A method for estimating driving range based on unit charge range and battery aging correction.

Abstract

The driving range estimation method based on unit charge range and battery aging correction mainly includes four steps: online battery SOC estimation, periodic battery SOH update, unit SOC range estimation, and driving range estimation. This invention uses periodic battery SOH estimation to correct the battery SOC estimation, and combines unit SOC range estimation to achieve accurate driving range estimation. It has the advantages of simple model structure, low computational load, high estimation accuracy, and easy implementation in vehicle BMS. Under normal driving conditions, the maximum driving range estimation error of this invention is within 3km, which meets the operating requirements of electric vehicles.

Description

Technical Field

[0001] This invention relates to the field of battery management technology for electric vehicles, and in particular to a method for estimating driving range based on unit charge range and battery aging correction. Background Technology

[0002] Despite the rapid development of electric vehicles in recent years, the problem of inaccurate range estimation remains severe, significantly impacting the driving experience and hindering the widespread adoption of electric vehicles. Accurate and real-time range estimation under real-world driving conditions can help alleviate "range anxiety" among drivers and passengers, boosting confidence in the further promotion of electric vehicles.

[0003] Currently, practical driving range estimation is generally calculated using the remaining battery energy and the average energy consumption per kilometer under NEDC conditions. Patent application CN202011375247.3 discloses a method for calculating electric vehicle driving range based on driving condition identification, which uses an initial mileage determined by NEDC conditions and then corrected by real-time driving condition identification. However, in reality, the error in estimating the remaining battery energy accumulates continuously during the integration process, and due to the complex and variable driving conditions in actual driving, using a constant average energy consumption per kilometer cannot guarantee estimation accuracy. On the other hand, some studies have used machine learning methods to calculate energy consumption from driving data segments, but this requires cumbersome preprocessing steps for training data, resulting in high costs, computational complexity, and high development technical requirements, making it difficult to apply in practice. Furthermore, battery aging significantly affects battery performance, and current driving range estimation methods do not consider battery aging. The invention disclosed in patent application CN202210689260.9, which predicts the driving range of an electric vehicle, uses an adjustment method and related device based on the ambient temperature of the electric vehicle to correct the remaining battery charge, thereby predicting the vehicle's driving range. However, this method only considers the current ambient temperature and ignores the battery's lifespan for energy correction, which is clearly inaccurate. Therefore, there is a need for a driving range estimation method that is simple in structure, requires less computation, and considers the impact of battery aging to improve estimation accuracy. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, the present invention aims to provide a driving range estimation method based on unit charge range and battery aging correction. This method is based on unit charge range and battery aging correction, takes into account the impact of battery aging, has low computational load, and is easy to apply in practice.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] The driving range estimation method based on unit charge range and battery aging correction includes the following steps:

[0007] Step (1), online estimation of battery SOC

[0008] First, an equivalent circuit model is established, and the model parameters are identified online using a recursive algorithm based on measured voltage and current data. Then, the dynamic battery model is combined with a Kalman filter algorithm to complete the online SOC estimation.

[0009] Step (2), battery SOH is updated periodically.

[0010] First, a cyclic aging experiment is conducted on the battery to collect a battery aging dataset. Then, based on the constant current charging process of the battery, health indicator features strongly correlated with SOH are extracted, a machine learning model is selected for training and saved. Finally, at fixed date nodes, the trained model is called to estimate the current SOH of the battery, completing the periodic update of the battery SOH.

[0011] Step (3), Unit SOC travel estimation

[0012] The distance traveled per unit SOC is calculated as an empirical value based on historical driving data; the distance traveled per unit SOC is calculated based on current trip data; finally, the two sets of information are fused to obtain the final estimated trip value per unit SOC.

[0013] Step (4), estimated driving range

[0014] First, at a pre-set update time node, the current SOH value of the vehicle battery system is updated using the method described in step (2); during the vehicle's journey, the SOC of the vehicle is estimated in real time using the method described in step (1), and the SOC estimate is corrected using the SOH value; then, based on the current journey data, the current unit SOC journey estimate is calculated using the method described in step (3); finally, the product of the SOC estimate and the SOC journey estimate is the final estimated driving range.

[0015] The specific steps of step (1) include:

[0016] Step (1.1): Establish the equivalent circuit model of the battery; in the established first-order RC equivalent circuit model, E0 is the open-circuit voltage, R1 and R2 are the ohmic internal resistance and polarization internal resistance, respectively, C2 is the polarization capacitor, V0 and V2 are the battery terminal voltage and polarization voltage, respectively, and I is the current. The space state equation of this model is:

[0017]

[0018] in:

[0019] Step (1.2): Online identification of model parameters; Based on the battery voltage and current data collected in real time by the sensors built into the vehicle, the recursive least squares method is used to identify the model parameters online, and a dynamic battery equivalent model updated in real time is obtained.

[0020] Step (1.3) Online SOC estimation: Based on the dynamic battery equivalent model, the battery voltage and current data collected by the sensor are used as input, and the extended Kalman filter algorithm is used to realize online SOC estimation.

[0021] The specific steps of step (2) include:

[0022] Step (2.1): Collect battery aging data; conduct cyclic aging experiments on the battery, record the voltage and current data of each aging cycle, and form a battery aging dataset;

[0023] Step (2.2) Battery health feature extraction: Based on the constant current charging process of the aging dataset, analyze the evolution law of voltage and current data and battery aging, and extract health features strongly correlated with battery aging, including charging time and average voltage features at a fixed voltage range.

[0024] Step (2.3): Data-driven model training; Select the support vector machine data-driven model, establish the mapping relationship between the extracted health features and the battery SOH, train the model offline and save it;

[0025] Step (2.4) Periodic update of battery SOH: The update cycle is set in advance. At the time node that needs to be updated, the current SOH of the battery is estimated by the offline trained model based on the charging process of the battery, and the periodic update of battery SOH is completed.

[0026] The specific steps of step (3) include:

[0027] Step (3.1): Calculate the empirical value of each unit SOC trip; based on the vehicle's historical travel data, calculate the average distance traveled per unit SOC, which is used as the empirical value of the unit SOC trip. past Regularly updated;

[0028] Step (3.2): Calculate the distance traveled per unit SOC; based on the travel data of this trip, calculate the distance traveled per unit SOC. now ;

[0029] Step (3.3), Unit SOC travel estimation; The unit SOC travel distances from steps (3.1) and (3.2) are fused to obtain the final unit SOC travel estimate. The calculation method is as follows:

[0030] l=α·l past +(1-α)lnow

[0031] Where α is the weight of the unit SOC travel.

[0032] Compared with the prior art, the present invention, by adopting the above technical solution, has the following technical effects:

[0033] (1) The Kalman filter algorithm is used for online SOC estimation, which has a fast convergence speed and a maximum estimation error of less than 2.5%. A machine learning model is used for online SOH estimation, thereby correcting the online SOC estimate and overcoming the impact of battery capacity decay caused by battery aging on the accuracy of SOC estimation. This ensures the accuracy of SOC estimation throughout the battery's life cycle and further guarantees the accuracy of subsequent driving range estimation.

[0034] (2) The driving range estimation method provided by this invention has a maximum estimation error of less than 3km under normal driving conditions, and the estimation error at the end of the trip is kept within 1km. The overall estimation accuracy is high, which greatly alleviates the user's "range anxiety".

[0035] This invention estimates the vehicle's driving range by using the current SOC and historical unit SOC of the battery, and periodically performs battery SOH estimation to correct the SOC estimation error caused by battery capacity decay. It features a simple model structure, low computational load, high estimation accuracy, and is easy to implement in an on-board BMS. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating the electric vehicle range estimation method based on unit charge travel and battery aging correction according to the present invention.

[0037] Figure 2 This is a schematic diagram of the first-order RC equivalent circuit model established in step (1) of the present invention.

[0038] Figure 3 The figure shows the results of estimating the SOC using the extended Kalman filter algorithm in the verification case of this invention.

[0039] Figure 4 This is a graph showing the battery SOH estimation results in the verification case of this invention.

[0040] Figure 5(a) shows the range estimation results of the electric vehicle range estimation method based on unit charge range and battery aging correction of the present invention; Figure 5(b) shows the estimation error. Detailed Implementation

[0041] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.

[0042] This invention can be implemented in many different forms and should not be considered limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully express the scope of the invention to those skilled in the art.

[0043] like Figure 1 As shown, this invention provides a method for estimating the driving range of electric vehicles based on unit charge range and battery aging correction, including the following steps:

[0044] Step 1), Online SOC estimation of the battery. First, an equivalent circuit model is established, and based on measured voltage and current data, a recursive algorithm is used to identify the model parameters online. Then, the dynamic battery model is combined with a filtering algorithm to complete the online SOC estimation.

[0045] The specific implementation method is as follows:

[0046] Step (1.1): Establish the equivalent circuit model of the battery. The established first-order RC equivalent circuit model is as follows: Figure 2 As shown in the diagram. Here, E0 is the open-circuit voltage, R1 and R2 are the ohmic internal resistance and polarization internal resistance, respectively, C2 is the polarization capacitance, V0 and V2 are the battery terminal voltage and polarization voltage, respectively, and I is the current. The space state equation of this model is:

[0047]

[0048] in:

[0049] Step (1.2) Online identification of model parameters. Based on the battery voltage and current data collected in real time by the vehicle's built-in sensors, the model parameters are identified online using the recursive least squares method to obtain a dynamic battery equivalent model that is updated in real time.

[0050] Step (1.3) Online SOC estimation. Based on the dynamic battery equivalent model, using the battery voltage and current data collected by the sensor as input, the extended Kalman filter (EKF) algorithm is used to achieve online SOC estimation.

[0051] Step (2): Periodically update the battery SOH. First, conduct a cyclic aging experiment on the battery to collect the battery aging dataset; then, extract health indicator features that are strongly correlated with SOH based on the constant current charging process of the battery, select a suitable data-driven model for training and save it; finally, call the trained model at fixed date nodes to estimate the current battery SOH, and complete the periodic update of the battery SOH.

[0052] The specific implementation method is as follows:

[0053] Step (2.1): Collect battery aging data; conduct cyclic aging experiments on the battery, record the voltage and current data of each aging cycle, and form a battery aging dataset.

[0054] Step (2.2) Battery health feature extraction: Based on the constant current charging process of the aging dataset, analyze the evolution of voltage and current data with battery aging, and extract health features strongly correlated with battery aging, including charging time and average voltage features at a fixed voltage range.

[0055] Step (2.3): Data-driven model training; Select the support vector machine data-driven model, establish the mapping relationship between the extracted health features and the battery SOH, train the model offline and save it.

[0056] Step (2.4) Regularly update battery SOH: Pre-set the update cycle (e.g., monthly). At the time node that needs to be updated, based on the battery charging process, use the offline trained model to estimate the current SOH of the battery to complete the regular update of battery SOH.

[0057] Step (3), unit SOC trip estimation. The distance traveled per unit SOC is calculated as an empirical value based on historical driving data; the distance traveled per unit SOC is calculated based on current trip data; finally, the two information are fused to obtain the final unit SOC trip estimate.

[0058] The specific implementation method is as follows:

[0059] Step (3.1): Calculate the empirical value of each unit SOC trip; based on the vehicle's historical travel data, calculate the average distance traveled per unit SOC, which is used as the empirical value of the unit SOC trip. past Regularly updated;

[0060] Step (3.2): Calculate the distance traveled per unit SOC; based on the travel data of this trip, calculate the distance traveled per unit SOC. now ;

[0061] Step (3.3), Unit SOC travel estimation; The unit SOC travel distances from steps (3.1) and (3.2) are fused to obtain the final unit SOC travel estimate. The calculation method is as follows:

[0062] l=α·l past +(1-α)l now

[0063] Where α is the weight of the unit SOC travel.

[0064] Step (4), driving range estimation. Based on the battery's SOH periodically corrected during the charging process, the final driving range estimate is calculated using real-time online SOC estimation combined with the current unit SOC range estimate.

[0065] The specific implementation method is as follows: According to step (2), the current SOH of the battery is updated at the specified time node; during the vehicle travel process, the battery SOC is estimated online in real time by step (1); according to step (3), the unit SOC range estimate is calculated; finally, the real-time SOC estimate is multiplied by the unit SOC range estimate to obtain the final driving range estimate result.

[0066] In summary, this invention discloses a method for estimating the driving range of electric vehicles based on unit charge range and battery aging correction. It mainly includes four steps: online estimation of battery state of charge (SOC), periodic updating of battery state of health (SOH), unit SOC range estimation, and driving range estimation. The driving range estimation method disclosed in this invention first estimates the battery's SOC, then uses periodic battery SOH estimation to correct for SOC estimation errors caused by battery capacity decay, and combines this with unit SOC range estimation to achieve accurate driving range estimation. This method has a simple model structure, low computational cost, high estimation accuracy, and is easy to implement in an onboard BMS. This invention can replace existing driving range estimation techniques in real-world scenarios, helping to alleviate users' range anxiety and facilitating industrial promotion and application.

[0067] To verify the feasibility and effectiveness of this method, data from a single full-trip journey of a real electric motorcycle was used. During this journey, the electric motorcycle traveled from a fully charged state to a fully discharged state.

[0068] Figure 3 The figure shows the results of online SOC estimation using the extended Kalman filter algorithm in the verification case of this invention. As can be seen from the figure, the estimated value of the method can track the true SOC value very well, and the SOC estimation error remains within 2.5% throughout the entire operation of the vehicle.

[0069] like Figure 4 The figure shows the SOH estimation results based on support vector machine in the verification case of this invention. As can be seen from the figure, the proposed method can accurately estimate the current SOH value of the battery throughout its entire lifespan (SOH>70%), with an average absolute estimation error of only 0.81%, achieving accurate SOH estimation. Regularly updated SOH (aging correction) ensures accurate SOC estimation, further guaranteeing the accuracy of subsequent driving range estimation.

[0070] Figure 5 shows the range estimation results of the electric vehicle range estimation method based on unit charge range and battery aging correction according to the present invention. Figure 5(a) shows that the estimated value can track the actual driving range well, especially in the initial stage and the last segment of the vehicle's journey. The error diagram in Figure 5(b) shows that the maximum estimation error is only 3km and the average absolute estimation error is 1.15km throughout the entire driving process, meeting practical requirements. It is worth noting that in the last 20 minutes of the vehicle's journey, the range estimation error remains within 1km, which greatly enhances the driver's confidence and alleviates the user's "range anxiety".

[0071] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0072] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for estimating driving range based on unit charge range and battery aging correction, characterized in that, Includes the following steps: Step (1), online estimation of battery SOC First, an equivalent circuit model is established, and the model parameters are identified online using a recursive algorithm based on measured voltage and current data. Then, the dynamic battery model is combined with a Kalman filter algorithm to complete the online SOC estimation. Step (2), battery SOH is updated regularly. First, a cyclic aging experiment is conducted on the battery to collect the battery aging dataset; then, a support vector machine data-driven model is selected for training and saved; finally, the trained model is called at fixed date nodes to estimate the current SOH of the battery, thus completing the periodic update of the battery SOH. Step (3), Unit SOC travel estimation The distance traveled per unit SOC is calculated as an empirical value based on historical driving data; the distance traveled per unit SOC is calculated based on current trip data; finally, the two sets of information are fused to obtain the final estimated trip value per unit SOC. Step (4), estimated driving range First, at a pre-set update time node, the current SOH value of the vehicle battery system is updated using the method described in step (2); during the vehicle's journey, the SOC of the vehicle is estimated in real time using the method described in step (1), and the SOC estimate is corrected using the current SOH value; then, based on the current journey data, the current unit SOC journey estimate is calculated using the method described in step (3); finally, the product of the SOC estimate and the SOC journey estimate is the final estimated driving range. The specific steps of step (3) include: Step (3.1): Calculate the empirical value of each unit SOC trip; based on the vehicle's historical travel data, calculate the average distance traveled per unit SOC as the empirical value of each unit SOC trip. Regularly updated; Step (3.2): Calculate the distance traveled per unit SOC; based on the travel data of this trip, calculate the distance traveled per unit SOC. ; Step (3.3), Unit SOC travel estimation; The unit SOC travel distances from steps (3.1) and (3.2) are fused to obtain the final unit SOC travel estimate. The calculation method is as follows: in To calculate the weight of a unit SOC travel.

2. The driving range estimation method based on unit charge range and battery aging correction according to claim 1, characterized in that: The specific steps of step (1) include: Step (1.1): Establish the equivalent circuit model of the battery; in the established first-order RC equivalent circuit model, It is open-circuit voltage. , These are the ohmic internal resistance and the polarization internal resistance, respectively. It is a polarized capacitor. , These are the battery terminal voltage and polarization voltage, respectively. It is current, and the space state equation of this model is: in: ; ; ; ; ; ; ; Step (1.2): Online identification of model parameters; Based on the battery voltage and current data collected in real time by the sensors built into the vehicle, the recursive least squares method is used to identify the model parameters online, and a dynamic battery equivalent model updated in real time is obtained. Step (1.3) Online SOC estimation: Based on the dynamic battery equivalent model, the battery voltage and current data collected by the sensor are used as input, and the extended Kalman filter algorithm is used to realize online SOC estimation.

3. The driving range estimation method based on unit charge range and battery aging correction according to claim 1, characterized in that: The specific steps of step (2) include: Step (2.1): Collect battery aging data; conduct cyclic aging experiments on the battery, record the voltage and current data of each aging cycle, and form a battery aging dataset; Step (2.2) Battery health feature extraction: Based on the constant current charging process of the aging dataset, analyze the evolution law of voltage and current data and battery aging, and extract health features strongly correlated with battery aging, including charging time and average voltage features at a fixed voltage range. Step (2.3): Data-driven model training; Select a support vector machine data-driven model, establish the mapping relationship between the extracted health features and the battery SOH, train the model offline and save it; Step (2.4) Periodic update of battery SOH: The update cycle is set in advance. At the time node that needs to be updated, the current SOH of the battery is estimated by the offline trained model based on the charging process of the battery, and the periodic update of battery SOH is completed.

Images