SOH (state of health) evaluation system and method applied to lithium battery of electric vehicle

Abstract

The invention discloses a health state assessment system and method applied to an electric vehicle lithium battery, and relates to the technical field of battery health assessment, and the system comprises a battery data collection module, a battery data preprocessing module, a sub-health degree analysis module, a battery health early warning module and a health cycle prediction module. The method comprises the following steps: acquiring a battery data set through a sensor group and a battery management system, preprocessing battery data to obtain a feature data set, analyzing the feature data set, calculating the sub-health degree of a battery, triggering a battery health decline early warning according to the decline trend of the sub-health degree, and calculating the number of battery health cycles. According to the method, resource waste caused by premature replacement is avoided through a dynamic early warning mechanism, the problem of lag caused by maintenance is also prevented, closed-loop management of the lithium battery from state monitoring to active maintenance is realized, the service life of the battery is effectively prolonged, and the running safety of the electric vehicle is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery health assessment, and in particular to a health status assessment system and method for lithium batteries of electric vehicles. Background Art

[0002] Amid the rapid development of the new energy vehicle industry, lithium-ion batteries, as a core power source, have a significant impact on vehicle safety and endurance. However, existing assessment technologies suffer from significant deficiencies in data processing, anomaly detection, and status prediction, making them incapable of meeting the high-precision assessment requirements under complex operating conditions. Traditional lithium-ion battery management systems, when collecting sensor data such as voltage, current, temperature, and internal resistance, are susceptible to electromagnetic interference and ambient temperature fluctuations during vehicle operation, leading to high-frequency noise and occasional outliers, which directly impact assessment accuracy. This is particularly true during the initial battery life or system startup phase, when insufficient data collection cycles and early data points lack sufficient historical information. Conventional denoising methods, due to insufficient data volume, can lead to analytical bias and fail to accurately capture true trends in battery performance. Furthermore, independent processing of multi-sensor data can lead to time series desynchronization, failing to fully reflect the correlations between multi-dimensional battery parameters, creating data-level risks for health assessment. Existing outlier detection methods often rely on preset fixed thresholds, making them incapable of adapting to nonlinear anomalies caused by aging and sudden operating conditions during the lithium-ion battery charging and discharging process. For example, in dynamic anomaly scenarios such as sudden voltage changes and abnormally high temperatures, fixed thresholds cannot adaptively adjust to data distribution characteristics, easily leading to missed or misidentified anomalies and, in turn, distorting the assessment model. Furthermore, traditional methods lack a coordinated mechanism for handling anomalies in multi-dimensional data. Independently monitoring each parameter may overlook correlations between different features, making it impossible to identify comprehensive anomaly patterns reflecting battery health degradation at a system level, reducing the reliability of assessment results. Early assessment technologies often relied on a single parameter (such as voltage or cycle count) to determine battery status, failing to fully capture the multiple aspects of battery performance degradation (such as internal resistance changes and temperature consistency), resulting in one-sided assessment results. Furthermore, traditional methods focus on static analysis of the current status and lack dynamic tracking of health degradation trends, making them unable to promptly capture accelerated declines in battery health. In practical applications, excessive battery wear due to delayed maintenance decisions or resource waste due to overreliance on a single indicator leads to premature replacement. The lack of precise guidance on maintenance time windows hinders the formation of a closed-loop monitoring, assessment, and maintenance system for battery lifecycle management, hindering the improvement of electric vehicle performance. Summary of the Invention

[0003] The purpose of the present invention is to provide a health status assessment system and method for lithium batteries of electric vehicles to solve the problems raised in the above background technology.

[0004] In order to solve the above technical problems, the present invention provides the following technical solution: a health status assessment method for electric vehicle lithium batteries, comprising the following steps: S1, collecting battery data sets through sensor groups and battery management systems; S2. Preprocess the battery data to obtain a feature data set; S3. Analyze the characteristic data set and calculate the sub-health level of the battery; S4. Trigger a battery health degradation warning based on the decreasing trend of the sub-health level; S5. Calculate the number of battery health cycles and prompt the user to set a time limit for battery replacement.

[0005] Furthermore, in step S1, the sensor group includes a voltage sensor, a current sensor, a temperature sensor and an internal resistance sensor, and sets a data collection time interval a. In any collection cycle, the voltage A is collected in real time by the voltage sensor. t , real-time voltage B is collected through the current sensor t , collect real-time temperature C through temperature sensor t , collect real-time internal resistance D through internal resistance sensor t , collect the cycle number E through the battery management system t , the battery is charged and discharged once, which is recorded as one charge and discharge cycle. The number of cycles refers to the number of charge and discharge cycles of the lithium battery, and then the voltage data set {A1, A2, ..., A t ,…,A T}、Current data set {B1,B2,…,B t ,…,B T}、Temperature data set {C1,C2,…,C t ,…,C T}、Internal resistance data set {D1,D2,…,D t ,…,D T} and the cycle number data set {E1,E2,…,E t ,…,E T}, where T represents the number of times data is collected, and t represents the tth time data is collected; in data processing, sliding filtering and borrowing of data from previous cycles are used to effectively filter out noise and solve the problem of insufficient initial data, making the voltage, current and other data collected by multiple sensors more stable and reliable, laying a solid data foundation for evaluation. When identifying anomalies, a reasonable range is dynamically determined based on the data's own distribution, and outliers are replaced with the median to accurately capture nonlinear problems such as voltage mutations and temperature anomalies, reducing false positives and missed judgments, and making the evaluation results closer to the actual state of the battery. The evaluation system integrates multi-dimensional indicators, observing both the current state and analyzing trend changes. It can provide a specific maintenance time window when the battery health declines rapidly, realizing closed-loop management from monitoring to maintenance, and improving battery life and vehicle operation safety.

[0006] Furthermore, in step S2, the voltage data set is analyzed, considering the voltage data A t The continuous n voltage data of the end point are the voltage data A t After denoising, the voltage data after denoising is A t_n : ; Where n is the number of reference data greater than or equal to 3, A k Indicates A t_n The kth reference data of , and then the denoised voltage data set {A 1_n ,A 2_n ,…,A t_n ,…,A T_n}, if in the voltage data set, in voltage data A t If there are no n voltage data for denoising, the voltage data in the previous acquisition cycle are borrowed in reverse order; The current data set, temperature data set and internal resistance data set are processed synchronously to obtain the denoised current data set {B 1_n ,B 2_n ,…,B t_n ,…,B T_n}、Denoised temperature data set {C 1_n ,C 2_n ,…,C t_n ,…,C T_n} and the denoised internal resistance data set {D 1_n ,D 2_n ,…,D t_n ,…,D T_n}; Analyze the denoised voltage data set and calculate the first quartile of the denoised voltage data as Q A_1 , the third quartile of the denoised voltage data is Q A_3 , set the minimum threshold for voltage abnormality judgment to QA_MIN , Q A_MIN =2.5*Q A_1 -1.5*Q A_3 , the maximum threshold for voltage abnormality judgment is Q A_MAX , Q A_MAX =2.5*Q A_3 -1.5*Q A_1 , if Q A_MIN <A t_n <Q A_MAX , then judge A t_n is the normal voltage data, otherwise the median of the denoised voltage data set is used to replace A t_n As normal voltage data, the normal voltage data set {a 1_n ,a 2_n ,…,a t_n ,…,a T_n}; The de-noised current data set, the de-noised temperature data set, the de-noised internal resistance data set and the cycle number data set are judged synchronously. The de-noised current data set, the de-noised temperature data set, the de-noised internal resistance data set and the cycle number data set are obtained in the same way as the normal voltage data, and then the normal current data set {b 1_n ,b 2_n ,…,b t_n ,…,b T_n}、Normal temperature data set {c 1_n ,c 2_n ,…,c t_n ,…,c T_n}、Normal internal resistance data set {d 1_n ,d 2_n ,…,d t_n ,…,d T_n} and the normal cycle number data set {e 1_n ,e 2_n ,…,e t_n ,…,e T_n}, and then obtain the normalized voltage feature set {a1, a2, …, a t ,…,a T}、Current feature set {b1,b2,…,b t ,…,b T}、Temperature feature set {c1,c2,…,c t ,…,c T}、Internal resistance feature set {d1,d2,…,d t ,…,d T} and the cycle number feature set {e1,e2,…,e t ,…,e TStep S2 uses a sliding window and historical borrowing denoising strategy to smooth the frequently fluctuating voltage and current data: when the new cycle data is insufficient, the historical data of the previous charge is automatically retrieved to fill the gap, avoiding the hidden dangers of missed judgments due to the small number of samples in the early collection. The subsequent dynamic threshold anomaly detection based on the data's own distribution does not rely on fixed standards, but makes flexible judgments based on the fluctuations of the battery's current data: for example, when the voltage suddenly jumps, it does not simply alarm, but replaces the abnormal point with the intermediate value, which not only retains the true trend but also eliminates accidental interference. More importantly, this processing acts simultaneously on the four sensors of voltage, current, temperature, and internal resistance to ensure that the multi-dimensional data "resonates at the same frequency" and avoid misdiagnosis caused by misjudgment of a single parameter. The clean data finally output is like laying a solid foundation for battery health assessment, making subsequent sub-health calculations and trend analysis closer to the actual state of the battery. It is especially suitable for the use scenarios of electric vehicles with frequent start-stop and complex working conditions, and improves the accuracy of full life cycle management from the source.

[0007] Furthermore, in step S3, the voltage feature set is analyzed, and the maximum value of the voltage feature is a max , the minimum value of the voltage characteristic is a min , the average value of the voltage characteristic is a0, and the lowest threshold of healthy voltage is set to a min +β*(a0-a min ) and the highest threshold of sub-health voltage a max -β*(a max -a0), when a min +β*(a0-a min )<a t <a max -β*(a max -a0), judge a t is a healthy voltage, otherwise it is judged as a t is the sub-health voltage, and the number of sub-health voltage is a 2_x β is the established sub-health judgment index, which can be set to 0.1, and x is the number of the acquisition cycle, and then the voltage sub-health level A of the xth acquisition cycle is obtained. 2_x =a 2_x / T; Synchronously analyze the current feature set, temperature feature set, internal resistance feature set and cycle number feature set to obtain the current sub-health level b of the xth acquisition cycle 2_x / T, the temperature sub-health level c of the xth acquisition cycle 2_x / T, the sub-health level of internal pressure in the xth collection cycle d 2_x / T, the number of cycles in the xth acquisition cycle and the degree of sub-health e 2_x / T, the average value of voltage sub-health level, current sub-health level, temperature sub-health level, internal pressure sub-health level and cycle number sub-health level is taken as the sub-health level W of the xth acquisition cycle x Conduct a comprehensive analysis of the multi-dimensional characteristics of voltage, current, temperature, internal resistance and number of cycles, dynamically set health thresholds, break free from the constraints of fixed standards, and better adapt to the complex and changing operating conditions of batteries. By calculating the sub-health level of each dimension separately and then taking the average, a deep fusion evaluation of multi-dimensional data is achieved, avoiding the one-sidedness and risk of misjudgment of single-dimensional judgments. This method comprehensively captures the battery performance status and accurately identifies potential anomalies, laying a solid foundation for subsequent health trend analysis and maintenance decisions, effectively improving the accuracy and comprehensiveness of battery health assessments, ensuring battery life and electric vehicle operation safety, reducing improper maintenance problems caused by assessment deviations, and making battery management more scientific and reliable.

[0008] Furthermore, in step S4, the battery health decline trend V at the xth cycle is calculated. x , V x =W x -W x-1 , when V x > V0, the battery health is judged to be declining rapidly, triggering a battery health decline warning, otherwise the detection of the next collection cycle will continue. V0 is the established battery health decline trend threshold; Furthermore, in step S5, when the detection of the next acquisition cycle is continued, the battery health decline trend V at the x-1th cycle is calculated. x-1 , if V x >V x-1 , calculate the number of battery health cycles U x , U x To meet U x *(V x -V x-1 )>V0, prompt the user to x Replace the battery within a certain period of time; track battery health trends in real time. By comparing the sub-health levels of adjacent periods, it can quickly identify accelerated health decline and trigger an early warning, allowing early detection of battery anomalies and avoiding sudden failures during operation. Furthermore, based on trend changes, it accurately calculates the minimum recommended battery replacement period, providing users with a clear maintenance time window, preventing both resource waste caused by premature replacement and safety impacts of delayed maintenance. This mechanism forms a closed-loop management system for monitoring, early warning, and maintenance, effectively extending battery life and improving the safety of electric vehicle operation. It allows users to enjoy scientific and efficient battery management, ensures stable and reliable vehicle use, and reduces the inconvenience and risks caused by battery problems.

[0009] The health status assessment system for lithium batteries in electric vehicles includes: battery data acquisition module, battery data preprocessing module, sub-health degree analysis module, battery health warning module and health cycle prediction module; The battery data acquisition module is used to collect battery data sets through the sensor group and the battery management system; The battery data preprocessing module is used to preprocess the battery data to obtain a feature data set; The sub-health analysis module is used to analyze the characteristic data set and calculate the sub-health level of the battery; The battery health warning module is used to trigger a battery health degradation warning based on the downward trend of the sub-health level; The health cycle prediction module is used to calculate the number of battery health cycles and prompt the user to replace the battery within a limited time.

[0010] Compared with the existing technology, the beneficial effects achieved by the present invention are: on the one hand, in the data acquisition stage, the multi-dimensional data such as voltage, current, temperature, internal resistance, etc. are denoised through sliding median filtering, effectively filtering out high-frequency noise and accidental interference. In response to the problem of insufficient data in the early acquisition cycle, a previous cycle data borrowing mechanism is designed to ensure that each data point can be denoised based on sufficient historical information to avoid analysis bias caused by insufficient data volume. This method of synchronous denoising of multi-sensor data not only retains the true trend of the data, but also improves the stability of the data, laying a reliable data foundation for subsequent health status assessment, and is particularly suitable for high-frequency data acquisition scenarios under complex working conditions during the operation of electric vehicles.

[0011] On the one hand, the outlier detection method based on the interquartile range can adaptively identify abnormal data that deviates from the normal range and correct outliers through a median replacement strategy to prevent individual extreme values ​​from interfering with the overall assessment. This method does not rely on preset fixed thresholds, but instead dynamically determines the reasonable range based on the data's inherent distribution characteristics. It has strong recognition capabilities for nonlinear anomalies such as sudden voltage changes and abnormal temperature increases that may occur during the lithium battery charging and discharging process. It simultaneously processes outliers in multi-dimensional data to ensure the consistency and coordination of various characteristic parameters, making the health status assessment results more closely aligned with the actual battery operating status and reducing the risk of misjudgments and missed detections.

[0012] On the other hand, a multi-dimensional sub-health assessment system covering voltage, current, temperature, internal resistance and number of cycles has been constructed. By comprehensively analyzing the sub-health levels in each dimension, it comprehensively reflects the various characteristics of battery performance degradation. On this basis, a health decline trend analysis is introduced, which not only focuses on the absolute value changes of the current state, but also judges the acceleration of the degradation rate by comparing the trends of adjacent cycles. When a rapid decline in health status or an intensified degradation trend is detected, the minimum cycle for recommended battery replacement can be accurately calculated, providing users with a specific maintenance time window. This dynamic early warning mechanism not only avoids the waste of resources caused by premature replacement, but also prevents the lag problem caused by maintenance, realizing closed-loop management of lithium batteries from status monitoring to active maintenance, effectively improving battery service life and the safety of electric vehicle operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is a structural diagram of the health status assessment system for lithium batteries of electric vehicles applied to the present invention; Figure 2 The present invention is a flow chart of a method for evaluating the health status of a lithium battery in an electric vehicle. DETAILED DESCRIPTION

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0015] See also Figure 1 and Figure 2 The present invention provides a technical solution: a method for evaluating the health status of a lithium battery in an electric vehicle, comprising the following steps: S1, collecting battery data sets through sensor groups and battery management systems; S2. Preprocess the battery data to obtain a feature data set; S3. Analyze the characteristic data set and calculate the sub-health level of the battery; S4. Trigger a battery health degradation warning based on the decreasing trend of the sub-health level; S5. Calculate the number of battery health cycles and prompt the user to set a time limit for battery replacement.

[0016] In step S1, the sensor group includes a voltage sensor, a current sensor, a temperature sensor and an internal resistance sensor, and sets the data collection time interval a. In any collection cycle, the voltage A is collected in real time by the voltage sensor. t , real-time voltage B is collected through the current sensor t , collect real-time temperature C through temperature sensor t , collect real-time internal resistance D through internal resistance sensor t , collect the cycle number E through the battery management system t , the battery is charged and discharged once, which is recorded as one charge and discharge cycle. The number of cycles refers to the number of charge and discharge cycles of the lithium battery, and then the voltage data set {A1, A2, ..., A t ,…,A T}、Current data set {B1,B2,…,B t ,…,B T}、Temperature data set {C1,C2,…,C t ,…,C T}、Internal resistance data set {D1,D2,…,D t ,…,D T} and the cycle number data set {E1,E2,…,E t ,…,E T}, where T represents the number of times data is collected, and t represents the tth time data is collected; in data processing, sliding filtering and borrowing of data from previous cycles are used to effectively filter out noise and solve the problem of insufficient initial data, making the voltage, current and other data collected by multiple sensors more stable and reliable, laying a solid data foundation for evaluation. When identifying anomalies, a reasonable range is dynamically determined based on the data's own distribution, and outliers are replaced with the median to accurately capture nonlinear problems such as voltage mutations and temperature anomalies, reducing false positives and missed judgments, and making the evaluation results closer to the actual state of the battery. The evaluation system integrates multi-dimensional indicators, observing both the current state and analyzing trend changes. It can provide a specific maintenance time window when the battery health declines rapidly, realizing closed-loop management from monitoring to maintenance, and improving battery life and vehicle operation safety.

[0017] In step S2, the voltage data set is analyzed, considering the voltage data A t The continuous n voltage data of the end point are the voltage data A t After denoising, the voltage data after denoising is A t_n : ; Where n is the number of reference data greater than or equal to 3, A k Indicates A t_n The kth reference data of , and then the denoised voltage data set {A 1_n ,A2_n ,…,A t_n ,…,A T_n}, if in the voltage data set, in voltage data A t If there are no n voltage data for denoising, the voltage data in the previous acquisition cycle are borrowed in reverse order; The current data set, temperature data set and internal resistance data set are processed synchronously to obtain the denoised current data set {B 1_n ,B 2_n ,…,B t_n ,…,B T_n}、Denoised temperature data set {C 1_n ,C 2_n ,…,C t_n ,…,C T_n} and the denoised internal resistance data set {D 1_n ,D 2_n ,…,D t_n ,…,D T_n}; Analyze the denoised voltage data set and calculate the first quartile of the denoised voltage data as Q A_1 , the third quartile of the denoised voltage data is Q A_3 , set the minimum threshold for voltage abnormality judgment to Q A_MIN , Q A_MIN =2.5*Q A_1 -1.5*Q A_3 , the maximum threshold for voltage abnormality judgment is Q A_MAX , Q A_MAX =2.5*Q A_3 -1.5*Q A_1 , if Q A_MIN <A t_n <Q A_MAX , then judge A t_n is the normal voltage data, otherwise the median of the denoised voltage data set is used to replace A t_n As normal voltage data, the normal voltage data set {a 1_n ,a 2_n ,…,a t_n ,…,a T_n}; The denoised current data set, denoised temperature data set, denoised internal resistance data set and cycle number data set are judged synchronously to obtain the normal current data set {b 1_n ,b 2_n ,…,b t_n ,…,b T_n}、Normal temperature data set {c 1_n ,c 2_n ,…,c t_n ,…,cT_n}、Normal internal resistance data set {d 1_n ,d 2_n ,…,d t_n ,…,d T_n} and the normal cycle number data set {e 1_n ,e 2_n ,…,e t_n ,…,e T_n}, and then obtain the normalized voltage feature set {a1, a2, …, a t ,…,a T}、Current feature set {b1,b2,…,b t ,…,b T}、Temperature feature set {c1,c2,…,c t ,…,c T}、Internal resistance feature set {d1,d2,…,d t ,…,d T} and the cycle number feature set {e1,e2,…,e t ,…,e T Step S2 uses a sliding window and historical borrowing denoising strategy to smooth the frequently fluctuating voltage and current data: when the new cycle data is insufficient, the historical data of the previous charge is automatically retrieved to fill the gap, avoiding the hidden dangers of missed judgments due to the small number of samples in the early collection. The subsequent dynamic threshold anomaly detection based on the data's own distribution does not rely on fixed standards, but makes flexible judgments based on the fluctuations of the battery's current data: for example, when the voltage suddenly jumps, it does not simply alarm, but replaces the abnormal point with the intermediate value, which not only retains the true trend but also eliminates accidental interference. More importantly, this processing acts simultaneously on the four sensors of voltage, current, temperature, and internal resistance to ensure that the multi-dimensional data "resonates at the same frequency" and avoid misdiagnosis caused by misjudgment of a single parameter. The clean data finally output is like laying a solid foundation for battery health assessment, making subsequent sub-health calculations and trend analysis closer to the actual state of the battery. It is especially suitable for the use scenarios of electric vehicles with frequent start-stop and complex working conditions, and improves the accuracy of full life cycle management from the source.

[0018] In step S3, the voltage feature set is analyzed and the maximum value of the voltage feature is a max , the minimum value of the voltage characteristic is a min , the average value of the voltage characteristic is a0, and the lowest threshold of healthy voltage is set to a min +β*(a0-a min ) and the highest threshold of sub-health voltage a max -β*(a max -a0), when a min +β*(a0-a min )<a t <a max-β*(a max -a0), judge a t is a healthy voltage, otherwise it is judged as a t is the sub-health voltage, and the number of sub-health voltage is a 2_x β is the established sub-health judgment index, which can be set to 0.1, and x is the number of the acquisition cycle, and then the voltage sub-health level A of the xth acquisition cycle is obtained. 2_x =a 2_x / T; Synchronously analyze the current feature set, temperature feature set, internal resistance feature set and cycle number feature set to obtain the current sub-health level b of the xth acquisition cycle 2_x / T, the temperature sub-health level c of the xth acquisition cycle 2_x / T, the sub-health level of internal pressure in the xth collection cycle d 2_x / T, the number of cycles in the xth acquisition cycle and the degree of sub-health e 2_x / T, the average value of voltage sub-health level, current sub-health level, temperature sub-health level, internal pressure sub-health level and cycle number sub-health level is taken as the sub-health level W of the xth acquisition cycle x Conduct a comprehensive analysis of the multi-dimensional characteristics of voltage, current, temperature, internal resistance and number of cycles, dynamically set health thresholds, break free from the constraints of fixed standards, and better adapt to the complex and changing operating conditions of batteries. By calculating the sub-health level of each dimension separately and then taking the average, a deep fusion evaluation of multi-dimensional data is achieved, avoiding the one-sidedness and risk of misjudgment of single-dimensional judgments. This method comprehensively captures the battery performance status and accurately identifies potential anomalies, laying a solid foundation for subsequent health trend analysis and maintenance decisions, effectively improving the accuracy and comprehensiveness of battery health assessments, ensuring battery life and electric vehicle operation safety, reducing improper maintenance problems caused by assessment deviations, and making battery management more scientific and reliable.

[0019] In step S4, the battery health decline trend V at the xth cycle is calculated. x , V x =W x -W x-1 , when V x When V0 is higher than V0, it is determined that the battery health is deteriorating rapidly, and a battery health degradation warning is triggered. Otherwise, the detection of the next collection cycle is continued. The V0 is the established battery health degradation trend threshold.

[0020] In step S5, when the detection of the next acquisition cycle is continued, the battery health decline trend V at the x-1th cycle is calculated. x-1 , if V x >V x-1 , calculate the number of battery health cycles U x , Ux To meet U x *(V x -V x-1 )>V0, prompt the user to x Replace the battery within a certain period of time; track battery health trends in real time. By comparing the sub-health levels of adjacent periods, it can quickly identify accelerated health decline and trigger an early warning, allowing early detection of battery anomalies and avoiding sudden failures during operation. Furthermore, based on trend changes, it accurately calculates the minimum recommended battery replacement period, providing users with a clear maintenance time window, preventing both resource waste caused by premature replacement and safety impacts of delayed maintenance. This mechanism forms a closed-loop management system for monitoring, early warning, and maintenance, effectively extending battery life and improving the safety of electric vehicle operation. It allows users to enjoy scientific and efficient battery management, ensures stable and reliable vehicle use, and reduces the inconvenience and risks caused by battery problems.

[0021] A health status assessment system for lithium batteries in electric vehicles, comprising a battery data acquisition module, a battery data preprocessing module, a sub-health analysis module, a battery health warning module, and a health cycle prediction module; The battery data acquisition module is used to collect battery data sets through the sensor group and the battery management system; The battery data preprocessing module is used to preprocess the battery data to obtain a feature data set; The sub-health analysis module is used to analyze the characteristic data set and calculate the sub-health level of the battery; The battery health warning module is used to trigger a battery health degradation warning based on the downward trend of the sub-health level; The health cycle prediction module is used to calculate the number of battery health cycles and prompt the user to replace the battery within a limited time.

[0022] Example 1: A user starts a vehicle, and the onboard battery management system immediately begins operating. Sensors installed in the battery pack collect multiple data points in real time: a voltage sensor continuously senses the battery's output status, a current sensor tracks energy flow during driving, a temperature sensor records the battery's temperature fluctuations in the environment, and an internal resistance sensor monitors internal conductivity. The system also simultaneously counts the number of charge and discharge cycles of the battery. This data is continuously collected at fixed intervals, providing basic information for subsequent health assessments. While driving, the sensor occasionally captures brief voltage fluctuations. The system activates a data smoothing mechanism, and for each data point, it performs a comprehensive calculation based on the most recent historical data to eliminate instantaneous outliers. For example, when a voltage value collected at a certain time suddenly jumps, the system will not use this data directly, but will combine it with the previously stable data for weighted averaging to generate more reliable denoised data. This process is applied simultaneously to all collected data such as current, temperature, and internal resistance to ensure the stability and accuracy of each parameter. When the vehicle was charging in the afternoon, the system analyzed the denoised data and found that the temperature data for a certain period of time was significantly outside the normal range. Based on the distribution pattern of historical data, the system set a reasonable value range, and abnormal data outside this range was automatically identified. For these abnormal points, the system does not directly discard or retain them, but replaces them with the median value of similar data to ensure that all data is within a reasonable physical range. For example, the abnormal value of the current suddenly increased during a fast charge was corrected and returned to a reasonable level that conforms to the charging curve; After multiple consecutive monitoring cycles, the system collects statistics on abnormal conditions in each parameter. For example, low temperatures can cause the voltage to drop below the healthy standard multiple times, frequent fast charging can cause current fluctuations to exceed the limit, and long-term use can cause the internal resistance to gradually increase. The system calculates the abnormal proportion of each parameter within the total monitoring cycle, such as the abnormal voltage ratio and the abnormal current ratio, and then averages these ratios to obtain the overall sub-health level of the current battery, which directly reflects the performance degradation of the battery in multiple key parameters. While driving the next day, the system compared the sub-health level of the current cycle with the previous cycle and found that the increase exceeded the preset warning threshold. Further analysis of the trend showed that the rate of health decline was accelerating. Based on this trend, the system automatically calculated the number of cycles remaining before the battery's critical performance value and issued a reminder to the driver through the on-board screen: "Battery health is declining at an accelerated rate. Maintenance and inspection are recommended in the near future." The driver promptly went to the service station according to the prompt. Professional equipment verified that some cells in the battery pack had experienced capacity degradation. Premature replacement avoided a possible subsequent drop in battery life or safety hazards. Through real-time data collection, intelligent noise reduction and correction, comprehensive status assessment and trend warning, dynamic monitoring of the health status of electric vehicle lithium batteries is achieved, providing users with scientific maintenance recommendations and effectively ensuring the safe and efficient use of batteries.

[0023] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A health status assessment method for lithium batteries in electric vehicles, characterized by: The method comprises the following steps: S1, collecting battery data sets through sensor groups and battery management systems; S2. Preprocess the battery data to obtain a feature data set; S3. Analyze the characteristic data set and calculate the sub-health level of the battery; S4. Trigger a battery health degradation warning based on the decreasing trend of the sub-health level; S5. Calculate the number of battery health cycles and prompt the user to set a time limit for battery replacement.

2. The health status assessment method for electric vehicle lithium batteries according to claim 1, characterized in that: In step S1, the sensor group includes a voltage sensor, a current sensor, a temperature sensor and an internal resistance sensor, and sets the data collection time interval a. In any collection cycle, the voltage A is collected in real time by the voltage sensor. t , real-time voltage B is collected through the current sensor t , collect real-time temperature C through temperature sensor t , collect real-time internal resistance D through internal resistance sensor t , collect the cycle number E through the battery management system t , the number of cycles refers to the number of charge and discharge times of the lithium battery, and then the voltage data set {A1, A2, ..., A t ,…,A T }、Current data set {B1,B2,…,B t ,…,B T }、Temperature data set {C1,C2,…,C t ,…,C T }、Internal resistance data set {D1,D2,…,D t ,…,D T } and the cycle number data set {E1,E2,…,E t ,…,E T }, where T represents the number of times data is collected, and t represents the tth time data is collected.

3. The health status assessment method for electric vehicle lithium batteries according to claim 2, characterized in that: In step S2, the voltage data set is analyzed, considering the voltage data A t The continuous n voltage data of the end point are the voltage data A t After denoising, the voltage data after denoising is A t_n : ； Where n is the number of reference data greater than or equal to 3, A k Indicates A t_n The kth reference data of , and then the denoised voltage data set {A 1_n ,A 2_n ,…,A t_n ,…,A T_n }, if in the voltage data set, in voltage data A t If there are no n voltage data for denoising, the voltage data in the previous acquisition cycle are borrowed in reverse order; The current data set, temperature data set and internal resistance data set are processed synchronously to obtain the denoised current data set {B 1_n ,B 2_n ,…,B t_n ,…,B T_n }、Denoised temperature data set {C 1_n ,C 2_n ,…,C t_n ,…,C T_n } and the denoised internal resistance data set {D 1_n ,D 2_n ,…,D t_n ,…,D T_n }.

4. The health status assessment method for lithium batteries in electric vehicles according to claim 3, characterized in that: Analyze the denoised voltage data set and calculate the first quartile of the denoised voltage data as Q A_1 , the third quartile of the denoised voltage data is Q A_3 , set the minimum threshold for voltage abnormality judgment to Q A_MIN , the maximum threshold for voltage abnormality judgment is Q A_MAX , if Q A_MIN <A t_n <Q A_MAX , then judge A t_n is the normal voltage data, otherwise the median of the denoised voltage data set is used to replace A t_n As normal voltage data, the normal voltage data set {a 1_n ,a 2_n ,…,a t_n ,…,a T_n }; The denoised current data set, denoised temperature data set, denoised internal resistance data set and cycle number data set are judged synchronously to obtain the normal current data set {b 1_n ,b 2_n ,…,b t_n ,…,b T_n }、Normal temperature data set {c 1_n ,c 2_n ,…,c t_n ,…,c T_n }、Normal internal resistance data set {d 1_n ,d 2_n ,…,d t_n ,…,d T_n } and the normal cycle number data set {e 1_n ,e 2_n ,…,e t_n ,…,e T_n }, and then obtain the normalized voltage feature set {a1, a2, …, a t ,…,a T }、Current feature set {b1,b2,…,b t ,…,b T }、Temperature feature set {c1,c2,…,c t ,…,c T }、Internal resistance feature set {d1,d2,…,d t ,…,d T } and the cycle number feature set {e1,e2,…,e t ,…,e T }.

5. The health status assessment method for lithium batteries in electric vehicles according to claim 4, characterized in that: In step S3, the voltage feature set is analyzed and the maximum value of the voltage feature is a max , the minimum value of the voltage characteristic is a min , the average value of the voltage characteristic is a0, and the lowest threshold of healthy voltage is set to a min +β*(a0-a min ) and the highest threshold of sub-health voltage a max -β*(a max -a0), when a min +β*(a0-a min )<a t <a max -β*(a max -a0), judge a t is a healthy voltage, otherwise it is judged as a t is the sub-health voltage, and the number of sub-health voltage is a 2_x β is the established sub-health judgment index, x is the number of the acquisition cycle, and the voltage sub-health level A of the xth acquisition cycle is obtained. 2_x =a 2_x / T.

6. The health status assessment method for lithium batteries in electric vehicles according to claim 5, characterized in that: Synchronously analyze the current feature set, temperature feature set, internal resistance feature set and cycle number feature set to obtain the current sub-health level b of the xth acquisition cycle 2_x / T, the temperature sub-health level c of the xth acquisition cycle 2_x / T, the sub-health level of internal pressure in the xth collection cycle d 2_x / T, the number of cycles of the xth acquisition cycle and the degree of sub-health e 2_x / T, the average value of voltage sub-health level, current sub-health level, temperature sub-health level, internal pressure sub-health level and cycle number sub-health level is taken as the sub-health level W of the xth acquisition cycle x .

7. The health status assessment method for lithium batteries in electric vehicles according to claim 6, characterized in that: In step S4, the battery health decline trend V at the xth cycle is calculated. x , V x =W x -W x-1 , when V x When V0 is higher than V0, it is determined that the battery health is deteriorating rapidly, and a battery health degradation warning is triggered. Otherwise, the detection of the next collection cycle is continued. The V0 is the established battery health degradation trend threshold.

8. The health status assessment method for lithium batteries in electric vehicles according to claim 6, characterized in that: In step S5, when the detection of the next acquisition cycle is continued, the battery health decline trend V at the x-1th cycle is calculated. x-1 , if V x >V x-1 , calculate the number of battery health cycles U x , U x To meet U x *(V x -V x-1 )>V0 minimum positive integer, prompt the user to x Replace the battery within a certain period.

9. A health status assessment system for a lithium battery in an electric vehicle, wherein the system is applied to the health status assessment method for a lithium battery in an electric vehicle according to any one of claims 1 to 7, characterized in that: The system includes: a battery data acquisition module, a battery data preprocessing module, a sub-health degree analysis module, a battery health warning module and a health cycle prediction module; The battery data acquisition module is used to collect battery data sets through a sensor group and a battery management system; The battery data preprocessing module is used to preprocess the battery data to obtain a feature data set; The sub-health degree analysis module is used to analyze the characteristic data set and calculate the sub-health degree of the battery; The battery health warning module is used to trigger a battery health degradation warning according to the downward trend of the sub-health level; The health cycle prediction module is used to calculate the number of battery health cycles and prompt the user to change the battery within a limited time.

Images