System and Method for Evaluating Performance and Predicting Residual Value of Electric Vehicle Battery

Abstract

The present invention may provide an electric vehicle battery performance evaluation and residual value prediction system comprising: an evaluation data collection device for collecting evaluation data to evaluate the performance and predict the residual value of a battery installed in an electric vehicle; and a battery evaluation server for receiving the evaluation data by communicating with the evaluation data collection device and evaluating the battery, wherein the evaluation data includes the initial rated capacity, current charge amount, SOC, and battery voltage data of the battery, and the battery evaluation server includes a battery residual value evaluation unit that calculates the battery SOH using the evaluation data.

Description

Technology Field

[0001] The present invention relates to a system for evaluating the performance and predicting the residual value of a battery used for driving motors of various means of transportation, such as electric vehicles and personal mobility (PM), and a method for evaluating the performance and predicting the residual value of a battery. Background Technology

[0002] With the expansion of electric vehicle adoption, active technological development is underway to extend driving range through high-energy density batteries. While the performance of electric vehicle batteries is continuously improving due to these technological advancements, there are limitations to the battery's usable range under the conditions of ensuring safety and maintaining reliability. Generally, when a battery's health drops below a certain level (e.g., 80%), it is deemed unsuitable for driving within an electric vehicle and becomes subject to disposal.

[0003] As such, the performance degradation of vehicles goes beyond a simple replacement issue and serves as a critical indicator for ensuring operational safety, establishing periodic inspection standards, and evaluating residual value during used car transactions. Therefore, there is a demand for battery performance evaluation and residual value prediction technologies that comprehensively consider actual usage history—such as charge / discharge history, operating environment, and driving habits—rather than just battery status information including State of Charge (SOC), State of Health (SOH), State of Balance (SOB), and State of Power (SOP).

[0004] Existing battery diagnostic methods rely on static measurements at specific points in time or testing after removal, and therefore fail to adequately reflect the patterns of performance degradation that occur in actual driving environments. Consequently, there is a growing need for technology that evaluates electric vehicle batteries based on non-removal data while they remain mounted in the vehicle. The problem to be solved

[0005] Therefore, there is a need for technology capable of collecting and analyzing data based on the actual operational history of batteries and quantitatively predicting their residual value and lifespan. In particular, a data-driven prediction system that comprehensively considers various factors—such as key state variables including battery State of Health (SOH), charge / discharge history, and driving conditions—is recognized as an essential technology for ensuring vehicle safety as well as for used car transactions.

[0006] However, existing battery performance evaluation technologies have the following problems.

[0007] First, traditional diagnostic methods evaluate performance by removing the battery from the vehicle and conducting repeated charge-discharge tests; however, this process is time-consuming and costly, and fails to reflect actual driving conditions. Additionally, removing the battery carries the risk of adverse effects on vehicle electronic components such as the ECU and BMS.

[0008] Second, existing methods statically diagnose only the current state of the battery and do not reflect dynamic operating conditions such as usage history, driving habits, temperature, and route, resulting in low accuracy in predicting performance degradation trends or remaining lifespan.

[0009] Third, various driving environments such as urban driving, high-speed driving, inclines, low / high temperature conditions, and rapid acceleration have a significant impact on battery degradation, but there is a lack of systems to analyze this contextual information in conjunction with battery status information.

[0010] Fourth, the evaluation method based on charging time and temperature, which is based on a fully charged state, has limitations in practical application due to the long measurement time and high data throughput.

[0011] Accordingly, the present invention provides a non-invasive data-based performance prediction system capable of diagnosing the battery condition and degree of degradation while the battery is mounted in the vehicle without separating the battery. Furthermore, the invention aims to provide a highly effective system and method capable of various applications by collecting and analyzing various dynamic data accumulated during operation, such as SOH, SOC, charge / discharge history, temperature, current, driving speed, and driving habits, to predict residual value, such as battery performance and remaining lifespan, with high accuracy. means of solving the problem

[0012] An electric vehicle battery performance evaluation and residual value prediction system according to one embodiment of the present invention comprises: an evaluation data collection device for collecting evaluation data to evaluate the performance and predict the residual value of a battery installed in an electric vehicle; and a battery evaluation server for receiving the evaluation data by communicating with the evaluation data collection device and evaluating the performance of the battery; wherein the evaluation data includes the initial rated capacity, current charge amount, State of Charge (SOC), and battery voltage data of the battery, and the battery evaluation server may include a battery residual value evaluation unit that calculates the battery SOH using the evaluation data.

[0013] In addition, the battery residual value evaluation unit calculates the SOH based on the ratio of the current available capacity to the initial rated capacity of the electric vehicle, and the initial rated capacity and the current available capacity can be calculated using the charge unit (Ah).

[0014] In addition, the battery residual value evaluation unit can calculate SOH only when the SOC is in the range of 20 or more and 80 or less.

[0015] In addition, the evaluation data collection device further includes a vehicle data collection unit that collects at least one of the battery usage environment data and vehicle operation data, and the battery residual value evaluation unit can calculate the battery's SOH using at least one regression model among Random Forest Regression, Polynomial Regression, LSTM-based Time Series Regression, Gradient Boosting Regression, Linear Regression, and Decision Tree Regression, using at least one of the data among the total driving distance of the electric vehicle, battery temperature, and ambient temperature collected by the vehicle data collection unit as independent variables.

[0016] In addition, the battery evaluation server may further include an item-specific score calculation unit that calculates a reliability score for the battery SOH calculated by the battery residual value evaluation unit.

[0017] In addition, the battery residual value evaluation unit can continuously update the regression model by reflecting driving history and sensor data.

[0018] A method for evaluating electric vehicle battery performance and predicting residual value according to one embodiment of the present invention comprises: a first step of collecting initial rated capacity data of a battery installed in an electric vehicle; a second step of collecting evaluation data including the current charge amount, SOC, and battery voltage of the battery; and a third step of calculating the state of health (SOH) of the battery using the initial rated capacity, current charge amount, SOC, and battery voltage data of the battery, wherein the third step is characterized by using a regression model in which at least one of the data collected in the second step is used as an independent variable and the SOH of the battery to be measured is used as a dependent variable. Effects of the invention

[0019] According to the present invention, since battery performance evaluation and residual value can be estimated based on data collected during vehicle operation, there is an advantage that diagnosis is possible without removing the battery or performing a discharge test with external equipment.

[0020] In addition, battery performance evaluation and residual value prediction are possible even in a partially charged state without fully charging the battery, which has the effect of reducing measurement time and data throughput.

[0021] Furthermore, unlike existing threshold-based or empirical diagnostic methods, the present invention has the advantage of being able to numerically predict accurate and continuous SOH values ​​by learning the mathematical relationship between sensor data (voltage, current, SOC, temperature, etc.) and SOH using a regression model. In addition, since regression analysis allows for modeling by reflecting various variables such as temperature, driving patterns, and charge / discharge history, it can possess robust predictive performance even with environmental changes; moreover, as data is continuously accumulated, retraining to improve model performance is possible, thereby enhancing accuracy over the long term. Brief explanation of the drawing

[0022] Figure 1 is a schematic diagram of the electric vehicle battery performance evaluation and residual value prediction system of the present invention. Figure 2 is a configuration diagram of the electric vehicle battery performance evaluation and residual value prediction system of the present invention. Figure 3 is a graph showing the current available capacity of the battery according to the change in SOC in terms of power (Wh), and compares the results under two average temperature conditions. Figure 4 is a graph showing the current available capacity of the battery according to the change in SOC based on the charge (Ah), and compares the results under two average temperature conditions. Figure 5 is a graph showing the current available capacity of the battery according to changes in ambient temperature in terms of power (Wh). Figure 6 is a graph showing the current available capacity of the battery according to changes in ambient temperature based on the charge (Ah). Figure 7 is a graph showing the current available capacity of a battery according to changes in SOC based on the charge (Ah). Figures 8 to 11 are graphs showing the current available capacity of a battery measured while fixing the SOC and ambient temperature to a specific range, recorded by date. Figure 12 shows the SOH predicted using a Random Forest regression model with the collected evaluation data as independent variables and compared with the SOH measured through BMS, etc. Figure 13 is a graph showing the importance of variables provided by Random Forest. Figure 14 is a graph showing the importance of variables and their correlation with SOH through SHAP value. Figure 15 is a graph showing the influence of each input variable on the result in a SOH prediction model using the Partial Dependence Plot (PDP) technique. Figure 16 is a graph showing the predicted SOH values ​​according to the external temperature. Specific details for implementing the invention

[0023] Hereinafter, embodiments of the system and method according to the present invention will be described with reference to the attached drawings. At this time, the present invention is not limited or restricted by the embodiments below. Furthermore, in describing the present invention, specific descriptions of known functions or configurations may be omitted to clarify the gist of the present invention.

[0024] FIG. 1 is a schematic diagram of the electric vehicle battery performance evaluation and residual value prediction system of the present invention, and FIG. 2 is a configuration diagram of the electric vehicle battery performance evaluation and residual value prediction system of the present invention.

[0025] Referring to FIGS. 1 and 2, the electric vehicle battery performance evaluation and residual value prediction system (10) according to the present invention may include an evaluation data collection device (100) for collecting evaluation data to evaluate the residual value of a battery mounted on an electric vehicle (1), and a battery evaluation server (200) for receiving evaluation data by communicating with the evaluation data collection device (100) and predicting the performance evaluation and residual value of the battery.

[0026] In the present invention, the electric vehicle (1) may include a pure electric vehicle (EV), a hybrid electric vehicle (HEV), a plug-in hybrid electric vehicle (PHEV), etc.

[0027] The evaluation data collection device (100) may include a vehicle specification collection unit (110), a battery information collection unit (120), and a vehicle data collection unit (130). As an embodiment of the present invention, the evaluation data collection device (100) may include an OBD (On-Board Diagnostics) terminal, which is an electric vehicle terminal installed in an electric vehicle.

[0028] The battery evaluation server (200) may include a battery residual value evaluation unit (210) for evaluating the performance and predicting the residual value of the battery using evaluation data provided from the evaluation data collection device (100).

[0029] The evaluation data collection device (100) and the battery evaluation server (200) may each be provided with a communication unit that communicates with each other to transmit and receive data.

[0030] The evaluation data collection device (100) can receive vehicle and battery-related data from various sensors installed in the electric vehicle (1).

[0031] The vehicle specification collection unit (110) included in the evaluation data collection device (100) can collect evaluation data such as the electric vehicle manufacturer, vehicle type, year of production, battery capacity (distinguished as standard / long range, 58~77.4kWh, etc.), standard fuel efficiency (kw / kWh) (cumulative driving distance (km) / cumulative energy consumed in driving the vehicle (kWh)) (distinguished as city / highway, etc.).

[0032] In addition, the vehicle specification collection unit (110) can collect the vehicle model name, chassis number, year of manufacture, serial number of the vehicle (electric vehicle) terminal (basic terminal information such as modem type), IP of the vehicle (electric vehicle) terminal, firmware version of the vehicle (electric vehicle) terminal, etc.

[0033] The battery information collection unit (120) can collect information about the battery that is subject to residual value determination. The battery information may include information that serves as a basis for battery performance evaluation and residual value determination, such as battery identification information and battery usage information.

[0034] Battery identification information is information for distinguishing batteries and may include battery type (lithium-ion / solid-state, etc.) and battery capacity (indicated as standard / long-range 58~77.4kWh, etc.). Battery usage information may include data regarding the power of the battery consumed during the operation of the electric vehicle.

[0035] In one embodiment, the battery information collection unit (120) may include a Battery Management System (BMS). The BMS may be a system configuration that senses relevant information such as the current, voltage, and temperature of the battery and controls the battery to perform optimally. Such a BMS can prevent overcharging and / or over-discharging of the battery by monitoring relevant information, and can cut off the power through control when overcharging and / or over-discharging occurs.

[0036] The vehicle data collection unit (130) can collect battery charging record data, battery usage environment data, and vehicle operation data of the electric vehicle (1).

[0037] The charging record data is a charging record for the battery and may include data regarding whether fast / slow charging was performed, the number of charges, the amount charged during charging, etc.

[0038] Battery usage environment data may include information regarding the usage environment in which the battery is used, such as the environment in which the battery is cooled (air-cooled, water-cooled, etc.), the maximum / minimum temperature at which the battery is used, and the vehicle's ambient temperature (maximum / minimum / average temperature of the travel area).

[0039] Vehicle operation data is data on vehicle operation and may include information on vehicle driving distance (driving distance per full charge), vehicle driving speed, number of sudden acceleration / sudden deceleration / sudden start / sudden stop / sudden turn, etc.

[0040] In addition, vehicle operation data may include information regarding the vehicle's location information via GPS, driving time per trip, driving distance per trip, electricity consumption, fuel efficiency, maximum vehicle speed, idling time, fuel cost, and the ratio of vehicle speed within a trip.

[0041] The battery residual value evaluation unit (210) can extract information necessary for battery performance evaluation and residual value prediction, such as SOC, SOH, SOB, and SOP, based on various evaluation data collected from the evaluation data collection device (100), to output nodes.

[0042] Here, SOC is an indicator representing the state of charge of a battery, which can display the amount of power remaining in the battery as a percentage. For example, an SOC of 80% means that the battery has 80% of its energy. Electric vehicle users can check the vehicle's remaining driving range and plan when to charge using the SOC.

[0043] In one embodiment, the battery residual value evaluation unit (210) can calculate SOC, SOH, SOB, and SOP values ​​by analyzing data collected through an analysis / control program to extract multiple feature information and training it using at least one of a plurality of machine learning algorithms. The machine learning algorithm used in the analysis / control program may be one of a Decision Tree (DT) classification algorithm, a Random Forest classification algorithm, or a Support Vector Machine (SVM; a binary linear classification model that determines which category a given data belongs to). However, the prediction model used in the analysis / control program is not limited to machine learning techniques and may include various artificial intelligence and data-based methodologies such as deep learning, reinforcement learning, large language models (LLM), expert system-based rule-based models, statistical regression models, and hybrid models. That is, the present invention is not limited to a specific algorithm or learning method and may include various forms of prediction systems or algorithms capable of predicting remaining lifespan or performance based on battery state information.

[0044] A battery residual value evaluation unit (210) that receives charge / discharge data of a battery to be measured installed in an electric vehicle (1) and performs performance evaluation and residual value prediction of the battery can extract the same section based on the SOC or time of the charge / discharge data of the battery to be measured and the standard charge / discharge data of a new battery using residual value measurement software.

[0045] The battery residual value evaluation unit (210) can calculate a residual value including one or more of the state of charge (SOC: State of Charge, %), remaining life (SOH: State of Health, %), performance level (S, A, B, C, D, E grades), and time required to measure residual value for the battery being measured by comparing the amount of change in SOC in the same section.

[0046] The battery residual value evaluation unit (210) compares the residual value measurement result with the measurement indicator to verify the residual value measurement result and determine whether an error value occurs. If an error value occurs, the battery residual value evaluation unit (210) can compensate for the error value through correction.

[0047] The battery evaluation server (200) may further include a data classification unit (220).

[0048] The data classification unit (220) can classify various data collected through the vehicle specification collection unit (110), battery information collection unit (120), and vehicle data collection unit (130) by category. The classified categories may include a first category in which real-time information related to the vehicle's driving environment is classified, and a second category in which cumulative statistical information related to the vehicle's driving environment is classified.

[0049] The first item may include a 1-1 detailed item related to environmental information at the time the evaluation item data was acquired, a 1-2 detailed item related to driving information at the time the evaluation item data was acquired, and a 1-3 detailed item related to the vehicle battery at the time the evaluation item data was acquired.

[0050] Detail 1-1 may include weather data, temperature data, and seasonal data. Detail 1-2 may include vehicle speed data, vehicle mileage data, and vehicle location information data. Detail 1-3 may include battery temperature data, battery voltage data, and battery current data.

[0051] The second item may include a second-1 detailed item related to the average temperature of the battery and a second-2 detailed item related to the average voltage of the battery. The second-1 detailed item may include data on the average temperature of the battery over a preset period. The second-2 detailed item may include data on the average voltage of the battery over a preset period.

[0052] Here, evaluation item data may include at least one of vehicle operation information, environmental information, and battery information, and may include a plurality of first data related to a preset item based on at least one of vehicle operation information, environmental information, and battery information.

[0053] A plurality of first data may include at least one of weather data, temperature data, season data, vehicle speed data, vehicle mileage data, vehicle location information data, battery temperature data, battery voltage data, battery current data, data on the average battery temperature over a preset period, and data on the average battery voltage over a preset period. In one embodiment, the preset period may be, for example, 12 months.

[0054] The data classification unit (220) can classify these multiple first data by item.

[0055] Additionally, the battery evaluation server (200) may further include an item-specific score calculation unit (230). The item-specific score calculation unit (230) can calculate scores based on a plurality of first data classified by item from the data classification unit (220).

[0056] The item-specific score calculation unit (230) can calculate item-specific score indicators to calculate scores. The item-specific score indicators may be criteria for calculating scores. When item-specific score indicators are calculated, item-specific scores can be calculated based on the calculated item-specific score indicators.

[0057] The item-specific score calculation unit (230) can calculate a reliability score for the battery SOH calculated by the battery residual value evaluation unit (210). Additionally, the item-specific score calculation unit (230) can calculate a reliability score for the SOB and SOP calculated by the battery residual value evaluation unit (210). The reliability score can be calculated by comparing the values ​​of SOH, SOB, SOP, etc. calculated by the item-specific score calculation unit (230) with the measured values. In one embodiment, the measured values ​​may be values ​​measured through a BMS installed in the electric vehicle (1) or values ​​measured after removing the battery.

[0058] The battery evaluation server (200) may further include an evaluation score calculation unit (240).

[0059] The evaluation score calculation unit (240) can calculate the battery evaluation score by inputting the scores calculated for each item by the item-by-item score calculation unit (230) into a preset evaluation model.

[0060] The evaluation model may have at least one of the scores calculated for each item input, and the evaluation model may be a model in which at least one detailed item is input.

[0061] The evaluation model may be a model that calculates a battery evaluation score based on multiple first data related to each of the 1-1 detailed item, 1-2 detailed item, 1-3 detailed item, 2-1 detailed item, and 2-2 detailed item.

[0062] Accordingly, the evaluation score calculation unit (240) can calculate the battery evaluation score based on the scores for each of the 1-1 detailed item, 1-2 detailed item, 1-3 detailed item, 2-1 detailed item, and 2-2 detailed item.

[0063] The battery evaluation score may include at least one of the vehicle's battery's SOH score, a state score indicating the condition of the cells included in the battery, and a safety index related to the safety of the battery.

[0064] Meanwhile, the Remaining Useful Life (RUL) can be derived based on when the predicted SOH curve reaches a predefined threshold (e.g., 70% or 60%). Through this, users can predict how many kilometers or months later the battery will need to be replaced.

[0065] In addition, the calculated RUL information can be scored according to certain criteria. For example, it can be set so that 100 points represent the initial state and 0 points represent a state requiring replacement. In addition to simple linear methods, the scoring algorithm can utilize curved depreciation functions (e.g., log / exponential functions) to enhance user perception. For instance, intuitive information can be provided to the vehicle driver via a mobile application or dashboard in the form of "Battery health score 84 points, usable for approximately 2.5 years," and vehicle manufacturers, leasing companies, insurance companies, etc., can utilize this information for warranty management, lease recovery determination, and insurance rate adjustments.

[0066] A status score indicating the condition of the cells included in the battery may be a score calculated based on battery cell voltage deviation, battery module temperature deviation, battery pack temperature, or battery pack current, etc. For example, the status score may be expressed as a score such as 0 to 100 points.

[0067] A safety index related to battery safety may be an index representing a safety rating of the battery. For example, the safety index may be calculated based on scores related to sudden acceleration and sudden braking of the vehicle. Data related to sudden acceleration and sudden braking may be calculated based on vehicle speed data. For example, the safety index may be expressed as Grade A, Grade B, or Grade C.

[0068] SOH is an indicator representing the remaining lifespan of a battery, which can be calculated by comparing the initial battery state with the current battery state. Additionally, SOH can be calculated based on the accumulated charge amount or accumulated mileage. SOH can be expressed as a percentage. For example, an SOH of 90% may mean that the battery retains only 90% of its performance from when it was first manufactured. The battery evaluation score calculated through an evaluation model may include SOH.

[0069] Below, a method for calculating SOH in the battery residual value evaluation unit (210) is described.

[0070] The battery residual value evaluation unit (210) can calculate the SOH (%) value using the ratio of the current available capacity to the initial rated capacity of the battery, and the formula is as follows.

[0071]

[0072] Here, Q0 is the initial rated capacity, meaning the maximum charge capacity measured under standard conditions immediately after battery manufacturing, and Q max represents the maximum chargeable capacity of the battery at the time of measurement, based on the currently available capacity.

[0073] The initial rated capacity and current available capacity can be calculated based on the amount of electricity (Wh) or the amount of electric charge (Ah). However, the method of calculating the SOH value in the battery residual value evaluation unit (210) is not limited to Equation 1, and methods such as multiplying the ratio of the current available capacity to the initial rated capacity by a proportionality constant or adding other constants to the above formula may be applied.

[0074] The initial rated capacity of the battery can be obtained using data collected from the vehicle specifications collection unit (110) or the battery information collection unit (120). The current available capacity of the battery can be calculated by the battery residual value evaluation unit (210).

[0075] Figures 3 and 4 are graphs showing the current available capacity of the battery according to changes in SOC in terms of energy (Wh) and charge (Ah), respectively, and compare the results under two average temperature conditions. Figures 5 and 6 are graphs showing the current available capacity of the battery according to changes in ambient temperature in terms of energy (Wh) and charge (Ah), respectively.

[0076] Below, the relationship between the ambient temperature and the current available capacity of the battery will be explained using Figures 3 to 6.

[0077] Referring to the graph in Figure 3, the current available capacity of the battery according to the change in SOC is calculated in units of electrical energy (Wh), and the values ​​when the ambient temperature is 20℃ or higher are shown in blue, and the values ​​when it is less than 10℃ are shown in red.

[0078] In one embodiment, the battery residual value evaluation unit (210) can calculate the current available capacity of the battery in units of power (Wh) through the following mathematical formula.

[0079]

[0080] Here, Q max(Wh) represents the current available capacity of the battery, Q(Wh) represents the current charge amount of the battery, and SOC(%) represents the charge state, and values ​​collected from the battery information collection unit (120) can be used.

[0081] When calculating the current available capacity of the battery in units of electrical energy (Wh), it can be seen that the value (indicated in blue) when the ambient temperature is 20°C or higher is generally higher than the value (indicated in red) when the ambient temperature is less than 10°C. The graph in Figure 5 shows the current available capacity of the battery calculated in units of electrical energy (Wh) according to changes in ambient temperature, and it can be seen that the current available capacity increases as the ambient temperature increases.

[0082] Although the current available capacity of a battery should be measured as constant regardless of ambient temperature, it has been confirmed that when calculated in units of electrical energy, variations occur depending on the ambient temperature. Therefore, when calculating SOH by determining the battery's current available capacity in units of electrical energy (Wh), deviations in the value due to changes in ambient temperature may lead to reliability issues.

[0083] Referring to the graph in Figure 4, the current available capacity of the battery according to the change in SOC is calculated in units of electric charge (Ah), and the values ​​when the ambient temperature is 20℃ or higher are shown in blue, and the values ​​when it is less than 10℃ are shown in red.

[0084] In one embodiment, the battery residual value evaluation unit (210) can calculate the current available capacity of the battery in units of electric charge (Wh) through the following mathematical formula.

[0085]

[0086] Here, Q max(Ah) represents the current available charge of the battery, Q(Wh) represents the current charging power of the battery, V represents the battery voltage, and SOC(%) represents the charging state, and values ​​collected from the battery information collection unit (120) can be used. The current available capacity can be estimated based on data such as power, voltage, and temperature measured in a specific SOC range.

[0087] When calculating the current available capacity of the battery in units of electric charge (Ah), it can be seen that there is no difference between the value when the ambient temperature is 20°C or higher (indicated in blue) and the value when it is less than 10°C (indicated in red) (t-test result for temperature difference p-value = 0.7). The graph in Figure 6 shows the current available capacity of the battery calculated in units of electric charge (Ah) according to changes in ambient temperature, and it can be seen that the current available capacity is not significantly related to the ambient temperature.

[0088] The formula for calculating total battery capacity in Wh does not consider battery voltage, which is interpreted to be because battery voltage is significantly affected by ambient temperature. Therefore, to calculate SOH while minimizing the influence of temperature, it is advisable to calculate the current available capacity in Ah (electric charge) rather than Wh. In other words, converting both the initial rated capacity and the current available capacity to Ah during SOH calculation can improve the accuracy of the estimate.

[0089] Accordingly, the battery residual value evaluation unit (210) can estimate the current available capacity of the battery in units of charge (Ah) using the current charge amount (Wh) and battery voltage of the battery at a specific SOC.

[0090] Referring to Fig. 7, it can be seen that as the SOC increases, the battery current available capacity (Battery capa) value calculated by the battery residual value evaluation unit (210) tends to decrease as the SOC increases.

[0091] Therefore, a specific SOC range (e.g., 86 to 88%) with low correlation between SOC and total battery capacity can be fixed, and the rate of decrease over time can be used to calculate the SOH indicator.

[0092] Meanwhile, since the battery's current available capacity (Battery capa) is sensitive to measurement errors or changes in the external environment (e.g., temperature, load conditions, etc.), errors may occur in the estimated value even if the SOC range is limited to a specific area.

[0093] Referring to Figures 8 to 11, when calculating the current available battery capacity (Battery capa) by limiting the external temperature to a specific range without specifying the SOC range, a decreasing trend in the current available capacity over time was confirmed.

[0094] Specifically, by expanding the SOC range to an area where the deterioration effect is clearly evident, such as from 20% to 80% or from 30% to 90%, and analyzing only the data in that range, it is possible to confirm that the SOH estimate value decreases over time. Therefore, in the present invention, a method of estimating the SOH value in the above manner can be used by expanding the range of the SOC range to 20% to 80% or from 30% to 90%.

[0095] Furthermore, this technology is not limited to calculating SOH at a specific single point in time; by applying periodic or cumulative-based SOH calculation models, it is possible to monitor the trend of SOH reduction over time in real time.

[0096] Meanwhile, since a greater voltage drop occurs as the battery degrades for every equal decrease in capacity (charge), the SOH value can be estimated by comparing the voltage drop rate relative to the charge discharge rate in a specific discharge range. This can be expressed mathematically as follows.

[0097]

[0098] Here, represents the voltage drop rate and represents the electric charge discharge rate.

[0099] Referring to FIGS. 12 and 13, the battery residual value evaluation unit (210) can predict SOH by using at least one of the following data collected from the vehicle data collection unit of the evaluation data collection device (100) as an independent variable, such as the total driving distance of the electric vehicle, battery temperature, ambient temperature, maximum voltage among all cells, and initial rated capacity of the battery, and applying a regression model with SOH as the dependent variable. As a result of applying the regression model for predicting SOH, R based on the Random Forest Regression model 2 A prediction performance of 0.96 was achieved. Therefore, the trained model can be used to use the predicted SOH value as a surrogate for data where SOH is not collected.

[0100] Here, in addition to Random Forest Regression, the regression model can learn past SOH trend data through time series-based or machine learning-based regression analysis techniques and use it to predict future SOH reduction curves. Specifically, the battery residual value evaluation unit (210) can calculate the SOH of the battery using at least one regression model among various regression models such as polynomial regression, LSTM-based time series regression, gradient boosting regression, linear regression, and decision tree regression.

[0101] Referring to Figures 14 to 16, the analysis of the SHAP value revealed that the total distance and expected battery capacity have a positive correlation with SOH, while the remaining variables—outside temperature, battery temperature max / min, and cell max / min—have a negative correlation.

[0102] Among them, the temperature-related variable showed a negative correlation with SOH, suggesting that there is a dependency on time, as SOH decreases as temperature increases over time during the collection period (approximately January to April). Meanwhile, since the temperature-related variable showed high values ​​in both variable importance and SHAP value, there is room to suspect overfitting.

[0103] A similar relationship between SOH and temperature can also be confirmed through the Partial Dependence Plot.

[0104] This allows us to verify whether the model is actually overfitted by comparing data collected from all seasons of the year, and if there is an actual tendency toward overfitting, it can be resolved by collecting data over a long period of time.

[0105] Although preferred embodiments of the present invention have been illustrated and described above, the present disclosure is not limited to the specific embodiments described above. It is understood that various modifications can be made by those skilled in the art without departing from the essence of the invention as claimed in the claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention. Explanation of the symbols

[0106] 1: Electric vehicle 10: Electric Vehicle Battery Performance Evaluation and Residual Value Prediction System 100: Evaluation data acquisition device 110: Vehicle Specifications Collection Department 120: Battery Information Collection Unit 130: Vehicle Data Collection Unit 200: Battery Evaluation Server 210: Battery Residual Value Assessment Department 220: Data Classification Section 230: Item-by-item score calculation unit 240: Evaluation Score Calculation Department

Claims

Claim 1 An electric vehicle battery performance evaluation and residual value prediction system comprising: an evaluation data collection device for collecting evaluation data to evaluate the performance and predict the residual value of a battery installed in an electric vehicle; and a battery evaluation server for receiving the evaluation data by communicating with the evaluation data collection device and evaluating the performance of the battery; wherein the evaluation data includes the initial rated capacity (Ah) of the battery, the current charging power of the battery, the State of Charge (SOC) of the battery, and voltage data of the battery; and wherein the battery evaluation server includes a battery residual value evaluation unit that calculates the SOH of the battery based on the evaluation data; wherein the battery residual value evaluation unit calculates the current available capacity of the battery in units of electric charge (Ah) using the current charging power of the battery, the SOC of the battery, and the voltage of the battery, calculates the SOH of the battery based on the ratio of the current available capacity (Ah) to the initial rated capacity (Ah), and calculates the SOH of the battery only when the SOC of the battery is in the range of 20 or more and 80 or less. Claim 2 delete Claim 3 delete Claim 4 In claim 1, the evaluation data collection device further includes a vehicle data collection unit that collects at least one of the battery usage environment data and the electric vehicle operation data, and the battery residual value evaluation unit calculates the SOH of the battery using at least one regression model among Random Forest Regression, Polynomial Regression, LSTM-based Time Series Regression, Gradient Boosting Regression, Linear Regression, and Decision Tree Regression, using at least one of the data collected by the vehicle data collection unit, the total driving distance of the electric vehicle, the temperature of the battery, and the ambient temperature of the electric vehicle as independent variables. Claim 5 In claim 1, the battery evaluation server further comprises an item-specific score calculation unit that calculates a reliability score for the SOH of the battery calculated by the battery residual value evaluation unit, an electric vehicle battery performance evaluation and residual value prediction system. Claim 6 delete Claim 7 A method for evaluating electric vehicle battery performance and predicting residual value, comprising: a first step in which an evaluation data collection device installed in an electric vehicle collects data on the initial rated capacity (Ah) of a battery; a second step in which the evaluation data collection device collects the current charging power of the battery, the SOC of the battery, and the voltage of the battery; and a third step in which a battery residual value evaluation unit receives data on the initial rated capacity (Ah), the current charging power of the battery, the SOC of the battery, and the voltage of the battery from the evaluation data collection device and calculates the SOH of the battery; wherein in the third step, the current available capacity of the battery is calculated in units of electric charge (Ah) using the current charging power of the battery, the SOC of the battery, and the voltage of the battery, and the SOH of the battery is calculated based on the ratio of the current available capacity (Ah) to the initial rated capacity (Ah), and the SOH of the battery is calculated only when the SOC of the battery is in the range of 20 or more and 80 or less.

Images