An intelligent fault prediction method and system for an electric vehicle charging pile

Abstract

The present invention discloses an intelligent fault prediction method and system for an electric vehicle charging pile, relating to the technical field of intelligent fault prediction of charging piles, and comprising the following steps: quantifying the output of a Hall sensor for the sensitivity during historical magnetic flux detection; using the outputs of a plurality of different Hall sensors as a plurality of magnetic flux detection sensitivities; obtaining interference data of current load changes on the magnetic field and detection quality data; and constructing a Hall sensor output evaluation model based on the bp neural network algorithm; corresponding the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor to construct a two-dimensional display model, and performing data analysis on the two-dimensional display model to obtain an improved Hall sensor output, which is used as the magnetic flux detection sensitivity for charging pile cable fault prediction; solving the problem that in the existing cable fault detection by magnetic flux, the accuracy of fault detection and the false alarm rate cannot be effectively balanced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent fault prediction for charging piles. More specifically, the present invention relates to an intelligent fault prediction method and system for electric vehicle charging piles. Background Art

[0002] With the popularization of electric vehicles, as a key part of the charging infrastructure, the stability of the operation of charging piles and the importance of fault prediction have become more prominent. However, due to the complex charging pile system, which includes multiple components such as power transmission, cable connection, and charging guns, and these components need to operate stably for a long time, the fault prediction and detection of charging piles are particularly important.

[0003] Modern intelligent fault prediction technology for charging piles mainly realizes the early identification of faults through the real-time collection and analysis of various sensor data. For example, in the characteristic parameters of charging equipment, the state of the power supply equipment characteristic parameters and the auxiliary power supply equipment characteristic parameters can be monitored through voltage acquisition and voltage detection circuits. By analyzing the voltage fluctuations and abnormal conditions, the possible fault risks of the equipment characteristic parameters can be identified. For the charging gun and gun holder, the state sensor or the change of resistance can be used to judge whether the charging gun is in a normal connection state, so as to realize the real-time detection of the interface part.

[0004] For charging pile cables, intelligent fault prediction technology can also use magnetic flux detection and image analysis technology to perform multi-level detection on the surface defects and internal structures of the cables. For example, magnetic flux detection is used to identify whether there are abnormal magnetic field changes in the cable to judge whether there are physical defects. At the same time, combined with image recognition technology to detect defects in the cable surface or cross-sectional image, it helps to quickly locate the fault location and evaluate its severity. This intelligent detection method can not only improve the accuracy of fault detection but also timely discover and prevent potential safety hazards.

[0005] For example, the charging pile, charging pile equipment characteristic parameter detection method and device announced in the invention patent announcement with the announcement number of: CN114113833B, including a charger circuit, including: a charging circuit, a first voltage acquisition circuit connected between the characteristic parameters of the DC power supply device, a second voltage acquisition circuit connected between the characteristic parameters of the auxiliary power supply device, and a first voltage detection circuit connected to the first charging connection confirmation device characteristic parameter CC1; a charging gun, connected to the charger circuit; a gun holder for placing the charging gun, including: a travel switch for detecting the placement state of the charging gun, a variable resistor connected between the device characteristic parameter CC1 and the ground wire, a third voltage acquisition circuit connected between the characteristic parameters of the DC power supply device and the auxiliary power supply device, a second voltage detection circuit connected to the second charging connection confirmation device characteristic parameter CC2, and a third voltage detection circuit connected to the ground wire; a controller, connected to the charger circuit and the gun holder, for controlling the charging circuit. This application solves the technical problems that the process of detecting faults in the equipment characteristic parameters of the charging pile is complex and the user experience is poor in the related art.

[0006] For example, the fault detection method, device, equipment and storage medium for cable production announced in the invention patent announcement with the announcement number of: CN115144704A, including improving the accuracy of fault detection in cable production and the safety of the cable. The method includes: if the root cause analysis result is that the charging pile cable is abnormal, then perform magnetic flux detection on the charging pile cable to obtain a magnetic flux detection result; if there are defects in the charging pile cable, collect multiple cable surface images corresponding to the charging pile cable, and perform cable surface defect detection on the charging pile cable based on the multiple cable surface images to obtain a surface defect detection result; perform fault location on the charging pile cable according to the surface defect detection result to obtain a fault location result, and intercept the faulty cable corresponding to the charging pile cable according to the fault location result and collect a cross-sectional image of the faulty cable; input the cross-sectional image into a preset cable fault detection model for cable fault detection to obtain a cable fault detection result.

[0007] In the above disclosed technical solutions, there are at least the following technical problems: In the process of cable fault analysis, there is a lack of an effective adjustment mechanism for setting the magnetic flux detection sensitivity of the cable; resulting in insufficient or excessive adjustment of the sensitivity in different application scenarios.

[0008] On the one hand, too low magnetic flux detection sensitivity may not be able to capture weak fault signals in the cable, missing the opportunity for early fault diagnosis and affecting the timeliness of the warning system;

[0009] On the other hand, too high magnetic flux detection sensitivity may lead to excessive response to current fluctuations in the normal working state, generating too many false alarms or noise interferences, thus reducing the reliability and accuracy of the system.

[0010] In the existing cable fault detection through magnetic flux, it is impossible to effectively balance the accuracy of fault detection and the false alarm rate, which affects the effect of cable fault diagnosis and the user experience. In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0011] To overcome the above defects of the prior art, an embodiment of the present invention provides an intelligent fault prediction method and system for an electric vehicle charging pile. By continuously adjusting the gain of the Hall sensor, the fault detection effect under different magnetic flux detection sensitivities is analyzed, so as to solve the problem that in the existing cable fault detection through magnetic flux, it is impossible to effectively balance the accuracy of fault detection and the false alarm rate, which affects the effect of cable fault diagnosis and the user experience.

[0012] To achieve the above object, the present invention provides the following technical solutions:

[0013] An intelligent fault prediction method for an electric vehicle charging pile includes the following steps: performing magnetic flux detection on the charging pile cable; if the magnetic flux detection result indicates that there are potential hazards in the charging pile cable, collecting the cable surface image, performing fault location and analyzing the cross-sectional image of the cable at the located fault position, so as to perform fault prediction; quantifying the sensitivity during historical magnetic flux detection through the output of the Hall sensor, and averaging to obtain the initial magnetic flux detection sensitivity; performing a random number of Hall sensor gains at the initial magnetic flux detection sensitivity to obtain the outputs of a number of different Hall sensors as a number of magnetic flux detection sensitivities; obtaining the interference data of the current load change on the magnetic field and the detection quality data of the charging pile cable fault prediction under a number of magnetic flux detection sensitivities; constructing a Hall sensor output evaluation model based on the interference data of the current load change on the magnetic field and the detection quality data by using the bp neural network algorithm; corresponding the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor, constructing a two-dimensional display model, and performing data analysis on the two-dimensional display model to obtain an improved Hall sensor output, which is used as the magnetic flux detection sensitivity for the charging pile cable fault prediction.

[0014] In a preferred embodiment, the magnetic flux detection of the charging pile cable is specifically: using a Hall sensor to perform real-time detection on the magnetic field around the charging pile cable to obtain the magnetic flux data of the cable in the working state; comparing the collected magnetic flux data with the reference magnetic flux in the normal working state to identify the deviation of the magnetic field; analyzing the amplitude and frequency of the magnetic flux change, constructing a change curve, analyzing the mutation and periodic fluctuation of the change curve, and comparing with a preset safety range. When the preset safety range is exceeded, it indicates that there are potential hazards in the charging pile cable.

[0015] In a preferred embodiment, the sensitivity during the detection of historical magnetic flux is quantified through the output of a Hall sensor, and the initial magnetic flux detection sensitivity is obtained by averaging, specifically as follows: Obtain the historical magnetic flux data of the Hall sensor at different time points from the historical record; Quantify the historical magnetic flux data at different time points by calculating the change amount per unit time of the magnetic flux value at each data point; Calculate the average of the change amounts per unit time of all historical magnetic flux detection data as the initial magnetic flux detection sensitivity.

[0016] In a preferred embodiment, a number of random Hall sensor gains are performed at the initial magnetic flux detection sensitivity, and the outputs of a number of different Hall sensors are obtained as a number of magnetic flux detection sensitivities, specifically as follows: According to the historical data, obtain the gain range of the Hall sensor, and randomly obtain a number of gain values within the gain range of the Hall sensor based on computer random numbers; Perturb the initial magnetic flux detection sensitivity with the number of gain values to obtain the outputs of a number of different Hall sensors as a number of magnetic flux detection sensitivities.

[0017] In a preferred embodiment, the detection quality data includes a detection response fluctuation coefficient, and the specific method for obtaining the detection response fluctuation coefficient is as follows: Obtain the signal fluctuation amplitudes during the magnetic flux detection of the charging pile cable; Obtain the signal fluctuation amplitude standard deviation and the signal fluctuation amplitude average value of the signal fluctuation amplitudes at the same time interval during the magnetic flux detection of the charging pile cable; Calculate the signal fluctuation amplitude variation coefficient according to the signal fluctuation amplitude standard deviation and the signal fluctuation amplitude average value; Calculate the detection response fluctuation coefficient based on the signal fluctuation amplitude variation coefficient using a preset detection response fluctuation coefficient calculation formula.

[0018] In a preferred embodiment, the interference data of the current load on the magnetic field includes a current load influence coefficient and a magnetic flux misdetection anomaly coefficient. The specific method for obtaining the current load influence coefficient is as follows: Obtain the waveforms of the magnetic induction intensity vector and the current load signal changing with time; Obtain the harmonic compositions of the magnetic induction intensity vector signal waveform and the current load signal waveform respectively at a preset frequency, and perform harmonic component analysis to obtain the basic waveform components; Convert the basic waveform components to harmonic components based on the inverse Fourier transform; And calculate the floating value of the harmonic according to the quotient of the harmonic component and the basic waveform component; Perform data analysis on the floating value of the harmonic to calculate the current load influence coefficient.

[0019] In a preferred embodiment, the specific method for obtaining the magnetic flux misdetection anomaly coefficient is as follows: Multiply the device characteristic parameters obtained in the charging pile system by the corresponding influence coefficients to obtain the actual electromagnetic interference influence. The device characteristic parameters include current, power, and frequency; Calculate the actual electromagnetic interference influence of each device characteristic parameter according to the device characteristic parameters in combination with a preset electromagnetic interference source calculation formula; Calculate the electromagnetic interference influence in different frequency bands in combination with a preset interference frequency calculation formula; Determine the standard electromagnetic interference influence of various device characteristic parameters according to historical monitoring data; Compare and add the actual electromagnetic interference influence with the standard electromagnetic interference influence to obtain the magnetic flux misdetection anomaly coefficient.

[0020] In a preferred embodiment, corresponding the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor to construct a two-dimensional display model, specifically: Construct a two-dimensional array with the output of the Hall sensor output evaluation model and the corresponding magnetic flux detection sensitivity; Arrange the magnetic flux detection sensitivities of several two-dimensional arrays from largest to smallest as the horizontal axis, and the output of the corresponding Hall sensor output evaluation model as the vertical axis to construct a two-dimensional coordinate system in which the output of the Hall sensor output evaluation model changes with the magnetic flux detection sensitivity; Mark several two-dimensional arrays on the two-dimensional coordinate system and connect them with a smooth curve to construct a two-dimensional display model.

[0021] In a preferred embodiment, perform data analysis on the two-dimensional display model to obtain an improved Hall sensor output and apply it as the magnetic flux detection sensitivity to the charging pile cable fault prediction, specifically: According to the two-dimensional display model, obtain the magnetic flux detection sensitivities corresponding to several peaks; Calculate the second derivative of several peaks to judge the rate of change of the output of the Hall sensor output evaluation model with the magnetic flux detection sensitivity, and use the magnetic flux detection sensitivity corresponding to the peak with the smallest absolute value of the second derivative as the output of the improved Hall sensor.

[0022] The technical effects and advantages of the intelligent fault prediction method and system for an electric vehicle charging pile of the present invention:

[0023] 1. The present invention realizes intelligent fault prediction and hidden danger monitoring of charging pile cables by combining magnetic flux detection, image acquisition, and fault location. First, the system uses magnetic flux detection to identify potential cable problems, and locates and analyzes faults by collecting surface images and cross-sectional images of the cable to ensure accurate prediction of the fault occurrence location. Second, regarding the sensitivity of the Hall sensor, the detection accuracy of the sensor is further optimized by quantifying historical detection data and adjusting the gain. By analyzing the data of magnetic field interference caused by current load changes under different sensitivities, a Hall sensor output evaluation model is constructed based on the BP neural network algorithm, effectively improving the accuracy and robustness of fault prediction. This method not only improves the accuracy and reliability of charging pile cable fault detection, but also enables efficient fault diagnosis in complex environments, with significant advantages such as intelligence, strong adaptability, and strong anti-interference ability.

[0024] 2. The present invention can effectively evaluate and improve the output performance of the Hall sensor by constructing a Hall sensor output evaluation model based on the BP neural network algorithm. By considering multiple factors such as current load, magnetic field interference, and detection quality, a Hall sensor output evaluation coefficient is generated, and in-depth analysis of the data is carried out through a two-dimensional display model to find the optimal magnetic flux detection sensitivity. This method can improve the stability and accuracy of the Hall sensor under dynamic load and complex environments. Especially in the application of charging pile cable fault prediction, the improved sensor output can provide higher sensitivity and more accurate fault diagnosis, thus enhancing the reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic structural diagram of an intelligent fault prediction method for an electric vehicle charging pile according to the present invention.

[0026] Figure 2 It is a schematic structural diagram of an intelligent fault prediction system for an electric vehicle charging pile according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0028] Embodiment 1 Figure 1 An intelligent fault prediction method for an electric vehicle charging pile according to the present invention is given, including:

[0029] S1. Conduct magnetic flux detection on the charging pile cable. If the magnetic flux detection result indicates potential hazards in the charging pile cable, collect the cable surface image, perform fault location, and analyze the cross-sectional image of the cable at the located fault position to predict faults.

[0030] The magnetic flux detection of the charging pile cable is specifically as follows:

[0031] Use a Hall sensor to detect the magnetic field around the charging pile cable in real time and obtain the magnetic flux data of the cable in the working state.

[0032] Compare the collected magnetic flux data with the reference magnetic flux in the normal working state to identify the deviation of the magnetic field.

[0033] Analyze the amplitude and frequency of the magnetic flux change, construct a change curve, analyze the mutation and periodic fluctuation of the change curve, and compare it with the preset safety range. When it exceeds the preset safety range, it indicates potential hazards in the charging pile cable.

[0034] It should be noted that in the normal working state, the magnetic flux data of the charging pile cable will be collected and recorded in advance as a reference value. By comparing the currently collected magnetic flux data with the reference magnetic flux, it is possible to identify whether there is a deviation in the current magnetic field. The existence of a deviation may mean that there is an abnormal current distribution or a fault phenomenon around the cable. Perform time series analysis on the magnetic flux data to construct a magnetic flux change curve. By analyzing the fluctuation pattern of the curve, identify the amplitude and frequency characteristics of the magnetic flux change. If the change amplitude of the magnetic flux suddenly increases or shows abnormal periodic fluctuations, it may indicate that the cable has a fault or potential damage. Determine whether there are potential hazards by comparing the magnetic flux change curve with the preset safety range (such as the maximum fluctuation amplitude, frequency range, etc.). When the magnetic flux change exceeds the safety range, the system will issue an alarm, indicating that the cable may have a fault risk or damage and further detection or maintenance is required.

[0035] If the magnetic flux detection result indicates potential hazards in the charging pile cable, collect the cable surface image, perform fault location, and analyze the cross-sectional image of the cable at the located fault position to predict faults. The specific steps are as follows:

[0036] If the magnetic flux detection result indicates potential hazards in the charging pile cable, first collect images of the cable surface. Obtain the image data of the cable surface through high-precision imaging equipment for further analysis of potential surface defects or damages.

[0037] Analyze the collected cable surface image through image processing technology to detect whether there are cracks, charred marks, color changes, or other visible damages on the surface. These surface defects may be external manifestations of cable faults, indicating that there may be deeper problems.

[0038] Based on the results of surface defect detection, locate the cable fault area. Determine the specific location of the abnormality on the cable surface and provide accurate location data for the next fault analysis.

[0039] After locating the fault area, collect cross-sectional images of the cable at the fault location. Cross-sectional images can provide detailed information about the inside of the cable, such as insulation layer damage, conductor corrosion, or other structural problems.

[0040] Input the cross-sectional image into a pre-set cable fault detection model for further analysis to determine the type, location, and severity of the fault. Evaluate the possible causes of the fault through the model and predict the fault development trend, ultimately providing a basis for repair and maintenance decisions.

[0041] It should be noted that there are mature technologies for fault prediction through image analysis, which will not be elaborated here.

[0042] S2. Quantify the sensitivity during historical magnetic flux detection through the output of the Hall sensor and calculate the average to obtain the initial magnetic flux detection sensitivity.

[0043] The quantification of the sensitivity during historical magnetic flux detection through the output of the Hall sensor and the calculation of the average to obtain the initial magnetic flux detection sensitivity is specifically as follows:

[0044] Obtain the historical magnetic flux data of the Hall sensor at different time points from the historical records.

[0045] Perform quantization processing on the historical magnetic flux data at different time points by calculating the change amount per unit time of the magnetic flux value at each data point;

[0046] Calculate the average of the change amounts per unit time of all historical magnetic flux detection data as the initial magnetic flux detection sensitivity.

[0047] It should be noted that the magnetic flux data of the Hall sensor at different time points cover the magnetic field changes of the cable under various working conditions, including normal working and possible fault states; if necessary, sensitivity calibration and correction can be performed based on the magnetic flux data under known faults or cable abnormal states to ensure that the obtained initial sensitivity value has high representativeness and reliability.

[0048] S3. Conduct a number of random Hall sensor gains at the initial magnetic flux detection sensitivity to obtain the outputs of a number of different Hall sensors as a number of magnetic flux detection sensitivities.

[0049] Performing a number of random Hall sensor gains at the initial magnetic flux detection sensitivity to obtain the outputs of a number of different Hall sensors as a number of magnetic flux detection sensitivities, specifically:

[0050] According to historical data, obtain the gain range of the Hall sensor, and randomly obtain a number of gain values within the gain range of the Hall sensor based on computer random numbers;

[0051] Use a number of gain values to perturb the initial magnetic flux detection sensitivity to obtain the outputs of a number of different Hall sensors as a number of magnetic flux detection sensitivities.

[0052] The above method obtains the gain range of the Hall sensor based on historical data, and uses computer random numbers to generate a number of gain values for perturbation, thereby obtaining multiple different Hall sensor outputs. The advantages of this method are as follows: By introducing randomness, it can comprehensively cover the magnetic flux detection sensitivities under different gain configurations, avoiding the limitations that may be brought by a single gain value, and providing a richer output data set. These diverse output data can help analyze the fault detection performance under different gains, thus providing a more comprehensive basis for optimizing the Hall sensor gain selection, helping to improve the sensitivity and accuracy of the system, and reducing the risks of false alarms and missed alarms.

[0053] S4. Obtain the interference data of the current load change on the magnetic field and the detection quality data for the prediction of the charging pile cable fault under a number of magnetic flux detection sensitivities.

[0054] The interference data of the current load change on the magnetic field includes the current load influence coefficient and the magnetic flux misdetection anomaly coefficient;

[0055] The detection quality data includes the detection response fluctuation coefficient.

[0056] The current load influence coefficient is used to quantify the influence of the interference of the current load change on the magnetic field on the fault detection result, especially on the false alarm rate. Since during the operation of the charging pile, the fluctuation of the current load may introduce additional electromagnetic interference, and these interference signals may be misjudged as cable faults by the Hall sensor, resulting in false alarms. The current load influence coefficient helps to evaluate the degree of over-response of the system to noise under different sensitivity settings by measuring the correlation between the detection error of the magnetic flux and the actual fault under different current loads. This coefficient helps to optimize the magnetic flux detection sensitivity to reduce false alarms caused by current load fluctuations, thereby improving the accuracy and reliability of fault prediction.

[0057] Analyzing the influence of the current load influence coefficient on the fault prediction effect under different magnetic flux detection sensitivities has the following advantages in solving the problem that in the existing cable fault detection through magnetic flux, it is impossible to effectively balance the accuracy of fault detection and the false alarm rate:

[0058] Optimize sensitivity settings: By quantifying the impact of current load changes on magnetic field interference, the current load impact coefficient can help identify the overresponse of the magnetic flux detection system to electrical environmental noise at different sensitivities, thereby adjusting the sensitivity of the Hall sensor and optimizing the system's performance in the face of current load fluctuations.

[0059] Improve fault diagnosis accuracy: This coefficient can help distinguish real faults from false fault signals caused by current load, reduce false alarms, improve the accuracy of fault detection, and ensure more reliable cable fault prediction.

[0060] Reduce false alarm rate: Fluctuations in current load may cause interference signals in the magnetic field to be misjudged as faults. By evaluating the false alarm rate at different sensitivities, the current load impact coefficient can effectively prevent the system from overresponding to interference signals, reduce the occurrence of false alarms, and thus enhance the user experience.

[0061] Enhance system stability: By precisely adjusting the magnetic flux sensitivity and combining the evaluation of the current load impact coefficient, it is possible to maintain the stability of fault detection under different working environments and load changes, avoiding detection instability problems caused by current load changes.

[0062] Optimize user experience: Reducing the false alarm rate and improving the accuracy of fault prediction make the maintenance and management of charging piles more efficient, reduce unnecessary inspections and downtime, and enhance the overall user satisfaction of the system.

[0063] The specific method for obtaining the current load impact coefficient is as follows:

[0064] Obtain the waveforms of the magnetic induction intensity vector and the current load signal changing with time;

[0065] Obtain the harmonic compositions of the magnetic induction intensity vector signal waveform and the current load signal waveform at a preset frequency respectively, and conduct harmonic component analysis to obtain the basic waveform components;

[0066] Inverse Fourier transform the basic waveform components into harmonic components; and calculate the floating value of the harmonics according to the quotient of the harmonic components and the basic waveform components;

[0067] Conduct data analysis on the floating value of the harmonics and calculate to obtain the current load impact coefficient.

[0068] Among them, the specific calculation formula of the current load impact coefficient is as follows:

[0069]

[0070] In the formula, is the current load impact coefficient, is the number of harmonic compositions, is the current load data volume, z is the current load data volume, and j is the label of the harmonic composition. is the inverse Fourier transform. is the magnetic induction intensity vector waveform. is the current load waveform. is the average value of the magnetic induction intensity vector. is the average value of the current load.

[0071] The magnetic flux misdetection anomaly coefficient is used to measure the influence of external electromagnetic interference on the output signal of the Hall sensor, especially the false alarms that may be caused under high-sensitivity settings. Sources of electromagnetic interference include electromagnetic noise from other devices (such as air conditioners, power tools, motors, etc.) near the charging pile, lightning, etc. By monitoring the change in magnetic flux of the cable when these devices are working, the magnetic flux misdetection anomaly coefficient can analyze the interference of external electromagnetic noise on the magnetic field and help evaluate how the Hall sensor responds to these interference signals under different sensitivity settings. High-sensitivity settings may generate excessive signal fluctuation amplitudes for external noise, resulting in an increase in the false alarm rate, while low sensitivity may not be able to accurately capture the interference signal, thus affecting the accuracy of fault detection. This coefficient provides a basis for sensitivity optimization to ensure reducing false alarms and improving the reliability of fault prediction in a high-interference environment.

[0072] Analyzing the magnetic flux misdetection anomaly coefficient for the fault prediction effect under different magnetic flux detection sensitivities and solving the problem that in the existing cable fault detection through magnetic flux, it is impossible to effectively balance the accuracy of fault detection and the false alarm rate has the following advantages:

[0073] Optimizing the sensitivity setting: By quantifying the influence of electromagnetic noise on the Hall sensor, it can help adjust the sensitivity setting. When reducing the response to environmental noise, it can still maintain the sensitivity to actual fault signals, thus improving the accuracy of fault prediction.

[0074] Reducing false alarms: The magnetic flux misdetection anomaly coefficient can reveal the excessive response to external electromagnetic interference under high-sensitivity settings, helping the system avoid unnecessary false alarms when the current load changes or external devices are operating, ensuring that the fault detection system is more accurate.

[0075] Enhancing the anti-interference ability: By monitoring the influence of electromagnetic noise on the detection result under different sensitivities, it can identify and reduce false alarms caused by electromagnetic interference (such as from other devices near the charging pile or natural factors such as lightning), thus improving the stability and reliability of the system in a high-interference environment.

[0076] Enhanced Fault Detection Reliability: By effectively identifying and filtering interference signals, the magnetic flux misdetection anomaly coefficient helps the system balance the sensitivity and accuracy of detection in cable fault detection, ensuring accurate fault prediction results can still be provided in a complex electrical environment.

[0077] The specific method for obtaining the magnetic flux misdetection anomaly coefficient is as follows:

[0078] The specific method for obtaining the magnetic flux misdetection anomaly coefficient is as follows:

[0079] Multiply the device characteristic parameters obtained in the charging pile system by the corresponding influence coefficients to obtain the actual electromagnetic interference influence. The device characteristic parameters include current, power, and frequency;

[0080] According to the device characteristic parameters, calculate the actual electromagnetic interference influence of each device characteristic parameter in combination with the preset electromagnetic interference source calculation formula;

[0081] Calculate the electromagnetic interference influence of different frequency bands in combination with the preset interference frequency calculation formula;

[0082] Determine the standard electromagnetic interference influence of various device characteristic parameters based on historical monitoring data;

[0083] Compare and add the actual electromagnetic interference influence with the standard electromagnetic interference influence to obtain the magnetic flux misdetection anomaly coefficient;

[0084] The specific calculation formula for the actual electromagnetic interference influence of each device characteristic parameter is as follows:

[0085]

[0086] The specific calculation formula for the electromagnetic interference influence of different frequency bands is as follows:

[0087]

[0088] The specific calculation formula for the magnetic flux misdetection anomaly coefficient is as follows:

[0089]

[0090] In the formula, is the actual electromagnetic interference influence of each device characteristic parameter, m is the number of device characteristic parameters, is the topological distance of the device characteristic parameter, is the power, is the frequency, is the current, is the electromagnetic interference influence of different frequency bands, is the total number of preset interference frequencies, is the magnitude of the interference frequency, is the difference between the upper and lower limits of the interference frequency, is the frequency response of the interference frequency, is the bandwidth of the interference frequency; is the abnormal coefficient of magnetic flux misdetection, is the influence of standard electromagnetic interference on the characteristic parameters of each device, is the influence of standard electromagnetic interference in different frequency bands.

[0091] The detection response fluctuation coefficient is used to quantify the response of different fault types (such as short circuit, open circuit, poor contact, etc.) to magnetic field changes under different sensitivity settings. By analyzing the fluctuation of magnetic flux when a fault occurs, it is possible to determine whether the sensitivity setting can effectively capture the characteristic signals of different fault types. For short circuit faults, strong magnetic field changes usually occur, and high sensitivity can respond faster, but too high sensitivity may lead to excessive signal fluctuations for normal fluctuations; for poor contact faults, low sensitivity may not be able to capture subtle current fluctuations, while high sensitivity may misjudge them as faults. Therefore, the detection response fluctuation coefficient helps to evaluate the rationality of the sensitivity setting, ensuring that the system can distinguish fault signals from normal operating states and improving the accuracy and reliability of fault diagnosis.

[0092] Analyzing the influence coefficient of current load on the fault prediction effect under different magnetic flux detection sensitivities has the following advantages for solving the problem that in the existing cable fault detection through magnetic flux, it is impossible to effectively balance the accuracy of fault detection and the false alarm rate:

[0093] Reducing false alarm responses to current load fluctuations: By analyzing the interference of current load changes on the magnetic field, the sensitivity setting can be optimized to avoid excessive responses to normal current load fluctuations at high sensitivity, thereby reducing the false alarm rate. High load current changes usually cause changes in the magnetic field, but too high sensitivity may misjudge these changes as fault signals. When adjusting the sensitivity, reasonably setting the influence coefficient of current load can effectively avoid this situation.

[0094] Improving the accuracy of fault prediction: Through the influence coefficient of current load, the sensitivity of fault prediction under different current load conditions can be evaluated, so as to more accurately distinguish between faults and normal operating states. For example, when the current load is large, a low-sensitivity setting may not be able to respond to fault signals in a timely manner, while too high sensitivity may introduce unnecessary false alarms. By balancing this coefficient, it can be ensured that faults are identified promptly and accurately when they occur.

[0095] Enhance the anti-interference ability of the system against interference signals: Current load changes are often accompanied by phenomena such as the switching of electrical equipment and power fluctuations. These changes can interfere with the magnetic field detection of Hall sensors. By reasonably calculating the current load influence coefficient, the anti-interference ability of the system against external electromagnetic interference can be improved on the premise of ensuring normal fault detection, and the output signal of the Hall sensor can be optimized.

[0096] Optimize the balance between sensitivity adjustment and false alarm rate: The current load influence coefficient can provide a quantitative basis for sensitivity setting, helping the system find the optimal sensitivity setting point under different current loads, thus avoiding false alarms or missed alarms caused by too high or too low sensitivity. This can not only ensure the timely identification of faults under high loads but also avoid misjudgments of irrelevant interference signals.

[0097] Improve the reliability of cable fault diagnosis: By accurately analyzing the relationship between changes in current load and magnetic field interference, the fault prediction algorithm can be further optimized, enabling the system to maintain a high accuracy and reliability under different working environments and load conditions. By optimizing the current load influence coefficient, external factors can be effectively distinguished from fault signals, thereby improving the credibility of fault prediction.

[0098] With the combined effect of these advantages, it can effectively solve the problem of finding a reasonable balance between accuracy and false alarm rate in the detection of charging pile cable faults, thereby improving the performance of fault detection and the user experience.

[0099] The specific method for obtaining the detection response fluctuation coefficient is as follows:

[0100] Obtain the signal fluctuation amplitudes of several times during the magnetic flux detection of the charging pile cable;

[0101] The standard deviation and average value of the signal fluctuation amplitudes of the signal fluctuation amplitudes at the same time interval during the magnetic flux detection of the charging pile cable;

[0102] Calculate the signal fluctuation amplitude coefficient of variation based on the standard deviation of the signal fluctuation amplitude and the average value of the signal fluctuation amplitude;

[0103] Calculate the detection response fluctuation coefficient by calculating the signal fluctuation amplitude coefficient of variation based on a preset detection response fluctuation coefficient calculation formula.

[0104] The specific calculation formula for the signal fluctuation amplitude coefficient of variation is as follows:

[0105]

[0106] The specific calculation formula for the detection response fluctuation coefficient is as follows:

[0107]

[0108] In the formula, is the coefficient of variation of the signal fluctuation amplitude, is the signal fluctuation amplitude at the j-th moment, is the total number of moments collected during the magnetic flux detection of the charging pile cable, and j is the moment label; is the detection response fluctuation coefficient.

[0109] In this embodiment, by combining magnetic flux detection, image acquisition, and fault location, intelligent fault prediction and hidden danger monitoring of the charging pile cable are realized. First, the system uses magnetic flux detection to identify potential cable problems, and locates and analyzes faults by collecting surface images and cross-sectional images of the cable to ensure accurate prediction of the fault occurrence location. Second, regarding the sensitivity of the Hall sensor, the detection accuracy of the sensor is further optimized by quantifying historical detection data and adjusting the gain. By analyzing the data of the magnetic field interference caused by the current load change under different sensitivities, a Hall sensor output evaluation model is constructed based on the BP neural network algorithm, effectively improving the accuracy and robustness of fault prediction. This method not only improves the accuracy and reliability of the charging pile cable fault detection, but also can perform efficient fault diagnosis in complex environments, and has significant advantages such as intelligence, strong adaptability, and strong anti-interference ability.

[0110] Embodiment 2, S5, construct a Hall sensor output evaluation model based on the BP neural network algorithm according to the magnetic field interference data caused by the current load change and the detection quality data.

[0111] The construction of the Hall sensor output evaluation model based on the BP neural network algorithm according to the magnetic field interference data caused by the current load change and the detection quality data is specifically as follows:

[0112] Construct a Hall sensor output evaluation model with the obtained current load influence coefficient, magnetic flux misdetection anomaly coefficient, and detection response fluctuation coefficient, and generate a Hall sensor output evaluation coefficient;

[0113] The specific calculation formula of the Hall sensor output evaluation coefficient is as follows:

[0114]

[0115] In the formula, is the Hall sensor output evaluation coefficient, is the preset proportional coefficient of the current load influence coefficient, is the preset proportional coefficient of the magnetic flux misdetection anomaly coefficient, is the preset proportional coefficient of the detection response fluctuation coefficient, is the current load influence coefficient, is the magnetic flux misdetection anomaly coefficient, is the detection response fluctuation coefficient.

[0116] S6. Correlate the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor, construct a two-dimensional display model, perform data analysis on the two-dimensional display model to obtain an improved Hall sensor output, and apply it as the magnetic flux detection sensitivity to the prediction of charging pile cable faults.

[0117] The specific method of correlating the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor and constructing a two-dimensional display model is as follows:

[0118] Construct a two-dimensional array with the output of the Hall sensor output evaluation model and the corresponding magnetic flux detection sensitivity.

[0119] Arrange the magnetic flux detection sensitivities of several two-dimensional arrays from largest to smallest as the horizontal axis, and the output of the corresponding Hall sensor output evaluation model as the vertical axis to construct a two-dimensional coordinate system showing the change of the output of the Hall sensor output evaluation model with the magnetic flux detection sensitivity.

[0120] Mark several two-dimensional arrays on the two-dimensional coordinate system and connect them with a smooth curve to construct a two-dimensional display model.

[0121] The specific method of performing data analysis on the two-dimensional display model to obtain an improved Hall sensor output and applying it as the magnetic flux detection sensitivity to the prediction of charging pile cable faults is as follows:

[0122] According to the two-dimensional display model, obtain the magnetic flux detection sensitivities corresponding to several peaks.

[0123] Calculate the second derivatives of several peaks, judge the rate of change of the output of the Hall sensor output evaluation model with the magnetic flux detection sensitivity, and use the magnetic flux detection sensitivity corresponding to the peak with the smallest absolute value of the second derivative as the output of the improved Hall sensor.

[0124] In this embodiment, by constructing a Hall sensor output evaluation model based on the BP neural network algorithm, the output performance of the Hall sensor can be effectively evaluated and improved. By considering multiple factors such as current load, magnetic field interference, and detection quality, a Hall sensor output evaluation coefficient is generated, and in-depth data analysis is performed through a two-dimensional display model to find the optimal magnetic flux detection sensitivity. This method can improve the stability and accuracy of the Hall sensor under dynamic load and complex environment. Especially in the application of charging pile cable fault prediction, the improved sensor output can provide higher sensitivity and more accurate fault diagnosis, thus enhancing the reliability of the system.

[0125] Example 3 Figure 2It is an intelligent fault prediction system for an electric vehicle charging pile, including a fault prediction module, a magnetic flux detection sensitivity quantification module, a magnetic flux detection sensitivity acquisition module, a data analysis module, a data processing module, and a magnetic flux detection sensitivity module;

[0126] The fault prediction module is used to detect the magnetic flux of the charging pile cable. If the magnetic flux detection result indicates potential hazards in the charging pile cable, it collects the cable surface image, locates the fault, and analyzes the cross-sectional image of the cable at the located fault position, thereby predicting the fault;

[0127] The magnetic flux detection sensitivity quantification module is used to quantify the sensitivity during historical magnetic flux detection through the output of the Hall sensor and obtain the initial magnetic flux detection sensitivity by averaging;

[0128] The magnetic flux detection sensitivity acquisition module is used to perform a random number of Hall sensor gains at the initial magnetic flux detection sensitivity and obtain the outputs of several different Hall sensors as several magnetic flux detection sensitivities;

[0129] The data analysis module is used to obtain the interference data of the current load change on the magnetic field and the detection quality data for the fault prediction of the charging pile cable under several magnetic flux detection sensitivities;

[0130] The data processing module is used to construct a Hall sensor output evaluation model based on the interference data of the current load change on the magnetic field and the detection quality data using the bp neural network algorithm;

[0131] The magnetic flux detection sensitivity module is used to correspond the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor, construct a two-dimensional display model, perform data analysis on the two-dimensional display model, obtain the improved Hall sensor output, and apply it as the magnetic flux detection sensitivity to the fault prediction of the charging pile cable.

[0132] All the above formulas are dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0133] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0134] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0135] In addition, the functional modules in each embodiment of this application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0136] As mentioned above, this is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

[0137] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent fault prediction method for an electric vehicle charging pile, characterized in that: The steps include: Perform magnetic flux detection on the charging pile cable. If the magnetic flux detection result shows that the charging pile cable has hidden dangers, collect the cable surface image, locate the fault and analyze the cross-sectional image of the cable at the located fault position to predict the fault. The historical magnetic flux detection sensitivity is quantified through the output of the Hall sensor, and the average is calculated to obtain the initial magnetic flux detection sensitivity; Performing random Hall sensor gains several times at the initial magnetic flux detection sensitivity to obtain outputs of several different Hall sensors as several magnetic flux detection sensitivities; Obtain interference data and detection quality data of current load changes on magnetic fields for charging pile cable fault prediction under several magnetic flux detection sensitivities; According to the interference data of current load change on magnetic field and detection quality data, a Hall sensor output evaluation model is constructed based on BP neural network algorithm; The output of the Hall sensor output evaluation model is matched with the output of the corresponding Hall sensor, a two-dimensional display model is constructed, and data analysis is performed on the two-dimensional display model to obtain the improved Hall sensor output, which is used as the magnetic flux detection sensitivity for charging pile cable fault prediction.

2. The intelligent fault prediction method for electric vehicle charging pile according to claim 1, characterized in that: The magnetic flux detection of the charging pile cable is specifically performed as follows: Use Hall sensors to detect the magnetic field around the charging pile cable in real time and obtain the magnetic flux data of the cable in working state; Compare the collected magnetic flux data with the reference magnetic flux under normal working conditions to identify the deviation of the magnetic field; Analyze the amplitude and frequency of magnetic flux changes, construct a change curve, analyze the sudden changes and periodic fluctuations of the change curve, and compare it with the preset safety range. When it exceeds the preset safety range, it indicates that there are hidden dangers in the charging pile cable.

3. The intelligent fault prediction method for electric vehicle charging pile according to claim 2 is characterized in that: The sensitivity of the historical magnetic flux detection is quantified by the output of the Hall sensor, and the initial magnetic flux detection sensitivity is obtained by averaging, which is specifically: Obtain historical magnetic flux data of the Hall sensor at different time points from historical records; By calculating the change in the magnetic flux value per unit time of each data point, the historical magnetic flux data at different time points are quantified; The average change per unit time of all historical magnetic flux detection data is calculated as the initial magnetic flux detection sensitivity.

4. The intelligent fault prediction method for electric vehicle charging pile according to claim 3 is characterized in that: The Hall sensor gain is randomly performed several times under the initial magnetic flux detection sensitivity to obtain outputs of several different Hall sensors as several magnetic flux detection sensitivities, specifically: According to historical data, the gain range of the Hall sensor is obtained, and a number of gain values ​​are randomly obtained within the gain range of the Hall sensor based on a computer random number; The initial magnetic flux detection sensitivity is disturbed by a plurality of gain values, and outputs of a plurality of different Hall sensors are obtained as a plurality of magnetic flux detection sensitivities.

5. The intelligent fault prediction method for electric vehicle charging pile according to claim 4 is characterized in that: The detection quality data includes a detection response fluctuation coefficient, and a specific method for obtaining the detection response fluctuation coefficient is as follows: Obtain the signal fluctuation amplitudes of several charging pile cable magnetic flux detections; The standard deviation of the signal fluctuation amplitude and the average value of the signal fluctuation amplitude at the same time interval during the magnetic flux detection of the charging pile cable; Calculate the coefficient of variation of signal fluctuation amplitude according to the standard deviation of signal fluctuation amplitude and the average value of signal fluctuation amplitude; The signal fluctuation amplitude variation coefficient is calculated based on a preset detection response fluctuation coefficient calculation formula to calculate the detection response fluctuation coefficient.

6. The intelligent fault prediction method for electric vehicle charging pile according to claim 5, characterized in that: The interference data of the current load change on the magnetic field includes the current load influence coefficient and the magnetic flux misdetection anomaly coefficient. The specific method for obtaining the current load influence coefficient is as follows: Obtaining the waveforms of the magnetic induction intensity vector and the current load signal changing with time; Obtaining the harmonic composition of the magnetic induction intensity vector signal waveform and the current load signal waveform at a preset frequency, and performing harmonic component analysis to obtain a basic waveform component; The basic waveform component is converted into a harmonic component based on an inverse Fourier transform; and the floating value of the harmonic is calculated according to the quotient of the harmonic component and the basic waveform component; The floating values ​​of harmonics are analyzed and the current load influence coefficient is calculated.

7. The intelligent fault prediction method for electric vehicle charging pile according to claim 6, characterized in that: The specific method for obtaining the magnetic flux misdetection anomaly coefficient is as follows: The actual electromagnetic interference impact is obtained by multiplying the device characteristic parameters obtained in the charging pile system by the corresponding influence coefficient, wherein the device characteristic parameters include current, power, and frequency; According to the equipment characteristic parameters, the actual electromagnetic interference impact of each equipment characteristic parameter is calculated in combination with the preset electromagnetic interference source calculation formula; Combine the preset interference frequency calculation formula to calculate the electromagnetic interference impact of different frequency bands; Determine the standard electromagnetic interference impact of characteristic parameters of various equipment based on historical monitoring data; The actual electromagnetic interference effect is compared with the standard electromagnetic interference effect and added together to obtain the magnetic flux misdetection anomaly coefficient.

8. The intelligent fault prediction method for electric vehicle charging pile according to claim 7, characterized in that: The output of the Hall sensor output evaluation model is matched with the output of the corresponding Hall sensor to construct a two-dimensional display model, specifically: The output of the Hall sensor output evaluation model and the corresponding magnetic flux detection sensitivity are used to construct a two-dimensional array; Arrange the magnetic flux detection sensitivities of several two-dimensional arrays from large to small as the horizontal axis, and the output of the corresponding Hall sensor output evaluation model as the vertical axis, and construct a two-dimensional coordinate system in which the output of the Hall sensor output evaluation model changes with the magnetic flux detection sensitivity; Mark several two-dimensional arrays onto a two-dimensional coordinate system and connect them with smooth curves to construct a two-dimensional display model.

9. The intelligent fault prediction method for electric vehicle charging pile according to claim 8, characterized in that: The two-dimensional display model is analyzed for data to obtain an improved Hall sensor output, which is used as the magnetic flux detection sensitivity for charging pile cable fault prediction, specifically: According to the two-dimensional display model, the magnetic flux detection sensitivity corresponding to several peak values ​​is obtained; The second-order derivatives of several peak values ​​are calculated to determine the rate at which the output of the Hall sensor output evaluation model changes with the flux detection sensitivity, and the flux detection sensitivity corresponding to the peak value with the smallest absolute value of the second-order derivative is used as the output of the improved Hall sensor.

10. A system using the intelligent fault prediction method for an electric vehicle charging pile according to any one of claims 1 to 9, characterized in that: It includes a fault prediction module, a magnetic flux detection sensitivity quantification module, a magnetic flux detection sensitivity acquisition module, a data analysis module, a data processing module and a magnetic flux detection sensitivity module; The fault prediction module is used to perform magnetic flux detection on the charging pile cable. If the magnetic flux detection result shows that the charging pile cable has hidden dangers, the cable surface image is collected to locate the fault and analyze the cross-sectional image of the cable at the located fault position, so as to predict the fault; The magnetic flux detection sensitivity quantification module is used to quantify the sensitivity of historical magnetic flux detection through the output of the Hall sensor and average it to obtain the initial magnetic flux detection sensitivity; A magnetic flux detection sensitivity acquisition module is used to perform random Hall sensor gains several times under the initial magnetic flux detection sensitivity to obtain outputs of several different Hall sensors as several magnetic flux detection sensitivities; A data analysis module is used to obtain interference data and detection quality data of the current load change on the magnetic field for charging pile cable fault prediction under several magnetic flux detection sensitivities; The data processing module is used to construct a Hall sensor output evaluation model based on the BP neural network algorithm according to the interference data of the current load change on the magnetic field and the detection quality data; The magnetic flux detection sensitivity module is used to match the output of the Hall sensor output evaluation model with the output of the corresponding Hall sensor, build a two-dimensional display model, and perform data analysis on the two-dimensional display model to obtain the improved Hall sensor output, which will be used as the magnetic flux detection sensitivity for charging pile cable fault prediction.

Images