Fault detection method, device and equipment based on secondary circuit and storage medium

Abstract

The embodiment of the invention provides a fault detection method and device based on a secondary circuit, equipment and a storage medium. The method comprises the following steps: acquiring voltage information and current information of a secondary circuit at a preset moment; the voltage information represents a voltage collected by a voltage sensor in a preset voltage measuring range under a preset sampling frequency; the current information represents a current collected by a current sensor in a preset current range at a preset sampling frequency; the number of the preset voltage measuring ranges is at least two; the number of the preset current range ranges is at least two; the preset voltage range ranges are in one-to-one correspondence with the preset current range ranges; the preset moments are in one-to-one correspondence with the preset sampling frequencies; and determining fault information of the secondary circuit according to the voltage information and the current information at each preset moment. The method is used for improving the fault detection accuracy of the secondary circuit.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a fault detection method, apparatus, device and storage medium based on a secondary circuit. Background Technology

[0002] Secondary circuit faults in power distribution network switchgear are a significant factor affecting power supply reliability.

[0003] Traditional fault diagnosis relies on on-site inspections and experience-based judgment by maintenance personnel. When transient faults (such as switch malfunctions or unauthorized tripping) occur, the transient characteristics of the fault signal are difficult to record completely, making it difficult to trace the cause of the fault. Therefore, improving the detection accuracy of transient faults in secondary circuits has become an urgent technical problem to be solved. Summary of the Invention

[0004] This application provides a fault detection method, apparatus, device, and storage medium based on secondary circuits, which aims to improve the accuracy of fault detection in secondary circuits.

[0005] In a first aspect, embodiments of this application provide a fault detection method based on a secondary circuit, comprising:

[0006] The system acquires voltage and current information of the secondary circuit at a preset time. The voltage information represents the voltage collected by a voltage sensor with a preset voltage range at a preset sampling frequency. The current information represents the current collected by a current sensor with a preset current range at a preset sampling frequency. There are at least two preset voltage ranges and at least two preset current ranges. The preset voltage range and the preset current range correspond one-to-one. The preset time corresponds one-to-one with the preset sampling frequency.

[0007] Based on the voltage and current information at each preset time, the fault information of the secondary circuit is determined.

[0008] In one possible implementation, the fault information of the secondary circuit is determined based on the voltage and current information at each preset time, including:

[0009] Based on a preset time length, all preset moments are grouped to obtain multiple moment groups; each moment group contains multiple preset moments within a preset time length; there are multiple preset time lengths.

[0010] For each time group corresponding to the preset time length, the characteristic information of the time group is determined based on the voltage and current information at each preset time within the time group; the characteristic information of the time group characterizes the voltage and current changes of the secondary circuit within the preset time length.

[0011] The fault information of the secondary circuit is determined based on the characteristic information of each time group corresponding to each preset time length.

[0012] In one possible implementation, the characteristic information of the time group is determined based on the voltage and current information at each preset time within the time group, including:

[0013] For each preset time within the time group, the first information at the preset time is determined based on the voltage information collected by the voltage sensor within each preset voltage range at the preset time; the first information at the preset time characterizes the voltage condition of the secondary circuit at the preset time.

[0014] Based on the current information collected by the current sensors within each preset current range at the preset time, the second information at the preset time is determined; the second information at the preset time characterizes the current situation of the secondary circuit at the preset time.

[0015] Based on the first information and the second information at the preset time, sub-information at the preset time is determined; the sub-information at the preset time represents the voltage and current conditions of the secondary circuit at the preset time.

[0016] The feature information of the time group is determined based on the sub-information at each preset time within the time group.

[0017] In one possible implementation, the first information at the preset time is determined based on the voltage information collected by voltage sensors within each preset voltage range at the preset time, including:

[0018] Based on the voltage information collected by voltage sensors within each preset voltage range at the preset time and the first preset weight of each preset voltage range, the first information at the preset time is determined; the first preset weight characterizes the degree of influence of the voltage information collected by voltage sensors within the preset voltage range on the first information.

[0019] Based on the current information collected by current sensors within each preset current range at the preset time, the second information at the preset time is determined, including:

[0020] The second information at the preset time is determined based on the current information collected by the current sensor in each preset current range at the preset time and the second preset weight of each preset current range; the second preset weight represents the degree of influence of the current information collected by the current sensor in the preset current range on the second information.

[0021] In one possible implementation, the characteristic information of the time group is determined based on the sub-information at each preset time within the time group, including:

[0022] Based on the sub-information at each preset time within the time group, the time-domain information of the time group is determined; the time-domain information represents the mean, standard deviation, and peak value of the sub-information at all preset times within the time group.

[0023] For each preset time within the time group, the sub-information at the preset time is subjected to frequency domain transformation to obtain transformed information; the transformed information represents the information of the sub-information at the preset time in the frequency domain.

[0024] Based on the conversion information at each preset time within the time group, the frequency domain information of the time group is determined; the frequency domain information represents the main frequency component and spectral energy of the conversion information at all preset times within the time group.

[0025] The characteristic information of the time group is determined based on the time domain information and the frequency domain information of the time group.

[0026] In one possible implementation, the preset time length includes at least one of the following: a first preset time length, a second preset time length, and a third preset time length; the first preset time length is less than the second preset time length, and the second preset time length is less than the third preset time length;

[0027] Based on the characteristic information of each time group corresponding to each preset time length, the fault information of the secondary circuit is determined, including:

[0028] For each time group corresponding to the first preset time length, determine the median of the first information at all preset times within the time group;

[0029] Based on the median corresponding to the time group, a first group is selected from all time groups corresponding to the second preset time length; and a second group is selected from all time groups corresponding to the third preset time length.

[0030] The scale information of the time group is determined based on the feature information of the time group, the feature information of the first group, and the feature information of the second group; the scale information of the time group represents information that integrates the feature information of all time groups with preset time lengths.

[0031] Based on the scale information corresponding to each time group corresponding to the first preset time length, the fusion feature information is determined; the fusion feature information represents information that fuses the scale information of all time groups.

[0032] Based on the fused feature information, the fault information of the secondary circuit is determined.

[0033] In one possible implementation, determining the fault information of the secondary circuit based on the fused feature information includes:

[0034] The fused feature information is input into the fault prediction model to obtain the fault information of the secondary loop. The fault prediction model is a trained neural network model, which is used to perform data reasoning on the input fused feature information to obtain the fault information of the secondary loop.

[0035] In one possible implementation, it also includes:

[0036] For each preset time, determine the neighboring times of the preset time; wherein, the neighboring times represent the two preset times that are preceding the preset time;

[0037] Based on the voltage and current information of each neighboring time corresponding to the preset time, the preset sampling frequency corresponding to the preset time is determined.

[0038] In one possible implementation, determining the preset sampling frequency corresponding to the preset time based on the voltage and current information at each neighboring time corresponding to the preset time includes:

[0039] Based on the voltage information at each neighboring time point, a first rate of change is determined; the first rate of change characterizes the rate of change of voltage information between two neighboring time points.

[0040] The second rate of change is determined based on the current information at each neighboring time point; the second rate of change characterizes the rate of change of the current information at two neighboring time points.

[0041] Based on the first rate of change and the second rate of change, the preset sampling frequency corresponding to the preset time is determined.

[0042] Secondly, embodiments of this application provide a fault detection device based on a secondary circuit, comprising:

[0043] The acquisition module is used to acquire voltage and current information of the secondary circuit at a preset time. The voltage information represents the voltage collected by a voltage sensor with a preset voltage range at a preset sampling frequency. The current information represents the current collected by a current sensor with a preset current range at a preset sampling frequency. There are at least two preset voltage ranges and at least two preset current ranges. The preset voltage ranges and preset current ranges correspond one-to-one. The preset time corresponds one-to-one with the preset sampling frequency.

[0044] The determination module is used to determine the fault information of the secondary circuit based on the voltage and current information at each preset time.

[0045] Thirdly, embodiments of this application provide a fault detection device based on a secondary loop, comprising: a memory and a processor;

[0046] The memory stores computer-executed instructions;

[0047] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0048] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0049] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0050] The fault detection method, apparatus, device, and storage medium based on secondary circuits provided in this application embodiment provide voltage and current information of the secondary circuit at preset times. The voltage information represents the voltage collected by a voltage sensor with a preset voltage range at a preset sampling frequency, and the current information represents the current collected by a current sensor with a preset current range at a preset sampling frequency. There are at least two preset voltage ranges and at least two preset current ranges, with a one-to-one correspondence between the preset voltage and current ranges. By simultaneously measuring the voltage of the secondary circuit using multiple voltage sensors with preset voltage ranges and simultaneously measuring the current of the secondary circuit using multiple current sensors with preset current ranges, the requirements for wide range coverage and high-precision measurement of the secondary circuit current and voltage are balanced, allowing for better adaptation to complex operating conditions. The one-to-one correspondence between preset times and preset sampling frequencies enables dynamic switching of the sampling frequency based on the signal change rate, ensuring that high-frequency signals are completely captured within a short period after a fault occurs. Based on the above technical principles, the fault information of the secondary circuit can be determined according to the voltage and current information at each preset time. This can significantly improve the integrity and accuracy of fault transient signal acquisition, thereby improving the fault detection accuracy of the secondary circuit. Attached Figure Description

[0051] 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.

[0052] Figure 1 A flowchart illustrating the fault detection method based on a secondary loop provided in this application embodiment. Figure 1 ;

[0053] Figure 2 A flowchart illustrating the fault detection method based on a secondary loop provided in this application embodiment. Figure 2 ;

[0054] Figure 3 A flowchart illustrating the fault detection method based on a secondary loop provided in this application embodiment. Figure 3 ;

[0055] Figure 4 A schematic diagram of the structure of a fault detection device based on a secondary loop provided in an embodiment of this application;

[0056] Figure 5 This is a schematic diagram of the structure of a fault detection device based on a secondary loop provided in an embodiment of this application.

[0057] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0058] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0059] Secondary circuit faults in power distribution network switchgear are a significant factor affecting power supply reliability. These faults encompass abnormal control power supply, relay failure, circuit disconnection, and poor contact. In power systems, secondary circuits are responsible for controlling the opening and closing operations, status monitoring, and protection functions of main equipment; their reliability directly impacts the safe operation of the power grid. However, traditional fault diagnosis relies on on-site inspection and experience-based judgment by maintenance personnel, which has significant drawbacks: First, when transient faults (such as switch failure to operate or unauthorized tripping) occur, the transient characteristics of the fault signal are difficult to record completely by traditional equipment, making it difficult to trace the cause of the fault. Second, manual inspection relies on a thorough understanding of equipment principles and fault modes, requiring high skill levels from maintenance personnel and resulting in low efficiency, easily leading to delays in recovery time due to misjudgment or omission. Furthermore, existing monitoring technologies mostly focus on steady-state parameters (such as average values ​​of voltage and current), lacking the ability to dynamically capture transient fault processes, making it difficult to identify hidden faults (such as sudden changes in contact resistance or relay contact jitter).

[0060] The application scenarios of this application include: troubleshooting secondary circuit faults in distribution network switchgear, specifically applicable to distribution automation terminals, smart substations, and operation and maintenance management systems. During power system operation, the secondary circuit is responsible for controlling the opening and closing operations, status monitoring, and protection functions of the main equipment. When faults such as abnormal control power supply, relay failure, or circuit disconnection occur, traditional methods rely on manual experience for judgment, making it difficult to locate hidden faults (such as poor contact or relay coil abnormalities) in a timely manner. This application utilizes automated devices deployed on terminal equipment to collect voltage, current, and status signals in real time, combines this with an artificial intelligence model for intelligent diagnosis, generates dynamic handling suggestions through a human-machine interface, and ultimately optimizes the model based on feedback data from maintenance personnel, forming a closed-loop management system that significantly improves the efficiency and reliability of fault troubleshooting.

[0061] The technical solution of this application and how it solves the above-mentioned technical problems will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0062] Figure 1 A flowchart illustrating the fault detection method based on a secondary loop provided in this application embodiment. Figure 1 ,like Figure 1 As shown, the method includes:

[0063] S101. Obtain voltage and current information of the secondary circuit at a preset time; the voltage information represents the voltage collected by a voltage sensor with a preset voltage range at a preset sampling frequency; the current information represents the current collected by a current sensor with a preset current range at a preset sampling frequency; there are at least two preset voltage ranges; there are at least two preset current ranges; the preset voltage ranges and preset current ranges correspond one-to-one; the preset time corresponds one-to-one with the preset sampling frequency.

[0064] For example, wide-range voltage sensors and narrow-range voltage sensors can be deployed simultaneously in the secondary circuit. For instance, a voltage sensor with a voltage range of 0-300V and an accuracy of 1% can be deployed; simultaneously, a voltage sensor with a preset voltage range of 0-150V and an accuracy of 0.1% can be deployed. By deploying multiple voltage sensors with different voltage ranges, complex operating conditions at different voltage levels can be accommodated. That is, for voltage sensors deployed simultaneously in the secondary circuit, there are at least two preset voltage ranges.

[0065] By deploying multiple voltage sensors synchronously, voltage information of the secondary circuit at preset times can be acquired. This voltage information represents the voltage collected by voltage sensors within a preset voltage range at a preset sampling frequency. In other words, voltage sensors within the preset voltage range sample at the preset sampling frequency to obtain the voltage information at the preset time. There are at least two preset voltage ranges; that is, each preset time includes at least two voltage information sets, each corresponding to a preset voltage range, and the corresponding voltage information is acquired by voltage sensors within that preset voltage range. The acquired voltage information can be used to analyze voltage changes in the secondary circuit, and based on these voltage changes, fault conditions in the secondary circuit can be analyzed.

[0066] The preset voltage range can be determined through the configuration information of the switching equipment. Specifically, this can be achieved by obtaining the switching equipment configuration information stored in the distribution automation terminal. For example, a connection can be established with the distribution automation terminal via the Modbus communication protocol to read the switching equipment configuration information stored in the distribution automation terminal, including key parameters such as the rated voltage of the control power supply, relay model, and circuit topology.

[0067] After obtaining the configuration parameter data of the switchgear, range mapping can be performed based on the rated voltage value of the control power supply. For example, a rated voltage of 48V corresponds to a voltage range of 0-100V; a rated voltage of 110V corresponds to a voltage range of 0-150V; and a rated voltage of 220V corresponds to a voltage range of 0-300V. By controlling the switching resistors within the sensor through relays in the switchgear, the voltage sensor range can be configured, thus obtaining the voltage sensor range setting parameters that match the rated voltage of the switchgear.

[0068] Based on all obtained voltage ranges, at least two preset voltage ranges can be determined. The largest preset voltage range encompasses all voltage ranges (i.e., the largest preset voltage range is greater than or equal to all voltage ranges). The smallest preset voltage range covers most voltage ranges (i.e., the smallest preset voltage range can cover most voltage ranges). This achieves the ability to measure a wide voltage range while obtaining high-precision voltage measurements within commonly used voltage ranges.

[0069] Based on the analysis of voltage changes in the secondary circuit, the analysis of current changes in the secondary circuit can be combined to obtain a more accurate fault diagnosis. Specifically, when deploying current sensors, the range configuration of the current sensors can be completed using the relay model information and preset voltage range of the switching equipment. For example, obtain the relay model information, calculate the required current range using the relay model information, estimate the peak inrush current range based on the rated current of the relay coil (typically 5-10 times the rated current), select the current sensor range covering this peak range, and obtain the maximum preset current range. The preset voltage range and preset current range correspond one-to-one, and the maximum preset current range corresponds to the maximum preset voltage range. Other preset current ranges are configured according to the number of preset voltage ranges.

[0070] By deploying multiple current sensors, current information of the secondary circuit at preset times can be acquired. This current information represents the current collected by current sensors within a preset current range at a preset sampling frequency. Specifically, current sensors within the preset current range sample at the preset sampling frequency to obtain the current information at the preset time. There are at least two preset current ranges; that is, each preset time includes at least two current information sets, each corresponding to a preset current range, and the corresponding current information is collected by current sensors within that range. The acquired current information can be used to analyze current changes in the secondary circuit, and based on these changes, fault conditions in the secondary circuit can be analyzed.

[0071] Each preset time point corresponds to a preset sampling frequency. Different preset time points may use the same or different preset sampling frequencies. For example, the rate of change of current and voltage information in the secondary circuit can be analyzed in real time. For preset time points where the rates of change of current and voltage information are both small, a lower sampling frequency is configured; for preset time points where the rates of change of current or voltage information are large, a higher sampling frequency is configured. This allows for high-frequency sampling to obtain more detailed information during the short period when a fault occurs in the secondary circuit, due to sudden changes in current or voltage causing changes in the opening and closing states of switching equipment, providing strong data support for subsequent fault analysis. Conversely, during stable operation of the secondary circuit, when current and voltage are relatively stable, low-frequency sampling can reduce the amount of data processed.

[0072] S102. Determine the fault information of the secondary circuit based on the voltage and current information at each preset time.

[0073] For example, for all preset time periods, the time period is divided into multiple sub-time periods. For each sub-time period, feature extraction is performed on the voltage information at all preset times within that sub-time period to obtain the feature value of the voltage corresponding to that sub-time period. For example, the mean and standard deviation of the voltage information at all preset times within that sub-time period can be extracted, and the extracted mean and standard deviation of the voltage information can be used as the feature value of the voltage corresponding to that sub-time period.

[0074] Simultaneously, feature extraction is performed on the current information at all preset times within the sub-time period to obtain the feature values ​​of the current corresponding to that sub-time period. For example, the mean and standard deviation of the current information at all preset times within the sub-time period can be extracted, and these extracted mean and standard deviation can be used as the feature values ​​of the current corresponding to that sub-time period. The feature values ​​of the voltage and current corresponding to that sub-time period are combined as the feature values ​​of that sub-time period. This process yields the feature values ​​corresponding to all sub-time periods.

[0075] Feature analysis can be performed on the feature values ​​of all sub-time periods. For example, an artificial intelligence model can be used to perform feature analysis on the feature values ​​of all sub-time periods to obtain fault information of the secondary circuit.

[0076] The fault detection method based on a secondary circuit provided in this application acquires voltage and current information of the secondary circuit at preset times. The voltage information represents the voltage collected by a voltage sensor with a preset voltage range at a preset sampling frequency, and the current information represents the current collected by a current sensor with a preset current range at a preset sampling frequency. There are at least two preset voltage ranges and at least two preset current ranges, with a one-to-one correspondence between the preset voltage and current ranges. By simultaneously measuring the voltage of the secondary circuit using multiple voltage sensors with preset voltage ranges and simultaneously measuring the current of the secondary circuit using multiple current sensors with preset current ranges, the requirements for wide range coverage and high-precision measurement of the secondary circuit's current and voltage are balanced, allowing for better adaptation to complex operating conditions. The one-to-one correspondence between preset times and preset sampling frequencies enables dynamic switching of the sampling frequency based on the signal change rate, ensuring that high-frequency signals are completely captured within a short period after a fault occurs. Combining the above technical principles, the fault information of the secondary circuit is determined based on the voltage and current information at each preset time, which can significantly improve the completeness and accuracy of fault transient signal acquisition, thereby improving the fault detection accuracy of the secondary circuit.

[0077] Figure 2 A flowchart illustrating the fault detection method based on a secondary loop provided in this application embodiment. Figure 2 ,like Figure 2 As shown, the method for determining secondary circuit fault information based on voltage and current information at each preset time point includes: grouping all preset time points into multiple time groups based on a preset time length; each time group contains multiple preset time points within a preset time length; there are multiple preset time lengths; for each time group corresponding to a preset time length, determining the characteristic information of the time group based on the voltage and current information at each preset time point within the time group; the characteristic information of the time group characterizes the voltage and current changes in the secondary circuit within the preset time length; and determining the secondary circuit fault information based on the characteristic information of each time group corresponding to each preset time length. This method includes:

[0078] S201. Obtain the voltage and current information of the secondary circuit at a preset time.

[0079] For example, digital filtering methods can be used to denoise the voltage and current information of the acquired secondary circuit at a preset time. For instance, a fourth-order Butterworth low-pass filter with a cutoff frequency of 2kHz can be used to perform forward and backward filtering on the acquired information to eliminate phase distortion and obtain denoised current and voltage information. The Butterworth filter is a filter with the flattest frequency response within its passband, characterized by the smoothest frequency response curve in the passband, free of ripple. The filter's transfer function is converted from the analog domain to the digital domain using a bilinear transform method.

[0080] After denoising the current and voltage information, a robust normalization method is used for amplitude standardization. For voltage information, voltage sensors with different preset voltage ranges collect voltage information from different channels; similarly, voltage sensors with different preset current ranges collect current information from different channels. For each channel's current or voltage information, the median and interquartile range can be calculated from the channel's historical operating data (during non-fault periods). The median is the value in the middle after arranging the data in ascending order, and the interquartile range is the difference between the third quartile and the first quartile. Normalization of the denoised current or voltage information is achieved by subtracting the median of the corresponding historical data from the denoised current or voltage information and then dividing by the interquartile range. This approach provides stronger robustness to outliers and fault signals.

[0081] In subsequent data processing, normalized current and voltage information were used for data processing.

[0082] It should be noted that, in order to record the current and voltage information for a short period of time when the switching state of the device changes, a circular buffer mechanism is employed to cache the acquired current and voltage information. For example, a circular buffer is pre-allocated in memory, with a size set to store sampling data for 10 seconds before and after the moment the switching state changes. The size of the circular buffer can be calculated based on the number of samples, the number of data bytes, the sampling frequency, and the storage time. Sampling data is written to the circular buffer in chronological order. When the circular buffer is full, the overwriting process begins again from the beginning. When a change in the switching state of the device is detected, the overwriting of the circular buffer stops after 10 seconds. Afterward, all data in the circular buffer is used as processing data; that is, all data in the circular buffer includes voltage and current information at all preset times.

[0083] The open / closed state of a switchgear can be determined using its open / closed position status signals and remote control command signals from the distribution automation terminal. For example, based on the open / closed position status signals, a preset state change detection method can be used to perform edge detection on the open and closed signals. When a signal changes from 0 (closed) to 1 (open) or from 1 to 0, it is determined that the open / closed state of the switchgear has changed. Simultaneously, the remote control command signals issued by the distribution automation terminal are detected. When a tripping or closing command is received, it is determined that the change in the open / closed state of the switchgear is caused by the issuance of an operation command.

[0084] S202. Based on the preset time length, all preset times are grouped to obtain multiple time groups; each time group contains multiple preset times within the preset time length; there are multiple preset time lengths.

[0085] For example, for all preset time points, a multi-scale sliding time window can be used for signal segmentation. Each scale corresponds to a preset time length, and multiple scales correspond to multiple preset time lengths. For instance, three time window scales can be set: a short window of 50ms for capturing fast transients, a medium window of 200ms for capturing regular dynamics, and a long window of 500ms for capturing slow changes. The sliding step is uniformly set to 25ms, and the data for 20 seconds before and after the fault event is segmented using a sliding window. This results in a sequence of approximately 800 signal segments within three time windows, containing signal segmentation data at the three time scales. That is, for all preset time lengths, approximately 800 time groups are obtained, and each time group contains multiple preset times within a preset time length, such as multiple preset times within a corresponding time window.

[0086] S203. For each time group corresponding to the preset time length, determine the characteristic information of the time group based on the voltage and current information at each preset time within the time group; the characteristic information of the time group represents the voltage and current changes of the secondary circuit within the preset time length.

[0087] For example, for each time group corresponding to a preset time length, feature extraction can be performed on the voltage information at all preset times within that time group to obtain the feature value of the voltage corresponding to that time group. For instance, the mean and standard deviation of the voltage information at all preset times within that time group can be extracted, and the extracted mean and standard deviation of the voltage information can be used as the feature value of the voltage corresponding to that time group.

[0088] Simultaneously, feature extraction is performed on the current information at all preset times within the time group to obtain the feature values ​​of the current corresponding to that time group. For example, the mean and standard deviation of the current information at all preset times within the time group can be extracted, and the extracted mean and standard deviation of the current information can be used as the feature values ​​of the current corresponding to that time group.

[0089] The characteristic values ​​of the voltage and current corresponding to this time group are determined as the characteristic information of that time group. The characteristic information of the time group represents the voltage and current changes in the secondary circuit within a preset time length. For example, the characteristic value of the voltage in the characteristic information of the time group represents the voltage change in the secondary circuit within a preset time length, and the characteristic value of the current in the characteristic information of the time group represents the current change in the secondary circuit within a preset time length.

[0090] In this embodiment, the determination of the characteristic information of the time group based on the voltage and current information at each preset time within the time group includes:

[0091] For each preset time within a time group, the first information for that preset time is determined based on the voltage information collected by voltage sensors within each preset voltage range at that preset time. The first information for that preset time represents the voltage condition of the secondary circuit at that preset time. The second information for that preset time is determined based on the current information collected by current sensors within each preset current range at that preset time. The second information for that preset time represents the current condition of the secondary circuit at that preset time. Sub-information for that preset time is determined based on the first information and the second information for that preset time. The sub-information for that preset time represents the voltage and current conditions of the secondary circuit at that preset time. The characteristic information of the time group is determined based on the sub-information for each preset time within the time group.

[0092] For example, for each preset time within a time group, voltage information collected by voltage sensors across all preset voltage ranges at that preset time can be fused to obtain first information at that preset time. This first information represents the voltage condition of the secondary circuit at that preset time; that is, it is information that fuses voltage values ​​collected by voltage sensors across all preset voltage ranges. Through this fusion process, wide-range, low-precision voltage information and narrow-range, high-precision voltage information can be combined to obtain wide-range, high-precision voltage information.

[0093] Simultaneously, the current information collected by current sensors within all preset current ranges at a preset time is fused to obtain the second information at the preset time. This second information represents the current situation in the secondary circuit at the preset time; that is, it is information that integrates the current values ​​collected by current sensors within all preset current ranges. Through this fusion process, wide-range, low-precision current information and narrow-range, high-precision current information can be combined to obtain wide-range, high-precision current information.

[0094] Sub-information at a preset time can be obtained by combining the first information and the second information at a preset time. The sub-information at the preset time represents the voltage status and changes in the secondary circuit at that preset time. For example, the first information in the sub-information at the preset time represents the voltage value of the secondary circuit at that preset time, and the second information in the sub-information at the preset time represents the current value of the secondary circuit at that preset time.

[0095] After obtaining the sub-information at each preset time within a time group, the feature information of that time group can be obtained by combining the sub-information at all preset times within that time group.

[0096] In this embodiment, determining the first information at the preset time based on the voltage information collected by voltage sensors within each preset voltage range at the preset time includes:

[0097] Based on the voltage information collected by voltage sensors within each preset voltage range at a preset time, and the first preset weight of each preset voltage range, a first piece of information at a preset time is determined; the first preset weight characterizes the degree of influence of the voltage information collected by voltage sensors within each preset voltage range on the first information; based on the current information collected by current sensors within each preset current range at a preset time, a second piece of information at a preset time is determined, including: based on the current information collected by current sensors within each preset current range at a preset time, and the second preset weight of each preset current range, the second preset weight characterizes the degree of influence of the current information collected by current sensors within each preset current range on the second information.

[0098] For example, for all voltage sensors within a preset voltage range at the preset time, zero-point correction and gain correction can be performed on each voltage sensor within its preset voltage range, so that the corrected data have the same dimensions and accuracy standards. The voltage information collected by all voltage sensors within the preset voltage range at the preset time is then weighted and fused to obtain the first information at that preset time.

[0099] For example, when there are two preset voltage ranges: a first range with wide voltage range and low precision, and a second range with narrow voltage range and high precision, if the voltage value of the secondary circuit is within the second range, the weight of the first range can be set to 0.2 and the weight of the second range to 0.8. That is, the first preset weight of the first range is 0.2 and the first preset weight of the second range is 0.8, to obtain a high-precision voltage value for the secondary circuit. If the voltage value of the secondary circuit exceeds the second range but is within the first range, the weight of the first range can be set to 1.0 and the weight of the second range is 0. That is, the first preset weight of the first range is 1.0 and the first preset weight of the second range is 0, to ensure that the voltage value of the secondary circuit can be obtained.

[0100] The voltage information collected by voltage sensors within each preset voltage range is weighted and summed using a first preset weight for each preset voltage range to obtain the first information at a preset time. The first preset weight characterizes the influence of the voltage information collected by the voltage sensors within each preset voltage range on the first information. Specifically, when the voltage value of the secondary circuit is within a narrow preset voltage range, the first preset weight for the narrow preset voltage range is greater than the first preset weight for the wide preset voltage range to ensure high-accuracy acquisition of the secondary circuit voltage value; when the voltage value of the secondary circuit exceeds the narrow preset voltage range but is within the wide preset voltage range, the first preset weight for the wide preset voltage range is greater than the first preset weight for the narrow preset voltage range to ensure that the voltage value of the secondary circuit can be acquired.

[0101] Similarly, the current information collected by current sensors within all preset current ranges at the preset time is weighted and fused to obtain the second information at the preset time.

[0102] For example, if there are two preset current ranges: a third range with wide current range and low precision, and a fourth range with narrow current range and high precision, then when the secondary circuit current value is within the fourth range, the weight of the third range can be set to 0.2 and the weight of the fourth range to 0.8. That is, the second preset weight of the third range is 0.2 and the second preset weight of the fourth range is 0.8, to obtain a high-precision secondary circuit current value. When the secondary circuit current value exceeds the fourth range but is within the third range, the weight of the third range can be set to 1.0 and the weight of the fourth range is 0. That is, the second preset weight of the third range is 1.0 and the second preset weight of the fourth range is 0, to ensure that the secondary circuit current value can be obtained.

[0103] The voltage information collected by voltage sensors within each preset voltage range is weighted and summed using a first preset weight for each preset voltage range to obtain the first information at a preset time. The first preset weight characterizes the influence of the voltage information collected by the voltage sensors within each preset voltage range on the first information. Specifically, when the voltage value of the secondary circuit is within a narrow preset voltage range, the first preset weight for the narrow preset voltage range is greater than the first preset weight for the wide preset voltage range to ensure high-accuracy acquisition of the secondary circuit voltage value; when the voltage value of the secondary circuit exceeds the narrow preset voltage range but is within the wide preset voltage range, the first preset weight for the wide preset voltage range is greater than the first preset weight for the narrow preset voltage range to ensure that the voltage value of the secondary circuit can be acquired.

[0104] In this embodiment, determining the feature information of the time group based on the sub-information of each preset time within the time group includes:

[0105] Based on the sub-information at each preset time within the time group, the time-domain information of the time group is determined; the time-domain information represents the mean, standard deviation, and peak value of the sub-information at all preset times within the time group; for each preset time within the time group, the sub-information at the preset time is subjected to frequency domain transformation processing to obtain transformed information; the transformed information represents the information of the sub-information at the preset time in the frequency domain; based on the transformed information at each preset time within the time group, the frequency-domain information of the time group is determined; the frequency-domain information represents the main frequency components and spectral energy of the transformed information at all preset times within the time group; based on the time-domain information and the frequency-domain information of the time group, the characteristic information of the time group is determined.

[0106] For example, the mean, standard deviation, peak value, etc. of all sub-information at preset times within the time group can be calculated, and the calculated mean, standard deviation, peak value, etc. can be combined to obtain the time domain information of the time group.

[0107] Simultaneously, the sub-information at each preset time within the time group undergoes frequency domain transformation processing. For example, a Fast Fourier Transform is used to transform the sub-information at each preset time within the time group to obtain the frequency domain representation of the sub-information at each preset time. This frequency domain representation of the sub-information at each preset time is determined as the transformed information at that preset time. The frequency components of the transformed information at all preset times within the time group are statistically analyzed, and the dominant frequency component and spectral energy distribution are extracted. The dominant frequency distribution and spectral energy distribution are used as the frequency domain information of the time group.

[0108] Then, the time-domain and frequency-domain information of the time group are combined, and the resulting multi-dimensional data is used as the feature information of the time group.

[0109] By extracting time-domain and frequency-domain information from time groups, deeper information about current and voltage can be extracted in both domains. The feature information obtained by combining time-domain and frequency-domain information, compared to the acquired voltage and current information alone, can provide more in-depth analytical data for subsequent fault analysis, revealing hidden causes of faults and thus improving the accuracy of subsequent fault information.

[0110] S204. Determine the fault information of the secondary circuit based on the characteristic information of each time group corresponding to each preset time length.

[0111] For example, feature information from each time group corresponding to all preset time lengths can be integrated, and an artificial intelligence model can be used to perform feature analysis on the integrated information to obtain fault information of the secondary circuit.

[0112] In this embodiment, the preset time length includes at least one of the following: a first preset time length, a second preset time length, and a third preset time length; the first preset time length is less than the second preset time length, and the second preset time length is less than the third preset time length; the above-mentioned determination of the secondary circuit fault information based on the feature information of each time group corresponding to each preset time length includes:

[0113] For each time group corresponding to the first preset time length, determine the median of the first information at all preset times within the time group; based on the median corresponding to the time group, select the first group from all time groups corresponding to the second preset time length; and select the second group from all time groups corresponding to the third preset time length; determine the scale information of the time group based on the feature information of the time group, the feature information of the first group, and the feature information of the second group; the scale information of the time group represents information that integrates the feature information of time groups of all preset time lengths; determine the fused feature information based on the scale information corresponding to each time group corresponding to the first preset time length; the fused feature information represents information that integrates the scale information of all time groups; and determine the fault information of the secondary circuit based on the fused feature information.

[0114] For example, the preset time length includes a first preset time length corresponding to a segment window length of 50ms, a second preset time length corresponding to a medium window length of 200ms, and a third preset time length corresponding to a long window length of 500ms, for a total of three time length scales.

[0115] For each time group corresponding to the first preset time length, the median of the first information at each preset time within that time group is calculated, which is the first median. Simultaneously, for each time group corresponding to the second preset time length, the median of the first information at each preset time within that time group is calculated, which is the second median. For each time group corresponding to the third preset time length, the median of the first information at each preset time within that time group is calculated, which is the third median.

[0116] For each time group corresponding to a first preset time length, using the first median corresponding to that time group, the second median closest to the first median is matched among all time groups corresponding to all second preset time lengths. The time group corresponding to the second median and the second preset time length is then designated as the first group. Simultaneously, among all time groups corresponding to all third preset time lengths, the third median closest to the first median is matched. The time group corresponding to the third median and the third preset time length is then designated as the second group. This ensures that each time group corresponding to the first preset time length can find a corresponding time group among all time groups corresponding to all second preset time lengths, and also finds a corresponding time group among all time groups corresponding to all third preset time lengths, facilitating better multi-scale data fusion.

[0117] Specifically, for each time group corresponding to the first preset time length, the feature information of that time group, the feature information of the first group corresponding to that time group, and the feature information of the second group corresponding to that time group are fused to obtain the scale information of that time group. The scale information of the time group represents the information that integrates the feature information of all time groups of the preset time length, that is, the multi-scale fused information of that time group.

[0118] It should be noted that the feature information of the aforementioned time groups can be standardized feature information. For example, for the feature information of all time groups corresponding to any preset time length, Z-score standardization can be performed. Z-score standardization is the process of converting data into a standard normal distribution with a mean of 0 and a standard deviation of 1, so that the standardized feature information has the same numerical scale.

[0119] For each time group corresponding to a first preset time length, the scale information of that time group can be obtained by fusing the feature information of that time group, the feature information of the first group corresponding to that time group, and the feature information of the second group corresponding to that time group, according to adaptive weights. Specifically, the rate of change of the first information at each preset time within the time group can be used to determine the speed of signal change. For example, for the first information of all preset times within the time group, the rate of change of the first information of any two adjacent preset times is calculated (the rate of change is obtained by the ratio of the difference to the time length), and the mean of the rate of change of the first information of all two adjacent preset times is calculated. The mean is compared with a preset mean range. If the mean is less than the minimum value of the preset mean range, it indicates that the signal change within the time group is slow; if the mean is within the preset mean range, it indicates that the signal change within the time group is a normal change; if the mean is greater than the maximum value of the preset mean range, it indicates that the signal change within the time group is fast.

[0120] For signals with rapid changes, the weight of the time group corresponding to the first preset length is set to 0.6, the weight of the first group corresponding to the second preset length is set to 0.3, and the weight of the second group corresponding to the third preset length is set to 0.1. For signals with regular changes, the weight of the time group corresponding to the first preset length is set to 0.33, the weight of the first group corresponding to the second preset length is set to 0.33, and the weight of the second group corresponding to the third preset length is set to 0.33. For signals with slow changes, the weight of the time group corresponding to the first preset length is set to 0.1, the weight of the first group corresponding to the second preset length is set to 0.3, and the weight of the second group corresponding to the third preset length is set to 0.6. Then, the feature information of the time group, the feature information of the first group, and the feature information of the second group are weighted and summed based on the set weights to obtain the scale information of the time group.

[0121] The advantage of this setup is that by fusing feature information corresponding to different time lengths and utilizing dynamic weights, it is possible to simultaneously capture dynamic features at different time scales in the fault signal. For example, for short time lengths, the focus is on short-term fault causes, such as electric shock jitter, while for long time lengths, the focus is on fault trends, such as voltage drop, thereby improving the sensitivity and accuracy of fault diagnosis.

[0122] In this embodiment, by grouping all preset time points based on a preset time length, multiple time groups are obtained. This facilitates subsequent individual data processing of the current and voltage information at each preset time point within each time group, yielding the feature information of that time group. Since the extracted feature information includes deeper-level information compared to the current and voltage information, analyzing the fault information of the secondary circuit using the feature information of each time group corresponding to each preset time length can yield more accurate fault information.

[0123] In this embodiment, determining the fault information of the secondary circuit based on the fused feature information includes:

[0124] The fused feature information is input into the fault prediction model to obtain the fault information of the secondary loop. The fault prediction model is a trained neural network model, which is used to perform data reasoning on the input fused feature information to obtain the fault information of the secondary loop.

[0125] For example, principal component analysis (PCA) can be used to process the fused feature information first, resulting in enhanced fused feature information. This enhanced fused feature information contains a compact feature representation of the fused multi-scale time-frequency domain. PCA is a dimensionality reduction technique that projects the original data into a lower-dimensional space through linear transformation. It can retain the main change information in the data, select principal components with a cumulative contribution rate of 95%, and apply a weight amplification factor of 2.0 to key feature dimensions (such as relay current peak value and contact state transition count).

[0126] After dimensionality reduction of the fused feature information, the enhanced fused feature information is input into the fault prediction model, which is a trained neural network model. In the fault prediction model, the enhanced fused feature information is arranged in chronological order to form a two-dimensional tensor of shape T multiplied by D, where T is the time step and D is the feature dimension. Based on the two-dimensional tensor data, batch normalization is used to calculate the mean and variance of the batch data for each feature channel. Batch normalization is a technique that accelerates deep network training by standardizing each mini-batch of data, reducing internal covariate bias. Batch normalization is used to standardize the two-dimensional tensor data, resulting in standardized tensor data.

[0127] For standardized tensor data, depthwise separable convolutional layers are used to extract local temporal features. Depthwise separable convolution is a technique that decomposes standard convolution into two steps: depthwise convolution and pointwise convolution, which reduces the number of parameters and computational cost. The first layer of depthwise separable convolution is performed, which includes two operations: channel-wise convolution and pointwise convolution. Channel-wise convolution performs a one-dimensional convolution on each feature channel independently, with a kernel size of 7; the output is the channel-wise convolution output, which is the feature representation after independent convolution of each feature channel.

[0128] Perform pointwise convolution, use a 1×1 convolution kernel to fuse cross-channel information on the channel-wise convolution output, and output 64 channels; output the first layer of convolutional features, which are 64-channel feature representations after depthwise separable convolution processing.

[0129] Based on the features from the first convolutional layer, activation and pooling methods are employed. The convolutional output is normalized using a batch normalization layer, and non-linearity is introduced through the ReLU activation function. ReLU is a simple and effective activation function that can alleviate the vanishing gradient problem. Downsampling is performed using a max pooling layer with a pooling window size of 2 and a stride of 2. The output is the first layer's features after activation and pooling, representing the features processed by normalization, activation, and pooling.

[0130] Based on the first layer of features after activation pooling, a multi-layer concatenated convolution strategy is employed, sequentially performing a second layer of depthwise separable convolution (kernel size 5, output channels 128) and a third layer of depthwise separable convolution (kernel size 3, output channels 256). Each convolutional layer is followed by batch normalization, ReLU activation, and max pooling to extract higher-level abstract features layer by layer. The output deep convolutional feature representation represents the 256-channel high-level abstract features processed by the multi-layer concatenated convolutions.

[0131] Next, a self-attention mechanism is used for feature enhancement. Self-attention is a technique that captures long-range dependencies within a sequence, assigning attention weights by calculating the correlation between each position and other positions in the sequence. By reshaping the convolutional features into a sequence form, the query matrix Q, key matrix K, and value matrix V are calculated, and the final attention output is obtained through attention weight calculation. The output is a self-attention-enhanced feature representation, resulting in a feature sequence enhanced by the attention mechanism.

[0132] Based on self-attention enhanced feature representation, a gated recurrent unit (GRU) is used for temporal dependency learning. GRU is a simplified recurrent neural network structure that controls the flow of information through gating mechanisms, including an update gate and a reset gate, which respectively control the retention of historical information and the fusion of new information. The outputs are z_t (update gate) and r_t (reset gate), representing the gating signals at each time step.

[0133] Based on the update gate output and reset gate output, the GRU state update method is used to calculate the candidate hidden state and the current hidden state. The update gate controls the fusion ratio of historical and new information. The output is the GRU hidden state at the current time step and the hidden state sequence at each time step.

[0134] Based on the GRU hidden state sequences at each time step, a bidirectional GRU structure is adopted, and forward and backward sequence modeling is performed simultaneously. The hidden states of the forward and backward GRUs are concatenated along the feature dimension. The output is a feature encoding containing bidirectional temporal information, resulting in a complete feature representation that integrates forward and backward temporal information.

[0135] Simultaneously, a multi-head output structure is employed for multi-task fault diagnosis, including three parallel fully connected classifiers: a fault type classification head, a fault location positioning head, and a confidence assessment head. The output shows the initialization state of the multi-head output structure and the network structure parameters of the three parallel classifiers.

[0136] The fault type classification head uses a fully connected layer to map feature encoding to the fault type space. The fully connected layer contains 256 neurons, which are activated by ReLU and then connected to the output layer. The output layer has 7 neurons, corresponding to 7 categories: normal, abnormal control power supply, relay fault, open circuit, short circuit, poor contact, and external interference. The probability distribution for each category is calculated using the Softmax function, which can convert any real-valued vector into a probability distribution. The output shows the fault location results, including the probability distribution of the 6 fault locations.

[0137] The fault location locator uses the same fully connected structure, with 6 neurons in the output layer, corresponding to 6 positions: control power supply, trip relay, closing relay, anti-pumping relay, trip auxiliary contact, and closing auxiliary contact. The probability distribution of each position is calculated using the Softmax function. The output shows the fault location result and the probability distribution of the 6 fault positions.

[0138] The confidence assessment head employs a regression output structure. Feature encoding is mapped to a single numerical value through a fully connected layer, and the output is then mapped to the 0-1 range after passing through a Sigmoid activation function. This numerical value represents the confidence score of the diagnostic result. The output is the confidence assessment result, specifically the confidence score of the diagnostic result.

[0139] Based on the fault type classification results, fault location results, and confidence assessment results, a result integration method is used to combine the three outputs into a complete diagnostic result, including fault type label, fault location label, and confidence score. The final diagnostic output of the output model contains a complete diagnostic result including fault type, fault location, and confidence score. In other words, the fault information of the secondary circuit includes a complete diagnostic result including fault type, fault location, and confidence score.

[0140] Figure 3 A flowchart illustrating the fault detection method based on a secondary loop provided in this application embodiment. Figure 3 ,like Figure 3 As shown, the method includes:

[0141] S301. For each preset time, determine the neighboring times of the preset time; wherein, the neighboring times represent the two preset times that are before the preset time.

[0142] For example, for each preset time, the neighboring times are determined according to the chronological order of the preset time. The neighboring times represent the two preset times preceding the current preset time; that is, the neighboring times are the two preset times preceding the current preset time.

[0143] S302. Determine the preset sampling frequency corresponding to the preset time based on the voltage and current information of each neighboring time corresponding to the preset time.

[0144] For example, for each preset time point, a preset sampling frequency is determined based on the voltage and current information of each of the two neighboring time points corresponding to that preset time point. For instance, the rate of change of voltage can be determined based on the voltage information of each of the two neighboring time points. Simultaneously, the rate of change of current can be determined based on the current information of each of the two neighboring time points. Based on the rates of change of voltage and current, the preset sampling frequency for that preset time point is set.

[0145] Specifically, the above-mentioned determination of the preset sampling frequency based on the voltage and current information of each neighboring time point at the preset time point includes:

[0146] Based on the voltage information at each neighboring time point, a first rate of change is determined. The first rate of change characterizes the rate of change of voltage information between two neighboring time points. Based on the current information at each neighboring time point, a second rate of change is determined. The second rate of change characterizes the rate of change of current information between two neighboring time points. Based on the first and second rates of change, a preset sampling frequency corresponding to a preset time point is determined.

[0147] For example, for each preset time point, the difference in voltage information between two neighboring time points corresponding to that preset time point is calculated. For instance, for any preset voltage range, a first difference in voltage information collected by the voltage sensor within that preset voltage range at two neighboring time points is calculated. Then, a first ratio of the first difference to the time interval between the two neighboring time points is calculated. Next, a first average of the first ratios corresponding to all preset voltage ranges is calculated, and this first average is used as the first rate of change.

[0148] Similarly, for each preset time point, a second difference in current information between two neighboring time points corresponding to that preset time point is calculated. For example, for any preset current range, a second difference in current information collected by the current sensor within that preset current range at two neighboring time points is calculated. Then, a second ratio of the second difference to the time interval between the two neighboring time points is calculated. Finally, a second average of the second ratios corresponding to all preset voltage ranges is calculated, and this second average is used as the second rate of change.

[0149] After obtaining the first and second rates of change, a preset sampling frequency corresponding to a preset time is determined based on these rates. For example, if the first rate of change is less than a first preset threshold and the second rate of change is less than a second preset threshold, the secondary loop is determined to be in a steady-state operation, and the preset sampling frequency can be set to 1 kHz to reduce the amount of data. When the first rate of change is greater than or equal to the first preset threshold, or the second rate of change is greater than or equal to the second preset threshold, the secondary loop is determined to be in a transient state, and the preset sampling frequency can be set to 10 kHz to accurately capture the transient process of the secondary loop. The sampling clock frequency can be adjusted in real time using a field-programmable gate array (FPGA) to obtain adaptive sampling frequency control information, and the control parameters of the sampling frequency can be dynamically adjusted based on this adaptive sampling frequency control information.

[0150] S303. Obtain the voltage and current information of the secondary circuit at a preset time.

[0151] S304. Determine the fault information of the secondary circuit based on the voltage and current information at each preset time.

[0152] In this embodiment, after determining the fault information of the secondary circuit, a hierarchical visualization interface can be used to display the monitoring results. For example, a hierarchical information organization method is adopted to divide the diagnostic information into three display levels according to importance and level of detail. The first level is set as the diagnostic summary level, displaying a textual description of the fault type and a confidence percentage; the second level is the fault location level, displaying a control loop topology diagram and fault location markings; and the third level is the detailed analysis level, displaying signal waveform curves and characteristic parameter values. This results in hierarchically organized display data, with diagnostic information data organized according to the three levels.

[0153] Based on hierarchically organized display data, a user interface can be drawn on a 7-inch 800×480 resolution color touchscreen using LCD touchscreen rendering methods. The first layer occupies the upper 1 / 4 of the screen, displaying a detected tripping relay fault in large font with a 92% confidence level. The second layer occupies the middle 1 / 2 of the screen, showing a simplified control circuit schematic, with the fault location highlighted in red. The third layer occupies the lower 1 / 4 of the screen, allowing users to view the signal waveforms of each channel by touching and sliding. This results in a user-friendly, hierarchical touchscreen interface.

[0154] Based on a visual display interface, a color-coding method is used to represent diagnostic confidence levels. A green background is used when the confidence level is above 90%; a yellow background is used when the confidence level is between 70% and 90% with a suggestion for manual review; and an orange background is used when the confidence level is below 70% with an "uncertain" warning requiring manual judgment, while also listing multiple possible causes of the fault for maintenance personnel to choose from. This provides an intuitive visualization of confidence levels using a color-coded confidence level display scheme.

[0155] In this embodiment, a device database query method can also be used to retrieve detailed information about the device from the device database based on the current switchgear number. This information includes the manufacturer's name, device model, manufacturing date, control circuit schematic diagram, and component list. This yields detailed parameter data for the switchgear, a dataset containing complete technical information about the switchgear.

[0156] Based on detailed parameter data and fault location information of the switching equipment, a schematic diagram analysis method is used to locate the circuit component corresponding to the fault location from the control circuit schematic diagram, and extract information such as the component's model, specifications, terminal number, and wiring relationship. For example, when the fault location is a tripping relay, the model, coil rated voltage, normally closed contact terminal number, and normally open contact terminal number of the tripping relay can be found from the schematic diagram. This yields the specific parameters of the faulty component and the detailed technical specifications of the component corresponding to the fault location.

[0157] It can also use a knowledge base matching method based on fault type information and specific parameters of the faulty component to retrieve standard handling procedure templates for the corresponding fault type from a preset handling knowledge base. The general component descriptions in the templates are replaced with specific component parameters to generate targeted handling suggestion text. For example, suggested handling steps: 1. Use a multimeter to measure the resistance between terminals 11-12; the normal value should be less than 100Ω; 2. Check if the terminal wiring is loose; 3. If the resistance is abnormal, replace the trip relay with model A. This provides actionable, dynamic handling suggestions and detailed handling steps for specific faults.

[0158] In this embodiment, a confidence level grading strategy can also be adopted, setting two thresholds: a high confidence threshold of 0.9 and a medium confidence threshold of 0.7. Based on the comparison between the confidence score and the threshold, three confidence levels—high, medium, and low—are determined. This yields confidence level identifiers and classification results for the high, medium, and low confidence levels.

[0159] Based on confidence level indicators, a human-computer interaction feedback mechanism is adopted. High-confidence diagnostic results are directly displayed and automatically recorded; medium-confidence results are displayed with "Confirm" and "Correct" buttons for maintenance personnel to choose from; low-confidence results display a list of multiple possible fault causes for maintenance personnel to select from, and the maintenance personnel's confirmation, correction, or selection actions are recorded as feedback information. This yields labeled data containing both diagnostic results and human feedback, combined with complete labeled information from machine diagnosis and human confirmation.

[0160] Based on labeled data, a multi-task loss function is used to train the fault prediction model. This includes a first loss function (which can be a cross-entropy loss function) to calculate the fault type classification loss; a second loss function (which can also be a cross-entropy loss function) to calculate the fault location loss; and a third loss function (which can be a mean squared error loss function) to calculate the confidence assessment loss. Cross-entropy loss is a common method for measuring the difference between two probability distributions. Mean squared error loss is used in regression tasks to measure the difference between predicted and true values. By assigning preset weights to the first, second, and third loss functions, and then summing them using weighted averages, a total loss function is obtained. The preset weights of each loss function can be adjusted based on the importance of each task. The total loss value is then used as the comprehensive loss function value for training the fault prediction model.

[0161] Based on labeled data and diagnostic results, a data classification and uploading method can be adopted: high-confidence diagnostic results are uploaded as diagnostic reports; medium-confidence results are uploaded as events awaiting review; and low-confidence results and feedback from maintenance personnel are uploaded as training samples. Data is sent to the distribution automation master station system via RS485 or Ethernet interfaces, and the master station system centrally stores and analyzes the collected data. This results in multi-level information sharing, classified diagnostic and feedback data uploaded to the master station system, and a hierarchical visual display interface, targeted dynamic handling suggestion text, maintenance personnel feedback labeled data, and diagnostic information uploaded to the master station.

[0162] In this embodiment, based on the dynamic handling suggestion text, a fault information verification method can be used to compare the fault type, fault location, and confidence score in the diagnostic results with the historical fault database. A similarity matching algorithm can be used to calculate the similarity between the current fault features and historical cases, using Euclidean distance as the metric. When the similarity exceeds a threshold of 0.85, it is classified as a known fault mode; when the similarity is below a threshold of 0.6, it is classified as a novel fault mode; and when the similarity is between the two thresholds, it is classified as a suspected fault mode. The resulting fault mode classification includes classification labels for known, suspected, and novel fault modes.

[0163] Based on the fault mode classification results and diagnostic confidence scores, a multi-dimensional verification strategy is adopted. For known fault modes and diagnostic results with a confidence score greater than 0.8, they are directly confirmed as credible faults. For suspected fault modes or diagnostic results with a confidence score between 0.6 and 0.8, maintenance personnel are required to conduct preliminary on-site verification. For novel fault modes or diagnostic results with a confidence score below 0.6, maintenance personnel are required to conduct detailed on-site inspections and report to professional technicians. A fault confirmation level identifier is generated according to the verification strategy. This results in fault confirmation levels and verification requirements, including confirmation results and corresponding verification requirements for three levels: credible faults, faults requiring verification, and faults requiring detailed investigation.

[0164] Based on fault confirmation levels and verification requirements, a tiered response mechanism is adopted to directly generate standardized verification checklists for trusted faults. These checklists include fault location checks, signal measurement items, and safety precautions. For faults requiring verification, an expanded verification checklist is generated, adding relevant equipment status checks and signal comparison measurements. For faults requiring detailed investigation, a comprehensive verification checklist is generated. This includes systematic troubleshooting steps and professional testing requirements, and the verification checklists are presented to maintenance personnel through a display interface. This results in tiered verification checklists and structured verification guidance documents for different fault confirmation levels.

[0165] Based on the tiered verification checklist and fault confirmation results, a knowledge base matching method is used to retrieve standard repair solutions matching the current fault type and location from the built-in fault handling knowledge base. The knowledge base contains standard handling procedures for seven types of faults, each categorized into three levels of severity: minor, moderate, and severe. Fault severity can be automatically determined based on fault characteristic parameters (such as voltage deviation amplitude, current abnormality duration, contact operation frequency, etc.). The resulting standard repair solution includes a fault severity assessment and the corresponding standard handling procedure.

[0166] Based on the matching standard repair scheme, an operation step generation method is used to convert the standard repair scheme into specific operation guidance steps. Each step includes four elements: operation content, safety requirements, expected results, and anomaly handling. Operation parameters are automatically adjusted according to the current equipment configuration parameters, such as adjusting the measurement range based on the control power supply voltage level and the test current value based on the relay model. The steps are ordered according to the order and logical relationship of the operations. This results in a personalized sequence of operation guidance steps, providing a complete operation guide containing specific operation content, safety requirements, expected results, and anomaly handling.

[0167] Based on a personalized operation guidance sequence, an interactive guidance method is adopted, displaying the operation steps step by step through the interface. After each step is completed, the operation results are required for confirmation by the operations and maintenance personnel, and the success of the operation is determined based on the feedback from the operations and maintenance personnel. When the operation results are consistent with expectations, the system automatically proceeds to the next step; when the operation results are abnormal, troubleshooting suggestions are provided or the emergency handling process is initiated. The execution time and result status of each operation step are recorded. An interactive operation execution record is obtained, which includes the execution status of each operation step, timestamps, and a complete operation log of result confirmation.

[0168] Based on interactive operation execution records, a functional verification test method is adopted, automatically initiating the verification test program after fault repair is completed. The test program consists of two phases: basic function testing and fault reproduction testing. Basic function testing verifies the normal operation of the control loop, including remote control opening and closing operations, position signal feedback, and relay action response. Fault reproduction testing simulates the original fault conditions under safe conditions to verify whether the fault has been completely eliminated. This results in a functional verification test plan, containing detailed test plans for both basic function testing and fault reproduction testing.

[0169] Based on the functional verification test scheme, an automated test execution method is adopted. A built-in test signal generator produces standard test signals to simulate control and feedback signals under normal operating conditions. Simultaneously, the response signals of all relevant test points, including voltage, current, and contact status, are monitored. The actual test results are compared with the expected standard values, and the deviation rate and response time are calculated. The test is considered passed when the deviation rate of all test items is less than 5% and the response time is within the normal range. An automated test report is obtained, including the actual value, standard value, deviation rate, and pass / fail status of each test item.

[0170] Based on the results of automated testing, a comprehensive evaluation method is used. Items that pass the test are marked as successfully repaired, while the reasons for failure are analyzed for items that fail. The overall repair success rate is calculated as the number of items that passed the test divided by the total number of items tested. When the success rate reaches 100%, the fault is confirmed to be completely repaired; when the success rate is between 80% and 99%, the fault is confirmed to be partially repaired and requires further processing; when the success rate is below 80%, the repair is confirmed to have failed and requires re-diagnosis. A repair verification report is generated, providing a comprehensive evaluation result that includes the repair success rate, analysis of failed items, and suggestions for subsequent processing.

[0171] Based on the repair verification and evaluation results, a structured recording method is used to organize the key information of the entire fault handling process in a structured manner. The recorded content includes the fault occurrence time, fault type, fault location, diagnostic confidence level, repair operation steps, verification test results, and repair time. Data is organized according to a standardized fault record format to ensure the completeness and traceability of the recorded information. This results in a structured fault handling record, containing standardized data records of the entire fault lifecycle.

[0172] Based on structured fault handling records, an intelligent report generation method is employed to automatically generate fault handling reports using preset report templates. The report comprises six parts: execution summary, fault description, diagnostic analysis, handling process, verification results, and experience summary. The report content is automatically populated based on the fault type and handling results. For complex faults or repair failures, professional technical suggestions and follow-up requirements are automatically added. The resulting automatically generated fault handling report is a standardized technical report containing a complete fault handling process and result analysis.

[0173] Based on automatically generated fault handling reports and raw diagnostic data, a multi-channel data upload method is employed to upload fault handling reports to the distribution automation master station system via RS485 or Ethernet interfaces. Simultaneously, detailed data such as raw signal data, diagnostic results, and operation records are uploaded to the fault analysis database. Successfully repaired fault cases are marked as positive samples; cases that fail to be repaired or are misdiagnosed are marked as negative samples, thus providing training data for subsequent model optimization. Multi-level data archiving and sharing are achieved, including data management feedback on report upload confirmation, database storage status, and sample labeling results. This also includes complete fault handling records, verification reports, and database storage confirmation.

[0174] Figure 4 This is a schematic diagram of the structure of a fault detection device based on a secondary loop provided in an embodiment of this application, as shown below. Figure 4 As shown, the fault detection device 40 based on a secondary circuit provided in this embodiment includes:

[0175] The acquisition module 401 is used to acquire voltage and current information of the secondary circuit at a preset time. The voltage information represents the voltage collected by a voltage sensor with a preset voltage range at a preset sampling frequency. The current information represents the current collected by a current sensor with a preset current range at a preset sampling frequency. There are at least two preset voltage ranges and at least two preset current ranges. The preset voltage ranges and preset current ranges correspond one-to-one. The preset time corresponds one-to-one with the preset sampling frequency.

[0176] The determination module 402 is used to determine the fault information of the secondary circuit based on the voltage and current information at each preset time.

[0177] In one possible implementation, the determining module 402 is further configured to:

[0178] Based on a preset time length, all preset times are grouped to obtain multiple time groups; each time group contains multiple preset times within a preset time length; there are multiple preset time lengths.

[0179] For each time group corresponding to a preset time length, the characteristic information of the time group is determined based on the voltage and current information at each preset time within the time group; the characteristic information of the time group represents the voltage and current changes of the secondary circuit within the preset time length.

[0180] Based on the characteristic information of each time group corresponding to each preset time length, the fault information of the secondary circuit is determined.

[0181] In one possible implementation, the determining module 402 is further configured to:

[0182] For each preset time within the time group, the first information at the preset time is determined based on the voltage information collected by the voltage sensor within each preset voltage range at the preset time; the first information at the preset time characterizes the voltage status of the secondary circuit at the preset time.

[0183] Based on the current information collected by the current sensors within each preset current range at a preset time, the second information at the preset time is determined; the second information at the preset time characterizes the current situation of the secondary circuit at the preset time.

[0184] Based on the first information and the second information at the preset time, the sub-information at the preset time is determined; the sub-information at the preset time represents the voltage and current conditions of the secondary circuit at the preset time.

[0185] The characteristic information of the time group is determined based on the sub-information at each preset time within the time group.

[0186] In one possible implementation, the determining module 402 is further configured to:

[0187] Based on the voltage information collected by the voltage sensor in each preset voltage range at a preset time and the first preset weight of each preset voltage range, the first information at the preset time is determined; the first preset weight represents the degree of influence of the voltage information collected by the voltage sensor in the preset voltage range on the first information.

[0188] Based on the current information collected by current sensors within each preset current range at a preset time, the second information at the preset time is determined, including:

[0189] The second information at the preset time is determined based on the current information collected by the current sensor in each preset current range at the preset time and the second preset weight of each preset current range; the second preset weight represents the degree of influence of the current information collected by the current sensor in the preset current range on the second information.

[0190] In one possible implementation, the determining module 402 is further configured to:

[0191] Based on the sub-information at each preset time within the time group, the time-domain information of the time group is determined; the time-domain information represents the mean, standard deviation, and peak value of the sub-information at all preset times within the time group.

[0192] For each preset time within the time group, the sub-information at the preset time is subjected to frequency domain transformation to obtain transformed information; the transformed information represents the information of the sub-information at the preset time in the frequency domain.

[0193] Based on the conversion information at each preset time within the time group, the frequency domain information of the time group is determined; the frequency domain information represents the main frequency component and spectral energy of the conversion information at all preset times within the time group.

[0194] The characteristic information of the time group is determined based on the time domain information and the frequency domain information of the time group.

[0195] In one possible implementation, the preset time length includes at least one of the following: a first preset time length, a second preset time length, and a third preset time length; the first preset time length is less than the second preset time length, and the second preset time length is less than the third preset time length; the determining module 402 is further configured to:

[0196] For each time group corresponding to the first preset time length, determine the median of the first information at all preset times within the time group;

[0197] Based on the median corresponding to the time group, the first group is selected from all time groups corresponding to the second preset time length; and the second group is selected from all time groups corresponding to the third preset time length.

[0198] The scale information of the time group is determined based on the feature information of the time group, the feature information of the first group, and the feature information of the second group; the scale information of the time group represents the information that integrates the feature information of all time groups with preset time lengths.

[0199] Based on the scale information corresponding to each time group corresponding to the first preset time length, the fusion feature information is determined; the fusion feature information represents the information that fuses the scale information of all time groups.

[0200] Based on the fusion feature information, the fault information of the secondary circuit is determined.

[0201] In one possible implementation, the determining module 402 is further configured to:

[0202] The fused feature information is input into the fault prediction model to obtain the fault information of the secondary loop. The fault prediction model is a trained neural network model, which is used to perform data reasoning on the input fused feature information to obtain the fault information of the secondary loop.

[0203] In one possible implementation, the fault detection device 40 based on a secondary circuit further includes a second determining module, used for:

[0204] For each preset time, determine the neighboring times of the preset time; where the neighboring times represent the two preset times that precede the preset time.

[0205] Based on the voltage and current information of each neighboring time corresponding to the preset time, the preset sampling frequency is determined.

[0206] In one possible implementation, the second determining module is further configured to:

[0207] The first rate of change is determined based on the voltage information at each neighboring time point; the first rate of change characterizes the rate of change of voltage information between two neighboring time points.

[0208] The second rate of change is determined based on the current information at each neighboring time; the second rate of change characterizes the rate of change of the current information at two neighboring time points.

[0209] The preset sampling frequency corresponding to the preset time is determined based on the first rate of change and the second rate of change.

[0210] The fault detection device based on a secondary circuit provided in this embodiment can execute the method provided in the above-described method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0211] Figure 5 This is a schematic diagram of the structure of a fault detection device based on a secondary loop, provided in an embodiment of this application. Figure 5 As shown, the secondary loop-based fault detection device 50 provided in this embodiment includes at least one processor 501 and a memory 502. Optionally, the secondary loop-based fault detection device 50 further includes a communication component 503. The processor 501, memory 502, and communication component 503 are connected via a bus.

[0212] In a specific implementation, at least one processor 501 executes computer execution instructions stored in memory 502, causing at least one processor 501 to perform the above-described method.

[0213] The specific implementation process of processor 501 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0214] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0215] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0216] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0217] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0218] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0219] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0220] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0221] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0222] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0223] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0224] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0225] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0226] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for detecting a fault based on a secondary circuit, characterized in that, The method comprises: acquiring voltage information and current information of the secondary circuit at preset time points; the voltage information represents the voltage collected by the voltage sensor of the preset voltage range at the preset sampling frequency; the current information represents the current collected by the current sensor of the preset current range at the preset sampling frequency; the preset voltage range has at least two; the preset current range has at least two; the preset voltage range corresponds to the preset current range one by one; the preset time point corresponds to the preset sampling frequency one by one; determining the fault information of the secondary circuit according to the voltage information and the current information at each preset time point.

2. The method of claim 1, wherein, determining the fault information of the secondary circuit according to the voltage information and the current information at each preset time point, comprising: grouping all preset time points based on a preset time length to obtain a plurality of time groups; the time group contains a plurality of preset time points within the preset time length; the preset time length has a plurality of time lengths; for each time group corresponding to the preset time length, determining the characteristic information of the time group according to the voltage information and the current information at each preset time point in the time group; the characteristic information of the time group represents the voltage change and the current change of the secondary circuit within the preset time length; determining the fault information of the secondary circuit according to the characteristic information of each time group corresponding to each preset time length.

3. The method of claim 2, wherein, determining the characteristic information of the time group according to the voltage information and the current information at each preset time point in the time group, comprising: for each preset time point in the time group, determining the first information at the preset time point according to the voltage information collected by the voltage sensor of each preset voltage range at the preset time point; the first information at the preset time point represents the voltage of the secondary circuit at the preset time point; determining the second information at the preset time point according to the current information collected by the current sensor of each preset current range at the preset time point; the second information at the preset time point represents the current of the secondary circuit at the preset time point; determining the sub information at the preset time point according to the first information at the preset time point and the second information at the preset time point; the sub information at the preset time point represents the voltage and the current of the secondary circuit at the preset time point; determining the characteristic information of the time group according to the sub information at each preset time point in the time group.

4. The method of claim 3, wherein, determining the first information at the preset time point according to the voltage information collected by the voltage sensor of each preset voltage range at the preset time point, comprising: determining the first information at the preset time point according to the voltage information collected by the voltage sensor of each preset voltage range at the preset time point and the first preset weight of each preset voltage range; the first preset weight represents the influence degree of the voltage information collected by the voltage sensor of the preset voltage range on the first information; determining the second information at the preset time point according to the current information collected by the current sensor of each preset current range at the preset time point, comprising: According to the current information of each preset time point in the time group, the feature information of the time group is determined, including:

5. The method of claim 3, wherein, According to the current information of each preset time point in the time group, the time domain information of the time group is determined; the time domain information represents the mean, standard deviation, and peak value of the current information of all preset time points in the time group; For each preset time point in the time group, the frequency domain conversion processing is performed on the current information of the preset time point to obtain conversion information; the conversion information represents the information of the current information of the preset time point in the frequency domain; According to the conversion information of each preset time point in the time group, the frequency domain information of the time group is determined; the frequency domain information represents the main frequency component and spectral energy of the conversion information of all preset time points in the time group; According to the time domain information of the time group and the frequency domain information of the time group, the feature information of the time group is determined. The preset time length includes at least one of the following: a first preset time length, a second preset time length, and a third preset time length; the first preset time length is less than the second preset time length, and the second preset time length is less than the third preset time length; 6. The method of claim 4, wherein, According to the feature information of each time group corresponding to each preset time length, the fault information of the secondary circuit is determined, including: For each time group corresponding to the first preset time length, the median of the first information of all preset time points in the time group is determined; According to the median corresponding to the time group, a first group is selected from all time groups corresponding to the second preset time length, and a second group is selected from all time groups corresponding to the third preset time length; According to the feature information of the time group, the feature information of the first group, and the feature information of the second group, the scale information of the time group is determined; the scale information of the time group represents the information of the feature information of the time group fused with all preset time lengths; According to the scale information corresponding to each time group corresponding to the first preset time length, the fusion feature information is determined; the fusion feature information represents the information fused with the scale information of all time groups; According to the fusion feature information, the fault information of the secondary circuit is determined. According to the fusion feature information, the fault information of the secondary circuit is determined, including:

7. The method of claim 6, wherein, The fusion feature information is input into a fault prediction model to obtain the output fault information of the secondary circuit; the fault prediction model is a trained neural network model, and the fault prediction model is used for data reasoning on the input fusion feature information to obtain the fault information of the secondary circuit. Further comprising:

8. The method according to any one of claims 1-7, characterized in that, For each preset time point, the neighborhood time point of the preset time point is determined; wherein the neighborhood time point represents two preset time points before the preset time point; ​ According to the voltage information and the current information at each adjacent time corresponding to the preset time, a preset sampling frequency corresponding to the preset time is determined.

9. The method of claim 8, wherein, According to the voltage information and the current information at each adjacent time corresponding to the preset time, a preset sampling frequency corresponding to the preset time is determined, including: According to the voltage information at each adjacent time, a first change rate is determined; the first change rate represents a change speed of the voltage information at two adjacent times; According to the current information at each adjacent time, a second change rate is determined; the second change rate represents a change speed of the current information at two adjacent times; According to the first change rate and the second change rate, a preset sampling frequency corresponding to the preset time is determined.

10. A secondary circuit based fault detection apparatus, characterized by, including: An acquisition module is configured to acquire voltage information and current information of a secondary circuit at a preset time; The voltage information represents a voltage collected by a preset voltage range of a voltage sensor at a preset sampling frequency; the current information represents a current collected by a preset current range of a current sensor at a preset sampling frequency; The preset voltage range has at least two; the preset current range has at least two; the preset voltage range corresponds to the preset current range one by one; the preset time corresponds to the preset sampling frequency one by one; A determination module is configured to determine fault information of the secondary circuit according to the voltage information and the current information at each preset time.

11. A secondary circuit based fault detection apparatus, characterized by, including: A memory and a processor; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory, so that the processor executes the method in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by the processor to implement the method in any one of claims 1-9.

13. A computer program product, characterised in that, The computer program is executed by the processor to implement the method in any one of claims 1-9.

Images