Electric leakage detection system for distribution box based on multi-sensor data fusion

Abstract

The application discloses a power distribution box electric leakage detection system based on multi-sensor data fusion and belongs to the technical field of power distribution box electric leakage detection. The system comprises the following steps: an electric leakage detection sensor unit collects multi-dimensional electric leakage related physical quantities of the incoming line end and the outgoing line end of a power distribution box; an electric leakage feature analysis unit constructs an incoming line end electric leakage reference vector and an outgoing line end electric leakage reference vector according to the numerical relationship between the physical quantities; a difference measurement unit obtains the structural difference between the two vectors and extracts fluctuation form parameters according to the change trajectory of the structural difference at multiple time points; an electric leakage state evaluation unit adjusts the comparison scale according to the stability of the fluctuation form parameters and generates a preliminary electric leakage confidence; and an electric leakage grade determination unit corrects the preliminary electric leakage confidence by using the distribution interval state of the physical quantities within the preset range, obtains the final electric leakage confidence and matches an electric leakage grade signal output. The scheme can reduce false positives caused by load fluctuations and harmonic interference and improve the reliability of electric leakage determination.

Description

Technical Field

[0001] This invention relates to the field of leakage current detection technology for distribution boxes, specifically to a leakage current detection system for distribution boxes based on multi-sensor data fusion. Background Technology

[0002] Leakage detection in distribution boxes is a crucial measure to ensure electrical safety. Most existing leakage detection solutions rely on residual current transformers to directly measure the difference between the vector sum of the currents at the incoming and outgoing terminals. An alarm is triggered when the differential current exceeds a fixed threshold. Some improved solutions identify leakage by comparing the differences in the effective values ​​or power of the currents in each phase at the incoming and outgoing terminals. However, these techniques face several limitations in practical applications. The widespread use of nonlinear loads in distribution circuits leads to increased harmonic current content. Harmonic currents may form paths through the distributed capacitance to ground, generating residual currents similar to actual leakage characteristics. Simply relying on current differences or effective value differences cannot distinguish between leakage caused by harmonic interference and insulation degradation. Furthermore, random load switching and fluctuations during normal operation can cause instantaneous asymmetric changes in the electrical quantity relationship between the incoming and outgoing terminals. Fixed threshold judgment methods can easily misjudge such normal operating conditions as leakage faults, resulting in frequent alarms. Existing methods fail to fully explore the intrinsic correlation structure between multidimensional electrical quantities such as voltage, current RMS values, and harmonic distortion rate, and also lack the ability to adaptively analyze the dynamic characteristics of time series of differential signals, resulting in decreased detection accuracy in interference environments.

[0003] How to construct a feature representation that is sensitive to leakage current and robust to normal load fluctuations from the multidimensional electrical quantities at the incoming and outgoing ends, and dynamically adjust the discrimination scale according to the stability of the feature change trajectory during the judgment process, is a key problem that needs to be solved to improve the anti-interference capability and accuracy of leakage current detection in distribution boxes. Summary of the Invention

[0004] This paper presents a leakage current detection system for distribution boxes based on multi-sensor data fusion to solve the problem of false leakage current alarms caused by load fluctuations and harmonic interference in the existing technology. The system enables adaptive adjustment of detection criteria according to different operating conditions, thereby improving the accuracy of leakage current status assessment and level determination.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a leakage current detection system for distribution boxes based on multi-sensor data fusion, including a leakage current detection sensing unit, a leakage current feature analysis unit, a difference measurement unit, a leakage current status assessment unit, and a leakage current level determination unit. The leakage current detection sensing unit is used to collect multi-dimensional leakage current related physical quantities at the incoming and outgoing ends of the distribution box, and transmit the collected physical quantities to the leakage current feature analysis unit. As a preferred embodiment, the multi-dimensional leakage current related physical quantities include the effective value of the incoming voltage, the effective value of the incoming current, the harmonic distortion rate of the incoming current, the effective value of the outgoing voltage, the effective value of the outgoing current, and the harmonic distortion rate of the outgoing current; the leakage current detection sensing unit includes at least a voltage transformer, a current transformer, and a harmonic analysis module, the voltage transformer and the current transformer are respectively installed at the incoming and outgoing ends, and the harmonic analysis module is used to extract the current harmonic distortion rate. By simultaneously acquiring multiple types of electrical quantities at the incoming and outgoing terminals, it is possible to comprehensively capture subtle changes in voltage, current, and harmonic characteristics caused by leakage, providing a multi-dimensional information basis for subsequent fusion judgment.

[0006] The leakage current characteristic analysis unit constructs an incoming-end leakage current reference vector based on the numerical relationships between various physical quantities at the incoming end at the same sampling time, and constructs an outgoing-end leakage current reference vector based on the numerical relationships between various physical quantities at the outgoing end at the same sampling time. In a specific technical solution of the present invention, when constructing the incoming-end leakage current reference vector, the ratio of the effective value of the incoming-end voltage to the effective value of the incoming-end current is used as the equivalent impedance of the incoming end, and the product of the effective value of the incoming-end current and the harmonic distortion rate of the incoming-end current is used as the harmonic current amplitude of the incoming end. Then, the equivalent impedance of the incoming end and the harmonic current amplitude of the incoming end are arranged and combined according to a preset dimension to form the incoming-end leakage current reference vector; the outgoing-end leakage current reference vector adopts the same construction method, that is, the equivalent impedance of the outgoing end and the harmonic current amplitude of the outgoing end are calculated and arranged and combined according to the same dimension. The leakage current reference vector obtained in this way integrates impedance characteristics and harmonic current amplitude, and can sensitively reflect the asymmetric characteristics caused by leakage faults at both ends of the power distribution line.

[0007] The difference measurement unit acquires the structural difference degree between the incoming leakage reference vector and the outgoing leakage reference vector at the same time. Preferably, the calculation process of the structural difference degree is as follows: calculate the absolute value of the difference between the sum of the values ​​of each dimension of the incoming leakage reference vector and the sum of the values ​​of each dimension of the outgoing leakage reference vector, as the total difference value; calculate the difference sequence of adjacent dimension values ​​in the incoming leakage reference vector and the difference sequence of adjacent dimension values ​​in the outgoing leakage reference vector, and obtain the sum of the absolute values ​​of the differences between the two difference sequences at corresponding positions, as the distribution difference value; sum the total difference value and the distribution difference value according to a preset weight to obtain the structural difference degree. This difference degree measures both the deviation of the electrical characteristics at both ends at the total level and captures the difference in their internal structural distribution, which can more robustly characterize the leakage state. On this basis, the difference measurement unit extracts the fluctuation morphology parameters according to the change trajectory of the structural difference degree at multiple consecutive times. As a preferred approach, the structural differences at multiple time points are arranged in ascending order to form a difference change sequence. The difference between structural differences at adjacent time points is calculated to obtain a difference difference sequence. The number of times the sign of the values ​​in the difference difference sequence changes, i.e., the number of times the product of two adjacent difference values ​​is negative, is counted as the fluctuation reversal frequency. Simultaneously, the average of all positive values ​​and the average of all absolute values ​​of negative values ​​in the difference difference sequence are calculated, and the ratio between the average of the positive values ​​and the average of the absolute values ​​of negative values ​​is used as the fluctuation amplitude balance. The fluctuation reversal frequency and the fluctuation amplitude balance are combined to form a fluctuation morphology parameter. This fluctuation morphology parameter can effectively characterize the frequency of direction changes in difference fluctuations and the symmetry of positive and negative fluctuations, providing a basis for adaptive adjustment judgment thresholds.

[0008] The leakage current assessment unit adjusts the comparison scale of the structural difference between the incoming leakage current reference vector and the outgoing leakage current reference vector based on the stability of the fluctuation morphology parameters within a preset time period prior to the current moment, generating a preliminary leakage current confidence level for the current moment. Specifically, it obtains the fluctuation turning frequency corresponding to each moment within the preset time period, calculates the coefficient of variation of the fluctuation turning frequency as a stability characterization value, matches the corresponding relaxation coefficient according to the numerical range of the stability characterization value, multiplies the structural difference at the current moment by the relaxation coefficient to obtain the adjusted structural difference, and then maps the adjusted structural difference to a preset confidence interval to obtain the preliminary leakage current confidence level. When the system is subjected to load fluctuations or harmonic interference causing drastic fluctuations in the difference, the stability is low, and a larger relaxation coefficient is automatically adopted to reduce the impact of instantaneous anomalies on the judgment; when the system state is stable, a smaller relaxation coefficient is used to maintain high detection sensitivity, thereby achieving a balance between detection real-time performance and anti-interference capability.

[0009] The leakage current level determination unit corrects the initial leakage current confidence level based on the distribution range of the multi-dimensional leakage current related physical quantities within a preset range at the current moment, thus obtaining the final leakage current confidence level at the current moment. In a preferred embodiment of the present invention, the effective values ​​of the incoming voltage, incoming current, outgoing voltage, and outgoing current at the current moment are each determined to belong to a specific interval level within a preset range, and the sum of all interval levels is calculated to obtain a range distribution characteristic value. When the range distribution characteristic value is greater than a preset correction trigger threshold, it indicates that the overall physical quantities are in a high operating range, and the leakage current risk is relatively prominent. The initial leakage current confidence level is multiplied by a first correction coefficient greater than 1. When the range distribution characteristic value is less than or equal to the correction trigger threshold, the initial leakage current confidence level is multiplied by a second correction coefficient less than 1 to obtain the final leakage current confidence level. This correction step uses the actual operating range of electrical quantities to weight the leakage current inference, which can effectively suppress false alarms caused by small signal disturbances and improve the confidence level of leakage current reports under heavy load conditions.

[0010] The leakage current level determination unit further matches the corresponding leakage current level signal output based on the final leakage current confidence level. Specifically, the final leakage current confidence level is compared with a preset first-level threshold and a second-level threshold, where the first-level threshold is greater than the second-level threshold. When the final leakage current confidence level is greater than or equal to the first-level threshold, a first leakage current level signal is output; when the final leakage current confidence level is less than the first-level threshold but greater than or equal to the second-level threshold, a second leakage current level signal is output; and when the final leakage current confidence level is less than the second-level threshold, a third leakage current level signal is output. Outputting graded signals through multi-level threshold comparisons facilitates differentiated handling based on the severity of leakage current. The overall system integrates multi-sensor data fusion, vector structure difference analysis, adaptive adjustment of fluctuation patterns, and range interval state correction, significantly improving the accuracy and reliability of leakage current detection in distribution boxes, and maintaining excellent anti-interference performance and alarm accuracy even in complex electrical environments.

[0011] The technical effects and advantages provided by the present invention in the above technical solution are as follows: By fusing the effective values ​​of the incoming voltage, current, and harmonic distortion rate into two dimensions—equivalent impedance and harmonic current amplitude—and arranging them along the same dimensions to construct an incoming leakage reference vector, and simultaneously constructing an outgoing leakage reference vector from the corresponding physical quantities at the outgoing end, the structural difference between the two vectors is calculated from two levels: total difference and distribution difference. Specifically, the absolute value of the difference between the sums of the values ​​of each dimension of the two vectors is taken as the total difference value, and the sum of the absolute values ​​of the differences at corresponding positions in the adjacent dimension difference sequences of the two vectors is calculated as the distribution difference value. These two values ​​are then weighted and summed. This multi-dimensional vector structural difference measurement method captures the asymmetric change in the ratio of equivalent impedance to harmonic current amplitude at both the incoming and outgoing ends when leakage occurs. Compared to comparing a single current difference or effective value, it can effectively distinguish between parameter changes caused by harmonic leakage paths and symmetrical load fluctuations, reducing the probability of misreporting normal operating conditions as leakage.

[0012] After obtaining the structural difference degree at each moment, a difference degree change sequence is generated along the time axis and a difference sequence is calculated. From this, the fluctuation turning frequency and fluctuation amplitude balance are extracted as fluctuation morphology parameters. The fluctuation turning frequency characterizes the frequency at which the structural difference degree changes direction between adjacent moments, and the fluctuation amplitude balance reflects the intensity symmetry between positive and negative fluctuations. The stability of the fluctuation morphology is measured using the coefficient of variation of the fluctuation turning frequency within a preset time period before the current moment, and a corresponding relaxation coefficient is matched accordingly. After scaling the structural difference degree at the current moment, a preliminary leakage current confidence level is obtained. Through this mechanism, during continuous monitoring, when the electrical environment is relatively stable and the fluctuation morphology parameters are stable, the relaxation coefficient automatically widens the comparison scale to maintain sensitivity to minor leakage defects. When the load fluctuates frequently and the fluctuation turning frequency and amplitude balance show large dispersion, the relaxation coefficient automatically tightens the comparison scale to avoid artificially high confidence levels due to process disturbances. Subsequently, by combining the sum of characteristic values ​​of the effective values ​​of each voltage and current within the preset range at the current moment, the initial leakage current confidence level is further corrected. If the range distribution characteristic value is too high, it indicates that the electrical quantity is approaching the range limit and the linearity of the sensing signal decreases. In this case, the confidence level is appropriately weakened by a correction coefficient, and vice versa. This strategy, which combines dynamic scale adjustment with range range correction, allows the leakage current determination criterion to adapt to changes in the actual operating state, reducing confidence level jumps caused by sensor operating point offsets and brief interferences, thereby outputting a more accurate leakage current level signal. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0014] Figure 1 This is a schematic diagram of a leakage current detection system for a distribution box based on multi-sensor data fusion. Figure 2 This is a flowchart of the difference measurement unit for calculating structural difference and wave morphology parameters; Figure 3 This is a flowchart of the frequency statistics for fluctuations and reversals; Figure 4 It is the time-domain variation curve of the equivalent impedance and harmonic current amplitude at the input and output terminals; Figure 5 It is the curve showing the difference in leakage current reference vector structure between the inlet and outlet terminals; Figure 6 It is a curve showing the change in structural difference adjustment and initial leakage confidence; Figure 7 This is a schematic diagram of the fluctuation reversal frequency of the structural difference sequence. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] See Figure 1This invention provides a leakage current detection system for distribution boxes based on multi-sensor data fusion, including a leakage current detection sensing unit, a leakage current feature analysis unit, a difference measurement unit, a leakage current status assessment unit, and a leakage current level determination unit. The leakage current detection sensing unit collects multi-dimensional leakage-related physical quantities at the inlet and outlet ends of the distribution box, and transmits the collected data to the leakage current feature analysis unit. The leakage current feature analysis unit constructs an inlet-end leakage current reference vector based on a first numerical relationship between the physical quantities at the inlet end at the same time, and constructs an outlet-end leakage current reference vector based on a second numerical relationship between the physical quantities at the outlet end at the same time. The difference measurement unit obtains the structural difference degree between the inlet-end and outlet-end leakage current reference vectors at the same time, and extracts the fluctuation morphology parameters of the change trajectory based on the change trajectory of the structural difference degree over multiple consecutive time periods. The leakage current status assessment unit adjusts the comparison scale of the structural difference degree between the inlet-end and outlet-end leakage current reference vectors based on the stability of the fluctuation morphology parameters within a preset time period before the current time, generating a preliminary leakage current confidence level for the current time. The leakage current level determination unit corrects the initial leakage current confidence level based on the distribution range of the multidimensional leakage current related physical quantities within the preset range at the current moment, obtains the final leakage current confidence level at the current moment, and outputs the corresponding leakage current level signal according to the final leakage current confidence level.

[0017] In practical implementation, the multidimensional leakage current related physical quantities include the RMS value of the input voltage, the RMS value of the input current, the harmonic distortion rate of the input current, the RMS value of the output voltage, the RMS value of the output current, and the harmonic distortion rate of the output current. The leakage current detection sensing unit includes at least a voltage transformer, a current transformer, and a harmonic analysis module. The voltage transformers are installed at both the input and output ends of the distribution box. The primary winding of the input voltage transformer is connected in parallel between the phase and neutral lines at the input end, and the secondary winding of the input voltage transformer outputs a low-voltage signal proportional to the RMS value of the input voltage. The primary winding of the output voltage transformer is connected in parallel between the phase and neutral lines at the output end, and the secondary winding of the output voltage transformer outputs a low-voltage signal proportional to the RMS value of the output voltage. The low-voltage signals output by the voltage transformers are processed by a signal conditioning circuit and an analog-to-digital converter to obtain the digital values ​​of the RMS values ​​of the input and output voltages.

[0018] Current transformers are installed at both the incoming and outgoing terminals. The primary winding of the incoming current transformer is connected in series with the incoming phase line, and its secondary winding outputs a low-current signal proportional to the effective value of the incoming current. Similarly, the primary winding of the outgoing current transformer is connected in series with the outgoing phase line, and its secondary winding outputs a low-current signal proportional to the effective value of the outgoing current. These low-current signals are then processed by a signal conditioning circuit and an analog-to-digital converter to obtain the digital values ​​of the effective values ​​of the incoming and outgoing currents.

[0019] The harmonic analysis module is electrically connected to both the incoming and outgoing current transformers. It receives the current signals output from both the incoming and outgoing current transformers, performs analog-to-digital conversion on the current signals, and then calculates the harmonic distortion rates of the incoming and outgoing currents using a Fast Fourier Transform (FFT). The module is implemented as follows: A FFT is performed on the discrete sampling sequence of the incoming current signal to obtain the fundamental amplitude and the amplitudes of each harmonic. The square root of the sum of the squares of each harmonic amplitude is calculated as the ratio of the fundamental amplitude to the input current harmonic distortion rate. Similarly, a FFT is performed on the discrete sampling sequence of the outgoing current signal to obtain the fundamental amplitude and the amplitudes of each harmonic. The square root of the sum of the squares of each harmonic amplitude is calculated as the ratio of the fundamental amplitude to the outgoing current harmonic distortion rate.

[0020] The leakage current detection sensing unit is internally equipped with a synchronous sampling controller. This controller provides a unified sampling clock signal to all voltage transformers, current transformers, and harmonic analysis modules, ensuring that the RMS values ​​of the input voltage, input current, input current harmonic distortion rate, output voltage, output current, and output current harmonic distortion rate are all obtained at the same sampling time. The leakage current detection sensing unit aligns all multidimensional leakage-related physical quantities collected at the same sampling time according to the sampling time, encapsulates them into a single data frame, and transmits the data frame to the leakage current feature analysis unit.

[0021] In specific implementation, the leakage current feature analysis unit receives the effective values ​​of the input terminal voltage, the effective value of the input terminal current, the harmonic distortion rate of the input terminal current, the effective value of the output terminal voltage, the effective value of the output terminal current, and the harmonic distortion rate of the output terminal current, all aligned at the same sampling time, transmitted by the leakage current detection sensing unit.

[0022] The process of constructing the input terminal leakage reference vector by the leakage current characteristic analysis unit is as follows: Extract the effective value of the input terminal voltage, the effective value of the input terminal current, and the harmonic distortion rate of the input terminal current at the same sampling time from the received data. Calculate the ratio of the effective value of the input terminal voltage to the effective value of the input terminal current to obtain the equivalent impedance of the input terminal. The effective value of the input terminal voltage is denoted as... The effective value of the incoming line current is denoted as Then the equivalent impedance at the input terminal Calculated using the following formula: in, This indicates the effective value of the input voltage, in volts. This indicates the effective value of the incoming line current, in amperes. This represents the equivalent impedance at the input terminal, in ohms. Under division-to-zero protection, when... When the value is zero, Set to a preset maximum value, the preset maximum value is The maximum preset value is set at ohms, which is more than two orders of magnitude greater than the maximum possible value of the equivalent impedance under normal operating conditions at the incoming terminal of the distribution box.

[0023] The product of the effective value of the input current and the harmonic distortion rate of the input current is taken as the amplitude of the input harmonic current. The harmonic distortion rate of the input current is stored as a percentage. Before calculating the product, the harmonic distortion rate of the input current is converted to a dimensionless decimal form, that is, the percentage value of the harmonic distortion rate of the input current is divided by... Then multiply it by the effective value of the incoming current.

[0024] The equivalent impedance at the input terminal and the amplitude of the harmonic current at the input terminal are combined according to a preset dimensional arrangement to form the input terminal leakage current reference vector. The preset dimensional arrangement is a two-dimensional vector, with the first dimension being the equivalent impedance at the input terminal and the second dimension being the amplitude of the harmonic current at the input terminal. The input terminal leakage current reference vector is represented as follows: ,in This indicates the amplitude of the harmonic current at the input terminal.

[0025] The process of constructing the leakage current reference vector at the output terminal by the leakage current characteristic analysis unit is as follows: Extract the effective value of the output terminal voltage, the effective value of the output terminal current, and the harmonic distortion rate of the output terminal current at the same sampling time from the received data. Calculate the ratio of the effective value of the output terminal voltage to the effective value of the output terminal current to obtain the equivalent impedance of the output terminal. The effective value of the output terminal voltage is denoted as... The effective value of the outgoing terminal current is denoted as Equivalent impedance at the output terminal pass and The ratio is obtained. Under division by zero protection, when When the value is zero, Set to the preset maximum value ohm.

[0026] The product of the effective value of the output current and the harmonic distortion rate of the output current is taken as the amplitude of the output harmonic current. The harmonic distortion rate of the output current is stored as a percentage. Before calculating the product, the harmonic distortion rate of the output current is converted to a dimensionless decimal form, that is, the percentage value of the harmonic distortion rate of the output current is divided by... Then multiply it by the effective value of the current at the output terminal.

[0027] The equivalent impedance at the outgoing terminal and the amplitude of the harmonic current at the outgoing terminal are combined in the same dimensional arrangement as the leakage reference vector at the incoming terminal to form the leakage reference vector at the outgoing terminal. The leakage reference vector at the outgoing terminal is represented as follows: ,in Indicates the equivalent impedance of the outgoing line. The output terminal harmonic current amplitude is represented by the first dimension of the output terminal leakage reference vector, which is the equivalent impedance of the output terminal, and the second dimension is the output terminal harmonic current amplitude.

[0028] In some embodiments, both the incoming-end leakage reference vector and the outgoing-end leakage reference vector contain two dimensions, and the arrangement of the two vector dimensions remains consistent at any given time. The leakage feature parsing unit sends the constructed incoming-end leakage reference vector and outgoing-end leakage reference vector to the difference measurement unit.

[0029] See Figure 4 In the graph, the horizontal axis represents time in seconds, ranging from 0 to 10 seconds. The left vertical axis represents the equivalent impedance of the input and output terminals in ohms (Ω), ranging from 21.0 to 23.5 Ω, corresponding to the orange curves (solid and dashed lines) in the graph. The right vertical axis represents the harmonic current amplitude of the input and output terminals in amperes (A), ranging from 0.22 to 0.34 A, corresponding to the purple curves (solid and dashed lines) in the graph.

[0030] The orange solid line in the diagram represents the equivalent impedance at the input terminal. The orange dashed line represents the equivalent impedance of the outgoing line. The solid purple line represents the amplitude of the harmonic current at the input terminal. The purple dashed line represents the amplitude of the harmonic current at the output terminal. The equivalent impedance is obtained by the ratio of the effective value of the voltage to the effective value of the current at the input or output terminal. The harmonic current amplitude is calculated by multiplying the effective value of the current at the corresponding terminal by the harmonic distortion rate of the current.

[0031] Within the time interval of 0 to 5 seconds, the equivalent impedance of the input and output terminals remained within the range of 21.0 to 22.5 Ω, with small fluctuations. Furthermore, the curves of both terminals showed relatively similar trends, indicating that the electrical parameters of the input and output terminals were in a relatively stable and consistent state. Correspondingly, the harmonic current amplitudes of the input and output terminals remained within the range of 0.24 to 0.30 A, with relatively gentle fluctuations.

[0032] From approximately 5 to 10 seconds, the equivalent impedance at the outgoing end (orange dashed line) gradually increases, with stronger fluctuations, reaching a maximum of nearly 23.5Ω. While the equivalent impedance at the incoming end (orange solid line) remains at a relatively low level overall, it shows a slight increase towards the end, highlighting the difference between the two. The harmonic current amplitude at the incoming end (purple solid line) exhibits a certain decreasing trend during this stage, with reduced fluctuations, while the harmonic current amplitude at the outgoing end (purple dashed line) fluctuates between 0.28 and 0.34 A, with an overall value slightly higher than that at the incoming end.

[0033] In specific implementation, please refer to Figure 2The difference measurement unit receives the incoming terminal leakage reference vector and the outgoing terminal leakage reference vector sent by the leakage current characteristic analysis unit at the same time. The incoming terminal leakage reference vector is composed of the incoming terminal equivalent impedance and the incoming terminal harmonic current amplitude in sequence, and the outgoing terminal leakage reference vector is composed of the outgoing terminal equivalent impedance and the outgoing terminal harmonic current amplitude in sequence.

[0034] The process by which the difference measurement unit obtains the structural difference between the incoming and outgoing leakage reference vectors at the same moment is as follows: First, calculate the first sum of the values ​​of each dimension in the incoming leakage reference vector, which is the sum of the equivalent impedance at the incoming end and the amplitude of the harmonic current at the incoming end. Second, calculate the second sum of the values ​​of each dimension in the outgoing leakage reference vector, which is the sum of the equivalent impedance at the outgoing end and the amplitude of the harmonic current at the outgoing end. Finally, calculate the absolute value of the difference between the first and second sums to obtain the total difference value.

[0035] Calculate the first difference sequence between adjacent dimensions in the incoming line leakage reference vector. The incoming line leakage reference vector contains two dimensions, and the difference between adjacent dimensions is only one item, namely the difference between the equivalent impedance of the incoming line and the amplitude of the harmonic current at the incoming line. The first difference sequence consists of this single difference. Calculate the second difference sequence between adjacent dimensions in the outgoing line leakage reference vector. The outgoing line leakage reference vector also contains two dimensions, and the second difference sequence consists of the difference between the equivalent impedance of the outgoing line and the amplitude of the harmonic current at the outgoing line. Obtain the sum of the absolute values ​​of the differences at corresponding positions between the first and second difference sequences to obtain the distribution difference value. Since each of the first and second difference sequences contains only one element, take the absolute value of the difference between the differences in the first and second difference sequences; this absolute value is the distribution difference value.

[0036] The specific process for the difference measurement unit to extract the fluctuation morphology parameters of the change trajectory is as follows: The structural differences at multiple consecutive time points are arranged in ascending order of time to form a difference change sequence. The number of consecutive time points is preset to a fixed value. , The range of values ​​is to In one implementation, the integers between [a certain value]. Values The setting is based on: when the sampling period is seconds, Each time corresponds to The time window is set to a number of seconds, which is long enough to cover several complete cycles of typical fluctuations during the development of a leakage fault in the distribution box, while avoiding the introduction of non-fault interference by an excessively long window.

[0037] The difference between structural differences at adjacent time points in the difference change sequence is calculated to obtain the difference difference sequence. The difference change sequence contains... The structural difference values, denoted in chronological order, are: The difference sequence contains... The difference value, the first The difference values ​​are determined by The calculation yielded, where Pick to Integers.

[0038] The frequency of fluctuation reversals is obtained by counting the number of times the sign of the numerical values ​​in the difference series changes. The frequency of fluctuation reversals is the number of times the product of two adjacent difference values ​​in the difference series is negative. This is done by traversing the difference series and, for the first... The difference value and the first If the product of the two differences is less than zero, then the frequency of fluctuation reversal is accumulated. ,in Pick to Integers.

[0039] Calculate the average of all positive values ​​and the average of the absolute values ​​of all negative values ​​in the difference sequence. Extract all difference values ​​greater than zero from the difference sequence, calculate the average of the positive values, and denot it as . Extract all difference values ​​less than zero from the difference sequence, calculate the absolute value of the negative values, and then average them. This average is denoted as... Average of positive values The average of the absolute values ​​of negative values The ratio of these values ​​is used to represent the degree of equilibrium in fluctuation amplitude, denoted as . , .when When the value is zero, the fluctuation amplitude equalization is... Set to the preset upper limit value, the preset upper limit value is The setting is based on the following: In leakage current detection scenarios, if no negative differential value appears, it indicates that the fluctuation shows a unidirectional increasing or decreasing trend, and the fluctuation amplitude balance tends to be extremely large. A preset upper limit value is set to represent the extreme imbalance state.

[0040] The frequency of fluctuation reversal and the uniformity of fluctuation amplitude are combined to form a fluctuation pattern parameter. This parameter is represented by a two-dimensional parameter set, where the first parameter is the frequency of fluctuation reversal and the second is the uniformity of fluctuation amplitude. The difference measurement unit sends the fluctuation pattern parameter to the leakage current status assessment unit.

[0041] See Figure 5In the graph, the horizontal axis represents time in seconds, ranging from 0 to 10 seconds; the vertical axis represents the dimensionless normalized value of the difference. The three curves in the graph represent the total difference (green solid line), the distribution difference (orange dashed line), and the structural difference (purple solid line), respectively.

[0042] As shown in the figure, both the total difference and the distribution difference fluctuate over time, with the total difference generally exhibiting a larger amplitude than the distribution difference. The structural difference curve closely follows the total difference curve, indicating that the total difference accounts for a larger proportion in the calculation of structural difference. The structural difference is obtained by weighting the total difference and the distribution difference according to preset weights, and the fluctuation trend of the structural difference in the figure takes into account the changing characteristics of both. Within the time interval of 0 to 10 seconds, the structural difference curve shows obvious fluctuations, with alternating peaks and troughs, reflecting the dynamic changes in the structural difference between the leakage reference vector at the incoming end and the leakage reference vector at the outgoing end over time. The fluctuation amplitude can reach a maximum of about 1.5 and a minimum close to zero, indicating that the difference varies over a wide range.

[0043] In practical implementation, the leakage current status assessment unit receives the fluctuation morphology parameters and structural difference degree sent by the difference measurement unit. The fluctuation morphology parameters include the fluctuation turning frequency and the fluctuation amplitude uniformity, while the structural difference degree is the structural difference between the leakage current reference vector at the incoming end and the leakage current reference vector at the outgoing end at the same time.

[0044] The process by which the leakage current status assessment unit generates the preliminary leakage current confidence level for the current moment is as follows: It obtains the fluctuation turning frequency corresponding to each moment within a preset time period prior to the current moment. The preset time period is set to a fixed duration, and the preset time period value is... Second, The range of values ​​is Instant Seconds. In one implementation, Values Seconds. The setting is based on: when the sampling period is... seconds, The second time window contains At each sampling time point, the time series sample size for the frequency of fluctuation reversals is sufficient to ensure statistical stability in the calculation of the coefficient of variation. The second window length can capture the transition characteristics of the initial fluctuation pattern of leakage current in the distribution box from stable to unstable.

[0045] The coefficient of variation of the frequency of wave reversals is calculated to obtain a value characterizing the stability. The coefficient of variation of the frequency of wave reversals is denoted as... The calculation formula is: in, The coefficient of variation represents the frequency of fluctuation reversals; This represents the standard deviation of the frequency values ​​of all fluctuations within a preset time period. This represents the average value of all fluctuation turning frequencies within a preset time period. When the value is zero, Set to the preset maximum upper limit value, the preset maximum upper limit value is The basis for this setting is: if no change in the fluctuation is detected in the initial stage of leakage current, and the average frequency of the fluctuation change is zero, it indicates that the system state is extremely stable, and the coefficient of variation should take a maximum value to characterize this extreme stable state.

[0046] The relaxation coefficient is matched to the numerical range in which the stability characterization value lies. Multiple pre-defined tables map the correspondence between numerical ranges and relaxation coefficients. The numerical ranges are divided as follows: Less than The first relaxation coefficient is matched to the first numerical interval. , Values ; Greater than or equal to and less than The second relaxation coefficient corresponds to the second numerical interval. , Values ; Greater than or equal to and less than The third numerical interval corresponds to the third relaxation coefficient. , Values ; Greater than or equal to The fourth numerical interval corresponds to the fourth relaxation coefficient. , Values The relaxation coefficient is chosen based on the following: when the stability indicator value is small, it indicates a low degree of variation in the frequency of fluctuation reversals and a stable system fluctuation pattern. Therefore, the comparison scale should be reduced to decrease sensitivity and avoid false alarms caused by minor disturbances. Thus, the relaxation coefficient is set to less than [value missing]. The value of the relaxation coefficient is considered. When the stability characteristic value is large, it indicates a high degree of variation in the frequency of fluctuation reversals and an unstable system fluctuation pattern. Therefore, the comparison scale should be increased to improve sensitivity and promptly detect potential leakage anomalies. Thus, the relaxation coefficient should be greater than or equal to [value missing]. The value.

[0047] Multiplying the structural dissimilarity by the matched relaxation coefficient yields the adjusted structural dissimilarity. The structural dissimilarity is denoted as... The relaxation coefficient is denoted as The adjusted structural difference is denoted as , The adjusted structural difference is mapped to a preset confidence interval to obtain the initial leakage current confidence level. The preset confidence interval is... The mapping method uses linear mapping, and the adjusted structural difference degree The calculation method for mapping to confidence intervals is as follows: preliminary leakage current confidence level. ,in The adjusted structural variability is the minimum value recorded in historical operations. This represents the maximum value of the adjusted structural difference recorded during historical operation. When... and When they are equal, Values The rationale is as follows: structural dissimilarity has not shown a range of variation in historical records, making it impossible to determine the tendency through linear mapping; therefore, the median of the confidence interval is used to represent the neutral confidence state. When the calculated... Less than At that time, Cut off as ;when Greater than At that time, Cut off as .

[0048] In some embodiments, the leakage current level determination unit receives the preliminary leakage current confidence level generated by the leakage current status assessment unit, and simultaneously receives the current effective values ​​of the incoming terminal voltage, incoming terminal current, outgoing terminal voltage, and outgoing terminal current collected by the leakage current detection sensing unit. The process by which the leakage current level determination unit obtains the final leakage current confidence level at the current moment involves determining the interval level to which each of the current effective values ​​of the incoming terminal voltage, incoming terminal current, outgoing terminal voltage, and outgoing terminal current belongs within a preset range. The preset range is set according to the rated electrical parameters of the distribution box; the preset range for the effective value of the incoming terminal voltage is [amount missing] of the rated voltage. Doubled The preset range of the effective value of the incoming line current is times the rated current. Doubled The preset range of the effective value of the output terminal voltage is times the rated voltage. Doubled The preset range of the effective value of the output terminal current is times the rated current. Doubled The preset range of each physical quantity is divided into three equal intervals: the first interval corresponds to 1 / 3 of the preset range. to Range, the second interval level corresponds to the preset range to The range, the third interval level corresponds to the preset range. to Range. The interval level number into which the current value of each physical quantity falls is the interval level to which that physical quantity belongs. The interval level number is determined by the numerical value. Numerical values Numerical values .

[0049] The sum of all interval levels is used to obtain the characteristic value of the range distribution. The calculation method is as follows: ,in This indicates the range level of the effective value of the incoming line voltage. This indicates the range level of the effective value of the incoming line current. Indicates the range level of the effective value of the output terminal voltage. This indicates the range level of the effective value of the output terminal current. The range of values ​​is to Integers between [a certain range].

[0050] The range distribution characteristic value is compared with a preset correction trigger threshold. The preset correction trigger threshold is set to... The basis for this setting is: when the sum of the interval levels of the four physical quantities equals When, the corresponding average interval level is This indicates that the overall physical quantities are at a moderate level within the preset range, forming a reference point separating normal and abnormal operating conditions. When the range distribution characteristic value exceeds the preset correction trigger threshold, it indicates that the electrical physical quantities at the incoming and outgoing ends are generally biased towards a higher range, possibly accompanied by an increasing leakage current trend due to increased load or insulation degradation. The initial leakage current confidence level should be multiplied by a factor greater than [value missing]. The first correction factor, the first correction factor is taken as: When the range distribution characteristic value is less than or equal to the correction trigger threshold, it indicates that the overall electrical physical quantities at the incoming and outgoing terminals are in the low to medium range, and the leakage risk is relatively low. The initial leakage confidence level is then multiplied by a factor less than [value missing]. The second correction factor, the second correction factor is taken as... First correction factor Second correction coefficient The setting is based on: through The correction range is designed to maintain a significant correction effect while avoiding drastic jumps in confidence level due to a single correction, ensuring a smooth transition in the final leakage current confidence level when the characteristic value of the range distribution crosses the correction trigger threshold. The value obtained after correction calculation is the final leakage current confidence level.

[0051] The process by which the leakage current level determination unit matches the corresponding leakage current level signal output based on the final leakage current confidence level is as follows: The final leakage current confidence level is compared with preset first-level and second-level thresholds. The first-level threshold is preset to... The second-level threshold is preset to The first-level threshold is greater than the second-level threshold. (First-level threshold) The setting is based on: when the final leakage current confidence level reaches At that time, the confidence level has exceeded the median value of the confidence level interval. Furthermore, it exhibits a significant positive bias, indicating that the possibility of leakage current is dominant. Second-level threshold. The setting is based on the following: when the final leakage current confidence level is lower than At that time, the confidence level was already lower than the median of the confidence interval. Furthermore, it exhibits a significant negative deviation, indicating a low probability of leakage.

[0052] When the final leakage confidence level is greater than or equal to the first-level threshold When the leakage current level is less than the first-level threshold, a first-level leakage current risk signal is output, triggering the highest-level alarm response. And greater than or equal to the second-level threshold When the leakage current level is less than the second-level threshold, a second leakage current level signal is output, corresponding to a medium leakage current risk level, triggering a warning response. At this time, a third leakage level signal is output, corresponding to a low leakage risk level, maintaining normal operation status indication. The output leakage level signal is sent to the distribution box monitoring terminal or the host computer system.

[0053] See Figure 6 In the graph, the horizontal axis represents time (in seconds), ranging from 0 to 10 seconds. The left vertical axis represents the structural variability (dimensionless), ranging from 0 to 1.5, and the right vertical axis represents the initial leakage confidence level (%), ranging from 0 to 100%. In the legend, the dashed line represents the structural variability, the solid blue line represents the adjusted structural variability, and the solid red line represents the initial leakage confidence level.

[0054] The structural difference curve exhibits a relatively volatile trend, with its peak value mainly concentrated between 0 and 1.5. The frequent and significant fluctuations reflect the dynamic changes in the structural difference between the incoming and outgoing leakage reference vectors over continuous time. Compared to the original structural difference curve, the adjusted curve shows a reduced overall fluctuation amplitude and a smoother curve, indicating that multiplying by a preset relaxation coefficient effectively adjusts the comparison scale of the structural difference to match the stability of the system's fluctuating state.

[0055] The initial leakage confidence curve generally follows the trend of the adjusted structural difference, but its value range extends to 0% to 100%. A linear mapping is used to convert the adjusted structural difference into leakage confidence, reflecting the dynamic assessment of leakage risk at different time points. Multiple peaks in the initial leakage confidence correspond to high values ​​in the structural difference, indicating significant leakage characteristics and a high system risk at these times. Conversely, troughs in the confidence correspond to lower structural difference, showing a relatively stable system state.

[0056] In specific implementation, please refer to Figure 3 The frequency of fluctuation reversals is the number of times the product of two adjacent difference values ​​in the difference degree sequence is negative. The difference degree sequence is generated by the difference measurement unit during the extraction of fluctuation morphology parameters of the change trajectory. The difference degree sequence is composed of the differences between adjacent structural difference degrees in multiple consecutive time periods arranged sequentially.

[0057] The difference sequence contains the following number of difference values: ,in Indicates the number of consecutive moments. Values The difference values ​​in the difference series are arranged in ascending order of time, the first... The difference value is denoted as , No. The difference value is denoted as ,in Pick to Integers.

[0058] The statistical process for determining the frequency of fluctuation reversals is as follows: Initialize a frequency counter, setting its initial value to zero. Starting from the first difference value in the difference sequence, iterate sequentially to the second-to-last difference value. In each iteration, obtain the difference value at the current iteration position. The difference between the next and the next adjacent position ,calculate and The product of the two numbers. Determine the sign of the product; if the product is less than zero, increment the frequency counter. If the product is greater than or equal to zero, the count value of the frequency counter remains unchanged.

[0059] The condition for a product to be less than zero is: and Opposite signs indicate that one difference value is positive and the other is negative. A positive difference value indicates that the structural dissimilarity increases from the current time to the next time step, while a negative difference value indicates that the structural dissimilarity decreases from the current time to the next time step. Opposite signs between two adjacent difference values ​​indicate a reversal in the direction of change in structural dissimilarity, i.e., from an increasing trend to a decreasing trend, or vice versa. After traversing the entire dissimilarity difference sequence, the final count value of the frequency counter is the fluctuation reversal frequency. The fluctuation reversal frequency is recorded in integer form.

[0060] In some embodiments, the difference sequence may contain difference values ​​that are zero. When or When any one of them is zero, the product is zero, which does not satisfy the condition that the product is less than zero, and the frequency counter does not increment. The difference value of zero indicates that the structural difference has not changed in adjacent time moments, and the trend of change has not reversed. The fluctuation reversal frequency, as a component of the fluctuation morphology parameter, is combined with the fluctuation amplitude balance and sent by the difference measurement unit to the leakage current state assessment unit for subsequent calculation of stability characterization value and relaxation coefficient matching.

[0061] See Figure 7 In the figure, the horizontal axis represents the index of the difference sequence, ranging from 0 to 500, indicating the temporal position of each difference value in the difference sequence; the vertical axis represents the difference value, in dimensionless units, with a value range of approximately -1.5 to 1.5. The gray curve represents the real-time trend of the difference sequence, reflecting the change in the difference of structural difference at continuous sampling times. The red scatter dots mark the turning points, specifically corresponding to the turning moments in the fluctuation turning frequency statistics, that is, the moments when the product of two adjacent difference values ​​is less than zero, indicating that the trend of structural difference has changed from increasing to decreasing or from decreasing to increasing.

[0062] This figure shows that the structural difference values ​​fluctuate frequently across 500 consecutive difference value samples, with alternating positive and negative values, a large number of turning points, and evenly distributed red dots. This indicates that the fluctuation trend in the structural difference change curve reverses multiple times, reflecting the high volatility and dynamic characteristics of structural difference during leakage current detection. The difference values ​​are mostly concentrated within the range of ±1, and the absolute value of the difference value corresponding to most turning points does not exceed 1, showing that the variation range of structural difference is at a moderate level.

[0063] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A leakage current detection system for distribution boxes based on multi-sensor data fusion, characterized in that, include: The leakage current detection sensing unit collects multi-dimensional leakage current related physical quantities at the inlet and outlet ends of the distribution box and transmits them to the leakage current feature analysis unit. The leakage current feature analysis unit constructs a leakage current reference vector at the input end based on the first numerical relationship between the physical quantities at the input end at the same time, and constructs a leakage current reference vector at the output end based on the second numerical relationship between the physical quantities at the output end at the same time. The difference measurement unit obtains the structural difference degree between the leakage reference vector at the input end and the leakage reference vector at the output end at the same time, and extracts the fluctuation morphology parameters of the change trajectory based on the change trajectory of the structural difference degree at multiple consecutive times. The leakage current status assessment unit adjusts the comparison scale of the structural difference between the leakage current reference vector at the input end and the leakage current reference vector at the output end based on the stability of the fluctuation morphology parameters within a preset time period before the current moment, and generates the preliminary leakage current confidence level at the current moment. The leakage current level determination unit corrects the initial leakage current confidence level based on the distribution range of the multidimensional leakage current related physical quantities within the preset range at the current moment, obtains the final leakage current confidence level at the current moment, and outputs the corresponding leakage current level signal based on the final leakage current confidence level.

2. The leakage current detection system for distribution boxes based on multi-sensor data fusion according to claim 1, characterized in that, The multidimensional leakage current related physical quantities include the effective value of the input voltage, the effective value of the input current, the harmonic distortion rate of the input current, the effective value of the output voltage, the effective value of the output current, and the harmonic distortion rate of the output current. The leakage current detection sensing unit includes at least a voltage transformer, a current transformer, and a harmonic analysis module. The voltage transformer is installed at the input terminal and the output terminal, respectively. The current transformer is installed at the input terminal and the output terminal, respectively. The harmonic analysis module is electrically connected to the current transformer at the input terminal and the current transformer at the output terminal, respectively.

3. The power distribution box leakage current detection system based on multi-sensor data fusion according to claim 2, characterized in that, The specific process for constructing the incoming line leakage reference vector in the leakage feature analysis unit is as follows: After aligning the effective values ​​of the input terminal voltage, the effective value of the input terminal current, and the harmonic distortion rate of the input terminal current obtained at the same sampling time according to the sampling time, the ratio of the effective value of the input terminal voltage to the effective value of the input terminal current is calculated to obtain the equivalent impedance of the input terminal. The product of the effective value of the input terminal current and the harmonic distortion rate of the input terminal current is taken as the amplitude of the input terminal harmonic current. The equivalent impedance of the input terminal and the amplitude of the harmonic current of the input terminal are arranged in a preset dimensional manner and combined to form the leakage reference vector of the input terminal.

4. The leakage current detection system for distribution boxes based on multi-sensor data fusion according to claim 3, characterized in that, The specific process for constructing the leakage reference vector at the outgoing terminal in the leakage feature analysis unit is as follows: After aligning the effective values ​​of the output terminal voltage, the effective values ​​of the output terminal current, and the harmonic distortion rate of the output terminal current obtained at the same sampling time according to the sampling time, the ratio of the effective value of the output terminal voltage to the effective value of the output terminal current is calculated to obtain the equivalent impedance of the output terminal. The product of the effective value of the output terminal current and the harmonic distortion rate of the output terminal current is taken as the amplitude of the output terminal harmonic current. The equivalent impedance of the outgoing terminal and the amplitude of the harmonic current of the outgoing terminal are arranged in the same dimension as the leakage reference vector of the incoming terminal, and combined to form the leakage reference vector of the outgoing terminal.

5. The power distribution box leakage detection system based on multi-sensor data fusion according to claim 4, characterized in that, The specific process by which the difference measurement unit obtains the structural difference between the inlet leakage reference vector and the outlet leakage reference vector at the same time is as follows: The absolute value of the difference between the first sum of the values ​​of each dimension in the leakage reference vector at the input end and the second sum of the values ​​of each dimension in the leakage reference vector at the output end is calculated to obtain the total difference value. Calculate the first difference sequence between adjacent dimension values ​​in the leakage reference vector at the incoming end, and calculate the second difference sequence between adjacent dimension values ​​in the leakage reference vector at the outgoing end. Obtain the sum of the absolute values ​​of the differences at corresponding positions between the first difference sequence and the second difference sequence to obtain the distribution difference value. The structural difference degree is obtained by summing the total difference value and the distribution difference value according to a preset weight.

6. The power distribution box leakage current detection system based on multi-sensor data fusion according to claim 5, characterized in that, The specific process by which the difference measurement unit extracts the fluctuation morphology parameters of the change trajectory is as follows: The structural differences at multiple consecutive time points are arranged in ascending order of time to form a difference change sequence, and the difference between the structural differences at adjacent time points in the difference change sequence is calculated to obtain a difference difference sequence. The frequency of fluctuation reversal is obtained by counting the number of times the numerical sign changes in the difference sequence. The average value of all positive values ​​and the average value of all negative values ​​in the difference sequence are calculated. The ratio of the average value of the positive values ​​to the average value of the absolute values ​​of the negative values ​​is used as the fluctuation amplitude balance. The frequency of the fluctuation turning point and the uniformity of the fluctuation amplitude are combined to form the fluctuation morphology parameter.

7. The power distribution box leakage current detection system based on multi-sensor data fusion according to claim 6, characterized in that, The specific process by which the leakage current status assessment unit generates the preliminary leakage current confidence level at the current moment is as follows: The fluctuation turning frequency corresponding to each time within a preset time period before the current time is obtained, the coefficient of variation of the fluctuation turning frequency is calculated to obtain the stability characterization value, and the corresponding relaxation coefficient is matched according to the numerical range of the stability characterization value. The structural difference is multiplied by the relaxation coefficient to obtain the adjusted structural difference, and the adjusted structural difference is mapped to a preset confidence interval to obtain the preliminary leakage confidence.

8. The power distribution box leakage detection system based on multi-sensor data fusion according to claim 7, characterized in that, The specific process by which the leakage current level determination unit obtains the final leakage current confidence level at the current moment is as follows: Determine the interval level of the current effective value of the input terminal voltage, the current effective value of the input terminal current, the current effective value of the output terminal voltage, and the current effective value of the output terminal current within the preset range, and calculate the sum of all interval levels to obtain the range distribution characteristic value. When the range distribution characteristic value is greater than the preset correction trigger threshold, the initial leakage confidence is multiplied by a first correction coefficient greater than 1. When the range distribution characteristic value is less than or equal to the correction trigger threshold, the initial leakage confidence is multiplied by a second correction coefficient less than 1 to obtain the final leakage confidence.

9. The power distribution box leakage current detection system based on multi-sensor data fusion according to claim 8, characterized in that, The specific process by which the leakage current level determination unit matches the corresponding leakage current level signal output based on the final leakage current confidence level is as follows: The final leakage confidence level is compared with a preset first-level threshold and a second-level threshold, wherein the first-level threshold is greater than the second-level threshold. When the final leakage confidence level is greater than or equal to the first level threshold, a first leakage level signal is output; when the final leakage confidence level is less than the first level threshold but greater than or equal to the second level threshold, a second leakage level signal is output; when the final leakage confidence level is less than the second level threshold, a third leakage level signal is output.

10. The power distribution box leakage current detection system based on multi-sensor data fusion according to claim 6, characterized in that, The fluctuation reversal frequency is the number of times the product of two adjacent difference values ​​in the difference sequence is negative.

Images