Method and device for determining guiding state of alternating current charging pile and alternating current charging pile

Abstract

The invention discloses a method and a device for determining a guide state of an alternating current charging pile and the alternating current charging pile, and particularly relates to the technical field of alternating current charging piles. Firstly, transmission and analysis of a charging request signal are completed through a communication interface between a charging pile and an electric vehicle, in the analysis process, filtering parameters are dynamically adjusted through an adaptive filtering algorithm, and meanwhile, abnormal signals and normal signals are accurately recognized based on a deep learning model; extracting multi-dimensional characteristic data from the screened normal signals by adopting a time-frequency domain conjoint analysis technology, calculating a signal accuracy index through a fuzzy logic algorithm, further judging whether the electric vehicle meets a charging condition or not for the high-accuracy normal signals, starting charging operation if the charging condition is met, and outputting an error prompt if the charging condition is not met. The signal analysis precision and adaptability of the charging pile are remarkably improved, the misjudgment rate and the operation and maintenance cost caused by signal reflection are reduced, and the reliability and safety of a charging system are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of AC charging piles, and particularly to a method and device for determining the guiding state of an AC charging pile and an AC charging pile. Background Art

[0002] The method for determining the guiding state of an AC charging pile refers to detecting the connection state between the charging pile and the electric vehicle through specific technical means and algorithms, and then judging whether charging can start or whether there are potential safety hazards. This process usually relies on technical means such as electrical signals, data communication protocols, and hardware interfaces to ensure a safe and stable connection between the charging pile and the electric vehicle. A common method for determining the guiding state is through the communication protocol between the charging pile and the electric vehicle. Modern charging piles and electric vehicles usually use standardized protocols (such as PLC or CAN communication protocols) for data exchange. The charging pile sends a detection command or a charging request signal to the electric vehicle, and after the electric vehicle responds, the charging pile can judge whether the charging conditions are met by analyzing the communication data. If the response data of the electric vehicle meets the requirements (such as normal battery state, voltage adaptation, etc.), the charging pile can determine the guiding state as the "chargeable" state, thus allowing charging to start. This method effectively ensures the intelligence and safety of the charging process.

[0003] The existing technology has the following deficiencies: In the case of long-distance signal transmission or when multiple devices are connected by branch cables, signal reflection may occur, resulting in waveform distortion or repeated reception of data. The charging pile may thus misjudge the vehicle connection state as abnormal or not connected, directly preventing the charging from starting. At the same time, the signal reflection problem often hides in the cable layout or interface design and is difficult to discover and repair through simple tests. It may be necessary to replace equipment or re-plan the wiring, significantly increasing the operation and maintenance costs and deployment complexity. Summary of the Invention

[0004] The object of the present invention is to provide a method and device for determining the guiding state of an AC charging pile and an AC charging pile to solve the deficiencies in the background art.

[0005] To achieve the above object, the present invention provides the following technical solution: A method for determining the guiding state of an AC charging pile, comprising the following steps: S1: Send a charging request signal through the communication interface between the charging pile and the electric vehicle, and the charging pile receives the charging request signal of the electric vehicle and analyzes the communication data; S2: During the process of analyzing the communication data, dynamically adjust the filtering parameters through an adaptive filtering algorithm, and automatically eliminate the reflection and distortion in the signal according to the change of the communication environment; S3: Based on the abnormal charging request signal recognition algorithm trained by the deep learning model, compare the historical data with the real-time charging request signal to accurately distinguish the normal charging request signal from the reflected interference charging request signal; S4: For the normal charging request signal, adopt the time-frequency domain joint analysis technology to extract the multi-dimensional feature data of the normal charging request signal respectively, and judge the accuracy of the normal charging request signal; S5: For the normal charging request signal with high accuracy, judge whether the electric vehicle meets the charging conditions. If the charging conditions are met, determine the guiding state as the rechargeable state and start charging; if the conditions are not met, output an error prompt and stop the charging operation.

[0006] Preferably, in S3, based on the abnormal charging request signal recognition algorithm trained by the deep learning model, compare the historical data with the real-time charging request signal to accurately distinguish the normal charging request signal from the reflected interference charging request signal. Specifically: Collect signal samples from normal charging and abnormal reflection interference scenarios, including time-domain waveforms and frequency-domain features. Receive the current charging request signal through the communication interface of the charging pile and perform sampling processing; extract the signal amplitude of each sampling point and calculate the time offset between the historical signal and the real-time signal through the cross-correlation method. Analyze the frequency distribution of the signal through Fourier transform, extract the spectral energy characteristics, and standardize the characteristics of the historical signal to obtain the feature mean and standard deviation for consistency comparison with the real-time signal; define the cross-correlation function to calculate the similarity and time delay between the historical signal and the real-time signal: ; where is the amplitude of the historical charging request signal at time t, is the amplitude of the real-time charging request signal at time , is the cross-correlation value, τ is the time delay, and determine the optimal delay : Find the time offset corresponding to the maximum correlation value: ; According to the actual delay and the mean and standard deviation of the normal delay distribution, define the delay anomaly index: ; is the delay anomaly index.

[0007] Preferably, in S3, input the delay anomaly index , amplitude characteristics and spectral characteristics of the real-time signal into the deep learning model, and the model calculates the signal matching score Pnormal according to the input characteristics: ; where f represents the prediction function of the deep learning model, is the normalized value of the signal amplitude characteristics, is the normalized value of the signal spectrum feature; Compare the calculated signal matching score with the reference threshold θ under normal conditions set according to historical data. If > θ, the signal is a normal charging request signal. If Pnormal ≤ θ, the signal is a reflected interference signal.

[0008] Preferably, in S4, for the normal charging request signal, a time-frequency domain joint analysis technology is adopted to extract multi-dimensional feature data of the normal charging request signal respectively to judge the accuracy of the normal charging request signal. Specifically: Extract the maximum amplitude of the signal and the minimum amplitude , calculate the change range of the signal amplitude , and the expression is: ; calculate the total energy of the signal in the time domain : ; where the amplitude of the signal at time t is denoted as S(t), and are the start and end points of the signal sampling time. Measure the stability of the signal amplitude through the mean value ( ) and variance of the signal , and the expression is: ; where N is the number of sampling points, is the time of each sampling point; calculate the spectral centroid , which represents the center frequency of the signal energy: ; where P(f) is the power of the signal at frequency f; calculate the effective bandwidth of the spectral energy distribution, which reflects the expansion of the frequency components, and the expression is: ; extract the frequency component with the largest energy in the spectrum: ; use the short-time Fourier transform STFT to calculate the energy distribution of the signal in the time-frequency domain , and extract the main energy concentration region: ; use the time-frequency domain feature data as the feature input.

[0009] Preferably, in S4, use the extracted time-frequency domain feature data as the input of the fuzzy logic system. Define rules according to the characteristic pattern of normal signals. The fuzzy inference adopts a rule activation mechanism: ; where is the activation membership degree of the i-th rule; Defuzzify the fuzzy result to obtain the comprehensive accuracy index Iaccuracy of the signal, and the expression is: ; is the weight of the i-th rule; Judge the accuracy of the signal according to the size of the comprehensive accuracy index Iaccuracy: : The signal is accurate; : The signal accuracy is average and further detection is required; : The signal is inaccurate.

[0010] Preferably, in S5, for the normal charging request signal with high accuracy, judge whether the electric vehicle meets the charging conditions. If the charging conditions are met, determine the guiding state as the rechargeable state and start charging; if the conditions are not met, output an error prompt and stop the charging operation. Specifically: Convert the signal matching score and the comprehensive accuracy index into a comprehensive feature vector, use the comprehensive feature vector as the input of the machine learning model, and use the machine learning model to predict the accuracy value label of the charging pile to judge the vehicle connection state for each group of comprehensive feature vectors as the prediction target, and use minimizing the sum of the prediction errors of the accuracy value labels of all charging piles to judge the vehicle connection state as the training target. Train the machine learning model until the sum of the prediction errors reaches convergence and then stop the model training. Determine the accuracy value of the charging pile to judge the vehicle connection state according to the model output result, where the machine learning model is a polynomial regression model.

[0011] Preferably, compare the obtained accuracy value of the charging pile to judge the vehicle connection state with the pre-set accuracy value reference threshold. If the accuracy value of the charging pile to judge the vehicle connection state is greater than or equal to the pre-set accuracy value reference threshold, it means that the accuracy of the charging pile to judge the vehicle connection state is high, and the charging conditions are met at this time and charging is started; if the accuracy value of the charging pile to judge the vehicle connection state is less than the pre-set accuracy value reference threshold, it means that the accuracy of the charging pile to judge the vehicle connection state is low, and the charging conditions are not met at this time, then output an error prompt and stop the charging operation.

[0012] The present invention also provides a guiding state determination device for an AC charging pile, including a communication interface module, a signal processing module, an abnormal signal detection module, a signal feature analysis module, and a charging condition judgment and control module; Communication interface module: Send a charging request signal through the communication interface between the charging pile and the electric vehicle, and the charging pile receives the charging request signal of the electric vehicle and parses the communication data; Signal processing module: During the process of parsing the communication data, dynamically adjust the filtering parameters through an adaptive filtering algorithm, and automatically eliminate the reflection and distortion in the signal according to the change of the communication environment; Abnormal signal detection module: Based on the abnormal charging request signal recognition algorithm trained by the deep learning model, compare the historical data with the real-time charging request signal to accurately distinguish the normal charging request signal from the reflected interference charging request signal; Signal Feature Analysis Module: For normal charging request signals, time-frequency domain joint analysis technology is adopted to extract multi-dimensional feature data of normal charging request signals respectively, and the accuracy of normal charging request signals is judged. Charging Condition Judgment and Control Module: For normal charging request signals with high accuracy, it is judged whether the electric vehicle meets the charging conditions. If the charging conditions are met, the guiding state is determined as the rechargeable state and charging is started; if the conditions are not met, an error prompt is output and the charging operation is stopped.

[0013] The present invention also provides an AC charging pile, including a guiding state determination device for an AC charging pile.

[0014] In the above technical solution, the technical effects and advantages provided by the present invention are as follows: 1. The present invention significantly improves the accuracy and robustness of signal analysis, especially for signal reflection problems in long-distance signal transmission or complex branch connection scenarios. The adaptive filtering algorithm is used to adjust the filtering parameters in real time, effectively eliminating the reflection interference and distortion in the signal; combined with the deep learning model, abnormal signals are accurately identified and eliminated, avoiding the failure of charging start caused by misjudgment. At the same time, the time-frequency domain joint analysis technology is used to extract multi-dimensional feature data of normal signals, and the fuzzy logic algorithm is combined to calculate the comprehensive accuracy index of the signal, further enhancing the accuracy of signal credibility judgment and laying a foundation for the efficient and stable operation of the charging pile.

[0015] 2. On the basis of ensuring high-accuracy normal signals, the present invention uses a machine learning model to further judge the vehicle connection state and charging conditions, accurately controlling the start and stop of charging, and greatly improving the intelligent level of the charging pile. The overall solution significantly enhances the safety, reliability and adaptability of the charging system, providing technical support for the deep integration of the future smart grid and vehicle-to-everything network while optimizing the user experience. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0017] Figure 1 It is a flowchart of the method of the present invention.

[0018] Figure 2 It is a block diagram of the device of the present invention. Detailed Embodiments

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] Embodiment 1. Refer to Figure 1 and Figure 2 As shown, a method for determining the guiding state of an AC charging pile in this embodiment includes the following steps: S1: Send a charging request signal through the communication interface between the charging pile and the electric vehicle. The charging pile receives the charging request signal of the electric vehicle and parses the communication data; S2: During the process of parsing the communication data, dynamically adjust the filtering parameters through an adaptive filtering algorithm, and automatically eliminate the reflection and distortion in the signal according to the change of the communication environment; S3: Based on the abnormal charging request signal recognition algorithm obtained by training the deep learning model, compare the historical data with the real-time charging request signal to accurately distinguish the normal charging request signal from the reflected interference charging request signal; S4: For the normal charging request signal, adopt the time-frequency domain joint analysis technology to extract the multi-dimensional feature data of the normal charging request signal respectively, and judge the accuracy of the normal charging request signal; S5: For the normal charging request signal with high accuracy, judge whether the electric vehicle meets the charging conditions. If the charging conditions are met, determine the guiding state as the rechargeable state and start charging; if the conditions are not met, output an error prompt and stop the charging operation.

[0021] In S1, sending a charging request signal through the communication interface between the charging pile and the electric vehicle, and the charging pile receiving the charging request signal of the electric vehicle and parsing the communication data is specifically as follows: After sending a charging request signal through the communication interface between the charging pile and the electric vehicle and receiving the charging request signal of the electric vehicle, the specific process of parsing the communication data is: Sending the charging request signal: The charging pile sends a standardized charging request signal to the electric vehicle through the communication interface (such as PLC, CAN or UART interface). The signal contains the basic information of the charging pile, such as the output voltage range, output current range, maximum power capacity, and charging protocol version.

[0022] Receiving the response signal of the electric vehicle: After the electric vehicle receives the charging request signal from the charging pile, it feeds back its own charging demand signal through the communication interface. The feedback signal includes information such as the current state of the vehicle battery (such as voltage, current, SOC (State of Charge)), the maximum acceptance power, and whether it supports the fast charging mode.

[0023] Analyzing the communication data: The charging pile analyzes the received response signal according to the communication protocol, specifically including: Protocol version verification: Confirm whether the response signal of the electric vehicle is consistent with the protocol version supported by the charging pile to ensure data compatibility.

[0024] Extracting battery parameters: Extract the battery parameters fed back by the electric vehicle, including information such as the current voltage, current, and maximum power, for matching the output capacity of the charging pile.

[0025] Detecting abnormal states: Detect whether there are abnormal state flags in the response signal, such as too high battery temperature, short - circuit risk, or vehicle interface failure.

[0026] Matching the charging mode: According to the charging demand signal fed back by the electric vehicle, determine whether high - power charging, ordinary AC charging, or other special modes are supported to ensure the selection of the optimal charging strategy.

[0027] S2: During the process of analyzing the communication data, dynamically adjust the filtering parameters through an adaptive filtering algorithm to automatically eliminate reflections and distortions in the signal according to changes in the communication environment.

[0028] The charging pile collects the signals transmitted in real - time from the communication interface, including the response signal of the electric vehicle and background noise. Based on the preset communication protocol and transmission environment, initialize the parameters of the adaptive filter (such as the filtering order, step - size factor, initial weight). Use the standardized communication protocol format or the historical communication data model as the target reference signal for comparison and optimization.

[0029] By detecting the time - delay characteristics of the signal, analyze whether there are reflected signals. Reflected signals usually appear as repeated waveforms similar to the original signal but with reduced amplitude and fixed delay. Conduct amplitude - characteristic analysis on the received signal, record the abnormal high or low amplitude change points, and mark the possible reflected - signal areas. Use the phase - analysis method to detect the phase shift in the signal to further confirm the existence of the reflected signal.

[0030] Compare the collected communication signal with the target reference signal to calculate the error signal. The error signal is used to guide the optimization process of the adaptive filter. According to the error signal, the filter parameters are dynamically adjusted through the following adaptive algorithms: LMS algorithm (Least Mean Square algorithm): Adjust the filtering weights according to the error signal to minimize the error amplitude. RLS algorithm (Recursive Least Squares algorithm): Quickly adjust the filter parameters by weighting historical data to improve the adaptability to dynamic environments. Time-frequency domain analysis method: Analyze the time and frequency characteristics of the signal simultaneously to optimize the response bandwidth of the filter. Update the filter weights using the optimized parameters to ensure real-time adaptation to signal environment changes.

[0031] Input the received communication signal into the optimized filter to eliminate the interference of reflected signals and signal distortion. Reconstruct the filtered signal to ensure that its waveform is consistent with the target reference signal.

[0032] Perform consistency verification on the corrected signal to ensure its data integrity is compatible with the communication protocol.

[0033] Use the recovered communication data after filtering as the final parsing result and transfer it to the subsequent guidance state judgment module. Record the real-time performance of the filter (such as error amplitude, filtering stability) for long-term optimization of the filtering algorithm.

[0034] S3: Based on the abnormal charging request signal recognition algorithm trained by the deep learning model, compare the historical data with the real-time charging request signal to accurately distinguish between normal charging request signals and reflected interference charging request signals.

[0035] Collect signal samples from normal charging and abnormal reflection interference scenarios, including time-domain waveforms and frequency-domain characteristics. Receive the current charging request signal through the communication interface of the charging pile and perform sampling processing; Extract the signal amplitude of each sampling point and calculate the time offset between the historical signal and the real-time signal through the cross-correlation method. Analyze the frequency distribution of the signal through Fourier transform and extract the spectral energy characteristics. Standardize the characteristics of the historical signal to obtain the feature mean and standard deviation for consistent comparison with the real-time signal; Define the cross-correlation function for calculating the similarity and time delay between the historical signal and the real-time signal: ; where is the amplitude of the historical charging request signal at time t, representing the time-domain representation of the historical signal, is the amplitude of the real-time charging request signal at time , representing the time-domain representation of the current signal after time offset. is the cross-correlation value, and τ is the time delay. Determine the optimal delay : Find the time offset corresponding to the maximum correlation value: ; According to the actual delay Deviation from the mean of the normal delay distribution and the standard deviation , define the delay anomaly index: ; is the delay anomaly index, the larger the value, the more it deviates from the normal delay range, and it may be a reflected interference signal. Input the delay anomaly index , amplitude characteristics and spectral characteristics of the real-time signal into the deep learning model. The model calculates the signal matching score Pnormal based on the input characteristics: ; where f represents the prediction function of the deep learning model, is the normalized value of the signal amplitude characteristic, is the normalized value of the signal spectral characteristic; the amplitude characteristic is normalized as: , where A is the measured characteristic, and are the historical sample mean and standard deviation. The prediction function f is a multi-layer perceptron (MLP) model built based on the TensorFlow framework, with the input being the feature vector , , and the output being the signal matching score Pnormal ∈ [0,1].

[0036] Compare the calculated signal matching score with the reference threshold θ under normal conditions set according to historical data. If > θ (the reference threshold is usually 0.8), then the signal is a normal charging request signal. If Pnormal ≤ θ, then the signal is a reflected interference signal.

[0037] S4: For normal charging request signals, use the time-frequency domain joint analysis technology to extract multi-dimensional feature data of the normal charging request signals respectively, and judge the accuracy of the normal charging request signals.

[0038] Extract the maximum amplitude and the minimum amplitude of the signal, calculate the change range of the signal amplitude, and the expression is: ; Calculate the total energy of the signal in the time domain: ; where the amplitude of the signal at time t is denoted as S(t), and are the start and end points of the signal sampling time, and the stability of the signal amplitude is measured by the mean ( ) and variance of the signal , and the expression is: ; where N is the number of sampling points, is the time of each sampling point; calculate the spectral centroid of the signal, which represents the center frequency of the signal energy: where P(f) is the power of the signal at frequency f; calculate the effective bandwidth of the spectral energy distribution , which reflects the expansibility of the frequency components, and the expression is: ; extract the frequency component with the maximum energy in the spectrum: ; use the short-time Fourier transform (STFT) to calculate the energy distribution of the signal in the time-frequency domain , and extract the main energy concentration region: ; use the key time-frequency points as features for input.

[0039] Use the extracted time-frequency domain feature data as the input of the fuzzy logic system, including: : amplitude range (low, medium, high); : time-domain energy (weak, medium, strong); : signal stability (low, medium, high); : spectral center (low frequency, medium frequency, high frequency); : bandwidth (narrow, medium, wide); : frequency peak (low frequency, medium frequency, high frequency).

[0040] Define rules according to the characteristic patterns of normal signals, for example: Rule 1: If is high, and is low, and is medium, then the signal accuracy is high.

[0041] Rule 2: If is weak, and is low, then the signal accuracy is low.

[0042] Rule 3: If is wide, and is high, then the signal accuracy is medium.

[0043] Fuzzy inference adopts the rule activation mechanism: ; where is the activation membership degree of the i-th rule; Defuzzify the fuzzy result to obtain the comprehensive accuracy index Iaccuracy of the signal, and the expression is: ; is the weight of the i-th rule (the scoring value corresponding to the accuracy, usually in the range of [0,1]).

[0044] Judge the accuracy of the signal according to the size of the comprehensive accuracy index Iaccuracy: Iaccuracy > 0.8: The signal is accurate and reliable. 0.5 ≤ Iaccuracy ≤ 0.8: The signal accuracy is average and further detection is required. Iaccuracy < 0.5: The signal is inaccurate and there may be reflection interference.

[0045] S5: For a normal charging request signal with high accuracy, determine whether the electric vehicle meets the charging conditions. If the charging conditions are met, determine the guiding state as the chargeable state and start charging; if the conditions are not met, output an error prompt and stop the charging operation.

[0046] Convert the signal matching score and the comprehensive accuracy index into a comprehensive feature vector, and use the comprehensive feature vector as the input of the machine learning model. The machine learning model takes predicting the accuracy value label of the charging pile to judge the vehicle connection status for each group of comprehensive feature vectors as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of all charging piles to judge the vehicle connection status as the training target. Train the machine learning model until the sum of the prediction errors reaches convergence and then stop the model training. Determine the accuracy value of the charging pile to judge the vehicle connection status according to the model output result. Among them, the machine learning model is a polynomial regression model.

[0047] The method for obtaining the accuracy value of the charging pile to judge the vehicle connection status is: obtain the corresponding function expression from the comprehensive feature vector training data of the trained machine learning model: ; In the formula, is the output function of the model, Pnormal is the signal matching score, Iaccuracy is the comprehensive accuracy index, is the accuracy value of the charging pile to judge the vehicle connection status.

[0048] Compare the obtained accuracy value of the charging pile to judge the vehicle connection status with the pre-set accuracy value reference threshold. If the accuracy value of the charging pile to judge the vehicle connection status is greater than or equal to the pre-set accuracy value reference threshold, it indicates that the accuracy of the charging pile to judge the vehicle connection status is high. At this time, the charging conditions are met and charging is started; if the accuracy value of the charging pile to judge the vehicle connection status is less than the pre-set accuracy value reference threshold, it indicates that the accuracy of the charging pile to judge the vehicle connection status is low. At this time, the charging conditions are not met, and an error prompt is output and the charging operation is stopped.

[0049] In this embodiment, a charging request signal is sent and received through the communication interface between the charging pile and the electric vehicle, and the communication data is parsed. During the parsing process, an adaptive filtering algorithm is used to dynamically adjust the filtering parameters, and the reflections and distortions in the signal are automatically eliminated according to the changes in the communication environment, improving the accuracy of the signal. At the same time, an abnormal signal recognition algorithm trained based on a deep learning model compares historical data with real-time signals to accurately distinguish normal signals from reflected interference signals. For the selected normal signals, time-frequency domain joint analysis technology is used to extract multi-dimensional feature data, and a fuzzy logic algorithm is combined to calculate the comprehensive accuracy index of the signal to determine whether the signal is reliable. Finally, for normal signals with high accuracy, it is further detected whether the electric vehicle meets the charging conditions. If it meets the conditions, the guiding state is determined to be the rechargeable state and charging is started; if it does not meet the conditions, an error prompt is output and the charging operation is stopped to ensure the safety and stability of the charging process.

[0050] Embodiment 2. The guiding state determination device of an AC charging pile in this embodiment includes a communication interface module, a signal processing module, an abnormal signal detection module, a signal feature analysis module, and a charging condition judgment and control module; Communication interface module: A charging request signal is sent through the communication interface between the charging pile and the electric vehicle, and the charging pile receives the charging request signal of the electric vehicle and parses the communication data; Signal processing module: During the process of parsing the communication data, the filtering parameters are dynamically adjusted through an adaptive filtering algorithm, and the reflections and distortions in the signal are automatically eliminated according to the changes in the communication environment; Abnormal signal detection module: Based on the abnormal charging request signal recognition algorithm trained by a deep learning model, historical data is compared with real-time charging request signals to accurately distinguish normal charging request signals from reflected interference charging request signals; Signal feature analysis module: For normal charging request signals, time-frequency domain joint analysis technology is used to extract multi-dimensional feature data of the normal charging request signals respectively, and the accuracy of the normal charging request signals is judged; Charging condition judgment and control module: For normal charging request signals with high accuracy, it is judged whether the electric vehicle meets the charging conditions. If the charging conditions are met, the guiding state is determined to be the rechargeable state and charging is started; if the conditions are not met, an error prompt is output and the charging operation is stopped.

[0051] Embodiment 3. The AC charging pile in this embodiment includes the guiding state determination device of the AC charging pile described in Embodiment 2.

[0052] The above formulas are all calculated by taking the numerical values after removing the dimensions. The formulas are obtained by collecting a large amount of data for software simulation to get a formula that is closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0053] It should be understood that the term "and / or" in this text is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. Additionally, the character " / " in this text generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship. The specific meaning can be understood by referring to the context before and after.

[0054] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this text 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.

[0055] As described 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.

Claims

1. A method for determining the guiding state of an AC charging pile, characterized in that: It includes the following steps: S1: Send a charging request signal through the communication interface between the charging pile and the electric vehicle. The charging pile receives the charging request signal of the electric vehicle and parses the communication data; S2: During the process of parsing the communication data, dynamically adjust the filtering parameters through an adaptive filtering algorithm, and automatically eliminate the reflection and distortion in the signal according to the change of the communication environment; S3: Based on the abnormal charging request signal recognition algorithm obtained by training the deep learning model, compare the historical data with the real-time charging request signal, and accurately distinguish the normal charging request signal from the reflected interference charging request signal; S4: For the normal charging request signal, adopt the time-frequency domain joint analysis technology, extract the multi-dimensional feature data of the normal charging request signal respectively, and judge the accuracy of the normal charging request signal; S5: For the normal charging request signal with high accuracy, judge whether the electric vehicle meets the charging conditions. If the charging conditions are met, determine the guiding state as the rechargeable state and start charging; if the conditions are not met, output an error prompt and stop the charging operation.

2. The guiding state determination method of an AC charging pile according to claim 1, characterized in that: In S3, based on the abnormal charging request signal recognition algorithm obtained by training the deep learning model, compare the historical data with the real-time charging request signal, and accurately distinguish the normal charging request signal from the reflected interference charging request signal. Specifically: Collect signal samples from normal charging and abnormal reflection interference scenarios, including time-domain waveforms and frequency-domain characteristics. Receive the current charging request signal through the communication interface of the charging pile and perform sampling processing; extract the signal amplitude of each sampling point and calculate the time offset between the historical signal and the real-time signal through the cross-correlation method, analyze the frequency distribution of the signal through Fourier transform, extract the spectral energy characteristics, standardize the characteristics of the historical signal to obtain the feature mean and standard deviation for consistency comparison with the real-time signal; define the cross-correlation function for calculating the similarity and time delay between the historical signal and the real-time signal: ; where is the amplitude of the historical charging request signal at time t, is the amplitude of the real-time charging request signal at time , is the cross-correlation value, τ is the time delay, and determine the optimal delay : find the time offset corresponding to the maximum correlation value: ; according to the actual delay and the deviation of the normal delay distribution mean and the standard deviation , define the delay anomaly index: ; is the delay anomaly index.

3. The guiding state determination method of an AC charging pile according to claim 2, wherein: In S3, the delay anomaly index of the real-time signal , the amplitude feature and the spectrum feature are input into the deep learning model, and the model calculates the signal matching score Pnormal according to the input features: ; where f represents the prediction function of the deep learning model, is the normalized value of the signal amplitude feature, is the normalized value of the signal spectrum feature; Compare the calculated signal matching score with the reference threshold θ under normal conditions set according to historical data. If > θ, the signal is a normal charging request signal. If Pnormal ≤ θ, the signal is a reflected interference signal.

4. A method for determining the guiding state of an AC charging pile according to claim 1, characterized in that: In S4, for the normal charging request signal, adopt the time-frequency domain joint analysis technology, extract the multi-dimensional feature data of the normal charging request signal respectively, and judge the accuracy of the normal charging request signal. Specifically: Extract the maximum amplitude of the signal and the minimum amplitude , and calculate the change range of the signal amplitude . The expression is: ; Calculate the total energy of the signal in the time domain : ; where the amplitude of the signal at time t is denoted as S(t), and are the start and end points of the signal sampling time, and the stability of the signal amplitude is measured by the mean and variance of the signal , and the expression is: ; where N is the number of sampling points, is the time of each sampling point; calculate the spectral centroid of the signal , which represents the center frequency of the signal energy: ; where P(f) is the power of the signal at frequency f; calculate the effective bandwidth of the spectral energy distribution , which reflects the extensibility of the frequency components, and the expression is: ; extract the frequency component with the maximum energy in the spectrum: ; use the short-time Fourier transform STFT to calculate the energy distribution of the signal in the time-frequency domain , and extract the main energy concentration region: ; Use the time-frequency domain feature data as the feature input.

5. A method for determining the guiding state of an AC charging pile according to claim 4, characterized in that: In S4, the extracted time-frequency domain feature data is used as the input of the fuzzy logic system, rules are defined according to the characteristic patterns of normal signals, and fuzzy inference adopts a rule activation mechanism: ; where is the activation membership degree of the i-th rule; Clarify the fuzzification result to obtain the comprehensive accuracy index Iaccuracy of the signal, and the expression is: ; is the weight of the i-th rule; Judge the accuracy of the signal according to the size of the comprehensive accuracy index Iaccuracy: Iaccuracy>0.8: the signal is accurate; 0.5≤Iaccuracy≤0.8: the signal accuracy is average and further detection is required; Iaccuracy<0.5: the signal is inaccurate.

6. A method for determining the guiding state of an AC charging pile according to claim 1, characterized in that: In S5, for the normal charging request signal with high accuracy, judge whether the electric vehicle meets the charging conditions. If the charging conditions are met, determine the guiding state as the rechargeable state and start charging; if the conditions are not met, output an error prompt and stop the charging operation. Specifically: Convert the signal matching score and the comprehensive accuracy index into a comprehensive feature vector, use the comprehensive feature vector as the input of the machine learning model, take the accuracy value label predicted by the machine learning model for each group of comprehensive feature vectors to judge the vehicle connection state of the charging pile as the prediction target, and take minimizing the sum of the prediction errors of the accuracy value labels for all charging piles to judge the vehicle connection state as the training target, train the machine learning model until the sum of the prediction errors reaches convergence and then stop the model training, and determine the accuracy value of the charging pile to judge the vehicle connection state according to the model output result. Among them, the machine learning model is a polynomial regression model.

7. A method for determining the guiding state of an AC charging pile according to claim 6, characterized in that: Compare the accuracy value of the charging pile's judgment on the vehicle connection status obtained with the pre-set accuracy value reference threshold. If the accuracy value of the charging pile's judgment on the vehicle connection status is greater than or equal to the pre-set accuracy value reference threshold, it indicates that the charging pile has a high accuracy in judging the vehicle connection status. At this time, the charging condition is met and charging is started. If the accuracy value of the charging pile's judgment on the vehicle connection status is less than the pre-set accuracy value reference threshold, it indicates that the charging pile has a low accuracy in judging the vehicle connection status. At this time, the charging condition is not met, and an error prompt is output and the charging operation is stopped.

8. A guiding state determination device for an AC charging pile, which is used to implement the guiding state determination method for an AC charging pile according to any one of claims 1-7, characterized in that: It includes a communication interface module, a signal processing module, an abnormal signal detection module, a signal feature analysis module, and a charging condition judgment and control module; Communication interface module: Send a charging request signal through the communication interface between the charging pile and the electric vehicle. The charging pile receives the charging request signal of the electric vehicle and analyzes the communication data; Signal processing module: During the process of analyzing the communication data, dynamically adjust the filtering parameters through an adaptive filtering algorithm, and automatically eliminate the reflection and distortion in the signal according to the change of the communication environment; Abnormal signal detection module: Based on the abnormal charging request signal recognition algorithm obtained by training the deep learning model, compare the historical data with the real-time charging request signal to accurately distinguish the normal charging request signal from the reflected interference charging request signal; Signal feature analysis module: For the normal charging request signal, adopt the time-frequency domain joint analysis technology to extract the multi-dimensional feature data of the normal charging request signal respectively, and judge the accuracy of the normal charging request signal; Charging condition judgment and control module: For the normal charging request signal with high accuracy, judge whether the electric vehicle meets the charging condition. If the charging condition is met, determine the guiding state as the rechargeable state and start charging. If the condition is not met, output an error prompt and stop the charging operation.

9. An AC charging pile, characterized in that: It includes a guiding state determination device for an AC charging pile as described in claim 8.

Images