High-power charging pile self-adjusting method, medium and system for high-temperature environment

Abstract

The invention provides a high-power charging pile self-adjusting method, medium and system for a high-temperature environment, and belongs to the technical field of electrical variable measurement, and the method comprises the steps: extracting time domain features and frequency domain features of an environment parameter, a charging temperature, an input parameter and an output parameter; a correlation matrix of each parameter is calculated through a Pearson's correlation coefficient and principal component analysis, and a stability matrix and a variability matrix are obtained through singular value decomposition. Based on the two matrixes, an autoregressive integrated moving average model and a conditional autoregressive model are respectively constructed, and the long-term trend and short-term fluctuation of each parameter are predicted. And then establishing a game model, and solving to obtain an optimal output parameter by taking the long-term stability and the short-term stability of the output parameter as a target, and taking the optimal output parameter as a charging pile output parameter in a next time period. The charging process is continuously optimized with the preset duration as the cycle period, and the technical problem that real-time optimization of charging adjustment is difficult to achieve due to the fact that an existing charging control technology is mainly based on a single temperature threshold value is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electrical variable measurement, and in particular, relates to a self-adjusting method, medium and system for a high-power charging pile used in a high-temperature environment. Background Art

[0002] As an important new energy vehicle, electric vehicles have been widely used around the world. Compared with traditional fuel vehicles, electric vehicles have the advantages of zero emissions, high energy utilization, and comfortable driving experience. However, the construction of charging infrastructure for electric vehicles has always been one of the key bottlenecks in the promotion of vehicles. In recent years, governments and enterprises have continuously increased the construction of charging piles, and high-power charging piles have gradually become popular. This high-power charging pile can significantly shorten the charging time of electric vehicles and improve the user's charging experience.

[0003] However, as the charging power continues to increase, the heat generated during the charging process is also increasing. Especially in high temperature environments, this heat is difficult to dissipate in time, which can easily cause the temperature of the charging pile and the battery to be too high, thus causing safety hazards. At present, the industry generally adopts passive temperature protection measures, such as temperature sensors, fan cooling, etc., which will stop charging once the temperature exceeds the limit. Although this method can prevent the risk of overtemperature in time, it also limits the use of charging power and reduces charging efficiency. On the other hand, there are complex dynamic correlations among many parameters involved in the charging process, such as ambient temperature, battery temperature, and charging pile temperature. The interaction between these parameters will affect the safety and stability of charging. However, the existing charging control technology is mainly based on a single temperature threshold, which cannot effectively perceive and predict the correlation changes between these parameters, and it is difficult to optimize charging adjustments in real time. Summary of the invention

[0004] In view of this, the present invention provides a high-power charging pile self-adjustment method, medium and system for high-temperature environment, which can solve the technical problem that the existing charging control technology is mainly based on a single temperature threshold and it is difficult to achieve real-time optimization of charging adjustment.

[0005] The present invention is achieved in that: A first aspect of the present invention provides a high-power charging pile self-adjustment method for a high-temperature environment, which comprises the following steps: S10, real-time acquisition of environmental parameters, charging temperature, input parameters, and output parameters of the charging pile during charging; S20, extracting the time domain features and frequency domain features of each time window by sliding the time window for each parameter obtained; S30, normalizing the acquired time domain features and frequency domain features of each parameter, calculating the Pearson correlation coefficient, applying principal component analysis to perform dimensionality reduction, and obtaining a two-dimensional correlation matrix of all parameters, wherein each element in the two-dimensional correlation matrix represents the correlation between the parameter corresponding to the row and the parameter corresponding to the column, and the element is represented by a correlation function; S40, using singular value decomposition to perform matrix decomposition on the correlation two-dimensional matrix to obtain a stability matrix and a variability matrix; the stability matrix represents the long-term stable relationship of the correlation between parameters; the variability matrix represents the short-term fluctuation relationship of the correlation between parameters; S50, based on the stability matrix, construct an autoregressive integrated moving average model to predict the long-term trend of the correlation between the parameters; at the same time, based on the volatility matrix, construct a conditional autoregressive model to predict the short-term fluctuation of the correlation between the parameters; S60, establishing a game model based on the autoregressive integrated moving average model and the conditional autoregressive model, including a long-term model with the long-term stability of the output parameters as the goal and a short-term model with the short-term stability of the output parameters as the goal, and determining the constraint conditions of the game model; S70, solving the game model to obtain the optimal output parameter as the output parameter of the charging pile in the next time period; S80, cyclically executing steps S10-S70 with a preset time period as the cycle period.

[0006] Specifically, the step S10 specifically includes: obtaining the environmental parameters, charging temperature, input parameters and output parameters of the charging process of the charging pile in real time. Among them, the environmental parameters include the ambient temperature and ambient humidity; the charging temperature includes the charging pile temperature, the battery temperature and the charging gun temperature, wherein the charging pile temperature is output by the temperature sensor built into the charging pile, the battery temperature is output by the temperature monitoring system built into the electric vehicle, and the charging gun temperature is collected by the temperature sensor on the charging gun. The input parameters are the parameters of the input power of the charging pile, including the input current and the input voltage; the output parameters are the parameters of the output power of the charging pile, including the output current, the output voltage and the output PWM waveform.

[0007] The specific steps of step S20 include: extracting the time domain features and frequency domain features of each time window by sliding the time window. The time domain features include characteristic parameters such as mean, second-order moment, standard deviation, skewness, kurtosis and root mean square value; the frequency domain features include characteristic parameters such as the first M main components of the power spectrum density, center frequency and mean frequency. This step aims to extract the time-frequency features that can describe the dynamic characteristics of the charging process from the time series data, laying the foundation for subsequent correlation analysis.

[0008] The specific steps of step S30 include: firstly, performing minimum-maximum normalization processing on the time domain and frequency domain features to eliminate the influence of dimension; then calculating the Pearson correlation coefficient between the parameters to obtain the correlation two-dimensional matrix; finally, applying principal component analysis (PCA) to reduce the dimension of the correlation matrix, selecting the main related features, and constructing the final correlation two-dimensional matrix. This step quantifies the correlation relationship between the parameters, which is conducive to the establishment of the subsequent game model.

[0009] The specific steps of step S40 include: performing singular value decomposition (SVD) on the correlation two-dimensional matrix obtained in step S30 to decompose it into a stability matrix and a variability matrix. The stability matrix represents the long-term stable relationship of the correlation between parameters, and the variability matrix represents the short-term fluctuation relationship of the correlation between parameters. This step separates the long-term and short-term characteristic components in the parameter correlation, providing a basis for the establishment of the subsequent prediction model.

[0010] The specific steps of step S50 include: constructing an autoregressive integrated moving average (ARIMA) model based on the stability matrix to predict the long-term trend of the correlation between the parameters; and constructing a conditional autoregressive (CAR) model based on the volatility matrix to predict the short-term fluctuation of the correlation between the parameters. The ARIMA model can capture the autoregressive and moving average characteristics in the time series, while the CAR model can establish the conditional dependency between the parameters, providing the required long-term and short-term predictions for the establishment of the subsequent game model.

[0011] The specific steps of step S60 include: constructing a game model based on the long-term and short-term prediction models obtained in step S50. The game model includes a long-term model with the long-term stability of the output parameters as the goal, and a short-term model with the short-term stability of the output parameters as the goal. There is a game relationship between the two models, and it is necessary to determine appropriate constraints, such as restrictions on output current and voltage, temperature, power, PWM waveform, and output parameter change rate, to ensure the safety and reliability of the charging process.

[0012] The specific steps of step S70 include: solving the game model constructed in step S60 to obtain the optimal output parameters as the output parameters of the charging pile in the next time period. The solution process can be iterative, such as the best response dynamic or policy gradient method, to find the Nash equilibrium solution. This step dynamically optimizes the charging strategy according to the current charging environment and historical charging data to meet the long-term and short-term stability requirements.

[0013] On the basis of the above technical solution, the self-adjusting method of a high-power charging pile for a high-temperature environment of the present invention can also be improved as follows: Among them, the environmental parameters include ambient temperature and ambient humidity; the charging temperature includes charging pile temperature, battery temperature and charging gun temperature; the charging pile temperature is the temperature output by the temperature sensor built into the charging pile, the battery temperature is the battery temperature output by the temperature monitoring system built into the electric vehicle, and the charging gun temperature is the temperature collected by the temperature sensor on the charging gun; the input parameters are the parameters of the electric energy input to the charging pile, including input current and input voltage; the output parameters are the parameters of the electric energy output by the charging pile, including output current, output voltage and output PWM waveform.

[0014] Furthermore, the output current and the output voltage are both floating point values.

[0015] Furthermore, the output PWM waveform is a vector, including the frequency, duty cycle, rise time, fall time, and dead time of the output PWM waveform, all of which are floating-point values.

[0016] Furthermore, the constraints of the game model specifically include: The output current and output voltage cannot exceed the rated output parameter range of the charging pile; The temperature of the charging pile, battery and charging gun cannot exceed the maximum operating temperature of the device; The output power cannot exceed the maximum output power of the charging pile; The output PWM waveform must meet the waveform requirements for charging electric vehicle batteries; The rate of change of the output parameters cannot exceed the preset rate of change threshold to avoid damage to the electric vehicle battery.

[0017] Wherein, the normalization process adopts the minimum-maximum normalization method.

[0018] Below is a formula describing the steps involved in the calculation: 1. Extraction of time domain features and frequency domain features: Time domain feature extraction formula: ; In the formula, is the time domain feature vector; is the mean; is the second-order moment; is the standard deviation; is the skewness; is the kurtosis; is the root mean square value.

[0019] The calculation formulas for these features are as follows: ; ; ; ; ; ; in, is the number of samples in the time window, For the Sample values.

[0020] Frequency domain feature extraction formula: ; In the formula, is the frequency domain eigenvector; The power spectral density main components; is the center frequency; is the mean frequency.

[0021] The power spectral density is calculated using the periodogram method: ; The calculation formula for center frequency and mean frequency is: ; ; in, For the frequency components, is the corresponding power spectral density value.

[0022] 2. Normalization processing: For time domain and frequency domain features, the minimum-maximum normalization method is used: ; In the formula, is the normalized value; is the original value; and are the minimum and maximum values ​​of the feature, respectively.

[0023] 3. Pearson correlation coefficient calculation: ; In the formula, for and Pearson correlation coefficient between two variables; and For the The value of samples; and They are and The average value of is the sample size.

[0024] 4. Principal Component Analysis (PCA) Dimensionality Reduction: First calculate the covariance matrix: ; In the formula, is the covariance matrix; For the The feature vector of samples; is the average feature vector of all samples; is the sample size.

[0025] Then solve the eigenvalue equation: ; In the formula, is the characteristic value, is the corresponding feature vector.

[0026] Before selection The eigenvectors corresponding to the largest eigenvalues ​​form the projection matrix , the data after dimension reduction is: ; In the formula, is the data after dimension reduction; is the original data; is the projection matrix.

[0027] 5. Construction of correlation two-dimensional matrix: ; In the formula, is a two-dimensional matrix of correlation; For the parameters and The correlation function between the parameters can be expressed as follows: ; in, For parameters and parameters The Euclidean distance of is the bandwidth parameter of the Gaussian kernel function.

[0028] 6. The formula for singular value decomposition (SVD) is: ; In the formula, is a two-dimensional matrix of correlation; and is an orthogonal matrix; is a diagonal matrix whose diagonal elements are singular values.

[0029] Calculation of stability matrix and volatility matrix: ; ; In the formula, is the stability matrix; is the volatility matrix; , and Respectively before The left singular vector, singular value, and right singular vector corresponding to the largest singular value. The choice can be based on the cumulative explained variance proportion, such as choosing the number of singular values ​​that reaches 95% of the cumulative explained variance.

[0030] 7. Autoregressive Integrated Moving Average (ARIMA) Model: ; In the formula, is a time series; is the lag operator; is an autoregressive polynomial; is the moving average polynomial; is the difference order; is a white noise sequence.

[0031] Specifically, ; ; in, is the autoregressive order, is the moving average order.

[0032] Parameter estimation can use the maximum likelihood estimation method: ; In the formula, is the log-likelihood function; is the sample size; is the white noise variance.

[0033] 8. Conditional Autoregressive (CAR) Model: ; ; In the formula, is a time series; is the conditional mean; is the autoregression coefficient; is a white noise sequence; is an exogenous variable; is the coefficient of the exogenous variable.

[0034] Parameter estimation can be done using conditional least squares: .

[0035] 9. Long-term model objective function: ; Short-term model objective function: ; In the formula, and are the objective functions of the long-term and short-term models respectively; and are the control strategies for long-term and short-term models respectively; is the system status; and are the reward functions for long-term and short-term models respectively; is the discount factor; The time horizon for short-term models; is the expectation operator.

[0036] Constraints: 10. Output current and voltage constraints: ; ; 11. Temperature constraints: ; ; ; 12. Output power constraints: ; 13.PWM waveform constraints: ; ; ; ; ; 14. Output parameter change rate constraints: ; ; In the formula, the subscripts min and max of all variables represent the minimum and maximum values, respectively, which can be determined according to the specific specifications of the charging pile and electric vehicle battery; the subscripts out and in represent the output and input, respectively.

[0037] The solution to the game model can adopt the concept of Nash equilibrium, that is, to find a set of strategies , so that: ; ; This problem can be solved iteratively, for example using best response dynamics or policy gradient methods. The specific solution algorithm needs to be selected according to the specific form and scale of the problem.

[0038] The relevant parameters are explained as follows: is the ambient temperature; is the ambient humidity; is the charging pile temperature; is the battery temperature; is the charging gun temperature; is the input current; is the input voltage; is the output current; is the output voltage; is the PWM frequency; is the PWM duty cycle; is the rise time; is the fall time; is the dead time; is the mean; is the second-order moment; is the standard deviation; is the skewness; is the kurtosis; is the root mean square value; is the power spectral density main components; is the center frequency; is the mean frequency; is the normalized value; is the minimum value of the feature; is the maximum value of the feature; for and Pearson correlation coefficient between two variables; is the covariance matrix; is the characteristic value; is the feature vector; is the projection matrix; is a two-dimensional matrix of correlation; For the parameters and Correlation function between parameters; is the bandwidth parameter of the Gaussian kernel function; , is an orthogonal matrix; is a diagonal matrix whose diagonal elements are singular values; is a time series; is the lag operator; is an autoregressive polynomial; is the moving average polynomial; is the difference order; is a white noise sequence; is the autoregressive order; is the moving average order; is the log-likelihood function; is the white noise variance; is a time series; is the conditional mean; is the autoregression coefficient; is an exogenous variable; is the coefficient of the exogenous variable; is the long-term model objective function; is the short-term model objective function; To model control strategies for the long term; for short-term model control strategies; is the system status; is the long-term model reward function; is the short-term model reward function; is the discount factor; The time horizon for short-term models; , is the minimum and maximum value of the output current; , is the minimum and maximum value of the output voltage; The maximum operating temperature of the charging pile; is the maximum operating temperature of the battery; The maximum operating temperature of the charging gun; is the maximum output power of the charging pile; , is the minimum and maximum value of PWM frequency; , is the minimum and maximum value of PWM duty cycle; , are the minimum and maximum values ​​of the rise time; , is the minimum and maximum value of the fall time; , is the minimum and maximum value of the dead time; is the maximum rate of change of output current; is the maximum rate of change of output voltage.

[0039] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are run in a computer, they are used to execute the above-mentioned high-power charging pile self-adjustment method for a high-temperature environment.

[0040] A third aspect of the present invention provides a high-power charging pile self-adjusting system for use in a high-temperature environment, which includes the above-mentioned computer-readable storage medium.

[0041] Compared with the prior art, the invention provides a high-power charging pile self-adjustment method, medium and system for high-temperature environment, which has the following beneficial effects: 1. Real-time monitoring and adaptive adjustment of the charging process are realized. This method can collect the environmental parameters, charging temperature, input and output parameters of the charging pile in real time, and use time-frequency feature analysis, correlation analysis and other means to deeply explore and model the dynamic characteristics of these parameters. On this basis, a game model of long-term and short-term charging goals is constructed, which can dynamically adjust the output parameters to meet the requirements of safety and charging efficiency in high temperature environments. Compared with the traditional charging control method based on temperature thresholds, this self-adjustment method is more intelligent and refined, and can better cope with complex charging scenarios.

[0042] 2. Improved the safety and reliability of high-power charging. In the prior art, high-power charging in a high-temperature environment is prone to cause safety hazards such as excessive temperature, and usually only simple temperature protection measures can be taken to limit the charging power. The self-adjusting method of the present invention can make full use of the potential of high-power charging, and maximize the output power, shorten the charging time, and improve the user experience while ensuring that key parameters such as the charging pile temperature, battery temperature, and charging gun temperature meet safety requirements. This charging control strategy based on multi-parameter dynamic optimization can more effectively prevent the risk of excessive temperature and improve the reliability of high-power charging.

[0043] 3. Enhanced the intelligent perception and decision-making capabilities of the charging system. The method of the present invention deeply explores the correlation rules between the parameters through the time-frequency characteristics analysis of the charging process parameters, and establishes a long-term and short-term correlation prediction model. These models can more accurately predict the dynamic relationship between parameters and provide a basis for subsequent charging strategy optimization. In addition, the method also constructs a game model that can seek the optimal balance between long-term and short-term charging goals and achieve intelligent decision-making. Compared with the traditional single condition constraint optimization, this charging strategy optimization method based on dynamic correlation analysis and game theory is more in line with the needs of complex charging scenarios and has stronger adaptability.

[0044] In general, the self-adjusting method proposed in the present invention integrates a variety of advanced mathematical and computer algorithms, which improves the charging efficiency while ensuring safe and reliable charging in a high temperature environment. The existing charging control technology is mainly based on a single temperature threshold, which makes it difficult to achieve real-time optimization of charging adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart of the method provided by the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0047] like Figure 1 As shown, it is a flow chart of a high-power charging pile self-adjustment method for a high-temperature environment provided by the first aspect of the present invention. The method comprises the following steps: S10, real-time acquisition of environmental parameters, charging temperature, input parameters, and output parameters of the charging pile during charging; S20, extracting the time domain features and frequency domain features of each time window by sliding the time window for each parameter obtained; S30, normalizing the acquired time domain features and frequency domain features of each parameter, calculating the Pearson correlation coefficient, applying principal component analysis to reduce the dimension, and obtaining a two-dimensional correlation matrix of all parameters, wherein each element in the two-dimensional correlation matrix represents the correlation between the parameter corresponding to the row and the parameter corresponding to the column, and the element is represented by a correlation function; S40, using singular value decomposition to decompose the correlation two-dimensional matrix to obtain a stability matrix and a variability matrix; the stability matrix represents the long-term stable relationship of the correlation between the parameters; the variability matrix represents the short-term fluctuation relationship of the correlation between the parameters; S50. Based on the stability matrix, an autoregressive integrated moving average model is constructed to predict the long-term trend of the correlation between the parameters; at the same time, based on the volatility matrix, a conditional autoregressive model is constructed to predict the short-term fluctuation of the correlation between the parameters; S60, establishing a game model based on an autoregressive integrated moving average model and a conditional autoregressive model, including a long-term model with the long-term stability of output parameters as a goal and a short-term model with the short-term stability of output parameters as a goal, and determining the constraints of the game model; S70, solving the game model to obtain the optimal output parameter as the output parameter of the charging pile in the next time period; S80, cyclically executing steps S10-S70 with a preset time period as the cycle period.

[0048] The specific implementation methods of the above steps are described in detail below: The specific implementation of step S10 is to obtain the environmental parameters, charging temperature, input parameters and output parameters of the charging process of the charging pile in real time. Among them, the environmental parameters include ambient temperature and ambient humidity; the charging temperature includes the charging pile temperature, battery temperature and charging gun temperature. The charging pile temperature is output by the built-in temperature sensor of the charging pile, the battery temperature is output by the built-in temperature monitoring system of the electric vehicle, and the charging gun temperature is collected by the temperature sensor on the charging gun. The input parameters are the parameters of the input electric energy of the charging pile, including input current and input voltage; the output parameters are the parameters of the electric energy output by the charging pile, including output current, output voltage and output PWM waveform.

[0049] The specific implementation of step S20 is to extract the time domain features and frequency domain features of each time window by sliding the time window. The time domain feature extraction formula is: ,in is the mean, is the second-order moment, is the standard deviation, is the skewness, is the kurtosis, is the root mean square value. The frequency domain feature extraction formula is ,in The power spectral density The main components, is the center frequency, The purpose of this step is to extract the time domain and frequency domain characteristics of each parameter to prepare for the subsequent correlation analysis.

[0050] The specific implementation of step S30 includes the following sub-steps: 1) performing minimum-maximum normalization processing on the time domain and frequency domain features so that each eigenvalue falls between 0 and 1; 2) calculating the Pearson correlation coefficient between each parameter to obtain a two-dimensional correlation matrix; 3) applying principal component analysis (PCA) to reduce the dimension of the correlation matrix to obtain the final two-dimensional correlation matrix. The elements in the two-dimensional correlation matrix Indicates Parameters and The correlation between the parameters is calculated using the Gaussian kernel function. The purpose of this step is to quantify the correlation between the parameters and lay the foundation for the subsequent game model establishment.

[0051] The specific implementation method of step S40 is to perform singular value decomposition (SVD) on the correlation two-dimensional matrix obtained in step S30 to obtain a stability matrix and the volatility matrix The stability matrix represents the long-term stable relationship of the correlation between parameters, and the volatility matrix represents the short-term fluctuation relationship of the correlation between parameters. This step aims to separate the long-term stable component and the short-term fluctuation component in the correlation between parameters, and further enrich the description of the correlation between parameters.

[0052] The specific implementation of step S50 includes the following sub-steps: 1) based on the stability matrix , construct an autoregressive integrated moving average (ARIMA) model to predict the long-term trend of the correlation between parameters; 2) based on the volatility matrix ,Construct a conditional autoregression (CAR) model to predict the short-term fluctuations of the correlation between parameters. The ARIMA model can capture the autoregression and moving average characteristics in the time series, while the CAR model can establish the conditional dependency between parameters. This step aims to predict the long-term trend and short-term fluctuations of parameter correlation respectively, providing input for the subsequent game model.

[0053] The specific implementation of step S60 is to construct a game model based on the long-term and short-term prediction models obtained in step S50. The game model includes a long-term model with the long-term stability of the output parameters as the goal, and a short-term model with the short-term stability of the output parameters as the goal. The objective function of the long-term model is to maximize the long-term expected benefits of the output parameters, and the objective function of the short-term model is to maximize the short-term expected benefits of the output parameters. There is a game relationship between the two models, and appropriate constraints need to be determined, including: 1) the output current and voltage cannot exceed the rated output parameter range of the charging pile; 2) the charging pile temperature, battery temperature and charging gun temperature cannot exceed the maximum operating temperature of the device; 3) the output power cannot exceed the maximum output power of the charging pile; 4) the output PWM waveform must meet the waveform requirements of the electric vehicle battery charging; 5) the rate of change of the output parameter cannot exceed the preset rate of change threshold to avoid damage to the electric vehicle battery. This step aims to build a charging control strategy that can balance long-term stability and short-term stability.

[0054] The specific implementation of step S70 is to solve the game model constructed in step S60 to obtain the optimal output parameters as the output parameters of the charging pile in the next time period. The solution process can be iterative, such as the best response dynamics or policy gradient method, to find the Nash equilibrium solution. The purpose of this step is to obtain the optimal charging strategy based on the current charging environment and charging history to meet the requirements of long-term stability and short-term stability.

[0055] The specific implementation of step S80 is to perform steps S10 to S70 in a loop with a preset duration as the cycle period. In this way, real-time self-regulation of the charging process can be achieved, and output parameters can be dynamically adjusted to meet the charging needs in a high temperature environment. This step ensures that the method can continue to operate, continuously optimize the charging strategy, and improve the reliability and safety of the charging pile in a high temperature environment.

[0056] In general, the self-adjustment method of a high-power charging pile for high-temperature environment proposed in the present invention acquires and analyzes various parameters of the charging process in real time, constructs long-term and short-term game models, and continuously solves and optimizes, and finally obtains a charging control strategy that meets the requirements of long-term and short-term stability. This method utilizes a variety of mathematical and computer algorithms such as time-frequency feature analysis, correlation analysis, and time series prediction, which can effectively cope with the complex dynamic characteristics of the charging process in a high-temperature environment and improve the safety and reliability of the charging pile.

[0057] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are run in a computer, they are used to execute the above-mentioned high-power charging pile self-adjustment method for a high-temperature environment.

[0058] A third aspect of the present invention provides a high-power charging pile self-adjusting system for use in a high-temperature environment, which includes the above-mentioned computer-readable storage medium.

[0059] Specifically, the principle of the present invention is: 1. Real-time perception of dynamic parameter characteristics of the charging process. This method can collect environmental parameters, charging temperature, input and output parameters of the charging pile in real time, and use time-frequency feature analysis to deeply explore these parameters. Specifically, the method first uses a sliding time window to extract the time domain characteristics (such as mean, variance, skewness, etc.) and frequency domain characteristics (such as power spectrum components, center frequency, etc.) of each time window. These time-frequency characteristics can more comprehensively describe the dynamic change law of the parameters. Subsequently, by normalizing and correlating these features, a two-dimensional matrix of correlation between the parameters is constructed. This step aims to quantify the dynamic coupling relationship between the parameters and provide a basis for subsequent prediction and optimization.

[0060] 2. Separate the long-term and short-term components in the parameter correlation. For the constructed correlation matrix, the method further applies the singular value decomposition (SVD) technique to perform matrix decomposition and obtains the stability matrix and volatility matrix. The former represents the long-term stable relationship of the correlation between parameters, while the latter represents the short-term fluctuation relationship of the correlation between parameters. This method of separating the long-term and short-term correlation components can more accurately characterize the complex dynamic coupling characteristics between parameters and provide richer information for subsequent predictions and decisions.

[0061] 3. Construct a game model for long-term and short-term charging targets. Based on the long-term and short-term correlation prediction model obtained above, a game model is established. The game model includes two optimization objectives: one is a long-term model with the long-term stability of the output parameters as the goal, and the other is a short-term model with the short-term stability of the output parameters as the goal. There is a certain game relationship between the two models, and it is necessary to seek the optimal balance under safety constraints. The construction of this game model fully considers the needs of long-term and short-term stability during charging in a high temperature environment, and can more comprehensively balance charging efficiency and safety.

[0062] 4. Dynamically optimize the charging strategy. Finally, the method solves the game model constructed above and obtains the optimal output parameters as the output parameters of the charging pile for the next time period. This solution process can use an iterative optimization algorithm, such as the best response dynamic or policy gradient method. Through continuous iterative solution, this method can dynamically adjust the charging output parameters to meet the requirements of long-term and short-term charging goals while ensuring that various safety constraints are met. This charging strategy optimization method based on dynamic correlation analysis and game theory is more in line with the needs of complex charging scenarios than traditional single condition constraint optimization and has strong adaptability.

[0063] In general, the technical principle of the method of the present invention is: through real-time perception of the dynamic parameter characteristics of the charging process, the long-term and short-term correlation components are separated, a game model of long-term and short-term charging goals is constructed, and the charging strategy is dynamically optimized to improve the charging efficiency while ensuring safe and reliable charging in a high temperature environment.

[0064] In order to better understand and implement the present invention, a specific embodiment 1 of the method of the present invention is provided below. The steps of this embodiment 1 are specifically described as follows: The specific implementation of step S10 is to obtain the environmental parameters, charging temperature, input parameters and output parameters of the charging process of the charging pile in real time. The environmental parameters include the ambient temperature and ambient humidity ;Charging temperature includes charging pile temperature , Battery temperature and charging gun temperature .in, Output from the temperature sensor built into the charging pile, Output from the temperature monitoring system built into the electric vehicle, The temperature is collected by the temperature sensor on the charging gun. The input parameters are the parameters of the charging pile input power, including the input current and input voltage ;The output parameters are the parameters of the electric energy output by the charging pile, including the output current , output voltage And output PWM waveform .

[0065] The specific implementation method of step S20 is to use a sliding time window to extract the time domain features and frequency domain features of each time window. The time domain feature extraction formula is: ; in, is the mean, is the second-order moment, is the standard deviation, is the skewness, is the kurtosis, is the root mean square value. The calculation formula is as follows: ; ; ; ; ; ; in, is the number of samples in the time window, For the Sample values. The frequency domain feature extraction formula is: ; in, The power spectral density The main components, is the center frequency, is the mean frequency. The power spectrum density is calculated using the periodogram method: ; The calculation formulas for the center frequency and mean frequency are: ; ; in, For the frequency components, is the corresponding power spectrum density value. The purpose of this step is to extract the time-frequency characteristics of each parameter and lay the foundation for subsequent correlation analysis.

[0066] The specific implementation of step S30 includes the following sub-steps: 1) performing minimum-maximum normalization processing on the time domain and frequency domain features: ; in, is the normalized value, is the original value, and are the minimum and maximum values ​​of the features respectively. 2) Calculate the Pearson correlation coefficient between each parameter: ; in, for and Pearson correlation coefficient between two variables; and For the The value of the samples; and They are and The average value of is the number of samples. 3) Apply principal component analysis (PCA) to reduce the dimension of the correlation matrix and construct the final two-dimensional correlation matrix ,in The Gaussian kernel function can be used to calculate: ; in, For parameters and parameters The Euclidean distance, is the bandwidth parameter of the Gaussian kernel function. This step quantifies the correlation between the parameters and lays the foundation for the subsequent game model establishment.

[0067] The specific implementation method of step S40 is to obtain the correlation two-dimensional matrix obtained in step S30. Perform singular value decomposition (SVD): ; in, and is an orthogonal matrix, is a diagonal matrix whose diagonal elements are singular values. Then the stability matrix and volatility matrix can be calculated: ; ; in, represents the long-term stable relationship between the correlations between parameters, Short-term fluctuations that represent correlations between parameters. , and Respectively before The left singular vector, singular value, and right singular vector corresponding to the largest singular value. The selection of can be based on the cumulative explained variance ratio, such as selecting the number of singular values ​​that reaches 95% of the cumulative explained variance. This step separates the long-term and short-term characteristic components in the parameter correlation, providing a basis for the establishment of the subsequent prediction model.

[0068] The specific implementation of step S50 includes the following sub-steps: 1) based on the stability matrix , build an autoregressive integrated moving average (ARIMA) model: ; in, is a time series; is the lag operator; is an autoregressive polynomial; is the moving average polynomial; is the difference order; is a white noise sequence. Specifically, ; ; in, is the autoregressive order, is the moving average order. Parameter estimation can use the maximum likelihood estimation method: ; in, is the log-likelihood function; is the sample size; is the white noise variance. 2) Based on the volatility matrix , build a conditional autoregressive (CAR) model: ; ; in, is a time series; is the conditional mean; is the autoregression coefficient; is a white noise sequence; is an exogenous variable; is the coefficient of the exogenous variable. Parameter estimation can use conditional least squares method: ; This step is based on long-term and short-term correlation prediction models, providing input for the establishment of subsequent game models.

[0069] The specific implementation method of step S60 is to construct a game model. The objective function of the long-term model is: ; The objective function of the short-term model is: ; in, and are the objective functions of the long-term and short-term models respectively; and Control strategies for long-term and short-term models respectively; is the system status; and are the reward functions for long-term and short-term models respectively; is the discount factor; is the time horizon of the short-term model. The constraints include: 1. Output current and voltage constraints: ; ; 2. Temperature constraints: ; ; ; 3. Output power constraints: ; 4.PWM waveform constraints: ; ; ; ; ; 5. Output parameter change rate constraints: ; ; This step constructs a game model that can balance long-term and short-term charging goals, laying the foundation for the subsequent solution of the optimal output parameters.

[0070] The specific implementation of step S70 is to solve the game model constructed in step S60 to obtain the optimal output parameters as the output parameters of the charging pile in the next time period. This solution process can be iterative, such as the best response dynamic or policy gradient method, to find the Nash equilibrium solution. , so that: ; ; This step dynamically optimizes the charging strategy based on the current charging environment and historical charging data to meet long-term and short-term stability requirements.

[0071] The specific implementation of step S80 is to perform steps S10 to S70 cyclically with a preset time period as the cycle period. In this way, real-time self-regulation of the charging process can be achieved, the charging strategy can be continuously optimized, and the reliability and safety of the charging pile in a high temperature environment can be improved.

[0072] In order to further better understand and implement the present invention, Example 2 of a specific application scenario of the present invention is provided below: A charging station adopts a high-power charging pile self-adjustment method for high-temperature environment proposed by the present invention.

[0073] 1. Overview of charging stations and environmental parameters The charging station has a total of 10 charging piles, each with a rated output power of 400kW. The charging piles are installed in an outdoor parking lot and need to cope with relatively harsh environmental conditions. According to the climate data of the local meteorological department, the area is summer from July to September each year, and the ambient temperature is often above 35°C. At the same time, the humidity is also high, often reaching more than 80%. In addition, there are some heat sources around the charging piles, such as photovoltaic power generation devices and transformers, which will cause certain heat interference to the charging environment.

[0074] In order to cope with this high temperature and high humidity charging environment, in addition to conventional temperature and humidity sensors, temperature sensors are installed inside the charging pile, in the battery compartment, and at the charging gun during the design of the charging station to monitor the temperature changes of key components in real time. At the same time, the input current, input voltage, output current, output voltage and other parameters of the charging pile are also continuously collected and recorded. The real-time collection of these environmental parameters, temperature parameters and electrical parameters provides the necessary data support for the subsequent self-regulation strategy.

[0075] 2. Dynamic characteristics analysis of charging process parameters Based on the above-mentioned real-time collected parameter data, the charging station uses the self-adjustment method proposed in the present invention to conduct an in-depth analysis of the dynamic characteristics of the parameters during the charging process. Specifically, it includes the following steps: (1) Time-frequency feature extraction. Taking 10 minutes as the time window, the time domain features and frequency domain features are extracted for each parameter time series. The time domain features include six indicators such as mean, second-order moment, standard deviation, skewness, kurtosis, and root mean square value; the frequency domain features include seven indicators such as the first five main components of the power spectrum density, center frequency, and mean frequency. In this way, each parameter corresponds to a 13-dimensional feature vector in each time window.

[0076] (2) Parameter correlation analysis. The time-frequency features extracted above were normalized by minimum-maximum normalization to eliminate the dimension effect. Then, the Pearson correlation coefficient between the parameters was calculated, and an 11×11 correlation two-dimensional matrix was constructed. This matrix quantitatively describes the dynamic coupling relationship between the parameters.

[0077] In order to further explore the long-term and short-term characteristics of parameter correlation, the charging station uses singular value decomposition (SVD) to decompose the correlation matrix and obtain the stability matrix and volatility matrix. The stability matrix represents the long-term stable relationship of the correlation between parameters, and the volatility matrix represents the short-term fluctuation relationship of the correlation between parameters.

[0078] (3) Predictive modeling of parameter correlation. Based on the stability matrix obtained above, the charging station constructed an autoregressive integrated moving average (ARIMA) model to predict the long-term trend of the correlation between the parameters. At the same time, based on the volatility matrix, a conditional autoregressive (CAR) model was constructed to predict the short-term fluctuations of the correlation between the parameters. These two prediction models provide the required long-term and short-term correlation change trends for the subsequent charging strategy optimization.

[0079] Through the above analysis, the charging station has mastered the time-frequency characteristics of various parameters in the charging process, as well as the complex dynamic coupling relationship between them. This has laid a solid foundation for the subsequent optimization of charging strategies based on correlation analysis.

[0080] 3. Charging strategy optimization based on game theory With the above in-depth analysis of the dynamic characteristics of charging process parameters, the charging station then built a game model to dynamically optimize the charging strategy and improve the charging efficiency while ensuring safety and reliability. The specific steps are as follows: (1) Objective function definition. The game model includes two objective functions: Long-term model objective function: ; in, is the objective function of the long-term model; Control strategies for long-term models; is the system status; is the reward function of the long-term model; is the discount factor. The objective function aims to maximize the long-term expected return of the output parameters.

[0081] Short-term model objective function: ; in, is the objective function of the short-term model; Control strategies for short-term models; is the reward function of the short-term model; is the time horizon of the short-term model. The objective function aims to maximize the short-term expected return of the output parameters.

[0082] There is a certain game relationship between the two models, and it is necessary to seek the optimal balance under safety constraints.

[0083] (2) Definition of constraints. When constructing the game model, the charging station set the following constraints: Temperature constraints: ; ; ; Current and voltage constraints: ; ; Power Constraints: ; PWM waveform constraints: ; ; ; ; Rate of change constraints: ; ; These constraints ensure the safety and reliability of the charging process and provide boundary limitations for subsequent strategy optimization.

[0084] (3) Game model solution. Based on the long-term and short-term objective functions and constraints defined above, the charging station uses an iterative optimization method to solve the Nash equilibrium solution of the game model. Specifically, the ARIMA and CAR models are used to predict the correlation trend between various parameters in the future. Then, on this basis, long-term and short-term objective functions are constructed, and the best response dynamic method is applied for iterative optimization until the optimal charging strategy that meets all constraints is found.

[0085] It should be noted that in each iteration cycle, the charging station will dynamically adjust the weights of the long-term and short-term objective functions according to the current charging environment and historical charging data to flexibly balance the needs of charging efficiency and safety. For example, in a high temperature environment, in order to ensure safety, the weight of the short-term stability goal can be appropriately increased; when the ambient temperature is low, the weight of the long-term efficiency goal can be appropriately increased to increase the charging power and shorten the charging time. This adaptive adjustment of the charging strategy enables the charging station to cope with various complex charging environment conditions.

[0086] 4. Analysis of Charging Station Operation By implementing the self-adjustment method proposed in the present invention, the operation of the charging station in a high temperature environment is summarized as follows: (1) Dynamic characteristics of charging process parameters. Table 1 lists the time-frequency characteristic statistics of various parameters in three 10-minute time windows from 10:00 to 10:30 during a certain charging process. It can be seen that as the charging time goes by, the characteristic values ​​of each parameter show a certain dynamic change pattern. For example, as the ambient temperature gradually increases over time, the charging pile temperature and the battery temperature also increase accordingly, but the change trends are not completely consistent. At the same time, there are certain fluctuations in the output current and output voltage. This complex dynamic coupling characteristic provides a basis for the subsequent optimization of the charging strategy.

[0087] (2) Parameter correlation analysis results. Table 2 lists the correlation matrix between the parameters in the charging process. It can be seen that the correlation between the ambient temperature and the charging pile temperature is strong, but the correlation with the battery temperature is weak; the correlation between the output current and the output voltage is strong, but the correlation with the charging pile temperature is weak. This complex correlation relationship between parameters provides a basis for the establishment of subsequent long-term and short-term prediction models.

[0088] (3) Charging strategy optimization effect. By solving the game model, the charging station dynamically adjusted the charging output parameters. Table 3 lists the optimization results at a certain moment. It can be seen that under the premise of meeting various constraints, the charging station appropriately reduced the charging power under the current high temperature environment conditions, and adjusted the output current and output voltage to 565A and 705V respectively to ensure that the temperature of the charging pile, battery and charging gun does not exceed the limit. At the same time, the charging pile also adjusted the output PWM waveform parameters accordingly to ensure the waveform requirements of battery charging. This adaptive charging strategy based on correlation analysis and game theory enables the charging station to maintain high charging efficiency and safety in a high temperature environment.

[0089] In general, by implementing the self-adjustment method proposed in the present invention, the charging station can perceive the dynamic parameter characteristics of the charging process in real time, establish a game model of long-term and short-term charging goals, and dynamically optimize the charging output parameters. Compared with the traditional charging control method based on temperature threshold, the self-adjustment method is more intelligent and refined, and greatly improves the charging efficiency and user experience while ensuring safe and reliable charging in a high temperature environment.

[0090] Table 1 Time-frequency characteristic statistics of each parameter during a charging process Time Window Ambient temperature Ambient humidity Charging station temperature Battery temperature Charging gun temperature Input Current Input voltage Output Current Output voltage Output PWM frequency Output PWM duty cycle Output PWM rise time Output PWM fall time Output PWM dead time 10:00-10:10 36.2 83.5% 48.3 32.7 43.2 612 712 580 722 385 0.51 18.2 20.1 10.5 10:10-10:20 37.1 84.2% 50.1 33.4 44.8 623 718 575 715 392 0.53 18.7 20.6 11.0 10:20-10:30 38.0 84.9% 51.9 34.1 46.5 635 724 570 708 400 0.55 19.3 21.2 11.5 Table 2 Correlation matrix between parameters Ambient temperature Ambient humidity Charging station temperature Battery temperature Charging gun temperature Input Current Input voltage Output Current Output voltage Output PWM frequency Output PWM duty cycle Ambient temperature 1.00 0.76 0.87 0.62 0.73 0.54 0.48 0.50 0.46 0.41 0.38 Ambient humidity 0.76 1.00 0.71 0.58 0.69 0.42 0.39 0.41 0.38 0.35 0.32 Charging station temperature 0.87 0.71 1.00 0.65 0.81 0.59 0.52 0.55 0.50 0.46 0.43 Battery temperature 0.62 0.58 0.65 1.00 0.72 0.47 0.43 0.45 0.42 0.38 0.35 Charging gun temperature 0.73 0.69 0.81 0.72 1.00 0.53 0.69 0.61 0.51 0.42 0.47 Input Current 0.54 0.42 0.59 0.47 0.53 1.00 0.83 0.92 0.81 0.62 0.58 Input voltage 0.48 0.39 0.52 0.43 0.49 0.83 1.00 0.75 0.88 0.55 0.52 Output Current 0.50 0.41 0.55 0.45 0.51 0.92 0.75 1.00 0.83 0.59 0.55 Output voltage 0.46 0.38 0.50 0.42 0.47 0.81 0.88 0.83 1.00 0.52 0.49 Output PWM frequency 0.41 0.35 0.46 0.38 0.43 0.62 0.55 0.59 0.52 1.00 0.71 Output PWM duty cycle 0.38 0.32 0.43 0.35 0.40 0.58 0.52 0.55 0.49 0.71 1.00 Table 3 Charging strategy optimization results at a certain moment parameter Numeric Output Current 565A Output voltage 705V Output Power 398kW Charging station temperature 67.2°C Battery temperature 41.8°C Charging gun temperature 57.3°C Output PWM frequency 438Hz Output PWM duty cycle 0.61 Output PWM rise time 21.5ms Output PWM fall time 23.8ms Output PWM dead time 12.8ms Through the above data analysis, it can be seen that in a high temperature environment, the charging station adopts the self-adjustment method proposed by the present invention, and through dynamic optimization of the charging strategy, the charging power is increased as much as possible under the premise of meeting various constraints, and the temperature of the charging pile, battery and charging gun is also controlled within a safe range. At the same time, the charging pile also adjusts the output PWM waveform parameters according to the battery charging requirements.

[0091] Compared with simple temperature protection measures, this self-adjustment method can more intelligently perceive and predict the dynamic characteristics of various parameters during the charging process, build a game model for long-term and short-term charging goals, and dynamically optimize the charging strategy. This not only improves the charging efficiency, but also effectively enhances the safety and reliability of the charging system.

[0092] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A high-power charging pile self-adjustment method for high-temperature environment, characterized in that: The following steps are involved: S10, real-time acquisition of environmental parameters, charging temperature, input parameters, and output parameters of the charging pile during charging; S20, extracting the time domain features and frequency domain features of each time window by sliding the time window for each parameter obtained; S30, normalizing the acquired time domain features and frequency domain features of each parameter, calculating the Pearson correlation coefficient, applying principal component analysis to perform dimensionality reduction, and obtaining a two-dimensional correlation matrix of all parameters, wherein each element in the two-dimensional correlation matrix represents the correlation between the parameter corresponding to the row and the parameter corresponding to the column, and the element is represented by a correlation function; S40, using singular value decomposition to perform matrix decomposition on the correlation two-dimensional matrix to obtain a stability matrix and a variability matrix; the stability matrix represents the long-term stable relationship of the correlation between parameters; The volatility matrix represents the short-term fluctuation relationship of the correlation between parameters; S50, based on the stability matrix, construct an autoregressive integrated moving average model to predict the long-term trend of the correlation between the parameters; at the same time, based on the volatility matrix, construct a conditional autoregressive model to predict the short-term fluctuation of the correlation between the parameters; S60, establishing a game model based on the autoregressive integrated moving average model and the conditional autoregressive model, including a long-term model with the long-term stability of the output parameters as the goal and a short-term model with the short-term stability of the output parameters as the goal, and determining the constraint conditions of the game model; S70, solving the game model to obtain the optimal output parameter as the output parameter of the charging pile in the next time period; S80, cyclically executing steps S10-S70 with a preset time period as the cycle period.

2. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 1, characterized in that: The environmental parameters include ambient temperature and ambient humidity; the charging temperature includes charging pile temperature, battery temperature and charging gun temperature; the charging pile temperature is the temperature output by the temperature sensor built into the charging pile, the battery temperature is the battery temperature output by the temperature monitoring system built into the electric vehicle, and the charging gun temperature is the temperature collected by the temperature sensor on the charging gun; The input parameters are the parameters of the electric energy input to the charging pile, including input current and input voltage; the output parameters are the parameters of the electric energy output by the charging pile, including output current, output voltage, and output PWM waveform.

3. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 2, characterized in that: The output current and the output voltage are both floating point values.

4. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 3, characterized in that: The output PWM waveform is a vector, including the frequency, duty cycle, rise time, fall time, and dead time of the output PWM waveform, all of which are floating-point values.

5. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 4, characterized in that: The constraints of the game model specifically include: The output current and output voltage cannot exceed the rated output parameter range of the charging pile; The temperature of the charging pile, battery and charging gun cannot exceed the maximum operating temperature of the device; The output power cannot exceed the maximum output power of the charging pile; The output PWM waveform must meet the waveform requirements for charging electric vehicle batteries; The rate of change of the output parameters cannot exceed the preset rate of change threshold to avoid damage to the electric vehicle battery.

6. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 1, characterized in that: The normalization process adopts the minimum-maximum normalization method.

7. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 6, characterized in that: The calculation of stability matrix and volatility matrix is ​​specifically expressed as: ; ; In the formula, is the stability matrix; is the volatility matrix; , and Respectively before The left singular vector, singular value, and right singular vector corresponding to the largest singular value.

8. A high-power charging pile self-adjustment method for a high-temperature environment according to claim 7, characterized in that: The objective function of the long-term model is expressed as: ; The objective function of the short-term model is expressed as: ; In the formula, and are the objective functions of the long-term and short-term models respectively; and are the control strategies for long-term and short-term models respectively; is the system status; and are the reward functions for long-term and short-term models respectively; is the discount factor; The time horizon for short-term models; is the expectation operator.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions, and when the program instructions are executed in a computer, the program instructions are used to execute the high-power charging pile self-adjustment method for a high-temperature environment as described in any one of claims 1 to 8.

10. A high-power charging pile self-adjusting system for high-temperature environments, characterized in that: A computer-readable storage medium comprising the computer-readable storage medium of claim 9.

Images