A fault diagnosis method and system for a station group heterogeneous circuit breaker

Abstract

The present application relates to the technical field of power system online monitoring, in particular to a fault diagnosis method and system for station group heterogeneous circuit breakers. By obtaining device parameter sets of each circuit breaker in the station group and grouping them into multiple device families, collecting switching operation data to extract characteristic vectors representing electrical and mechanical behaviors, converting the characteristic vectors to a unified dimension space according to the preset standardized mapping relationship of each device family, and constructing a group baseline model of each device family in the space, during fault diagnosis, inputting the standardized behavior vector corresponding to the target operation event of the circuit breaker to be diagnosed into the group baseline model of the device family to which it belongs, calculating its deviation degree and comparing it with the preset threshold to determine the operating state and fault diagnosis result, the circuit breaker state evaluation is expanded from single threshold judgment to comparative analysis based on the group behavior of device families, which is conducive to realizing unified and comparable diagnosis of heterogeneous circuit breaker operating behaviors in the station group.

Description

Technical Field

[0001] This invention relates to the field of power system online monitoring technology, specifically to a fault diagnosis method and system for heterogeneous circuit breakers in substation clusters. Background Technology

[0002] With the continuous expansion of power systems and the widespread application of intelligent devices, a large number of intelligent circuit breakers are deployed in substations and distribution stations to achieve line protection, control, and condition monitoring. Existing intelligent circuit breakers are typically equipped with functions such as coil current acquisition, operation time recording, and contact travel monitoring, which can collect relatively complete operating data during opening and closing operations, providing a data foundation for fault diagnosis and condition assessment.

[0003] In engineering practice, a power grid company often uses circuit breakers from different manufacturers, of different models, with different voltage levels, and different sampling configurations in multiple substations for extended periods. The differences in rated voltage, rated current, operating mechanism type, sampling frequency, and other device parameters among various circuit breakers result in inconsistencies in the dimensions, magnitudes, and dynamic response characteristics of the electrical and mechanical quantities during opening and closing processes. This makes it difficult to directly compare the operating data of different circuit breakers on the same scale.

[0004] In existing technologies, online monitoring and fault diagnosis of circuit breaker operating status are mostly based on single devices. They typically rely on threshold criteria, empirical rules, or feature-based discrimination models established for a specific circuit breaker model to evaluate its opening and closing behavior. Some solutions attempt to use historical operating data or waveform recordings for data-driven analysis, but these are often limited to a single manufacturer or model, or modeling only under the same device parameter configuration, making it difficult to uniformly evaluate the operating behavior of various heterogeneous circuit breakers across a network of substations.

[0005] On the other hand, within the same substation cluster, the number of circuit breakers is large and their operating conditions are varied. The characteristic deviation of a single circuit breaker may still be within the empirical threshold range from the perspective of a single unit, but it has already shown a significant deviation relative to the overall behavior of the same voltage level and type of equipment group. Existing single-unit diagnostic methods are difficult to use the statistical characteristics of equipment group behavior to detect such early potential problems in a timely manner. In addition, with the expansion and upgrading of substation clusters and equipment upgrades, new models of circuit breakers are constantly being added. The traditional method of relying on long-term historical data of a single unit to establish diagnostic criteria often lacks sufficient samples in the early stages of new equipment commissioning, making it difficult to form a reliable diagnostic basis in a timely manner. Existing historical data of the substation cluster is also difficult to directly transfer and utilize.

[0006] In summary, existing technologies generally lack a technical solution that can take into account the differences between circuit breaker devices from multiple manufacturers and of multiple models within a substation cluster, conduct comparative analysis of the group's operational behavior under a unified characteristic scale, and enable newly connected circuit breakers to be quickly incorporated into a unified diagnostic system even when a large amount of historical data is lacking. This makes it difficult to meet the engineering needs for consistent and scalable fault diagnosis of heterogeneous circuit breakers in a substation cluster. Summary of the Invention

[0007] (i) The technical problem to be solved by the present invention is that when multiple manufacturers, multiple models and different device parameter configurations of circuit breakers are used in a power plant cluster for a long time, although the existing technology can collect and analyze the opening and closing operation data of a single circuit breaker, it lacks a method to group circuit breakers based on device parameters, standardize the opening and closing behavior characteristics, build a baseline of normal behavior of the equipment group under a unified feature scale, and use the group deviation to carry out fault diagnosis. This makes it difficult to conduct unified and comparable status assessment and fault diagnosis of the operation behavior of heterogeneous circuit breakers within the power plant cluster.

[0008] (II) Technical Solution To address the aforementioned technical problems, this invention provides a fault diagnosis method for heterogeneous circuit breakers in a substation group, applicable to a substation group consisting of multiple circuit breakers from multiple substations, comprising the following steps: S1, obtain the device parameter set of each circuit breaker in the station group, and group the circuit breakers into multiple device families based on the device parameter set; S2, collect the operating data of each circuit breaker in the station group during the opening and closing operation events, and extract the behavioral feature vectors used to characterize the electrical and mechanical behavior of the opening and closing process; S3. Based on the device parameter set of each circuit breaker, a mapping relationship for standardization processing is preset for each device family. The mapping relationship is used to transform the behavior feature vector of the circuit breaker in the device family to a comparable unified dimension space, and the behavior feature vector extracted in step S2 is standardized according to the corresponding mapping relationship to obtain the standardized behavior vector corresponding to each opening operation event and closing operation event. S4. For each of the device families, construct a group baseline model representing the distribution of normal operating behavior based on the standardized behavior vector set obtained by the device family under preset normal operating conditions. S5. For the target opening and closing operation event of the circuit breaker to be diagnosed, obtain the corresponding standardized behavior vector, input the standardized behavior vector into the group baseline model of the device family to which the circuit breaker to be diagnosed belongs, calculate the group deviation of the standardized behavior vector relative to the group baseline model, and determine the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed based on the comparison result of the group deviation with the preset group threshold.

[0009] By acquiring the device parameter set of each circuit breaker within the substation cluster, and grouping the circuit breakers into multiple device families based on the device parameter set, behavioral feature vectors characterizing the electrical and mechanical behavior of the circuit breakers during opening and closing operation events are extracted from the operating data collected during these events. Furthermore, based on the pre-defined standardized mapping relationship of each device family, the behavioral feature vectors of the circuit breakers within that device family are transformed into a comparable unified dimension space. A group baseline model for each device family is constructed within this unified feature space. During fault diagnosis, the standardized behavioral vector corresponding to the target opening and closing operation event of the circuit breaker to be diagnosed is input into the group baseline model of its respective device family. The group deviation relative to the group baseline model is calculated and compared with a pre-defined group threshold to determine the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed. This facilitates unified and comparable fault diagnosis of the operating behavior of heterogeneous circuit breakers within the substation cluster.

[0010] According to an embodiment of the present invention, the standardized behavior vector set used to construct the population baseline model of each device family in step S4 is formed by selecting and storing the following standardized behavior vectors: Before the station group is put into online diagnostic operation, the operating data collected from the opening and closing operation events of each circuit breaker in the station group under preset normal operating conditions is standardized and processed to obtain and store multiple standardized behavior vectors; and During the online operation of the method, the standardized behavior vectors corresponding to multiple opening and closing operation events that occur under preset normal operating conditions and are not recorded within a preset observation time window are selected and stored in the standardized behavior vector set of the corresponding device family.

[0011] By constructing standardized behavior vector sets for each device family using the above method, the samples used to train the group baseline model come from two sources: firstly, basic test data collected under preset normal operating conditions before the station group is put into online diagnostic operation; and secondly, standardized behavior vectors that meet the conditions can be continuously supplemented from opening and closing operation events that have not been recorded and are associated with fault events during the online operation of the method. Thus, when modeling the group baseline, it can rely on high-confidence behavior data formed by factory tests, field handover tests, or initial operation to ensure that the initial baseline model has clear physical meaning and sufficient healthy sample support. Furthermore, in the subsequent long-term operation, as the operating conditions, load levels, and equipment aging status evolve, normal behavior data that has been verified by operation can be gradually introduced to dynamically expand and update the standardized behavior vector set. This allows the statistical characteristics of the group baseline model to gradually approach the normal behavior distribution under actual operating conditions. In this way, while avoiding the contamination of the baseline model by fault event samples, the representativeness and robustness of the baseline model to the actual operating state are improved, providing a more stable and reliable reference for subsequent fault diagnosis based on group deviation.

[0012] Furthermore, the method also includes the following during online operation: Within each preset update cycle, the model parameters of the population baseline model of the corresponding device family are updated based on the standardized behavior vectors of the standardized behavior vector set newly stored in each device family after the end of the previous update cycle.

[0013] By incrementally updating the model parameters of the group baseline model of the corresponding device family based solely on the standardized behavior vectors added to the standardized behavior vector set of each device family after the end of the previous update cycle within each preset update cycle, the correction of the group baseline model is carried out in batches at fixed time periods. This avoids frequent fluctuations in model parameters caused by a single running event immediately participating in the baseline recalculation, and also allows the group baseline model to be gradually corrected using normal behavior samples accumulated in each update cycle. Thus, while controlling the stability of the baseline, it allows it to evolve slowly with the running data, better reflecting the long-term changes in actual operating conditions.

[0014] According to an embodiment of the present invention, the method further includes the following during online operation: Within a preset statistical time window, for each circuit breaker in the substation group, a standardized behavior vector is selected corresponding to multiple opening and closing operation events that occur under preset normal operating conditions and are not recorded as associated with fault events within the preset observation time window. Based on the selected standardized behavior vector, multidimensional statistical features are calculated to construct an individual baseline model for characterizing the distribution of the circuit breaker's own operating behavior. By selecting standardized behavior vectors corresponding to multiple opening and closing operations that have been verified as non-fault events for each circuit breaker in the substation group within a preset statistical time window, and constructing an individual baseline model based on multidimensional statistical features calculated from these vectors, each circuit breaker can have a reference model that reflects its own long-term operating behavior distribution while sharing the baseline of the device family.

[0015] Step S5 further includes: based on the individual baseline model, calculating the individual deviation of the standardized behavior vector corresponding to the target opening and closing operation event relative to the individual baseline model, and using the individual deviation and the group deviation of the target opening and closing operation event relative to the group baseline model as the diagnostic basis for determining the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed.

[0016] When diagnosing target opening and closing operation events, the individual deviation is calculated based on the individual baseline model, and the group deviation of the event relative to the group baseline model is used as the diagnostic basis. This allows the diagnostic process to consider both the degree of change of the circuit breaker relative to its own historical behavior and its deviation from the normal behavior of a group of similar equipment. This helps to distinguish between behavioral changes caused by individual operating condition disturbances and suspected fault hazards that deviate from both individual and group baselines, thereby improving the precision of anomaly judgment and the reliability of diagnostic results.

[0017] Further, in step S5, based on the comparison results of the individual deviation degree and the group deviation degree with the preset individual threshold and the preset group threshold, respectively, the target opening and closing operation event is divided into at least one of the following modes: When the individual deviation is less than the preset individual threshold and the group deviation is less than the preset group threshold, the target opening and closing operation event is determined to be in normal operation mode. When the individual deviation is greater than or equal to the preset individual threshold and the group deviation is less than the preset group threshold, the target opening and closing operation event is determined to be in the operating condition disturbance mode. When the individual deviation is greater than or equal to the preset individual threshold and the group deviation is greater than or equal to the preset group threshold, the target opening and closing operation event is determined to be in the group out-of-group mode, and the circuit breaker corresponding to the target opening and closing operation event in the group out-of-group mode is marked as an object with potential fault.

[0018] By classifying the same type of deviation behavior into different modes such as normal operation, operating condition disturbance, and group outlier, the diagnostic results are clearer in distinguishing between general operating condition changes and suspected fault hazards. This helps maintenance personnel to conduct graded responses and key investigations according to the mode type, reducing the interference of false alarms on on-site operations.

[0019] According to an embodiment of the present invention, the device parameter set includes rated voltage, operating mechanism type and sampling frequency, and in step S1, circuit breakers with rated voltage of the same voltage level, the same operating mechanism type and the same sampling frequency are classified into the same device family.

[0020] According to an embodiment of the present invention, the operating data in step S2 includes coil current waveform data and operating time data. The behavioral feature vector extracted in step S2 includes: the coil current peak value and the duration of the current rise phase extracted from the coil current waveform data, and the total opening time and total closing time extracted from the operating time data.

[0021] Through the above settings, the device parameter set is clearly defined as rated voltage, operating mechanism type, and sampling frequency. Device families are divided according to the rules of having the same rated voltage level, operating mechanism type, and sampling frequency. This ensures high consistency of circuit breakers within the same device family in terms of rated operating conditions and data acquisition conditions, facilitating the subsequent establishment of a unified standardized mapping relationship and group baseline model within the device family, and reducing additional disturbances caused by differences in device parameters and sampling configurations. Furthermore, the operating data is limited to coil current waveform data and operating time data, from which features such as coil current peak value, current rise duration, total opening time, and total closing time are extracted. This allows the behavioral feature vector to centrally characterize the key electrical responses and mechanical action sequences of the opening and closing process. Under the premise of appropriate feature dimensions, it highlights dimensions closely related to mechanism action performance and fault symptoms, improving the engineering feasibility and diagnostic effectiveness of standardized processing and baseline modeling based on this feature space.

[0022] According to an embodiment of the present invention, when the parameter set of the access devices in the station group does not meet the parameter configuration requirements of any existing device family for the new circuit breaker, the method further includes: Configure the new circuit breaker as a circuit breaker in a new device family, and use the device parameter set of the new circuit breaker as the device parameter representative of the new device family; Based on the parameter differences between the device parameter representatives of the new device family and the device parameter representatives of each existing device family, the similarity between the new device family and each existing device family is calculated, and the existing device family with the highest similarity is selected as the target device family. Based on the differences in device parameters between the new device family and the target device family, the model parameters of the population baseline model of the target device family are corrected according to a preset parameter correction rule to obtain the initial model parameters of the population baseline model of the new device family.

[0023] Through the above steps, when a new circuit breaker with mismatched device parameter configurations is connected to a network of stations, the process no longer relies on accumulating sufficient samples after long-term operation to retrain the baseline model. Instead, the new circuit breaker is first assigned to a new device family, and using device parameter representatives as a bridge, the target device family with the closest similarity to the device parameter representatives of each existing device family is selected. Based on this, the parameters of the target device family's group baseline model are corrected according to preset parameter correction rules, directly generating the initial group baseline model of the new device family. Thus, considering the differences in device parameters, the group behavior knowledge of existing device families can be reused to provide an engineering-reasonable initial diagnostic baseline for the new circuit breaker. This shortens the cold start process from connection to usable fault diagnosis capability for the new device and avoids a prolonged state of "no baseline" or overly coarse criteria due to a lack of historical data.

[0024] This invention also provides a fault diagnosis system for heterogeneous circuit breakers in substation groups, applicable to substation groups consisting of multiple circuit breakers from multiple substations, including: An operation status acquisition unit is installed on each circuit breaker in the substation group. It is used to acquire operation data that characterizes the electrical and mechanical behavior of the opening and closing process during the opening and closing operation events of each circuit breaker. The operation data includes coil current waveform data and operation time data. The device parameter management and device family division module is used to obtain the device parameter set of each circuit breaker in the station group, and group the circuit breakers into multiple device families based on the device parameter set; The feature extraction and standardization module is used to extract behavioral feature vectors representing the electrical and mechanical behavior of the opening and closing process based on the operating data collected by the operating status acquisition unit, and to pre-set a mapping relationship for standardization processing for each device family based on the device parameter set of each circuit breaker. The mapping relationship is used to transform the behavioral feature vectors of the circuit breakers in the device family to a comparable unified dimension space, and to standardize the behavioral feature vectors according to the corresponding mapping relationship to obtain the standardized behavioral vectors corresponding to each opening operation event and closing operation event. The group baseline modeling module is used to construct a group baseline model representing the distribution of normal operation behavior for each device family based on the standardized behavior vector set obtained by the device family under preset normal operation conditions. The diagnosis and evaluation module is used to obtain the corresponding standardized behavior vector for the target opening and closing operation event of the circuit breaker to be diagnosed, input the standardized behavior vector into the group baseline model of the device family to which the circuit breaker to be diagnosed belongs, calculate the group deviation of the standardized behavior vector relative to the group baseline model, and output the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed based on the comparison result of the group deviation with the preset group threshold.

[0025] Through the above structural arrangement, the fault diagnosis system for heterogeneous circuit breakers in a substation group of the present invention forms a complete functional link on the hardware acquisition side and the substation group diagnosis side: the operation status acquisition unit provides the raw operation data required for the opening and closing process; the device parameter management and device family division module and the feature extraction and standardization module provide standardized behavior vector inputs organized by device family for group baseline modeling; the group baseline modeling module completes the modeling of the normal operation behavior distribution of each device family; and the diagnosis and evaluation module calculates the deviation of the target opening and closing operation events and outputs the diagnosis results based on this. Therefore, the system can directly support the execution of the above method steps in engineering deployment, enabling the integrated implementation of heterogeneous circuit breaker fault diagnosis within the substation group in a modular manner.

[0026] According to one embodiment of the present invention, the system further includes at least one of the following modules: The group baseline update module is used to add the standardized behavior vectors of multiple tripping and closing operation events that occur under preset normal operating conditions and are not recorded as associated with fault events within a preset observation time window during the online operation of the system to the standardized behavior vector set used to construct each device family. In each preset update cycle, the module updates the model parameters of the group baseline model of the corresponding device family based on the standardized behavior vectors added to the standardized behavior vector set after the end of the previous update cycle. The new device family migration module is used to configure a new circuit breaker as a circuit breaker in a new device family when the device parameter set of the new circuit breaker does not meet the device parameter configuration of any existing device family in the system. The module uses the device parameter set of the new circuit breaker as the device parameter representative of the new device family. Based on the parameter differences between the device parameter representative of the new device family and the device parameter representatives of each existing device family, the module calculates the similarity between the new device family and each existing device family. The module selects the existing device family with the highest similarity as the target device family. Based on the device parameter differences between the new device family and the target device family, the module corrects the model parameters of the group baseline model of the target device family according to a preset parameter correction rule to obtain the initial model parameters of the group baseline model of the new device family.

[0027] By setting up a group baseline update module and / or a new device family migration module in the system, the system can not only realize the basic group baseline modeling and diagnosis functions, but also periodically update the group baseline model of each device family according to preset rules using newly added standardized behavior vectors during online operation. On the other hand, when a new circuit breaker with access device parameters that do not belong to the existing device family is connected, the existing group baseline model is migrated and corrected based on the device parameter similarity to generate the initial group baseline model of the new device family. This provides support for the group baseline maintenance and new device family migration involved in the method from the system structure perspective.

[0028] (III) Beneficial effects of the present invention: When multiple circuit breakers with different device parameter configurations are used in the same station group, the normal behavior reference of the group can be constructed by using the operating behavior of multiple circuit breakers in the same device family under a unified characteristic scale, and fault diagnosis can be carried out based on the degree of deviation from the group behavior. This expands the evaluation of the circuit breaker operating status from threshold judgment of single data to comparative analysis based on the group behavior of the device family, which is conducive to unified and comparable fault diagnosis of the operating behavior of heterogeneous circuit breakers within the station group. Attached Figure Description

[0029] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram of a fault diagnosis system for heterogeneous circuit breakers in a station cluster, provided in one embodiment of the present invention. Figure 2 This is a flowchart illustrating a fault diagnosis method for heterogeneous circuit breakers in a station cluster, provided as an embodiment of the present invention. Detailed Implementation

[0031] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Specific Implementation

[0032] In this embodiment, the fault diagnosis system for heterogeneous circuit breakers in a substation cluster is preferably deployed on a substation cluster diagnosis platform, such as an industrial control server within a regional dispatch center or distribution automation master station, and interacts with smart circuit breakers in multiple substations through existing communication networks. This allows for centralized analysis and management of the operating status of circuit breakers within the substation cluster without altering the physical primary equipment structure.

[0033] like Figure 1 As shown, the fault diagnosis system includes functional units such as an operation status acquisition unit, a device parameter management and device family division module, a feature extraction and standardization module, a group baseline modeling module, and a diagnosis and evaluation module. The operation status acquisition unit is correspondingly installed on each circuit breaker within the substation group. It can be composed of a sampling module inside the circuit breaker's intelligent control unit, a current sensor connected in series or parallel with the coil circuit, and a contact position detection device for recording the opening and closing operation time. It is used to collect operating data characterizing the electrical and mechanical behavior of the opening and closing process during opening and closing operation events of each circuit breaker. The operating data includes at least coil current waveform data and operation time data, and is uploaded to the substation group diagnostic platform via fieldbus or Ethernet.

[0034] The device parameter management and device family division module is set on the station group diagnostic platform. It can consist of a database and supporting software processing programs. It is used to obtain the device parameter set of each circuit breaker in the station group from the equipment asset management system, the setting value management system or manual input. In this embodiment, it includes parameters such as rated voltage, operating mechanism type and sampling frequency. According to the preset division rules, circuit breakers with the same device parameter configuration or that meet the predetermined conditions are grouped into multiple device families so that unified modeling can be performed at the device family granularity in the future.

[0035] The feature extraction and standardization module also runs within the station group diagnostic platform. It can be implemented by the processor and its running feature processing program. Based on the coil current waveform data and operation time data uploaded by the operating status acquisition unit, it obtains behavioral feature vectors characterizing the electrical and mechanical behavior of the opening and closing processes according to preset feature extraction rules. Combined with the circuit breaker device parameter sets provided by the device parameter management and device family division module, it presets a mapping relationship for standardization processing for each device family. This mapping relationship is configured to transform the behavioral feature vectors of the circuit breakers within the device family to a comparable unified dimension space, and accordingly standardize the behavioral feature vectors to obtain the standardized behavioral vectors corresponding to each opening and closing operation event.

[0036] The group baseline modeling module can consist of a statistical modeling program and a model storage unit. For each device family, it calls the standardized behavior vector output by the feature extraction and standardization module, accumulates it under preset normal operating conditions to form a standardized behavior vector set, and constructs a group baseline model representing the distribution of normal operating behavior of the device family based on this standardized behavior vector set. The diagnosis and evaluation module is set on the same diagnostic platform. Upon receiving the standardized behavior vector corresponding to the target opening and closing operation event of the circuit breaker to be diagnosed, it inputs the standardized behavior vector into the group baseline model of the device family to which the circuit breaker to be diagnosed belongs, calculates its group deviation relative to the group baseline model, and outputs the operating status and fault diagnosis result of the circuit breaker to be diagnosed based on the comparison result of the group deviation with a preset group threshold. The operating status and diagnosis result can be displayed through a human-machine interface or recorded and distributed through the operation and maintenance system to support subsequent operation and maintenance decisions.

[0037] Based on the above system structure, such as Figure 2 As shown, this embodiment provides a fault diagnosis method for heterogeneous circuit breakers in a substation group, running on a substation group diagnostic platform. This method is used for centralized operational status assessment and fault diagnosis of a substation group consisting of multiple circuit breakers from multiple substations. The method can be completed collaboratively by software programs in the substation group diagnostic server, calling the aforementioned operational status acquisition unit and various functional modules. Specifically, it includes the following steps.

[0038] Step S1: Obtaining device parameter sets and dividing device families First, step S1 is executed to obtain the device parameter set of each circuit breaker in the station group, and the circuit breakers are grouped into multiple device families based on the device parameter set. The device parameter set includes rated voltage, operating mechanism type and sampling frequency. In step S1, circuit breakers with rated voltage of the same voltage level, the same operating mechanism type and the same sampling frequency are divided into the same device family.

[0039] In this embodiment, the device parameter set can be obtained through the equipment asset management system, relay protection setting management system, and / or input by maintenance personnel. To balance implementation cost and feature differentiation, the device parameter set includes at least three parameters: rated voltage, operating mechanism type, and sampling frequency. Optionally, it may also include parameters such as rated current, insulation medium type, control power supply voltage, coil rated voltage, and sampling range, used to further refine the device family classification in other embodiments. The rated voltage can be classified according to commonly used voltage levels in engineering. For example, circuit breakers with a nominal voltage in the 6kV–12kV range are classified as "10kV level," and circuit breakers in the 27kV–40kV range are classified as "35kV level." The corresponding voltage level can be pre-configured in the system parameters. The operating mechanism type can be recorded as one of spring-operated, electromagnetic, or hydraulic operating mechanisms. The sampling frequency is the sampling rate of the coil current and related signals, such as 2kHz, 5kHz, or 10kHz.

[0040] In specific classification, the station group diagnostic server can traverse all circuit breakers within the station group, using the triplet "voltage level, operating mechanism type, sampling frequency" as the grouping key to classify circuit breakers with identical triplet values ​​into the same device family. In cases of slight deviations in sampling frequency, such as when the actual sampling frequency of the field device is 2048Hz while the system configuration is 2kHz, the acquired waveform can be resampled or interpolated to convert it to a unified target sampling frequency. Circuit breakers that meet the target sampling frequency requirements after this preprocessing can still be classified into the same device family. For a device family containing only one circuit breaker, this embodiment still establishes a group baseline model corresponding to that device family. However, in subsequent diagnostics, it can be combined with the individual baseline model and individual deviation (described later) for auxiliary judgment to avoid statistical instability caused by insufficient sample size.

[0041] Step S2: Run data acquisition and construct behavioral feature vectors After completing the device family division, step S2 is executed to collect the operating data of each circuit breaker in the substation group during the opening and closing operation events, and extract behavioral feature vectors to characterize the electrical and mechanical behavior of the opening and closing process. The operating data in step S2 includes coil current waveform data and operation time data. The behavioral feature vectors extracted in step S2 include: the peak value of the coil current and the duration of the current rise phase extracted from the coil current waveform data, and the total opening time and total closing time extracted from the operation time data.

[0042] Specifically, when the circuit breaker receives a tripping or closing control command, the operation status acquisition unit can continuously sample the coil current within a preset pre-trigger time (e.g., tens of milliseconds) before the control command is issued and a preset delay time after the auxiliary contact or position switch detects the end of the action, forming a corresponding coil current waveform data sequence; at the same time, the control unit records the timestamp of the tripping or closing control command and the timestamp of the auxiliary contact switching from the initial state to the stable disconnecting / closing state, thereby obtaining the operation time data.

[0043] The peak value of the coil current can be defined as the maximum sampled value of the coil current waveform within the aforementioned sampling time window. In practical implementation, the waveform can be simply denoised or averaged before sampling to reduce the impact of single-point noise. The duration of the current rise phase can be defined as the time interval during which the coil current rises from an initial value close to zero to a value close to the peak value. For example, the moment when the current first exceeds the peak current by a certain percentage (e.g., 10% to 20%) can be taken as the start time of the rise, and the moment when it first reaches a certain percentage (e.g., 95% to 100%) of the peak current can be taken as the end time of the rise. The time difference between the two is the duration of the current rise phase. The total opening time and the total closing time can be defined as the time interval from the moment the corresponding opening / closing control command is issued until the moment when the auxiliary contact signal indicates that the main contacts of the circuit breaker have stably opened / closed. In some embodiments, the end time of the action can also be determined by the contact position signal output by the travel sensor.

[0044] In this embodiment, the behavioral feature vector can be constructed by concatenating the four scalar features in a fixed order to form a four-dimensional vector, namely, behavioral feature vector x=[I_peak,T_rise,T_open,T_close], where I_peak represents the peak value of the coil current, T_rise represents the duration of the current rise phase, T_open represents the total opening time, and T_close represents the total closing time. Since these features reflect both the electromagnetic driving force level and the overall operating speed of the mechanism, the electrical and mechanical behavior of the opening and closing process can be jointly characterized without introducing complex travel measurement devices. Those skilled in the art can further extend this by introducing additional features such as contact overtravel and contact rebound when needed, but this embodiment only requires the above four features to form an executable basic solution.

[0045] Step S3: Standardize mapping configuration and construct a unified dimensional space After obtaining the behavior feature vector, step S3 is executed. Based on the device parameter set of each circuit breaker, a mapping relationship for standardization processing is preset for each device family. The mapping relationship is used to transform the behavior feature vector of the circuit breaker in the device family to a comparable unified dimension space. The behavior feature vector extracted in step S2 is then standardized according to the corresponding mapping relationship to obtain the standardized behavior vector corresponding to each opening operation event and closing operation event.

[0046] In this embodiment, to eliminate the differences in characteristic dimensions and scales between different device families due to variations in rated voltage, mechanical configuration, and sampling configuration, the station group diagnostic platform can preset a set of standardized scale parameters for each device family k, denoted as s_k=[s_k1,s_k2,s_k3,s_k4]. Each scale parameter can be obtained from historical samples of the device family under preset normal operating conditions. For example, s_k1 can be taken as the reference value of the peak value of the coil current of the device family, and s_k2, s_k3, and s_k4 can be taken as reference values ​​for the duration of the current rise phase, the total opening time, and the total closing time, respectively. In specific implementation, the reference values ​​can be selected as the mean, mode, or median of the corresponding feature in historical normal samples.

[0047] For any behavioral feature vector x = [x1, x2, x3, x4] belonging to device family k, it can be standardized by scale normalization to obtain a standardized behavioral vector z = [z1, z2, z3, z4], where each component satisfies: z1=x1 / s_k1, z2 = x2 / s_k2, z3=x3 / s_k3, z4=x4 / s_k4.

[0048] In other embodiments, the standardized features can be centered based on scale normalization, for example, by subtracting the mean of each feature dimension or dividing by the standard deviation, so that the standardized behavior vectors have a more concentrated distribution in a unified dimension space. This embodiment configures standardized mapping relationships at the device family granularity, so that the behavior feature vectors of different circuit breakers within the same device family under different rated parameters and mechanical configurations are mapped to a comparable dimensionless feature space, facilitating subsequent baseline modeling and deviation calculation.

[0049] Step S4: Population Baseline Model Construction Then, step S4 is executed, whereby for each device family, a group baseline model representing the distribution of normal operating behavior is constructed based on the standardized behavior vector set obtained by the device family under preset normal operating conditions.

[0050] In this embodiment, the preset normal operating conditions can be set according to the distribution network operation standards, and generally at least meet the following conditions: the corresponding opening and closing operation event does not trigger protection tripping or abnormal alarms; the load current carried by the circuit breaker at the time of opening and closing is within the allowable range of the voltage level and circuit design; the bus voltage is within the allowable deviation range of the rated voltage; and no fault work order or defect record associated with the operation event is recorded within the preset observation time window. For each device family k, the system can accumulate standardized behavior vectors corresponding to multiple opening and closing operation events that meet the above conditions during the initial commissioning and online operation, and construct a standardized behavior vector set for that device family.

[0051] Based on this standardized behavior vector set, the group baseline modeling module can use a multidimensional statistical model to model the normal operation behavior of the device family. For example, it can be assumed that the standardized behavior vector of the device family under normal operating conditions approximately follows a multidimensional Gaussian distribution. The mean vector mu_k and covariance matrix Sigma_k of the normal samples of the device family can be calculated, and (mu_k, Sigma_k) can be used as the parameters of the group baseline model of the device family. In other implementations, the group baseline model can also be constructed using kernel density estimation or statistical models based on cluster centers, etc., which is not limited in this embodiment. In the above way, a benchmark model describing the statistical distribution of its normal opening and closing behavior can be obtained for each device family, which can be used to determine whether subsequent operating events deviate from the normal behavior distribution during the diagnostic phase.

[0052] Step S5: Calculation of Group Deviation and Determination of Operating Status Finally, step S5 is executed: for the target opening and closing operation event of the circuit breaker to be diagnosed, the corresponding standardized behavior vector is obtained, the standardized behavior vector is input into the group baseline model of the device family to which the circuit breaker to be diagnosed belongs, the group deviation of the standardized behavior vector relative to the group baseline model is calculated, and the operating status and fault diagnosis result of the circuit breaker to be diagnosed are determined based on the comparison result of the group deviation with the preset group threshold.

[0053] Specifically, when a circuit breaker in the substation group performs a tripping or closing operation, the system obtains the standardized behavior vector z corresponding to the target tripping or closing operation event according to the processing flow of steps S2 and S3. The diagnosis and evaluation module, based on the device family k to which the circuit breaker belongs, calls the corresponding group baseline model parameters (mu_k, Sigma_k) to calculate the group deviation D_g of the event. In this embodiment, the group deviation can be defined using Mahalanobis distance, for example: D_g=(z−mu_k)^T*inv(Sigma_k)*(z−mu_k), Where inv(Sigma_k) represents the inverse of the covariance matrix Sigma_k, and (·)^T represents the vector transpose. This deviation reflects the degree of dispersion of the current operational event within the normal behavior distribution of the device family; a larger value indicates a greater degree of deviation from normal group behavior. The preset group threshold can be set during the system tuning phase based on the group deviation distribution of historical normal samples. For example, the Mahalanobis distance quantile corresponding to a certain confidence level (such as 95% or 99%) can be selected as the threshold, or it can be appropriately relaxed or tightened based on the risk preferences of the operation and maintenance unit.

[0054] During the comparison phase, when the group deviation of a target opening / closing operation event is less than a preset group threshold, the diagnosis and evaluation module can determine the corresponding operating status of the event as a normal state from a group perspective and output a "normal operation" diagnostic result. When the group deviation is greater than or equal to the preset group threshold, the event can be determined as an abnormal event outside the normal behavior distribution of the device family. Without considering subsequent individual baseline constraints, the diagnosis and evaluation module can generate a diagnostic result of "abnormal deviation exists" or "suspected fault behavior" and mark the circuit breaker as an object that needs to be tracked closely. Combined with the individual baseline model and individual deviation introduced in subsequent embodiments, abnormal events can be further subdivided and interpreted to support more refined operation and maintenance decisions.

[0055] Building upon the aforementioned group baseline model constructed based on device families, this embodiment also includes a group baseline update module to prevent the group baseline model from remaining in an initial state for an extended period and gradually deviating from the actual operating characteristics of the equipment. This module dynamically maintains the standardized behavior vector set used for modeling and the corresponding model parameters during system online operation. Specifically, during system online operation, for multiple tripping and closing operation events that occur under preset normal operating conditions and are not recorded as associated with fault events within a preset observation time window, the group baseline update module adds their corresponding standardized behavior vectors to the standardized behavior vector set used to construct each device family. Furthermore, within each preset update cycle, based on the standardized behavior vectors added to the standardized behavior vector set after the end of the previous update cycle, the module updates the model parameters of the group baseline model for the corresponding device family.

[0056] In specific implementation, the standardized behavior vector set used to construct the group baseline model of each device family in step S4 is formed by selecting and storing the following standardized behavior vectors: On the one hand, before the station group is put into online diagnostic operation, the operation data of the manufacturer's type test records, field handover test records, and the opening and closing operation events that have been confirmed to be without abnormalities during the initial energized trial operation of the circuit breaker can be standardized according to the processing methods of steps S2 and S3 to obtain and store multiple standardized behavior vectors as the initial standardized behavior vector set of each device family; on the other hand, during the online operation of the method, the system can check the action type recorded by the protection device, the fault recording file, and the operation and maintenance work order system. The fault maintenance record corresponding to the timestamp of the target opening and closing operation event is used to determine whether the opening and closing operation event is associated with a fault event. For example, when the opening cause is marked as short-circuit fault protection action, ground fault protection action, or other types, or when there is a fault work order matching the event timestamp within the preset observation time window, the event is marked as associated with a fault event and is not included in the standardized behavior vector set. When the current and voltage of the circuit to which the circuit breaker belongs are within the preset normal operating range when the opening and closing operation event occurs, and no fault alarm or maintenance record associated with the event timestamp is recorded within the preset observation time window, the opening and closing operation event is regarded as a candidate normal event, and its corresponding standardized behavior vector is added to the standardized behavior vector set of the corresponding device family after standardization processing. To avoid a single circuit breaker or a certain operating condition having an excessively high proportion in the standardized behavior vector set, this embodiment can limit the number of standardized behavior vectors contributed by the same circuit breaker to no more than a preset upper limit within the preset statistical time window, and prioritize the retention of samples under different load ranges and different environmental conditions in the newly added samples, so as to improve the coverage of the standardized behavior vector set of the normal operating behavior of the device family.

[0057] Furthermore, to avoid increasing computational overhead by completely reconstructing the group baseline model every time a new standardized behavior vector is added, the method further includes the following during online operation: within each preset update cycle, based on the standardized behavior vectors added to the standardized behavior vector set of each device family after the end of the previous update cycle, the model parameters of the group baseline model for the corresponding device family are updated. Specifically, the model parameters of the current group baseline model and the cumulative new standardized behavior vectors since the end of the previous update cycle can be maintained for each device family in the station group diagnostic platform. When the preset update cycle is reached, the model parameters are corrected in a recursive manner. For example, for each new standardized behavior vector x_new, the mean vector can be updated in the form μ_new=(1-α)*μ_old+α*x_new, where μ_old is the mean vector at the previous time step, α is a weight coefficient tuned according to the number of new samples and the expected model response speed, and the covariance matrix can be corrected by referring to the mean update result using the corresponding recursive formula. The preset update cycle can be set to daily, weekly, or triggered when the number of newly added standardized behavior vectors reaches a preset threshold, based on the operating frequency and data volume of circuit breakers within the substation group. This achieves a balance between the model's ability to track slow changes in operating conditions and the consumption of computational resources. Through this design, the group baseline model can both establish an initial normal behavior distribution using high-quality prior samples before commissioning and continuously update during online operation while ensuring sample quality. This improves the applicability of the group baseline model throughout the circuit breaker's entire lifespan and the reliability of diagnostic results.

[0058] Building upon the aforementioned group baseline model constructed and updated online based on device families, this embodiment further introduces an individual baseline model to characterize the distribution of normal operating behavior of each circuit breaker within the substation group over a period of time. This allows for simultaneous consideration of two dimensions during diagnosis: whether it is abnormal compared to its past performance and whether it is out of the group compared to similar devices. To this end, the online operation of the method also includes: within a preset statistical time window, for each circuit breaker within the substation group, selecting standardized behavior vectors corresponding to multiple opening and closing operation events occurring under preset normal operating conditions and not recorded as associated with fault events within the preset observation time window; calculating multidimensional statistical features based on the selected standardized behavior vectors to construct an individual baseline model characterizing the distribution of the circuit breaker's own operating behavior; wherein, the preset statistical time window can be configured according to the substation group's operating characteristics by time length or by the number of events. For example, it can select operating records from the most recent period or select several recent opening and closing operation events judged as normal, ensuring sufficient sample size while avoiding excessive model lag behind the current state of the equipment. In practical implementation, the standardized behavior vectors selected by the circuit breaker within a preset statistical time window can be regarded as "normal behavior samples" of the circuit breaker. The mean vector and covariance matrix of each dimension of these samples, or other statistical quantities that can reflect the multidimensional distribution, are calculated to form an individual baseline model describing the distribution of the circuit breaker's normal operating behavior. In this embodiment, the individual baseline model and the aforementioned group baseline model can adopt the same modeling method in statistical form; however, the former's sample source is limited to normal events of a single circuit breaker within the statistical time window, while the latter's sample source is the normal events of multiple circuit breakers within the device family.

[0059] After obtaining the individual baseline model, step S5 further includes: based on the individual baseline model, calculating the individual deviation of the standardized behavior vector corresponding to the target opening and closing operation event relative to the individual baseline model, and using the individual deviation and the group deviation of the target opening and closing operation event relative to the group baseline model as the diagnostic basis for determining the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed. In this embodiment, the operating status and fault diagnosis result are used. The individual deviation can be calculated in the same or similar way as the group deviation. For example, a distance metric based on the mean vector and covariance matrix of the individual baseline model can be used to quantify the difference between the standardized behavior vector of the target event and its individual baseline model into a non-negative real number. The larger the value, the more abnormal the behavior of the event is relative to the historical normal behavior of the circuit breaker itself. By considering both the individual deviation and the group deviation simultaneously, it is possible to distinguish between pure load fluctuations, changes in the operating conditions of a single circuit breaker, and potential fault hazards with group outlier characteristics during the diagnosis process, thereby improving the pertinence of the diagnosis results.

[0060] To facilitate online application, in step S5 of this embodiment, based on the comparison results of the individual deviation and the group deviation with preset individual thresholds and preset group thresholds respectively, the target opening and closing operation event is divided into at least one of the following modes: When the individual deviation is less than the preset individual threshold and the group deviation is less than the preset group threshold, the target opening and closing operation event is determined to be in normal operation mode. At this time, the event belongs to the normal fluctuation range in both the circuit breaker's own historical behavior and the behavior of the device family group, and can be used as a candidate normal sample for subsequent baseline model updates; when the individual deviation is greater than or equal to the preset individual threshold and the group deviation is less than the preset group threshold, the target opening and closing operation event is determined to be in operating condition disturbance mode. Such events usually reflect the circuit breaker's own operating behavior. If a significant change has occurred relative to its historical level, such as a change in load level, wear along the operating mechanism, or a change in lubrication status, but this behavior is still interpretable within the normal distribution range of the corresponding device family, it can be used to alert maintenance personnel to pay attention to the trend of the circuit breaker's operating condition. When the individual deviation is greater than or equal to the preset individual threshold and the group deviation is greater than or equal to the preset group threshold, the target opening and closing operation event is determined to be in the group outlier mode, and the circuit breaker corresponding to the target opening and closing operation event in the group outlier mode is marked as an object with potential faults. At this time, the event deviates from both the normal behavior distribution of the circuit breaker itself and the group baseline distribution of similar circuit breakers, and is more likely to be related to early faults such as mechanical jamming of the mechanism and deterioration of coil insulation. Subsequently, it can be combined with alarm strategies to trigger in-depth inspections or maintenance work orders.

[0061] The aforementioned preset individual thresholds and preset group thresholds can be tuned based on the deviation distribution of historical normal operation events. For example, during the system trial operation phase, the individual and group deviations of a large number of confirmed fault-free opening and closing operation events can be statistically analyzed. The initial threshold value is determined based on the high quantile of the corresponding distribution (e.g., the quantile corresponding to a certain preset probability level), and appropriate adjustments are made during long-term operation in conjunction with the false alarm rate and the risk of missed alarms. Through this joint deviation discrimination mechanism based on the individual baseline model and the group baseline model, this embodiment can highlight high-risk events with obvious outlier characteristics while ensuring that normal operation events do not trigger excessive alarms, facilitating hierarchical handling by operation and maintenance personnel.

[0062] Building upon the aforementioned group baseline model and individual baseline model, this embodiment also considers that new types and configurations of circuit breakers will be continuously connected to the substation cluster throughout its lifecycle. To avoid accumulating samples and retraining the baseline model from scratch every time a new circuit breaker is connected, this embodiment provides a transfer modeling mechanism based on device parameter similarity. Therefore, the fault diagnosis system for heterogeneous circuit breakers in a substation cluster also includes a new device family migration module. This module is used to configure the new circuit breaker as a circuit breaker in a new device family when the device parameter set connected to the system does not meet the device parameter configuration of any existing device family, and to use the device parameter set of the new circuit breaker as the representative of the device parameters of that new device family.

[0063] Specifically, the device parameter representation can be viewed as a multi-dimensional parameter vector composed of parameters such as rated voltage, rated current, and sampling frequency in a fixed order. In practical applications, to eliminate the influence of different physical dimensions, each dimension of the parameter can be normalized first. For example, the rated voltage can be divided by the reference value of the corresponding voltage level, the rated current by the reference value of the rated current of the circuit breaker in the substation group, and the sampling frequency by the predetermined sampling frequency reference value, thereby obtaining dimensionless parameter components. Subsequently, the parameter differences between the new device family and each existing device family can be calculated using a weighted Euclidean distance method. For example, for each existing device family, the distance between its device parameter representation and the device parameter representation of the new device family in the normalized parameter space can be calculated, and the device family with the smaller distance can be regarded as having higher similarity. In an optional embodiment, weight coefficients can be configured for rated voltage, rated current, and sampling frequency to reflect the degree of consistency influence of different parameters on the behavioral feature scale, and a single similarity index can be obtained. This allows the existing device family with the highest similarity to be selected as the target device family, realizing automatic matching between the new device family and the existing device family.

[0064] After determining the target device family, this embodiment modifies the model parameters of the group baseline model of the target device family according to a preset parameter correction rule to obtain the initial model parameters of the group baseline model of the new device family. From an implementation perspective, the mean values ​​of current amplitude-related features in the group baseline model of the target device family can be scaled according to the ratio of the rated current of the old and new device families. For example, the mean component representing the peak value of the coil current can be linearly amplified or reduced according to the ratio of "rated current of the new device family / rated current of the target device family". For time-related features such as total opening time and total closing time, which are mainly determined by mechanical mechanisms, the mean components of the target device family can be directly inherited when the operating mechanism type is the same. When the operating mechanism type is different, a preset time compensation amount can be superimposed on the mean of the target device family to reflect the systematic differences brought about by different mechanism levels. Meanwhile, to reflect the objective situation of insufficient statistical samples and high model uncertainty in the early stage of new equipment family deployment, the statistical distribution of the target equipment family can be appropriately amplified at the covariance level. For example, the covariance matrix of the target equipment family's group baseline model can be scaled as a whole by a preset amplification factor greater than 1, so that the initial group baseline model of the new equipment family has a more lenient tolerance range for deviation behavior. Through the above parameter correction rules, the new equipment family can obtain a set of initial model parameters from similar equipment families and compensated for parameter differences even in the absence of its own historical data. Subsequently, as the standardized behavior vectors collected by the new equipment family continue to increase, its group baseline model can gradually transition from migration parameters to statistical distribution based on its own operating data through the aforementioned group baseline update mechanism, thereby suppressing early misjudgments and accelerating the convergence speed of online diagnosis of new circuit breakers.

[0065] The above are preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made to the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A fault diagnosis method for heterogeneous circuit breakers in a substation group, applicable to a substation group consisting of multiple circuit breakers from multiple substations, characterized in that, include: S1, obtain the device parameter set of each circuit breaker in the station group, and group the circuit breakers into multiple device families based on the device parameter set; S2, collect the operating data of each circuit breaker in the station group during the opening and closing operation events, and extract the behavioral feature vectors used to characterize the electrical and mechanical behavior of the opening and closing process; S3. Based on the device parameter set of each circuit breaker, a mapping relationship for standardization processing is preset for each device family. The mapping relationship is used to transform the behavior feature vector of the circuit breaker in the device family to a comparable unified dimension space, and the behavior feature vector extracted in step S2 is standardized according to the corresponding mapping relationship to obtain the standardized behavior vector corresponding to each opening operation event and closing operation event. S4. For each of the device families, construct a group baseline model representing the distribution of normal operating behavior based on the standardized behavior vector set obtained by the device family under preset normal operating conditions. S5. For the target opening and closing operation event of the circuit breaker to be diagnosed, obtain the corresponding standardized behavior vector, input the standardized behavior vector into the group baseline model of the device family to which the circuit breaker to be diagnosed belongs, calculate the group deviation of the standardized behavior vector relative to the group baseline model, and determine the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed based on the comparison result of the group deviation with the preset group threshold.

2. The fault diagnosis method for heterogeneous circuit breakers in a station group according to claim 1, characterized in that, The standardized behavior vector set used to construct the population baseline model of each device family in step S4 is formed by selecting and storing the following standardized behavior vectors: Before the station group is put into online diagnostic operation, the operation data collected from the opening and closing operation events of each circuit breaker in the station group under the preset normal operation conditions are processed and stored as multiple standardized behavior vectors. as well as During the online operation of the method, the standardized behavior vectors corresponding to multiple opening and closing operation events that occur under preset normal operating conditions and are not recorded within a preset observation time window are selected and stored in the standardized behavior vector set of the corresponding device family.

3. The fault diagnosis method for heterogeneous circuit breakers in a station group according to claim 2, characterized in that, The method also includes the following during online operation: Within each preset update cycle, the model parameters of the population baseline model of the corresponding device family are updated based on the standardized behavior vectors of the standardized behavior vector set newly stored in each device family after the end of the previous update cycle.

4. The fault diagnosis method for heterogeneous circuit breakers in a substation group according to any one of claims 1 to 3, characterized in that, The method also includes the following during online operation: Within a preset statistical time window, for each circuit breaker in the substation group, a standardized behavior vector is selected corresponding to multiple opening and closing operation events that occur under preset normal operating conditions and are not recorded as associated with fault events within the preset observation time window. Based on the selected standardized behavior vector, multidimensional statistical features are calculated to construct an individual baseline model for characterizing the distribution of the circuit breaker's own operating behavior. Step S5 further includes: based on the individual baseline model, calculating the individual deviation of the standardized behavior vector corresponding to the target opening and closing operation event relative to the individual baseline model, and using the individual deviation and the group deviation of the target opening and closing operation event relative to the group baseline model as the diagnostic basis for determining the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed.

5. The fault diagnosis method for heterogeneous circuit breakers in a station group according to claim 4, characterized in that, In step S5, based on the comparison results of the individual deviation and the group deviation with the preset individual threshold and the preset group threshold, respectively, the target opening and closing operation event is divided into at least one of the following modes: When the individual deviation is less than the preset individual threshold and the group deviation is less than the preset group threshold, the target opening and closing operation event is determined to be in normal operation mode. When the individual deviation is greater than or equal to the preset individual threshold and the group deviation is less than the preset group threshold, the target opening and closing operation event is determined to be in the operating condition disturbance mode. When the individual deviation is greater than or equal to the preset individual threshold and the group deviation is greater than or equal to the preset group threshold, the target opening and closing operation event is determined to be in the group out-of-group mode, and the circuit breaker corresponding to the target opening and closing operation event in the group out-of-group mode is marked as an object with potential fault.

6. The fault diagnosis method for heterogeneous circuit breakers in a station group according to claim 1, characterized in that, The device parameter set includes rated voltage, operating mechanism type and sampling frequency. In step S1, circuit breakers with rated voltage of the same voltage level, the same operating mechanism type and the same sampling frequency are classified into the same device family.

7. The fault diagnosis method for heterogeneous circuit breakers in a station group according to claim 1, characterized in that, The operating data in step S2 includes coil current waveform data and operation time data. The behavioral feature vector extracted in step S2 includes: the coil current peak value and the duration of the current rise phase extracted from the coil current waveform data, and the total opening time and total closing time extracted from the operation time data.

8. The fault diagnosis method for heterogeneous circuit breakers in a station group according to claim 1, characterized in that, When the parameter set of the access devices in the station group does not meet the parameter configuration of any existing device family for a new circuit breaker, the method further includes: Configure the new circuit breaker as a circuit breaker in a new device family, and use the device parameter set of the new circuit breaker as the device parameter representative of the new device family; Based on the parameter differences between the device parameter representatives of the new device family and the device parameter representatives of each existing device family, the similarity between the new device family and each existing device family is calculated, and the existing device family with the highest similarity is selected as the target device family. Based on the differences in device parameters between the new device family and the target device family, the model parameters of the population baseline model of the target device family are corrected according to a preset parameter correction rule to obtain the initial model parameters of the population baseline model of the new device family.

9. A fault diagnosis system for heterogeneous circuit breakers in a substation group, applied to a substation group consisting of multiple circuit breakers from multiple substations, characterized in that, include: An operation status acquisition unit is installed on each circuit breaker in the substation group. It is used to acquire operation data that characterizes the electrical and mechanical behavior of the opening and closing process during the opening and closing operation events of each circuit breaker. The operation data includes coil current waveform data and operation time data. The device parameter management and device family division module is used to obtain the device parameter set of each circuit breaker in the station group, and group the circuit breakers into multiple device families based on the device parameter set; The feature extraction and standardization module is used to extract behavioral feature vectors representing the electrical and mechanical behavior of the opening and closing process based on the operating data collected by the operating status acquisition unit, and to pre-set a mapping relationship for standardization processing for each device family based on the device parameter set of each circuit breaker. The mapping relationship is used to transform the behavioral feature vectors of the circuit breakers in the device family to a comparable unified dimension space, and to standardize the behavioral feature vectors according to the corresponding mapping relationship to obtain the standardized behavioral vectors corresponding to each opening operation event and closing operation event. The group baseline modeling module is used to construct a group baseline model representing the distribution of normal operation behavior for each device family based on the standardized behavior vector set obtained by the device family under preset normal operation conditions. The diagnosis and evaluation module is used to obtain the corresponding standardized behavior vector for the target opening and closing operation event of the circuit breaker to be diagnosed, input the standardized behavior vector into the group baseline model of the device family to which the circuit breaker to be diagnosed belongs, calculate the group deviation of the standardized behavior vector relative to the group baseline model, and output the operating status and / or fault diagnosis result of the circuit breaker to be diagnosed based on the comparison result of the group deviation with the preset group threshold.

10. The fault diagnosis system for heterogeneous circuit breakers in a station cluster according to claim 9, characterized in that, The system also includes at least one of the following modules: The group baseline update module is used to add the standardized behavior vectors of multiple tripping and closing operation events that occur under preset normal operating conditions and are not recorded as associated with fault events within a preset observation time window during the online operation of the system to the standardized behavior vector set used to construct each device family. In each preset update cycle, the module updates the model parameters of the group baseline model of the corresponding device family based on the standardized behavior vectors added to the standardized behavior vector set after the end of the previous update cycle. The new device family migration module is used to configure a new circuit breaker as a circuit breaker in a new device family when the device parameter set of the new circuit breaker does not meet the device parameter configuration of any existing device family in the system. The module uses the device parameter set of the new circuit breaker as the device parameter representative of the new device family. Based on the parameter differences between the device parameter representative of the new device family and the device parameter representatives of each existing device family, the module calculates the similarity between the new device family and each existing device family. The module selects the existing device family with the highest similarity as the target device family. Based on the device parameter differences between the new device family and the target device family, the module corrects the model parameters of the group baseline model of the target device family according to a preset parameter correction rule to obtain the initial model parameters of the group baseline model of the new device family.

Images