Intelligent charging pile system and method integrated with real-time battery state detection

Abstract

The application relates to the technical field of electric vehicle battery charging detection, and discloses an intelligent charging pile system integrated with real-time battery state detection and a method thereof. The system comprises a multi-sensor data acquisition module, a data fusion module, a battery health state prediction module, a charging optimization control module, a fault early warning module, a cloud platform and an edge computing collaborative processing module. The method comprises the following steps: obtaining battery data through a multi-sensor to construct a data frame; extracting electrical characteristics to generate a battery state vector; constructing a digital twin model to predict the health state; formulating a dynamic charging power adjustment strategy; comparing the battery health trend with the expected behavior to generate a fault early warning signal; uploading the data and the strategy to the cloud platform to update the control strategy. The application adopts a multi-sensor data fusion technology, combines a twin modeling and a deep learning algorithm, monitors the health state of the battery in real time, comprehensively evaluates the health state of the battery and accurately predicts the fault risk of the battery.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric vehicle battery charging detection, in particular to an intelligent charging pile system integrated with real-time battery state detection and a method thereof. BACKGROUND

[0002] Under the background of the continuous expansion of new energy vehicles and renewable energy applications, as the core energy storage unit, the charging efficiency, safety and service life of power batteries are attracting more and more attention. As an important intermediary connecting the power grid and the battery, the intelligent charging pile not only bears the function of energy transmission, but also gradually develops towards intelligence and informatization. With the rise of big data and artificial intelligence technology, the charging pile system gradually integrates health management, energy scheduling and risk prevention and control functions, promoting the evolution of the intelligent charging system.

[0003] The existing mainstream battery charging management system mainly adopts the architecture based on BMS. The system mainly performs basic protection control by monitoring basic parameters such as voltage, current and temperature. In this traditional system, the task of the charging pile is mainly concentrated in the energy transmission layer, and the perception and understanding of the battery state are still relatively passive. Many key state parameters rely on fixed threshold judgment, or only trigger protection strategies in extreme cases.

[0004] However, the existing charging pile battery state detection technology cannot comprehensively reflect the health status of the battery with single sensor data, and the existing system cannot effectively fuse data from different sensors for comprehensive analysis, resulting in insufficient accuracy of health assessment. The traditional battery management system often adopts a fixed charging strategy, which is difficult to dynamically adjust according to the real-time state of the battery, and is difficult to achieve personalized optimization, thereby wasting energy. Moreover, the existing fault warning mechanism mostly relies on manual or simple rule-based judgment, which is difficult to effectively predict potential battery failures, resulting in some problems being difficult to handle in a timely manner at an early stage. Therefore, the present application provides an intelligent charging pile system integrated with real-time battery state detection and a method thereof to solve the problems in the prior art. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides an intelligent charging pile system integrated with real-time battery state detection and a method thereof, which solves the problems of single battery monitoring, fixed charging strategy and untimely fault warning.

[0006] To achieve the above purpose, the present application realizes the following technical scheme: an intelligent charging pile system integrated with real-time battery state detection, comprising:

[0007] A multi-sensor data acquisition module is used to acquire the voltage, current, temperature and internal resistance parameters of the battery in real time through multi-source sensor data, to monitor the battery state in real time, and to generate raw data streams including various battery parameters.

[0008] a data fusion module for weighted average or Kalman filter fusion of battery parameters from different sensors according to the original data stream to generate a preliminary data stream integrating various battery states;

[0009] a battery health state prediction module for health state prediction of the preliminary data stream integrating battery states by digital twin modeling and machine learning algorithms to generate a health state prediction value of the battery;

[0010] a charging optimization control module for adjusting battery charging current and charging rate by using a nonlinear dynamic charging optimization algorithm and a multi-battery collaborative optimization strategy according to the health state prediction value of the battery and real-time collected battery parameters to generate an optimized charging strategy for the current charging environment;

[0011] a fault early warning module for analyzing and predicting the fault risk of the battery according to the optimized charging strategy for the current charging environment and the health state prediction value of the battery to generate a real-time fault early warning signal;

[0012] a cloud platform and edge computing collaborative processing module, the cloud platform being used for storing battery health data of charging piles, performing data analysis and state evaluation, and feeding back to the edge execution unit through cloud algorithms to process real-time data of the battery to generate immediate decisions.

[0013] Preferably, the multi-sensor data acquisition module comprises:

[0014] an acquisition control unit for initializing multi-source sensor task scheduling according to the working state of the charging pile, setting a data sampling period and a start trigger condition, the multi-source sensors including a voltage sensor, a current sensor, a temperature sensor, and a resistance sensor;

[0015] a data caching unit for dynamically creating a time series caching structure of battery acquisition data according to the multi-source sensor task scheduling;

[0016] a data identification unit for adding device number and time stamp information to each frame of data after caching is completed.

[0017] Preferably, the data fusion module comprises:

[0018] a data preprocessing unit for using the device number and time stamp information as an index to perform abnormality elimination and format standardization on data from multiple channels at the same time as a fusion initial set;

[0019] a feature extraction unit for extracting battery state evolution feature values including charging and discharging slope, temperature fluctuation rate, and impedance change trend from the fusion initial set.

[0020] A fusion generation unit is configured to perform a feature-level fusion operation after the battery state evolution feature value is extracted, to call a specified fusion rule based on a battery model and a working condition, and to generate a state description vector in a unified format.

[0021] Preferably, the battery health state prediction module comprises:

[0022] A twin modeling unit is configured to construct a battery virtual model according to the state description vector in the unified format and a corresponding historical running track, and to perform initial parameter configuration through the battery virtual model.

[0023] A training learning unit is configured to train a prediction model through a continuous input state vector, to learn an evolution mode of the battery at different aging stages, and to form a state mapping relationship adapted to a charging behavior.

[0024] A health prediction unit is configured to infer a current battery state vector through a trained model, to generate a current battery health state value SOH, and to label the current battery health state value SOH with a time label.

[0025] Preferably, the initial parameter configuration through the battery virtual model comprises:

[0026] The battery virtual model is constructed by combining historical data of the battery state with real-time collected data.

[0027] Based on a health degradation mode in the historical data, the battery virtual model fits a degradation feature of the battery in a modeling process, performs grouping processing in combination with a difference between different battery models, and optimizes modeling parameters.

[0028] The battery virtual model is updated according to real-time collected parameters of the battery voltage, current, temperature and internal resistance, and model parameters are dynamically adjusted.

[0029] Preferably, the charging optimization control module comprises:

[0030] A strategy generation unit is configured to dynamically construct a charging power scheduling scheme in combination with a current charging pile power supply capacity, a battery temperature state and a voltage window based on the battery health state value SOH.

[0031] A real-time regulation unit is configured to issue current and voltage set values to a charging interface module after the charging power scheduling scheme is generated, and to collect and correct and adjust a battery feedback in real time.

[0032] A collaborative optimization unit is configured to dynamically allocate charging priorities and resource proportions of the batteries by comparing SOH differences and power demands of the batteries, and to form a group charging strategy based on health degree sensing.

[0033] Preferably, the construction of the charging power scheduling scheme in combination with the current charging pile power supply capacity, battery temperature state and voltage window includes:

[0034] According to the SOH value output by the battery health state prediction module, the charging strategy of each battery is prioritized, and a conservative charging mode is generated for the battery with lower SOH;

[0035] By combining with the battery temperature and voltage range, the charging current and voltage are dynamically adjusted according to the thermal runaway risk of the battery;

[0036] Based on the health difference and charging demand between the batteries, a multi-objective optimization algorithm is used to consider the charging speed and safety at the same time, and balance is made between the battery health state, power allocation and time efficiency.

[0037] Preferably, the fault early warning module includes:

[0038] The abnormality identification unit is configured to compare the group charging strategy perceived by the health degree with the SOH state, analyze the response deviation degree of the battery to the normal control instruction, and identify the nonlinear behavior or response delay;

[0039] The risk modeling unit is configured to analyze the response deviation degree, refer to the historical failure case library for matching analysis, and construct a fault probability curve under the current state;

[0040] The early warning control unit is configured to set a warning threshold according to the fault probability curve under the current state, in combination with the current charging strategy level, and generate an instruction set including the warning level and the recommended control behavior.

[0041] Preferably, the cloud platform and edge computing cooperative processing module includes:

[0042] The cloud-side data management unit is configured to construct a time series data warehouse according to the health state prediction value, and the time series data warehouse supports horizontal battery comparison analysis;

[0043] The model updating unit is configured to perform data analysis and state evaluation on the data of the time series data warehouse, generate an optimization parameter package and model structure adjustment information through cloud training, and push them to the edge execution unit;

[0044] The edge execution unit is configured to receive the optimization parameter package and model structure adjustment information trained by the cloud, combine the real-time collected state data and the instruction set of the recommended control behavior to generate the final charging control parameter.

[0045] An intelligent charging pile method integrating real-time battery state detection is also provided, including the following steps:

[0046] The voltage, current, temperature and internal resistance data of the battery are acquired by multiple types of sensors respectively, and a time stamp and a device number are added to the acquired battery data to construct a multi-source heterogeneous state data frame;

[0047] Based on the constructed multi-source heterogeneous state data frame, the acquired battery data is aligned through a time synchronization mechanism, and then through an outlier elimination and feature extraction process, the electrical feature parameters closely related to battery aging are extracted, and a unified structure battery state description vector is generated by using a fusion algorithm;

[0048] According to the battery state description vector, a digital twin model is constructed, and historical operation data matching the battery type are selected for initialization, and a machine learning model is used to predict and infer the input data, and output the health state value and change trend of the current battery;

[0049] Based on the health state value of the current battery, combined with the real-time operation parameters of the current charging power capacity, battery position temperature and voltage window range, a dynamic charging power adjustment strategy is constructed, and the current and voltage set values of each charging port are generated in real time;

[0050] The dynamic charging power adjustment strategy is compared with the health change trend of the current battery, the deviation between the battery response and the expected behavior is identified, the warning level is generated through fault risk curve analysis, and the fault warning signal is sent to the high-risk single body;

[0051] The acquired multi-source heterogeneous state data frame, battery state description vector, battery health prediction value, dynamic charging power adjustment strategy and warning level are uploaded to the cloud platform in a structured manner, and the algorithm model is updated according to the cloud feedback, and the local control strategy replacement and charging behavior reconfiguration are completed.

[0052] The application provides an intelligent charging pile system and method integrated with real-time battery state detection.

[0053] 1、The application adopts multi-sensor data fusion technology, and combines twin modeling and deep learning algorithm to monitor the health state of the battery in real time. By fusing real-time data and historical data from different sensors, the health condition of the battery can be comprehensively evaluated, and the fault risk of the battery can be accurately predicted. Compared with the single sensor monitoring method in the prior art, the application can more comprehensively and real-timely acquire various parameters of the battery, and predict the future health trend of the battery through more intelligent data analysis, solving the problem that the traditional technology cannot comprehensively predict the health state of the battery.

[0054] 2、The application combines cloud computing and edge computing through the cloud-side data management and model updating module, ensuring the continuous updating of the battery state of health prediction model. The real-time collected charging pile data is stored and analyzed through the cloud platform, promoting the continuous optimization of the charging strategy. This innovation enables the battery charging strategy to be dynamically adjusted, providing real-time optimization solutions based on the actual status of the battery. Compared to the application of fixed models and static strategies in traditional technologies, the application improves the intelligent level of the charging pile system, making the charging process more flexible and efficient, and significantly reducing the faults caused by mismatched battery states.

[0055] 3、The application introduces a fault warning module, combining the state of health (SOH) of the battery and real-time monitoring data to timely identify potential fault risks of the battery. By predicting the occurrence of battery faults through intelligent algorithms, warnings are issued in advance to avoid safety problems caused by overcharging or battery damage. Unlike traditional technologies that rely on manual inspection or only judge battery health based on a single parameter, the application can provide real-time and accurate fault warnings, addressing the shortcomings of existing charging systems in terms of timely monitoring and warning of battery safety risks. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 The system architecture diagram of the application;

[0057] Figure 2 The method step flowchart of the application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the application will be described in detail below with reference to the accompanying drawings of the application. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the application.

[0059] Please refer to the accompanying drawings of the application Figure 1 The embodiments of the application provide an intelligent charging pile system integrated with real-time battery state detection, including:

[0060] A multi-sensor data acquisition module is used to acquire the voltage, current, temperature and internal resistance parameters of the battery in real time through multi-source sensor data, to monitor the battery state in real time and generate raw data streams including various battery parameters;

[0061] A data fusion module is used to perform weighted average or Kalman filter fusion on the battery parameters from different sensors based on the raw data streams, to generate a preliminary data stream that integrates various battery states;

[0062] a battery health state prediction module for predicting the health state of the battery by digital twin modeling and machine learning algorithms based on the preliminary data stream of the comprehensive battery state to generate a health state prediction value of the battery;

[0063] a charging optimization control module for adjusting the battery charging current and charging rate by using a nonlinear dynamic charging optimization algorithm and a multi-battery collaborative optimization strategy based on the health state prediction value of the battery and the real-time collected battery parameters to generate an optimized charging strategy for the current charging environment;

[0064] a fault early warning module for analyzing and predicting the fault risk of the battery based on the optimized charging strategy for the current charging environment and the health state prediction value of the battery to generate a real-time fault early warning signal;

[0065] a cloud platform and edge computing collaborative processing module, wherein the cloud platform is configured to store the battery health data of the charging pile, perform data analysis and state evaluation, and feed back to the edge execution unit through a cloud algorithm to process the real-time data of the battery to generate an immediate decision.

[0066] For the multi-sensor data collection module, in the embodiment, to realize comprehensive, accurate and dynamic collection of the battery state, the system first relies on the multi-sensor data collection module to collect the real-time data of the key parameters of the battery. The module serves as the basic support module of the entire system, and the high-precision data collected by the module provides necessary data input for the subsequent fusion analysis, health prediction and charging optimization function modules. In order to ensure the accuracy and adaptability of the data under different battery models and different use environments, the module also designs a multi-task scheduling and time synchronization mechanism to dynamically adjust the sampling period and trigger logic of the sensor.

[0067] In the embodiment, the multi-sensor data collection module includes a collection control unit, a data buffer unit and a data identification unit.

[0068] Specifically, the collection control unit is configured to initialize the collection task scheduling mechanism of the multi-source sensor according to the current working state of the charging pile. Generally, the control unit can set the trigger condition and the sampling period parameter, so that the sensor task starts data collection according to the charging interface power-on state or the battery connection state. As an option, the module supports a combination of voltage sensors, current sensors, temperature sensors and internal resistance sensors, which is suitable for various battery types such as lithium iron phosphate, ternary lithium and lead-acid batteries.

[0069] In one possible implementation, the system periodically samples the battery terminal voltage by using the following voltage sampling formula:

[0070] ;

[0071] wherein, represents the average voltage value of the i-th battery cell at the current time; represents the instantaneous voltage value at the i-th sampling time; represents the instantaneous voltage value at the i-th sampling time; is the window length; is the sampling time interval.

[0072] The above average voltage not only suppresses the error caused by instantaneous jitter, but also facilitates subsequent Kalman filtering or Bayesian fusion processing. As a further extension, the unit can also automatically adjust the sampling density according to the rate of change of the sampling data, forming an adaptive data acquisition strategy and improving the response speed of critical state changes.

[0073] The data buffering unit is used to dynamically establish a time series buffer structure for data according to the sampling instructions issued by the acquisition control unit. In some embodiments, the system uses a ring buffer structure to store voltage, current, temperature and internal resistance sampling values in a period of time, ensuring that data is not lost in the case of network or upper layer processing module delay, and ensuring the input integrity of synchronous analysis.

[0074] Further, to cooperate with the time correlation feature extraction in the subsequent state fusion and health prediction model, a four-dimensional matrix is established in the data buffer structure as follows:

[0075] ;

[0076] wherein, represents a certain sampling point data structure; is the data parameter type index (such as voltage, current, etc.); is the time stamp index; represents the sensor number; represents the sampling channel number.

[0077] The data identification unit is used to attach a unique device code and high-precision time stamp information to each frame of sampling data after data buffering. The time synchronization mechanism used can be GPS time synchronization, so as to realize time alignment and historical comparison analysis between edge nodes and cloud platforms. This unit can also support device identification redundancy mechanism for simultaneous sampling across multiple ports, avoiding multi-battery data crosstalk during charging.

[0078] In some embodiments, to improve the synchronization accuracy of multi-sensor asynchronous data streams, a method based on reference voltage pulse calibration is introduced. By periodically triggering a calibration pulse, the delay offset of each acquisition channel is measured and compensated, thereby achieving microsecond-level time alignment accuracy.

[0079] ​​In addition, the module also supports dynamic data compression encoding mechanism, such as using differential encoding (DPCM) or space-time compression technology to reduce the upload bandwidth load, and compresses the redundant data while maintaining the effective features without loss.

[0080] For the data fusion module, in this embodiment, the data obtained from each sensor is integrated, and necessary processing and analysis are performed to ensure accurate evaluation of the battery state. The data fusion processing module generates evaluation data of the battery health state by pre-processing, feature extraction, fusion calculation and other operations on the collected data, providing high-quality input data for charge optimization and battery life prediction. Specifically, the data fusion processing module includes a data preprocessing unit, a feature extraction unit and a data fusion unit.

[0081] In some embodiments, the function of the data preprocessing unit is to remove noise and outliers from the sensor data stream to ensure the accuracy of subsequent processing. This process usually includes data denoising, outlier detection and correction, and data standardization. The pre-processing formula is:

[0082] ;

[0083] In the formula, represents the pre-processed data; is the original data collected by the sensor; is the mean value of the data; is the standard deviation of the data. Through standardization processing, the dimension difference of different data sources can be eliminated, making the subsequent data fusion more effective.

[0084] The feature extraction unit is used to extract feature values closely related to the battery health state from the original data stream. In some embodiments, the goal of feature extraction is to identify the core operating parameters of the battery, such as the internal resistance, charge and discharge efficiency, and temperature change of the battery. The feature extraction process can be based on statistical methods, signal processing methods or machine learning algorithms.

[0085] For example, in terms of internal resistance extraction, Kalman filtering or wavelet transform technology can be used to accurately extract the internal resistance characteristics of the battery. The internal resistance extraction formula is:

[0086] ;

[0087] In the formula, is the open-circuit voltage of the battery; is the voltage under load; is the current value; is the internal resistance of the battery. This formula can effectively evaluate the health status of the battery and provide basic data for subsequent charge optimization and fault warning.

[0088] In this embodiment, the data fusion unit is responsible for integrating data from different sensors to generate a comprehensive evaluation value of the battery health status. Data fusion usually adopts techniques such as weighted average method, Kalman filter or particle filter. The formula for fusion calculation is:

[0089] ;

[0090] wherein, is the fused data; represents the processing data provided by the th sensor; is the weight coefficient of the corresponding sensor; is the number of sensors. The weight coefficient can be dynamically adjusted according to the accuracy and reliability of the sensor, so that the fused data is more representative and accurate.

[0091] For the battery health status prediction module, in this embodiment, by combining historical battery data and real-time data, based on twin modeling technology and deep learning algorithm, a virtual model of the battery is created, and the health status and SOH (state of health) of the battery are accurately predicted through training and learning. This module can realize the prediction and monitoring of the battery health status by introducing the historical data and real-time data fusion of the battery state, and then optimize the charging strategy, improve the use efficiency and safety of the battery, mainly composed of twin modeling unit, training and learning unit and battery state prediction unit.

[0092] In some embodiments, the twin modeling unit simulates the running state of the battery by establishing a virtual battery model, and updates the model according to historical data and real-time data. Specifically, the construction of the virtual battery model is based on the health degradation mode in the historical data, and considers different working conditions of the battery. The process of twin modeling is to establish a mathematical model to simulate the dynamic changes of key parameters such as charge and discharge process, temperature change, current and voltage of the battery.

[0093] The optimization formula of the twin model is as follows:

[0094] ;

[0095] wherein, represents the optimization of model parameters; is the actual battery state data of the th sample; is the input feature (such as voltage, current, temperature, etc.) of the th sample; is the prediction function of the virtual battery model; is the number of samples. By minimizing the error between the predicted value and the actual value, the model can continuously optimize its parameters and improve the prediction ability of the battery state of health.

[0096] The training learning unit trains the deep learning model based on the input battery historical data and real-time collected data. In specific implementation, deep neural network (DNN) or convolutional neural network (CNN) algorithm is used to predict the state of health of the battery. The goal of training is to establish the correlation between the battery state and the health degradation through multi-level feature learning, and then realize the real-time evaluation of the battery life and SOH.

[0097] The loss function of training can be expressed as:

[0098] ;

[0099] In the formula, is the loss function; represents the difference between the predicted SOH value and the actual SOH value ; is the number of training samples; is the training parameter. By minimizing the loss function, an accurate battery state of health prediction model can be effectively trained.

[0100] The battery state prediction unit performs real-time battery state of health prediction based on the trained model and outputs the SOH value. The calculation formula of SOH value is:

[0101] ;

[0102] In the formula, is the current actual available capacity of the battery; is the nominal capacity of the battery; represents the percentage of the state of health of the battery. Through real-time monitoring of SOH, the system can timely identify whether the battery is aging or damaged, thereby providing effective charging optimization and safety warning.

[0103] For the charging optimization control module, in this embodiment, the charging process is dynamically adjusted according to the state of health (SOH) of the battery and real-time monitoring data to ensure the safety and efficiency of the battery during the charging process. Through optimization of the power regulation scheme of battery charging and real-time monitoring of battery temperature and voltage, the charging optimization control module can accurately adjust the charging strategy of the battery to avoid battery damage caused by too fast, too slow or too high temperature charging. This module optimizes the charging rate through algorithm, and comprehensively considers the SOH and charging demand of multiple batteries to ensure that each battery is charged in the optimal state, mainly including a strategy generation unit, a charging control unit and a power regulation unit.

[0104] The strategy generation unit is responsible for calculating and generating the optimal charging strategy based on the state of health (SOH) of the battery and current battery charging capacity, temperature, and other data. Specifically, this unit determines the power adjustment scheme and temperature control strategy based on the SOH value and battery temperature information. The generation formula for the charging strategy is as follows:

[0105] ;

[0106] wherein, represents the charging power; is the state of health of the battery; is the temperature of the battery; is the voltage of the battery; represents the charging power strategy function. This formula shows that the charging power is a function of the battery state, temperature, and voltage. Through this formula, the system can dynamically adjust the charging power to keep the battery in the best state during the charging process.

[0107] The charging control unit implements the charging process control of the battery according to the charging strategy generated by the strategy generation unit. In actual operation, this unit optimizes the charging rate by adjusting the charging current and voltage to avoid overcharging, overheating, and other situations. In some embodiments, the charging control unit adjusts the charging current in real time based on real-time monitoring of battery temperature and voltage data.

[0108] The adjustment formula for the charging current is as follows:

[0109] ;

[0110] wherein, is the charging current; is the current battery voltage; is the target battery voltage; is the resistance of the charging path. Through this formula, the charging control unit can adjust the charging current in real time based on the actual voltage of the battery to ensure the stability of the charging process.

[0111] The power adjustment unit is responsible for adjusting the charging power based on real-time data during the charging process to ensure that the battery is always in the best charging state. In some cases, the power adjustment unit can perform dynamic power distribution based on the SOH and temperature of the battery to avoid damage caused by excessive charging or high battery temperature.

[0112] The optimization formula for power adjustment is as follows:

[0113] ;

[0114] wherein, is the optimized charging power; represents the charging power; This is the maximum safe temperature for the battery. This represents the current battery temperature. This formula indicates that when the battery temperature approaches its maximum safe temperature, the system will automatically reduce the charging power to prevent the battery from being damaged by overheating.

[0115] For the fault warning module, in this embodiment, by identifying the battery's State of Health (SOH) and combining it with real-time battery monitoring data, potential fault risks are identified in a timely manner, and the occurrence of battery faults is predicted through intelligent algorithms. Through anomaly identification and fault prediction algorithms, various deviations during the battery charging process are analyzed to predict potential battery faults and issue early warnings to avoid battery damage and extend battery life. This module mainly includes an anomaly identification unit and a risk analysis unit, which combine SOH prediction and data deviations during the charging process to provide fault warnings.

[0116] The anomaly detection unit determines whether there are any abnormal deviations during battery charging based on battery health status and real-time data. These deviations may include abnormal fluctuations in parameters such as battery voltage, temperature, or internal resistance. By comparing and analyzing monitoring data during battery charging, the anomaly detection unit can identify potential battery faults and issue early warnings.

[0117] The formula for detecting abnormal deviations is:

[0118] ;

[0119] In the formula, This indicates the difference between the battery's current voltage and its nominal voltage. Current battery voltage; This is the battery's nominal voltage. If the voltage deviation exceeds a preset threshold, the battery is considered to be at risk of failure.

[0120] The risk analysis unit is used to further analyze the correlation between battery health status and various data deviations during the charging process, and to assess the probability of battery failure. Based on historical and real-time monitoring data, the risk analysis calculates the likelihood of battery failure and proposes reasonable early warning measures accordingly. By comparing the state of health (SOH) of multiple batteries, the system can identify potential faults within a battery group and adjust the charging strategy accordingly.

[0121] The risk assessment formula is as follows:

[0122] ;

[0123] In the formula, This is the risk assessment value for battery failure; For the first Voltage deviation of individual batteries; For the first the failure probability of one battery; the nominal voltage of the battery; the number of batteries involved in the assessment. According to this formula, the risk analysis unit can comprehensively assess the health status of all batteries and generate a system-level failure risk value.

[0124] For the cloud platform and edge computing collaborative processing module, in this embodiment, real-time data is stored, processed and analyzed through the cloud computing platform to support dynamic updating of the intelligent charging pile health status prediction model. By combining real-time data and historical data, the battery health status prediction model is updated, the intelligentization and optimization degree of the charging process are improved, mainly including a cloud-side data management unit, a model updating unit and an edge execution unit. Through data analysis and prediction of the battery health status, the module can dynamically adjust the battery charging strategy, and continuously optimize the model through machine learning algorithms to achieve more accurate charging control.

[0125] The main function of the cloud-side data management unit is to store and manage the historical data and real-time monitoring data of the charging pile through the cloud computing platform. These data are not only used to support the updating of the battery health status prediction model, but also used for subsequent data analysis and model training. The storage and processing of data are managed through an efficient data architecture to ensure real-time updating and effective utilization of large-scale data.

[0126] The formula for data storage and management is as follows:

[0127] ;

[0128] In the formula, total data stored in the cloud; real-time data of the th charging pile; the number of charging piles. This formula shows that through the cloud-side data management unit, the data of all charging piles can be uploaded in time and aggregated to the cloud for storage and management.

[0129] The model updating unit is responsible for continuously updating the battery health status prediction model. By analyzing the stored historical data and real-time data, the model updating unit can continuously optimize the health status of the battery based on machine learning algorithms. This process includes adjusting the parameters of the existing model and retraining the model according to new data to improve the accuracy of prediction.

[0130] The edge execution unit is responsible for applying the cloud-updated model to the actual charging pile system. According to the results of model updating, the edge execution unit can make real-time decisions locally and dynamically adjust the charging strategy. Through this unit, the optimization strategy of the cloud can be quickly passed to each charging pile, realizing real-time optimization of the intelligent charging process.

[0131] The intelligent charging pile method integrating real-time battery state detection described below can be mutually corresponding with reference to the intelligent charging pile system integrating real-time battery state detection described above.

[0132] Please refer to the accompanying Figure 2 The application also provides an intelligent charging pile method integrating real-time battery state detection, comprising the following steps:

[0133] The voltage, current, temperature and internal resistance data of the battery are obtained by multiple types of sensors respectively, and a time stamp and a device number are added to the obtained battery data to construct a multi-source heterogeneous state data frame;

[0134] Based on the constructed multi-source heterogeneous state data frame, the obtained battery data is aligned through a time synchronization mechanism, and then through an outlier elimination and feature extraction process, the electrical feature parameters closely related to battery aging are extracted, and a unified structure battery state description vector is generated using a fusion algorithm;

[0135] A digital twin model is constructed according to the battery state description vector, and historical operation data matching the battery type are selected for initialization, and a machine learning model is used to predict and infer the input data, outputting the health state value and change trend of the current battery;

[0136] Based on the health state value of the current battery, combined with the real-time operating parameters of the current charging power capacity, battery location temperature and voltage window range, a dynamic charging power adjustment strategy is constructed, and the current and voltage set values of each charging port are generated in real time;

[0137] The dynamic charging power adjustment strategy is dynamically compared with the health change trend of the current battery to identify the deviation between the battery response and the expected behavior, generate a warning level through fault risk curve analysis, and issue a fault warning signal for high-risk single cells;

[0138] The obtained multi-source heterogeneous state data frame, battery state description vector, battery health prediction value, dynamic charging power adjustment strategy and warning level are uploaded to the cloud platform in a structured manner, and according to the algorithm model updated by the cloud feedback, the local control strategy replacement and charging behavior reconfiguration are completed.

[0139] The method of the present embodiment can be used to execute the above-mentioned system embodiment, and the principles and technical effects are similar, which will not be repeated here.

[0140] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. An intelligent charging pile system integrating real-time battery status detection, characterized in that, include: The multi-sensor data acquisition module is used to collect battery voltage, current, temperature and internal resistance parameters in real time through multi-source sensor data, monitor the battery status in real time, and generate raw data streams including various battery parameters. The data fusion module is used to perform weighted averaging or Kalman filtering fusion of battery parameters from different sensors based on the raw data stream to generate a preliminary data stream that integrates various battery states. The battery health status prediction module is used to predict the health status of the preliminary data stream of the comprehensive battery status through digital twin modeling and machine learning algorithms, and generate the predicted health status value of the battery. The charging optimization control module is used to adjust the battery charging current and charging rate based on the predicted health status of the battery and the real-time collected battery parameters, using a nonlinear dynamic charging optimization algorithm and a multi-battery collaborative optimization strategy, to generate an optimized charging strategy for the current charging environment. The fault warning module is used to analyze and predict the battery's fault risk based on the optimized charging strategy of the current charging environment and the predicted health status of the battery, and generate a real-time fault warning signal. The cloud platform and edge computing collaborative processing module, wherein the cloud platform is used to store the battery health data of the charging pile, perform data analysis and status assessment, and feed back to the edge execution unit through cloud algorithms to process the real-time data of the battery and generate instant decisions; The operation method of the intelligent charging pile system integrating real-time battery status detection includes the following steps: The battery's voltage, current, temperature, and internal resistance data are acquired by multiple types of sensors, and timestamps and device numbers are added to the acquired battery data to construct a multi-source heterogeneous state data frame. Based on the constructed multi-source heterogeneous state data frame, the acquired battery data is aligned through a time synchronization mechanism. Then, through outlier removal and feature extraction processes, electrical feature parameters closely related to battery aging are extracted, and a unified battery state description vector is generated using a fusion algorithm. A digital twin model is constructed based on the battery state description vector, and historical operating data matching the battery type is selected for initialization. A machine learning model is used to predict and infer the input data, and the current health status value and trend of the battery are output. Based on the current battery health status value, combined with the real-time operating parameters of the current charging power capacity, battery location temperature and voltage window range, a dynamic charging power adjustment strategy is constructed, and the current and voltage set values ​​of each charging port are generated in real time. By dynamically comparing the dynamic charging power adjustment strategy with the current health trend of the battery, the deviation between the battery response and the expected behavior is identified. The warning level is generated through fault risk curve analysis, and a fault warning signal is issued for high-risk cells. The acquired multi-source heterogeneous state data frames, battery state description vectors, battery health prediction values, dynamic charging power adjustment strategies, and warning levels are uploaded to the cloud platform in a structured manner. Based on the updated algorithm model fed back from the cloud, the local control strategy is replaced and the charging behavior is reconfigured.

2. The intelligent charging pile system with integrated real-time battery status detection according to claim 1, characterized in that, The multi-sensor data acquisition module includes: The data acquisition and control unit is used to initialize the multi-source sensor task scheduling according to the working status of the charging pile, and set the data sampling period and start-up trigger conditions. The multi-source sensors include voltage sensors, current sensors, temperature sensors and internal resistance sensors. The data caching unit is used to dynamically create a time-series cache structure for battery acquisition data according to the multi-source sensor task scheduling. The data identification unit is used to add device number and timestamp information to each frame of data after caching is completed.

3. The intelligent charging pile system with integrated real-time battery status detection according to claim 2, characterized in that, The data fusion module includes: The data preprocessing unit is used to perform anomaly removal and format standardization on data from multiple channels at the same time, using the device number and timestamp information as an index, as an initial set for fusion. The feature extraction unit is used to extract battery state evolution feature values ​​from the fused initial set, the evolution feature values ​​including charge / discharge slope, temperature fluctuation rate and impedance change trend; The fusion generation unit is used to perform feature-level fusion operation after the battery state evolution feature values ​​are extracted, and to generate a state description vector in a unified format by calling a specified fusion rule based on the battery model and operating conditions.

4. The intelligent charging pile system with integrated real-time battery status detection according to claim 3, characterized in that, The battery health status prediction module includes: The twin modeling unit is used to construct a battery virtual model based on the unified format state description vector and the corresponding historical running trajectory, and to configure the initial parameters through the battery virtual model. The training learning unit is used to train the prediction model by continuously inputting state vectors, learn the evolution pattern of the battery at different aging stages, and form a state mapping relationship adapted to charging behavior. The health prediction unit is used to infer the current battery state vector by calling the trained model, generate the current battery health state value (SOH), and add a time label.

5. The intelligent charging pile system with integrated real-time battery status detection according to claim 4, characterized in that, The initial parameter configuration via the battery virtual model includes: By combining historical battery status data with real-time collected data, a virtual battery model is constructed. Based on health degradation patterns in historical data, the battery virtual model fits the degradation characteristics of the battery during the modeling process, and groups the different battery models to optimize the modeling parameters. The battery virtual model is updated based on real-time collected battery voltage, current, temperature, and internal resistance parameters, and the model parameters are dynamically adjusted.

6. The intelligent charging pile system with integrated real-time battery status detection according to claim 4, characterized in that, The charging optimization control module includes: The strategy generation unit is used to dynamically construct a charging power scheduling scheme based on the battery health state value (SOH) and in combination with the current charging pile power supply capacity, battery temperature state and voltage window. The real-time control unit is used to send current and voltage setpoints to the charging interface module after the charging power scheduling scheme is generated, and to collect battery feedback in real time for correction and adjustment. The collaborative optimization unit is used to dynamically allocate the charging priority and resource ratio of each battery by comparing the SOH difference of multiple batteries with the power demand, so as to form a group charging strategy based on health perception.

7. The intelligent charging pile system with integrated real-time battery status detection according to claim 6, characterized in that, The dynamic charging power scheduling scheme, which combines the current charging pile power capacity, battery temperature status, and voltage window, includes: Based on the SOH value output by the battery health status prediction module, the charging strategy for each battery is prioritized. For batteries with low SOH, a conservative charging mode is generated. By combining the battery temperature and voltage range, the charging current and voltage are dynamically adjusted according to the risk of thermal runaway of the battery; Based on the health differences between batteries and charging requirements, a multi-objective optimization algorithm is used to simultaneously consider charging speed and safety, balancing battery health, power distribution, and time efficiency.

8. The intelligent charging pile system with integrated real-time battery status detection according to claim 6, characterized in that, The fault early warning module includes: An anomaly identification unit is used to compare the health-sensing group charging strategy with the SOH state, analyze the degree of deviation in the battery's response to normal control commands, and identify nonlinear behavior or response delay. The risk modeling unit is used to construct a failure probability curve for the current state by matching and analyzing the degree of response deviation and referring to the historical failure case library. The early warning control unit is used to set an early warning threshold based on the fault probability curve under the current state and in conjunction with the current charging strategy level, and to generate an instruction set including the early warning level and suggested control behavior.

9. The intelligent charging pile system with integrated real-time battery status detection according to claim 1, characterized in that, The cloud platform and edge computing collaborative processing module includes: The cloud-based data management unit is used to construct a time-series data warehouse based on the predicted health status values. The time-series data warehouse supports horizontal battery comparison analysis. The model update unit is used to perform data analysis and status evaluation on the data in the time series data warehouse, generate an optimization parameter package and model structure adjustment information after cloud training, and push them to the edge execution unit. The edge execution unit is used to receive the optimized parameter package and model structure adjustment information after training in the cloud, and generate the final charging control parameters by combining the real-time collected state data and the instruction set of suggested control behavior.

Images