Lightning protection grounding resistance real-time monitoring system based on internet of things

Abstract

The present application relates to the technical field of grounding resistance monitoring, in particular to a lightning protection grounding resistance real-time monitoring system based on the Internet of Things, which comprises a variable-frequency excitation collection module for generating a variable-frequency excitation signal superimposed with a low-frequency sinusoidal disturbance component according to a signal generator, loading the variable-frequency excitation signal to a lightning protection downlead input end, and generating a discrete voltage and current sequence. In the present application, the collected current response sequence is decomposed into multiple independent frequency band subspaces and the high-frequency node energy entropy is calculated, the information entropy theory is used to quantify the confusion degree of signal frequency energy distribution, the micro-arc discharge and high-frequency singular signals caused by mechanical connection loosening or contact surface oxidation can be captured, the static resonance resistance characteristics and dynamic contact entropy characteristics are combined, the health state discrimination of the grounding net from physical corrosion to contact stability is realized, and the situation that the lightning fire hazard caused by poor contact is covered due to the false qualified resistance value is avoided.

Description

Technical Field

[0001] This invention relates to the field of grounding resistance monitoring technology, and in particular to a lightning protection grounding resistance real-time monitoring system based on the Internet of Things. Background Technology

[0002] Grounding resistance monitoring technology is used to detect and evaluate the electrical performance of grounding devices in power systems, communication base stations, high-rise buildings and industrial facilities. It covers the continuous observation of the conduction status, impedance characteristics and physical integrity of grounding grids, down conductors and grounding bodies.

[0003] Most existing grounding resistance monitoring technologies are based on the average value of steady-state current and voltage, focusing only on the magnitude of macroscopic resistance. When there are loose bolts or surface oxidation at the grounding connection point but it is not completely disconnected, the readings of conventional instruments may still be within the acceptable range. This static measurement method is difficult to determine the stability of electrical connections under dynamic current impacts, which can lead to equipment damage or even fire accidents due to sudden changes in contact resistance during lightning strikes, resulting in blind spots in safety monitoring. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a real-time monitoring system for lightning protection grounding resistance based on the Internet of Things.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a real-time monitoring system for lightning protection grounding resistance based on the Internet of Things includes: The frequency conversion excitation acquisition module is used to generate a frequency conversion excitation signal with low-frequency sinusoidal disturbance components based on the signal generator, and to load the frequency conversion excitation signal onto the input terminal of the lightning protection down conductor to generate a discrete voltage and current sequence. The phase zero-point locking module is used to calculate the absolute value of the real-time phase difference between the voltage data series and the current data series based on the discrete voltage and current sequence, multiply the absolute value of the real-time phase difference by the low-frequency sinusoidal disturbance component to obtain the frequency gradient estimate, input the frequency gradient estimate into the integral controller to update the excitation frequency value at the next moment, drive the excitation frequency value to approach the frequency point where the phase difference is zero, lock the specific frequency when the phase difference is zero, and generate the resonant frequency resistor value.

[0006] Preferably, the system further includes: The energy entropy feature calculation module is used to decompose the current data column into multiple independent frequency band subspaces through a multi-layer frequency band decomposition filter bank according to the discrete voltage and current sequence, quantify the disorder of the signal frequency domain energy distribution, and generate high-frequency node energy entropy. The connection hazard identification module is used to compare the resonant frequency resistance value with a preset grounding body corrosion failure resistance threshold and the high-frequency node energy entropy with a preset metal connection loosening micro-arc light threshold based on the resonant frequency resistance value and the high-frequency node energy entropy. It extracts the judgment results of resistance exceeding the limit and the judgment results of energy entropy change, and combines the static impedance characteristics of the resistance dimension and the dynamic contact characteristics of the energy entropy dimension to generate the grounding grid health status.

[0007] Preferably, the frequency conversion excitation acquisition module includes: The excitation signal injection submodule generates a frequency conversion excitation signal superimposed with a low-frequency sinusoidal disturbance component. It sets the reference frequency parameter and superimposes the low-frequency sinusoidal disturbance waveform to synthesize a composite excitation waveform and drive the power amplifier circuit to increase the waveform power. The composite waveform with increased power is loaded to the input terminal of the lightning protection down conductor to establish a signal transmission loop and maintain a continuous excitation injection state to generate a disturbance excitation signal. The analog sampling submodule captures the instantaneous potential difference between the two ends of the grounding grid according to the disturbance excitation signal, drives the current sensor to capture the instantaneous current flowing through the down conductor, converts the continuous analog signal into digital instantaneous values, and generates original discrete data points. The timing alignment and reorganization submodule extracts the timestamp markers associated with the voltage sampling points and current sampling points based on the original discrete data points, compares the time index differences between the voltage data stream and the current data stream, fills the missing time slots through interpolation to synchronize the lengths of the two data streams, merges the synchronized voltage value column and current value column to construct a unified data matrix, and generates a discrete voltage and current sequence.

[0008] Preferably, the phase zero-point locking module includes: The phase gradient estimation submodule calculates the absolute value of the real-time phase difference between the voltage data series and the current data series based on the discrete voltage and current sequences, extracts the real-time phase difference value and performs calculations with the low-frequency sinusoidal disturbance component, removes the high-frequency noise component in the product result through low-pass filtering, extracts the filtered DC component as the slope reflecting the phase change trend, and generates the frequency adjustment gradient value. The zero-phase frequency search submodule inputs the frequency adjustment gradient value into the integral controller for accumulation calculation, adjusts the excitation frequency output value at the next moment according to the accumulation result, drives the excitation frequency to move in the direction that makes the gradient value zero, monitors the real-time phase difference change until it converges to zero, locks the specific frequency point that maintains the zero phase difference state, and generates the purely resistive resonant frequency. The impedance magnitude calculation submodule extracts the corresponding voltage amplitude data and current amplitude data based on the purely resistive resonant frequency, calculates the ratio of voltage amplitude to current amplitude, eliminates the influence of lead inductance and stray capacitance on the imaginary part of impedance measurement, obtains the impedance magnitude when the circuit exhibits purely resistive characteristics, and generates the resonant frequency resistance value.

[0009] Preferably, the energy entropy feature calculation module includes: The frequency band space decomposition submodule decomposes the current data column into multiple independent frequency band subspaces based on the discrete voltage and current sequence. Iterative filtering processing of the current data is performed using multi-level high-pass and low-pass filters to divide the broadband signal into multiple non-overlapping frequency segments. Detail coefficients and approximation coefficients in each frequency segment are extracted to generate a multi-layer frequency band coefficient set. The node energy quantization submodule accumulates the squared values ​​of the coefficients in each frequency band according to the multi-layer frequency band coefficient set to obtain the total energy of the frequency band, traverses all frequency band nodes and counts their respective energy values, constructs a feature vector reflecting the distribution of signal energy in different frequency intervals, and generates a frequency band energy distribution vector.

[0010] Preferably, the energy entropy feature calculation module further includes: The information entropy calculation submodule sums the energies of all frequency band nodes according to the frequency band energy distribution vector to obtain the total spectrum energy, calculates the proportion of the energy of frequency band nodes in multiple independent frequency band subspaces to the total spectrum energy, performs calculations on each proportion value, quantifies the disorder of the signal frequency domain energy distribution, and generates high-frequency node energy entropy.

[0011] Preferably, the connection hazard detection module includes: The resistive failure comparison submodule, based on the resonant frequency resistance value and the node energy entropy, calls the preset grounding body corrosion failure resistance threshold, compares the resonant frequency resistance value with the grounding body corrosion failure resistance threshold, determines whether the resistance value exceeds the allowable safe range, marks the resistance over-limit state to identify whether there is a risk of corrosion fracture in the grounding body, and generates a corrosion aging judgment item. The contact loosening assessment submodule, based on the corrosion aging judgment item, calls the preset metal connection loosening micro-arc light threshold, compares the high-frequency node energy entropy with the metal connection loosening micro-arc light threshold, detects whether there is a high-frequency energy dissipation phenomenon caused by poor contact, marks the energy entropy change state to identify whether there is a loosening or micro-arc light discharge hazard at the electrical connection point, and generates a contact instability judgment item.

[0012] Preferably, the connection hazard detection module further includes: The status level integration submodule, based on the contact instability judgment item, the comparison results of the combined resistance dimension and the energy entropy dimension, and the combined logic of whether the resistance value is normal and whether the contact is stable, looks up the corresponding risk level table, distinguishes between normal operation, simple aging, poor contact and compound fault conditions, determines the current operating level of the lightning protection grounding system, and generates the grounding grid health status.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, a frequency conversion excitation signal superimposed with a low-frequency sinusoidal disturbance component is generated and applied to the lightning protection down conductor. The excitation frequency is dynamically adjusted by feedback control using the product of the real-time absolute value of the phase difference and the disturbance component. The drive system automatically locks the resonant frequency point where the voltage and current phase difference is zero. The extreme value search control principle forces the circuit to operate in a purely resistive state, canceling the reactance component generated by the distributed inductance and stray capacitance of the down conductor. This eliminates the illusory impedance interference without disconnection, obtaining the resonant frequency resistance value that reflects the true corrosion degree of the grounding electrode, thus improving the electrical performance under complex conditions with long leads. The accuracy of resistance measurement is improved by decomposing the collected current response sequence into multiple independent frequency band subspaces and calculating the energy entropy of high-frequency nodes. The information entropy theory is used to quantify the disorder of the signal frequency domain energy distribution. This allows for the capture of micro-arc discharge and high-frequency singular signals caused by loose mechanical connections or oxidation of contact surfaces. By combining static resonant resistance characteristics with dynamic contact entropy characteristics, the health status of the grounding grid from physical corrosion to contact stability can be determined. This avoids the situation where the risk of lightning fire caused by poor contact is masked by false resistance values, thus improving the operational reliability of the lightning protection grounding system. Attached Figure Description

[0014] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0016] Please see Figure 1 The present invention provides a technical solution: a real-time monitoring system for lightning protection grounding resistance based on the Internet of Things, comprising: The frequency conversion excitation acquisition module is used to generate a frequency conversion excitation signal with low-frequency sinusoidal disturbance components based on the signal generator, and to load the frequency conversion excitation signal onto the input terminal of the lightning protection down conductor to generate a discrete voltage and current sequence. The phase zero-point locking module is used to calculate the absolute value of the real-time phase difference between the voltage data series and the current data series based on the discrete voltage and current sequences. It multiplies the absolute value of the real-time phase difference by a low-frequency sinusoidal disturbance component to obtain a frequency gradient estimate. The frequency gradient estimate is input to the integral controller to update the excitation frequency value at the next moment, driving the excitation frequency value to approach the frequency point where the phase difference is zero. It locks the specific frequency when the phase difference is zero and generates the resonant frequency resistor value. The energy entropy feature calculation module is used to decompose the current data column into multiple independent frequency band subspaces through a multi-layer frequency band decomposition filter bank based on the discrete voltage and current sequence, quantify the disorder of the signal frequency domain energy distribution, and generate high-frequency node energy entropy. The connection hazard identification module is used to compare the resonant frequency resistance value with the preset grounding body corrosion failure resistance threshold and the high-frequency node energy entropy with the preset metal connection loose micro-arc light threshold based on the resonant frequency resistance value and the high-frequency node energy entropy. It extracts the judgment results of resistance exceeding the limit and the judgment results of energy entropy change. Combining the static impedance characteristics of the resistance dimension and the dynamic contact characteristics of the energy entropy dimension, it generates the grounding grid health status.

[0017] The frequency converter excitation acquisition module includes: The excitation signal injection submodule generates a frequency conversion excitation signal superimposed with a low-frequency sinusoidal disturbance component. It sets the reference frequency parameter and superimposes the low-frequency sinusoidal disturbance waveform to synthesize a composite excitation waveform and drive the power amplifier circuit to increase the waveform power. The composite waveform with increased power is loaded to the input terminal of the lightning protection down conductor to establish a signal transmission loop and maintain a continuous excitation injection state to generate a disturbance excitation signal. The analog sampling submodule captures the instantaneous potential difference between the two ends of the grounding grid based on the disturbance excitation signal, drives the current sensor to capture the instantaneous current flowing through the down conductor, converts the continuous analog signal into digital instantaneous values, and generates original discrete data points. The timing alignment and reorganization submodule extracts the timestamp markers associated with voltage sampling points and current sampling points based on the original discrete data points, compares the time index differences between the voltage data stream and the current data stream, fills in the missing time slots through interpolation to synchronize the length of the two data streams, merges the synchronized voltage value column and current value column to construct a unified data matrix, and generates discrete voltage and current sequences.

[0018] Specifically, the excitation signal injection submodule configures the register parameters of the direct digital frequency synthesizer (DDS) and sets the scan start point of the reference frequency parameter to... The termination point is Set the sweep frequency step to Construct a low-frequency sinusoidal perturbation waveform, and set the frequency of this perturbation waveform to be... To avoid interference at the 50Hz power frequency, the amplitude of the disturbance waveform is set to the reference excitation amplitude. Multiply, perform waveform superposition operation, using the formula Calculate the instantaneous amplitude of the synthesized signal, where, This indicates that the synthesized composite excitation waveform is in Voltage value at time, This indicates the preset reference carrier amplitude. This indicates the current scanning reference frequency. Represents a time variable. This represents the disturbance coefficient. This represents the low-frequency disturbance frequency. The calculated digital waveform sequence is input into the digital-to-analog converter (DAC), and the conversion rate is set to [value missing]. To ensure waveform fidelity, the Class D power amplifier circuit connected to the back end is driven, and the gain coefficient is adjusted. This makes the peak voltage at the output terminal reach Using the formula Estimated output power, of which For output power, Peak voltage, To estimate the loop load impedance, an amplified analog voltage signal is applied to the test terminals of the lightning protection down conductor using a Kelvin four-wire connection. The loop current is then monitored to see if it exceeds the preset minimum conduction current threshold. If the gain is below this threshold, the gain coefficient will be gradually increased. Until the current meets the requirements, the signal is continuously output and the waveform is kept undistorted, generating a disturbance excitation signal.

[0019] The analog sampling submodule configures the differential input channels of the analog front-end AFE, connecting them to the voltage and current test points of the grounding grid loop, respectively. For voltage signals, a resistor divider network attenuates the high-voltage signal to the input range of the ADC. For current signals, read the output voltage of the Hall current sensor and set the sampling frequency of the analog-to-digital converter (ADC) to [value missing]. This frequency setting is based on the highest excitation frequency. To satisfy the Nyquist sampling theorem, a synchronous sampling clock is activated to trigger simultaneous data acquisition in both the voltage and current channels. The acquired analog voltage and current values ​​are quantized and encoded using a quantization formula. Calculate discrete numerical values, where, For the first The digital quantization value of each sampling point For the first The analog input voltage at each sampling time. This is the positive reference voltage for the ADC. This is the negative reference voltage for the ADC. The quantization bit depth is set here. To ensure resolution, This represents the rounding function, which stores the quantized raw binary data into a First-In-First-Out (FIFO) queue. Whenever the queue depth reaches... A direct memory access (DMA) transfer is triggered every time a data point is reached, moving the data block to the microprocessor's static random access memory (SRAM) to remove the DC bias component caused by the sensor's zero-point drift, thus generating the original discrete data points.

[0020] The timing alignment and reassembly submodule reads the timestamp metadata from the headers of the voltage and current channel data packets. This timestamp is marked by a high-precision 64-bit counter inside the FPGA at the moment of sampling trigger, with a counting clock frequency of [frequency missing]. Calculate the start time of the voltage sampling sequence With the start time of the current sampling sequence Deviation value between Set the maximum allowable synchronization error threshold as follows: ,like If the absolute value is less than the threshold, sequence pairing is performed directly. If the absolute value is greater than the threshold, then Lagrange interpolation is performed on the current series based on the time axis of the voltage series, using the three-point interpolation formula. Calculate the current value at the alignment moment, where, For the target time The interpolated current, For the known sampling current points before and after the target time, For the Lagrange basis functions, specifically: , and Given the time coordinates of the sampling points, interpolation calculations are used to fill in the time gaps caused by hardware trigger delays, resulting in aligned voltage values. With current value Index by row The data are filled into a two-dimensional array matrix in a one-to-one correspondence. Data segments that do not completely overlap at the beginning and end of the matrix are trimmed to ensure that each row in the matrix contains the voltage amplitude and current amplitude at the same time, thus generating a discrete voltage and current sequence.

[0021] The phase zero-point locking module includes: The phase gradient estimation submodule calculates the absolute value of the real-time phase difference between the voltage data series and the current data series based on the discrete voltage and current sequences. It extracts the real-time phase difference value and performs calculations with the low-frequency sinusoidal disturbance component. It removes the high-frequency noise component in the product result through low-pass filtering and extracts the filtered DC component as the slope reflecting the phase change trend to generate the frequency adjustment gradient value. The zero-phase frequency search submodule inputs the frequency adjustment gradient value into the integral controller for accumulation calculation. Based on the accumulation result, it adjusts the excitation frequency output value at the next moment, drives the excitation frequency to move in the direction that makes the gradient value zero, monitors the real-time phase difference change until it converges to zero, locks the specific frequency point that maintains the zero phase difference state, and generates the purely resistive resonant frequency. The impedance magnitude calculation submodule extracts the corresponding voltage and current amplitude data based on the purely resistive resonant frequency, calculates the ratio of voltage amplitude to current amplitude, eliminates the influence of lead inductance and stray capacitance on the imaginary part of impedance measurement, obtains the impedance magnitude when the circuit exhibits purely resistive characteristics, and generates the resonant frequency resistance value.

[0022] Specifically, the phase gradient estimation submodule performs Hilbert transforms on the synchronized voltage and current data sequences respectively, constructs analytic signals to extract the instantaneous phase, and uses the formula... Calculate the phase angle, where For instantaneous phase, For signal The Hilbert transform result, Given the original voltage or current signal, calculate the instantaneous phase of the voltage. Phase with current The absolute value of the difference is used to obtain the real-time phase difference sequence. Read the low-frequency sinusoidal disturbance waveform set in the previous steps. The real-time phase difference sequence is multiplied point-by-point with the disturbance waveform to obtain the modulation signal. Design a second-order Butterworth low-pass filter and set its cutoff frequency. For the perturbation frequency 0.1 times that is Using difference equations For modulated signals Perform digital filtering, where This is the current output value. This is the current input value. For historical output, For historical input, To extract the DC component remaining after filtering, based on the filter coefficients calculated from the cutoff frequency, the magnitude and sign of this DC component reflect the distance and direction of the current excitation frequency relative to the resonant point. This DC component is then used as the gradient slope. Generate frequency adjustment gradient values.

[0023] The zero-phase frequency search submodule establishes a discrete-time integral controller and sets the integral gain coefficient. experience points This value is set based on the system response speed requirements. The adjustment is completed within the time frame without overshoot oscillation; the current gradient slope is then read. Using iterative formulas Calculate the excitation frequency at the next time step, where For the updated output frequency, For the current excitation frequency, For integral gain, The gradient value is adjusted based on the input frequency, and the updated frequency value is fed back to the signal generation unit to adjust the output waveform frequency, while continuously monitoring the real-time phase difference. The moving average is used to set the convergence criterion threshold as... This threshold is set based on the phase resolution limit of the ADC. When continuous The average phase difference of each sampling period is less than When the system has converged to the zero-phase state, the corresponding frequency point is the frequency at which the circuit exhibits pure resistive characteristics. The frequency update is stopped and the frequency value is recorded to generate the pure resistive resonant frequency.

[0024] The impedance magnitude calculation submodule, based on a locked frequency point, extracts the value that remains stable at that frequency. The effective values ​​(RMS) of the voltage and current sequences for each period of discrete voltage and current data were calculated using the formula. Calculate the effective value of the voltage, where This is the effective value of the voltage. The total number of data points captured. For the first The voltage values ​​at each sampling point are calculated, and the effective current value is calculated similarly. Using Ohm's Law formula Calculate the impedance magnitude of the circuit, where The resistance at the resonance point. To calculate the effective value of the voltage, To calculate the effective value of the current, since the inductive and capacitive reactances cancel each other out at the resonant frequency, the imaginary impedance is zero. Therefore, the calculated impedance magnitude is the real part of the pure resistance. The resistance value of the pre-calibrated test lead is then read. Subtract the lead resistance value from the calculated result to obtain the true grounding resistance. This generates the resistance value at the resonant frequency.

[0025] The energy entropy feature calculation module includes: The frequency band space decomposition submodule decomposes the current data column into multiple independent frequency band subspaces based on the discrete voltage and current sequence. It uses multi-level high-pass and low-pass filters to perform iterative filtering on the current data, divides the broadband signal into multiple non-overlapping frequency segments, extracts the detail coefficients and approximation coefficients in each frequency segment, and generates a multi-layer frequency band coefficient set. The node energy quantization submodule accumulates the squared values ​​of the coefficients in each frequency band based on the multi-layer frequency band coefficient set to obtain the total energy of the frequency band. It traverses all frequency band nodes and counts their respective energy values, constructs a feature vector that reflects the distribution of signal energy in different frequency intervals, and generates a frequency band energy distribution vector. The information entropy calculation submodule sums the energies of all frequency band nodes based on the frequency band energy distribution vector to obtain the total spectrum energy, calculates the proportion of the energy of frequency band nodes in multiple independent frequency band subspaces to the total spectrum energy, performs calculations on each proportion value, quantifies the disorder of the signal frequency domain energy distribution, and generates the high-frequency node energy entropy.

[0026] Specifically, the frequency band spatial decomposition submodule extracts the current value sequence from the discrete voltage and current sequence as the signal to be processed. The Daubechies4 wavelet basis function was selected as the mother wavelet for decomposition, and the number of layers for wavelet packet decomposition was set to [value missing]. The signal is decomposed layer by layer using the Mallat pyramid algorithm. In each layer, the approximation coefficients of the previous layer are passed through a low-pass filter. and high-pass filter The convolution operation is performed using the following formula: and ,in Indicates the first The first layer Each node coefficient For discrete-time indexing, For convolution index, and The impulse response coefficients of the low-pass and high-pass filters are decomposed respectively. The convolution result is downsampled by a factor of two to remove redundant data, i.e., a value is taken at every other point. After three layers of iterative processing, the original broadband signal is completely divided into... Eight independent frequency band subspaces that do not overlap, each corresponding to a different frequency component from low to high frequency. The coefficient arrays of each subspace are rearranged according to frequency order to form a structured array containing eight sets of coefficient sequences, thus generating a multi-layer frequency band coefficient set.

[0027] The node energy quantization submodule obtains the energy by traversing and decomposing the multi-layer frequency band coefficient set. Each independent frequency band node is used to extract the coefficient sequence for each node. ,in The frequency band number ranges from 1 to 8. For each frequency band, the signal energy it contains is calculated using a sum of squares formula. Perform calculations, where Indicates the first The total energy of each frequency band node The length of the coefficient sequence within this frequency band. For the first The first frequency band The amplitude of each wavelet packet coefficient is checked after the energy calculation of all nodes is completed to prevent mathematical errors in subsequent logarithmic operations. Value, if it exists In the case of [the specific case], a very small non-zero bias value is assigned. All calculated energy values to The frequency bands are sequentially filled into a one-dimensional vector in ascending order of frequency. This vector intuitively reflects the distribution density of current signal energy in different frequency ranges, generating a frequency band energy distribution vector.

[0028] The information entropy calculation submodule sums the energy values ​​of all frequency bands in the frequency band energy distribution vector to obtain the total spectral energy. ,in For total energy, For the first Energy of each frequency band Given the total number of frequency bands, calculate the proportion of energy in the total spectral energy of each independent frequency band subspace, i.e., the normalized probability. ,in Indicates the first The energy probability density of each frequency band satisfies The disorder of energy distribution is quantified using the Shannon entropy principle, employing the formula... Calculate the energy entropy across the entire frequency band, where The calculated entropy value, For the natural logarithm, special attention is paid to energy changes in the high-frequency range because micro-arc light generated by poor contact is usually accompanied by high-frequency noise. The high-frequency band (i.e., the index) is extracted. The energy proportion of the frequency band (5 to 8) is weighted and corrected, or the full-spectrum entropy value is directly used as the disorder index. Here, the full-spectrum entropy is directly used. It is used to characterize the degree of disorder in the frequency domain energy distribution of a signal. When a micro-arc or loosening occurs, energy will diverge from the fundamental wave to the high frequency, causing the entropy value to increase significantly and generating high-frequency node energy entropy.

[0029] The connection hazard identification module includes: The resistive failure comparison submodule calls the preset grounding body corrosion failure resistance threshold based on the resonant frequency resistance value and the node energy entropy. It compares the resonant frequency resistance value with the grounding body corrosion failure resistance threshold to determine whether the resistance value exceeds the allowable safe range, marks the resistance over-limit state to identify whether there is a risk of corrosion fracture in the grounding body, and generates corrosion aging judgment items. The contact loosening assessment submodule, based on the corrosion aging judgment item, calls the preset metal connection loosening micro-arc light threshold, compares the high-frequency node energy entropy with the metal connection loosening micro-arc light threshold, detects whether there is a high-frequency energy dissipation phenomenon caused by poor contact, marks the energy entropy change state to identify whether there is a loosening or micro-arc light discharge hazard at the electrical connection point, and generates contact instability judgment item. The status level comprehensive submodule, based on the contact instability judgment item, the comparison results of the combined resistance dimension and the energy entropy dimension, and the combined logic of whether the resistance value is normal and whether the contact is stable, looks up the corresponding risk level table, distinguishes between normal operation, simple aging, poor contact and compound fault conditions, determines the current operating level of the lightning protection grounding system, and generates the grounding grid health status.

[0030] Specifically, the resistive failure comparison submodule reads the calculated true resistance value based on the resonant frequency resistance value and the node energy entropy. The reference resistance value of the lightning protection grounding system, entered during the initial installation and acceptance, is retrieved from the device's non-volatile memory. The allowable resistance aging deviation coefficient is set to (That is, a maximum resistance value increase of 50% is allowed), using the formula The threshold resistance of grounding electrode corrosion failure was calculated. ,in To determine the critical resistance value at which failure occurs, numerical comparison logic is executed to make a judgment. Is it greater than ,like This indicates that the cross-sectional area of ​​the grounding down conductor or grounding grid metal conductor has decreased due to electrochemical corrosion, or that there is a breakage. The corrosion status flag will be set accordingly. Set it to 1 if necessary, otherwise set it to 0 to generate a corrosion aging judgment item.

[0031] The contact loosening assessment submodule reads the high-frequency node energy entropy calculated in real time based on the corrosion aging judgment item. The system invokes a preset threshold for loose metal connections using micro-arc light. This threshold is set based on statistical analysis of historical entropy data from a large amount of data collected during normal operation, calculating the average entropy value of the historical data. and standard deviation According to 3 The rule sets an abnormal threshold. ,in This refers to the threshold of micro-arc light due to loose metal connection, which will be the current... and Perform a comparison, if This indicates the presence of nonlinear high-frequency components in the signal spectrum, a typical characteristic of microscopic arc discharge caused by loose electrical connection points. The time window continuously exceeding this threshold is statistically analyzed; if the duration exceeds... Then the contact status flag will be set. Set it to 1, otherwise keep it at 0, and generate a contact instability judgment item.

[0032] The status level synthesis submodule, based on the contact instability judgment item, summarizes the corrosion status flags output by the resistive failure comparison submodule. Contact status flags output by the contact loosening assessment submodule Construct a two-dimensional state decision logic table, if and If the condition is determined to be "Level 1: Normal Operation", it means that the grounding grid resistance is qualified and the connection is reliable; if and The classification is "Level 2: Simple Aging," indicating that corrosion alone causes an increase in resistance, but there is no contact spark. and The classification is "Level 3: Poor Contact," indicating that the resistance is still within the acceptable range but there is a risk of loosening and arcing. and If the fault is classified as "Level 4: Composite Fault", it indicates that the system has a serious safety risk. The system outputs the corresponding text description and risk level code, and generates the grounding grid health status.

Claims

1. A real-time monitoring system for lightning protection grounding resistance based on the Internet of Things, characterized in that, The system includes: The frequency conversion excitation acquisition module is used to generate a frequency conversion excitation signal with low-frequency sinusoidal disturbance components based on the signal generator, and to load the frequency conversion excitation signal onto the input terminal of the lightning protection down conductor to generate a discrete voltage and current sequence. The phase zero-point locking module is used to calculate the absolute value of the real-time phase difference between the voltage data series and the current data series based on the discrete voltage and current sequence, multiply the absolute value of the real-time phase difference by the low-frequency sinusoidal disturbance component to obtain the frequency gradient estimate, input the frequency gradient estimate into the integral controller to update the excitation frequency value at the next moment, drive the excitation frequency value to approach the frequency point where the phase difference is zero, lock the specific frequency when the phase difference is zero, and generate the resonant frequency resistor value.

2. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 1, characterized in that, The system also includes: The energy entropy feature calculation module is used to decompose the current data column into multiple independent frequency band subspaces through a multi-layer frequency band decomposition filter bank according to the discrete voltage and current sequence, quantify the disorder of the signal frequency domain energy distribution, and generate high-frequency node energy entropy. The connection hazard identification module is used to compare the resonant frequency resistance value with a preset grounding body corrosion failure resistance threshold and the high-frequency node energy entropy with a preset metal connection loosening micro-arc light threshold based on the resonant frequency resistance value and the high-frequency node energy entropy. It extracts the judgment results of resistance exceeding the limit and the judgment results of energy entropy change, and combines the static impedance characteristics of the resistance dimension and the dynamic contact characteristics of the energy entropy dimension to generate the grounding grid health status.

3. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 1, characterized in that, The frequency conversion excitation acquisition module includes: The excitation signal injection submodule generates a frequency conversion excitation signal superimposed with a low-frequency sinusoidal disturbance component. It sets the reference frequency parameter and superimposes the low-frequency sinusoidal disturbance waveform to synthesize a composite excitation waveform and drive the power amplifier circuit to increase the waveform power. The composite waveform with increased power is loaded to the input terminal of the lightning protection down conductor to establish a signal transmission loop and maintain a continuous excitation injection state to generate a disturbance excitation signal. The analog sampling submodule captures the instantaneous potential difference between the two ends of the grounding grid according to the disturbance excitation signal, drives the current sensor to capture the instantaneous current flowing through the down conductor, converts the continuous analog signal into digital instantaneous values, and generates original discrete data points. The timing alignment and reorganization submodule extracts the timestamp markers associated with the voltage sampling points and current sampling points based on the original discrete data points, compares the time index differences between the voltage data stream and the current data stream, fills the missing time slots through interpolation to synchronize the lengths of the two data streams, merges the synchronized voltage value column and current value column to construct a unified data matrix, and generates a discrete voltage and current sequence.

4. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 1, characterized in that, The phase zero-point locking module includes: The phase gradient estimation submodule calculates the absolute value of the real-time phase difference between the voltage data series and the current data series based on the discrete voltage and current sequences, extracts the real-time phase difference value and performs calculations with the low-frequency sinusoidal disturbance component, removes the high-frequency noise component in the product result through low-pass filtering, extracts the filtered DC component as the slope reflecting the phase change trend, and generates the frequency adjustment gradient value. The zero-phase frequency search submodule inputs the frequency adjustment gradient value into the integral controller for accumulation calculation, adjusts the excitation frequency output value at the next moment according to the accumulation result, drives the excitation frequency to move in the direction that makes the gradient value zero, monitors the real-time phase difference change until it converges to zero, locks the specific frequency point that maintains the zero phase difference state, and generates the purely resistive resonant frequency. The impedance magnitude calculation submodule extracts the corresponding voltage amplitude data and current amplitude data based on the purely resistive resonant frequency, calculates the ratio of voltage amplitude to current amplitude, eliminates the influence of lead inductance and stray capacitance on the imaginary part of impedance measurement, obtains the impedance magnitude when the circuit exhibits purely resistive characteristics, and generates the resonant frequency resistance value.

5. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 2, characterized in that, The energy entropy feature calculation module includes: The frequency band space decomposition submodule decomposes the current data column into multiple independent frequency band subspaces based on the discrete voltage and current sequence. Iterative filtering processing of the current data is performed using multi-level high-pass and low-pass filters to divide the broadband signal into multiple non-overlapping frequency segments. Detail coefficients and approximation coefficients in each frequency segment are extracted to generate a multi-layer frequency band coefficient set. The node energy quantization submodule accumulates the squared values ​​of the coefficients in each frequency band according to the multi-layer frequency band coefficient set to obtain the total energy of the frequency band, traverses all frequency band nodes and counts their respective energy values, constructs a feature vector reflecting the distribution of signal energy in different frequency intervals, and generates a frequency band energy distribution vector.

6. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 5, characterized in that, The energy entropy feature calculation module also includes: The information entropy calculation submodule sums the energies of all frequency band nodes according to the frequency band energy distribution vector to obtain the total spectrum energy, calculates the proportion of the energy of frequency band nodes in multiple independent frequency band subspaces to the total spectrum energy, performs calculations on each proportion value, quantifies the disorder of the signal frequency domain energy distribution, and generates high-frequency node energy entropy.

7. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 2, characterized in that, The connection vulnerability detection module includes: The resistive failure comparison submodule, based on the resonant frequency resistance value and the node energy entropy, calls the preset grounding body corrosion failure resistance threshold, compares the resonant frequency resistance value with the grounding body corrosion failure resistance threshold, determines whether the resistance value exceeds the allowable safe range, marks the resistance over-limit state to identify whether there is a risk of corrosion fracture in the grounding body, and generates a corrosion aging judgment item. The contact loosening assessment submodule, based on the corrosion aging judgment item, calls the preset metal connection loosening micro-arc light threshold, compares the high-frequency node energy entropy with the metal connection loosening micro-arc light threshold, detects whether there is a high-frequency energy dissipation phenomenon caused by poor contact, marks the energy entropy change state to identify whether there is a loosening or micro-arc light discharge hazard at the electrical connection point, and generates a contact instability judgment item.

8. The real-time monitoring system for lightning protection grounding resistance based on the Internet of Things according to claim 7, characterized in that, The connection vulnerability detection module also includes: The status level integration submodule, based on the contact instability judgment item, the comparison results of the combined resistance dimension and the energy entropy dimension, and the combined logic of whether the resistance value is normal and whether the contact is stable, looks up the corresponding risk level table, distinguishes between normal operation, simple aging, poor contact and compound fault conditions, determines the current operating level of the lightning protection grounding system, and generates the grounding grid health status.

Images