Charging pile insulation resistance self-detection and early warning system

Abstract

The invention discloses a charging pile insulation resistance self-detection and early warning system, and relates to the technical field of power equipment state detection, and the system comprises a vector construction module which is used for extracting characteristic frequency point displacement and dielectric loss gradient based on frequency domain characteristic data, and constructing a material microscopic aging state vector; the strength analysis module is used for calculating a characteristic frequency point displacement acceleration based on the material micro aging state vector, obtaining a risk coupling strength tensor in combination with a dielectric loss gradient, and generating a risk level label; the trajectory prediction module is used for constructing an insulation state trajectory prediction model and generating an insulation resistance state change trajectory according to the risk level label; and the early warning decision-making module is used for generating a grading response instruction through a dynamic early warning decision-making algorithm based on the insulation resistance state change track, and obtaining an operation and maintenance response strategy. Molecular-level micro-vibration is accurately captured through the quantum displacement acceleration vector, and feature recognition of the electrochemical corrosion incubation period is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment state detection, and in particular to a charging pile insulation impedance self-detection and early warning system. BACKGROUND

[0002] With the progress of charging pile insulation state monitoring technology, the correlation modeling of microscopic aging state vector and macroscopic insulation performance is gradually mature. The feature frequency point displacement acceleration calculation combined with the dielectric loss gradient generates the risk coupling strength tensor, which promotes the development of risk level labels to dynamic and fine. The construction of the insulation state trajectory prediction model realizes the cross-scale deduction from the risk level label to the insulation resistance state change trajectory. The risk mutation time positioning technology based on trajectory combined with quantum state encoding generates a hierarchical response instruction, which marks the transformation of the early warning decision mechanism from rule-driven to intelligent-driven, forming a closed-loop system of detection-early warning-decision.

[0003] The current charging pile insulation state monitoring technology still faces two major bottlenecks. In the aspect of sudden risk identification, the traditional threshold alarm method relies on fixed rules and is difficult to effectively capture the nonlinear mutation characteristics of the insulation state. Especially in complex working conditions such as sudden changes in temperature and humidity, millisecond-level resistance steep drop phenomena often cause missed reports due to response delay; in the field of cross-scale state modeling, the correlation deduction of microscopic aging state and macroscopic insulation performance still has theoretical limitations. The classical physical model simplifies the influence of quantum tunneling effect on feature frequency point displacement acceleration, resulting in a deviation in the mapping relationship between dielectric loss gradient and risk coupling strength tensor. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a charging pile insulation impedance self-detection and early warning system to solve the problems of static threshold being difficult to capture dynamic mutation risk and cross-scale modeling distortion.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] The present application provides a charging pile insulation impedance self-detection and early warning system, which comprises,

[0008] The frequency domain processing module collects impedance response data and generates frequency domain feature data through preprocessing; the vector construction module extracts feature frequency point displacement and dielectric loss gradient based on the frequency domain feature data, and constructs a material micro aging state vector; the strength analysis module calculates feature frequency point displacement acceleration based on the material micro aging state vector, obtains a risk coupling strength tensor by combining the dielectric loss gradient, and generates a risk level label; the trajectory prediction module constructs an insulation state trajectory prediction model, and generates an insulation resistance state change trajectory according to the risk level label; and the early warning decision module generates a graded response instruction by a dynamic early warning decision algorithm based on the insulation resistance state change trajectory, and obtains an operation and maintenance response strategy.

[0009] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the impedance response data includes insulation resistance value, capacitance value, frequency domain amplitude and phase information.

[0010] The preprocessing includes signal amplification, filtering and noise reduction, and analog-to-digital conversion.

[0011] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the frequency domain processing module includes the following modules:

[0012] The vector construction module extracts feature frequency point displacement and dielectric loss gradient based on the frequency domain feature data, and the specific steps are as follows:

[0013] The displacement trajectory of the feature frequency point is captured in real time based on the frequency domain response data, and feature frequency point displacement data is generated.

[0014] The feature frequency point displacement is input into a displacement loss field coupling equation to calculate the dielectric loss gradient.

[0015] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the construction of the material micro aging state vector means that the feature frequency point displacement and the dielectric loss gradient are fused to generate a multi-dimensional fusion feature tensor, and the material micro aging state vector is constructed through an optical projection algorithm.

[0016] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the strength analysis module calculates feature frequency point displacement acceleration based on the material micro aging state vector, and the specific steps are as follows:

[0017] The feature frequency point displacement acceleration is calculated based on the material micro aging state vector to generate a quantum displacement acceleration vector.

[0018] The quantum displacement acceleration vector and the dielectric loss gradient are input into a nonlinear coupling equation to calculate the risk coupling strength tensor.

[0019] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the risk level label is generated by mapping the risk coupling strength tensor through the three-dimensional grid coordinates to generate a risk intensity distribution map, and generating a risk level label through phase change feature analysis.

[0020] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the insulation state trajectory prediction model is constructed, and the specific steps are as follows,

[0021] The risk type in the risk level label is extracted, and a time dimension prediction framework is constructed by binding a time stamp;

[0022] Based on the time dimension prediction framework, a multi-source fusion training tensor is generated by combining frequency domain feature data;

[0023] Based on the multi-source fusion training tensor, a coupling coefficient matrix is generated through a quantum annealing algorithm, and an insulation state trajectory prediction model is constructed in parallel through quantum tunneling.

[0024] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the insulation resistance state change trajectory is generated according to the risk level label, which means that the risk level label is input into the insulation state trajectory prediction model, the evolution path of the micro aging state vector is generated through the quantum tunneling parallel algorithm, and the insulation resistance state change trajectory is output through coordinate conversion.

[0025] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, based on the insulation resistance state change trajectory, a hierarchical response instruction is generated through a dynamic early warning decision algorithm, and the specific steps are as follows,

[0026] Based on the insulation resistance state change trajectory, the risk mutation time is located, and a risk inflection point data set is obtained;

[0027] According to the risk inflection point data set, a risk intensity distribution field is constructed to obtain a dynamic risk intensity tensor, and a hierarchical response instruction is generated through quantum bit entanglement state coding.

[0028] As a preferred scheme of the charging pile insulation impedance self-detection and early warning system, the operation and maintenance response strategy is obtained by combining the hierarchical response instruction and the insulation resistance state change trajectory, and an operation and maintenance response strategy is output through a non-equilibrium state statistical optimization algorithm.

[0029] The present application has the advantages that: the quantum displacement acceleration vector accurately captures the molecular level micro-vibration, dynamically associates the characteristic frequency point displacement with the quantum tunneling effect, realizes the feature recognition of the electrochemical corrosion incubation period; at the same time, the quantum entanglement state coding decision maintains the consistency of the instruction through coherence, realizes the low delay response of the resistance steep drop, and suppresses the interference of temperature and humidity fluctuations. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a schematic diagram of a charging pile insulation impedance self-detection and early warning system.

[0032] Figure 2 A flowchart for constructing the microscopic aging state vector of a material.

[0033] Figure 3 A flowchart for generating the trajectory of insulation resistance state changes.

[0034] Figure 4 The flowchart for executing the dynamic early warning decision algorithm. Detailed Implementation

[0035] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0036] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0037] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0038] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a charging pile insulation impedance self-detection and early warning system, including the following steps: like Figure 1 The aforementioned charging pile insulation impedance self-detection and early warning system includes: a frequency domain processing module, a vector construction module, an intensity analysis module, a trajectory prediction module, and an early warning decision module.

[0039] The frequency domain processing module acquires impedance response data and generates frequency domain characteristic data through preprocessing.

[0040] The impedance response data includes insulation resistance value, capacitance value, frequency domain amplitude and phase information.

[0041] It should be noted that the insulation resistance value is directly collected by a high-precision sensor from the original resistance measurement value between the DC bus of the charging pile and the ground, which represents the ability of the insulation material to prevent current leakage. The higher the insulation resistance value, the better the insulation performance. As a basic parameter, it evaluates the overall health status of the insulation material and detects potential aging or defect risks.

[0042] The capacitance value is the insulation material capacitance characteristic data measured by the sensor in real time, which reflects the dielectric performance of the insulation material storing electric charge. The capacitance value is related to the polarization response of the material, and it is used to assist in analyzing the dynamic behavior of the insulation material under alternating electric field and providing input for dielectric loss calculation.

[0043] The frequency domain amplitude is the impedance amplitude value in the frequency domain feature data generated by the frequency domain sensor, which represents the size and intensity of the impedance signal at different frequencies. The frequency domain amplitude reflects the frequency response characteristics, and it is used to identify the displacement of the key feature frequency point and support the extraction of the dielectric loss gradient.

[0044] The phase information is also part of the impedance response data, which is obtained by converting the phase signal collected by the phase sensor. It represents the phase difference angle between the voltage and current waveforms. The phase signal reveals the phase shift of the signal in the frequency domain, and it is used to quantify the dielectric loss and correlate the material micro aging state. For example, abnormal phase angle can indicate the incubation period of electrochemical corrosion.

[0045] The preprocessing includes signal amplification, filtering and noise reduction, and analog-to-digital conversion.

[0046] It should be noted that an instrument amplifier is used to perform gain processing on the impedance response data to enhance the amplitude of weak signals. The instrument amplifier amplifies the current signal to a volt-level effective signal to ensure the signal strength in the subsequent processing steps.

[0047] Filtering and noise reduction are applied to separate the noise from the amplified impedance response data. The filtering reference is set for power frequency interference and high frequency white noise respectively to achieve signal-to-noise separation effect with interference attenuation greater than the filtering reference.

[0048] The filtered impedance response data is digitized by an analog-to-digital converter to generate a discrete digital signal sequence.

[0049] The frequency domain feature data includes impedance amplitude, phase angle and feature frequency point.

[0050] Specifically, the impedance amplitude in the frequency domain feature data is derived from the processing result of performing fast Fourier transform on the preprocessed frequency domain signal, representing the size intensity of the impedance signal in the frequency domain, and the impedance amplitude embodies the amplitude variation characteristics of the impedance at different frequencies, which serves as the basis data for feature frequency point identification and supports the subsequent capture of displacement trajectory.

[0051] The phase angle is also generated from the phase signal through fast Fourier transform, representing the angle difference between the voltage and current waveforms, and the phase signal reveals the phase shift degree in the frequency domain, which serves to quantify the dielectric loss behavior and associate the material aging state, for example, abnormal phase angle can indicate the risk of electrochemical corrosion.

[0052] The feature frequency point is the processing result of locating the extreme point in the curve composed of the phase angle, and the feature frequency point is the frequency coordinate at a specific frequency, which serves as the anchor point for displacement trajectory analysis and is used to calculate displacement acceleration and construct the aging state vector.

[0053] The vector construction module extracts the feature frequency point displacement and dielectric loss gradient based on the frequency domain feature data, and constructs the material micro-aging state vector;

[0054] Based on the frequency domain response data, the displacement trajectory of the feature frequency point is captured in real time, and the feature frequency point displacement data is generated;

[0055] It should be noted that the phase angle in the frequency domain feature data is screened; for each phase angle, the amplitude comparison is performed, comparing the phase angle of the current frequency point with the phase angles of the left and right adjacent frequency points, if the phase angle value of the current frequency point is greater than the phase angle values of the left and right adjacent frequency points at the same time, it is marked as a local maximum point, if it is smaller than the phase angle values of the left and right adjacent frequency points at the same time, it is marked as a local minimum point; noise suppression processing is performed on the marked local maximum and minimum points, and the frequency coordinate sequence of the feature frequency point is output.

[0056] For each feature frequency point frequency coordinate, a synchronous time stamp of data acquisition is bound to form a time-frequency coordinate pair sequence. Based on the feature frequency point frequency coordinates of adjacent time stamps, the difference between the feature frequency point frequency coordinates of the previous time stamp and the feature frequency point frequency coordinates of the subsequent time stamp is taken as the time stamp-frequency offset, and the feature frequency point displacement data is generated.

[0057] The feature frequency point displacement is input into the displacement loss field coupling equation to calculate the dielectric loss gradient;

[0058] It should be noted that the frequency offset in the feature frequency point displacement data is extracted; the displacement loss field coupling equation is called, the frequency offset is substituted into the displacement loss field coupling equation to perform algebraic operation to calculate the dielectric loss gradient; the time stamp-dielectric loss gradient value sequence is output, and the dielectric loss gradient is generated by identifying the inflection point of the time stamp-dielectric loss gradient value sequence.

[0059] Further, the physical meaning of the displacement loss field coupling equation is to quantify the mapping relationship between the characteristic frequency point displacement and the dielectric loss gradient. The characteristic frequency point displacement represents the characteristic frequency shift caused by the micro-vibration of the molecular chain of the insulating material. The dielectric loss gradient reflects the spatial decay rate of the dielectric performance of the material. The structure of the loss field coupling equation reflects the linear positive correlation between the dielectric loss gradient and the characteristic frequency point displacement. The loss field coupling equation works to realize the cross-scale conversion from the micro-molecular vibration to the dielectric performance index, highlights the potential risk of small displacement through linear amplification effect, and supports the early identification of the incubation period of electrochemical corrosion.

[0060] The expression for calculating the dielectric loss gradient is:

[0061] ;

[0062] wherein, is the dielectric loss gradient, is the molecular chain polarization rate, is the relative dielectric constant, is the dielectric relaxation time, is the electronic transition energy threshold, is the electronic transition energy value, is the characteristic frequency point displacement, is the time interval, is the natural logarithm function.

[0063] The electronic transition energy threshold is based on the core parameters of the quantum effect of the insulating material. For example, the electronic transition energy threshold of epoxy resin is in the range of , the electronic transition energy threshold of silicone rubber is in the range of , the electronic transition energy threshold of high-temperature ceramics is in the range of , and the electronic transition energy threshold of polytetrafluoroethylene is in the range of .

[0064] More preferably, the displacement loss field coupling equation is a core mathematical model connecting the quantum tunneling vibration of the micro-molecular chain and the decay of the macroscopic dielectric performance. The physical nature of the displacement loss field coupling equation is a nonlinear mapping process from the quantified characteristic frequency point displacement to the dielectric loss gradient. The displacement loss field coupling equation realizes the spatial transmission of molecular-level micro-vibration to the dielectric performance of the insulating material through the displacement loss field coupling mechanism, and works to convert the frequency point drift signal of quantum scale into a detectable dielectric gradient value, solving the cross-scale modeling distortion problem of traditional methods under temperature and humidity fluctuations. In specific operation, the displacement loss field coupling equation calculates the dielectric loss gradient in real time based on the characteristic frequency point displacement data, supports the construction of the material micro-aging state vector, and finally realizes the early warning of the incubation period of electrochemical corrosion, forming a technical closed loop of insulating state trajectory prediction and risk level label generation.

[0065] The characteristic frequency point displacement is combined with the dielectric loss gradient to generate a multi-dimensional fusion feature tensor, and a material micro-aging state vector is constructed through an optical projection algorithm, as shown in the following formula: Figure 2 The characteristic frequency point displacement is combined with the dielectric loss gradient to generate a multi-dimensional fusion feature tensor, and a material micro-aging state vector is constructed through an optical projection algorithm, as shown in the following formula:

[0066] It should be noted that the characteristic frequency point displacement and the dielectric loss gradient are aligned with the same timestamp, the frequency offset in the characteristic frequency point displacement and the dielectric loss gradient value in the dielectric loss gradient are extracted, the matching timestamp, frequency offset and dielectric loss gradient value are integrated into a single data point (timestamp, frequency offset and dielectric loss gradient value) according to a three-dimensional structure, and all single data points are arranged in time sequence, with the row dimension being the timestamp sequence and the column dimension including the timestamp, frequency offset and dielectric loss gradient value three data dimensions to construct a multi-dimensional fusion feature tensor.

[0067] The maximum value of the frequency offset, the maximum value of the timestamp and the maximum value of the dielectric loss gradient in the multi-dimensional fusion feature tensor are counted, the coordinate mapping is performed on each data point, the ratio of the frequency offset to the maximum value of the frequency offset is taken as the first-dimensional coordinate value (quantifying the molecular chain micro-vibration intensity), the ratio of the dielectric loss gradient value to the maximum value of the dielectric loss gradient value is taken as the second-dimensional coordinate value (representing the dielectric performance attenuation rate), and the ratio of the current timestamp to the maximum value of the timestamp is taken as the time attenuation factor to generate the third-dimensional coordinate value (revealing the time dimension aging process), and a three-dimensional coordinate vector is output to constitute a material micro-aging state vector.

[0068] Preferably, the optical projection algorithm is a mathematical transformation method for mapping high-dimensional feature data to a low-dimensional state vector, and the core of the optical projection algorithm is to realize the physical meaning fusion of multi-source heterogeneous data through coordinate normalization processing. The optical projection algorithm originates from the microstructure visualization demand in material science, and generates a fusion vector that quantifies the molecular chain micro-vibration intensity, dielectric performance attenuation rate and aging process at the same time by projecting the characteristic frequency point displacement, dielectric loss gradient and time attenuation factor in three-dimensional space.

[0069] The intensity analysis module calculates the characteristic frequency point displacement acceleration based on the material micro-aging state vector, obtains a risk coupling intensity tensor by combining the dielectric loss gradient, and generates a risk level label.

[0070] The feature frequency point displacement acceleration is calculated based on the material micro-aging state vector, and a quantum displacement acceleration vector is generated;

[0071] It should be noted that the first dimension coordinate value sequence and the corresponding time stamp sequence are extracted from the material micro-aging state vector, the adjacent time stamp difference operation is performed on the first dimension coordinate value sequence, and the difference between the first dimension coordinate value of the previous time stamp and the first dimension coordinate value of the subsequent time stamp is taken as the instantaneous speed sequence (time stamp-speed value sequence); the adjacent time stamp difference operation is performed on the instantaneous speed sequence, and the difference between the instantaneous speed of the previous time stamp and the instantaneous speed of the subsequent time stamp is taken as the feature frequency point displacement acceleration sequence (time stamp-acceleration value sequence); the displacement acceleration sequence is bound with the time stamp, and the quantum displacement acceleration vector is output.

[0072] The quantum displacement acceleration vector and the dielectric loss gradient are input into the nonlinear coupling equation to calculate the risk coupling strength tensor;

[0073] It should be noted that the data points of the same time stamp of the quantum displacement acceleration vector and the dielectric loss gradient are aligned, the acceleration value in the quantum displacement acceleration vector and the dielectric loss gradient value in the dielectric loss gradient are extracted; the acceleration value and the dielectric loss gradient value are substituted into the hyperbolic tangent operation of the nonlinear coupling equation, and the risk coupling strength value is output; the time stamp and the risk coupling strength value are bound to generate the risk coupling strength tensor.

[0074] The expression for calculating the risk coupling strength value is:

[0075]

[0076] Among them, is the risk coupling strength value, is the material temperature coefficient, is the phase angle offset, is the reference phase angle, is the quantum displacement acceleration value, is the reference acceleration, is the dielectric loss gradient, is the hyperbolic tangent function.

[0077] The value range of the material temperature coefficient is , which is derived from the frequency domain phase angle temperature drift experiment calibration combined with the temperature response characteristics of the molecular chain polarizability. For example, the material temperature coefficient of epoxy resin is in the range of , the material temperature coefficient of silicone rubber is in the range of , the material temperature coefficient of polytetrafluoroethylene is in the range of , and the material temperature coefficient of high-temperature ceramics is in the range of .

[0078] ​The nonlinear coupling equation is a core mathematical model for realizing dynamic correlation between quantum bit displacement acceleration and dielectric loss gradient in the early warning system. The nonlinear coupling equation has a physical nature of cross-scale fusion of molecular level micro-vibration signals and macroscopic dielectric attenuation behavior through the saturation nonlinear characteristics of the hyperbolic tangent function. The nonlinear coupling equation solves the dual pain points of "sudden risk missed report" and "cross-scale modeling distortion", and realizes adaptive compensation of environmental disturbance by introducing a material temperature coefficient.

[0079] The risk coupling intensity tensor is mapped through three-dimensional grid coordinates to generate a risk intensity distribution map, and a risk level label is generated through phase transition feature analysis.

[0080] It should be noted that a three-dimensional coordinate system is constructed, with the timestamp as the horizontal axis, the dielectric loss gradient value as the vertical axis, and the risk coupling intensity value as the vertical axis; the timestamp value, dielectric loss gradient value and risk coupling intensity value of each data point in the risk coupling intensity tensor are mapped into a three-dimensional coordinate system; the number and total value of risk coupling intensity of all coordinate points in each three-dimensional coordinate system are counted, and the ratio of the total value of risk coupling intensity to the number of all coordinate points in the grid area is taken as the average value of risk coupling intensity.

[0081] Taking the average value of risk coupling intensity as the benchmark, the risk level above the average value of risk coupling intensity is marked as high risk level, and the risk level below the average value of risk coupling intensity is marked as low risk level. The risk level of the corresponding grid area is bound for each timestamp to generate a timestamp-risk level sequence, and the risk level label is output.

[0082] The trajectory prediction module constructs an insulation state trajectory prediction model to generate an insulation resistance state change trajectory according to the risk level label.

[0083] The risk type in the risk level label is extracted, and a time dimension prediction framework is constructed by binding the timestamp.

[0084] It should be noted that the risk level label is obtained as input data, and each data point in the risk level label contains a timestamp and a risk level. When analyzing the risk level, if the risk level value is high risk level, it is marked as risk type code high; if the risk level value is low risk level, it is marked as risk type code low. The risk type code is bound with the corresponding timestamp, the original timestamp of the risk level label is kept unchanged, and the risk type code replaces the original risk level to generate a timestamp-risk type code sequence, which constitutes a time dimension prediction framework.

[0085] Based on the time dimension prediction framework, a multi-source fusion training tensor is generated in combination with frequency domain feature data.

[0086] It should be noted that the time dimension prediction framework is obtained with the frequency domain feature data; each timestamp in the time dimension prediction framework is traversed, and the data point of the same timestamp in the frequency domain feature data is located; the risk type code (risk type code high or risk type code low) in the corresponding time dimension prediction framework of the same timestamp and the impedance amplitude, phase angle and characteristic frequency point in the frequency domain feature data are extracted; the timestamp value, risk type code, impedance amplitude, phase angle and characteristic frequency point are integrated into five-dimensional data points; all five-dimensional data points are arranged in ascending order of timestamp, the row dimension is the timestamp sequence, and the column dimension includes five-dimensional data, and a multi-source fusion training tensor is constructed.

[0087] According to the multi-source fusion training tensor, an insulation state trajectory prediction model is constructed by a quantum annealing algorithm.

[0088] It should be noted that the risk type code, impedance amplitude, phase angle and characteristic frequency point of each data point in the multi-source fusion training tensor are analyzed; the quantum annealing energy function includes a quantum spin variable product term and a linear bias term.

[0089] The quantum annealing energy function is initialized to a high temperature state, and the quantum spin variable encodes the risk type code (for example, the risk type code high corresponds to the quantum spin variable positive one, and the risk type code low corresponds to the quantum spin variable negative one) under the constraint of the multi-source fusion training tensor data; the temperature of the quantum annealing energy function is decreased, and the temperature is linearly decreased from a high temperature state to a low temperature final state; the quantum tunneling effect is used to penetrate the energy barrier, and the quantum annealing energy function collapses to a low temperature ground state at the low temperature final state; when the quantum annealing energy function collapses to the low temperature ground state, a global minimum value is reached.

[0090] The quantum annealing energy function value reaches the global minimum value, and the optimal quantum spin variable configuration, the linear bias term and the interaction strength of the associated impedance amplitude, phase angle, characteristic frequency point and risk type code are output; the interaction strength of the associated impedance amplitude, phase angle, characteristic frequency point and risk type code is constructed into a coupling coefficient matrix; the coupling coefficient matrix and the linear bias term are integrated into a set of insulation state trajectory prediction model parameters, and an insulation state trajectory prediction model is constructed.

[0091] Preferably, the quantum annealing algorithm is derived from the solution of combinatorial optimization problems in the field of quantum computing, and is designed to handle the complex nonlinear correlation between frequency domain features and risk types. By constructing an energy function including a quantum spin variable coupling term and a linear bias term, a quantum annealing process from a high temperature initial state to a low temperature ground state is simulated, and an optimal solution of the coupling coefficient matrix is output.

[0092] The risk level label is input into the insulation state trajectory prediction model, an evolution path of a microscopic aging state vector is generated by a quantum tunneling parallel algorithm, and an insulation resistance state change trajectory is output through coordinate conversion, such as Figure 3The risk type in the extracted risk level label is bound with a time stamp to construct a time dimension prediction framework, and a multi-source fusion training tensor is generated in combination with frequency domain feature data; an insulation state trajectory prediction model is constructed based on the multi-source fusion training tensor through a quantum annealing algorithm; the risk level label is input into the insulation state trajectory prediction model, and an evolution path at all time steps is output through quantum parallel calculation, and an insulation resistance state change trajectory is output.

[0093] It should be noted that the risk level label time stamp risk level sequence is input into the insulation state trajectory prediction model to obtain a coupling coefficient matrix and a linear bias parameter set; the first group of values of the micro aging state vector are initialized through a quantum tunneling parallel algorithm, the micro aging state vector at the current time stamp is encoded into a quantum bit superposition state, so that a single quantum bit simultaneously represents multiple evolution states; the coupling coefficient matrix of the insulation state trajectory prediction model is called to construct an evolution Hamiltonian, so that the evolution paths at all time steps are synchronously output in the quantum superposition state, the first dimension coordinate value (molecular chain microvibration intensity) of the micro aging state vector corresponding to each time stamp in the evolution path is extracted, and a time stamp insulation resistance value sequence is output to form an insulation resistance state change trajectory.

[0094] Preferably, the insulation state trajectory prediction model is a core prediction model constructed based on a quantum annealing algorithm, which aims to generate a multi-source fusion training tensor by processing risk level labels, time stamps and frequency domain feature data, and deduce the evolution path of the micro aging state vector using a quantum tunneling parallel algorithm to realize future change trajectory prediction of the insulation resistance state; the insulation state trajectory prediction model means to correlate micro quantum effects and macro performance degradation across scales, solves the problems of sudden risk false negatives and modeling distortion, and functions to early warn resistance steep drop, drive graded response decisions, and optimize overall closed-loop performance.

[0095] The early warning decision module generates graded response instructions based on the insulation resistance state change trajectory through a dynamic early warning decision algorithm to obtain an operation and maintenance response strategy.

[0096] Based on the insulation resistance state change trajectory, the risk mutation time is located, and a risk inflection point data set is obtained.

[0097] It should be noted that the insulation resistance value corresponding to each time stamp in the insulation resistance state change trajectory is analyzed; the difference between the insulation resistance value at the next time stamp and the insulation resistance value at the previous time stamp is divided by the time interval to obtain the differential change rate of the insulation resistance value between adjacent time stamps.

[0098] arranging the absolute value of the current differential change rate in ascending order based on the statistical distribution characteristics of the current differential change rate sequence itself; positioning the risk mutation moment as a dynamic boundary value, comparing the current differential change rate with the dynamic boundary value, if the current differential change rate is greater than the dynamic boundary value, marking the corresponding time stamp as the risk mutation moment, if the current differential change rate is less than the dynamic boundary value, not marking; extracting the time stamp and insulation resistance value of the risk mutation moment; binding the time stamp and the insulation resistance value to form a time stamp-insulation resistance value sequence to constitute a risk inflection point data set.

[0099] According to the risk inflection point data set, a risk intensity distribution field is constructed to obtain a dynamic risk intensity tensor, and a hierarchical response instruction is generated through quantum bit entanglement state coding, such as Figure 4 Based on the insulation resistance state change trajectory, the risk mutation moment is located, and the risk inflection point data set is obtained; according to the risk inflection point data set, a risk intensity distribution field is constructed to obtain a dynamic risk intensity tensor, and a hierarchical response instruction is generated through quantum bit entanglement state coding: primary response executes continuous monitoring followed by increasing monitoring frequency, intermediate response executes local maintenance to reinforce the material, and highest level response executes shutdown maintenance to replace the insulation module.

[0100] It should be noted that the time stamp in the risk inflection point data set is taken as the time coordinate axis, and the insulation resistance value is taken as the risk intensity value to construct a risk intensity distribution field through spatial interpolation technology, and output a continuous matrix of time stamp risk intensity value;

[0101] Extracting the risk intensity value corresponding to each time stamp in the risk intensity value continuous matrix, arranging to generate a dynamic risk intensity tensor in time sequence; traversing each risk intensity value in the dynamic risk intensity tensor to execute quantum bit entanglement state coding operation, normalizing the risk intensity value, mapping the normalized risk intensity value to quantum state probability amplitude (ground state probability amplitude is negatively correlated with normalized risk intensity value, excited state probability amplitude is positively correlated with normalized risk intensity value), applying quantum superposition to obtain a binary string; converting the binary string into an instruction code, outputting a hierarchical response instruction (the instruction code corresponds to primary response continuous monitoring, the instruction code corresponds to intermediate response maintenance within a specific time, and the instruction code corresponds to highest level response immediate shutdown maintenance).

[0102] Combined with the hierarchical response instruction and the insulation resistance state change trajectory, an operation and maintenance response strategy is output through a non-equilibrium state statistical optimization algorithm.

[0103] It should be noted that when the insulation resistance state change trajectory presents a smooth oscillation, the primary response instruction is executed, corresponding to increasing the monitoring frequency and starting the auxiliary diagnosis strategy set; when the insulation resistance state change trajectory presents a continuous inflection point, the intermediate response instruction is executed, corresponding to the local component replacement and material reinforcement processing strategy set; when the insulation resistance state change trajectory presents a cliff-like drop, the highest level response instruction is executed, corresponding to the whole machine shutdown maintenance and replacement of the insulation module strategy set; the candidate strategy set is obtained;

[0104] An initial strategy solution in the candidate strategy set is randomly selected, and a neighboring strategy solution is generated by fine-tuning the strategy parameters in the initial strategy solution; according to the performance change (such as risk suppression effect or execution efficiency improvement) of the neighboring strategy solution, an operation and maintenance response strategy is output.

[0105] Further, the operation and maintenance response strategy includes a primary response strategy, an intermediate response strategy and a highest level response strategy; the primary response strategy executes continuous monitoring and auxiliary diagnosis, generates real-time frequency domain feature data by collecting impedance response data at a benchmark frequency, activates a quantum tunneling parallel algorithm to track the evolution path of the material micro-aging state vector, dynamically determines the characteristic frequency point displacement acceleration and the dielectric loss gradient, and maintains monitoring or triggers upgrading.

[0106] The intermediate response strategy executes local component replacement and material reinforcement, locates the risk mutation time space coordinates according to the insulation resistance state change trajectory, injects reinforcement materials in the abnormal area of the dielectric loss gradient, and verifies the characteristic frequency point displacement regression benchmark.

[0107] The highest level response strategy executes whole machine shutdown maintenance, cuts off the main power supply of the charging pile, comprehensively scans the risk coupling strength tensor and compares the insulation state trajectory prediction model output, replaces the whole set of insulation materials when the quantum displacement acceleration vector is continuously dangerous, and finally ensures that the characteristic frequency point displacement acceleration is zero through full-band impedance scanning.

[0108] Preferably, the dynamic early warning decision algorithm is the core decision of the charging pile insulation impedance early warning system, and is specially used for converting the insulation resistance state change trajectory into a hierarchical response instruction. The technical meaning of the dynamic early warning decision algorithm is to realize the recognition of the risk mutation time and the dynamic optimization of the response strategy through the quantum bit entangled state encoding mechanism.

[0109] In summary, the present application realizes the feature recognition of the electrochemical corrosion incubation period by: the quantum displacement acceleration vector accurately captures the molecular level micro-vibration, dynamically associates the characteristic frequency point displacement with the quantum tunneling effect, and realizes the feature recognition of the electrochemical corrosion incubation period; at the same time, the quantum entangled state encoding decision realizes the low delay response of the resistance steep drop through the coherence maintenance instruction consistency, and suppresses the temperature and humidity fluctuation interference.

[0110] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A charging pile insulation impedance self-detection and early warning system, characterized in that: comprising, a frequency domain processing module that collects impedance response data and generates frequency domain feature data through preprocessing; a vector construction module that extracts feature frequency point displacement and dielectric loss gradient based on the frequency domain feature data and constructs a material micro-aging state vector; a strength analysis module that calculates feature frequency point displacement acceleration based on the material micro-aging state vector, obtains a risk coupling strength tensor by combining the dielectric loss gradient, and generates a risk level label; a trajectory prediction module that constructs an insulation state trajectory prediction model and generates an insulation resistance state change trajectory according to the risk level label; an early warning decision module that generates a graded response instruction through a dynamic early warning decision algorithm based on the insulation resistance state change trajectory and obtains an operation and maintenance response strategy.

2. The charging pile insulation impedance self-detecting and early warning system according to claim 1, characterized in that: The impedance response data includes insulation resistance value, capacitance value, frequency domain amplitude, and phase information. The preprocessing includes signal amplification, filtering and noise reduction, and analog-to-digital conversion.

3. The charging pile insulation impedance self-detection and early warning system of claim 2, wherein: The feature frequency point displacement and dielectric loss gradient are extracted based on the frequency domain feature data, and the specific steps are as follows, Real-time capture the displacement trajectory of the feature frequency point based on the frequency domain response data, and generate feature frequency point displacement data; Input the feature frequency point displacement into the displacement loss field coupling equation to calculate the dielectric loss gradient.

4. The charging pile insulation impedance self-detecting and early warning system of claim 3, wherein: The construction of the material micro-aging state vector refers to the fusion of the feature frequency point displacement and the dielectric loss gradient to generate a multi-dimensional fusion feature tensor, and the construction of the material micro-aging state vector through an optical projection algorithm.

5. The charging station insulation impedance self-detecting and pre-warning system of claim 4, wherein: Based on the material micro-aging state vector, the feature frequency point displacement acceleration is calculated, and the risk coupling strength tensor is obtained by combining the dielectric loss gradient, and the specific steps are as follows, Calculate the feature frequency point displacement acceleration based on the material micro-aging state vector to generate a quantum displacement acceleration vector; Input the quantum displacement acceleration vector and the dielectric loss gradient into the nonlinear coupling equation to calculate the risk coupling strength tensor.

6. The charging station insulation impedance self-detecting and pre-warning system of claim 5, wherein: The generation of the risk level label refers to mapping the risk coupling strength tensor through a three-dimensional grid coordinate to generate a risk intensity distribution map, and generating a risk level label through phase change feature analysis.

7. The charging station insulation impedance self-detecting and pre-warning system of claim 6, wherein: The specific steps of constructing the insulation state trajectory prediction model are as follows, Extract the risk type in the risk level label, and construct a time dimension prediction framework by binding a time stamp; Based on the time dimension prediction framework, generate a multi-source fusion training tensor by combining the frequency domain feature data; Based on the multi-source fusion training tensor, generate a coupling coefficient matrix through a quantum annealing algorithm, and construct an insulation state trajectory prediction model through quantum tunneling parallel.

8. The charging station insulation impedance self-detecting and pre-warning system of claim 7, wherein: The generation of the insulation resistance state change trajectory according to the risk level label refers to inputting the risk level label into the insulation state trajectory prediction model, generating the evolution path of the micro-aging state vector through the quantum tunneling parallel algorithm, and outputting the insulation resistance state change trajectory through coordinate conversion.

9. The charging station insulation impedance self-detecting and pre-warning system of claim 8, wherein: The specific steps of generating a graded response instruction through a dynamic early warning decision algorithm based on the insulation resistance state change trajectory are as follows, Locate the risk mutation time based on the insulation resistance state change trajectory to obtain a risk inflection point data set; According to the risk inflection point data set, a risk intensity distribution field is constructed to obtain a dynamic risk intensity tensor, and a hierarchical response instruction is generated through quantum bit entanglement state coding.

10. The charging station insulation impedance self-detecting and pre-warning system of claim 9, wherein: The obtaining of the operation and maintenance response strategy is that, in combination with the hierarchical response instruction and the insulation resistance state change trajectory, an operation and maintenance response strategy is output through a non-equilibrium state statistical optimization algorithm.

Images