DC Charging Pile Fault Diagnosis and Repair Method, Device and Storage Medium

Abstract

The present invention relates to a method, device and storage medium for fault diagnosis and repair of a DC charging pile. The method: performing Hilbert transform on the voltage, current and temperature parameters of the charging pile to obtain envelope data of multi-channel analytic signals; performing frequency domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain a spectrum feature vector; inputting the spectrum feature vector into a dual control law observer with continuous terms and discontinuous terms for state tracking to obtain a system residual matrix; performing threshold detection and cumulative error integral operation on the system residual matrix to obtain fault feature data; performing λ consensus iterative operation on the fault feature data to obtain a control parameter compensation scheme. The present invention improves the recognition accuracy of fault features and ensures the optimality and reliability of the repair strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of DC charging piles, and particularly to a method, device and storage medium for fault diagnosis and repair of DC charging piles. Background Art

[0002] With the rapid development of the new energy vehicle industry, the operational reliability of DC charging piles, as key infrastructure, has attracted increasing attention. During long-term operation, charging piles are prone to various faults, which not only affect the charging efficiency but may also lead to safety accidents. Therefore, an efficient and accurate fault diagnosis and repair mechanism needs to be established. Currently, the mainstream fault diagnosis methods for charging piles are mainly based on the spectral analysis of operating parameters. However, due to limited sampling time, traditional methods have obvious deficiencies in diagnostic accuracy and real-time performance.

[0003] Existing charging pile fault diagnosis systems generally adopt a single controller structure, making it difficult to balance the accuracy and response speed of diagnosis simultaneously. Traditional multi-channel data acquisition schemes face problems such as low resolution and signal interference, which affect the quality of fault feature extraction. In terms of fault repair, existing repair strategies often use fixed control parameters and lack the ability to adaptively adjust to the dynamic characteristics of the system, making it difficult to meet the real-time repair requirements of large-scale charging stations. Summary of the Invention

[0004] The main objective of the present invention is to provide a method, device and storage medium for fault diagnosis and repair of DC charging piles, which improves the recognition accuracy of fault features and ensures the optimality and reliability of the repair strategy.

[0005] To achieve the above objective, the present invention provides a method for fault diagnosis and repair of DC charging piles, including the following steps:

[0006] Perform Hilbert transform on the voltage, current, and temperature parameters of the charging pile to obtain the envelope data of multi-channel analytic signals;

[0007] Perform frequency domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain spectral feature vectors;

[0008] Input the spectral feature vectors into a dual control law observer with continuous and discontinuous terms for state tracking to obtain a system residual matrix;

[0009] Perform threshold detection and cumulative error integral operation on the system residual matrix to obtain fault feature data;

[0010] Perform λ consensus iterative operation on the fault feature data to obtain a control parameter compensation scheme.

[0011] The present invention also provides a device for fault diagnosis and repair of DC charging piles, including:

[0012] A transformation module, configured to perform Hilbert transform on the voltage, current, and temperature parameters of the charging pile to obtain the envelope data of the multi-channel analytic signal;

[0013] A decomposition module, configured to perform frequency-domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain a spectral feature vector;

[0014] A state tracking module, configured to input the spectral feature vector into a dual control law observer with continuous terms and discontinuous terms for state tracking to obtain a system residual matrix;

[0015] An integral operation module, configured to perform threshold detection and cumulative error integral operation on the system residual matrix to obtain fault feature data;

[0016] An iteration module, configured to perform λ-consensus iteration operation on the fault feature data to obtain a control parameter compensation scheme.

[0017] The present invention also provides a computer device, including a memory and a processor, where a computer program is stored in the memory, and when the processor executes the computer program, the steps of the method described in any one of the above are implemented.

[0018] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in any one of the above are implemented.

[0019] In summary, the technical solution provided by the present invention effectively improves the accuracy of fault feature extraction and reduces the interference of background noise by introducing a multi-channel data preprocessing mechanism based on Hilbert transform and combining frequency-domain analysis with a 0.1Hz high-resolution frequency grid. By adopting a dual control law observer structure and realizing the unity of fast convergence and accurate tracking through the adaptive weight allocation of discontinuous control law and continuous control law, the robustness of fault diagnosis is enhanced. The feature extraction method based on sparse envelope spectrum analysis improves the recognition accuracy of fault features through matrix sparsification processing constrained by the L1 norm. Innovatively applying the λ-consensus algorithm to repair strategy optimization significantly improves the computational efficiency of parameter optimization through Lagrangian dual decomposition and parallel computing architecture. A two-layer detector structure is designed, and by combining real-time threshold detection and cumulative error integral analysis, comprehensive monitoring of instantaneous faults and gradual faults is realized. By establishing a multi-dimensional optimization model of the repair objective function and combining the decoupled separation of control parameters and compensator parameters, the optimality and reliability of the repair strategy are ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a schematic diagram of the steps of the DC charging pile fault diagnosis and repair method in an embodiment of the present invention;

[0021] Figure 2 is the structural block diagram of the DC charging pile fault diagnosis and repair device in an embodiment of the present invention;

[0022] Figure 3 is the structural schematic block diagram of a computer device in an embodiment of the present invention.

[0023] The realization, functional features and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed implementation manners

[0024] In order to make the object, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0025] Referring to Figure 1 , this embodiment provides a method for diagnosing and repairing faults of a DC charging pile, including the following steps:

[0026] S1, perform Hilbert transform on the voltage, current, and temperature parameters of the charging pile to obtain the envelope data of the multi-channel analytic signal;

[0027] Among them, the voltage, current, and temperature parameters are input into a high-precision analog-to-digital converter for digital conversion to convert the analog signal into digital parameter data. The Hilbert kernel function based on the Fourier series is used to perform convolution calculation on the digital parameter data. The Hilbert kernel function generates the orthogonal component of the signal through the symmetry in the frequency domain, so that the real part and the imaginary part of the signal can maintain a 90-degree difference in phase. The Fourier series provides a mathematical framework to represent the signal as a linear combination of a series of sine and cosine functions, and the Hilbert kernel function uses this representation form to extract the orthogonal component data of the signal through convolution operation. These orthogonal components contain the instantaneous characteristics of the input signal. The orthogonal component data and the original digital parameter data are used for complex synthesis to generate analytic signal data. The orthogonal component is used as the imaginary part, and the digital parameter data is used as the real part, and the two together constitute the complex analytic signal. To extract the envelope amplitude data, perform the sum-of-squares operation on the real part and the imaginary part of the complex analytic signal data, and then calculate the square root value of the sum of squares. The mathematical expression of this operation is , where represents the envelope amplitude, and respectively represent the real part and the imaginary part of the signal. The envelope amplitude data generated in this step is an intuitive reflection of the signal energy change. Perform a time-domain alignment operation on the envelope amplitude data. By means of a linear interpolation-based method, map the data with different sampling frequencies or time points to a unified time axis to solve the problem of temporal inconsistency. Linear interpolation utilizes the linear relationship between adjacent data points to interpolate the unaligned signal, obtaining envelope data with time synchronization, which reflects the coordination and consistency of signals in different channels. Group and merge the synchronized envelope data according to the voltage channel, current channel, and temperature channel respectively. By organizing different types of signals, the data organization form becomes clearer, facilitating subsequent multi-channel signal analysis. The data structure after grouping and merging effectively supports subsequent fault feature extraction and diagnosis algorithms, and finally forms multi-channel analytic signal envelope data containing all channel information.

[0028] S2. Perform a frequency-domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain spectral feature vectors;

[0029] Specifically, perform a discrete Fourier transform on the envelope data. By setting a frequency resolution grid of 0.1 Hz, map the envelope data in the time domain to the frequency domain to generate initial spectral data. Through the reasonable design of the sampling frequency and the frequency resolution, capture the possible low-frequency fault characteristics in the signal while avoiding the interference of high-frequency noise. The discrete Fourier transform can transform the signal from the time domain to the frequency domain and generate a complex spectrum containing amplitude and phase information. Use the method of L1 norm constraint to perform matrix sparsification on the initial spectral data. The L1 norm constraint effectively compresses the low-intensity spectral components by introducing sparse regularization while retaining the significant frequency components with important information. The goal of sparsification is to generate a sparse spectral matrix such that only a few frequency points with significant amplitudes are retained in the matrix, and the values of the rest approach zero. Input the sparse spectral matrix into the Bayesian prior probability model for parameter optimization and statistical modeling of spectral characteristics. The introduction of the Bayesian prior probability model is to combine prior knowledge and data distribution characteristics to enhance the description ability of spectral data through statistical inference. Perform parameter estimation of the prior distribution for the sparse spectral matrix, for example, assume that its amplitude follows a Gaussian distribution or a Laplace distribution, and use these assumptions to establish a spectral prior distribution model. Iteratively calculate the spectral prior distribution through the maximum a posteriori probability criterion to obtain the posterior probability density function. In this process, the maximum a posteriori probability criterion captures the statistical characteristics of the signal and eliminates noise and interference by continuously optimizing the model parameters. Based on the posterior probability density function, use the maximum likelihood method to solve for the parameters to generate consistent data for multi-channel spectra. The core idea of maximum likelihood estimation is to find a set of parameter values that maximize the likelihood value of the given data under this model. This result ensures that the generated consistent data for multi-channel spectra can reflect the common characteristics between different signal channels while remaining sensitive to the unique information of a specific channel. Perform a real-value transformation and a dimensionality reduction operation on the consistent data. The real-value transformation converts the spectral data into a real number form by removing the complex part, facilitating subsequent analysis and processing; the dimensionality reduction operation converts the high-dimensional data into a low-dimensional feature vector through principal component analysis or other dimensionality reduction algorithms, extracting the most representative spectral feature information to obtain a spectral feature vector.

[0030] S3. Input the spectral feature vector into a dual-control-law observer with continuous and discontinuous terms for state tracking to obtain a system residual matrix;

[0031] It should be noted that the spectral feature vector is input into the state space model to analyze the dynamic characteristics of the charging pile system, and the system state equation is obtained. The state space model reflects the time-varying characteristics during the operation of the charging pile by describing the dynamic relationship between system state variables. Through the analysis of the spectral feature vector, the dynamic relationship between the input and output of the system is expressed in the form of a state equation. After establishing the system state equation, a sliding mode surface based on the sign function is designed to generate a discontinuous control law. Sliding mode control is a robust nonlinear control method that guides the system state to the sliding mode surface through the design of the sliding mode surface and moves along the sliding mode surface to achieve the control goal. The design of the discontinuous control law uses the sign function to feedback and adjust the state deviation of the system, so that the tracking accuracy and stability of the system can still be guaranteed in the presence of strong disturbances or uncertainties. However, due to the inherent discontinuity of the sign function, chattering phenomenon exists in the control input, which is improved by introducing a continuous control law. In order to achieve the continuity and robustness of system control, the system state equation is continuously processed based on the hyperbolic tangent function to generate a continuous control law. Due to its smooth transition characteristics, the hyperbolic tangent function can effectively alleviate the chattering problem caused by the discontinuous control law while maintaining sensitivity to system disturbances. The design of the continuous control law replaces the sign function with the hyperbolic tangent function, so that the control input is gradually adjusted with the change of the state error, thereby achieving smooth control of the system within a certain range. According to the error magnitude of the system state equation, an adaptive weight allocation is performed on the discontinuous control law and the continuous control law to generate a combined control law. The adaptive weight allocation dynamically adjusts the action ratio of the discontinuous control law and the continuous control law according to the amplitude and frequency changes of the system state error. When the error is large, the discontinuous control law can quickly respond and eliminate large deviations; while when the error decreases, the continuous control law gradually dominates to achieve smooth transition of the control input. The adaptive characteristics of the combined control law can balance the rapidity and stability of system control. The combined control law is input into the repetitive controller to compensate for periodic disturbances. The repetitive controller is a control strategy that can effectively suppress periodic disturbances and achieve precise control of system operation by memorizing and compensating for repetitive errors. During the compensation process, the generated compensation control amount is used to correct the deviation of the input signal, so that the system can still operate stably in a complex disturbance environment. In order to ensure the effectiveness of the compensation control amount and the stability of the system, a convergence analysis based on the Lyapunov stability theory is carried out on it. The Lyapunov stability theory verifies whether the system state can tend to be stable over time by constructing a Lyapunov function, thereby ensuring that the designed compensation control strategy has theoretical stability guarantee. After completing the state observation, a residual calculation is performed based on the observation result and the input spectral feature vector to generate an error sequence. The error sequence is the deviation performance between the actual operating state of the system and the theoretical model state, and is used to reflect the dynamic change characteristics of the system under different operating conditions.To enhance the feature representation ability of the residual, a sliding reconstruction is performed on the error sequence. The sliding reconstruction eliminates high-frequency noise interference by introducing a sliding window to perform local weighted averaging or filtering on the error sequence, while retaining the significant feature information existing in the system operation. The result after sliding reconstruction is organized into a system residual matrix.

[0032] S4. Perform threshold detection and cumulative error integral operation on the system residual matrix to obtain fault feature data;

[0033] Specifically, the real-time data segment of the system residual matrix is input into the threshold logic detector. The threshold logic detector instantaneously analyzes the real-time residual data through boundary comparison operations to determine whether it exceeds the pre-set safe operating range and generates an instantaneous fault detection result accordingly. The main objective of instantaneous fault detection is to quickly identify emerging abnormal situations and ensure that the system can capture potential fault signals at an early stage. Based on the instantaneous fault detection result, a probability density function based on the normal distribution is established to quantify the probability characteristics of fault occurrence. By statistically analyzing the residual data, its distribution parameters such as the mean and variance are estimated to construct a normal distribution model. This probability density function reflects the likelihood of a fault occurring in the current operating state of the system. On this basis, by performing interval segmentation processing on the fault probability distribution data, the threshold ranges corresponding to different fault levels are defined. To enhance the perception of the fault evolution trend, a data accumulation operation is performed on the system residual matrix to generate a sequence containing historical data. By accumulating historical data, the operating state of the system is observed on a broader time scale, and potential long-term problems can be identified more accurately. An exponentially weighted cumulative integration operation is performed on the historical data sequence to calculate the fault accumulation eigenvalue. The exponentially weighted integration assigns different time weights to historical data, highlighting the influence of recent data while avoiding the excessive interference of long-term data on the current state. After obtaining the fault accumulation eigenvalue, a multi-dimensional feature fusion operation is performed based on the relationship between the fault level threshold and the fault accumulation eigenvalue to generate a fault type flag. The multi-dimensional feature fusion comprehensively considers fault information in different dimensions, such as the magnitude of the cumulative error, the fault probability, and the operating time, to achieve a comprehensive assessment and classification of system faults. Feature extraction operations are performed on the fault type flag through sparse envelope spectrum analysis to obtain the frequency components of the fault. Sparse envelope spectrum analysis is a signal analysis method based on sparse representation that extracts key features from complex spectrum data, avoiding unnecessary spectrum redundancy. This analysis process accurately identifies the frequency components related to the fault, such as the specific frequency of mechanical vibration or the characteristic frequency of current fluctuations. The matching degree between the extracted fault frequency components and the pre-set fault characteristic patterns is calculated to determine the fault characteristic data. The pre-set fault characteristic patterns are a template library based on historical data and experience summaries, used to describe the typical characteristics of various known faults. The matching degree calculation determines whether the current fault matches a certain known fault pattern by comparing the similarity between the current frequency components and the template library, and accordingly generates the final fault characteristic data, including the category and severity of the fault, and provides a preliminary analysis result of the fault cause.

[0034] S5. Perform a λ consensus iteration operation on the fault characteristic data to obtain a control parameter compensation scheme.

[0035] Among them, the fault feature data is input into the repair target optimization model to construct a cost function for describing the charging pile fault repair target. The repair target optimization model needs to comprehensively consider various factors, such as operating efficiency, power loss, maintenance cost, and system stability, etc., to construct a multi-repair target function. Decouple and separate the control parameters and compensator parameters of the multi-repair target function to clarify the variable space involved in the optimization process. The purpose of decoupling and separation is to separate the control parameters (such as the control weights of voltage and current) from the compensator parameters (such as disturbance correction values and feedback gains), to avoid the mutual interference between variables in the optimization process. Use the Lagrange duality principle to decompose and optimize the optimization variable space to generate a set of sub-problems. The Lagrange duality principle transforms the constraint conditions of the original problem into a dual problem by introducing Lagrange multipliers, realizing the relaxation and decomposition of the problem. This step can effectively decompose the global optimization problem into a set of smaller sub-problems, and each sub-problem only focuses on part of the variables and constraint relationships. For the generated set of sub-problems, perform λ-consensus parallel computing to obtain a distributed solution set. By introducing the λ-consensus mechanism, establish a consistency constraint between sub-problems, so that the solution results of each sub-problem tend to the global optimal solution. The λ-consensus calculation performs multiple rounds of iteration among each sub-problem by sharing key parameters and update information to achieve the coordination and consistency of the solution and obtain a distributed solution set. Perform gradient descent iteration processing based on an adaptive step size on the distributed solution set to obtain an iterative convergence result. The adaptive step size mechanism dynamically adjusts the update amplitude according to the current gradient magnitude, thus avoiding the oscillation and over-jumping problems in the traditional gradient descent method. In each iteration, gradually approximate the optimal solution by reducing the error term and finally achieve global convergence. After multiple adjustments of the iterative convergence result, perform stability analysis to ensure that the obtained result has steady-state characteristics in practical applications. After obtaining the steady-state solution, optimize the parameters of the charging pile control system according to this solution. According to the repair plan provided by the steady-state solution, readjust the key parameters in the charging pile control system, such as controller gain, current response time, and voltage setting value, etc., to improve the system performance. After optimization, the control system can more efficiently cope with the impact of faults, while improving the stability and energy efficiency of operation, and obtain the final control parameter compensation scheme.

[0036] Perform convex optimization decomposition of the repair objective function on the optimized variable space to generate normalized constraint conditions. By analyzing and separating the objective function and constraint conditions, convert the non-convex problem into multiple solvable convex problems. Through this step, clarify the convexity characteristics of each part in the variable space, providing mathematical feasibility and decoupling for subsequent optimization. At the same time, the generated normalized constraint conditions can reasonably limit the range of solutions, ensuring the stability and convergence of the optimization process. Input the normalized constraint conditions into the Lagrange multiplier dual converter to construct the dual problem and generate the dual constraint equations. Lagrange duality theory decomposes the original optimization problem into multiple more tractable subproblems by introducing Lagrange multipliers, and at the same time transforms the constraint conditions into the form of dual variables. The core of the transformation lies in using the properties of the Lagrangian function, so that the solution of the problem is obtained by jointly solving the primal problem and the dual problem. In the dual constraint equations, the relationships between variables are systematically expressed as a set of equations. According to the constructed dual constraint equations, decompose the optimization problem based on the KKT conditions and complementary slackness to generate a sequence of subproblems. The KKT conditions are the necessary conditions for the optimization problem. Combining complementary slackness analysis, decompose the complex global problem into multiple local subproblems. These subproblems are relatively independent and can be solved in parallel. The generated sequence of subproblems significantly reduces the complexity of the global optimization problem by clarifying the variable ranges and constraint conditions of each subproblem. Deploy the sequence of subproblems to the parallel computing architecture to form a parallel optimization subsystem. The introduction of the parallel computing architecture can significantly improve the computing efficiency and support the distributed processing of large-scale problems, making the optimization process more efficient and scalable. In the parallel optimization subsystem, in order to achieve collaborative solution of subproblems, perform λ-consensus iterative calculation on local variables based on neighbor nodes. The λ-consensus algorithm realizes coordinated update of local variables by introducing a communication mechanism between neighbor nodes. Each node calculates the consensus update value according to its own local variables and the information of neighbor nodes, making the system as a whole gradually converge to the consensus solution. This process coordinates the solutions of each node to be near the global optimal solution through multiple rounds of iterative calculation, while taking into account local constraints and global consistency. The parallel nature of the λ-consensus algorithm enables it to make full use of distributed computing resources to ensure the efficiency and robustness of the solution. In the λ-consensus iterative calculation, in addition to the update of local variables, it is also necessary to synchronize global variables and determine convergence for the consensus solution. The goal of global variable synchronization is to aggregate the local solutions of each node into an overall result, thus ensuring the overall consistency of the optimization process. At the same time, by defining convergence determination conditions, such as the change amplitude of the solution or the convergence rate of the objective function, dynamically monitor the progress of the solution process and terminate the iteration when the convergence conditions are met. After completing global synchronization and convergence determination, generate a distributed solution set containing the global optimal solution.

[0037] In one example, Hilbert transform is performed on the voltage, current, and temperature parameters of a charging pile to obtain the envelope data of multi-channel analytic signals, including:

[0038] The voltage, current, and temperature parameters are input into a high-precision analog-to-digital converter for digital conversion to obtain digital parameter data;

[0039] Convolution calculation is performed on the digital parameter data through a Hilbert kernel function based on Fourier series to obtain quadrature component data;

[0040] According to the quadrature component data and the digital parameter data, complex synthesis is performed to obtain analytic signal data, and square root operation of the sum of squares of the real part and the imaginary part of the analytic signal data is performed to obtain envelope amplitude data;

[0041] Time-domain alignment based on linear interpolation is performed on the envelope amplitude data to obtain synchronous envelope data, and the synchronous envelope data is grouped and merged according to the voltage, current, and temperature channels to obtain the envelope data of multi-channel analytic signals.

[0042] In this example, the key analog signals (voltage , current , temperature ) of the charging pile are converted into corresponding digital signals through a high-precision analog-to-digital converter. These parameters are the core indicators reflecting the operating state of the charging pile. Assuming the sampling frequency is , then at any time point , the analog signal is discretized into a digital signal , where is the sampling value at the corresponding time point. Taking the voltage signal as an example, if represents a sine wave signal, after being converted by the analog-to-digital converter, its discretized form is , where is the signal frequency, is the sampling frequency. Convolution calculation is performed on the digital parameter data through a Hilbert kernel function based on Fourier series to obtain quadrature component data. The Hilbert transform generates the quadrature component (imaginary part) of the signal through a phase shift operation in the frequency domain. For the discretized signal , its Hilbert kernel function is expressed as (when ) and . By performing convolution calculation on the signal and , the quadrature component is obtained, that is:

[0043] ;

[0044] Among them, is the orthogonal component (imaginary part) of the signal, is the sampled value of the original signal, is the corresponding value of the kernel function. For a voltage signal , its orthogonal component is a cosine signal with a phase shift of , that is . By obtaining the orthogonal component data and the original signal , complex synthesis is performed to obtain the analytic signal data . The formula for complex synthesis is:

[0045] ;

[0046] where, is the analytic signal, is the imaginary unit, is the real part of the signal, is the imaginary part of the signal. For example, for a voltage signal and its orthogonal component , the analytic signal is written as:

[0047] ;

[0048] Performing the square root operation of the sum of the squares of the real and imaginary parts on the analytic signal to obtain the envelope amplitude data . The formula for the envelope amplitude is:

[0049] ;

[0050] For the above voltage signal , its envelope amplitude is:

[0051] ;

[0052] Performing time domain alignment based on linear interpolation on the envelope amplitude data . Unifying the data of different channels (such as voltage, current, temperature) to the same time reference. Assuming the sampling frequency of the voltage signal is , the current signal is , and the temperature signal is , through linear interpolation, map the envelope amplitude data of these signals to the unified time axis. For example, the linear interpolation formula is:

[0053] ;

[0054] where is the interpolated value, and are the two nearest sampling points, and are the corresponding sampling values. By interpolation, the envelope amplitudes of voltage, current, and temperature are aligned to the same time axis to form synchronous envelope data. The synchronous envelope data is grouped and merged according to the voltage, current, and temperature channels to generate the envelope data of the multi-channel analytical signal. For example, for the envelope data of the voltage channel , the envelope data of the current channel , and the envelope data of the temperature channel , the multi-channel envelope data is represented in matrix form as:

[0055] .

[0056] In one example, the envelope data is subjected to frequency-domain conversion and maximum a posteriori probability matrix factorization to obtain spectral feature vectors, including:

[0057] The envelope data is subjected to discrete Fourier transform through a frequency grid with a resolution of 0.1 Hz to obtain initial spectral data, and the initial spectral data is subjected to matrix sparsification with L1 norm constraint to obtain a sparse spectral matrix;

[0058] The sparse spectral matrix is input into a Bayesian prior probability model for parameter optimization to obtain a spectral prior distribution, and the spectral prior distribution is iteratively calculated through the maximum a posteriori probability criterion to obtain a posterior probability density function;

[0059] According to the posterior probability density function, maximum likelihood parameter solving is performed to obtain multi-channel spectral consistency data, and the multi-channel spectral consistency data is subjected to real-value transformation and dimensionality reduction to obtain spectral feature vectors.

[0060] In this example, for the envelope data a discrete Fourier transform is performed, a frequency grid with a resolution of 0.1 Hz is selected, and the interval within the frequency range is 0.1 Hz. Assume the sampling frequency is , and the number of sampling points is . The frequency resolution is defined as . To meet the resolution of 0.1 Hz, it must be ensured that is large enough such that Hz. The formula for discrete Fourier transform is:

[0061] ;

[0062] where is the spectral value corresponding to the frequency , is the value of the envelope data at the th sampling point, is the imaginary unit, is the frequency index. By calculating , the initial spectral data is obtained, which contains the amplitude and phase information of the signal. Matrix sparsification with L1-norm constraint is performed on the initial spectral data to obtain a sparse spectral matrix. The L1-norm constraint makes the optimization objective function preferentially select a small number of important frequency components during solution by adding a sparse regularization term, while compressing the remaining irrelevant components to zero. Let be the initial spectral matrix, and the sparsification problem is expressed as the following optimization problem:

[0063] ;

[0064] where is the reconstruction error term, representing the difference between the sparse matrix and the original spectral matrix ; is the sparse regularization term, representing the L1-norm of the matrix (i.e., the sum of the absolute values of all elements); is the sparsification weight parameter, controlling the balance between sparsity and reconstruction error. By solving this optimization problem, the sparse spectral matrix is obtained, which only retains the frequency components that are significantly meaningful for the signal. The sparse spectral matrix is input into the Bayesian prior probability model for parameter optimization to obtain the spectral prior distribution. The Bayesian model defines the prior distribution by assuming that the amplitude of the sparse spectrum follows a certain probability distribution (such as Gaussian distribution or Laplace distribution) and combining historical data or expert knowledge. Let the likelihood function of the measurement data be , and the posterior distribution is obtained through Bayes' formula:

[0065] ;

[0066] where is the marginal distribution, ensuring the normalization of the posterior distribution. Through iterative calculation of the spectral prior distribution by the maximum a posteriori probability criterion, a set of parameters is found to maximize . The optimization problem of the maximum a posteriori probability is:

[0067] ;

[0068] In each iteration, the parameters are updated to approximate the optimal posterior probability density function. Based on the posterior probability density function, the maximum likelihood estimation method is used to solve the multi-channel spectral consistency data. The maximum likelihood method maximizes the probability of the data appearance by selecting a set of parameters . The likelihood function of the consistency data is defined as , and its optimization objective is:

[0069] ;

[0070] Among them, represents the spectrum consistency parameter, represents the number of channels, is the sparse spectrum data of the th channel. The consistency data of the multi-channel spectrum is obtained by solving this problem. The real-valued transformation and dimensionality reduction processing are performed on the multi-channel spectrum consistency data to generate the spectrum feature vector. The real-valued transformation takes half of the frequency band of the sparse spectrum to convert the complex spectrum into a symmetric real form, and the formula is:

[0071] ;

[0072] Among them, and represent the real part and the imaginary part of the spectrum respectively. Through the dimensionality reduction algorithm (such as principal component analysis), the most representative low-dimensional feature vectors are extracted from the high-dimensional data. For example, for the spectrum data matrix , principal component analysis finds the main direction of the data through eigenvalue decomposition and generates the spectrum feature vector , that is:

[0073] ;

[0074] Among them, is the dimensionality reduction matrix, which defines the weights of the main components. Through the above steps, the conversion, sparsification, Bayesian modeling and feature extraction of the spectrum data are realized, and the spectrum feature vector for fault diagnosis is generated.

[0075] In an example, the spectrum feature vector is input into a dual control law observer with continuous terms and discontinuous terms for state tracking, and the system residual matrix is obtained, including:

[0076] The spectrum feature vector is input into the state space model for dynamic characteristic analysis to obtain the system state equation;

[0077] According to the system state equation, the sliding mode surface design based on the sign function is established to obtain the discontinuous control law, and the continuous processing based on the hyperbolic tangent function is performed on the system state equation to obtain the continuous control law;

[0078] According to the error magnitude of the system state equation, the adaptive weight allocation is performed on the discontinuous control law and the continuous control law to obtain the combined control law;

[0079] The combined control law is input into the repetitive controller for periodic disturbance compensation to obtain the compensation control quantity, and the convergence analysis based on Lyapunov stability is performed on the compensation control quantity to obtain the state observation result;

[0080] Based on the state observation results and the spectral feature vectors, residual calculation is performed to obtain an error sequence, and the error sequence is slid and reconstructed to obtain a system residual matrix.

[0081] In this example, the spectral feature vectors are input into the state space model to analyze the dynamic characteristics of the system. Assume that the dynamic behavior of the system is described by a discrete-time state space model in the form of:

[0082] ;

[0083] ;

[0084] where represents the state vector of the system at the -th moment; is the control input; is the output of the system; are the system matrix, input matrix, and output matrix, respectively, describing the dynamic relationship of the system; and represent process noise and measurement noise, respectively. By analyzing the distribution and evolution of the spectral feature vectors, the parameters of are estimated to obtain the state equation describing the dynamic characteristics of the system. After obtaining the system state equation, a sliding mode surface is designed based on the sign function to construct a discontinuous control law. Sliding mode control is to force the system state onto the sliding mode surface and slide along the sliding mode surface to achieve the control objective. The definition of the sliding mode surface is:

[0085] ;

[0086] where is the sliding mode surface function, and its zero point represents the ideal system trajectory. The discontinuous control law is defined by the sign function as:

[0087] ;

[0088] where is the sliding mode gain; is the sign function, defined as:

[0089] ;

[0090] The role of the sign function is to quickly respond to the deviation, but its inherent discontinuity has a chance to introduce chattering problems. To alleviate the chattering phenomenon caused by the discontinuous control law, the system state equation is continuously processed based on the hyperbolic tangent function, thereby designing a continuous control law. The hyperbolic tangent function is a smooth nonlinear function, defined as:

[0091] ;

[0092] Based on this, the expression of the continuous control law is as follows:

[0093] ;

[0094] Among them, is the continuous control gain. Compared with the sign function, the hyperbolic tangent function can achieve a smooth transition, reduce the jitter of the control input, and is suitable for dealing with small-range errors. According to the error magnitude of the system state equation, an adaptive weight allocation is performed on the discontinuous control law and the continuous control law to generate a combined control law. The strategy of the adaptive weight allocation is to dynamically adjust the ratio of the two control laws based on the error amplitude. Assume that the amplitude of the error is , then the weight allocation is defined as:

[0095] ;

[0096] Among them, and are the weights of the discontinuous control law and the continuous control law; is a small positive number used to prevent the denominator from being zero. The expression of the combined control law is:

[0097] ;

[0098] Through weight allocation, the combined control law combines the rapidity of discontinuous control and the smoothness of continuous control. The combined control law is input into the repetitive controller to compensate for periodic disturbances. The repetitive controller is a control method optimized for periodic signals. By memorizing and compensating for repetitive errors, the steady-state deviation of the system is gradually reduced. Let be the compensation control quantity, and its update rule is:

[0099] ;

[0100] Among them, is the desired output, is the compensation gain. After multiple rounds of iteration, the output of the repetitive controller can effectively eliminate periodic disturbances. A convergence analysis of the compensation control quantity based on the Lyapunov stability theory is performed to verify the stability of the system. Define the Lyapunov function as:

[0101] ;

[0102] Among them, is a symmetric positive definite matrix. If monotonically decreases with time, then the system is Lyapunov stable. According to the obtained state observation results and the spectral feature vector a residual calculation is performed to generate an error sequence:

[0103] ;

[0104] Perform sliding reconstruction on the error sequence by introducing a sliding window and perform weighted averaging:

[0105] ;

[0106] The error sequence after sliding reconstruction is used to generate the system residual matrix , and its form is:

[0107] ;

[0108] The system residual matrix is used to characterize the deviation in the system dynamic characteristics.

[0109] In one example, perform threshold detection and cumulative error integration operation on the system residual matrix to obtain fault feature data, including:

[0110] Input the real-time data segment of the system residual matrix into a threshold logic detector for boundary comparison operation to obtain instantaneous fault detection results;

[0111] Establish a probability density function based on normal distribution according to the instantaneous fault detection results to obtain fault probability distribution data;

[0112] Perform interval segmentation processing on the fault probability distribution data to obtain fault level thresholds, and perform data accumulation operation on the system residual matrix to obtain a historical data sequence;

[0113] Perform exponentially weighted cumulative integration operation on the historical data sequence to obtain fault cumulative eigenvalues, and perform multi-dimensional feature fusion operation according to the fault level thresholds and fault cumulative eigenvalues to obtain fault type flags;

[0114] Perform feature extraction operation on the fault type flags through sparse envelope spectrum analysis to obtain fault frequency components, and calculate the matching degree according to the fault frequency components and preset fault feature patterns to obtain fault feature data.

[0115] In this example, the real-time data segment of the system residual matrix is input into a threshold logic detector, and boundary comparison operation is performed on it to identify instantaneous faults. Assume that the normal operating state of the system corresponds to an allowable residual range , then for each time point , the judgment condition is defined as:

[0116] ;

[0117] where is the instantaneous fault detection result. A value of 1 indicates an anomaly, and a value of 0 indicates normal. By real-time analysis , quickly determine whether the system exhibits abnormal behavior. Based on the instantaneous fault detection result, establish a probability density function based on the normal distribution to quantify the probability characteristics of fault occurrence. Assume that the residuals of the system follow a normal distribution under normal conditions, and the probability density function is expressed as:

[0118] ;

[0119] where is the mean of the residuals, reflecting the central value under normal conditions; is the standard deviation, describing the degree of data dispersion; is the residual value. By statistically analyzing the instantaneous fault detection results, estimate the values of and , thereby constructing the fault probability distribution. Perform interval segmentation processing on the fault probability distribution data and define different fault level thresholds. For example, according to the distribution characteristics of the probability density, it is divided into three levels: minor fault, medium fault, and severe fault, corresponding to the threshold intervals , and . The segmentation processing can convert continuous fault probability information into discrete fault levels. To capture the long-term behavior characteristics of the system, perform data accumulation operation on the system residual matrix to generate a historical data sequence . Assume that is the length of the cumulative time window, then the definition of the historical data sequence is:

[0120] ;

[0121] where represents the cumulative residual from time to . By introducing a time window, smooth the short-term fluctuations of the data and highlight the long-term trend. Perform exponentially weighted cumulative integration operation on the historical data sequence to calculate the fault cumulative eigenvalue . The definition of exponentially weighted integration is:

[0122] ;

[0123] where is the weighting factor, used to control the importance decay of historical data; is the length of the time window. Through exponential weighting, highlight the contribution of recent data to fault characteristics and reduce the impact of long-term data. According to the fault level threshold and the fault cumulative eigenvalue , perform multi-dimensional feature fusion operations to generate fault type flags. Assume that exceeds different thresholds and corresponds to different fault types respectively, then the definition of the fault type flag is:

[0124] ;

[0125] Through feature fusion operations, dynamically identify the fault type according to the changing trend of the cumulative feature value. Based on sparse envelope spectrum analysis, perform feature extraction operations on the fault type flag to extract fault frequency components. Let the fault signal have a sparse spectrum of , and its calculation formula is:

[0126] ;

[0127] Through sparsification processing, highlight the key frequency components in the spectrum and filter out noise interference. For the fault type flag , the extracted frequency components represent the characteristic frequencies of system faults. Compare the fault frequency components with the preset fault characteristic patterns to calculate the matching degree and obtain the fault characteristic data. The definition of the matching degree uses a similarity measurement formula, such as cosine similarity:

[0128] ;

[0129] Among them, is the extracted frequency component vector, is the preset fault mode vector. By calculating the similarity, judge the matching degree of the current fault with the known pattern, and then output the fault characteristic data.

[0130] In an example, perform λ consensus iterative operations on the fault characteristic data to obtain a control parameter compensation scheme, including:

[0131] Input the fault characteristic data into the repair target optimization model to construct a cost function and obtain a multi-repair target function;

[0132] Decouple and separate the control parameters and compensator parameters of the multi-repair target function to obtain an optimization variable space;

[0133] Decompose and optimize the optimization variable space through the Lagrangian dual principle to obtain a set of sub-problems, and perform λ consensus parallel calculations on the set of sub-problems to obtain a distributed solution set;

[0134] Perform gradient descent iterative processing with an adaptive step size on the distributed solution set to obtain an iterative convergence result, and perform stability analysis on the iterative convergence result to obtain a steady-state solution;

[0135] Optimize the control parameters of the charging pile control system according to the steady-state solution to obtain a control parameter compensation scheme.

[0136] In this example, the fault feature data is input into the repair target optimization model to construct a multi-repair target function. Considering the multi-dimensional performance indicators of the charging pile operation comprehensively, including power stability , energy loss , temperature deviation , etc. Assume that these indicators are defined as follows:

[0137] ;

[0138] ;

[0139] ;

[0140] Among them, is the actual power, is the target power, is the loss power, and are the actual and target temperatures respectively. The constructed multi-repair target function is a weighted combination of the above performance indicators:

[0141] ;

[0142] Among them, is the weight coefficient, reflecting the importance of each indicator. Decouple and separate the control parameters and the compensator parameters of the multi-repair target function to obtain an optimization variable space. The control parameters include voltage, current and temperature control variables; the compensator parameters are used for dynamic compensation of disturbances and faults. The optimization variable space is defined as:

[0143] ;

[0144] Among them, is the inequality constraint, representing the physical boundary conditions, such as the upper and lower limits of voltage and current; is the equality constraint, representing the system balance condition. Through decoupling and separation, the optimization problem is divided into two independent sub-spaces, and the control parameters and compensator parameters are optimized respectively. To solve this optimization problem, use the Lagrangian duality principle to decompose and optimize the optimization variable space to generate a set of sub-problems. The Lagrangian function of the optimization problem is defined as:

[0145] ;

[0146] Among them, and are Lagrange multipliers used to introduce constraint conditions. Through the duality principle, the original problem is decomposed into multiple sub-problems:

[0147] ;

[0148] ;

[0149] Through this step, the global optimization problem is decomposed into a set of independent sub-problems. For the generated set of sub-problems, the consensus parallel computing method is adopted to achieve decoupled cooperation among multiple nodes. Consensus computing achieves consistency through the following update rules:

[0150] ;

[0151] ;

[0152] Among them, and are step size parameters, and are the gradients of the control parameter and the compensator parameter respectively. In each iteration, each node shares update information through neighbor nodes and gradually achieves coordination of the global solution. The distributed solution set is processed by gradient descent iteration with an adaptive step size to improve the solution accuracy. The update rule of the adaptive step size strategy is:

[0153] ;

[0154] ;

[0155] Among them, is the step size of the th iteration, is the step size adjustment factor. By dynamically adjusting the step size, a balance is achieved between fast convergence and stability. After the iteration converges, a stability analysis is performed on the result to verify the reliability of the system solution. The stability analysis is achieved through the Lyapunov function, defined as:

[0156] ;

[0157] Among them, is the state variable, is a positive definite matrix. If , then the system is stable. According to the steady-state solution, the parameters of the charging pile control system are optimized to obtain a control parameter compensation scheme.

[0158] In one example, the optimization variable space is decomposed and optimized through the Lagrangian dual principle to obtain a set of sub-problems, and the λ-consensus parallel calculation is performed on the set of sub-problems to obtain a distributed solution set, including:

[0159] Perform convex optimization decomposition of the repair objective function on the optimization variable space to obtain normalized constraint conditions;

[0160] Input the normalized constraint conditions into the Lagrange multiplier dual transformer to construct the dual problem and obtain the dual constraint equation;

[0161] According to the dual constraint equation, perform KKT condition and complementary slackness decomposition to obtain a sequence of sub-problems, and perform distributed deployment of the sequence of sub-problems based on the parallel computing architecture to obtain a parallel optimization subsystem;

[0162] Perform λ-consensus iterative calculation on the local variables in the parallel optimization subsystem based on neighbor nodes to obtain a consistent solution, and perform global variable synchronization and convergence determination on the consistent solution to obtain a distributed solution set.

[0163] In this example, perform convex optimization decomposition of the repair objective function on the optimization variable space to obtain normalized constraint conditions. Assume the repair objective function is , where represents the state variable, represents the control variable, and the optimization objective is to minimize the value of while satisfying the constraint conditions. To facilitate the solution, the non-convex problem is decomposed into convex sub-problems. The optimization problem is expressed as:

[0164] ;

[0165] where, and represent inequality constraints and equality constraints respectively. Through convex optimization decomposition, the objective function is divided into multiple convex sub-functions , and the constraint conditions are normalized into convex constraints:

[0166] ;

[0167] where, and are the local variables of the -th sub-problem respectively, and are the local constraint functions of the sub-problem. The decomposed problem is convenient for parallel solution. Input the normalized constraint conditions into the Lagrange multiplier dual transformer to construct the dual problem and generate the dual constraint equation. By introducing the Lagrange multipliers and , the Lagrangian function of the optimization problem is defined as:

[0168] ;

[0169] in, is the Lagrange multiplier corresponding to the inequality constraint, is the Lagrange multiplier corresponding to the equality constraint. The goal of the dual problem is to maximize the dual function:

[0170] ;

[0171] Through The dual decomposition of the global problem is transformed into multiple local dual constraint equations, each of which contains only local variables and constraints. According to the dual constraint equations, the dual problem is decomposed into a sequence of sub-problems using the KKT condition (Karush-Kuhn-Tucker condition) and complementary slackness decomposition. The KKT condition includes the following four parts: the original problem constraint and ; Dual feasibility ; Gradient conditions ; complementary slackness Through these conditions, the global problem is decomposed into several local sub-problems:

[0172] ;

[0173] in, is a local Lagrangian function, which only contains the The subproblems are deployed to a parallel computing architecture to improve computational efficiency, forming parallel optimization subsystems. Each subsystem solves its corresponding local optimization problem independently, while sharing key information with neighboring nodes to achieve collaborative solutions. In the parallel optimization subsystem, local variables are optimized based on neighboring nodes to ensure the consistency of the global solution. Consensus iteration calculation. The goal of consensus iteration is to make all local solutions gradually converge to the same through the interaction of neighboring nodes. and There is a communication link between them, and the consensus update rules are:

[0174] ;

[0175] ;

[0176] in, Representation Node The neighbor set of and is the step size parameter. Through multiple rounds of iteration, the variables of each node gradually tend to be consistent. Global variable synchronization and convergence determination are performed on the consistent solution to generate a distributed solution set. The synchronization process is to summarize each local solution and into a global solution:

[0177] ;

[0178] where is the total number of subsystems. Convergence determination is completed by evaluating whether the variable change amount and the objective function change amount are less than the preset threshold .

[0179] Referring to Figure 2 , this embodiment provides a DC charging pile fault diagnosis and repair device, including:

[0180] Transformation module 1, used to perform Hilbert transformation on the voltage, current, and temperature parameters of the charging pile to obtain the envelope data of the multi-channel analytic signal;

[0181] Decomposition module 2, used to perform frequency domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain the spectral feature vector;

[0182] State tracking module 3, used to input the spectral feature vector into a dual control law observer with continuous terms and discontinuous terms for state tracking to obtain the system residual matrix;

[0183] Integral operation module 4, used to perform threshold detection and cumulative error integral operation on the system residual matrix to obtain the fault feature data;

[0184] Iteration module 5, used to perform λ consensus iteration operation on the fault feature data to obtain the control parameter compensation scheme.

[0185] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to that described in the above method embodiment, and details are not described herein again.

[0186] Referring to Figure 3 , this embodiment of the present invention also provides a computer device, which can be a server, and its internal structure can be as Figure 3As shown in the figure. The computer device includes a processor, a memory, a display screen, an input device, a network interface, and a database connected through a system bus. Among them, the processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program, when executed by the processor, implements the above method.

[0187] Those skilled in the art can understand that Figure 3 the structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0188] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0189] Those of ordinary skill in the art can understand that all or part of the processes in the above method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above method embodiments. Among them, any reference to memory, storage, database, or other media provided by the present invention and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or an external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0190] It should be noted that in this text, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, apparatus, article or method comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article or method. Without further limitation, an element defined by the phrase "comprising an..." does not exclude the presence of additional identical elements in the process, apparatus, article or method comprising such element.

[0191] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall equally be included in the patent protection scope of the present invention.

Claims

1. A method for diagnosing and repairing faults of a DC charging pile, characterized in that, Including the following steps: Perform Hilbert transform on the voltage, current, and temperature parameters of the charging pile to obtain the envelope data of the multi-channel analytic signal; Perform frequency-domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain the spectral feature vector; Input the spectral feature vector into a dual control law observer with continuous and discontinuous terms for state tracking to obtain the system residual matrix; Perform threshold detection and cumulative error integration operation on the system residual matrix to obtain the fault feature data; Perform λ consensus iterative operation on the fault feature data to obtain the control parameter compensation scheme; specifically including: input the fault feature data into a repair target optimization model to construct a cost function to obtain a multi-repair target function; decouple and separate the control parameters and compensator parameters of the multi-repair target function to obtain an optimization variable space; decompose and optimize the optimization variable space through the Lagrange duality principle to obtain a set of sub-problems, and perform λ consensus parallel calculation on the set of sub-problems to obtain a distributed solution set; for the generated set of sub-problems, perform λ consensus parallel calculation to obtain a distributed solution set. By introducing the λ consensus mechanism, establish a consistency constraint between sub-problems, so that the solution results of each sub-problem tend to the global optimal solution. The λ consensus parallel calculation performs multiple rounds of iteration between each sub-problem by sharing key parameters and update information to achieve the coordination and consistency of the solution, and obtain a distributed solution set; perform gradient descent iterative processing based on an adaptive step size on the distributed solution set to obtain an iterative convergence result, and perform stability analysis on the iterative convergence result to obtain a steady-state solution; optimize the parameters of the control system of the charging pile according to the steady-state solution to obtain the control parameter compensation scheme.

2. The DC charging pile fault diagnosis and repair method according to claim 1, wherein The performing Hilbert transform on the voltage, current, and temperature parameters of the charging pile to obtain the envelope data of the multi-channel analytic signal includes: Input the voltage, current, and temperature parameters into a high-precision analog-to-digital converter for digital conversion to obtain digital parameter data; Perform convolution calculation on the digital parameter data through a Hilbert kernel function based on Fourier series to obtain quadrature component data; Perform complex synthesis based on the quadrature component data and the digital parameter data to obtain analytic signal data, and perform square root operation on the sum of the squares of the real part and the imaginary part of the analytic signal data to obtain envelope amplitude data; Perform time-domain alignment based on linear interpolation on the envelope amplitude data to obtain synchronous envelope data, and group and merge the synchronous envelope data according to the voltage, current, and temperature channels to obtain the envelope data of the multi-channel analytic signal.

3. The method for diagnosing and repairing faults of a DC charging pile according to claim 1, wherein The performing frequency-domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain the spectral feature vector includes: Perform discrete Fourier transform on the envelope data through a frequency grid with a resolution of 0.1 Hz to obtain initial spectral data, and perform matrix sparsification with L1 norm constraint on the initial spectral data to obtain a sparse spectral matrix; Input the sparse spectrum matrix into the Bayesian prior probability model for parameter optimization to obtain the spectral prior distribution, and perform iterative calculation on the spectral prior distribution through the maximum a posteriori probability criterion to obtain the posterior probability density function; Solve the maximum likelihood parameters according to the posterior probability density function to obtain multi-channel spectral consistency data, and perform real-value transformation and dimensionality reduction on the multi-channel spectral consistency data to obtain spectral feature vectors.

4. The method for diagnosing and repairing faults of a DC charging pile according to claim 1, characterized in that The step of inputting the spectral feature vector into a dual control law observer with continuous and discontinuous terms for state tracking to obtain a system residual matrix includes: Input the spectral feature vector into a state space model for dynamic characteristic analysis to obtain a system state equation; Establish a sliding mode surface design based on the sign function according to the system state equation to obtain a discontinuous control law, and perform continuous processing on the system state equation based on the hyperbolic tangent function to obtain a continuous control law; Perform adaptive weight allocation on the discontinuous control law and the continuous control law according to the error magnitude of the system state equation to obtain a combined control law; Input the combined control law into a repetitive controller for periodic disturbance compensation to obtain a compensation control quantity, and perform convergence analysis based on Lyapunov stability on the compensation control quantity to obtain a state observation result; Perform residual calculation based on the state observation result and the spectral feature vector to obtain an error sequence, and perform sliding reconstruction on the error sequence to obtain a system residual matrix.

5. The method for diagnosing and repairing faults of a DC charging pile according to claim 1, characterized in that, The step of performing threshold detection and cumulative error integral operation on the system residual matrix to obtain fault feature data includes: Input the real-time data segment of the system residual matrix into a threshold logic detector for boundary comparison operation to obtain an instantaneous fault detection result; Establish a probability density function based on the normal distribution according to the instantaneous fault detection result to obtain fault probability distribution data; Perform interval segmentation processing on the fault probability distribution data to obtain fault level thresholds, and perform data accumulation operation on the system residual matrix to obtain a historical data sequence; Perform exponentially weighted cumulative integral operation on the historical data sequence to obtain a fault cumulative eigenvalue, and perform multi-dimensional feature fusion operation based on the fault level threshold and the fault cumulative eigenvalue to obtain a fault type flag; Perform feature extraction operation on the fault type flag through sparse envelope spectrum analysis to obtain fault frequency components, and perform matching degree calculation based on the fault frequency components and a preset fault feature pattern to obtain fault feature data.

6. The method for diagnosing and repairing faults of a DC charging pile according to claim 1, wherein The step of decomposing and optimizing the optimization variable space through the Lagrangian dual principle to obtain a set of sub-problems and performing λ consensus parallel calculation on the set of sub-problems to obtain a distributed solution set includes: Perform convex optimization decomposition of the repair objective function on the optimization variable space to obtain a normalized constraint condition; Input the normalized constraint condition into a Lagrangian multiplier dual converter for dual problem construction to obtain a dual constraint equation; Perform KKT condition and complementary slackness decomposition according to the dual constraint equation to obtain a sequence of sub-problems, and perform distributed deployment of the sequence of sub-problems based on a parallel computing architecture to obtain a parallel optimization subsystem; Perform λ consensus iterative calculation of local variables in the parallel optimization subsystem based on neighbor nodes to obtain a consistent solution, and perform global variable synchronization and convergence determination on the consistent solution to obtain a distributed solution set; Among them, in the parallel optimization subsystem, in order to achieve collaborative solution of sub-problems, perform λ consensus iterative calculation of local variables based on neighbor nodes. The λ consensus algorithm realizes coordinated update of local variables by introducing a communication mechanism between neighbor nodes. Each node calculates a consistent update value according to its own local variables and the information of neighbor nodes, so that the system as a whole gradually converges to a consistent solution. Through multiple rounds of iterative calculation, the solutions of each node are coordinated to the vicinity of the global optimal solution, while taking into account local constraints and global consistency. In the λ consensus iterative calculation, in addition to the update of local variables, it is also necessary to perform global variable synchronization and convergence determination on the consistent solution. The goal of global variable synchronization is to summarize the local solutions of each node into an overall result, so as to ensure the overall consistency of the optimization process. At the same time, by defining convergence determination conditions and terminating the iteration when the convergence conditions are met, after completing global synchronization and convergence determination, a distributed solution set including the global optimal solution is generated.

7. A DC charging pile fault diagnosis and repair device, characterized in that, For implementing the steps of the method according to any one of claims 1 to 6, the device includes: A transformation module, configured to perform Hilbert transform on the voltage, current, and temperature parameters of the charging pile to obtain envelope data of multi-channel analytic signals; A decomposition module, configured to perform frequency domain conversion and maximum a posteriori probability matrix decomposition on the envelope data to obtain a spectral feature vector; A state tracking module, configured to input the spectral feature vector into a dual control law observer with continuous terms and discontinuous terms for state tracking to obtain a system residual matrix; An integral operation module, configured to perform threshold detection and cumulative error integral operation on the system residual matrix to obtain fault feature data; An iteration module, configured to perform λ consensus iterative operation on the fault feature data to obtain a control parameter compensation scheme.

8. A computer device, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 6.

Images