Low-voltage switch cabinet intelligent monitoring system and method based on internet of things

Abstract

The application discloses a low-voltage switch cabinet intelligent monitoring system and method based on the Internet of Things, and particularly relates to the technical field of power equipment state monitoring, and is used for solving the problem that the arc electromagnetic pulse causes the cumulative damage of the sensor and triggers the distortion of the monitoring data in the prior art; the current voltage waveform is collected and the arc event is identified; the electromagnetic coupling weight coefficient is calculated according to the arc position and the sensor coordinate; the working parameters of the sensor are captured and compared with the historical reference value, and the key parameter offset is calculated; the information entropy value mutation of the offset of the same type sensor group is analyzed, and the nonlinear damage mark is triggered; the segmented mapping relationship is established based on the offset and the damage mark, and the cumulative damage evaluation value is output; the monitoring data is corrected according to the evaluation value, and the fault diagnosis model is input to execute the early warning; the quantitative evaluation of the sensor hidden damage and the data self-correction are realized, and the accuracy of the fault early warning is guaranteed from the source.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment state monitoring, and more particularly to an intelligent monitoring system and method for a low-voltage switch cabinet based on the Internet of Things. BACKGROUND

[0002] As a key device of a power distribution system, the running state of a low-voltage switch cabinet directly affects power supply reliability. With the popularization of the Internet of Things technology, remote intelligent monitoring through the deployment of multiple types of sensors (such as temperature, current, and partial discharge monitoring devices) inside the switch cabinet has become the mainstream solution in the industry. The existing technology usually transmits sensor data to a cloud platform using wired or general Internet of Things communication protocols (such as Modbus and MQTT), and realizes fault early warning by combining data analysis algorithms. Especially for transient high-risk events such as arc faults, the system relies on high-precision sensors to capture transient characteristics to trigger the protection mechanism. The current technical solution focuses on optimizing the arc detection algorithm, improving the sampling frequency and communication real-time performance to ensure a quick response to switch cabinet body faults.

[0003] However, the existing monitoring solution ignores the defect that the strong electromagnetic pulse (EMP) generated by the arc fault will invade the internal circuit of the sensor through conduction or radiation coupling, causing irreversible cumulative damage to microelectronic devices. This damage does not cause device failure at the initial stage, but causes slow drift in sensor measurement parameters (such as temperature readings and current waveforms), and conventional calibration cannot identify this hidden degradation. The direct consequence is that the fault diagnosis and early warning of the monitoring system based on distorted data will gradually deviate from the real working condition, eventually leading to false negatives or false positives, and cannot eliminate the monitoring device failure risk caused by the monitored object (arc fault) at the data source. SUMMARY

[0004] To overcome the above-mentioned defects of the prior art, the present application provides an intelligent monitoring system and method for a low-voltage switch cabinet based on the Internet of Things to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution:

[0006] The intelligent monitoring method for a low-voltage switch cabinet based on the Internet of Things comprises:

[0007] S1, collecting current and voltage waveform data, and determining that an arc event has occurred when the current mutation rate exceeds a preset mutation threshold and is accompanied by high-frequency oscillation characteristics;

[0008] S2, extracting the skin effect characteristics of the cabinet body and the parasitic capacitance of the sensor power supply loop according to the arc occurrence position and the sensor installation coordinates, and calculating the coupling weight coefficient of the electromagnetic pulse on each sensor;

[0009] S3, capture the working parameters of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and compare the working parameters with the historical reference parameters of the corresponding sensor to calculate the key parameter offset;

[0010] S4, calculate the information entropy value of the key parameter offset of the same type sensor group, and when the mutation amplitude of the information entropy value exceeds the preset entropy change threshold, trigger the nonlinear damage marker;

[0011] S5, establish a segmented mapping relationship based on the key parameter offset and the nonlinear damage marker, and output the cumulative damage evaluation value of the sensor group;

[0012] S6, correct the sensor monitoring data according to the cumulative damage evaluation value and input the fault diagnosis model to perform early warning.

[0013] Further, collect current and voltage waveform data, when the current mutation rate exceeds the preset mutation threshold and is accompanied by high-frequency oscillation characteristics, determine that an arc event occurs, including:

[0014] Based on the current waveform data, calculate the current differential sequence, and identify the interval in the current differential sequence that continuously exceeds the preset mutation threshold as the arc event determination interval;

[0015] Analyze the frequency spectrum characteristics of the voltage waveform in the arc event determination interval to confirm the existence of high-frequency oscillation components that meet the arc characteristics;

[0016] Determine the spatial coordinates of the arc occurrence position according to the time domain center position of the arc event determination interval.

[0017] Further, according to the arc occurrence position and the sensor installation coordinates, extract the skin effect characteristics of the cabinet and the parasitic capacitance of the sensor power supply loop, calculate the coupling weight coefficient of electromagnetic pulse on each sensor, including:

[0018] Based on the spatial coordinates of the arc occurrence position and the spatial distance vector of each sensor installation coordinate, calculate the skin depth distribution characteristic value of the electromagnetic wave propagation path on the surface of the cabinet;

[0019] Measure the parasitic capacitance parameters of each sensor power supply loop through an impedance analyzer;

[0020] Input the skin depth distribution characteristic value and the parasitic capacitance parameter of each sensor into the preset electromagnetic coupling equation to output the coupling weight coefficient of each sensor.

[0021] Further, capture the working parameters of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and compare the working parameters with the historical reference parameters of the corresponding sensor to calculate the key parameter offset, including:

[0022] Screening sensors with coupling weight coefficients exceeding a preset weight threshold as target sensors;

[0023] Synchronously recording real-time values of working parameters of the target sensors when an arc event occurs;

[0024] Retrieving corresponding parameter reference values in a historical reference parameter database of the target sensors;

[0025] Calculating a relative offset between the real-time value of the working parameter and the parameter reference value as a key parameter offset.

[0026] Further, calculating an information entropy value of the key parameter offsets of the group of sensors of the same type, and when a mutation amplitude of the information entropy value exceeds a preset entropy variation threshold, triggering a nonlinear damage marker, including:

[0027] Normalizing the key parameter offsets of each sensor in the group of sensors of the same type to generate a normalized offset probability distribution;

[0028] Calculating a Shannon information entropy value based on the normalized offset probability distribution as an information entropy value at the current time;

[0029] Obtaining an information entropy value of the group at the previous time, and calculating an absolute difference between the information entropy value at the current time and the information entropy value at the previous time as an entropy value mutation amplitude;

[0030] When the entropy value mutation amplitude exceeds the preset entropy variation threshold, generating a nonlinear damage marker triggering signal.

[0031] Further, establishing a segmented mapping relationship based on the key parameter offset and the nonlinear damage marker, and outputting a cumulative damage evaluation value of the group of sensors, including:

[0032] Selecting a linear mapping mode or a nonlinear mapping mode according to a state of the nonlinear damage marker triggering signal;

[0033] In the linear mapping mode, performing a weighted summation operation on the key parameter offsets of the group of sensors of the same type to obtain the cumulative damage evaluation value;

[0034] In the nonlinear mapping mode, extracting a maximum key parameter offset in the group of sensors of the same type to perform an exponential function transformation to obtain the cumulative damage evaluation value.

[0035] Further, correcting sensor monitoring data according to the cumulative damage evaluation value and inputting a fault diagnosis model to perform early warning, including:

[0036] Obtaining the cumulative damage evaluation value of the group of sensors and original monitoring data of the corresponding group of sensors;

[0037] Converting the cumulative damage evaluation value into a data correction coefficient through a material damage transfer function;

[0038] multiplying the data correction coefficient with the original monitoring data to generate corrected monitoring data;

[0039] inputting the corrected monitoring data into the pre-trained fault diagnosis model to generate a warning signal.

[0040] In another aspect, the present application provides an Internet of Things-based low-voltage switch cabinet intelligent monitoring system, comprising:

[0041] An arc detection module is configured to collect current and voltage waveform data, and determine that an arc event occurs when the current mutation rate exceeds a preset mutation threshold and is accompanied by high-frequency oscillation characteristics.

[0042] A coupling weight module is configured to extract the skin effect characteristics of the cabinet body and the parasitic capacitance of the sensor power supply loop according to the arc occurrence position and the sensor installation coordinates, and calculate the coupling weight coefficient of the electromagnetic pulse on each sensor.

[0043] An offset calculation module is configured to capture the working parameters of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and compare the working parameters with the historical reference parameters of the corresponding sensor to calculate the key parameter offset.

[0044] An entropy change monitoring module is configured to calculate the information entropy value of the key parameter offset of the same type of sensor group, and trigger a nonlinear damage marker when the information entropy value mutation amplitude exceeds a preset entropy change threshold.

[0045] A segmented mapping module is configured to establish a segmented mapping relationship based on the key parameter offset and the nonlinear damage marker, and output the cumulative damage evaluation value of the sensor group.

[0046] A warning correction module is configured to correct the sensor monitoring data according to the cumulative damage evaluation value and input the fault diagnosis model to perform warning.

[0047] Compared with the prior art, the present application has the following beneficial effects:

[0048] 1. Through electromagnetic damage perception and data self-correction mechanism, the monitoring distortion problem caused by sensor implicit degradation is solved; a dynamic correlation model of arc event and sensor health status is established: the electromagnetic coupling weight is calculated based on the physical space characteristics, the interference intensity of electromagnetic pulse on each sensor is accurately quantified, and the limitation of traditional scheme ignoring the electromagnetic propagation characteristics of device structure is broken through; relative offset analysis is adopted, the dynamic deviation of real-time working parameters and long-term reference values is compared, and the parameter drift of sensor caused by cumulative damage is effectively identified; an information entropy change detection mechanism is introduced, the nonlinear damage critical point is captured by using the distribution characteristic mutation of sensor group offset, and this characteristic has unique sensitivity to early implicit degradation.

[0049] 2. Real-time evaluation of the health status of the monitoring device is realized, the sensor is upgraded from a simple data acquisition unit to an intelligent node with self-diagnosis function, and the risk of misjudgment caused by failure of the monitoring device is eliminated from the source; the accumulated damage amount is converted into a data correction coefficient through a segmented mapping relationship, and the monitoring data input into the fault diagnosis model is dynamically corrected to ensure that the early warning decision is always based on the real working condition; under the premise of keeping the existing Internet of Things architecture unchanged, the long-term reliability of the system in a strong electromagnetic environment is significantly improved, and it is especially suitable for severe working conditions with high-frequency arc. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 A flowchart of the intelligent monitoring method for the low-voltage switch cabinet based on the Internet of Things is provided.

[0051] Figure 2 A structural schematic diagram of the intelligent monitoring system for the low-voltage switch cabinet based on the Internet of Things is provided. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0053] Embodiment 1: Figure 1 The intelligent monitoring method for the low-voltage switch cabinet based on the Internet of Things is provided, which comprises:

[0054] S1, current and voltage waveform data are collected, and when the current mutation rate exceeds the preset mutation threshold and is accompanied by high-frequency oscillation characteristics, it is determined that an arc event occurs;

[0055] S2, according to the arc occurrence position and the sensor installation coordinates, the skin effect characteristics of the cabinet body and the parasitic capacitance of the sensor power supply loop are extracted, and the coupling weight coefficient of the electromagnetic pulse to each sensor is calculated;

[0056] S3, the working parameters of the corresponding sensor when the arc event occurs are captured based on the coupling weight coefficient, and the working parameters are compared with the historical reference parameters of the corresponding sensor to calculate the key parameter offset;

[0057] S4, the information entropy value of the key parameter offset of the same type sensor group is calculated, and when the mutation amplitude of the information entropy value exceeds the preset entropy change threshold, a nonlinear damage mark is triggered;

[0058] S5, a segmented mapping relationship is established based on the key parameter offset and the nonlinear damage mark, and the cumulative damage evaluation value of the sensor group is output;

[0059] S6、According to the cumulative damage evaluation value, the sensor monitoring data is corrected and input into the fault diagnosis model to perform early warning.

[0060] S1, Collect current and voltage waveform data. When the current mutation rate exceeds the preset mutation threshold and is accompanied by high-frequency oscillation characteristics, it is determined that an arc event has occurred. The specific implementation is as follows:

[0061] When collecting current and voltage waveform data, the current waveform raw data is obtained through the Rogowski coil sensor installed on the switch cabinet busbar, and the voltage waveform raw data is collected through the capacitor voltage divider. The sampling frequency of the current waveform raw data is set to 1 million sampling points per second. The current differential sequence is generated by differential calculation of the current waveform raw data. The current differential sequence represents the current change rate in a unit time window, and its dimension is ampere per microsecond. The preset mutation threshold is set according to the proportion of the rated current of the switch cabinet. For example, for a switch cabinet with a rated current of 1000 amperes, the preset mutation threshold is set to 50 amperes per microsecond. The interval in the current differential sequence that continuously exceeds the preset mutation threshold for a certain time is identified as the arc event determination interval, and the start and end time stamps of the arc event determination interval are recorded by the time synchronization unit of the data acquisition system.

[0062] In the determined arc event determination interval, the voltage waveform raw data corresponding to the time period is extracted, and the time domain voltage waveform is converted to frequency domain representation using a frequency spectrum conversion algorithm. Analyze the energy distribution of the frequency domain data in a specific high frequency range, for example, in the frequency band range of 300 kilohertz to 30 megahertz. When there is a high-frequency oscillation component with an amplitude significantly higher than the background noise and a duration that meets the minimum requirement, it is confirmed that there is a high-frequency oscillation component that meets the arc characteristics. The determination condition of the high-frequency oscillation component is set based on the physical characteristics of the arc, and the background noise level is obtained by continuously collecting voltage waveform data in the system under no load state and statistical analysis.

[0063] The time domain center position is calculated according to the start and end time stamps of the arc event determination interval. The specific calculation method is to add the start time stamp value and the end time stamp value and then divide by 2. The calculated time domain center position is input into the spatial position mapping database, which stores the correspondence between time information and spatial position. By querying the physical position information matching the time stamp, the spatial coordinates of the arc occurrence position are output, which are represented in a three-dimensional rectangular coordinate system.

[0064] The generation process of the current differential sequence is specifically: performing a difference operation on adjacent sampling points of the current waveform original data, and the difference step is fixed as 1 sampling interval. The preset mutation threshold is set according to the device technical specification, and is dynamically adjusted for switch cabinets of different specifications. The adjusted threshold is stored in the device memory. In the determination condition of the high-frequency oscillation component, the specific high-frequency range is determined according to the physical characteristics of arc discharge, the background noise level is obtained by actual measurement, and the minimum requirement for the duration is set based on the power system frequency cycle.

[0065] The mapping of the time domain center position to the spatial coordinates depends on the discretization model of the physical structure of the switch cabinet, which divides the conductive components into multiple physical segments, and the center point coordinates of each physical segment are pre-measured and stored in a database. The determination method of the spatial coordinates includes: first matching the physical segment closest to the time domain center position, and then outputting the center point coordinates of the physical segment as the spatial coordinates of the arc occurrence position. To improve the positioning accuracy, a coordinate correction mechanism can be added, for example, the distance correction amount is calculated according to the electromagnetic wave propagation speed and the time difference.

[0066] The duration verification method of the high-frequency oscillation component is: in the arc event determination interval, detecting the continuous time period during which the amplitude of the high-frequency oscillation component exceeds the specific decibel value of the background noise. The error compensation of the spatial coordinates can be realized through a physical model, for example, considering the electromagnetic wave propagation speed factor. All calculation processes are based on physically measurable parameters, and the parameter setting values can be determined through device debugging process calibration. The dynamic adjustment method of the preset mutation threshold is: according to the rated current value marked on the nameplate of the switch cabinet, the specific threshold value is calculated by a fixed proportional coefficient.

[0067] S2, according to the arc occurrence position and the sensor installation coordinates, extracting the skin effect characteristics of the cabinet and the parasitic capacitance of the sensor power supply loop, calculating the coupling weight coefficient of the electromagnetic pulse to each sensor, which is specifically implemented as:

[0068] After obtaining the spatial coordinates of the arc occurrence position and the installation coordinates of each sensor, the spatial distance vector is calculated in a three-dimensional rectangular coordinate system. The calculation process of the spatial distance vector is: for each sensor, subtracting the X-axis component of the arc occurrence position spatial coordinates from the X-axis component of the sensor installation coordinates to obtain ΔX value, subtracting the Y-axis component to obtain ΔY value, and subtracting the Z-axis component to obtain ΔZ value. The length of the spatial distance vector is obtained by the formula: square root of the sum of the squares of ΔX, ΔY and ΔZ. The unit of the calculation result is meter. The length represents the straight line path length of the electromagnetic wave propagating from the arc occurrence position to the sensor.

[0069] The calculation of the skin depth distribution characteristic value of the electromagnetic wave propagation path on the surface of the cabinet body is implemented according to the electromagnetic field propagation theory. The method for determining the skin depth distribution characteristic value is as follows: first, the skin depth reference value is calculated, which is equal to the reciprocal of the square root of the product of the constant term, the circumference, the electromagnetic wave frequency value, the magnetic permeability of the cabinet material and the electrical conductivity of the cabinet material. The center frequency value of the high-frequency oscillation component detected during the arc event is used as the electromagnetic wave frequency value, and the magnetic permeability and electrical conductivity of the cabinet material are obtained through the material performance test report. The skin depth distribution characteristic value is finally equal to the skin depth reference value multiplied by the distance correction coefficient, which is adjusted according to the spatial distance vector length.

[0070] The setting rule of the distance correction coefficient is as follows: when the spatial distance vector length is less than a certain distance threshold, the distance correction coefficient is 1; when the spatial distance vector length is greater than or equal to the certain distance threshold, the distance correction coefficient decreases linearly with the increase of the spatial distance vector length. The certain distance threshold is set according to the cabinet size, for example, 1 meter for a standard switch cabinet. The proportional relationship of linear decrease is as follows: for every 1 meter increase in the spatial distance vector length, the distance correction coefficient decreases by 0.1, and the minimum is not less than 0.7.

[0071] When measuring the parasitic capacitance parameters of each sensor power supply loop by the impedance analyzer, the following operations need to be performed: connect the test terminals of the impedance analyzer between the positive and negative electrodes of the sensor power supply loop. The measurement frequency is set to the center frequency value of the arc high-frequency oscillation component, for example, 300 kHz. The measurement environment temperature is controlled within the range of 20 to 30 degrees Celsius, and the power supply loop load current does not exceed 10% of the rated current of the sensor. The impedance analyzer outputs the admittance value in complex form, and the parasitic capacitance parameter is equal to the imaginary part of the admittance value divided by the angular frequency, which is equal to 2 times the circumference times the measurement frequency value. The measurement result is stored in farad.

[0072] The preset electromagnetic coupling equation adopts a product relationship structure, and the specific expression form is as follows: the coupling weight coefficient is equal to the reciprocal of the skin depth distribution characteristic value multiplied by the parasitic capacitance parameter and then multiplied by the proportional coefficient. The proportional coefficient is determined through the experimental calibration process, and the calibration method is as follows: in the standard electromagnetic pulse test environment, the sensor output signal amplitude is recorded, the coupling weight coefficient is equal to the output signal amplitude divided by the standard electromagnetic pulse field intensity, and the proportional coefficient is obtained by statistical average of multiple test data. The value range of the proportional coefficient is set to be between 0.5 and 2.0, and the specific value is determined according to the sensor type, for example, 1.2 for current transformer type sensors and 0.8 for voltage divider type sensors.

[0073] The electromagnetic wave frequency value required for calculating the characteristic value of the skin depth distribution is derived from the center frequency of the high-frequency oscillation component determined in the arc event detection process, which is obtained by fast Fourier transform spectral analysis. The magnetic permeability of the cabinet material adopts the vacuum magnetic permeability constant, with a value of 4π x 10 -7 Henry / meter. The electrical conductivity of the cabinet material is valued according to the material test report provided by the switch cabinet manufacturer, for example, the electrical conductivity of a copper cabinet is 5.8 x 10 7 Siemens / meter. The vector length of the spatial distance is calculated based on three-dimensional coordinate measurement data, with a coordinate measurement accuracy of millimeter level.

[0074] The quality control in the measurement process of the stray capacitance parameter includes: performing zero point calibration on the impedance analyzer before measurement; measuring each sensor three times and taking the arithmetic mean value; recording the measurement results with three significant digits. The experimental calibration environment of the scaling coefficient requires: the standard electromagnetic pulse field intensity is 1 volt per meter; the sensor output signal sampling rate is 10 megahertz; the statistical average data sample size is not less than 10 groups. The calibration process is completed in the initial installation and debugging stage of the equipment.

[0075] The coupling weight coefficient of each sensor is calculated independently, and the data validity verification is performed in the calculation process: checking whether the characteristic value of the skin depth distribution is greater than 0; verifying whether the stray capacitance parameter is within the reasonable range of 1 picofarad to 100 nanofarad; confirming whether the scaling coefficient is within the preset range. If any parameter exceeds the reasonable range, the re-measurement process is triggered. The final output coupling weight coefficient is dimensionless, with a value of four significant digits.

[0076] The measurement method of the spatial coordinates includes: using a laser range finder to measure the horizontal and vertical distances of the sensor installation position relative to the reference point of the cabinet; determining the angular offset in the installation height direction by an angle measuring instrument; combining the absolute coordinates of the reference points marked on the cabinet design drawings to convert into spatial coordinates in a three-dimensional rectangular coordinate system. All coordinate data are stored in the equipment configuration database, and the coordinate update period is the equipment overhaul and maintenance period.

[0077] The acquisition approach of the electrical conductivity of the cabinet material includes: directly reading from the factory test report of the switch cabinet; or using a portable conductivity meter to measure on site. When measuring on site, select three different positions on the surface of the cabinet to measure and take the average value. The linear adjustment rule of the distance correction coefficient is implemented in the software algorithm, which specifically includes: when the vector length of the spatial distance is between 1 meter and 3 meters, first calculate the difference between the vector length of the spatial distance and 1, then calculate the product of 0.1 and the difference, and finally subtract the product from 1.0 to obtain the distance correction coefficient; when the vector length of the spatial distance is greater than 3 meters, the distance correction coefficient is fixed at 0.7.

[0078] The proportional coefficient in the electromagnetic coupling equation establishes a regular calibration mechanism, and the calibration period is set to 12 months. The calibration process includes: using a standard signal generator to generate a test signal with a frequency of 300 kHz; measuring the actual electromagnetic field intensity by a standard field intensity probe; recording the sensor output signal and recalculating the proportional coefficient. When the deviation of the newly calculated proportional coefficient from the original stored value exceeds 5%, the stored value is automatically updated. All calculations use the International System of Units, intermediate variables are operated with six significant digits, and the final result is rounded.

[0079] In the calculation process of the spatial distance vector module length, the units of ΔX, ΔY and ΔZ are meters, and the square root operation result is rounded to three decimal places. The constant term in the calculation of the skin depth reference value is the reciprocal of the square root of 2, with a value of about 0.707. When measuring the electrical conductivity of the cabinet material on site, the measurement points are selected to avoid joints and flat areas of the coating, and each measurement point is read five times to remove abnormal values and take the average. The impedance analyzer zero point calibration is performed before each measurement, and the test terminal is short-circuited during calibration to perform the automatic zeroing program.

[0080] The data verification mechanism for calculating the coupling weight coefficient includes range verification and consistency verification: range verification ensures that the calculation result is within the theoretical interval of 0.01 to 5.0; consistency verification compares the current calculation result with the historical record, and triggers an alarm if the deviation of the calculation result exceeds 10% for three consecutive times. The environmental temperature monitoring during the measurement process is realized by real-time acquisition of the temperature sensor, and the measurement is automatically paused when the temperature exceeds the range of 20 to 30 degrees Celsius. The spatial coordinate mapping data contains a timestamp, and the coordinate data version is checked after each device maintenance.

[0081] The linear adjustment rule of the distance correction coefficient is realized as a conditional judgment function in the software: input the spatial distance vector module length value, output 1.0 when the module length is less than 1 meter; when the module length is between 1 meter and 3 meters, perform a three-step calculation process; when the module length is greater than 3 meters, output 0.7. The placement position of the standard field intensity probe during the calibration of the proportional coefficient is specified to be 10 centimeters away from the measured sensor, and the direction is perpendicular to the sensor sensing surface. The final output coupling weight coefficient is applied to subsequent calculations, and normalization processing is performed to make the sum of the coupling weight coefficients of the same batch of sensor groups equal to 1.

[0082] S3, based on the coupling weight coefficient, capture the working parameters of the corresponding sensor when the arc event occurs, and compare the working parameters with the historical reference parameters of the corresponding sensor, calculate the key parameter offset, which is specifically implemented as:

[0083] According to the coupling weight coefficient filtering target sensor, the following operations are performed: reading S2 step calculation of all sensor coupling weight coefficient, the coefficient value is compared with the preset weight threshold, the preset weight threshold is set to 0.1-0.3 empirical value, for example, 0.2 can be taken. Only select the coupling weight coefficient of the sensor as the target sensor which exceeds the preset weight threshold, the sensor which does not reach the threshold is excluded in this step. The screening result of target sensor is stored in target device list register, which is refreshed every 100 milliseconds.

[0084] At the precise moment of arc event determination, the target sensor operating parameter capture operation is performed synchronously: the real-time value of the operating parameter of the target sensor is obtained by the high-speed data acquisition unit, and the operating parameter includes but is not limited to current effective value, voltage peak value, signal signal-to-noise ratio and other electrical characteristic quantities. The data acquisition time window is aligned with the arc event trigger time, and the time synchronization error is controlled within 50 microseconds. The real-time value is stored in the temporary buffer in 32-bit floating point format, with millisecond time stamp and sensor number label.

[0085] The operation process of calling historical reference parameter is as follows: according to the sensor number, the historical reference parameter database is retrieved from the non-volatile memory. The database structure includes three core fields: sensor number field, parameter type field and reference value field. The reference value is determined by statistical analysis method: the monitoring data of the sensor in the fault-free state of the device for 30 days is obtained, the arithmetic mean value is taken as the reference value after removing the maximum value and the minimum value. The reference value updating mechanism is to automatically recalculate on the first day of each month, and the alarm is triggered when the deviation of the new reference value from the original value is more than 5%.

[0086] The calculation process of key parameter offset is implemented as follows: the real-time value of the operating parameter is read from the temporary buffer, and the corresponding parameter reference value is obtained from the database. The mathematical relationship for calculating the relative offset is: the difference value is obtained by subtracting the reference value from the real-time value, then the difference value is divided by the reference value, and finally multiplied by 100 to convert to percentage value. The calculation result is a signed real number, positive value indicates positive offset, negative value indicates negative offset. The calculation formula performs data verification: when the reference value is zero, it is automatically switched to absolute difference calculation; when the calculation result exceeds ±200%, it is forced to clamp to the boundary value. The calculation process is completed in digital signal processor, and the single calculation time is not more than 10 microseconds.

[0087] When the target sensor communication is interrupted, the moving average of the last 5 valid real-time values is automatically called; when the database retrieval fails, the default benchmark value of the factory default of the device is enabled; when the calculation process has a zero division error, a special error code 0xFFFF is output and the system diagnostic program is triggered. All key parameter offset results are stored in a dedicated result register, which includes four fields: a timestamp field (accurate to milliseconds), a sensor number field, a parameter type field, and an offset value field.

[0088] The setting of the preset weight threshold is based on the electromagnetic interference propagation attenuation model, and the lower limit of the threshold is 0.1, corresponding to the maximum coupling effect value of the sensor farthest from the arc occurrence point. The historical benchmark parameter database uses a circular storage structure, retaining the last 12 months of benchmark value records for trend analysis. The percentage conversion step in the relative offset calculation can be configured to be turned off, and when turned off, the ratio is directly output in decimal form. The offset data is transmitted to the central processing unit through the industrial bus, with a transmission delay of no more than 2 milliseconds.

[0089] S4, calculate the information entropy value of the key parameter offset of the same type sensor group, and when the mutation amplitude of the information entropy value exceeds the preset entropy change threshold, trigger a nonlinear damage marker, which is implemented as follows:

[0090] After obtaining the key parameter offsets of the same type sensor group, normalization processing is performed to generate a normalized offset probability distribution. The specific implementation of this normalization processing is as follows: first, calculate the sum of the absolute values of all key parameter offsets of the sensor group, which is used as the normalization denominator; then divide the key parameter offset of each sensor by the normalization denominator to obtain the normalized offset of each sensor; the normalized offsets of all sensors form a probability distribution that satisfies the condition that the sum of all normalized offsets is equal to 1. The normalized offset has a value range of 0 to 1 and is a dimensionless real number. The key parameter offset is derived from the sensor state analysis result output by step S3, and the offset is a signed physical quantity offset value.

[0091] When calculating the Shannon information entropy value based on the normalized offset probability distribution, the calculation process is implemented according to the basic principles of information theory. The specific calculation method of the Shannon information entropy value is as follows: for each normalized offset in the probability distribution, first calculate the natural logarithm value of the offset; then multiply the offset value by the corresponding natural logarithm value; finally, add up all the product results of the sensors, and multiply the cumulative result by -1 to obtain the final entropy value. When the normalized offset value is 0, the product term corresponding to the offset is defined as 0. The calculation result of the Shannon information entropy value is output as the information entropy value at the current time, which is a dimensionless positive real number, and the calculation result is rounded to four decimal places.

[0092] When acquiring the information entropy value of the same group at the previous time, the latest valid record is retrieved from the device history database. The history database stores information entropy value data indexed by time stamp, and the retrieval conditions include: the sensor group number is completely matched; the time stamp is before the current time and closest to the current time; the time difference is not more than 72 hours. The method for calculating the entropy value mutation amplitude is: taking the current time information entropy value and the retrieved previous time information entropy value, calculating the absolute value of the difference between the two, which is the entropy value mutation amplitude. The entropy value mutation amplitude is also a dimensionless positive real number.

[0093] The setting method of the preset entropy change threshold includes two modes: in the initial setting mode, the initial threshold value of the current sensor group is set to 0.3 according to the experience value provided by the device manufacturer; in the adaptive adjustment mode, the system automatically collects at least 100 consecutive records of entropy value mutation amplitude data under normal working conditions, calculates the arithmetic mean and standard deviation of these data, and sets the preset entropy change threshold equal to the arithmetic mean plus three times the standard deviation. When the entropy value mutation amplitude value is greater than the preset entropy change threshold, a nonlinear damage marker trigger signal with a Boolean value of true is generated; otherwise, a signal with a Boolean value of false is generated.

[0094] The data verification mechanism in the normalization process includes: verifying whether all key parameter offsets have been correctly input; checking whether the normalization denominator is greater than 0.000001 (if less than this value, it is determined that all sensors have no significant offset); confirming that the sum of all normalization offsets is within a reasonable error range of 0.999 to 1.001. When the verification fails, the original data reacquisition process is triggered, and the output result of step S3 is reacquired.

[0095] The logarithmic operation in the calculation of Shannon information entropy value is implemented using the natural logarithm function, and the calculation process uses double-precision floating-point number operation. The history information entropy retrieval process includes exception handling: when there is no matching record in the database, the current information entropy value is stored in the database as the first record, and the mutation amplitude calculation step is skipped; when multiple valid records are retrieved, the single record with the time stamp closest to the current time is selected.

[0096] The dynamic adjustment rule of the preset entropy change threshold is: after adding 50 new valid arc event records, the statistical characteristic value of the historical data set is recalculated; when the difference between the new threshold and the original threshold exceeds 10%, the stored value is updated; the lower limit of the threshold is limited to 0.1, and the upper limit is limited to 1.0. After the nonlinear damage marker trigger signal is generated, the corresponding flag bit is set in the device state register, and the flag bit remains valid until it is manually reset or automatically reset after 72 hours.

[0097] The quality control of the entropy value mutation amplitude calculation includes: checking whether the input two information entropy values are within the theoretical range of 0 to 10; confirming that the time interval is within the effective interval of 1 minute to 72 hours; triggering the data review program when the calculation result is an abnormal large value. All numerical calculations use IEEE 754 floating point standard, and overflow protection mechanism is added to the key operation steps.

[0098] The data structure of the normalized offset probability distribution is stored in a dynamic array, and the array length is strictly consistent with the number of sensor group members. The historical database storage format includes four fields: a time stamp field accurate to milliseconds, a sensor group number field, an information entropy value field, and a data validity flag field. The initialization of the preset entropy change threshold is completed when the device is first put into operation, and the initial value is input through the human-machine interface and the adaptive adjustment function is enabled.

[0099] The boundary condition processing rules are specifically: when the sum of the absolute values of the key parameters offset is less than 0.000001, the normalized offset of each sensor in the group is assigned equally (i.e. 1 divided by the number of sensors); when the normalized offset is less than 0.0000000001, the product term corresponding to the offset is forced to be zero in the Shannon entropy calculation. The natural logarithm calculation module has built-in input value lower limit protection, and returns a logarithmic value of -23.025851 when the input value is lower than 10⁻¹ 0

[0100] The application logic of the nonlinear damage marker trigger signal includes: shielding repeated triggering within the signal validity period; starting the associated sensor special detection program after the signal is generated; and displaying the corresponding alarm indication on the system status panel. The information entropy value historical data uses a circular storage strategy, with a maximum of 1000 records, and the earliest record is automatically overwritten when exceeded. The threshold comparison operation sets a hysteresis interval of 0.001, and the previous judgment result is maintained when the entropy value mutation amplitude is within the threshold ± 0.001 range.

[0101] The statistical calculation process of the preset entropy change threshold is refined as follows: extracting the entropy value mutation amplitude data of the last 200 valid records from the database; eliminating the extreme values of the maximum 5% and the minimum 5%; calculating the arithmetic mean μ and the standard deviation σ of the remaining data; and finally setting the threshold value as μ+3σ. This statistical calculation is automatically performed every 90 days. The same type determination of the sensor group is based on the model matching rules defined in the device configuration file, and the same type sensors in different installation positions are classified into the same group.

[0102] ​The key parameter offset is obtained in the following manner: within 1 second after the completion of step S3, the relative offset values of each sensor are read from the data buffer. The normalization module is triggered synchronously with the arc event, and a separate probability distribution dataset is generated for each arc event. The Shannon entropy calculation unit starts operating within 200 milliseconds after the probability distribution is generated, ensuring real-time performance. All time parameters are derived from the device's high-precision real-time clock, with a synchronization error of less than 1 millisecond.

[0103] S5, based on the key parameter offset and the nonlinear damage marker, a segmented mapping relationship is established, and the cumulative damage evaluation value of the sensor group is output. The specific implementation is as follows:

[0104] After obtaining the key parameter offset and the nonlinear damage marker trigger signal, a segmented mapping relationship is established to output the cumulative damage evaluation value of the sensor group. In specific implementation, according to the Boolean state of the nonlinear damage marker trigger signal, the mapping mode is selected: when the signal is false, the linear mapping mode is enabled; when the signal is true, the nonlinear mapping mode is enabled. Mode selection is realized through conditional judgment logic, and the judgment result is stored in the mode state register. The key parameter offset is derived from the signed real value output by step S3, and the nonlinear damage marker trigger signal is derived from the Boolean signal output by step S4, both of which are transmitted synchronously through the data bus to the processing unit.

[0105] The specific operation process of the linear mapping mode is as follows: the key parameter offset values are read from the same type of sensor group one by one, and the coupling weight coefficients calculated and stored in step S2 are obtained from the non-volatile memory. For each sensor, weighted calculation is performed: multiply the key parameter offset value by the corresponding coupling weight coefficient to obtain the weighted offset of the sensor. The cumulative damage evaluation value is equal to the algebraic sum of all weighted offsets. This calculation process must satisfy the weight coefficient normalization condition, i.e. the sum of the coupling weight coefficients of all sensors in the same group is equal to 1. Data validity verification is performed during calculation: check whether each coupling weight coefficient is within the closed interval of 0 to 1; confirm whether the key parameter offset is within the ±200% relative variation range allowed by the device. The calculation result is a signed real value, and its dimension remains consistent with the original key parameter offset.

[0106] The specific operation process of the non-linear mapping mode is as follows: traverse all key parameter offsets of the same type sensor group, and extract the value with the largest absolute value as the basic input value. Perform exponential function transformation on the basic input value: first multiply the basic input value by the scaling coefficient to obtain an intermediate variable; then calculate the intermediate variable power of the base of natural logarithm; then subtract constant 1 from the calculation result; finally, multiply the difference by the proportion factor to obtain the cumulative damage evaluation value. The calculation result is a dimensionless real number, and the output range is limited to 0 to 1.0. The scaling coefficient is obtained according to the device tolerance characteristic experiment calibration, and the calibration method includes: performing gradient test on not less than 10 sensor samples of the same type in the laboratory environment, gradually increasing the key parameter offset by 10% until the sensor output deviation exceeds 20% of the range, and recording the failure critical point data; the scaling coefficient is equal to the natural logarithm function value divided by the critical offset value. The typical value of the scaling coefficient may be, for example, between 0.5 and 2.0, and the current sensor may be 1.2 in specific implementation. The proportion factor is fixed at 0.632, which corresponds to the statistical feature point of the failure probability of 63.2%.

[0107] The time sequence control mechanism of mode switching and data processing is as follows: within 5 milliseconds after the arc event trigger signal arrives, the non-linear damage marker trigger signal state output by the S4 step is obtained; the mapping mode is selected according to the signal state; the cumulative damage evaluation value calculation of the selected mode is completed within 20 milliseconds. The values output by the two modes are converted into dimensionless damage indicators in the range of 0-1.0: the result of the linear mapping mode is divided by the preset reference value and multiplied by the conversion coefficient 0.5; the result of the non-linear mapping mode is directly used as the output value. The preset reference value takes 90% of the maximum key parameter offset in the historical record of the same type sensor group, and the reference value is automatically updated every month.

[0108] The boundary condition processing rules specifically include: in the linear mapping mode, when a negative weight coefficient is detected, the absolute value is automatically taken and the system alarm is triggered; when the weighted sum result exceeds ±300%, the output value is clamped to the ±3.0 boundary value. In the non-linear mapping mode, when the basic input value is less than 0.01, it is forced to be 0; when it is higher than 0.8, it is fixed at 1.0. The abnormal processing mechanism covers the following scenarios: when the input data is missing, the last valid value is automatically called; when the mode state register is abnormal, the linear mapping mode is used by default; when the calculation is timed out, the simplified algorithm is enabled (linear mode takes the arithmetic mean value, and non-linear mode directly outputs the maximum offset).

[0109] The source and precision control measures of the key parameters are as follows: the key parameter offset is updated within 100 milliseconds after the arc event trigger, and the storage format is 32-bit floating point number. The calibration data of the scaling coefficient is saved in the device security storage area, and password authentication is required for access. The exponential function calculation uses a table lookup method combined with linear interpolation to achieve a step size of 0.001, ensuring that the maximum error is less than 0.1%. The final output of the cumulative damage assessment value is rounded to three decimal places, stored in a dedicated register and synchronized to the monitoring system.

[0110] The parameter dynamic adjustment mechanism includes: the preset reference value is automatically updated on the first day of each month, scanning the moving maximum value in the data of the previous 30 days, and the new reference value is set to 90% of the value. The scaling coefficient is calibrated every year during device maintenance, and the new value is obtained through a backup sensor calibration test, and when the deviation between the new and old values exceeds 10%, the database is updated. The weight coefficient is re-measured and entered after the device hardware is modified. All adjustment operations generate detailed log records, including operation time, values before and after modification, and operator information.

[0111] The application logic of the output result is specifically: after the cumulative damage assessment value is output, according to the value range, trigger three levels of state markers: 0 to 0.3 is green normal state, 0.3 to 0.7 is yellow warning state, and 0.7 to 1.0 is red alarm state. The state marker is persistently stored in the device operation database, which is used to generate a life trend analysis report. When the red alarm state is detected for three consecutive times, a maintenance work order is automatically generated in the maintenance system. Data transmission uses industrial bus protocol to ensure that it reaches the central monitoring platform within 100 milliseconds.

[0112] S6, according to the cumulative damage assessment value, correct the sensor monitoring data and input the fault diagnosis model to execute the early warning, which is specifically implemented as:

[0113] After completing the cumulative damage assessment value calculation, the monitoring data correction and fault warning process is executed. First, obtain the cumulative damage assessment value and the corresponding original monitoring data of the same type sensor group, where the cumulative damage assessment value is derived from the 0 to 1.0 range of dimensionless real numbers output by S5 step, and the original monitoring data is the physical quantity data collected by the sensor in real time, such as current value in amperes or voltage value in volts, with a sampling rate of 10 kilohertz, and data format using 32-bit floating point number. The two types of data are aligned and matched through time stamps accurate to milliseconds, with a time synchronization error controlled within 1 millisecond.

[0114] The construction method of the material damage transfer function comprises: the function represents a quantitative relationship between material damage degree and monitoring data distortion degree, and is obtained through material accelerated aging test calibration. The calibration process is implemented as follows: more than 30 sensor samples of the same type are subjected to a stepwise increasing mechanical stress in a laboratory environment, and the stress gradient is 10% of the rated value of the equipment per step; the electron microscope is used to observe the microcrack propagation rate of the material, and when the crack propagation rate reaches 1 micrometer per hour, the sensor output deviation value is recorded synchronously; the least square method is used to fit the experimental data, and a mapping relationship between the cumulative damage evaluation value and the data correction coefficient is established. The specific conversion rule is: when the cumulative damage evaluation value is in the range of 0 to 0.3, the output correction coefficient is 1.0; when the cumulative damage evaluation value is in the range of 0.3 to 0.7, the correction coefficient linearly decreases from 1.0 to 0.7; when the cumulative damage evaluation value is greater than 0.7, the correction coefficient decays according to an exponential law to the lower limit value 0.5. The correction coefficient is a dimensionless real number, and the calculation process retains four decimal places of accuracy.

[0115] The implementation process of the data correction operation is as follows: the original monitoring data array at the current time is read, and the corresponding data correction coefficient is obtained. For each sampling point of the original monitoring data, multiplication operation is performed: the sampling point value is multiplied by the data correction coefficient to obtain the corrected monitoring data. The calculation is performed in real time in the digital signal processor, and 1000 data points are processed per millisecond. The corrected monitoring data maintains the original physical dimension, for example, the current data still uses ampere as the unit, and the data storage format is consistent with the original data.

[0116] The input and output rules of the pre-trained fault diagnosis model are as follows: the model input is the corrected multi-sensor monitoring data group, and the data group contains waveform data of all sensors in the same group within a 200 millisecond time window, and the total number of sampling points is fixed at 2000 points. The model output is a three-level early warning signal: the value 0 represents a normal state, the value 1 represents a pre-warning state, and the value 2 represents an alarm state. The model architecture adopts a convolutional neural network composed of three convolutional layers and two fully connected layers, and is trained on a historical fault data set containing 10,000 labeled samples. The model execution period is synchronized with the data window, and the output result is updated once every 200 milliseconds.

[0117] Key parameter sources and processing details include: cumulative damage assessment value is read from the result register of S5 step, and the update frequency is synchronized with the arc event trigger. The parameters of the material damage transfer function are stored in the device read-only memory, including five characteristic mapping points, such as cumulative damage assessment value 0.3 corresponding to correction coefficient 1.0, cumulative damage assessment value 0.5 corresponding to correction coefficient 0.85, cumulative damage assessment value 0.7 corresponding to correction coefficient 0.7, cumulative damage assessment value 0.8 corresponding to correction coefficient 0.6, and cumulative damage assessment value 1.0 corresponding to correction coefficient 0.5. The correction calculation is performed in parallel using a hardware multiplier array, and the operation delay is less than 10 microseconds.

[0118] The boundary condition and abnormal processing mechanism covers the following scenarios: when the cumulative damage assessment value exceeds the range of 0 to 1.0, it is forced to be set to the nearest boundary value (i.e. less than 0 is processed as 0, and greater than 1.0 is processed as 1.0); when the transfer function output coefficient is less than 0.5, 0.5 is used as the correction coefficient; when part of the original monitoring data is missing, the arithmetic mean of the effective data in the previous 5 milliseconds is used instead; when the model input data is abnormal, the current window data is discarded and the next window data is enabled.

[0119] The training and updating rules of the fault diagnosis model are as follows: the training data set covers normal working conditions and 12 typical fault modes, with a total of 10,000 samples. The model parameters are updated online every 90 days, and the last 100 arc event data are added to retrain the full connection layer parameters. The model output layer adds confidence check: when the maximum probability value of the output layer is less than 0.6, the current warning signal is marked as invalid result.

[0120] The technical effect verification method of data correction includes: injecting standard test signals during device maintenance, and verifying the effect by comparing the total harmonic distortion rate index of the waveform before and after correction; when the cumulative damage assessment value is greater than 0.7, the harmonic component of the corrected signal should be reduced by more than 30%. The engineering application rules of the warning signal are: output value 1 triggers the device state indicator light to flash yellow; output value 2 triggers the red constant light and sends a short message notification; three consecutive output values 2 trigger the device automatic power-off protection program.

[0121] The real-time guarantee measures include: setting two levels of buffer registers in the data correction pipeline to ensure continuous data processing without interruption; using fixed-point optimization algorithm in model reasoning process, single execution time is controlled within 180 milliseconds; establishing a timeout fuse mechanism, when the calculation time is 250 milliseconds, the last valid result is directly output.

[0122] The historical data management function is implemented as follows: all monitoring data before and after correction, along with millisecond-level timestamps and cumulative damage assessment value labels, are stored in a circular buffer, retaining complete records for the most recent 72 hours. The early warning decision log contains five fields: timestamp field, cumulative damage assessment value field, correction coefficient field, model input data fingerprint field, output result field, and confidence level field, used for post-event fault tracing analysis.

[0123] The application scope of the correction factor is limited to analog sensor data, such as current, voltage, and temperature; digital signals are not corrected. For multi-range sensors, the raw data is automatically converted to percentage range values ​​for unified processing before correction calculation. The final output warning signal is transmitted to the monitoring center via industrial Ethernet, with an end-to-end transmission delay of no more than 50 milliseconds.

[0124] The technical solution of this embodiment achieves a significant improvement in the accuracy of equipment fault early warning through multi-step synergy. In step S1, when establishing the coupled weighting system, unlike the conventional average weighting method, weights are dynamically allocated based on the spatial relationship between the physical location of the sensor and the arc occurrence point. This allocation logic needs to be combined with the equipment structural topology. Step S3 uses relative offset analysis instead of absolute value comparison, effectively eliminating the influence of individual equipment differences. Step S4 introduces an entropy mutation detection mechanism, capturing early signs of nonlinear material damage through probability distribution characteristics. This characteristic is fundamentally different from the traditional linear cumulative damage model. The piecewise mapping relationship established in step S5 is not a simple mode switch, but rather based on material failure mechanism research, using exponential transformation to highlight the dominant effect of local damage in the nonlinear stage. Step S6 uses the material damage transfer function, empirically constructed through accelerated aging tests. Its mapping relationship, which converts structural damage into electrical signal correction coefficients, needs rigorous calibration and verification. The technical closed loop formed by these steps breaks through the limitations of traditional monitoring systems that only focus on electrical parameters, achieving early fault early warning through a mechanical-electrical damage correlation model.

[0125] Example 2: Figure 2 A schematic diagram of the IoT-based intelligent monitoring system for low-voltage switchgear is provided. The IoT-based intelligent monitoring system for low-voltage switchgear includes:

[0126] The arc detection module is used to collect current and voltage waveform data. When the current mutation rate exceeds the preset mutation threshold and is accompanied by high-frequency oscillation characteristics, it is determined that an arc event has occurred.

[0127] The coupling weight module is used to extract the skin effect characteristics of the cabinet and the parasitic capacitance of the sensor power supply circuit based on the location of the electric arc and the sensor installation coordinates, and to calculate the coupling weight coefficient of the electromagnetic pulse for each sensor.

[0128] An offset calculation module is configured to capture the working parameter of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and compare the working parameter with the historical reference parameter of the corresponding sensor, and calculate the key parameter offset;

[0129] An entropy change monitoring module is configured to calculate the information entropy value of the key parameter offset of the same type sensor group, and trigger the nonlinear damage marking when the mutation amplitude of the information entropy value exceeds the preset entropy change threshold;

[0130] A segmented mapping module is configured to establish a segmented mapping relationship based on the key parameter offset and the nonlinear damage marking, and output the cumulative damage evaluation value of the sensor group.

[0131] A warning correction module is configured to correct the sensor monitoring data according to the cumulative damage evaluation value and input the fault diagnosis model to perform the warning.

[0132] In the embodiments, all the calculations are dimensionless numerical calculations, and the preset parameters and threshold values in the calculations are set by the person skilled in the art according to the actual situation.

[0133] It should be noted that the present application can be deployed in the device itself to realize embedded application, or run on a PC terminal or other terminal with a user interface, thereby meeting various hardware environments and use requirements.

[0134] The above embodiments can be realized by software, hardware, firmware or any combination thereof, in whole or in part. When realized by software, the above embodiments can be realized in the form of a computer program product in whole or in part. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another through wireless or wired direction; the wired transmission mode includes optical fiber, twisted pair, coaxial cable, etc.; the wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0135] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and module described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0136] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and actual implementation can have another division manner, for example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed modules can be indirect coupling or communication connection through some interfaces, devices or modules, and can be electrical, mechanical or other forms.

[0137] The modules described as separate components can or can not be physically separated, and the components shown as modules can or can not be physical modules, which can be located in one place or distributed on a plurality of network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0138] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically, or two or more modules can be integrated in one module.

[0139] If the functions are realized in the form of software function modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the essential part or part of the technical solutions that make contributions to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0140] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0141] Finally, the above merely provides the preferred embodiments of the present application, but is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A low-voltage switch cabinet intelligent monitoring method based on the Internet of Things, characterized in that, The method comprises the following steps: S1, collect current and voltage waveform data, and determine that an arc event occurs when the current mutation rate exceeds a preset mutation threshold and is accompanied by high-frequency oscillation characteristics; S2, extract the skin effect characteristics of the cabinet and the parasitic capacitance of the sensor power supply loop according to the arc occurrence position and the sensor installation coordinates, and calculate the coupling weight coefficient of the electromagnetic pulse on each sensor; S3, capture the working parameters of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and compare the working parameters with the historical reference parameters of the corresponding sensor to calculate the key parameter offset; S4, calculate the information entropy value of the key parameter offset of the same type sensor group, and trigger a nonlinear damage marker when the mutation amplitude of the information entropy value exceeds the preset entropy change threshold; S5, establish a segmented mapping relationship based on the key parameter offset and the nonlinear damage marker, and output the cumulative damage evaluation value of the sensor group; S6, correct the sensor monitoring data according to the cumulative damage evaluation value and input the fault diagnosis model to perform early warning. 2.The low-voltage switch cabinet intelligent monitoring method based on Internet of Things according to claim 1, characterized in that, Collect current and voltage waveform data, and determine that an arc event occurs when the current mutation rate exceeds a preset mutation threshold and is accompanied by high-frequency oscillation characteristics, including: Calculate the current differential sequence based on the current waveform data, and identify the interval in the current differential sequence that continuously exceeds the preset mutation threshold as the arc event determination interval; Analyze the frequency spectrum characteristics of the voltage waveform in the arc event determination interval to confirm the existence of high-frequency oscillation components that meet the arc characteristics; Determine the spatial coordinates of the arc occurrence position according to the time domain center position of the arc event determination interval. 3.The low-voltage switch cabinet intelligent monitoring method based on Internet of Things according to claim 2, characterized in that, According to the arc occurrence position and the sensor installation coordinates, extract the skin effect characteristics of the cabinet and the parasitic capacitance of the sensor power supply loop, and calculate the coupling weight coefficient of the electromagnetic pulse on each sensor, including: Based on the spatial distance vector between the spatial coordinates of the arc occurrence position and the installation coordinates of each sensor, calculate the skin depth distribution characteristic value of the electromagnetic wave propagation path on the surface of the cabinet; Measure the parasitic capacitance parameters of each sensor power supply loop through an impedance analyzer; Input the skin depth distribution characteristic value and the parasitic capacitance parameters of each sensor into a preset electromagnetic coupling equation to output the coupling weight coefficient of each sensor. 4.The low-voltage switch cabinet intelligent monitoring method based on Internet of Things according to claim 3, characterized in that, Capture the working parameters of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and compare the working parameters with the historical reference parameters of the corresponding sensor to calculate the key parameter offset, including: Select the sensors with coupling weight coefficients exceeding a preset weight threshold as target sensors; Synchronously record the real-time values of the working parameters of the target sensors when the arc event occurs; Retrieve the corresponding parameter reference values in the historical reference parameter database of the target sensors; Calculate the relative offset between the real-time values of the working parameters and the parameter reference values as the key parameter offset.

5. The low-voltage switchgear intelligent monitoring method based on the Internet of Things according to claim 4, characterized in that, Calculate the information entropy value of the key parameter offset of the same type sensor group, and trigger a nonlinear damage marker when the mutation amplitude of the information entropy value exceeds the preset entropy change threshold, including: Normalize the key parameter offset of each sensor in the same type sensor group to generate a normalized offset probability distribution; Calculate the Shannon information entropy value based on the normalized offset probability distribution as the information entropy value at the current time. Obtaining the information entropy value of the same group at the last time, calculating the absolute difference value between the information entropy value at the current time and the information entropy value at the last time as the entropy mutation amplitude; When the entropy mutation amplitude exceeds the preset entropy change threshold, a nonlinear damage marker trigger signal is generated. 6.The low-voltage switch cabinet intelligent monitoring method based on Internet of Things according to claim 5, characterized in that, Based on the key parameter offset and the nonlinear damage marker, a segmented mapping relationship is established, and the cumulative damage evaluation value of the sensor group is output, including: According to the state of the nonlinear damage marker trigger signal, a linear mapping mode or a nonlinear mapping mode is selected; In the linear mapping mode, the weighted sum operation is performed on the key parameter offset of the same type sensor group to obtain the cumulative damage evaluation value; In the nonlinear mapping mode, the maximum key parameter offset in the same type sensor group is extracted for exponential function transformation to obtain the cumulative damage evaluation value. 7.The low-voltage switch cabinet intelligent monitoring method based on Internet of Things according to claim 6, characterized in that, According to the cumulative damage evaluation value, the sensor monitoring data is corrected and input into the fault diagnosis model to execute early warning, including: Obtaining the cumulative damage evaluation value of the sensor group and the original monitoring data of the corresponding sensor group; Through the material damage transfer function, the cumulative damage evaluation value is converted into a data correction coefficient; The data correction coefficient is multiplied by the original monitoring data to generate corrected monitoring data; The corrected monitoring data is input into the pre-trained fault diagnosis model to generate an early warning signal.

8. The low-voltage switch cabinet intelligent monitoring system based on the Internet of Things, used to realize the low-voltage switch cabinet intelligent monitoring method based on the Internet of Things according to any one of claims 1-7, characterized in that, Including: The arc detection module is used for collecting current and voltage waveform data, and when the current mutation rate exceeds the preset mutation threshold and is accompanied by high-frequency oscillation characteristics, it is determined that an arc event occurs; The coupling weight module is used for extracting the skin effect characteristics of the cabinet and the parasitic capacitance of the sensor power supply loop according to the arc occurrence position and the sensor installation coordinates, and calculating the coupling weight coefficient of electromagnetic pulse to each sensor; The offset calculation module is used for capturing the working parameters of the corresponding sensor when the arc event occurs based on the coupling weight coefficient, and comparing the working parameters with the historical reference parameters of the corresponding sensor to calculate the key parameter offset; The entropy change monitoring module is used for calculating the information entropy value of the key parameter offset of the same type sensor group, and when the information entropy value mutation amplitude exceeds the preset entropy change threshold, the nonlinear damage marker is triggered; The segmented mapping module is used for establishing a segmented mapping relationship based on the key parameter offset and the nonlinear damage marker, and outputting the cumulative damage evaluation value of the sensor group; The early warning correction module is used for correcting the sensor monitoring data according to the cumulative damage evaluation value and inputting it into the fault diagnosis model to execute early warning.

Images