Charging pile and system based on power regulation

Abstract

The present invention discloses a charging pile and system based on power regulation, including a state information acquisition module, a data preprocessing and feature extraction module, a dynamic change pattern analysis module, a dynamic change prediction module and a power regulation optimization module; the state information acquisition module acquires various state information of the battery in real time during the period when the charging pile charges the vehicle battery; the data preprocessing and feature extraction module preprocesses the original data in the data set. The present invention acquires battery state information in real time, extracts the SOC nonlinear change amplitude and the thermal management demand change rate, combines data analysis and machine learning prediction, accurately captures the dynamic changes of battery charging, and uses fuzzy logic algorithm to dynamically adjust the power regulation frequency, realizes efficient charging power distribution, significantly improves charging efficiency, reduces power waste and safety risks, prolongs battery life, optimizes charging pile resource utilization, and provides innovative technical solutions for intelligent electric vehicle charging.

Description

Technical Field

[0001] The present invention relates to the technical field of charging pile power regulation, and in particular to a charging pile and a system based on power regulation. Background Art

[0002] A charging pile based on power regulation is an electric vehicle charging device that can intelligently adjust the output power. It achieves dynamic power allocation by real-time monitoring of grid load, charging demand, battery status and user settings. Its core technologies include power regulation algorithms, smart grid communication protocols and energy management systems, which can optimize charging speed and efficiency while ensuring grid stability. This technology is suitable for distributed power management during peak load periods, which can relieve grid pressure, reduce charging costs, support demand response plans, and improve power resource utilization.

[0003] The existing charging piles generally rely on fixed adjustment frequency in power regulation. Their power control algorithms are designed based on static rules and low computational complexity, aiming to simplify system implementation and reduce hardware costs. However, this fixed-frequency regulation method is difficult to adapt to the dynamic changes of vehicle battery status (such as remaining power SOC, temperature changes or charging mode requirements), and cannot flexibly optimize the charging rate according to real-time needs, resulting in slower charging speeds for some vehicles, longer user waiting times, and lower overall charging efficiency. At the same time, the non-optimality of power distribution not only causes waste of electricity, but also increases user costs, weakens the resource utilization of charging piles, and is difficult to meet the complex charging needs of modern electric vehicles.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a charging pile and system based on power control, which obtains battery status information in real time, extracts the SOC nonlinear change amplitude and the change rate of thermal management system requirements, and combines data analysis within the monitoring window period to accurately capture the dynamic mode of battery charging, inputs feature data into a pre-trained machine learning model, and provides accurate prediction of the dynamic changes of battery charging. Based on the prediction results, a fuzzy logic algorithm is used to dynamically adjust the power control frequency, which is flexibly converted into an optimized control strategy, thereby realizing dynamic response and precise control of charging power, greatly improving charging efficiency, shortening user waiting time, reducing energy waste and operating costs, while reducing the risks of overcharging and overheating, extending battery life, optimizing charging pile resource utilization, meeting the complex charging needs of modern electric vehicles, and providing innovative solutions for intelligent charging to solve the problems in the above-mentioned background technology.

[0006] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: a charging system based on power regulation, comprising a state information acquisition module, a data preprocessing and feature extraction module, a dynamic change pattern analysis module, a dynamic change prediction module and a power regulation optimization module;

[0007] The status information collection module obtains various status information of the battery in real time while the charging pile is charging the vehicle battery;

[0008] The data preprocessing and feature extraction module preprocesses the raw data in the data set, and then extracts the nonlinear change amplitude of the battery SOC in a short period of time and the change rate of the battery's demand for the thermal management system;

[0009] The dynamic change pattern analysis module analyzes the nonlinear change amplitude of the battery SOC in a short period of time within the set monitoring window period, captures the change pattern of the SOC during the charging process, analyzes the change rate of the battery's demand for the thermal management system, and evaluates the change in the battery's demand for the thermal management system based on the change trend of the battery temperature;

[0010] The dynamic change prediction module inputs the analyzed feature data into a pre-trained machine learning model. The machine learning model predicts the dynamic changes of the battery during the charging process based on the input data.

[0011] The power control optimization module uses a fuzzy logic algorithm to dynamically adjust the power control frequency based on the prediction results of the machine learning model. Specifically, the prediction results are converted into control actions through a fuzzy logic algorithm, and the adjustment step size is set through fuzzy rules.

[0012] Preferably, within the set monitoring window period, after analyzing the nonlinear change amplitude of the battery SOC in a short period of time, an SOC transient change reference value is generated, and by capturing the change pattern of the SOC during the charging process and analyzing the change rate of the battery's demand for the thermal management system, a thermal management demand surge reference value is generated, and the change in the battery's demand for the thermal management system is evaluated through the thermal management demand surge reference value.

[0013] Preferably, the steps for generating the SOC transient change reference value are as follows:

[0014] The instantaneous acceleration of SOC during the monitoring window is calculated to capture the nonlinear characteristics of SOC changes. The instantaneous acceleration is expressed by the first-order derivative and second-order derivative of SOC. The specific expression is: ,in, is the instantaneous acceleration of SOC, is the remaining battery power at time t, is the first-order derivative of SOC, that is, the instantaneous rate of change of SOC, is the second-order derivative of SOC, that is, the rate of change of SOC change rate;

[0015] In order to further capture the high-frequency fluctuation characteristics of SOC changes, wavelet transform is used to perform time-frequency analysis on SOC signals. The SOC signal is decomposed into sub-signals of different frequencies by continuous wavelet transform. The decomposition expression is: ,in, For scale a and location b The wavelet coefficients under is the selected wavelet basis function, a is the scale parameter, which determines the frequency resolution of the analysis, b is the time shift parameter, which determines the time position of the analysis;

[0016] The instantaneous acceleration is obtained by weighted fusion method. and high frequency fluctuation wavelet coefficients Combined, the final SOC transient change reference value is generated, and the generated expression is: ,in, Indicates the SOC transient change reference value, and is the weight coefficient, reflecting the relative importance of instantaneous acceleration and high-frequency fluctuations, is the absolute value of instantaneous acceleration, quantifying the acceleration of SOC change, is the sum of the energy of high-frequency wavelet coefficients at different scales, reflecting the high-frequency fluctuation intensity of SOC changes. is the energy amplitude of high-frequency fluctuations, which is used to normalize the contribution of high-frequency features.

[0017] Preferably, in order to extract high-frequency fluctuation features, the high-frequency energy of the wavelet coefficients within the high-frequency range at each time point is calculated, and the calculation formula for the high-frequency energy is: ,in, is the high frequency energy at time t, is the selected high frequency scale range, The square of the wavelet coefficient representing each high-frequency scale reflects the local energy of the signal at that frequency;

[0018] The high frequency energy With energy threshold Compare and extract high-frequency fluctuation features, specifically: , then mark the time point as high-frequency fluctuation, retain the high-frequency wavelet coefficient of the time point as a feature, and record the high-frequency wavelet coefficient of the time point as the high-frequency fluctuation wavelet coefficient ; Otherwise, set it to zero; High-frequency fluctuation wavelet coefficient It is the high-frequency wavelet feature after threshold screening.

[0019] Preferably, the steps for obtaining the thermal management demand sudden increase reference value are as follows:

[0020] During the monitoring window, the current thermal management demand of the battery is obtained at each time step. The thermal management demand change rate is defined as the rate of change of the thermal management demand between two consecutive time steps. The calculation expression of the thermal management demand change rate is: ,in, It represents the instantaneous demand of the battery on the thermal management system at time t, reflecting the power required for cooling or heating. Indicated in Thermal management needs at all times, is the time step, which indicates the time interval between two data samplings;

[0021] According to the absolute value of the instantaneous change rate, the nonlinear cumulative change within the monitoring window period is calculated. The calculation expression is: ,in, It is the nonlinear cumulative change of thermal management demand during the monitoring window period, reflecting the dynamic change intensity of thermal management demand during the entire monitoring window period. is the start time of the monitoring window period, T is the total duration of the monitoring window period, is a nonlinear adjustment coefficient used to amplify the impact of drastic changes. is the time decay factor, which is used to reduce the weight of changes in the latter part of the monitoring window period. is the time decay function, which decreases as t increases;

[0022] Normalize the nonlinear cumulative change with the total demand baseline value during the monitoring window to generate a thermal management demand surge reference value. The generated expression is: ,in, It is the reference value of thermal management demand surge, reflecting the change rate and sudden increase intensity of thermal management demand during the monitoring window period. is the total demand benchmark value during the monitoring window period, defined as: , which represents the total demand level of the thermal management system during the entire window period.

[0023] Preferably, during the set monitoring window period, after obtaining the SOC transient change reference value generated by analyzing the nonlinear change amplitude of the battery SOC in a short period of time and the thermal management demand surge reference value generated by analyzing the change rate of the battery's demand on the thermal management system, the SOC transient change reference value and the thermal management demand surge reference value are input into a pre-trained machine learning model to generate a charging dynamic change index, and the dynamic changes of the battery during the charging process are predicted by the charging dynamic change index.

[0024] Preferably, based on the prediction results of the machine learning model, the power control frequency is dynamically adjusted using a fuzzy logic algorithm, and the specific steps are as follows:

[0025] Step 1: Define fuzzy variables, set the reference threshold of charging dynamic change index and the starting power control frequency;

[0026] Specifically, the charging dynamic change index obtained from the machine learning model is used as the input variable of the fuzzy logic algorithm. According to the reference threshold of the charging dynamic change index, the charging dynamic change index is divided into different fuzzy sets. At the same time, a starting power control frequency is determined. The power control frequency is used as the output fuzzy variable and is divided into fuzzy sets.

[0027] Step 2: Apply fuzzy logic rules to convert the charging dynamic change index into control actions;

[0028] Specifically, a fuzzy rule base is established to describe the relationship between input variables and output variables, the real-time charging dynamic change index value is fuzzified, the membership of the charging dynamic change index in each fuzzy set is calculated, and then, according to the fuzzy rules, the input membership is inferred to obtain the fuzzy result of the output variable;

[0029] Step 3: Defuzzify the output, set the adjustment step size and dynamically adjust the power control frequency;

[0030] Specifically, the output result obtained by fuzzy reasoning is defuzzified and the fuzzy output is converted into a specific numerical value, that is, the adjustment step of the power control frequency. According to this adjustment step, the power control frequency of the charging pile is dynamically updated, so that the power control frequency of the charging pile adjusts the power output more frequently or less frequently to adapt to the real-time demand changes of the battery.

[0031] Charging piles based on power regulation, charging systems based on power regulation.

[0032] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0033] The present invention obtains battery status information in real time, extracts the nonlinear change amplitude of SOC and the change rate of thermal management system requirements, and combines data analysis within the monitoring window period to accurately capture the key patterns of dynamic changes in the battery charging process, and inputs the characteristic data obtained by analysis into a pre-trained machine learning model, thereby providing accurate prediction capabilities for dynamic changes in the battery charging process. Based on the prediction results, the fuzzy logic algorithm is used to dynamically adjust the power regulation frequency, thereby realizing flexible transformation of complex prediction results into optimized regulation strategies, effectively ensuring dynamic response and precise control of charging power distribution, greatly improving charging efficiency, shortening user waiting time, reducing energy waste and operating costs, and significantly reducing the risk of overcharging or overheating, extending battery life, and successfully meeting the complex charging needs of modern electric vehicles, optimizing the resource utilization of charging piles, and providing innovative and forward-looking technical solutions for intelligent electric vehicle charging. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0035] Figure 1 It is a module schematic diagram of the charging system based on power regulation of the present invention. DETAILED DESCRIPTION

[0036] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0037] The present invention provides Figure 1 The charging system based on power regulation shown includes a state information acquisition module, a data preprocessing and feature extraction module, a dynamic change mode analysis module, a dynamic change prediction module and a power regulation optimization module;

[0038] The status information collection module obtains various status information of the battery in real time while the charging pile is charging the vehicle battery;

[0039] The battery status information includes but is not limited to the remaining capacity (SOC), battery temperature, voltage, current, and charging mode requirements. This data can be obtained through the vehicle's battery management system (BMS) or the communication interface installed between the charging station and the vehicle. All collected data is systematically stored in a dynamically updated data set;

[0040] This step ensures that the system can continuously monitor the real-time status of the battery and provide comprehensive and accurate basic data for subsequent data analysis and prediction.

[0041] The data preprocessing and feature extraction module preprocesses the raw data in the data set, and then extracts the nonlinear change amplitude of the battery SOC in a short period of time and the change rate of the battery's demand for the thermal management system;

[0042] The dynamic change pattern analysis module analyzes the nonlinear change amplitude of the battery SOC in a short period of time within the set monitoring window period, captures the change pattern of the SOC during the charging process, analyzes the change rate of the battery's demand for the thermal management system, and evaluates the change in the battery's demand for the thermal management system (cooling or heating) based on the change trend of the battery temperature;

[0043] Preprocessing the raw data in a data set is a key step in data analysis and modeling, which mainly includes the following: data cleaning (removing outliers and noise data to ensure data accuracy and consistency), missing value filling (filling missing values ​​through interpolation, averaging or predictive models to prevent incomplete data from affecting analysis results), normalization and standardization (converting data of different dimensions into a unified scale to ensure that different features have the same weight in the analysis), and data smoothing (reducing random fluctuations in data to more clearly reflect trends).

[0044] Data preprocessing ensures the high quality and relevance of input data, eliminates noise and bias, and optimizes the results of subsequent analysis and model training, thus laying a solid foundation for accurate prediction and decision-making.

[0045] A large nonlinear change in the battery SOC (State of Charge) in a short period of time usually indicates that the dynamic change of the battery state is more drastic. The nonlinear change of SOC reflects the battery's ability to accept the charging current, the battery's chemical reaction rate, and the rapid fluctuation of the internal resistance. For example, in the initial charging stage after deep discharge, the battery may quickly absorb a large amount of current, resulting in a significant change in SOC; when it is close to full, the SOC change may show a nonlinear sharp increase or fluctuation due to the voltage platform effect. If the SOC changes greatly in a short period of time, it often means that the battery charging characteristics or external conditions (such as temperature, power input) have changed dramatically. In this case, the charging pile needs to dynamically adjust the power control strategy to avoid a decrease in charging efficiency or damage to battery life. Therefore, the short-term nonlinear change amplitude of SOC is an important indicator to measure the dynamic state of the battery.

[0046] During the set monitoring window, the nonlinear change amplitude of the battery SOC in a short period of time is analyzed to generate a SOC transient change reference value. By capturing the change pattern of the SOC during the charging process, the steps for generating the SOC transient change reference value are as follows:

[0047] The instantaneous acceleration of SOC during the monitoring window is calculated to capture the nonlinear characteristics of SOC changes. The instantaneous acceleration is expressed by the first-order derivative and second-order derivative of SOC. The specific expression is: ,in, is the instantaneous acceleration of SOC, is the remaining battery power at time t, is the first-order derivative of SOC, that is, the instantaneous rate of change of SOC, is the second-order derivative of SOC, that is, the rate of change of SOC change rate;

[0048] By calculating the instantaneous acceleration of SOC, the nonlinearity of SOC change can be quantified. When the change of SOC shows an acceleration or deceleration trend, the instantaneous acceleration will deviate significantly from zero, reflecting the dynamic fluctuation of the battery state. This indicator helps to identify potential rapid change stages during the charging process.

[0049] In order to further capture the high-frequency fluctuation characteristics of SOC changes, wavelet transform is used to perform time-frequency analysis on SOC signals. The appropriate wavelet basis function is selected, and the SOC signal is decomposed into sub-signals of different frequencies through continuous wavelet transform (CWT). The decomposition expression is: ,in, For scale a and location b The wavelet coefficients under is the selected wavelet basis function (e.g. Morlet wavelet),a is the scale parameter, which determines the frequency resolution of the analysis, b is the time shift parameter, which determines the time position of the analysis;

[0050] Wavelet transform can effectively separate different frequency components in SOC signal, especially high-frequency fluctuations. By extracting high-frequency fluctuation features, subtle dynamic changes and sudden events in SOC changes can be identified. These high-frequency features reflect the rapid response and nonlinear behavior of the battery during charging.

[0051] The instantaneous acceleration is obtained by weighted fusion method. and high frequency fluctuation wavelet coefficients Combined, the final SOC transient change reference value is generated, and the generated expression is: ,in, Indicates the SOC transient change reference value, and is the weight coefficient, reflecting the relative importance of instantaneous acceleration and high-frequency fluctuations, is the absolute value of instantaneous acceleration, quantifying the acceleration of SOC change, is the sum of the energy of high-frequency wavelet coefficients at different scales, reflecting the high-frequency fluctuation intensity of SOC changes. is the energy amplitude of high-frequency fluctuations, which is used to normalize the contribution of high-frequency features;

[0052] By weighted fusion of instantaneous acceleration and high-frequency fluctuation energy, STCI can comprehensively reflect the nonlinear and high-frequency dynamic characteristics of SOC changes. and It can be adjusted according to the specific application scenario to cope with the sensitivity of different change characteristics. Ultimately, the SOC transient change reference value provides a comprehensive indicator for the charging system to evaluate the dynamic change degree of the battery state, thereby supporting more accurate power regulation and charging strategy optimization.

[0053] It can be seen from the SOC transient change reference value that within the set monitoring window period, the larger the performance value of the SOC transient change reference value generated after analyzing the nonlinear change amplitude of the battery SOC in a short period of time, the more drastic the dynamic change of the battery state, and vice versa, the more stable the battery state. This is because the SOC transient change reference value comprehensively reflects the dynamic characteristics of the battery charging process by quantifying the nonlinear change amplitude, high-frequency fluctuation characteristics and change rate of the battery SOC in a short period of time. When the SOC transient change reference value is large, it usually means that the SOC change pattern contains rapid nonlinear fluctuations and significant high-frequency characteristics, which may be caused by a sharp change in charging rate, a sharp fluctuation in battery chemical reactions, or the influence of external environment (such as temperature). When the SOC transient change reference value is small, it indicates that the SOC change is relatively stable, the charging state is closer to the linear law, and the battery charging process tends to be stable. Therefore, the SOC transient change reference value can be used as a key indicator to measure the degree of battery state fluctuation, and provide a scientific basis for the adjustment of charging pile control strategy.

[0054] In order to extract high-frequency fluctuation characteristics, the high-frequency energy of the wavelet coefficients in the high-frequency range at each time point is calculated. The calculation formula of high-frequency energy is: ,in, is the high frequency energy at time t, For the selected high frequency scale range, for example, the high frequency part can be defined as the scale range, where is the critical scale, which determines the boundary of the high-frequency part. The square of the wavelet coefficient representing each high-frequency scale reflects the local energy of the signal at that frequency.

[0055] The high frequency energy With energy threshold Compare and extract high-frequency fluctuation features, specifically: , then mark the time point as high-frequency fluctuation, retain the high-frequency wavelet coefficient of the time point as a feature, and record the high-frequency wavelet coefficient of the time point as the high-frequency fluctuation wavelet coefficient ; Otherwise, set it to zero; High-frequency fluctuation wavelet coefficient It is the high-frequency wavelet feature after threshold screening.

[0056] During the period when the charging station is charging the vehicle battery, if the rate of change of the battery's demand for the thermal management system is high, it usually indicates that the dynamic changes of the battery state are large. This is because the rapid fluctuations in thermal management requirements reflect the significant changes in battery temperature in a short period of time, and the temperature change is closely related to the battery's charging and discharging characteristics, internal electrochemical reaction rates, and power requirements. For example, during fast charging, high currents will cause increased heating inside the battery, requiring the cooling system to respond quickly; similarly, in low temperature environments, the battery may need to be heated quickly to improve charging efficiency. The high rate of change of thermal management requirements means that the battery state is in a highly dynamic adjustment, and the power output needs to be optimized to ensure the safety and efficiency of the charging process.

[0057] During the set monitoring window, after analyzing the change rate of the battery's demand for the thermal management system, a thermal management demand surge reference value is generated. The thermal management demand surge reference value is used to evaluate the change in the battery's demand for the thermal management system (cooling or heating). The steps for obtaining the thermal management demand surge reference value are as follows:

[0058] During the monitoring window, the current thermal management demand of the battery is obtained at each time step. The thermal management demand change rate is defined as the rate of change of the thermal management demand between two consecutive time steps. The calculation expression of the thermal management demand change rate is: ,in, It represents the instantaneous demand of the battery on the thermal management system at time t, reflecting the power required for cooling or heating. Indicated in Thermal management needs at all times, is the time step, which indicates the time interval between two data samples (e.g. 1 second);

[0059] This step quantifies the rate of change of thermal management requirements in a short period of time, providing basic data for subsequent evaluation of sudden increases.

[0060] According to the absolute value of the instantaneous change rate, the nonlinear cumulative change within the monitoring window period is calculated. The calculation expression is: ,in, It is the nonlinear cumulative change of thermal management demand during the monitoring window period, reflecting the dynamic change intensity of thermal management demand during the entire monitoring window period. is the start time of the monitoring window period, T is the total duration of the monitoring window period, is a nonlinear adjustment factor used to amplify the impact of drastic changes (permanent , such as 2 or 3), is a time decay factor used to reduce the weight of changes in the later stages of the monitoring window to simulate the greater impact of early changes. is the time decay function, which decreases as t increases;

[0061] By accumulating the nonlinear values ​​of the instantaneous rate of change, the contribution of sudden and early changes in thermal management demand to the overall dynamics is emphasized, and smooth change patterns are filtered out.

[0062] Normalize the nonlinear cumulative change with the total demand baseline value during the monitoring window to generate a thermal management demand surge reference value. The generated expression is: ,in, It is the reference value of thermal management demand surge, reflecting the change rate and sudden increase intensity of thermal management demand during the monitoring window period. is the total demand benchmark value during the monitoring window period, defined as: , represents the total demand level of the thermal management system during the entire window period;

[0063] The thermal management demand surge reference value generated by the normalization operation can quantify the sudden increase in the battery's demand for the thermal management system, highlight the situation where the demand changes dramatically, and provide an accurate indicator for evaluating the dynamic state of the battery.

[0064] It can be seen from the thermal management demand surge reference value that within the set monitoring window period, the larger the performance value of the thermal management demand surge reference value generated after analyzing the rate of change of the battery's demand for the thermal management system, the higher the rate of change of the battery's demand for the thermal management system, and the more significant the nonlinear cumulative change, indicating that the battery's temperature characteristics and charging behavior have fluctuated violently, and the state adjustment has become more frequent; on the contrary, if the thermal management demand surge reference value is small, it means that the battery's demand for thermal management has a low rate of change, a small fluctuation range, a stable dynamic change, and a relatively stable state. Therefore, the size of the thermal management demand surge reference value can be used as an important indicator to judge the severity of the dynamic change of the battery state, providing a key reference for optimizing charging strategies and ensuring battery safety.

[0065] The dynamic change prediction module inputs the analyzed feature data into a pre-trained machine learning model. The machine learning model predicts the dynamic changes of the battery during the charging process based on the input data.

[0066] During the set monitoring window, after obtaining the SOC transient change reference value generated by analyzing the nonlinear change amplitude of the battery SOC in a short period of time and the thermal management demand surge reference value generated by analyzing the change rate of the battery's demand for the thermal management system, the SOC transient change reference value and the thermal management demand surge reference value are input into a pre-trained machine learning model to generate a charging dynamic change index, and the dynamic change of the battery during the charging process is predicted by the charging dynamic change index.

[0067] A pre-trained machine learning model refers to an algorithm model that has completed the model training process using a large amount of historical data in the charging dynamic change prediction system and has a high prediction ability. The core of this model is that it establishes a mathematical mapping that can accurately predict the state change trend during battery charging by analyzing and learning the complex nonlinear relationship between key input features such as the battery SOC transient change reference value and the thermal management demand surge reference value and the battery dynamic change. The model training process usually includes steps such as data collection, feature engineering, model selection, parameter optimization, and performance evaluation. Through these steps, the model captures the laws of various dynamic change characteristics during battery charging in the training data, such as SOC change patterns, temperature fluctuation characteristics, and their impact on charging performance and safety, so that it can cope with complex working conditions in actual environments.

[0068] In this scenario, a pre-trained model means that the machine learning model has been fully trained based on a large amount of historical data including SOC changes, thermal management requirements, and other related parameters before it is officially applied. Through these training data, the model learns how to map input features such as SOC transient change reference values ​​and thermal management demand surge reference values ​​to the output result of the charging dynamic change index, thereby predicting the dynamic changes of the battery. The model may take the form of deep learning (such as multi-layer neural networks), tree models (such as random forests, XGBoost), or probabilistic statistics-based algorithms (such as Bayesian models), and the specific choice depends on the complexity of the problem and the characteristics of the data. The pre-trained model has the ability to quickly generate reliable predictions when new data is input through optimized parameter settings and weight distribution. Its significance lies in that it not only significantly improves the prediction efficiency and accuracy, but also ensures that the model can quickly adapt to the real-time state changes of the battery in actual operation, providing a scientific basis for the power regulation and optimization strategy of the charging pile.

[0069] The machine learning model is not specifically limited here, and can realize the SOC transient change reference value and thermal management demand sudden increase reference value Comprehensive analysis to generate charging dynamic change index In order to realize the technical solution of the present invention, the present invention provides a specific implementation method; the charging dynamic change index The resulting calculation formula is: , where , They are SOC transient change reference values and thermal management demand sudden increase reference value The preset scaling factor of , Both are greater than 0.

[0070] It can be seen from the charging dynamic change index that within the set monitoring window period, the greater the performance value of the SOC transient change reference value generated after analyzing the nonlinear change amplitude of the battery SOC in a short period of time, the greater the performance value of the thermal management demand sudden increase reference value generated after analyzing the change rate of the battery's demand for the thermal management system. That is, the greater the performance value of the charging dynamic change index generated when the battery charging process is predicted by a pre-trained machine learning model, the more drastic the dynamic change of the battery is, and vice versa.

[0071] The power control optimization module uses the fuzzy logic algorithm to dynamically adjust the power control frequency based on the prediction results of the machine learning model. Specifically, the prediction results are converted into control actions through the fuzzy logic algorithm, and the adjustment step size is set through fuzzy rules.

[0072] Based on the prediction results of the machine learning model, the fuzzy logic algorithm is used to dynamically adjust the power control frequency. The specific steps are as follows:

[0073] Step 1: Define fuzzy variables, set the reference threshold of charging dynamic change index and the starting power control frequency;

[0074] First, the charging dynamic change index obtained from the machine learning model is used as the input variable of the fuzzy logic algorithm. According to experience and system requirements, the reference threshold of the charging dynamic change index is set and divided into different fuzzy sets, such as "low change", "medium change" and "high change". At the same time, a starting power regulation frequency is determined, which is the basic frequency for the charging pile to adjust the power under normal circumstances. The power regulation frequency, as an output fuzzy variable, is also divided into fuzzy sets, such as "reduce frequency", "maintain frequency" and "increase frequency".

[0075] The purpose of this step is to establish the fuzzy variables and their membership functions of the input and output for the fuzzy logic system, build the basis of fuzzy reasoning, and enable the system to identify different degrees of battery dynamic changes and make corresponding adjustments.

[0076] Step 2: Apply fuzzy logic rules to convert the charging dynamic change index into control actions;

[0077] Next, a set of fuzzy rule bases are established. These rules are based on expert experience or system optimization requirements and describe the relationship between input variables (charging dynamic change index) and output variables (power control frequency adjustment). Typical fuzzy rules may include:

[0078] Rule 1: If the charging dynamic change index is "high change", the power control frequency should be "increase frequency".

[0079] Rule 2: If the charging dynamic change index is "medium change", the power control frequency should be "maintain frequency".

[0080] Rule 3: If the charging dynamic change index is "low change", the power control frequency should be "reduced frequency".

[0081] The real-time charging dynamic change index value is processed by fuzzification to calculate its membership in each fuzzy set. Then, according to the fuzzy rules, the input membership is inferred to obtain the fuzzy result of the output variable (power control frequency adjustment).

[0082] The purpose of this step is to use fuzzy logic reasoning to convert the complex, nonlinear prediction results of the charging dynamic change index into a specific control strategy, so that the system can determine whether the power control frequency needs to be increased, maintained unchanged, or decreased according to the degree of dynamic change of the battery state.

[0083] Step 3: Defuzzify the output, set the adjustment step size and dynamically adjust the power control frequency;

[0084] The output results obtained by fuzzy reasoning are defuzzified. Common methods include the centroid method or the average maximum membership method to convert the fuzzy output into a specific value, that is, the adjustment step of the power control frequency. For example, it is calculated that the power control frequency needs to be increased by 5 Hz or decreased by 3 Hz. According to this adjustment step, the power control frequency of the charging pile is dynamically updated to adjust the power output more frequently or less frequently to adapt to the real-time demand changes of the battery.

[0085] The purpose of this step is to apply the fuzzy reasoning results to actual control. Through precise numerical adjustment, it can achieve dynamic optimization of power regulation frequency, improve charging efficiency, ensure battery safety, and meet users' charging needs.

[0086] The present invention obtains battery status information in real time, extracts the nonlinear change amplitude of SOC and the change rate of thermal management system requirements, and combines data analysis within the monitoring window period to accurately capture the key patterns of dynamic changes in the battery charging process. The characteristic data obtained by analysis is input into a pre-trained machine learning model, providing accurate prediction capabilities for dynamic changes in the battery charging process. Based on the prediction results, the fuzzy logic algorithm is used to dynamically adjust the power regulation frequency, realizing the flexible transformation of complex prediction results into optimized regulation strategies, effectively ensuring the dynamic response and precise control of charging power distribution. In practical applications, the system has greatly improved charging efficiency, shortened user waiting time, reduced energy waste and operating costs, and significantly reduced the risk of overcharging or overheating, extending battery life. In addition, the solution successfully meets the complex charging needs of modern electric vehicles, optimizes the resource utilization of charging piles, and provides an innovative and forward-looking technical solution for intelligent electric vehicle charging.

[0087] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0088] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0089] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A charging system based on power regulation, characterized in that: It includes a state information acquisition module, a data preprocessing and feature extraction module, a dynamic change pattern analysis module, a dynamic change prediction module, and a power control optimization module; The status information collection module obtains various status information of the battery in real time while the charging pile is charging the vehicle battery; The data preprocessing and feature extraction module preprocesses the raw data in the data set, and then extracts the nonlinear change amplitude of the battery SOC in a short period of time and the change rate of the battery's demand for the thermal management system; The dynamic change pattern analysis module analyzes the nonlinear change amplitude of the battery SOC in a short period of time within the set monitoring window period, captures the change pattern of the SOC during the charging process, analyzes the change rate of the battery's demand for the thermal management system, and evaluates the change in the battery's demand for the thermal management system based on the change trend of the battery temperature; The dynamic change prediction module inputs the analyzed feature data into a pre-trained machine learning model. The machine learning model predicts the dynamic changes of the battery during the charging process based on the input data. The power control optimization module uses the fuzzy logic algorithm to dynamically adjust the power control frequency based on the prediction results of the machine learning model. Specifically, the prediction results are converted into control actions through the fuzzy logic algorithm, and the adjustment step size is set through fuzzy rules. During the set monitoring window, the nonlinear change amplitude of the battery SOC in a short period of time is analyzed to generate a SOC transient change reference value. By capturing the change pattern of the SOC during the charging process and analyzing the change rate of the battery's demand for the thermal management system, a thermal management demand surge reference value is generated. The thermal management demand surge reference value is used to evaluate the change in the battery's demand for the thermal management system. During the set monitoring window, after obtaining the SOC transient change reference value generated by analyzing the nonlinear change amplitude of the battery SOC in a short period of time and the thermal management demand surge reference value generated by analyzing the change rate of the battery's demand for the thermal management system, the SOC transient change reference value and the thermal management demand surge reference value are input into a pre-trained machine learning model to generate a charging dynamic change index, and the dynamic change of the battery during the charging process is predicted by the charging dynamic change index.

2. The charging system based on power regulation according to claim 1, characterized in that: The steps for generating the SOC transient change reference value are as follows: The instantaneous acceleration of SOC during the monitoring window is calculated to capture the nonlinear characteristics of SOC changes. The instantaneous acceleration is expressed by the first-order derivative and second-order derivative of SOC. The specific expression is: ,in, is the instantaneous acceleration of SOC, is the remaining battery power at time t, is the first-order derivative of SOC, that is, the instantaneous rate of change of SOC, is the second-order derivative of SOC, that is, the rate of change of SOC change rate; In order to further capture the high-frequency fluctuation characteristics of SOC changes, wavelet transform is used to perform time-frequency analysis on SOC signals. The SOC signal is decomposed into sub-signals of different frequencies by continuous wavelet transform. The decomposition expression is: ,in, For scale a and location b The wavelet coefficients under is the selected wavelet basis function, a is the scale parameter, which determines the frequency resolution of the analysis, b is the time shift parameter, which determines the time position of the analysis; The instantaneous acceleration is obtained by weighted fusion method. and high frequency fluctuation wavelet coefficients Combined, the final SOC transient change reference value is generated, and the generated expression is: ,in, Indicates the SOC transient change reference value, and is the weight coefficient, reflecting the relative importance of instantaneous acceleration and high-frequency fluctuations, is the absolute value of instantaneous acceleration, quantifying the acceleration of SOC change, is the sum of the energy of high-frequency wavelet coefficients at different scales, reflecting the high-frequency fluctuation intensity of SOC changes. is the energy amplitude of high-frequency fluctuations, which is used to normalize the contribution of high-frequency features.

3. The charging system based on power regulation according to claim 2, characterized in that: In order to extract high-frequency fluctuation characteristics, the high-frequency energy of the wavelet coefficients in the high-frequency range at each time point is calculated. The calculation formula of high-frequency energy is: ,in, is the high frequency energy at time t, is the selected high frequency scale range, The square of the wavelet coefficient representing each high-frequency scale reflects the local energy of the signal at that frequency; The high frequency energy With energy threshold Compare and extract high-frequency fluctuation features, specifically: , then mark the time point as high-frequency fluctuation, retain the high-frequency wavelet coefficient of the time point as a feature, and record the high-frequency wavelet coefficient of the time point as the high-frequency fluctuation wavelet coefficient ; Otherwise, set it to zero; High-frequency fluctuation wavelet coefficient It is the high-frequency wavelet feature after threshold screening.

4. The charging system based on power regulation according to claim 3, characterized in that: The steps for obtaining the thermal management demand sudden increase reference value are as follows: During the monitoring window, the current thermal management demand of the battery is obtained at each time step. The thermal management demand change rate is defined as the rate of change of the thermal management demand between two consecutive time steps. The calculation expression of the thermal management demand change rate is: ,in, It represents the instantaneous demand of the battery on the thermal management system at time t, reflecting the power required for cooling or heating. Indicated in Thermal management needs at all times, is the time step, which indicates the time interval between two data samplings; According to the absolute value of the instantaneous change rate, the nonlinear cumulative change within the monitoring window period is calculated. The calculation expression is: ,in, It is the nonlinear cumulative change of thermal management demand during the monitoring window period, reflecting the dynamic change intensity of thermal management demand during the entire monitoring window period. is the start time of the monitoring window period, T is the total duration of the monitoring window period, is a nonlinear adjustment coefficient used to amplify the impact of drastic changes. is the time decay factor, which is used to reduce the weight of changes in the latter part of the monitoring window period. is the time decay function, which decreases as t increases; Normalize the nonlinear cumulative change with the total demand baseline value during the monitoring window to generate a thermal management demand surge reference value. The generated expression is: ,in, It is the reference value of thermal management demand surge, reflecting the change rate and sudden increase intensity of thermal management demand during the monitoring window period. is the total demand benchmark value during the monitoring window period, defined as: , which represents the total demand level of the thermal management system during the entire window period.

5. The charging system based on power regulation according to claim 4, characterized in that: Based on the prediction results of the machine learning model, the fuzzy logic algorithm is used to dynamically adjust the power control frequency. The specific steps are as follows: Step 1: Define fuzzy variables, set the reference threshold of charging dynamic change index and the starting power control frequency; Specifically, the charging dynamic change index obtained from the machine learning model is used as the input variable of the fuzzy logic algorithm. According to the reference threshold of the charging dynamic change index, the charging dynamic change index is divided into different fuzzy sets. At the same time, a starting power control frequency is determined. The power control frequency is used as the output fuzzy variable and is divided into fuzzy sets. Step 2: Apply fuzzy logic rules to convert the charging dynamic change index into control actions; Specifically, a fuzzy rule base is established to describe the relationship between input variables and output variables, the real-time charging dynamic change index value is fuzzified, the membership of the charging dynamic change index in each fuzzy set is calculated, and then, according to the fuzzy rules, the input membership is inferred to obtain the fuzzy result of the output variable; Step 3: Defuzzify the output, set the adjustment step size and dynamically adjust the power control frequency; Specifically, the output result obtained by fuzzy reasoning is defuzzified and the fuzzy output is converted into a specific numerical value, that is, the adjustment step of the power control frequency. According to this adjustment step, the power control frequency of the charging pile is dynamically updated, so that the power control frequency of the charging pile adjusts the power output more frequently or less frequently to adapt to the real-time demand changes of the battery.

6. A charging pile based on power regulation, comprising a charging system based on power regulation as described in any one of claims 1 to 5.

Images