New energy vehicle charger automatic test system

Abstract

The present application discloses an automated testing system for a new energy vehicle charger, and relates to the field of automated testing technology. The system includes: a data acquisition part, which is used to collect the operating data of the new energy vehicle charger in real time, pre-process the collected operating data, and obtain pre-processed operating data; a charger model construction part, which is used to construct the charger state equation and charger measurement equation of the charger based on the pre-processed operating data using the extended Kalman filter method; an automated testing part, which is used to calculate the charging efficiency and the fault probability based on the charger state equation and the charger measurement equation. The present invention realizes high-precision calculation of charging efficiency and high-precision evaluation of fault probability, thereby improving the accuracy of the test.

Description

Technical Field

[0001] The present application relates to the field of automated testing technology, and in particular to an automated testing system for a new energy vehicle charger. Background Art

[0002] The development of new energy vehicles has made significant progress in the past decade. As one of the key technologies of new energy vehicles, charging technology has received widespread attention. With the continuous increase in the number of new energy vehicles, the requirements for charging equipment are also increasing, especially in terms of charging speed, efficiency, safety and intelligence. However, there are still many problems with existing charging technology, which limits the further popularization and application of new energy vehicles.

[0003] In the existing charging technology, safety and intelligence level are also important considerations. Traditional charging equipment usually relies on fixed charging modes and simple control strategies, which makes it difficult to achieve real-time monitoring and intelligent adjustment of the charging process. With the increase in battery capacity and energy density of new energy vehicles, the risk of overheating, overvoltage, overcurrent and other faults during charging is also increasing. If these faults cannot be detected and handled in time, it will not only affect the charging efficiency and battery life, but may even cause safety accidents. In the existing technology, the testing and fault detection of chargers mainly rely on manual operation and simple testing instruments. This method is not only inefficient, but also easily affected by human factors, and the accuracy and reliability of the test results cannot be fully guaranteed. Especially in the construction and operation of large-scale charging stations, traditional testing and maintenance methods are difficult to meet the requirements of efficient and safe operation. Therefore, how to improve the level of automated testing of chargers and achieve real-time monitoring and intelligent management of the charging process has become a problem that the industry needs to solve urgently. Summary of the invention

[0004] The embodiment of the present application provides an automated testing system for a new energy vehicle charger. The present invention achieves high-precision calculation of charging efficiency and high-precision evaluation of failure probability, thereby improving the accuracy of the test.

[0005] An embodiment of the present application provides an automated testing system for a new energy vehicle charger, the system comprising: a data acquisition part, for real-time acquisition of operating data of a new energy vehicle charger, preprocessing the acquired operating data to obtain preprocessed operating data; a charger model construction part, for constructing a charger state equation and a charger measurement equation of the charger based on the preprocessed operating data using an extended Kalman filter method; an automated testing part, for calculating the charging efficiency and the fault probability based on the charger state equation and the charger measurement equation.

[0006] Furthermore, the operating data at least includes: voltage, current and temperature.

[0007] Furthermore, suppose that the operating data of the new energy vehicle charger collected in real time by the data collection part is Where V i represents the voltage at the i-th time step, I i represents the current at the i-th time step, T i represents the temperature of the i-th time step; N is the total number of time steps; suppose the preprocessing operation data is The following formula is used to calculate V i After processing, we get:

[0008]

[0009] in, is the Fourier series expansion term, which is used to convert V i Perform frequency domain analysis and reconstruction; a k is the cosine coefficient of the kth term of the Fourier series expansion, which represents the amplitude of the signal on the cosine basis function; b k is the sine coefficient of the kth term of the Fourier series expansion, which represents the amplitude of the signal on the sine basis function; k is the index of the number of terms in the Fourier series, ranging from 1 to M, where M is the total number of terms in the Fourier series; L is the period of the Fourier series expansion; is a nonlinear transformation integral term, which is used to capture the nonlinear characteristics in voltage data; λ is the exponential attenuation coefficient, which controls the attenuation rate of the exponential term. The larger the λ, the faster the attenuation; t is the integral variable, ranging from 0 to V i .

[0010] Furthermore, I is calculated by the following formula i After processing, we get:

[0011]

[0012] Among them, ψ represents the mother wavelet function, a j and b j are the scale and displacement parameters of the wavelet transform, α and β are the fractal dimension correction parameters; J represents the upper limit of the scale parameter of the wavelet transform; The following formula is used to calculate T i After processing, we get:

[0013]

[0014] Among them, P n represents the nth order Legendre polynomial, σ is the standard deviation of the adaptive filter; τ is the integral variable, and its range is the entire real number domain.

[0015] Furthermore, the charger model is constructed based on the preprocessed operation data, and the charger state equation of the charger constructed using the extended Kalman filter method is expressed using the following formula:

[0016]

[0017] Among them, A k and B k are the state transfer matrix and the control input matrix respectively; ∈ is a positive number to prevent division by zero; x k is the extended Kalman filter result of the kth state; x k+1 is the extended Kalman filter result of the k+1th state; u k is the control input vector in the k state; w k represents process noise, which follows a zero-mean normal distribution. Where Q k is the covariance matrix of process noise; P k is the covariance matrix of the extended Kalman filter of the kth state; the total number of states of the extended Kalman filter is equal to the total number of time steps N.

[0018] Furthermore, the charger model is constructed based on the preprocessed operation data, and the charger measurement equation of the charger is constructed using the extended Kalman filter method and is expressed using the following formula:

[0019]

[0020] Among them, H k is the measurement matrix; v k To measure noise, Where R k is the covariance matrix of the measurement noise.

[0021] Furthermore, the transition probability matrix P of the extended Kalman filter k+1|k Use the following formula to calculate:

[0022]

[0023] Among them, F k is the state transfer function of the extended Kalman filter for x k The Jacobian matrix of k is the covariance matrix of the extended Kalman filter of the kth state; the Kalman gain K of the extended Kalman filter k+1 Use the following formula to update:

[0024]

[0025] Among them, Kx +1is the Kalman gain of the extended Kalman filter in the k+1th state; the covariance matrix of the extended Kalman filter is updated using the following formula:

[0026] P k+1 =(IK k+1 H k ) k+1|k ;

[0027] Among them, P k+1 is the covariance matrix of the extended Kalman filter of the k+1th state.

[0028] Furthermore, in the automated test part, the charging efficiency η is calculated using the following formula based on the charger state equation and the charger measurement equation:

[0029]

[0030] Among them, || is an operation for calculating the modulus of a vector or the determinant of a matrix.

[0031] Furthermore, in the automated test part, the fault probability is calculated using the following formula based on the charger state equation and the charger measurement equation:

[0032]

[0033] Where p is the failure probability.

[0034] The automatic test system for new energy vehicle charger provided by the present application improves the quality and reliability of data through high-precision data acquisition and preprocessing technology. The system integrates a variety of high-precision sensors, such as current sensors, voltage sensors, temperature sensors, etc., which can monitor the operating status of the charger in real time. Data preprocessing includes steps such as filtering, denoising and normalization, which effectively removes noise and interference in the data and improves the accuracy and stability of the data. For example, the voltage data is preprocessed by Fourier series expansion and nonlinear transformation integral, the current data is preprocessed by wavelet transform and fractal dimension correction, and the temperature data is preprocessed by Legendre polynomial expansion and adaptive filtering, which can fully capture the frequency domain characteristics and nonlinear characteristics of the signal, thereby providing high-quality data support for subsequent model construction and state estimation. Secondly, the present invention adopts the extended Kalman filter (EKF) technology to construct a high-precision charger state model. EKF can improve the accuracy of state estimation while ensuring computational efficiency by linearizing the nonlinear system. Specifically, EKF iteratively performs prediction and update steps at each time step, combines the system state and measurement data through the state transfer equation and the measurement equation, and realizes the dynamic estimation of the system state. The state transfer equation comprehensively considers the nonlinear relationship between key parameters such as voltage, current and temperature, and can accurately describe the dynamic characteristics of the charger by introducing multiple nonlinear transformations such as Fourier series and wavelet transform. The measurement equation uses the Kalman gain to correct the measurement error to ensure a high degree of match between the state estimation and the actual measurement data, thereby improving the accuracy and robustness of the state estimation. In addition, the present invention has significant advantages in fault detection and prediction. Traditional fault detection methods mainly rely on simple threshold judgments and empirical rules, and it is difficult to accurately identify and predict potential faults in the charging process. The present invention uses intelligent algorithms and machine learning techniques to conduct in-depth analysis of pre-processed operating data and construct a fault detection and prediction model. For example, through the Mahalanobis distance of the measurement error, the Kalman gain correction and the comparison of the state estimation with the expected state, the operating state of the system can be accurately evaluated, potential fault signals can be identified, and early warnings can be given. In this way, not only the sensitivity and accuracy of fault detection are improved, but also preventive measures can be taken before a fault occurs to ensure the safety of the charging process. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The technical solution and other beneficial effects of the present application will be made apparent by describing in detail the specific implementation methods of the present application in conjunction with the accompanying drawings.

[0036] Figure 1 A schematic diagram of the system structure of an automatic testing system for a new energy vehicle charger provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0038] Example 1: Reference Figure 1 , an automatic test system for a new energy vehicle charger, the system comprising: a data acquisition part, used to acquire the operating data of the new energy vehicle charger in real time, pre-process the acquired operating data to obtain pre-processed operating data; a charger model building part, used to construct a charger state equation and a charger measurement equation of the charger based on the pre-processed operating data using an extended Kalman filter method; an automatic test part, used to calculate the charging efficiency and the fault probability according to the charger state equation and the charger measurement equation.

[0039] Specifically, the data acquisition part comprehensively monitors the operating status of the charger by integrating a variety of high-precision sensors, such as current sensors, voltage sensors, temperature sensors, power sensors, etc. These sensors measure the working parameters of the charger in real time and transmit the collected signals to the data processing module. Since the raw data collected by the sensor usually contains certain noise and interference, directly using these data may lead to error accumulation and deviation of the analysis results. Therefore, the raw data needs to be preprocessed. The preprocessing process includes a variety of data processing technologies, the main purpose of which is to improve data quality and reduce the impact of noise on the analysis results. First, filtering technologies such as Kalman filtering, low-pass filtering and mean filtering are used to filter the data to remove high-frequency noise and sudden interference. Kalman filtering is a dynamic recursive filtering method that can estimate the optimal state sequence in a noisy background and is particularly suitable for processing time series data. Secondly, the data is normalized to convert data of different dimensions into the same dimension range, which is convenient for subsequent unified processing and analysis. Normalization can use methods such as maximum and minimum value normalization and Z-score standardization to ensure the scale consistency of the data. In addition, filling missing values ​​in data is also an important part of preprocessing. Missing data can be reasonably filled by interpolation, mean filling, prediction model and other methods to ensure data integrity. The data acquisition part also needs to have certain intelligent analysis capabilities to identify and process abnormal data. By setting reasonable thresholds and rules, abnormal data that deviates from the normal range can be detected and marked or removed to prevent abnormal data from misleading the analysis results. On the basis of data preprocessing, the generated preprocessed data will be used as input for subsequent model construction and state estimation. The design and implementation of the entire data acquisition part fully considers the complexity and variability of the operating environment of new energy vehicle chargers, and adopts advanced data processing technology and intelligent analysis methods to ensure high accuracy and reliability of data. The data acquisition part is not only the basic module of the system, but also the key link to realize automated testing and intelligent decision-making. Its efficient and accurate data processing capabilities provide a solid guarantee for charger status monitoring, fault diagnosis and performance evaluation. Through real-time and accurate data acquisition and preprocessing, the system can dynamically respond to changes in the operating status of the charger, improve the automation level and intelligence level of the test, and provide important support for the performance optimization and safe operation of new energy vehicle chargers.

[0040] The charger model construction part plays a core role in the automatic test system of the charger for new energy vehicles. Its task is to construct the state equation and measurement equation of the charger based on the preprocessed operation data to achieve accurate estimation and prediction of the charger operation state. In order to effectively deal with the nonlinear dynamic characteristics of the charger, this part adopts the extended Kalman filter (EKF) method. EKF is an algorithm widely used in the state estimation of nonlinear systems. It realizes recursive estimation and correction of the system state by introducing a linearization step in the nonlinear system. With the preprocessed operation data as input, EKF predicts the operation state of the charger through the state equation. The state equation describes the dynamic evolution of the charger system in time, that is, how to transition from the current state to the next state. Due to the complexity and variability of the charger system, the state equation is usually nonlinear. EKF linearizes the nonlinear state equation near the current estimated state through Taylor expansion, thereby simplifying the problem to a linear system. At each time step, EKF uses the current state estimate and control input to predict the state at the next moment, while considering the process noise of the system. This prediction provides a priori state estimate for the next moment. Subsequently, the measurement equation uses the real-time collected charger operation data to correct the prior state estimate. The measurement equation describes how the system state is reflected in the actual observation data. Also due to the nonlinear relationship, the measurement equation needs to be linearized near the current state estimate. EKF combines the difference between the measurement data and the prior estimate, and adjusts the prior estimate to make it closer to the true state by calculating the Kalman gain. The Kalman gain reflects the relative credibility of the prediction model and the measurement data. The prior state estimate is weighted updated according to the Kalman gain to obtain the posterior state estimate. In this process, the influence of measurement noise is also taken into account, thereby reducing the influence of noise while improving the estimation accuracy. An important feature of the charger model construction part is its recursiveness and real-time performance. EKF iteratively performs the prediction and update steps at each time step, and continuously corrects the state estimate. This dynamic update mechanism enables the system to respond to changes in the charger operation state in real time. By continuously correcting the state equation and measurement equation, the model can adaptively adjust the estimation strategy and gradually approach the true system state. Compared with the traditional static model method, EKF not only considers the dynamic characteristics of the system, but also provides robust state estimation in a noisy and uncertain environment. In the process of building the charger model, the key is to correctly establish the state equation and measurement equation to ensure that they can accurately reflect the physical characteristics and operating laws of the charger. The state equation needs to consider the main influencing factors in the charging process, such as the change law of parameters such as voltage, current, temperature, and the electrical and thermal dynamic characteristics inside the charger. The measurement equation needs to reflect the relationship between sensor data and system status to ensure that the measurement data can effectively correct the state estimation.In practical applications, the parameter selection and adjustment of the EKF algorithm are crucial. Appropriate parameter settings can significantly improve the accuracy and convergence speed of state estimation. The charger model construction part converts the preprocessed operating data into system state estimation through the EKF method, providing high-precision, real-time monitoring and prediction capabilities of the charger operation status. This part not only provides an accurate state estimation basis for the automated test system, but also lays a solid foundation for subsequent charging efficiency calculations and fault probability assessments. Through scientific and reasonable model construction and state estimation, the charger model construction part effectively improves the intelligence level and operating efficiency of the system, and provides important technical support for performance optimization and fault diagnosis of new energy vehicle chargers.

[0041] During the operation of the charger, the automated test part first uses the state equation and measurement equation to track and predict the real-time state of the charger. The state equation describes the dynamic behavior of the charger, including the changing rules of key parameters such as charging voltage and charging current; the measurement equation reflects the relationship between the actual observed data and the system state. Through these two equations, the automated test part can obtain the state estimation of the charger in real time, compare it with the actual measured data, and identify possible deviations and anomalies. The calculation of charging efficiency is an important function of the automated test part. Charging efficiency is defined as the ratio of output power to input power and is an important indicator for measuring the performance of the charger. The automated test part calculates the charging efficiency at each moment by monitoring the input power and output power of the charger in real time. In order to improve the calculation accuracy, the system not only relies on the directly measured data, but also makes corrections based on the state estimation results. This can effectively reduce the impact of measurement noise and system uncertainty on efficiency calculation and ensure the accuracy and reliability of the results. In terms of the evaluation of the probability of failure, the automated test part uses advanced statistical analysis and machine learning algorithms to predict potential failure risks through a comprehensive analysis of the charger operation data. First, the system establishes a fault mode library through statistical analysis of historical data and identifies common fault features and patterns. During real-time monitoring, the system matches the current operating data with the features in the fault mode library and calculates the probability of fault occurrence. In order to further improve the accuracy of fault prediction, the automated test part also introduces machine learning technology to automatically learn the complex relationship between fault features and operating status through training models, so that signs of faults can be detected at an early stage and preventive measures can be taken in advance. A notable feature of the automated test part is its intelligence and efficiency. Through the automated test system, real-time monitoring and dynamic adjustment of the charger's operating status can be achieved, reducing the need for manual intervention and improving test efficiency. The system can automatically identify abnormal data and fault conditions, and issue alarms in a timely manner to prompt maintenance personnel to inspect and repair. This not only improves the operating safety of the charger, but also extends the service life of the equipment and reduces maintenance costs. In addition, the automated test part also has good scalability and adaptability. With the continuous development of new energy vehicle charging technology, the design and operating environment of the charger may change. Through modular design and flexible algorithm architecture, the automated test system can be easily upgraded and adjusted to adapt to new technical requirements and application scenarios. For example, when new sensors or data sources are introduced, the system can integrate the new data and incorporate it into the scope of state estimation and fault analysis through simple configuration updates.

[0042] Embodiment 2: The operating data at least includes: voltage, current and temperature.

[0043] Specifically, voltage is one of the important parameters to measure the performance of the charger. When charging the battery, the charger needs to adjust the output voltage according to the current voltage state of the battery to ensure the stability and efficiency of the charging process. Voltage monitoring can help identify abnormal conditions during the charging process, such as voltage fluctuations, overvoltage or undervoltage. If the voltage is unstable, it may lead to reduced battery charging efficiency or even damage the battery. Therefore, real-time voltage monitoring and data analysis are the basis for ensuring the normal operation of the charger. Current data is also crucial. When the charger is working, the change in current directly reflects the charging speed and the charging state of the battery. Excessive charging current may cause the battery to overheat, affect the battery life, and even cause safety problems; while too small charging current will prolong the charging time and reduce the charging efficiency. By monitoring the charging current in real time, the system can dynamically adjust the charging strategy to ensure the optimal charging efficiency within a safe range. In addition, current data can also be used to detect whether there are faults such as short circuits and open circuits during the charging process to improve the safety and reliability of the system. Temperature is another key parameter that reflects the thermal state of the charger and battery. During the charging process, the internal components of the battery and charger will generate heat. If the temperature is too high, it may cause damage to the components or thermal runaway of the battery, posing a serious safety hazard. By monitoring the temperature in real time, the system can take corresponding cooling measures when the temperature is too high, such as reducing the charging current or starting the cooling device to prevent overheating. In addition, temperature data can also be used to diagnose the health status of the charger and battery, helping to identify potential faults or performance degradation problems. In the charger automation test system, these operating data are not only basic monitoring parameters, but also can provide support for state estimation and fault prediction through data fusion and intelligent analysis. The system integrates voltage, current and temperature data to construct the state equation and measurement equation of the charger to achieve accurate estimation of the operating state of the charger. For example, the extended Kalman filter (EKF) method continuously updates and corrects the state estimation through the real-time input of these data, improving the accuracy and robustness of the model. In addition, voltage, current and temperature data can also be used for charging efficiency calculation and fault probability assessment. The calculation of charging efficiency depends on voltage and current data, and the efficiency of the charging process is evaluated by analyzing the ratio of input and output electrical energy. Fault probability assessment requires comprehensive consideration of voltage, current and temperature data, and the use of pattern recognition and machine learning techniques to identify potential fault characteristics, provide early warning of possible faults, and take preventive measures.

[0044] Example 3: Assume that the operating data of the new energy vehicle charger collected in real time by the data collection part is Where V i represents the voltage at the i-th time step, I i represents the current at the i-th time step, T irepresents the temperature of the i-th time step; N is the total number of time steps; suppose the preprocessing operation data is The following formula is used to calculate V i After processing, we get:

[0045]

[0046] in, is the Fourier series expansion term, which is used to convert V i Perform frequency domain analysis and reconstruction; a k is the cosine coefficient of the kth term of the Fourier series expansion, which represents the amplitude of the signal on the cosine basis function; b k is the sine coefficient of the kth term of the Fourier series expansion, which represents the amplitude of the signal on the sine basis function; k is the index of the number of terms in the Fourier series, ranging from 1 to M, where M is the total number of terms in the Fourier series; L is the period of the Fourier series expansion; is a nonlinear transformation integral term, which is used to capture the nonlinear characteristics in voltage data; λ is the exponential attenuation coefficient, which controls the attenuation rate of the exponential term. The larger the λ, the faster the attenuation; t is the integral variable, ranging from 0 to V i .

[0047] Specifically, the first part of this formula is the Fourier series expansion, which is used to decompose the voltage data into a superposition of a series of sine and cosine functions. The Fourier series is a powerful tool for processing periodic signals. It can convert complex time domain signals into the frequency domain, thereby revealing the frequency components in the signal. For the voltage signal of the new energy vehicle charger, the Fourier series expansion can effectively capture its periodic changes and frequency characteristics. By converting the voltage data V i Expanding into multiple frequency components enables analysis and reconstruction of periodic patterns in the signal. This is important for identifying periodic fluctuations and anomalies in the charging process. Cosine and sine coefficients a in Fourier series expansion k and b k Respectively represent the amplitude of the signal on the corresponding basis function, and they provide information about the energy distribution of the signal at different frequencies. In this way, through frequency domain analysis, we can gain a deep understanding of the frequency characteristics of the voltage signal and identify possible periodic interference or harmonic components. However, the voltage signal not only contains periodic characteristics, but also may have complex nonlinear characteristics. These nonlinear characteristics may not be fully displayed in simple frequency domain analysis. Therefore, the second part of the preprocessing formula introduces a nonlinear transformation integral term to capture the nonlinear characteristics in the voltage data. This integral term is obtained by calculating V i Perform nonlinear function This is achieved by integrating. The nonlinear function introduces an exponential decay factor λ to control the decay rate of the integral term. A larger λ value causes the integral term to decay rapidly, thereby capturing short-term changes in the voltage signal; a smaller λ value causes the integral term to decay more slowly, capturing long-term trends in the voltage signal. In this way, the nonlinear transformation integral term can effectively enhance the ability to describe the nonlinear characteristics of the signal. This is especially important for the complex voltage signals of new energy vehicle chargers, because the voltage changes during the charging process may be affected by a variety of nonlinear factors, including battery characteristics, charger control strategies, etc. Combining the Fourier series expansion and the nonlinear transformation integral, the preprocessed voltage data It not only contains the frequency domain characteristics of the original signal, but also enhances its nonlinear characteristics. This comprehensive preprocessing method can comprehensively improve the quality and information content of the data, and provide richer and more accurate input for subsequent state estimation and analysis. Specifically, in the state estimation process, the extended Kalman filter (EKF) method takes the preprocessed voltage data as input, and improves the estimation accuracy of the charger operation state through recursive prediction and correction. The frequency domain characteristics and nonlinear characteristics in the preprocessed data can help EKF better capture the dynamic changes and complex relationships of the system, thereby providing more reliable state estimation results. In addition, the preprocessed voltage data can also be used for fault detection and performance evaluation. By analyzing the frequency components and nonlinear characteristics in the preprocessed data, the system can identify potential anomalies and signs of faults. For example, if the energy of a certain frequency component suddenly increases, it may indicate that there is harmonic interference or circuit failure in the system; while changes in nonlinear characteristics may indicate that the performance of the battery or charger has deteriorated. By timely detecting and analyzing these anomalies, the system can take preventive measures to prevent the fault from further deteriorating and ensure the safety and stability of the charging process.

[0048] Embodiment 4: I is calculated by the following formula i After processing, we get:

[0049]

[0050] Wavelet transform is a powerful signal processing tool that can analyze signals in both time and frequency domains. Unlike Fourier transform, wavelet transform has the ability of multi-resolution analysis and can effectively capture local features in the signal. For the current signal of the new energy vehicle charger, the wavelet transform can be used to decompose it into components of different scales and positions, thereby revealing the details and global characteristics of the signal. Specifically, by using different scale parameters a j and displacement parameter b j, wavelet transform can analyze signals in different time windows, which is very useful for capturing mutations and transient features in current signals. The wavelet function ψ is used as a mother wavelet, and different forms can be selected according to specific applications to adapt to the characteristics of current signals. In the preprocessing of current data, the first part of the wavelet transform is to transform the wavelet coefficients c of different scales and displacements. j The multi-scale features of the current signal are reconstructed by weighted sum. This multi-scale feature helps to reveal the complex patterns and change laws in the current signal, providing richer information for subsequent state estimation and fault detection. For example, during the charging process, the current signal may contain various frequency components and sudden changes. Wavelet transform can effectively separate and analyze these components to identify possible anomalies and faults. However, the current signal not only contains periodic and local characteristics, but also may have complex nonlinear characteristics and self-similar structures. These nonlinear characteristics may not be fully displayed in simple frequency domain analysis. To this end, the second part of the preprocessing formula introduces fractal dimension correction to describe the complexity and self-similar characteristics of the current signal. Fractal dimension is an indicator of signal complexity. By introducing fractal dimension correction, the inherent laws of the current signal can be better characterized. Specifically, the fractal dimension correction term controls the nonlinear transformation of the current signal through parameters α and β, so that the details and overall characteristics of the signal are better balanced. Such a transformation can amplify or compress the different amplitude ranges of the current signal, thereby enhancing the characteristic expression ability of the signal. Fractal dimension correction helps to reveal the complex structure and inherent laws of the current signal and improve the expression ability of the signal. This processing method is particularly suitable for processing current data with highly nonlinear and complex dynamic characteristics. For example, during the charging process, the current signal may be affected by multiple factors such as battery characteristics and charger control strategy, showing complex nonlinear characteristics. Through fractal dimension correction, these complex characteristics can be better captured and described, providing more accurate and comprehensive data support for state estimation and fault detection. Combining wavelet transform and fractal dimension correction, the preprocessed current data not only contains the multi-scale characteristics of the original signal, but also enhances its nonlinear characteristics. This comprehensive preprocessing method can comprehensively improve the quality and information content of the data, and provide richer and more accurate input for subsequent state estimation, fault detection and performance evaluation. Through this method, the system can better capture the complex patterns and change laws in the current signal, improve the monitoring and analysis capabilities of the charger operation status, and ensure the efficiency and safety of the charging process. This preprocessing method makes full use of the advantages of modern signal processing technology, realizes efficient processing and intelligent analysis of complex current signals, and provides important technical support for the automated testing and optimization of new energy vehicle chargers.

[0051] Among them, ψ represents the mother wavelet function, a j and b jare the scale and displacement parameters of wavelet transform, α and β are fractal dimension correction parameters; J represents the upper limit of the scale parameter of wavelet transform; The following formula is used to calculate T i After processing, we get:

[0052]

[0053] Legendre polynomial expansion is an effective signal processing method for approximating and representing complex functions. Legendre polynomials are a set of orthogonal polynomials that are mutually orthogonal in a given interval, which means that they can be used to decompose and reconstruct arbitrarily complex signals. For the temperature signal of a new energy vehicle charger, the Legendre polynomial expansion can be used to represent it as a weighted sum of multiple polynomials. This representation method can capture the global characteristics and trends of the temperature signal, thereby revealing the main patterns and changes in the signal. By converting the temperature data T i Expand into different orders P nThe sum of polynomials can analyze and reconstruct the complex features in the signal. This is of great significance for identifying temperature change trends and anomalies during the charging process. In the preprocessing of temperature data, the Legendre polynomial expansion part reconstructs the global features of the temperature signal by weighted sum of polynomials of different orders. Polynomial expansion helps to capture the long-term changes and overall trends of the temperature signal, providing richer information for subsequent state estimation and fault detection. For example, during the charging process, the temperature signal may show a slow upward or downward trend. The polynomial expansion can effectively separate and analyze these trends and identify possible anomalies and faults. However, the temperature signal not only contains global features and trends, but also may have complex local features and noise. These local features may not be fully displayed in a simple polynomial expansion. To this end, the second part of the preprocessing formula introduces adaptive filtering to smooth the temperature signal and reduce the impact of noise. The adaptive filter is smoothed by the Gaussian kernel function, which has good localization properties and can retain the main features of the signal while ensuring signal smoothness. Specifically, the Gaussian kernel function controls the degree of smoothing through the standard deviation σ, integrates the entire real number domain, and obtains the smoothed temperature signal. This processing method can effectively suppress noise, improve the smoothness and stability of the signal, and retain the main characteristics of the signal. Adaptive filtering helps to smooth local changes in the temperature signal and reduce the interference of noise on the signal. This is particularly important for processing temperature signals with high-frequency noise and local mutations. For example, during the charging process, the temperature signal may be affected by multiple factors such as ambient temperature fluctuations and equipment heat dissipation, showing high-frequency noise and local mutations. Through adaptive filtering, these noises can be effectively smoothed, the main change trend of the signal can be retained, and more accurate and reliable data support can be provided for state estimation and fault detection. Combining Legendre polynomial expansion and adaptive filtering, the preprocessed temperature data not only contains the global characteristics of the original signal, but also enhances its smoothness and stability. This comprehensive preprocessing method can comprehensively improve the quality and information content of the data, and provide richer and more accurate input for subsequent state estimation, fault detection and performance evaluation. With this method, the system can better capture the complex patterns and change laws in the temperature signal, improve the monitoring and analysis capabilities of the charger operation status, and ensure the efficiency and safety of the charging process. This preprocessing method makes full use of the advantages of modern signal processing technology, realizes efficient processing and intelligent analysis of complex temperature signals, and provides important technical support for the automated testing and optimization of new energy vehicle chargers.

[0054] Among them, P n represents the nth order Legendre polynomial, σ is the standard deviation of the adaptive filter; τ is the integral variable, and its range is the entire real number domain.

[0055] Embodiment 5: The charger model construction part is based on the pre-processed operation data, using the extended Kalman filter method, and the charger state equation of the charger is expressed by the following formula:

[0056]

[0057] Among them, A k and B k are the state transfer matrix and the control input matrix respectively; ∈ is a positive number to prevent division by zero; x k is the extended Kalman filter result of the kth state; x k+1 is the extended Kalman filter result of the k+1th state; u k is the control input vector in the k state; w k represents process noise, which follows a zero-mean normal distribution. Where Q k is the covariance matrix of process noise; P k is the covariance matrix of the extended Kalman filter of the kth state; the total number of states of the extended Kalman filter is equal to the total number of time steps N.

[0058] Specifically, the linear part of the state equation is represented by the state transfer matrix A k and the control input matrix B k It describes the state changes of the system under normal operating conditions. k Indicates how the system state evolves naturally when there is no external disturbance and control input; the control input matrix B k Then the control input u is introduced k This part provides a basic framework for state estimation, describing the basic dynamic behavior of the system through linear relationships. However, the actual operating environment of the charger is often highly nonlinear and complex, so it is necessary to introduce multiple nonlinear terms to more accurately describe the system state.

[0059] The first nonlinear term is obtained by passing the preprocessed voltage Current and temperature To reflect the comprehensive impact of these key parameters on the system state, this term uses a combination of fractions and trigonometric functions: Among them, P kis the covariance matrix, which reflects the uncertainty of state estimation, and ∈ is a positive number to prevent division by zero. This item is designed to capture the complex interaction between voltage, current and temperature. The voltage data has been preprocessed by Fourier series and nonlinear transformation, and has included frequency domain and nonlinear characteristics; the current data has been preprocessed by wavelet transform and fractal dimension correction to reflect multi-scale characteristics and complex structures; the temperature data has been preprocessed by Legendre polynomial expansion and adaptive filtering to include global characteristics and local smoothness characteristics. By introducing these preprocessed data into the state equation, the actual dynamic characteristics of the charger can be better reflected and the accuracy of state estimation can be improved. In practical applications, the voltage It is obtained through Fourier series expansion and nonlinear transformation integral preprocessing. Fourier series expansion can capture the frequency domain characteristics of the voltage signal, while nonlinear transformation integral is used to describe the nonlinear characteristics in the voltage signal. These preprocessing steps can enhance the expression ability of voltage data and make it contain more useful information. Through wavelet transform and fractal dimension correction preprocessing, the multi-scale characteristics and complex structure of the signal can be captured, thereby improving the quality and information content of the data. The global characteristics and local smoothness characteristics of the signal can be captured through Legendre polynomial expansion and adaptive filtering preprocessing. By introducing these preprocessed data into the state equation, the actual dynamic characteristics of the charger can be better reflected and the accuracy of state estimation can be improved. The second nonlinear term is a combination of exponential function and trigonometric function: Among them, λ is the exponential decay factor, which reflects the nonlinear characteristics of the voltage data. Through the rapid decay characteristics of the exponential function and the periodic characteristics of the trigonometric function, this term can better describe the impact of current and temperature on the system state. The exponential function is used to capture short-term changes in the voltage signal, while the trigonometric function enhances the description of current and temperature through its periodic characteristics. This combination enables the state equation to more comprehensively reflect the dynamic characteristics of the charger, especially in the face of rapidly changing input signals, providing more accurate state estimates. The design of this nonlinear term takes into account the combined effects of various dynamic factors in the operation of the charger. After the voltage data is processed by exponential decay, it can effectively capture the short-term change characteristics in the voltage signal, while the current and temperature data are enhanced by the periodic processing of trigonometric functions to enhance the description of these parameters. By introducing these complex nonlinear terms, the state equation can more comprehensively reflect the actual dynamic behavior of the charger and improve the accuracy of state estimation. The third nonlinear term is an integral term, which further enhances the model's ability to describe the nonlinear characteristics of the current signal by integrating the current data: This integral term combines the current and voltage data, capturing the long-term changes and overall trends in the signal through the integration operation. The integrand in the integral term is expressed by The form balances the short-term changes and long-term trends of the signal, so that the model can better reflect the dynamic behavior of the current signal. Through the integration operation, the high-frequency noise in the data can be effectively smoothed, while retaining the main characteristics and change trends of the signal, thereby improving the stability and robustness of the state estimation. The design of this integral term takes into account the complex nonlinear characteristics and dynamic changes in the current signal. After the current data is integrated, it can effectively capture the long-term change characteristics in the signal, thereby providing a more stable and reliable state estimation. This integral processing method can not only smooth high-frequency noise, but also retain the main change trend of the signal, so that the state equation can more comprehensively reflect the dynamic behavior of the current signal and improve the accuracy and robustness of the state estimation. In the entire state equation, the process noise w k is modeled as a zero-mean normal distribution with covariance matrix Q k The process noise reflects the random disturbance and unmodeled dynamics in the system. By introducing process noise into the state equation, the robustness and adaptability of the model can be improved. The covariance matrix Q kIt is used to describe the intensity and correlation of process noise, and provides a statistical description for the extended Kalman filter. Through the combination of the above nonlinear terms, the state equation not only describes the basic dynamic behavior of the charger, but also captures the complex nonlinear characteristics and interactions in the voltage, current and temperature data. These nonlinear terms enable the extended Kalman filter to estimate the system state more accurately and improve the monitoring and prediction capabilities of the charger operation state. The preprocessed operation data is effectively integrated into the state equation through these complex nonlinear transformations, so that the model can more comprehensively and accurately reflect the actual operation characteristics of the charger. By combining a variety of nonlinear transformations and preprocessing methods, the state equation can comprehensively consider various complex dynamic characteristics in the operation of the charger. The extended Kalman filter method achieves efficient and accurate state estimation by linearizing the nonlinear system. By introducing the complex nonlinear characteristics of voltage, current and temperature data, the state equation provides a comprehensive description of the system state, providing solid technical support for the automated testing and optimization of the charger. This method not only improves the accuracy and robustness of state estimation, but also enhances the system's adaptability to complex dynamic environments, providing an important driving force for the development of new energy vehicle charging technology. In addition, the extended Kalman filter can correct the state estimate in real time by recursively executing the prediction and update steps at each time step, thereby improving the dynamic response capability of the system. In the prediction step, the state equation uses the current state and control input to calculate the prior state estimate of the next moment; in the update step, the Kalman gain is calculated in combination with the actual measurement data, and the Kalman gain is used to correct the prior state estimate to obtain the posterior state estimate. Through this dynamic correction mechanism, the EKF can continuously adjust the state estimate during the operation of the charger to ensure the accuracy and reliability of the estimation results. In practical applications, the operating state of the charger is affected by many factors, including voltage fluctuations, current changes, and temperature changes. The state equation provides a comprehensive description of the system state by comprehensively considering these factors. The voltage data is preprocessed by Fourier series expansion and nonlinear transformation integral to enhance the expression of frequency domain features and nonlinear features; the current data is preprocessed by wavelet transform and fractal dimension correction to capture multi-scale features and complex structures; the temperature data is preprocessed by Legendre polynomial expansion and adaptive filtering to capture global features and local smoothness characteristics. By introducing these preprocessed data into the state equation, the actual dynamic characteristics of the charger can be better reflected and the accuracy of state estimation can be improved.

[0060] Embodiment 6: The charger model construction part is based on the pre-processed operation data, using the extended Kalman filter method, and the charger measurement equation of the charger is expressed by the following formula:

[0061]

[0062] Among them, H k is the measurement matrix; v k To measure noise, Where R k is the covariance matrix of the measurement noise.

[0063] Specifically, in the automatic test system of new energy vehicle chargers, the extended Kalman filter (EKF) method is used to construct and optimize the measurement equation of the charger to achieve accurate estimation and prediction of the charger operation state. The core of the measurement equation is to associate the system state with the actual measurement data, and to capture the key characteristics and nonlinear behaviors of the charger operation by combining the preprocessed operation data, thereby improving the accuracy of state estimation and the accuracy of the test. The first part of the measurement equation contains the measurement matrix H k and the system state x k , describes the basic linear relationship between the system state and the measurement data. The measurement matrix H k It is used to map the state variables to the measurement space, thus realizing the basic connection between the state and the measurement data. However, due to the high nonlinearity and complexity of the charger operating environment, the linear part is not enough to fully and accurately describe the relationship between the system state and the measurement data. Therefore, it is necessary to introduce multiple nonlinear terms to enhance the descriptive ability of the measurement equation. The first nonlinear term is obtained by preprocessing the temperature. and voltage To reflect the impact of these key parameters on the measured data, this term uses a combination of logarithmic functions and fractional forms. Where P k is the covariance matrix, which reflects the uncertainty of the state estimation, and ∈ is a positive number that prevents the logarithmic function from taking zero. The logarithmic function is used to process the nonlinear characteristics of the voltage data, while the fractional form is used to balance the influence of the temperature data. This combination can capture the complex nonlinear relationship between temperature and voltage, so that the measurement equation can more comprehensively reflect the true dynamic characteristics of the system. The temperature data is preprocessed by Legendre polynomial expansion and adaptive filtering to enhance the global characteristics and local smoothness, while the voltage data is preprocessed by Fourier series and nonlinear transformation integral to include frequency domain and nonlinear characteristics. By introducing these preprocessed data, the measurement equation can more accurately reflect the actual operating state of the charger, thereby improving the correlation accuracy between the measurement data and the system state, and ultimately improving the accuracy of the test. The second nonlinear term is obtained by the preprocessed current Voltage and temperature The combination of further enhances the ability of the measurement equation to describe complex dynamic characteristics. The specific form is This term combines trigonometric functions and fractional forms. The trigonometric processing of current and temperature can capture the periodic characteristics of these parameters, while the fractional form processing of voltage and current is used to describe the complex interaction between them. The current data is preprocessed by wavelet transform and fractal dimension correction to capture the multi-scale characteristics and complex structure of the signal, thereby improving the quality and information content of the data. By introducing these preprocessed current data, the measurement equation can better capture the complex nonlinear characteristics and dynamic changes of the current signal, and enhance the ability to describe the system state. The introduction of these nonlinear terms enables the measurement equation to more comprehensively reflect the actual dynamic characteristics of the charger, thereby improving the correlation accuracy between the measurement data and the system state, and ultimately improving the accuracy of the test. The last part of the measurement equation is the measurement noise v k , which is modeled as a zero-mean normal distribution with a covariance matrix R k The measurement noise reflects the random errors and unmodeled dynamics in the actual measurement process. By introducing the measurement noise into the measurement equation, the robustness and adaptability of the model can be improved. The covariance matrix R k It is used to describe the intensity and correlation of measurement noise, and provides a statistical description for the extended Kalman filter. The measurement equation effectively captures the key characteristics and dynamic behaviors of the charger operation by comprehensively considering the linear and nonlinear relationship between the system state and the measurement data. The extended Kalman filter method achieves efficient and accurate state estimation by linearizing the nonlinear system. By introducing pre-processed temperature, voltage and current data, the measurement equation can more comprehensively and accurately reflect the actual operating characteristics of the charger, improve the correlation accuracy between the measurement data and the system state, and thus significantly improve the accuracy of the test.

[0064] Embodiment 7: The transition probability matrix P of the extended Kalman filter k+1|k Use the following formula to calculate:

[0065]

[0066] Among them, F k is the state transfer function of the extended Kalman filter for x k The Jacobian matrix of k is the covariance matrix of the extended Kalman filter of the kth state; the Kalman gain K of the extended Kalman filter k+1 Use the following formula to update:

[0067]

[0068] Among them, K k+1 is the Kalman gain of the extended Kalman filter in the k+1th state; the covariance matrix of the extended Kalman filter is updated using the following formula:

[0069] P k+1 =(IK k+1 H k ) k+1|k ;

[0070] Among them, P k+1 is the covariance matrix of the extended Kalman filter of the k+1th state.

[0071] Specifically, the transition probability matrix P k+1∣k The calculation formula is Here, F k is the state transition function for the current state x k The Jacobian matrix F reflects the evolution of the system state over time and its sensitivity to small perturbations. k The introduction of is to approximate the linear change of state in nonlinear systems, so that EKF can be linearized in a nonlinear framework. The covariance matrix P of the current state k Reflecting the uncertainty of the state, after the state transfer, the new state covariance matrix P k+1∣k It combines the dynamic characteristics of the system itself (described by the Jacobian matrix) and the process noise Q k The process noise covariance matrix Q k It describes the random disturbances and unmodeled dynamics of the system during the transfer process. By considering process noise, the robustness and adaptability of state estimation can be improved. Kalman gain K k+1 The update formula is

[0072] The Kalman gain is a key parameter in the EKF, which determines the extent to which the measurement update corrects the state estimate. The calculation of the Kalman gain is based on the prediction covariance matrix P k+1∣k and the measurement matrix H k and the measurement noise covariance matrix R k By calculating the Kalman gain, the weighted ratio between the predicted state and the measured data can be determined, thereby effectively combining the predicted and measured information during the update process and improving the accuracy of the state estimation. Specifically, the larger the Kalman gain, the higher the weight of the measured data in the state estimation, and vice versa. The update formula of the covariance matrix is ​​P k+1 =(IK k+1 H k ) k+1∣k The update of the covariance matrix reflects the correction of the state estimation uncertainty by the measurement update. By introducing the Kalman gain K k+1 , which can effectively reduce the uncertainty of state estimation and improve the accuracy and stability of estimation. The updated covariance matrix P k+1Represents the state uncertainty after combining the measurement data. Through this update process, EKF can dynamically adjust the accuracy of state estimation, thereby providing reliable state tracking and prediction in complex nonlinear systems. In the entire EKF process, state transfer and measurement update are two complementary steps. State transfer provides a prediction of the dynamic characteristics of the system by considering the Jacobian matrix and process noise, while measurement update introduces actual measurement data into the state estimation through the combination of Kalman gain and measurement noise, thereby correcting the prediction error. This dynamic prediction and update mechanism enables EKF to maintain high-precision state estimation in complex nonlinear environments. Specifically for the application scenario of new energy vehicle chargers, EKF can accurately estimate the operating state of the charger by considering the nonlinear relationship between key parameters such as voltage, current and temperature. The preprocessed voltage data contains rich frequency domain and nonlinear features after Fourier series and nonlinear transformation integration; the current data reflects multi-scale characteristics and complex structures after wavelet transformation and fractal dimension correction; the temperature data captures global characteristics and local smoothness after Legendre polynomial expansion and adaptive filtering. By introducing these preprocessed data into the state transfer and measurement update process of EKF, the dynamic characteristics of the charger can be more accurately reflected, and the accuracy and robustness of state estimation can be improved. The advantage of the EKF method is that it can provide approximate optimal state estimation in nonlinear systems, and adapt to complex dynamic environments through linearization processing and dynamic updates. In the actual operation of the charger, the changes in parameters such as voltage, current and temperature are highly nonlinear and complex. The EKF method can effectively capture these dynamic changes and provide high-precision state tracking and prediction. This not only improves the safety and reliability of the system, but also provides important technical support for optimizing the charging process and improving charging efficiency.

[0073] Embodiment 8: In the automated test part, the charging efficiency η is calculated using the following formula according to the charger state equation and the charger measurement equation:

[0074]

[0075] Among them, || is an operation for calculating the modulus of a vector or the determinant of a matrix.

[0076] Specifically, the calculation formula of charging efficiency integrates multiple key parameters, including system state, measurement data, control input, measurement noise and process noise. The numerator contains the state estimate x k , measurement matrix H k , measurement noise covariance matrix R k and the actual measured data z k These parameters are defined by and K k (z k-H k x k ) form, which reflects the relationship between the system state estimation and the actual measurement data. k Provided by the Extended Kalman Filter, it is the best estimate of the true state of the system. The measurement matrix H k Map the state vector to the measurement space so that the state estimate can be compared with the actual measurement data. The measurement noise covariance matrix P k It reflects the uncertainty and error in the measurement process. Kalman gain K k It plays a key role in the formula. It is used to correct the difference between the state estimate and the measured data, thereby improving the accuracy of the estimate. Specifically, the Kalman gain updates and corrects the state estimate by combining the prediction error with the actual measurement error. This process ensures that the state estimate can reflect the actual operating state of the system in real time, thereby providing more accurate input data for the calculation of charging efficiency. The numerator also includes an exponential term This exponential term measures the degree of match between the measured data and the state estimate by calculating the Mahalanobis distance of the measurement error. The Mahalanobis distance takes into account the covariance structure of the measurement noise, making the error assessment more accurate and comprehensive. The introduction of the exponential term increases the sensitivity to the measurement error, ensuring that the calculation result of the charging efficiency can accurately reflect the uncertainty of the system state estimate when the measurement error is large. Through these treatments, the numerator of the formula can fully capture the complex relationship between the system state and the measured data, ensuring the high accuracy of the charging efficiency calculation. The measured data undergoes multiple preprocessing steps, such as Fourier series expansion and nonlinear transformation integration of voltage data, wavelet transform and fractal dimension correction of current data, and Legendre polynomial expansion and adaptive filtering of temperature data, making the input data more accurate and informative. These preprocessing steps enhance the quality and information content of the input data, making the state estimate more reliable, thereby improving the accuracy of the charging efficiency calculation.

[0077] The denominator of the formula combines the control input u k , control input matrix B k , process noise covariance matrix Q k and the state covariance matrix P k Through and The combined effect of control input and state uncertainty is measured in the form of . The processing of the control input part ensures that the energy input during the charging process is accurately calculated, while the introduction of the state covariance matrix reflects the uncertainty and estimation error of the system. Through the square root operation, the influence between different dimensions can be balanced, making the result of the denominator more reasonable and stable. Specifically, the control input matrix B kControl input u k Transformed into the impact on the system state, through the covariance matrix Q of the control input k To describe the uncertainty and disturbance in the control process. The state covariance matrix P k It reflects the accuracy of state estimation. Through the introduction of Kalman gain, this part of the calculation can dynamically adjust the uncertainty of state estimation, thereby improving the robustness and adaptability of the system. By combining state estimation, measurement data and control input, the formula realizes a comprehensive evaluation of charging efficiency. The state estimation part ensures the accuracy and reliability of the estimation through the correction of Kalman gain and measurement error; the measurement data part reflects the accuracy and matching degree of the measurement data through the evaluation of Mahalanobis distance; the control input part ensures the rationality and stability of the charging efficiency through the accurate calculation of energy input. Through this comprehensive calculation method, the evaluation of charging efficiency can not only reflect the current performance of the system, but also provide optimization suggestions for the system state and control strategy through the analysis of dynamic data. The advantage of this method is that it can make full use of the key parameters in the extended Kalman filter, and improve the accuracy and robustness of the charging efficiency calculation through the comprehensive analysis of state estimation, measurement data and control input. In the actual application of new energy vehicle chargers, the changes in parameters such as voltage, current and temperature are highly nonlinear and complex. Through this comprehensive calculation method, the dynamic changes of the system can be effectively captured and high-precision charging efficiency evaluation can be provided. This not only helps optimize the charging process and improve charging efficiency, but also prevents potential failures and performance degradation through accurate state estimation, ensuring the safety and reliability of the charger. By utilizing the extended Kalman filter method, the formula can dynamically adjust the relationship between state estimation and measurement data, thereby providing accurate charging efficiency evaluation in real-time operation. This method not only enhances the accuracy of charging efficiency calculation, but also improves the system's adaptability to complex dynamic environments, providing solid technical support for the automated testing and optimization of new energy vehicle chargers.

[0078] Embodiment 9: In the automated test part, the fault probability is calculated using the following formula according to the charger state equation and the charger measurement equation:

[0079]

[0080] Where p is the failure probability.

[0081] Specifically, in the automatic test system of new energy vehicle chargers, the calculation of failure probability is a key link to ensure the safe operation and optimized performance of the chargers. Example 8 proposes a formula for calculating the failure probability using an extended Kalman filter (EKF). This formula integrates multiple factors such as state estimation, measurement data, and control input through complex mathematical operations, providing a comprehensive and high-precision method for calculating failure probability. Through the detailed explanation, it can be seen how this formula accurately calculates the failure probability and improves the accuracy of the results. The calculation formula for the failure probability p evaluates the operating status and potential failure risks of the system by integrating multiple key parameters, including system status, measurement data, control input, measurement noise, and process noise. The numerator of the formula first introduces a logical function It uses the state estimate x k and the measurement matrix H k The combination of maps the amplitude of the state estimate to a probability value, reflecting the nonlinear characteristics and matching degree of the state estimate in the measurement space. This term, in the form of a logical function, enables the change of the state estimate to be sensitively reflected in the calculation of the fault probability, thereby capturing the nonlinear dynamic changes in the system. Next, the formula calculates the sum of squares of the measurement error, and through the term Here, the measurement data z k With the state estimate H k x k The difference between the measured noise covariance matrix R k The Mahalanobis distance of the measurement error is obtained by weighting the Mahalanobis distance in different dimensions. The Mahalanobis distance is reasonably weighted in different dimensions, making the error assessment more accurate and comprehensive. By calculating the sum of squares of the measurement errors, the formula can capture anomalies and potential faults in system operation, improving the sensitivity and accuracy of fault detection. In addition, the formula introduces the term |K k (z k -H k x k )|, where the Kalman gain K k The measurement error is weighted to further correct the difference between the state estimate and the measured data. The Kalman gain plays a key role in the extended Kalman filter, combining the prediction error with the actual measurement error to dynamically update the state estimate. This process ensures that the state estimate can reflect the actual operating state of the system in real time and provide more accurate input data for the calculation of the probability of failure. The exponential term in the formula Then the state estimate x is calculated k and the expected state μ k The difference between the expected state μ and the expected state μ further measures the dynamic characteristics of the system state. kIt represents the standard state of the system under normal operating conditions, and is estimated by comparing it with the actual state x k By comparing the control input and state uncertainty, the deviation in system operation can be identified. The introduction of the exponential term enhances the dynamic evaluation capability of state estimation, allowing the formula to more comprehensively reflect the actual operating state of the system. The denominator combines the influence of control input and state uncertainty, and and The control input matrix B is a combination of k Control input u k Transformed into the impact on the system state through the process noise covariance matrix Q k To describe the uncertainty and disturbance in the control process. The state covariance matrix P k It reflects the accuracy of state estimation. Through the introduction of Kalman gain, this part of the calculation can dynamically adjust the uncertainty of state estimation, thereby improving the robustness and adaptability of the system. Through these processes, the numerator of the formula can fully capture the complex relationship between the system state and the measurement data, ensuring the high accuracy of fault probability calculation. The measurement data undergoes a variety of preprocessing steps, such as Fourier series expansion and nonlinear transformation integration of voltage data, wavelet transform and fractal dimension correction of current data, and Legendre polynomial expansion and adaptive filtering of temperature data, making the input data more accurate and informative. These preprocessing steps enhance the quality and information content of the input data, making the state estimation more reliable, thereby improving the accuracy of fault probability calculation.

[0082] The above is a detailed introduction to the automatic testing system for the new energy vehicle charger provided by the embodiment of the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. The automatic test system for new energy vehicle charger is characterized by: The system includes: a data acquisition part, which is used to collect the operating data of the new energy vehicle charger in real time, pre-process the collected operating data, and obtain pre-processed operating data; a charger model construction part, which is used to construct the charger state equation and charger measurement equation of the charger based on the pre-processed operating data by using the extended Kalman filter method; an automated test part, which is used to calculate the charging efficiency and the fault probability according to the charger state equation and the charger measurement equation; the operating data at least includes: voltage, current and temperature; assuming that the operating data of the new energy vehicle charger collected by the data acquisition part in real time is Where V i represents the voltage at the i-th time step, I i represents the current at the i-th time step, T i represents the temperature of the i-th time step; N is the total number of time steps; suppose the preprocessing operation data is The following formula is used to calculate V i After processing, we get: in, is the Fourier series expansion term, which is used to convert V i Perform frequency domain analysis and reconstruction; a k is the cosine coefficient of the kth term of the Fourier series expansion, which represents the amplitude of the signal on the cosine basis function; b k is the sine coefficient of the kth term of the Fourier series expansion, which represents the amplitude of the signal on the sine basis function; k is the index of the number of terms in the Fourier series, ranging from 1 to M, where M is the total number of terms in the Fourier series; L is the period of the Fourier series expansion; is a nonlinear transformation integral term, which is used to capture the nonlinear characteristics in voltage data; λ is the exponential attenuation coefficient, which controls the attenuation rate of the exponential term. The larger the λ, the faster the attenuation; t is the integral variable, ranging from 0 to V i ; I is calculated by the following formula i After processing, we get: Among them, ψ represents the mother wavelet function, a j and b j are the scale and displacement parameters of the wavelet transform, α and β are the fractal dimension correction parameters; J represents the upper limit of the scale parameter of the wavelet transform; T is obtained by processing it with the following formula: Among them, P n represents the nth order Legendre polynomial, σ is the standard deviation of the adaptive filter; τ is the integral variable, and its range is the entire real number domain; the charger model is constructed based on the preprocessed operation data, and the charger state equation of the charger constructed using the extended Kalman filter method is expressed using the following formula: Among them, A k and B k are the state transfer matrix and the control input matrix respectively; ∈ is a positive number to prevent division by zero; x k is the extended Kalman filter result of the kth state; x k+1 is the extended Kalman filter result of the k+1th state; u k is the control input vector in the k state; w k represents process noise, which follows a zero-mean normal distribution. Where Q k is the covariance matrix of process noise; P k is the covariance matrix of the extended Kalman filter of the kth state; the total number of states of the extended Kalman filter is equal to the total number of time steps N; the charger model construction part is based on the pre-processed operation data, using the extended Kalman filter method, and the charger measurement equation of the charger constructed is expressed by the following formula: Among them, H k is the measurement matrix; v k To measure noise, Where R k is the covariance matrix of the measurement noise; the transition probability matrix P of the extended Kalman filter k+1|k Use the following formula to calculate: Among them, F k is the state transfer function of the extended Kalman filter for x k The Jacobian matrix of k is the covariance matrix of the extended Kalman filter of the kth state; the Kalman gain K of the extended Kalman filter k+1 Use the following formula to update: Among them, K k+1 is the Kalman gain of the extended Kalman filter in the k+1th state; the covariance matrix of the extended Kalman filter is updated using the following formula: P k+1 =(I-K k+1 H k )P k+1|k ; Among them, P k+1 is the covariance matrix of the extended Kalman filter of the k+1th state; in the automated test part, the charging efficiency η is calculated using the following formula according to the charger state equation and the charger measurement equation: Among them, || is an operation for calculating the modulus of a vector or the determinant of a matrix.

2. The automatic testing system for the new energy vehicle charger according to claim 1, characterized in that: In the automated test part, the fault probability is calculated using the following formula based on the charger state equation and charger measurement equation: Where p is the failure probability.

Images