop resistance state is shown in Table 1.

[0150] Table 1

[0151]

[0152] As shown in Table 1, when the three-phase resistance change rates are all within limits and the three-phase imbalance rate is also within limits, the switchgear is generally healthy, and the loop resistance status is judged as normal. When the three-phase resistance change rates are all within limits, but the three-phase imbalance rate is not, it indicates synchronous deterioration of the three phases, which is an overall degradation, and the loop resistance status is judged as a three-phase common degradation state. When only one phase's resistance change rate exceeds the limit, and the three-phase imbalance rate also exceeds the limit, it indicates that there is a local defect in that phase, and the loop resistance status is judged as a single-phase local defect state. When two phases' resistance change rates exceed the limit, one phase does not exceed the limit, and the three-phase imbalance rate exceeds the limit, it indicates that there is synchronous deterioration in the two phases, and the loop resistance status is judged as a two-phase local defect state. Common phase degradation state; when the three-phase resistance change rate is within the limit, but the three-phase imbalance rate is within the limit, it indicates a contact quality problem rather than material deterioration, and the loop resistance state is judged as a poor contact state; when the three-phase resistance change rate is within the limit and the three-phase imbalance rate is also within the limit, it indicates that the overall equipment is severely degraded, and special attention is required at this time, and the loop resistance state is judged as a severely degraded state or a state of common phase degradation plus local defects (adjusted according to the actual situation); when the resistance change rate of some phases exceeds the limit but the three-phase imbalance rate does not exceed the limit (only one or two phases exceed the limit), it is a logical contradiction combination, and the loop resistance state is judged as a data abnormal state, triggering the data review process.

[0153] The comparison result between the mechanical health score and the preset health score is used as the third comparison result; the comparison result between the toughness score and the preset toughness score is used as the fourth comparison result. In one embodiment of this application, the third comparison result includes: the mechanical health score is greater than or equal to the preset health score (i.e., the mechanical health score meets the standard), and the mechanical health score is less than the preset health score (i.e., the mechanical health score does not meet the standard). The fourth comparison result includes: the toughness score is greater than or equal to the preset toughness score (i.e., the toughness score meets the standard), and the toughness score is less than the preset toughness score (i.e., the toughness score does not meet the standard). The preset health score can be determined based on the switchgear's factory parameters, historical operating data statistics, or industry standards. The preset toughness score can be determined based on factors such as the type of switchgear, its operating years, and environmental conditions.

[0154] Based on the third comparison result, the fourth comparison result, and the loop resistance status, the first mechanical state of the switching device is determined. In one embodiment of this application, the first mechanical state includes: normal state, warning state, defective state, and pending confirmation state; the mapping relationship between the third comparison result, the fourth comparison result, the loop resistance status, and the first mechanical state is shown in Table 2.

[0155] Table 2

[0156]

[0157] As can be seen from Table 2, when the loop resistance status is in an abnormal state, regardless of the results of the third and fourth comparisons, the first mechanical status is determined to be a pending confirmation state; when the loop resistance status is in a normal state, and both the third and fourth comparison results meet the standards, the first mechanical status is determined to be a normal state; when at least one of the third and fourth comparison results is non-compliant, and the loop resistance status is in a normal state or a severely degraded state, the first mechanical status is determined to be a warning state; when at least one of the third and fourth comparison results is non-compliant, and the loop resistance status is in a three-phase common degradation state, a two-phase common degradation state, a single-phase local defect state, or a poor contact state, the first mechanical status is determined to be a defect state.

[0158] The confidence levels for evidence consistency, resistance deviation, toughness support, and three-phase consistency are calculated for the first mechanical state. In one embodiment of this application, the confidence level for evidence consistency is determined based on the third comparison result, the fourth comparison result, and the loop resistance state; the mapping relationship between the third comparison result, the fourth comparison result, the loop resistance state, and the confidence level for evidence consistency is shown in Table 3.

[0159] Table 3

[0160]

[0161] As can be seen from Table 3, when the third and fourth comparison results are consistent and both meet the standards, the confidence level of evidence consistency decreases in the direction of normal state, warning state, and defective state of the first mechanical state; when the third and fourth comparison results are inconsistent, the confidence level decreases in the direction of warning state, defective state, and normal state of the first mechanical state; when the third and fourth comparison results are consistent and both fail to meet the standards, the confidence level of evidence consistency decreases in the direction of defective state, warning state, and normal state of the first mechanical state.

[0162] In one embodiment of this application, by combining the evidence from the third comparison result, the fourth comparison result, and the loop resistance state, an evidence consistency confidence value in the range of 0.20 to 1.00 is output to quantify the degree of mutual support among the current multi-source evidence, providing a reliable basis for the determination of the first mechanical state.

[0163] In one embodiment of this application, the confidence level of the resistance deviation is determined based on the resistance change rate. The confidence level of the resistance deviation includes: the confidence level of the three-phase resistance deviation, with the third phase as the basis. The formula for calculating the confidence level of the resistance deviation is as follows:

[0164] Equation (28)

[0165] in, Indicates the first Confidence level of phase resistance deviation Indicates the first Phase resistance change rate, This represents the attenuation coefficient, used to control the descent speed.

[0166] In one embodiment of this application, the toughness support confidence level is determined based on a toughness score, which is the ratio of the toughness score to a baseline score (e.g., 100 points). The three-phase consistency confidence level is determined based on the three-phase imbalance rate; the formula for calculating the three-phase consistency confidence level is as follows:

[0167] Equation (29)

[0168] in, This indicates the confidence level of three-phase consistency. This represents the three-phase imbalance rate. This represents the maximum value of the three-phase imbalance rate.

[0169] The confidence scores for evidence consistency, resistance deviation, toughness support, and three-phase consistency are weighted and summed to obtain the confidence score for the first mechanical state. In one embodiment of this application, after obtaining the confidence scores for the three-phase resistance deviation, the minimum value among the three-phase resistance deviation confidence scores is selected to participate in the confidence score calculation. That is, the confidence scores for evidence consistency, the minimum value among the three-phase resistance deviation confidence scores, the toughness support confidence score, and the three-phase consistency confidence score are weighted and summed to obtain the confidence score for the first mechanical state. The weights of the confidence scores for evidence consistency, the minimum value among the three-phase resistance deviation confidence scores, the toughness support confidence score, and the three-phase consistency confidence score are set according to the actual situation. For example, the weight of the confidence score for evidence consistency is set to 0.35, the weight of the minimum value among the three-phase resistance deviation confidence scores is set to 0.30, the weight of the toughness support confidence score is set to 0.20, and the weight of the three-phase consistency confidence score is set to 0.15.

[0170] In one embodiment of this application, the first mechanical state of the switching equipment is determined by the third comparison result, the fourth comparison result, and the loop resistance state, effectively avoiding misjudgments and omissions caused by a single scoring dimension, and significantly improving the accuracy of the first mechanical state diagnosis. The four dimensions of resistance change rate (reflecting contact state), three-phase imbalance rate (reflecting three-phase balance), mechanical health score (reflecting overall mechanical characteristics), and toughness score (reflecting equipment's anti-disturbance capability) are organically integrated, and the electrical characteristic indicators are correlated and verified with the mechanical characteristic scores, realizing deep coupling and cross-verification between the electrical and mechanical dimensions, and improving the reliability of the first mechanical state diagnosis. By weighted summing of four sub-confidences—evidence consistency confidence, resistance deviation confidence, toughness support confidence, and three-phase consistency confidence—the credibility of the first mechanical state diagnosis can be reflected more comprehensively and more precisely.

[0171] In one embodiment of this application, the process of determining the resistivity change rate threshold based on a toughness score includes:

[0172] If the absolute value of the difference between the toughness score and the benchmark toughness score is less than a preset score difference threshold, then the benchmark threshold is used as the resistance change rate threshold. In one embodiment of this application, the benchmark toughness score is determined based on historical toughness scores under normal conditions, for example, the median of historical toughness scores under normal conditions. The preset score difference threshold is set according to actual conditions (e.g., The benchmark threshold is set based on actual conditions and with reference to industry standards.

[0173] If the absolute value of the difference between the toughness score and the benchmark toughness score is greater than or equal to a preset score difference threshold, and the toughness score is greater than the benchmark toughness score, then a resistance change rate threshold is calculated based on the benchmark judgment threshold and the first preset slope. In one embodiment of this application, the first preset slope is set according to actual conditions; for example, the value range of the first preset slope is 0.01 to 0.1.

[0174] When the absolute value of the difference between the toughness score and the benchmark toughness score is greater than or equal to a preset score difference threshold, and the toughness score is less than the benchmark toughness score, a resistance change rate threshold is calculated based on the benchmark judgment threshold and the second preset slope. In one embodiment of this application, the second preset slope is set according to actual conditions; for example, the value range of the second preset slope is -0.2 to -0.05. The absolute value of the second preset slope is greater than the absolute value of the first preset slope, so that when the toughness score is lower than the benchmark toughness score, the resistance change rate threshold tightens at a faster rate, thereby improving detection sensitivity and reducing the risk of missed defect detection when the equipment's anti-disturbance capability decreases. At the same time, it limits the relaxation range when the toughness score is higher than the benchmark toughness score, avoiding excessive relaxation of detection standards due to good equipment condition, and realizing an asymmetric risk response strategy. If the absolute value of the difference between the toughness score and the benchmark toughness score is less than the preset score difference threshold, the resistance change rate threshold remains unchanged, avoiding frequent threshold switching caused by small fluctuations.

[0175] In one embodiment of this application, the formula for calculating the resistance change rate threshold is as follows:

[0176] Equation (30)

[0177] in, Indicates the threshold of the rate of change of resistance. Indicates the benchmark judgment threshold. Indicates the preset slope. Indicates resilience score, This represents the baseline resilience score. If the absolute value of the difference between the resilience score and the baseline resilience score is greater than or equal to a preset score difference threshold, and the resilience score is greater than the baseline resilience score, the preset slope is the first preset slope; when the absolute value of the difference between the resilience score and the baseline resilience score is greater than or equal to the preset score difference threshold, and the resilience score is less than the baseline resilience score, the preset slope is the second preset slope. The calculation logic for the three-phase imbalance rate threshold is the same as the calculation logic for the resistance change rate threshold.

[0178] Figure 2 This is another exemplary embodiment of the present application illustrating a flowchart for determining the toughness score of a switchgear based on its shock resistance, fluctuation resistance, and self-recovery capability. Figure 2In this study, the process of determining the resilience score of switchgear based on its shock resistance, fluctuation resistance, and self-recovery includes: (1) weighted summation of shock resistance, fluctuation resistance, and self-recovery to obtain the current resilience score of the switchgear; (2) obtaining the historical resilience score of the switchgear and splicing the current resilience score and the historical resilience score in chronological order to obtain a resilience score sequence; (3) selecting the first score segment from the resilience score sequence according to the first time span; and selecting the second score segment from the resilience score sequence according to the second time span; (4) performing linear regression analysis on the resilience score in the first score segment to obtain the first resilience trend quantification index; and performing linear regression fitting on the resilience score in the second score segment to obtain the second regression slope; (5) correcting the current resilience score based on the first resilience trend quantification index and the second regression slope to obtain the resilience score.

[0179] Figure 3 This is another exemplary embodiment of the present application illustrating a flowchart for determining the mechanical state of a switchgear based on a mechanical health score and a toughness score. Figure 3 The process of determining the mechanical state of a switchgear based on its mechanical health score and resilience score includes: (1) obtaining the contact resistance value, cumulative number of operations, and environmental parameter data of the switchgear; (2) calculating the resistance change rate and three-phase imbalance rate of the contact resistance value based on the contact resistance value; (3) determining the first mechanical state and confidence level of the switchgear based on the comparison results of the mechanical health score with the preset health score, the resilience score with the preset resilience score, the resistance change rate with the resistance change rate threshold, and the three-phase imbalance rate with the three-phase imbalance rate threshold; (4) inputting the contact resistance value, resistance change rate, three-phase imbalance rate, mechanical health score, resilience score, environmental parameter data, and cumulative number of operations into the AI ​​model to obtain the second mechanical state and state probability of the switchgear; (5) determining the mechanical state of the switchgear based on the second mechanical state, state probability, first mechanical state, and confidence level.

[0180] Figure 4 This is an exemplary embodiment of the present application illustrating the operational interface diagram of a high- and low-voltage switchgear status assessment based on multi-dimensional features and AI recognition. Figure 4 The interface includes: a display window for mechanical characteristic parameters, a display window for health level scores, a display window for toughness scores, and a display window for mechanical status. The mechanical characteristic parameter display window shows the mechanical characteristic parameter data; the health level score display window shows the mechanical health level score of the switchgear; the toughness score display window shows the toughness score of the switchgear; and the mechanical status display window determines and displays the mechanical status of the switchgear based on the mechanical health level score and toughness score. Figure 4 The mechanical characteristic parameter data shows that although the speed parameter in the stroke curve of the circuit breaker (10kV level, cumulative operation 1,284 times) is good, indicating that the mechanical transmission chain and contact wear are normal, the time characteristic parameter (opening and closing time) and the bounce characteristic parameter (bounce duration) both show abnormal phenomena exceeding the reference value, triggering the system's reference alarm. This suggests that there is slight jamming in the operating mechanism, poor lubrication, or changes in electrical circuit parameters. It is recommended to include it in the key observation items or arrange for recent maintenance.

[0181] The following describes an embodiment of the apparatus described in this application, which can be used to execute the high- and low-voltage switchgear condition assessment system based on multi-dimensional features and AI recognition described in the above embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the high- and low-voltage switchgear condition assessment method based on multi-dimensional features and AI recognition described above in this application.

[0182] Figure 5 This is a block diagram illustrating a high- and low-voltage switchgear condition assessment system based on multi-dimensional features and AI recognition, as shown in an exemplary embodiment of this application.

[0183] like Figure 5 As shown, this exemplary high and low voltage switchgear condition assessment system 500 based on multi-dimensional features and AI recognition includes:

[0184] The data acquisition module 501 is used to acquire mechanical characteristic parameter data generated by the switchgear in the high and low voltage switchgear during the opening and closing operations.

[0185] The health score calculation module 502 is used to calculate the degree of parameter deviation, the degree of action integrity, and the degree of action coordination of the switchgear during the opening and closing operation based on the mechanical characteristic parameter data; and to determine the mechanical health score of the switchgear based on the degree of parameter deviation, the degree of action integrity, and the degree of action coordination.

[0186] The resilience rating calculation module 503 is used to calculate the degree of shock resistance, fluctuation resistance, and self-recovery of the switchgear during the opening and closing operation based on the mechanical characteristic parameter data; and to determine the resilience rating of the switchgear based on the degree of shock resistance, fluctuation resistance, and self-recovery.

[0187] The status determination module 504 is used to determine the mechanical status of the switchgear based on the mechanical health score and the toughness score, and to use the mechanical status as the evaluation result of the switchgear in terms of mechanical operation in high and low voltage switchgear.

[0188] In one embodiment of this application, the switching device includes a circuit breaker; mechanical characteristic parameter data includes: stroke curve, opening and closing time, bounce, synchronization difference, overtravel, etc. The stroke curve is the relationship curve of the displacement of the moving contact of the switch changing with time during the entire opening and closing movement (the horizontal axis is time, and the vertical axis is the travel displacement). The slope of the relationship curve reflects the movement speed of the moving contact of the switch; a sudden change in the slope reflects the jamming or buffering abnormality of the mechanism; the plateau segment of the relationship curve reflects the overtravel compression process after the moving contact of the switch; the oscillating segment of the relationship curve reflects the closing bounce or opening rebound. The real-time displacement data of the moving contact of the switch is directly collected by a displacement sensor, and the corresponding timestamp is recorded synchronously to obtain timestamp-displacement data. The stroke curve is plotted using the timestamp-displacement data. The displacement data is collected synchronously at a sufficiently high sampling rate (e.g., a sampling rate of not less than 10kHz) to ensure that short-term dynamic processes such as closing bounce, opening rebound, and buffering abnormality can be distinguished. Displacement measurement can be performed using a linear displacement sensor, a rotation angle sensor, etc.; or the displacement can be indirectly obtained by collecting acceleration signals through an accelerometer and performing integration calculations. The opening and closing times include: opening time and closing time. Opening time is the time interval from the moment the circuit breaker receives the opening command (when the opening coil is energized) until the moment all contacts on all poles separate. Opening time reflects the speed at which the circuit breaker clears fault current; the shorter the time, the better the protection effect. Closing time is the time interval from the moment the circuit breaker receives the closing command (when the closing coil is energized) until the moment all contacts on all poles make contact. Closing time reflects the response speed of the circuit breaker when it is put into operation. The acquisition process for the opening (or closing) time includes: capturing the energization moment of the opening coil (or closing coil) through Hall sensor or coil voltage detection; starting a high-precision timer (usually based on a DSP (Digital Signal Processor) or microcontroller, with a time resolution of up to 0.1ms); after the timer starts, synchronously sampling the break signal (or transient signal, displacement signal, etc.) at high speed; identifying the end point of the contact action through signal processing algorithms (e.g., short-time energy method, change-point detection method, high-frequency signal acquisition method, etc.); and calculating the time difference between the end point and the start point to obtain the opening (or closing) time. The break signal can be used to identify the moment of change in the on / off state of the moving and stationary contacts; the transient signal can be used to identify the transient characteristics generated at the moment of contact or separation of the moving and stationary contacts; and the displacement signal can be used to identify the key position changes during the movement of the moving contact.Bounce characterizes the phenomenon of repeated separation and contact between the moving and stationary contacts of a switch during the closing operation, caused by mechanical impact and incomplete dissipation of kinetic energy. Bounce includes: the number of bounces, bounce duration, and bounce amplitude. The number of bounces represents the number of times the moving and stationary contacts repeatedly separate (close) during the closing process. The bounce duration represents the time from the first closing of the moving and stationary contacts to the end of the last bounce (i.e., the contacts are stably closed and no longer separate). The bounce amplitude represents the maximum displacement value of the moving contact rebounding from the closed position during the closing operation. The calculation and acquisition of the number of bounces are achieved through methods such as the break-point signal method, displacement sensor method, and vibration sensor method. The calculation and acquisition of the bounce duration are achieved through methods such as the break-point signal method and displacement sensor method. The bounce amplitude is extracted from the stroke curve. The process involves differentiating the stroke curve to obtain the velocity curve. The moment the velocity curve abruptly changes from a positive value to zero (or the difference from zero is less than a difference threshold) is the moment the moving contact first reaches the closing position. The displacement value at this moment is the reference position for the bounce amplitude. Within the bounce interval (from the first closing to the last closing), the minimum displacement value in the stroke curve is found and taken as the farthest position the moving contact bounces back to during the bounce process. The difference between the reference position and the minimum displacement value is recorded as the bounce amplitude. Synchronization difference characterizes the synchronicity of the three-phase contact actions, including: inter-phase synchronization difference and intra-phase synchronization difference. Inter-phase synchronization difference represents the difference between the maximum and minimum closing times or the maximum and minimum opening times among the three phases. Intra-phase synchronization difference refers to the difference between the maximum and minimum closing times or the maximum and minimum opening times among multiple breaks within the same phase. Intra-phase synchronization difference is applicable to circuit breaker structures with multiple breaks within the same phase. Overtravel (also known as contact travel or overtravel) refers to the extra distance that the operating mechanism continues to push the moving contact forward after the moving and stationary contacts have made contact during the closing operation (at which time the contact pressure spring is compressed). Overtravel is extracted from the travel curve. The process of extracting overtravel includes: determining the point where the moving and stationary contacts just close by using the abrupt change point of the slope of the travel curve, the characteristic point of speed change, or the moment of on / off change of the break signal; determining the final travel of the moving contact at its final stopping position; and the displacement difference between the final travel and the travel at the point of just close is taken as overtravel.

[0189] In one embodiment of this application, the degree of parameter deviation is calculated based on the stroke curve, opening and closing time, bounce, and overtravel in the mechanical characteristic parameter data. The calculation formula for the degree of parameter deviation is shown in equation (1). The formulas for calculating the deviation of mechanical characteristic parameter data of each type are shown in Equation (2), the formulas for calculating the completeness of the action are shown in Equation (3), and the formulas for calculating the coordination of the action are shown in Equation (4).

[0190] In one embodiment of this application, when the synchronization difference is the phase-to-phase synchronization difference, formula (4) calculates the phase-to-phase coordination degree, and the average synchronization difference is the average of the phase-to-phase synchronization differences; when the synchronization difference is the same-phase synchronization difference, formula (4) calculates the same-phase coordination degree, and the average synchronization difference is the average of the same-phase synchronization differences. For a three-phase circuit breaker, the same-phase synchronization differences of phases A, B, and C can be calculated separately, resulting in three values.

[0191] In one embodiment of this application, after obtaining the coordination degree of interphase actions and the coordination degree of three in-phase actions, the minimum coordination degree among the interphase actions and the three in-phase actions is selected as the final coordination degree; or, the average value of the three in-phase actions is calculated, and the average value is weighted and summed with the interphase actions to obtain the final coordination degree; or, the minimum value among the three in-phase actions is selected, and the minimum value is weighted and summed with the interphase actions to obtain the final coordination degree.

[0192] The resilience score is used to characterize the dynamic response capability of switchgear during the opening and closing operations. The formulas for calculating the impact resistance are shown in Equation (5), the first speed factor is shown in Equation (6), the bounce factor is shown in Equation (7), the overtravel factor is shown in Equation (8), the fluctuation resistance is shown in Equation (9), the time factor is shown in Equation (10), the self-recovery factor is shown in Equation (11), the no-current time factor is shown in Equation (12), the short-time factor is shown in Equation (13), the secondary closing factor is shown in Equation (14), and the secondary closing bounce factor is shown in Equation (15).

[0193] In one embodiment of this application, when the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the opening-closing-opening action sequence, or when the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the closing-opening-closing action sequence, the degree of self-recovery can be calculated by formula (11). When the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the closing-opening-closing action sequence, the degree of self-recovery is calculated using the no-current time, the secondary closing time, the secondary closing bounce time, and the short time obtained from the last measurement.

[0194] In another embodiment of this application, the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the opening-closing-opening action sequence, and the self-recovery degree is calculated by formula (11). When the mechanical characteristic parameter data is the mechanical characteristic parameter data formed by the closing-opening-closing action sequence, the self-recovery degree is obtained by weighted summation of the no-current time, the secondary closing time, and the secondary closing bounce time.

[0195] In another embodiment of this application, the mechanical characteristic parameter data includes at least mechanical characteristic parameter data formed by the opening-closing-opening action sequence, or at least includes mechanical characteristic parameter data formed by the closing-opening-closing action sequence. When calculating the self-recovery level, the calculation is performed based on the mechanical characteristic parameter data formed by the opening-closing-opening action sequence most recently at the current calculation time, or the calculation is performed based on the mechanical characteristic parameter data formed by the closing-opening-closing action sequence most recently at the current calculation time. When calculating the mechanical health level score, the calculation is performed based on the mechanical characteristic parameter data formed by the opening action most recently at the current calculation time, or the calculation is performed based on the mechanical characteristic parameter data formed by the closing action most recently at the current calculation time.

[0196] In one embodiment of this application, mechanical characteristic parameter data generated by the switchgear during the opening and closing operations are acquired. The mechanical health of the switchgear is comprehensively evaluated from three dimensions: parameter deviation, operational integrity, and operational coordination. The dynamic response resilience of the switchgear is evaluated from three dimensions: shock resistance, fluctuation resistance, and self-recovery. The mechanical health score and resilience score are combined to determine the mechanical state of the switchgear. This achieves comprehensive perception and evaluation of multiple types of monitoring parameters, effectively overcoming the shortcomings of single-parameter evaluation methods that can only reflect a certain local state information of the switchgear and have insufficient ability to identify early deterioration characteristics. At the same time, it avoids the problems of pure data-driven models lacking physical mechanism support, poor interpretability, and insufficient dynamic adaptability. This improves the accuracy, reliability, and interpretability of the mechanical state evaluation of switchgear, providing a scientific basis for preventive maintenance and fault early warning of high and low voltage switchgear, and effectively ensuring the continuity and safety of power system operation.

[0197] It should be noted that the high- and low-voltage switchgear status assessment system based on multi-dimensional features and AI recognition provided in the above embodiments and the high- and low-voltage switchgear status assessment method based on multi-dimensional features and AI recognition provided in the above embodiments belong to the same concept. The specific operation methods of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the high- and low-voltage switchgear status assessment system based on multi-dimensional features and AI recognition provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0198] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for assessing the condition of high and low voltage switchgear based on multi-dimensional features and AI recognition, characterized in that: The method includes: Acquire mechanical characteristic parameter data generated by the switching equipment in the high and low voltage switchgear during the opening and closing operations; the mechanical characteristic parameter data includes: stroke curve, opening and closing time, bounce, synchronization difference, and overtravel; the switching equipment includes: circuit breaker; Based on the mechanical characteristic parameter data, the parameter deviation, action integrity, and action coordination of the switchgear during the opening and closing operation are calculated; and based on the parameter deviation, action integrity, and action coordination, the mechanical health score of the switchgear is determined. Based on the mechanical characteristic parameter data, the impact resistance, fluctuation resistance, and self-recovery degree of the switchgear during the opening and closing operation are calculated; and based on the impact resistance, fluctuation resistance, and self-recovery degree, the toughness score of the switchgear is determined; the toughness score is used to characterize the dynamic response capability of the switchgear during the opening and closing operation. Based on the mechanical health score and the toughness score, the mechanical condition of the switchgear is determined, and the mechanical condition is used as the evaluation result of the switchgear in the high and low voltage switch cabinet in terms of mechanical operation.

2. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 1, characterized in that, The process of determining the mechanical health score of the switching equipment based on the degree of deviation of the parameters, the degree of completeness of the action, and the degree of coordination of the action includes: The current health score of the switching device is obtained by weighted summing of the deviation of the parameters, the completeness of the action, and the consistency of the action. The current health score is fuzzified to obtain the membership degree of each health status level; The deviation of the parameters, the completeness of the action, and the coordination of the action are fuzzified to obtain the membership degree of each dimension parameter; and according to the preset fuzzy rules, fuzzy inference is performed on the membership degree of each health status level and the membership degree of each dimension parameter to obtain the activation intensity of each inferred health status level; and the activation intensity is normalized to obtain the belonging probability distribution of each inferred health status level. The mechanical health status score is obtained by weighting the representative scores of each reasoning health status level by using the probability distribution of each reasoning health status level as the weight; the representative score of each reasoning health status level is determined according to the score range of the corresponding reasoning health status level.

3. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 1 or 2, characterized in that, The process of determining the toughness score of the switchgear based on the shock resistance, the fluctuation resistance, and the self-recovery degree includes: The current toughness score of the switching device is obtained by weighted summing of the shock resistance, the fluctuation resistance, and the self-recovery. Obtain the historical resilience score of the switching device, and concatenate the current resilience score and the historical resilience score in chronological order to obtain a resilience score sequence; A first score segment is obtained by filtering from the resilience score sequence according to a first time span; and a second score segment is obtained by filtering from the resilience score sequence according to a second time span; the first time span is greater than the second time span, and the last resilience score of the first score segment and the last resilience score of the second score segment are both the current resilience score; Linear regression analysis is performed on the resilience scores in the first scoring segment to obtain a first resilience trend quantification index; and linear regression fitting is performed on the resilience scores in the second scoring segment to obtain a second regression slope; the first resilience trend quantification index includes: first regression slope, goodness of fit, and significance probability value. Based on the first resilience trend quantification index and the second regression slope, the current resilience score is corrected to obtain the resilience score.

4. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 3, characterized in that, The process of revising the current resilience score based on the first resilience trend quantification index and the second regression slope to obtain the resilience score includes: Based on the first resilience trend quantification index, determine the first correction factor; Multiply the current resilience score by the first correction factor to obtain the first resilience score; The second correction factor is determined based on the comparison between the first regression slope and the second regression slope. The first toughness score is multiplied by the second correction factor to obtain the toughness score.

5. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 4, characterized in that, The process of determining the second correction factor based on the comparison between the first regression slope and the second regression slope includes: If the second regression slope is less than the first regression slope, and the first regression slope is less than the first preset slope threshold, then the first correction factor is used as the second correction factor; the first correction factor is less than the benchmark factor. If the first regression slope is less than the second regression slope, and the second regression slope is less than the first preset slope threshold, then the second correction factor is used as the second correction factor; the second correction factor is greater than the benchmark factor. If the second regression slope is greater than the first preset slope threshold and the first regression slope is less than the first preset slope threshold, then the third correction factor is used as the second correction factor; the second correction factor is less than the third correction factor. If the absolute value of the difference between the second regression slope and the first preset slope threshold is less than the preset absolute value threshold, and the absolute value of the difference between the first regression slope and the first preset slope threshold is less than the preset absolute value threshold, then the benchmark ratio is used as the second correction factor. If the absolute value of the difference between the first regression slope and the first preset slope threshold is less than the preset absolute value threshold, and the absolute value of the second regression slope is greater than the second preset slope threshold, then the fourth correction factor is used as the second correction factor; the second preset slope threshold is greater than the first preset slope threshold; the fourth correction factor is less than the first correction factor. If the second regression slope is greater than the first regression slope, and the first regression slope is greater than the first preset slope threshold, then the fifth correction factor is used as the second correction factor; the fifth correction factor is greater than the third correction factor. If the second regression slope is less than the first regression slope, and the first regression slope is greater than the first preset slope threshold, then the sixth correction factor is used as the second correction factor; the sixth correction factor is greater than the first correction factor, and the sixth correction factor is less than the benchmark factor.

6. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 1 or 2, characterized in that, The process of determining the mechanical condition of the switchgear based on the mechanical health score and the toughness score includes: Acquire the contact resistance value, cumulative number of operations, and environmental parameter data of the switching device; the environmental parameter data includes: ambient temperature and humidity; Based on the contact resistance value, calculate the resistance change rate and the three-phase imbalance rate of the contact resistance value; Based on the comparison results of the mechanical health score and the preset health score, the comparison results of the toughness score and the preset toughness score, the comparison results of the resistance change rate and the resistance change rate threshold, and the comparison results of the three-phase imbalance rate and the three-phase imbalance rate threshold, the first mechanical state and confidence level of the switchgear are determined; the first mechanical state includes: normal state, warning state, defective state, and pending confirmation state; the resistance change rate threshold is determined based on the toughness score; the three-phase imbalance rate threshold is determined based on the toughness score. The contact resistance value, the resistance change rate, the three-phase imbalance rate, the mechanical health score, the toughness score, the environmental parameter data, and the cumulative number of operations are input into the AI ​​model to obtain the second mechanical state and state probability of the switchgear; the AI ​​model is obtained by training a preset model based on sample data; the second mechanical state includes: normal state, warning state, defect state, and pending confirmation state. The mechanical state of the switching device is determined based on the second mechanical state, the state probability, the first mechanical state, and the confidence level.

7. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 6, characterized in that, The process of determining the mechanical state of the switching device based on the second mechanical state, the state probability, the first mechanical state, and the confidence level includes: If the mechanical state with the highest confidence in the first mechanical state is the same as the mechanical state with the highest probability in the second mechanical state, then the mechanical state with the highest confidence or the mechanical state with the highest probability shall be taken as the mechanical state. If the mechanical state with the highest confidence in the first mechanical state is inconsistent with the mechanical state with the highest state probability in the second mechanical state, then the historical mechanical states of the switching device and the state transition probability vector under the historical mechanical states are obtained. The historical mechanical states and the state transition probability vector are multiplied to obtain the probability distribution under the predicted mechanical state. The probability value of the mechanical state with the highest confidence in the probability distribution is used as the matching degree between the predicted mechanical state and the mechanical state with the highest confidence, denoted as the first matching degree. The probability value of the mechanical state with the highest state probability in the probability distribution is used as the matching degree between the predicted mechanical state and the mechanical state with the highest state probability, denoted as the second matching degree. If the difference between the first matching degree and the second matching degree is greater than a preset difference threshold, then the mechanical state corresponding to the larger matching degree between the first matching degree and the second matching degree is taken as the mechanical state; If the difference between the first matching degree and the second matching degree is less than or equal to the preset difference threshold, then the mechanical state with the higher anomalous level between the mechanical state with the highest confidence and the mechanical state with the highest state probability is taken as the mechanical state.

8. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 6, characterized in that, The process of determining the first mechanical state and confidence level of the switchgear based on the comparison results of the mechanical health score with the preset health score, the toughness score with the preset toughness score, the resistance change rate with the preset resistance change rate, and the three-phase imbalance rate with the preset three-phase imbalance rate includes: The comparison result between the resistance change rate and the resistance change rate threshold is taken as the first comparison result; and the comparison result between the three-phase imbalance rate and the three-phase imbalance rate threshold is taken as the second comparison result. The loop resistance status is determined by the first comparison result and the second comparison result; the loop resistance status includes: normal status, three-phase common degradation status, single-phase local defect status, and data abnormal status. The comparison result between the mechanical health score and the preset health score is taken as the third comparison result; the comparison result between the toughness score and the preset toughness score is taken as the fourth comparison result. Based on the third comparison result, the fourth comparison result, and the loop resistance status, the first mechanical state of the switching device is determined; the first mechanical state includes: normal state, warning state, defect state, and pending confirmation state. Calculate the confidence level of evidence consistency, resistance deviation, toughness support, and three-phase consistency for the first mechanical state; the confidence level of evidence consistency is determined based on the third comparison result, the fourth comparison result, and the loop resistance state; the confidence level of resistance deviation is determined based on the resistance change rate; the confidence level of toughness support is determined based on the toughness score; and the confidence level of three-phase consistency is determined based on the three-phase imbalance rate. The confidence level of the first mechanical state is obtained by weighted summing of the confidence levels of the evidence consistency, resistance deviation, toughness support, and three-phase consistency.

9. The method for assessing the status of high and low voltage switchgear based on multi-dimensional features and AI recognition according to claim 6, characterized in that, The process of determining the resistance change rate threshold based on the toughness score includes: If the absolute value of the difference between the toughness score and the benchmark toughness score is less than the preset score difference threshold, then the benchmark judgment threshold is used as the resistance change rate threshold. If the absolute value of the difference between the toughness score and the benchmark toughness score is greater than or equal to the preset score difference threshold, and the toughness score is greater than the benchmark toughness score, then the resistance change rate threshold is calculated based on the benchmark judgment threshold and the first preset slope. When the absolute value of the difference between the toughness score and the benchmark toughness score is greater than or equal to the preset score difference threshold, and the toughness score is less than the benchmark toughness score, then the resistance change rate threshold is calculated based on the benchmark judgment threshold and the second preset slope; the absolute value of the second preset slope is greater than the absolute value of the first preset slope.

10. A high- and low-voltage switchgear condition assessment system based on multi-dimensional features and AI recognition, characterized in that: include: The data acquisition module is used to acquire mechanical characteristic parameter data generated by the switchgear in high and low voltage switchgear during the opening and closing operations; The mechanical characteristic parameter data includes: stroke curve, opening and closing time, bounce, synchronization difference, and overtravel; the switching equipment includes: circuit breaker; The health score calculation module is used to calculate the degree of parameter deviation, the degree of action integrity, and the degree of action coordination of the switchgear during the opening and closing operation based on the mechanical characteristic parameter data; and to determine the mechanical health score of the switchgear based on the degree of parameter deviation, the degree of action integrity, and the degree of action coordination. The resilience score calculation module is used to calculate the impact resistance, fluctuation resistance, and self-recovery of the switchgear during the opening and closing operation based on the mechanical characteristic parameter data; and to determine the resilience score of the switchgear based on the impact resistance, fluctuation resistance, and self-recovery; the resilience score is used to characterize the dynamic response capability of the switchgear during the opening and closing operation. The status determination module is used to determine the mechanical status of the switchgear based on the mechanical health score and the resilience score, and to use the mechanical status as the evaluation result of the switchgear in the high and low voltage switch cabinet in terms of mechanical operation.

Images