Electric vehicle power battery vehicle cloud collaborative management and control method

Abstract

The invention relates to an electric vehicle power battery vehicle cloud collaborative management and control method, and belongs to the technical field of vehicle big data management. A vehicle end battery management system uploads battery data to a cloud end battery management platform according to a set uploading period and uploading field scale, the uploading period and the uploading field scale are adjusted according to needs, and when the cloud end battery management platform needs to update model parameters, the uploading period and the uploading field scale are improved; and the cloud battery management platform processes the received historical battery data of the same vehicle and the battery data of other vehicles of the same type as the vehicle by using the model so as to evaluate the battery state of the vehicle. According to the method, the uploading period and the uploading field scale can be adjusted according to actual requirements, server overload caused by too large data volume can be avoided, and the processing precision of the vehicle end battery management system can be guaranteed.

Description

Technical Field

[0001] The present invention relates to a cloud collaborative management and control method for an electric vehicle power battery vehicle, belonging to the technical field of vehicle big data management. Background Technique

[0002] The performance of the battery is not constant. The performance of the battery will continuously decay with continuous charging and discharging during its life cycle, and the decay rate is related to many factors, such as whether it is overcharged or over-discharged, whether the ambient temperature is appropriate, how long the charging and discharging process has been experienced, etc.; once overcharged or over-discharged, or the temperature is too high or too low, it will seriously affect the battery life and even cause permanent damage. Therefore, it is necessary to strictly control the charging and discharging process during the entire warranty period to avoid overcharging, over-discharging, and overheating. The traditional vehicle-end BMS has the following defects: the management parameters are rigid and cannot accurately follow the battery changes, which is not conducive to the management of the battery during the entire warranty period; limited by storage and computing capabilities, it is unable to utilize the historical data of this vehicle and the data of other vehicles of the same model; the fault diagnosis is relatively basic and cannot apply complex technologies and predictions; it is impossible to implement personalized management strategies.

[0003] 1. The accuracy of the traditional vehicle-end battery management method is poor: Apply this parameter identification method (i.e., offline identification) in the computer to identify the battery model parameters of the battery used in the actual vehicle, and burn the identified parameters into the vehicle-end BMS as the initial identification parameters of the battery model in the vehicle-end BMS. As the vehicle runs, the characteristics of the battery will change, resulting in an increasing error in the SOC estimated by the vehicle-end BMS, leading to inaccurate battery management.

[0004] The existing cloud battery management method has an uncoordinated relationship between data upload density and management accuracy: Use a vehicle-end data upload terminal to periodically upload the specified data of the battery to the cloud server, establish a vehicle operation monitoring platform on the cloud server, and further establish a twin model of the vehicle-end battery model on the vehicle operation monitoring platform and inject the parameters identified offline to achieve cloud management of the battery; however, there is an uncoordinated relationship between data upload density and management accuracy. Theoretically, the higher the data upload density, the higher the management accuracy of the battery; when the uploaded battery data density is high, the data throughput suddenly increases. When the number of connected vehicles reaches 100,000 or a higher level, the data generated every day will reach the TB level, resulting in server overload. When the uploaded battery data density is low, the cloud management model cannot obtain the accurate state of the battery.

[0005] For example, the Chinese patent application document with the application number CN201910162345.X discloses a battery management system and method based on a big data cloud platform. The battery information of multiple users is collected through the big data cloud platform, the main features are extracted through PCA principal feature analysis, and the monitoring model is trained through a neural network. When managing the battery, the battery information is uploaded to extract features and compared with the monitoring model, and the battery is managed based on the comparison result. However, it cannot dynamically adjust the data upload mechanism, nor can it intelligently update the parameters of the twin model. The Chinese patent application document with the application number CN201910656602.5 discloses a power battery pack management system and method based on digital twin. This solution establishes a digital twin system by coupling the physical entity and the virtual entity, uses the twin cloud data platform to analyze the physical entity and the virtual entity through the rolling optimization method, processes the twin cloud data through the cloud computing system to obtain the state of the physical battery pack and the twin virtual battery pack under the full life cycle of the power battery pack, and realizes the full life cycle management of the power battery pack through the vehicle-cloud battery management system interaction and communication. However, in the face of a vehicle scale of hundreds of thousands, this solution has no corresponding solution. Therefore, in the current solutions, the upload cycle and the upload field scale cannot be adjusted, which may lead to too large data volume or relatively low data processing accuracy. Summary of the Invention

[0006] The object of the present invention is to provide a method for collaborative vehicle-cloud control of electric vehicle power batteries to solve the problems of too large data volume or low cloud data processing accuracy caused by the inability to adjust the data upload cycle and upload field scale of the vehicle-side battery management system.

[0007] The present invention provides a method for collaborative vehicle-cloud control of electric vehicle power batteries to solve the above technical problems. The method includes the following steps:

[0008] 1) The vehicle-side battery management system uploads battery data to the cloud battery management platform according to the set upload cycle and upload field scale, and the upload cycle and upload field scale are adjusted as needed. When the cloud battery management platform needs to update the model parameters, the upload cycle and upload field scale are increased;

[0009] 2) The cloud battery management platform uses the model to process the historical battery data of the same vehicle received and the battery data of other vehicles of the same model as this vehicle to evaluate the battery state of this vehicle.

[0010] Furthermore, the model adopted by the cloud battery management platform is a digital twin model, and the parameters of this digital twin model are identified online. During online identification, the parameter identification is converted into an optimization problem, with the minimum of the battery parameters output by the digital twin model and the real battery parameters as the objective function of the optimization problem, and the objective function is solved to obtain the parameters of the digital twin model.

[0011] Furthermore, the digital twin model is a battery model, including an electrochemical P2D model, a fractional-order impedance model, a life model, an integer-order equivalent circuit model, a battery thermal model, and a thermal runaway model.

[0012] Furthermore, the cloud battery management platform is also used to collect user habit data.

[0013] Furthermore, when the cloud battery management platform receives the occurrence of a battery failure fault in a vehicle, the cloud battery management platform is used to analyze the historical battery data of vehicles of the same model, identify the data feature differences between normal vehicles and abnormal vehicles, so as to achieve accurate fault location.

[0014] Furthermore, the cloud battery management platform adopts power battery characteristic parameter outlier detection technology to accurately locate faults by setting multi-segment parallel screening thresholds or by using machine learning models for accurate fault location.

[0015] Furthermore, the vehicle-end battery management system is also used to update battery parameters according to seasons and altitude, and upload them to the cloud battery management platform.

[0016] Furthermore, the cloud battery management platform is also used to formulate management strategies according to the battery state assessment results, and send the management strategies to the vehicle-end battery management system.

[0017] The maximum scale of the upload fields in the upload field scale is to upload all data and upload it according to the first period; the minimum scale of the upload fields is to only upload the current field and voltage field required by the model and upload it according to the second period, and the first period is less than the second period.

[0018] The beneficial effects of the present invention are as follows: The vehicle-end battery management system of the present invention uploads battery data to the cloud battery management platform according to the set upload cycle and upload field scale, and the upload cycle and upload field scale are adjusted as needed. When the cloud battery management platform needs to update the model parameters, the upload cycle and upload field scale are increased; the cloud battery management platform uses the model to process the historical battery data of the same vehicle received and the battery data of other vehicles of the same model as this vehicle to evaluate the battery state of this vehicle. The present invention can adjust the upload cycle and upload field scale according to actual needs, which can not only avoid server overload caused by excessive data volume, but also ensure the processing accuracy of the vehicle-end battery management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a schematic diagram of the principle of the vehicle-cloud collaborative control method for the power battery of an electric vehicle according to the present invention;

[0020] Figure 2 is a schematic diagram of the vehicle-cloud implementation of model parameter identification and precise early warning and control of battery self-maintenance according to the present invention;

[0021] Figure 3 is a flow chart of model parameter identification in the cloud battery management platform according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The following further describes the specific embodiments of the present invention with reference to the drawings.

[0023] The vehicle-end battery management system of the present invention uploads battery data to the cloud battery management platform according to the set upload cycle and upload field scale, and the upload cycle and upload field scale are adjusted as needed. When the cloud battery management platform needs to update the model parameters, the upload cycle and upload field scale are increased; the cloud battery management platform uses the model to process the historical battery data of the same vehicle received and the battery data of other vehicles of the same model as this vehicle to evaluate the battery state of this vehicle. The present invention can adjust the upload cycle and upload field scale according to actual needs, which can not only avoid server overload caused by excessive data volume, but also ensure the processing accuracy of the vehicle-end battery management system.

[0024] Such as Figure 1As shown in the figure, the principle of vehicle-cloud collaborative control adopted by the electric vehicle power battery vehicle cloud collaborative control method is to collect battery-related data from the vehicle-side battery management system (referred to as in-vehicle BMS) and upload it to the cloud battery management platform (referred to as cloud BMS) through Icard. After the cloud BMS gives a targeted solution based on the uploaded battery data, it is then sent to the electronic control unit of the in-vehicle BMS, thereby ensuring the accuracy of battery management and improving the battery life. The present invention adopts a composite battery management architecture of "vehicle-side BMS" + "cloud BMS", and the in-vehicle BMS and the cloud BMS perform task division and complementary advantages. Simply put, the cloud BMS is more inclined to long-term planning and prediction, while the in-vehicle BMS is more inclined to short-term analysis and execution.

[0025] The cloud battery management platform relies on the cloud big data platform and theoretically has unlimited storage and computing capabilities. It can use the historical data of the vehicle life cycle and the battery data of other vehicles of the same model to perform battery algorithm calculations and predictions. At the same time, it can adopt one strategy and one management for each vehicle. Through big data analysis and optimization of both the BMS and user habits, it can periodically identify the battery state and formulate corresponding management strategies. This patent invents a cloud battery management platform that uses the data of the cloud platform (such as the estimated battery capacity information), through a periodic task scheduling and management mechanism, and sends management instructions (such as updating capacity information) to the vehicle-side BMS through the MCD platform, achieving a high-precision and high-security management effect for the entire battery warranty cycle, thereby improving the performance and service life of the battery. The model adopted by the cloud battery management platform is a digital twin model, and the parameters of this digital twin model can be identified online. Here, the digital twin model includes an electrochemical P2D model, a fractional-order impedance model, a life model, an integer-order equivalent circuit model, a battery thermal model, and a thermal runaway model. There is a set of these models on the vehicle and a set on the cloud, running independently. The input of the model is the current, voltage, and temperature information collected during vehicle operation; the output of the model is the battery SOC, terminal voltage, SOH, internal temperature of the battery cell, battery safety status, and other information calculated by the model.

[0026] The vehicle-end battery management system is based on a 4G / 5G vehicle-cloud interactive high-speed information channel and a human-machine remote intelligent interaction platform. In order to improve the estimation management accuracy while reducing the data upload volume and calculation amount, an estimation model is designed at multiple scales. Different-scale estimation models require data at different scales, and the multiple scales are reflected in two aspects: the data upload cycle and the data upload fields. The upload cycle and upload field scales are adjusted as needed. When the cloud battery management platform needs to update the model parameters, the upload cycle and upload field scales are increased. For example, the upload cycle changes from 0.1 second to 20 seconds. High-density data is required when updating the model parameters, low-density data is required during the sleep period, and normal data is required during other periods. The maximum scale of the upload fields is to upload all the data and upload it at a cycle of 0.1 second, such as when the cloud performs parameter identification; however, the data transmission volume is large, the cost is high, and the vehicle's power consumption is high. The minimum scale of the upload fields is to only upload the voltage and current fields required by the model, and the upload cycle is 20 seconds, which can meet the monitoring needs of the safety model and save traffic and power consumption; that is, the scale of the uploaded fields changes with the needs of the cloud model.

[0027] The data uploaded at the maximum scale of the upload fields ranges from the basic total battery voltage, temperature, current to the voltages, temperatures, voltage maximum values and their numbers, temperature maximum values and their numbers, current maximum values and their branch numbers, system insulation resistance, system cell voltage difference (V), battery high-voltage positive insulation resistance, battery high-voltage negative insulation resistance, charging current limit value, discharging current limit value, battery system rated capacity, battery system rated energy, BMS-controlled charging current, battery management system status, ON-gear signal, vehicle speed, standby time, standby duration, vehicle position dimension, ambient temperature, humidity, air pressure, humidity inside the box, air pressure, concentration of organic compounds, etc.

[0028] The model adopted by the cloud battery management platform is a digital twin model, and the parameters of this digital twin model can be identified online. Parameter identification is to find a set of parameters to make the voltage output by the digital twin model as close as possible to the voltage of the real battery. Therefore, the present invention transforms the parameter identification problem into an optimization problem, and a set of solutions to the optimization problem is a set of estimated parameters.

[0029] The optimization problem includes design parameters x, objective function F(x), parameter boundary constraints, and constraint functions. The optimization problem solver adjusts the parameter x within the constraint range to meet the specified objective function. Taking the first-order RC circuit model as an example, the model parameters Em, R0, R1, and C1 are all stored in a two-dimensional table and change with temperature and SOC. At a certain temperature, SOC can be divided into 10 segments. According to the charge and discharge curve at this temperature, a set of battery parameters is estimated using the least squares parameter estimation algorithm. The estimation process Figure 3As shown in the figure. Online identification means that after the cloud platform receives the battery operation data, when the difference between the collected voltage and the voltage output by the battery model reaches a certain value, automatic correction is triggered, and battery parameter identification is performed in the cloud. After the battery model parameter identification is completed, it is sent to the vehicle end to update the battery model parameters in the vehicle-end battery management system (BMS).

[0030] The vehicle-end battery management system is also used to update battery parameters according to seasons and altitude and upload them to the cloud battery management platform. For example, when the vehicle drives to a high-altitude area, it is necessary to pay attention to whether the air pressure in the battery box is the same as the outside air pressure. If not, the explosion-proof valve needs to be diagnosed. The updated battery parameters are the parameters of the battery model. Taking the first-order RC circuit model as an example, the model parameters are Em, R0, R1, and C1.

[0031] The cloud battery management platform uses the above model to process the historical battery data of the same vehicle received and the battery data of other vehicles of the same model as this vehicle to evaluate the battery state of this vehicle. When the cloud battery management platform receives a battery failure fault of a vehicle, the cloud battery management platform is used to analyze the historical battery data of vehicles of the same model, identify the data feature differences between normal vehicles and abnormal vehicles, so as to achieve accurate fault location. The cloud battery management platform adopts the outlier detection technology of power battery characteristic parameters and performs accurate fault location by setting multi-segment parallel screening thresholds or uses a machine learning model for accurate fault location.

[0032] The cloud battery management platform downloads data acquisition instructions, diagnostic instructions, equalization control instructions, calibration instructions, and model parameter updates as needed; develops a remote upgrade technology for battery management software, and actively updates the vehicle-end BMS software through the air interface of mobile communication, meeting the wireless upgrade of the vehicle-end battery management model and parameters; with the help of a remote monitoring platform and a big data fault warning platform, develops a BMS cloud MCD intelligent control platform. Through this platform, remote measurement (key data such as BMS operation status), calibration (cell parameters, fault thresholds, etc.), and diagnosis (loop interlock, insulation detection fault auxiliary location analysis, etc.) of the vehicle-end BMS are carried out, and finally, active and efficient early warning, remote online processing, and non-intrusive services of BMS system faults are realized, improving the comprehensiveness and convenience of BMS diagnosis and management, improving after-sales maintenance efficiency, and reducing after-sales maintenance costs. Calibration means sending the battery model parameters identified by the cloud platform to the vehicle-end BMS to update the model parameters of the vehicle-end BMS. For example Figure 2As shown in the figure, the BMS model at the vehicle end and the BMS model in the cloud are a pair of twin models; the cloud BMS and the vehicle-end BMS can run multiple pairs of battery twin models. When the cloud BMS receives a problem with the battery of a certain vehicle, such as a fault of abnormal capacity attenuation, abnormal increase in resistance, or abnormal increase in temperature, it will extract the data characteristics of the faulty battery before and after the fault according to the corresponding model and the data phenomenon of the battery at the fault moment. Then, based on this feature, it will identify the faults of the batteries of the same batch of vehicles, and through multi-algorithm fusion SOC prediction, predict the potential failure trend and failure point; and through the vehicle-cloud data channel, issue control instructions to the BMS end to perform self-repair on the battery. Therefore, the cloud battery management platform of the present invention can perform battery failure prediction, potential abnormal fault perception, self-maintenance management, and intelligent management based on the battery data twin model and multi-algorithm fusion SOC prediction means, and can send the failure prediction results and potential abnormal faults to the in-vehicle BMS for active safety control by the vehicle end.

[0033] By using the coordinated cooperation between the cloud battery management platform and the in-vehicle BMS, the present invention can achieve a high-precision and high-safety management effect throughout the battery full warranty period, thereby improving the performance and service life of the battery. Moreover, the present invention can adjust the upload cycle and upload field scale according to actual needs, which can not only avoid server overload caused by excessive data volume but also ensure the processing accuracy of the vehicle-end battery management system.

Claims

1. A cloud collaborative control method for an electric vehicle power battery vehicle, characterized in that, The method includes the following steps: 1) The on-vehicle battery management system uploads battery data to the cloud battery management platform according to the set upload period and upload field scale, where the upload period and upload field scale are adjusted as needed. When the cloud battery management platform needs to update the model parameters, the upload period and upload field scale are increased; 2) The cloud battery management platform uses the model to process the historical battery data of the same vehicle received and the battery data of other vehicles of the same model as this vehicle to evaluate the battery state of this vehicle.

2. The cloud collaborative control method for an electric vehicle power battery vehicle according to claim 1, wherein, The model adopted by the cloud battery management platform is a digital twin model, and the parameters of the digital twin model are identified online. When identifying the parameters online, the parameter identification is converted into an optimization problem, and the minimum of the battery parameters output by the digital twin model and the real battery parameters is used as the objective function of the optimization problem, and the objective function is solved to obtain the parameters of the digital twin model.

3. The method for cloud collaborative management and control of an electric vehicle power battery vehicle according to claim 2, wherein The digital twin model is a battery model, including an electrochemical P2D model, a fractional-order impedance model, a life model, an integer-order equivalent circuit model, a battery thermal model, and a thermal runaway model.

4. The cloud collaborative control method for an electric vehicle power battery vehicle according to claim 1, characterized in that, The cloud battery management platform is also used to collect user habit data.

5. The cloud collaborative control method for an electric vehicle power battery vehicle according to claim 1, wherein, When the cloud battery management platform receives the occurrence of a battery failure fault of a vehicle, the cloud battery management platform is used to analyze the historical battery data of vehicles of the same model, identify the data feature differences between normal vehicles and abnormal vehicles, so as to achieve accurate fault location.

6. The method for cloud collaborative control of an electric vehicle power battery vehicle according to claim 5, wherein The cloud battery management platform adopts a power battery characteristic parameter outlier detection technology, and performs accurate fault location by setting a multi-segment parallel screening threshold or uses a machine learning model for accurate fault location.

7. The method for cloud collaborative management and control of an electric vehicle power battery vehicle according to claim 1, wherein The on-vehicle battery management system is also used to update the battery parameters according to the season and altitude, and upload them to the cloud battery management platform.

8. The cloud collaborative control method for an electric vehicle power battery vehicle according to claim 1, characterized in that, The cloud battery management platform is also used to formulate a management strategy according to the battery state evaluation result, and send the management strategy to the on-vehicle battery management system.

9. The method for cloud collaborative control of an electric vehicle power battery vehicle according to claim 1, wherein, The maximum scale of the upload field in the upload field scale is to upload all data and upload it according to the first period; the minimum scale of the upload field is to only upload the current field and voltage field required by the model and upload it according to the second period, and the first period is less than the second period.

Images