A data-driven based super charging pile charging process segmentation regression fitting method

Abstract

The application discloses a kind of based on data-driven super charging pile charging process segmentation regression fitting method, it is related to electrical engineering technical field, including the data of super charging pile charging process is collected and is standardized processing, generates multidimensional time series data set, based on multidimensional time series data set using equivalent capacitance-resistance model establishes the segmentation model including rising stage, constant current stage, constant voltage stage and trickle stage;Using segmentation model calculates the fitting value of each time point, obtains power sequence, constructs bayesian hierarchical model to carry out change point detection to power sequence, and uses Markov chain monte carlo method to sample change point after distribution, obtains stage boundary result.The application establishes bayesian hierarchical model by combining change point detection technology, accurately optimizes the stage division and fitting accuracy of charging process, improves the prediction accuracy, efficiency and stability of charging process.

Description

Technical Field

[0001] This invention relates to the field of electrical engineering technology, and in particular to a data-driven piecewise regression fitting method for the charging process of supercharging piles. Background Technology

[0002] With the rapid development of electric vehicles and charging infrastructure, supercharging stations have become key equipment for fast charging of electric vehicles. To improve the efficiency and accuracy of the charging process, researchers and engineers have proposed various modeling and control methods for the charging process. Traditional battery charging modeling typically uses equivalent circuit models based on electrochemical models. These models usually employ capacitors and resistors to simulate the dynamic behavior of the battery charging process. However, they have certain limitations, such as inaccurate parameter estimation, cumbersome modeling processes, and a lack of dynamic identification of changing points, making it difficult to accurately reflect the instantaneous changes in the complex charging process.

[0003] Most existing battery charging models rely on empirical formulas or simplified electrochemical equivalent models for fitting, but these methods often ignore the subtle differences between different stages of battery charging. Traditional models often use only a single regression method for full-process fitting, lacking sufficient modeling of the changing patterns of the charging process, resulting in limited predictive power. When there are significant variations in the charging process, existing methods often struggle to effectively capture instantaneous changes, thus affecting the improvement of charging efficiency and the accurate control of the charging process. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a data-driven segmented regression fitting method for the charging process of supercharging piles, which solves the problems of inaccurate change points at each stage, large regression fitting errors, and imprecise stage division in the charging process model.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a data-driven piecewise regression fitting method for the charging process of supercharging piles, comprising, Data from the charging process of supercharging piles is collected and standardized to generate a multidimensional time series dataset. Based on the multidimensional time series dataset, a segmented model including the rising phase, constant current phase, constant voltage phase and trickle phase is established using the equivalent capacitance-resistance model. The piecewise model is used to calculate the fitted value at each time point to obtain the power sequence. A Bayesian hierarchical model is constructed to detect change points in the power sequence. The Markov chain Monte Carlo method is used to sample the distribution after the change points to obtain the stage boundary results. Based on the stage boundary results, a corresponding regression algorithm is used for each stage to obtain a set of piecewise regression models. The fitting error and goodness of fit of the set of piecewise regression models are calculated, and the set of piecewise regression models is optimized based on the calculation results. The charging pile operation data and the piecewise regression fitting data are uploaded to the database for storage.

[0007] As a preferred embodiment of the data-driven piecewise regression fitting method for the charging process of supercharging piles described in this invention, the step of collecting data from the supercharging pile charging process and performing standardization processing to generate a multidimensional time series dataset refers to collecting current, voltage, power, and temperature data while the supercharging pile is running, cleaning the collected data, removing outliers, reducing noise, and performing standardization processing using the Z-score method. Based on continuous multidimensional data within a period, a multidimensional time series dataset is obtained.

[0008] As a preferred embodiment of the data-driven segmented regression fitting method for the charging process of supercharging piles described in this invention, the segmented model consisting of rising phase, constant current phase, constant voltage phase and trickle phase is established using an equivalent capacitance-resistance model based on a multidimensional time series dataset. This means using the least squares method to fit and train the current and voltage data in the multidimensional time series dataset to obtain the trained equivalent capacitance-resistance model. Based on the equivalent capacitance-resistance model, the current and voltage relationships in the rising phase, constant current phase, constant voltage phase, and trickle phase are established respectively to obtain the piecewise model.

[0009] As a preferred embodiment of the data-driven segmented regression fitting method for the charging process of supercharging piles described in this invention, the step of using a segmented model to calculate the fitting value at each time point and obtain the power sequence refers to multiplying the current and voltage of each stage of the segmented model, calculating the fitting values ​​of the rising stage, constant current stage, constant voltage stage and trickle stage respectively, and obtaining the power sequence based on the continuous fitting values ​​within the period.

[0010] As a preferred embodiment of the data-driven segmented regression fitting method for the charging process of supercharging piles described in this invention, the method of constructing a Bayesian hierarchical model to detect change points in the power sequence and using the Markov chain Monte Carlo method to sample the distribution after the change points to obtain the stage boundary results assumes that the power of each stage follows a Gaussian distribution, obtains the likelihood function of each stage, and obtains the overall likelihood function based on the likelihood function of each stage. The overall likelihood function is used as the observation layer of the Bayesian hierarchical model, and the number and location of the variable points in each stage are set as the parameter layer of the Bayesian hierarchical model. The Bayesian hierarchical model is obtained based on the observation layer and the parameter layer. Initialize the variable point positions, update the random walk proposals for the variable point positions, and calculate the acceptance probability of the random walk proposals. The updated variable point positions are used as variable point samples. When the number of iterations reaches the set upper limit, the median of all change point samples is calculated as the change point estimate and a confidence interval is set as the stage boundary result of the change point.

[0011] As a preferred embodiment of the data-driven segmented regression fitting method for the charging process of a supercharging pile described in this invention, the step of obtaining a segmented regression model set by adopting a corresponding regression algorithm for each stage based on the stage boundary results refers to using exponential regression to fit the current and voltage for the rising stage, linear regression to fit the voltage and current for the constant current stage, exponential decay regression to fit the current change for the constant voltage stage, and constant regression to fit the current change for the trickle stage. The regression models of each stage are spliced ​​together based on the charging cycle to obtain the segmented regression model set.

[0012] As a preferred embodiment of the data-driven segmented regression fitting method for the charging process of supercharging piles described in this invention, wherein: calculating the fitting error and goodness of fit of the segmented regression model set, and optimizing the segmented regression model set according to the calculation results refers to calculating the fitting error and goodness of fit of each stage of the segmented regression model set. Set the fitting error threshold ,like Then continue to update the variable point walk proposal until the fitting error is less than the set fitting error threshold; Set the goodness-of-fit threshold ,like If the regression fitting algorithm for a given stage is changed, the goodness of fit will continue until it exceeds the set goodness of fit threshold.

[0013] As a preferred embodiment of the data-driven piecewise regression fitting method for the charging process of supercharging piles described in this invention, the step of uploading the charging pile operation data and piecewise regression fitting data to the database for storage refers to uploading the state data of the charging pile during operation and the data generated during the piecewise regression fitting process as empirical data to the database for storage.

[0014] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the data-driven piecewise regression fitting method for the charging process of a supercharging pile as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the data-driven piecewise regression fitting method for the charging process of a supercharging pile as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: By standardizing the data of the supercharging pile charging process and generating a multi-dimensional time series dataset, an accurate piecewise model is established by combining the equivalent capacitance-resistance model, which can more accurately capture the charging characteristics of each stage; a Bayesian hierarchical model is used to detect change points in the power sequence, and the distribution after sampling the change points is sampled using the Markov chain Monte Carlo method, which accurately identifies the stage boundaries in the charging process. The most suitable regression algorithm can be selected for each stage, further improving the fitting accuracy of the regression model; an optimized piecewise regression model set is established, which not only improves the fitting accuracy of each stage, but also enhances the stability and adaptability of the overall model, improves the prediction accuracy and charging efficiency of the charging process, and ensures the efficient operation and precise control of the charging pile in practical applications. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Fig. 1 The flowchart shows a piecewise regression fitting method for the charging process of a data-driven supercharging pile.

[0019] Fig. 2 A schematic diagram for establishing a segmented model.

[0020] Fig. 3 A schematic diagram illustrating the stage boundary determination for variable point detection. Detailed Implementation

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0023] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0024] Reference Figs. 1-3 This is one embodiment of the present invention, which provides a data-driven piecewise regression fitting method for the charging process of supercharging piles, including the following steps: Data from the charging process of supercharging piles is collected and standardized to generate a multidimensional time series dataset. Based on the multidimensional time series dataset, a segmented model including the rising phase, constant current phase, constant voltage phase, and trickle phase is established using the equivalent capacitance-resistance model.

[0025] Specifically, during the operation of the supercharging pile, current, voltage, power, and temperature data are collected. The collected data is cleaned, outliers are removed, and noise is reduced. The Z-score method is used for standardization. Based on continuous multidimensional data within a period, a multidimensional time series dataset is obtained.

[0026] By collecting current, voltage, power, and temperature data from supercharging piles during operation, and performing data cleaning, outlier removal, and noise reduction, the quality and reliability of the data were ensured. The Z-score method was used to standardize the data, eliminating dimensional differences and achieving a unified scale for all data. A multidimensional time-series dataset was constructed, providing an accurate and clear data foundation for subsequent modeling.

[0027] Furthermore, the least squares method is used to fit and train the current and voltage data in the multidimensional time series dataset to obtain the trained equivalent capacitance-resistance model. The formula for the equivalent capacitance-resistance model is: ; in, For time index, For charging piles in Voltage at time, For charging piles in Current at any moment Equivalent resistance This is the equivalent capacitance.

[0028] Based on the equivalent capacitance-resistance model, the current and voltage relationships in the rising phase, constant current phase, constant voltage phase and trickle phase are established respectively to obtain the piecewise model; In the rising phase of the piecewise model, the current increases exponentially, while the voltage increases linearly in sync, as shown in the formula: ; in, The maximum current value is set. It is a time constant. The initial voltage, This represents the rate of voltage rise.

[0029] In the constant current phase of the piecewise model, the current remains constant while the voltage increases linearly, as shown in the formula: ; in, The set constant current value, This represents the rate of change of voltage.

[0030] In the constant-voltage phase of the piecewise model, the voltage remains constant, and the current decays exponentially, as shown in the formula: ; in, This marks the start of the constant pressure phase. The current attenuation coefficient affected by temperature. This is the set constant voltage value.

[0031] In the trickle phase of the piecewise model, the current maintains a minimum safe value, and the voltage fluctuates slightly. The formula is: ; in, For minimum safe current, The voltage cutoff threshold. This is a fluctuation term.

[0032] An accurate equivalent capacitance-resistance model was established by fitting and training current and voltage data from a multidimensional time series dataset using the least squares method. Based on this model, the current-voltage relationship was established for the rising, constant current, constant voltage, and trickle charging stages, forming a piecewise model. This effectively and accurately models different stages of the charging process, ensuring a precise description of the current-voltage relationship at each stage and improving the prediction accuracy and stability of the charging process.

[0033] The power sequence is obtained by calculating the fitted value at each time point using a segmented model. A Bayesian hierarchical model is then constructed to detect change points in the power sequence. Finally, the distribution after the change points is sampled using the Markov chain Monte Carlo method to obtain the stage boundary results.

[0034] Specifically, the current and voltage of each stage of the segmented model are multiplied together to calculate the fitted values ​​for the rising stage, constant current stage, constant voltage stage and trickle stage respectively. Based on the continuous fitted values ​​within the period, the power sequence is obtained.

[0035] By multiplying the current and voltage data of each stage in the piecewise model, the fitted values ​​of the rising stage, constant current stage, constant voltage stage and trickle stage are calculated respectively, ensuring that the power calculation of each stage can be fitted according to the accurate current and voltage relationship, providing high-precision power data for subsequent analysis.

[0036] Furthermore, assuming that the power in each stage follows a Gaussian distribution, the likelihood function for each stage is obtained as follows: ; in, For stage index, For the first The likelihood function for each stage, For the first The turning points of each stage For the first The observed variance at each stage, For charging piles in Actual power at time, For the piecewise model in the first The charging piles are calculated in each stage. Power at any given moment.

[0037] The overall likelihood function is obtained based on the likelihood functions at each stage, and the formula is: ; in, For the overall power sequence, For a set of variable points, For the total number of variable points, The charging piles calculated for the segmented model are in Power at any given moment.

[0038] Using the overall likelihood function as the observation layer of the Bayesian hierarchical model, and setting priors for the number and location of change points in each stage as the parameter layer of the Bayesian hierarchical model, the Bayesian hierarchical model is obtained based on the observation layer and the parameter layer, as shown in the formula: ; in, For Bayesian hierarchical models, Given the prior distribution of the number of variable points, The prior distribution is the location of the variable point.

[0039] The formula for initializing the variable point position is: ; in, , , These represent the initial positions of the change points in the constant flow stage, constant pressure stage, and trickle flow stage, respectively. , , These are the time percentage coefficients for the constant current stage, constant pressure stage, and trickle stage, respectively. This represents the historical average charging time for each charging session at the charging station.

[0040] The proposal to update the variable point location is based on a random walk. Each updated variable point location is used as a variable point sample, and the formula is as follows: ; in, For variable point indexes, For the first The new location of the changing point Index for iteration count, For the first The variable point is at the 1st Position at the next iteration The random perturbation is set, and Follows a normal distribution. The proposed distribution variance for random walks.

[0041] The probability of accepting a random walk proposal is calculated using the following formula: Pro ; Where Pro is the probability of accepting the random walk proposal. For the first The posterior probability of a new position at a variable point. For the first The posterior probability of the current position of each variable point.

[0042] When the number of iterations reaches the set upper limit, the median of all change point samples is calculated as the change point estimate and a confidence interval is set as the stage boundary result of the change point.

[0043] By assuming that the power in each stage follows a Gaussian distribution, the likelihood function for each stage is obtained. Based on these likelihood functions, an overall likelihood function is constructed as the observation layer of the Bayesian hierarchical model. Priors for the number and location of change points in each stage are set as the parameter layer of the Bayesian hierarchical model, forming a complete model. By initializing the change point locations and performing random walk proposal updates, calculating and accepting change point updates, and progressively sampling change point locations, the median of the change point samples is calculated after multiple iterations, and confidence intervals are set to obtain the stage boundary results. This achieves accurate detection of change points and stage division during the charging process, improves the accuracy of change point identification, and ensures the stability and reliability of stage division.

[0044] Based on the stage boundary results, a corresponding regression algorithm is used for each stage to obtain a set of piecewise regression models. The fitting error and goodness of fit of the set of piecewise regression models are calculated, and the set of piecewise regression models is optimized based on the calculation results. The charging pile operation data and the piecewise regression fitting data are uploaded to the database for storage.

[0045] Specifically, exponential regression is used to fit the current and voltage during the rising phase, linear regression is used to fit the voltage and current during the constant current phase, exponential decay regression is used to fit the current change during the constant voltage phase, and constant regression is used to fit the current change during the trickle phase. The regression models for each phase are then spliced ​​together based on the charging cycle to obtain a set of piecewise regression models.

[0046] By employing different regression algorithms for fitting at different stages—exponential regression for the rising stage, linear regression for the constant current stage, exponential decay regression for the constant voltage stage, and constant regression for the trickle stage—accurate modeling of current and voltage changes in each stage is achieved. The regression models for each stage are then concatenated based on the charging cycle to form a complete set of piecewise regression models. This ensures that the most suitable regression method is used for each stage, enabling precise fitting of each stage of the charging process and improving the overall model's accuracy and predictive power.

[0047] Furthermore, the fitting error of each stage of the piecewise regression model set is calculated using the following formula: ; in, For the first The fitting error value for the stage.

[0048] The goodness of fit for each stage of the piecewise regression model set is calculated using the following formula: ; in, For the first Goodness of fit at each stage For the first The average actual power of the stage.

[0049] Set the fitting error threshold ,like Then continue to update the variable point walk proposal until the fitting error is less than the set fitting error threshold; Set the goodness-of-fit threshold ,like If the regression fitting algorithm for a given stage is changed, the goodness of fit will continue until it exceeds the set goodness of fit threshold.

[0050] By calculating the fitting error and goodness of fit at each stage of the piecewise regression model ensemble, and setting thresholds for both fitting error and goodness of fit, the model is ensured to achieve the required accuracy. When the fitting error at a certain stage exceeds the set threshold, further optimization is achieved through a change-point walk proposal until the error falls below the threshold. When the goodness of fit falls below the set threshold, the regression fitting algorithm is promptly replaced until the goodness of fit reaches the standard. This effectively improves the accuracy and stability of the regression model at each stage, ensuring the overall model's efficient and accurate predictions during the charging process.

[0051] Furthermore, the status data of the charging pile during operation and the data generated during the piecewise regression fitting process are used as empirical data and uploaded to the database for storage.

[0052] By uploading the status data of the charging pile during operation and the data generated during the piecewise regression fitting process as empirical data to the database for storage, key information in the charging process can be systematically stored, providing reliable data support for subsequent charging process analysis, optimization and model iteration.

[0053] This embodiment also provides a computer device applicable to the data-driven piecewise regression fitting method for the charging process of supercharging piles, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the data-driven piecewise regression fitting method for the charging process of supercharging piles as proposed in the above embodiment.

[0054] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0055] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the piecewise regression fitting method for the charging process of a supercharging pile based on data-driven principles as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0056] In summary, this invention achieves the following: First, it standardizes the data from the supercharging process of a supercharger and generates a multi-dimensional time-series dataset. Second, it establishes an accurate piecewise model using an equivalent capacitance-resistance model, enabling more precise capture of the charging characteristics at each stage. Third, it employs a Bayesian hierarchical model to detect change points in the power sequence and uses a Markov chain Monte Carlo method to sample the distribution after these change points, accurately identifying the stage boundaries in the charging process. This allows for the selection of the most suitable regression algorithm for each stage, further improving the fitting accuracy of the regression model. Finally, it establishes an optimized set of piecewise regression models, which not only improves the fitting accuracy of each stage but also enhances the stability and adaptability of the overall model, improving the prediction accuracy and charging efficiency of the charging process. This ensures the efficient operation and precise control of the supercharger in practical applications.

[0057] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A data-driven piecewise regression fitting method for the charging process of supercharging piles, characterized in that: include, Data from the charging process of supercharging piles is collected and standardized to generate a multidimensional time series dataset. Based on the multidimensional time series dataset, a segmented model including the rising phase, constant current phase, constant voltage phase and trickle phase is established using the equivalent capacitance-resistance model. The piecewise model is used to calculate the fitted value at each time point to obtain the power sequence. A Bayesian hierarchical model is constructed to detect change points in the power sequence. The Markov chain Monte Carlo method is used to sample the distribution after the change points to obtain the stage boundary results. Based on the stage boundary results, a corresponding regression algorithm is used for each stage to obtain a set of piecewise regression models. The fitting error and goodness of fit of the set of piecewise regression models are calculated, and the set of piecewise regression models is optimized based on the calculation results. The charging pile operation data and the piecewise regression fitting data are uploaded to the database for storage.

2. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 1, characterized in that: The process of collecting and standardizing data from the supercharging pile during its charging process to generate a multidimensional time series dataset refers to collecting current, voltage, power, and temperature data while the supercharging pile is running, cleaning the collected data, removing outliers, reducing noise, and standardizing it using the Z-score method. Based on continuous multidimensional data within a period, a multidimensional time series dataset is obtained.

3. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 2, characterized in that: The establishment of a piecewise model based on a multidimensional time series dataset using an equivalent capacitance-resistance model, which includes a rising phase, a constant current phase, a constant voltage phase, and a trickle phase, refers to using the least squares method to fit and train the current and voltage data in the multidimensional time series dataset to obtain the trained equivalent capacitance-resistance model. Based on the equivalent capacitance-resistance model, the current and voltage relationships in the rising phase, constant current phase, constant voltage phase, and trickle phase are established respectively to obtain the piecewise model.

4. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 3, characterized in that: The process of using a piecewise model to calculate the fitted value at each time point and obtain the power sequence involves multiplying the current and voltage of each stage of the piecewise model, calculating the fitted values ​​for the rising stage, constant current stage, constant voltage stage, and trickle stage respectively, and obtaining the power sequence based on the continuous fitted values ​​within the period.

5. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 4, characterized in that: The construction of the Bayesian hierarchical model to detect change points in the power sequence and the use of the Markov chain Monte Carlo method to sample the distribution after the change points to obtain the stage boundary results means assuming that the power of each stage follows a Gaussian distribution, obtaining the likelihood function of each stage, and obtaining the overall likelihood function based on the likelihood function of each stage. The overall likelihood function is used as the observation layer of the Bayesian hierarchical model, and the number and location of the variable points in each stage are set as the parameter layer of the Bayesian hierarchical model. The Bayesian hierarchical model is obtained based on the observation layer and the parameter layer. Initialize the variable point positions, update the random walk proposals for the variable point positions, and calculate the acceptance probability of the random walk proposals. The updated variable point positions are used as variable point samples. When the number of iterations reaches the set upper limit, the median of all change point samples is calculated as the change point estimate and a confidence interval is set as the stage boundary result of the change point.

6. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 5, characterized in that: Based on the stage boundary results, a corresponding regression algorithm is adopted for each stage to obtain a set of piecewise regression models. For the rising stage, exponential regression is used to fit the current and voltage; for the constant current stage, linear regression is used to fit the voltage and current; for the constant voltage stage, exponential decay regression is used to fit the current change; and for the trickle stage, constant regression is used to fit the current change. The regression models of each stage are spliced ​​together based on the charging cycle to obtain a set of piecewise regression models.

7. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 6, characterized in that: The calculation of the fitting error and goodness of fit of the piecewise regression model set, and the optimization of the piecewise regression model set based on the calculation results, refers to the calculation of the fitting error and goodness of fit of each stage of the piecewise regression model set. Set the fitting error threshold ,like Then continue to update the variable point walk proposal until the fitting error is less than the set fitting error threshold; Set the goodness-of-fit threshold. ,like If the regression fitting algorithm for a given stage is changed, the goodness of fit will continue until it exceeds the set goodness of fit threshold.

8. The piecewise regression fitting method for the charging process of a supercharging pile based on data-driven methods as described in claim 7, characterized in that: The step of uploading charging pile operation data and piecewise regression fitting data to the database for storage refers to uploading the state data of the charging pile during operation and the data generated during the piecewise regression fitting process as empirical data to the database for storage.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the data-driven piecewise regression fitting method for the charging process of a supercharging pile as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the data-driven piecewise regression fitting method for the charging process of a supercharging pile as described in any one of claims 1 to 8.

Images