Power battery power jump early warning method, device and equipment

Abstract

The application discloses a power battery power jump early warning method, device and equipment, and belongs to the computer technical field. The method comprises the following steps: acquiring the current driving parameter, the battery state and the environment parameter of the environment where the vehicle is located; determining whether the residual power of the power battery of the vehicle will jump after a preset time length based on the driving parameter, the battery state and the environment parameter; and sending early warning information to the vehicle in the case that the residual power of the power battery jumps after the preset time length, wherein the early warning information is used for indicating the correction of the residual power of the vehicle. According to the application, the early warning information is sent to the vehicle in the case that the residual power will jump after the preset time length, so that the vehicle can timely correct the residual power of the power battery, thereby ensuring that the residual mileage of the vehicle can be accurately determined and displayed subsequently, and the user experience is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, and device for early warning of power level fluctuations in a power battery. Background Technology

[0002] With the development of science and technology, the emergence of electric vehicles has brought great convenience to people's daily lives. Since electric vehicles are powered by batteries, their environmental friendliness is a major selling point and has made them popular.

[0003] Normally, the remaining battery charge is calculated by the electric vehicle's BMS (Battery Management System). However, in certain scenarios, such as when the electric vehicle is driving on bumpy roads or in low-temperature environments, the BMS's calculation of the remaining charge may be inaccurate. This can cause the remaining battery charge to fluctuate, resulting in a change in the remaining driving range. Consequently, the displayed remaining range may not match the actual range, thus degrading the user experience. Summary of the Invention

[0004] This application provides a method, device, equipment, and storage medium for warning of battery charge level fluctuations. It can predict whether the remaining charge of the power battery will fluctuate in the future and, if such fluctuation is expected, provide a warning to the vehicle, allowing the vehicle to adjust its remaining charge level in advance. This ensures accurate display of the remaining range and improves the user experience. The technical solution is as follows:

[0005] Firstly, a method for early warning of sudden changes in the power level of a power battery is provided, the method comprising:

[0006] Obtain the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's surroundings;

[0007] Based on the driving parameters, the battery status, and the environmental parameters, determine whether the remaining charge of the vehicle's power battery changes abruptly after a preset time.

[0008] If the remaining charge of the power battery changes abruptly after a preset period of time, a warning message is sent to the vehicle, which is used to instruct the vehicle to correct the remaining charge.

[0009] In this application, the current driving parameters of the vehicle, the battery status, and the environmental parameters of the vehicle's environment are first obtained; that is, parameters that can affect the remaining charge of the power battery are obtained. Then, based on these driving parameters, battery status, and environmental parameters, it is determined whether the remaining charge of the vehicle's power battery will change after a preset time period. Finally, if the remaining charge of the power battery changes after the preset time period, a warning message is sent to the vehicle, i.e., a charge change warning is given, allowing the vehicle to subsequently correct the remaining charge. Thus, by sending a warning message to the vehicle when it is determined that the remaining charge will change after a preset time period, the vehicle can promptly correct the remaining charge, ensuring accurate determination and display of the vehicle's remaining range, thereby improving the user experience.

[0010] Optionally, the step of sending a warning message to the vehicle when the remaining charge of the power battery changes abruptly after a preset period of time includes:

[0011] If the remaining charge of the power battery changes abruptly after a preset time, a correction coefficient for the charging and discharging efficiency of the power battery is determined based on the driving parameters, the battery state, and the environmental parameters.

[0012] Send a warning message carrying the charge / discharge efficiency correction coefficient to the vehicle to instruct the vehicle to correct the remaining charge of the power battery based on the charge / discharge efficiency correction coefficient.

[0013] Optionally, the driving parameters include the vehicle's current speed and current mileage, the battery status includes the current remaining charge and current health status of the power battery, and the environmental parameters include the current temperature of the environment in which the vehicle is located. The step of determining whether the remaining charge of the vehicle's power battery changes abruptly after a preset time period based on the driving parameters, the battery status, and the environmental parameters includes:

[0014] The current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature are input into the battery charge jump prediction model. The battery charge jump prediction model processes the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and outputs the probability that the remaining battery power will jump after a preset time.

[0015] If the probability is greater than or equal to a preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will change after a preset time.

[0016] If the probability is less than the preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will not change after a preset time period.

[0017] Optionally, the battery level jump prediction model includes multiple leaf nodes. The process of using the battery level jump prediction model to process the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and outputting the probability of a jump, includes:

[0018] The multiple leaf nodes are used to make decisions and judgments on the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and output the probability that the remaining battery power will change after a preset time.

[0019] Optionally, the training method for the power level jump prediction model includes:

[0020] Obtain a first training dataset, which includes multiple sets of sample data and labels for the multiple sets of sample data. Each set of sample data includes the values ​​of multiple feature attributes, including historical driving parameters, historical battery status, and historical environmental parameters. The labels are used to indicate whether the remaining power capacity of the power battery changes within a target time period under the sample data.

[0021] For any set of sample data among the multiple sets of sample data, feature extraction is performed on the sample data to obtain the features of the sample data;

[0022] The features of the sample data are concatenated with the sample data to obtain a second training dataset, which includes the multiple sets of sample data, the features of the multiple sets of sample data, and the labels of the multiple sets of sample data.

[0023] The power fluctuation prediction model is trained based on the second training dataset.

[0024] Optionally, the plurality of feature attributes include continuous and discrete feature attributes, and the step of extracting features from the sample data to obtain the features of the sample data includes:

[0025] For any continuous type feature attribute in the sample data, determine the mean, maximum, minimum and variance of the continuous type feature attribute in the sample data;

[0026] For any discrete feature attribute in the sample data, determine the median corresponding to the discrete feature attribute in the sample data.

[0027] Optionally, training the power jump prediction model based on the second training dataset includes:

[0028] For the root node in the first level of n levels, the second training dataset is input into the root node, the information gain of multiple feature attributes under the root node is determined, and the second training dataset is divided based on the feature attribute with the largest information gain, so as to obtain multiple leaf nodes based on the root node.

[0029] For each leaf node in the i-th level of the n levels, determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node by the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the (i+1)-th level, where i is an integer greater than or equal to 2 and less than or equal to n.

[0030] Let i = i+1, and execute the following steps for each leaf node in the i-th level of the n levels: determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node with the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the i+1-th level, until the predicted jump probability is determined based on the dataset contained in the split leaf nodes.

[0031] Based on the difference between the predicted jump probability and the tag, the parameters of the power jump prediction model are adjusted.

[0032] Optionally, obtaining the first training dataset includes:

[0033] Acquire historical vehicle data, which includes historical driving parameters, historical battery status, and historical environmental parameters;

[0034] The historical vehicle data is divided into multiple sets of sample data by using the target time period as the time interval. Each set of sample data includes historical driving parameters, historical battery status, and historical environmental parameters for the corresponding time period.

[0035] For any set of sample data from the multiple sets of sample data, the label of the sample data is determined based on the historical battery status of the corresponding time period.

[0036] Optionally, the historical battery status of the corresponding time period in the sample data includes the historical remaining battery power within the corresponding time period. Determining the label of the sample data based on the historical battery status of the corresponding time period includes:

[0037] Determine the difference between the historical remaining electricity at the start time and the historical remaining electricity at the end time of the corresponding time period;

[0038] If the power difference is greater than or equal to a preset power threshold, the label of the sample data is determined to be a jump.

[0039] If the power difference is less than the preset power threshold, the label of the sample data is determined to be "no change".

[0040] Secondly, a power battery charge level jump warning device is provided, the device comprising:

[0041] The first acquisition module is used to acquire the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's environment.

[0042] The determination module is used to determine, based on the driving parameters, the battery status, and the environmental parameters, whether the remaining charge of the vehicle's power battery changes abruptly after a preset time.

[0043] The warning module is used to send a warning message to the vehicle when the remaining power of the power battery changes abruptly after a preset time. The warning message is used to instruct the vehicle to correct the remaining power of the battery.

[0044] Optionally, the early warning module is used for:

[0045] If the remaining charge of the power battery changes abruptly after a preset time, a correction coefficient for the charging and discharging efficiency of the power battery is determined based on the driving parameters, the battery state, and the environmental parameters.

[0046] Send a warning message carrying the charge / discharge efficiency correction coefficient to the vehicle to instruct the vehicle to correct the remaining charge of the power battery based on the charge / discharge efficiency correction coefficient.

[0047] Optionally, the determining module is used to:

[0048] The current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature are input into the battery charge jump prediction model. The battery charge jump prediction model processes the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and outputs the probability that the remaining battery power will jump after a preset time.

[0049] If the probability is greater than or equal to a preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will change after a preset time.

[0050] If the probability is less than the preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will not change after a preset time period.

[0051] Optionally, the power level jump prediction model includes multiple leaf nodes, and the determining module is used for:

[0052] The multiple leaf nodes are used to make decisions and judgments on the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and output the probability that the remaining battery power will change after a preset time.

[0053] Optionally, the device further includes:

[0054] The second acquisition module is used to acquire a first training dataset. The first training dataset includes multiple sets of sample data and labels for the multiple sets of sample data. Each set of sample data includes the values ​​of multiple feature attributes. The multiple feature attributes include historical driving parameters, historical battery status, and historical environmental parameters. The labels are used to indicate whether the remaining power of the power battery changes within a target time period under the sample data.

[0055] The feature extraction module is used to extract features from any one set of sample data among the multiple sets of sample data to obtain the features of the sample data.

[0056] The concatenation module is used to concatenate the features of the sample data with the sample data to obtain a second training dataset. The second training dataset includes the multiple sets of sample data, the features of the multiple sets of sample data, and the labels of the multiple sets of sample data.

[0057] The training module is used to train the power jump prediction model based on the second training dataset.

[0058] Optionally, the plurality of feature attributes include continuous and discrete feature attributes, and the feature extraction module is used for:

[0059] For any continuous type feature attribute in the sample data, determine the mean, maximum, minimum and variance of the continuous type feature attribute in the sample data;

[0060] For any discrete feature attribute in the sample data, determine the median corresponding to the discrete feature attribute in the sample data.

[0061] Optionally, the training module is used for:

[0062] For the root node in the first level of n levels, the second training dataset is input into the root node, the information gain of multiple feature attributes under the root node is determined, and the second training dataset is divided based on the feature attribute with the largest information gain, so as to obtain multiple leaf nodes based on the root node.

[0063] For each leaf node in the i-th level of the n levels, determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node by the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the (i+1)-th level, where i is an integer greater than or equal to 2 and less than or equal to n.

[0064] Let i = i+1, and execute the following steps for each leaf node in the i-th level of the n levels: determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node with the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the i+1-th level, until the predicted jump probability is determined based on the dataset contained in the split leaf nodes.

[0065] Based on the difference between the predicted jump probability and the tag, the parameters of the power jump prediction model are adjusted.

[0066] Optionally, the second acquisition module is used for:

[0067] Acquire historical vehicle data, which includes historical driving parameters, historical battery status, and historical environmental parameters;

[0068] The historical vehicle data is divided into multiple sets of sample data by using the target time period as the time interval. Each set of sample data includes historical driving parameters, historical battery status, and historical environmental parameters for the corresponding time period.

[0069] For any set of sample data from the multiple sets of sample data, the label of the sample data is determined based on the historical battery status of the corresponding time period.

[0070] Optionally, the second acquisition module is used for:

[0071] Determine the difference between the historical remaining electricity at the start time and the historical remaining electricity at the end time of the corresponding time period;

[0072] If the power difference is greater than or equal to a preset power threshold, the label of the sample data is determined to be a jump.

[0073] If the power difference is less than the preset power threshold, the label of the sample data is determined to be "no change".

[0074] Thirdly, a computer device is provided, the computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the above-described method for warning of power battery charge fluctuations.

[0075] Fourthly, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for warning of power battery charge level fluctuations.

[0076] Fifthly, a computer program product containing instructions is provided, which, when run on a computer, causes the computer to execute the steps of the aforementioned power battery charge level jump warning method.

[0077] It is understood that the beneficial effects of the second, third, fourth, and fifth aspects mentioned above can be found in the relevant descriptions in the first aspect above, and will not be repeated here. Attached Figure Description

[0078] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0079] Figure 1 This is a schematic diagram of a scenario for a power battery charge jump warning method provided in an embodiment of this application;

[0080] Figure 2 This is a flowchart of a power battery charge jump warning method provided in an embodiment of this application;

[0081] Figure 3 This is a schematic diagram of the structure of a power jump prediction model provided in an embodiment of this application;

[0082] Figure 4 This is a flowchart illustrating the training process of a power level jump prediction model provided in an embodiment of this application.

[0083] Figure 5 This is a schematic diagram of the structure of a power battery charge jump warning device provided in an embodiment of this application;

[0084] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0085] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0086] It should be understood that "multiple" as mentioned in this application refers to two or more. In the description of this application, unless otherwise stated, " / " indicates "or," for example, A / B can mean A or B; "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist, for example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, to facilitate a clear description of the technical solutions of this application, the terms "first," "second," etc., are used to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or execution order, and that "first," "second," etc., do not necessarily imply differences.

[0087] Before providing a detailed explanation of the embodiments of this application, the application scenarios of these embodiments will be described first.

[0088] The remaining charge of a vehicle's battery is affected by various factors. In certain scenarios, such as high or low temperatures, the BMS's calculation of the remaining battery charge can be affected, leading to inaccurate calculations. Furthermore, since the vehicle's remaining range is calculated based on the remaining battery charge, inaccurate remaining charge will also result in inaccurate range calculations. Consequently, users will not see the true remaining range, thus degrading the user experience.

[0089] Therefore, this application provides a method for predicting power battery charge level fluctuations, which can be applied to scenarios where the remaining power charge of a vehicle's power battery is to be predicted.

[0090] For example, Figure 1 This is a schematic diagram illustrating a method for early warning of sudden changes in battery power level. (See attached diagram) Figure 1 , Figure 1 It includes vehicle 101 and cloud platform 102, with vehicle 101 having a power battery 103.

[0091] Vehicle 101 can be an electric vehicle, and a T-BOX (Telematics BOX, vehicle-to-everything intelligent terminal) is deployed on vehicle 101. The T-BOX can connect to the CAN (Controller Area Network) bus of vehicle 101 to receive various parameters of vehicle 101. In addition, the T-BOX also communicates with cloud platform 102, so that vehicle 101 can communicate with cloud platform 102 through T-BOX.

[0092] The cloud platform 102 can be a TSP (Telematics Service Provider) platform or other cloud platforms, as long as they can implement the power battery charge jump warning method provided in this application embodiment. This application embodiment does not limit this.

[0093] Specifically, when issuing a warning about a sudden drop in battery power, each controller on vehicle 101 first obtains the current driving parameters, the current battery status, and the environmental parameters of the environment in which vehicle 101 is located. Then, it sends the current driving parameters, the current battery status, and the environmental parameters of the environment in which vehicle 101 is located to T-BOX. Afterward, T-BOX sends the current driving parameters, the current battery status, and the environmental parameters of the environment in which vehicle 101 is located to cloud platform 102.

[0094] After receiving various parameters from the T-BOX of vehicle 101, cloud platform 102 determines whether the remaining charge of vehicle 101's power battery 103 will fluctuate after a preset time period based on current driving parameters, current battery status, and current environmental parameters. If the remaining charge of power battery 103 is expected to fluctuate after the preset time period, cloud platform 102 sends a warning message to vehicle 101, allowing vehicle 101 to adjust its remaining charge level upon receiving the warning message.

[0095] Thus, by sending a warning message to the vehicle 101 when it is determined that the remaining charge of the power battery 103 will change after a preset time, the vehicle 101 can promptly correct the remaining charge of the power battery 103, thereby ensuring that the remaining range of the vehicle 101 can be accurately determined and displayed, thereby improving the user experience.

[0096] The following is a detailed explanation of the power battery charge jump warning method provided in the embodiments of this application.

[0097] Figure 2 This is a flowchart illustrating a method for providing early warning of sudden power level fluctuations in a power battery, as provided in an embodiment of this application. This method can be applied to computer devices; for example, the computer device can be a server deployed with the aforementioned cloud platform. See also... Figure 2 The method includes the following steps.

[0098] Step 201: The computer device acquires the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's environment.

[0099] In this embodiment of the application, the vehicle may be an electric vehicle.

[0100] The driving parameters refer to parameters generated during vehicle operation, used to indicate the vehicle's current driving status. Optionally, these driving parameters may include parameters such as the vehicle's current speed and current mileage. The current mileage refers to the total mileage traveled up to the current moment. Optionally, the current mileage can be obtained by acquiring the mileage from the vehicle's odometer.

[0101] Alternatively, the current vehicle speed can be determined in any of the following ways.

[0102] Example 1: A vehicle speed sensor can be installed on the vehicle, and the vehicle's current speed can be obtained by the vehicle speed sensor.

[0103] Example 2: The vehicle's current speed can also be calculated using wheel speed conversion. Optionally, methods for calculating the current speed based on wheel speed include the average wheel speed method and the maximum wheel speed method. The average wheel speed method takes the average of the wheel speeds of the two rear wheels as the current speed; the maximum wheel speed method takes the maximum value of the wheel speeds of all four wheels as the current speed.

[0104] Example 3: The vehicle's current speed can also be obtained using another wheel speed conversion method. The specific calculation process is: Vehicle speed = Wheel circumference * Wheel speed. Here, wheel speed can be obtained through wheel speed sensors, and wheel circumference is an inherent parameter of the tire.

[0105] It should be understood that the above methods are merely illustrative examples, and the calculated vehicle speeds are not significantly different; all can be taken as the actual vehicle speed. Any method used to calculate vehicle speed falls within the scope of protection of this application.

[0106] The battery status refers to the state of the vehicle's power battery. Optionally, the battery status may include the vehicle's power battery's SOC (State of Charge), which is the remaining capacity of the power battery, and may also include the power battery's State of Health (SOH), which indicates the degree of aging of the power battery, that is, reflects the healthy lifespan of the power battery. The remaining capacity and state of health of the power battery can be calculated by the BMS (Battery Management System).

[0107] Optionally, the BMS can calculate the remaining charge of the power battery using the ampere-hour integration method. Alternatively, the BMS can calculate the healthy lifespan of the power battery using the following formula (1).

[0108]

[0109] Among them, R EoL R is the internal resistance of the power battery at the end of its life. BoL R is the internal resistance of the power battery when it leaves the factory, and R is the internal resistance of the power battery in its current state.

[0110] This environmental parameter can include temperature, that is, the ambient temperature of the environment in which the vehicle is located. Typically, a temperature sensor is located behind the front grille of the vehicle; optionally, the ambient temperature of the vehicle's surroundings can be detected by the temperature sensor behind the front grille.

[0111] In this case, after the vehicle detects the driving parameters, battery status, and environmental parameters, it sends these parameters to the computer device via the vehicle's T-BOX, thereby enabling the computer device to obtain the driving parameters, battery status, and environmental parameters.

[0112] As another implementation, the computer device can first send a parameter request to the vehicle via the TSP service. This parameter request requests the vehicle to send the driving parameters, battery status, and environmental parameters. Subsequently, upon receiving the parameter request, the vehicle collects its current driving parameters, battery status, and environmental parameters, and sends these parameters to the computer device via the T-BOX.

[0113] Step 202: Based on the driving parameters, the battery status, and the environmental parameters, the computer device determines whether the remaining charge of the vehicle's power battery will change after a preset time.

[0114] The preset duration can be set in advance, and technicians can adjust the preset duration according to actual needs. For example, if it is necessary to determine whether the remaining charge of the power battery will fluctuate after 5 minutes, the preset duration can be set to 5 minutes.

[0115] Specifically, given that the driving parameters include the vehicle's current speed and current mileage, the battery status includes the current remaining charge and current health status, and the environmental parameters include the current temperature of the vehicle's environment, step 202 can be performed as follows: inputting the current speed, current mileage, current remaining charge, current health status, and current temperature into the charge jump prediction model; processing the current speed, current mileage, current remaining charge, current health status, and current temperature through the charge jump prediction model; and outputting the probability that the remaining charge of the power battery will jump after a preset time. If the probability is greater than or equal to a preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will jump after the preset time. If the probability is less than the preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will not jump after the preset time.

[0116] This battery charge jump prediction model is used to predict whether the remaining charge of the power battery will jump after a preset time. Optionally, the battery charge jump prediction model can be a machine learning model such as a decision tree or random forest, or a deep learning-based neural network model such as a deep convolutional neural network or a fully connected neural network. This application embodiment does not limit the specific model.

[0117] The preset probability threshold can be set by technicians according to actual needs, and the preset probability threshold can be set relatively high. For example, the preset probability threshold can be set to 0.8. In this case, if the probability is greater than or equal to the preset probability threshold, it means that the battery charge change prediction model predicts a high probability that the remaining battery charge will change after a preset time period, thus confirming that the remaining battery charge will change after the preset time period. If the probability is less than or equal to the preset probability threshold, it means that the battery charge change prediction model predicts a low probability that the remaining battery charge will change after the preset time period, meaning that the remaining battery charge will not change after the preset time period.

[0118] Optionally, if the power consumption jump prediction model is a decision tree model, the power consumption jump prediction model may include multiple leaf nodes. These multiple leaf nodes constitute a decision tree model, which is the power consumption jump prediction model.

[0119] It should be understood that these multiple leaf nodes are obtained when the battery charge change prediction model is trained. That is, during the training process, these multiple leaf nodes and the relationships between them are gradually determined. Furthermore, the training process of this battery charge change prediction model is also a process of learning the changing patterns of the remaining battery charge and determining the decision conditions accordingly. Therefore, these multiple leaf nodes can be seen as decision conditions for determining whether a change in the remaining battery charge will occur after a preset time.

[0120] In this case, the computer device processes the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature through the battery power jump prediction model, and outputs the probability of a jump. The operation can be as follows: make decisions and judgments on the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature through the multiple leaf nodes, and output the probability that the remaining battery power will jump after a preset time.

[0121] The decision-making process involves determining whether the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature meet the decision conditions indicated by the multiple leaf nodes.

[0122] In this way, by learning the changing pattern of the remaining power of the power battery in advance and constructing a decision tree, the decision tree can be used to make decisions on the current vehicle speed, current driving range, current remaining power, current health status, and current temperature. This allows for an accurate determination of whether the remaining power of the power battery will change after a preset time.

[0123] For example, Figure 3 A schematic diagram of the structure of this power fluctuation prediction model is shown below. Figure 3 , Figure 3 It includes a root node 301, eight leaf nodes 302, and output data 303 (the probability of a jump). Specifically, the process of predicting whether the remaining battery power will jump after a preset time using this battery power jump prediction model includes the following steps.

[0124] (1) Input the current vehicle speed, current mileage, current remaining battery power, current health status and current temperature into the root node 301, and first determine whether the current vehicle speed is greater than the target vehicle speed threshold.

[0125] (2) When the current vehicle speed is greater than the target vehicle speed threshold, the current driving mileage, current remaining battery power, current health status and current temperature are input into the first leaf node 302, and the first leaf node 302 determines whether the current driving mileage is greater than the target mileage threshold.

[0126] (3) If the current driving mileage is greater than the target mileage threshold, the current remaining battery power, current health status and current temperature are input into the 4th leaf node 302, and the 4th leaf node 302 determines whether the current health status meets the target health conditions.

[0127] For example, if the current health status of a vehicle is 60%, while the target health status is 80%, it means that the current health status does not meet the target health conditions.

[0128] (4) If the current health status does not meet the target health conditions, input the current remaining power and the current temperature into the 6th leaf node 302, and the 6th leaf node 302 determines whether the current temperature is greater than the target temperature threshold.

[0129] (5) If the current temperature is greater than the target temperature threshold, input the current remaining power into the 7th leaf node, and the 7th leaf node 302 determines whether the current remaining power is less than the target power threshold.

[0130] (6) When the current remaining power is less than the target power threshold, the probability that the remaining power of the output power battery will change after a preset time is 303.

[0131] Among them, the target vehicle speed threshold, target mileage threshold, target health condition, target temperature threshold, and target battery level threshold can be determined during the training of the battery level jump prediction model.

[0132] It is worth noting that before using the power battery charge jump prediction model to predict whether the remaining power of the power battery will jump after a preset time, the power battery charge jump prediction model needs to be trained first.

[0133] Specifically, the process of training the power jump prediction model may include the following steps (1)-(4).

[0134] (1) Obtain the first training dataset.

[0135] The first training dataset includes multiple sets of sample data and their labels. Each set of sample data corresponds one-to-one with its corresponding label. The label for each set of sample data indicates whether the power battery undergoes a jump after a preset time period under the corresponding sample data.

[0136] Each set of sample data includes the values ​​of multiple feature attributes, including historical driving parameters, historical battery status, and historical environmental parameters.

[0137] Specifically, given that historical driving parameters include vehicle speed and mileage, historical battery status includes remaining charge and health status, and historical environmental parameters include the temperature of the vehicle's environment, these multiple characteristic attributes include vehicle speed, mileage, remaining charge, health status, and temperature. Therefore, each set of sample data can include specific values ​​for vehicle speed, mileage, remaining charge, health status, and temperature.

[0138] Furthermore, in this embodiment of the application, each set of sample data includes values ​​for vehicle speed, mileage, remaining battery power, health status, and temperature at multiple times. For example, the multiple sets of sample data include values ​​for vehicle speed, mileage, remaining battery power, health status, and temperature at multiple times within different time periods.

[0139] In this way, by training the battery charge change prediction model based on vehicle speed, mileage, remaining battery charge, health status, and temperature data collected in different time periods, the model can learn the SOC change pattern of the power battery in continuous time periods, and thus accurately predict whether the remaining battery charge will change after a preset time.

[0140] It is worth noting that the first training dataset can be a training dataset obtained by processing vehicle data collected from vehicles of the same model. Furthermore, the battery level jump prediction model trained based on the first training dataset is used to predict whether the power battery of that vehicle model will experience a jump after a preset time period.

[0141] In this embodiment, for different vehicle models, vehicle data collected from different models can be acquired separately, and after processing, training datasets corresponding to different models can be obtained. Subsequently, for different vehicle models, corresponding battery level jump prediction models are trained using the corresponding training datasets. These battery level jump prediction models are used to predict whether the power battery of the corresponding vehicle model will experience a jump after a preset time.

[0142] In this way, multiple battery level change prediction models can be trained. These models are used to predict whether the battery level of a vehicle of a corresponding model will change after a preset time. Furthermore, since these multiple battery level change prediction models are trained using training datasets corresponding to the corresponding vehicle models, they can accurately predict whether the battery level of the corresponding vehicle model will change after a preset time.

[0143] Furthermore, after training and obtaining these multiple battery level jump prediction models, these multiple battery level jump prediction models can be stored in a one-to-one correspondence with the vehicle model identifiers of multiple vehicle models.

[0144] This model designation is used to uniquely identify the vehicle's model. Vehicles of the same model share the same model designation, while vehicles of different models have different model designations.

[0145] In this scenario, when acquiring the driving parameters, battery status, and environmental parameters, the computer device can also obtain the vehicle model identifier. Subsequently, the computer device can select the battery level prediction model corresponding to the vehicle model identifier from among multiple battery level prediction models. Then, it uses this battery level prediction model to predict whether the remaining battery charge of the vehicle corresponding to that model will experience a sudden change.

[0146] Specifically, step (1) can be performed as follows: acquire historical vehicle data; divide the historical vehicle data into multiple sets of sample data by using the target time period as the time interval, and each set of sample data includes the historical driving parameters, historical battery status, and historical environmental parameters of the corresponding time period; for any set of sample data in the multiple sets of sample data, determine the label of the set of sample data based on the historical battery status of the corresponding time period.

[0147] Historical vehicle data refers to the driving parameters, battery status, and environmental parameters of a vehicle collected continuously at multiple times during its various journeys. In other words, historical vehicle data includes historical driving parameters, historical battery status, and historical environmental parameters.

[0148] The target time period can be preset, and can also be set by technicians according to actual needs. The target time period can be set to a small value; for example, setting the target time period to 10 minutes will divide the historical vehicle data into 10-minute intervals.

[0149] For example, if the target time period is 10 minutes, and the historical vehicle data includes driving parameters, battery status, and environmental parameters collected between 10:00 and 11:00, then 6 sets of sample data can be obtained by dividing the data into 10-minute intervals. Each of these 6 sets of sample data includes driving parameters, battery status, and environmental parameters for the corresponding time period. For example, the first set of sample data includes driving parameters, battery status, and environmental parameters for the vehicle between 10:01 and 10:10. The second set of sample data includes driving parameters, battery status, and environmental parameters for the vehicle between 10:11 and 10:20.

[0150] In this scenario, dividing historical vehicle data into multiple sets of sample data increases the number of samples for the battery level jump prediction model and reduces its input dimensionality. This provides more data support for training the model, leading to a more accurate battery level jump prediction model.

[0151] The historical battery status for a given time period can include the historical remaining battery power for that time period. In this case, for any set of sample data from the multiple sets of sample data, the operation of determining the label of this set of sample data based on the historical battery status for the corresponding time period can be as follows: determine the difference between the historical remaining battery power at the start time and the historical remaining battery power at the end time of the corresponding time period; if the difference is greater than or equal to a preset battery power threshold, determine that the label of this set of sample data is "transitional"; if the difference is less than the preset battery power threshold, determine that the label of this set of sample data is "no transition".

[0152] Because the battery is constantly discharging while the vehicle is in motion, there will be a difference between the remaining charge at the start and end of a short period of time. If this difference is too large, it indicates a jump in the remaining charge. Therefore, we can first determine the difference between the historical remaining charge at the start and end of the corresponding time period, and then determine the label for this set of sample data based on the difference.

[0153] A preset battery threshold can be set in advance, and this threshold can be set relatively high. In this case, if the battery difference is greater than or equal to the preset battery threshold, it indicates that the battery difference is large. This means there is a large difference between the historical remaining battery level at the beginning and end of the corresponding time period, indicating significant fluctuations in the remaining battery level within that time period. Therefore, it can be determined that a jump occurred in the remaining battery level within that time period, and the label for this set of sample data can be identified as a jump.

[0154] If the power difference is less than the preset power threshold, it indicates that the power difference is small. This means the difference between the historical remaining power at the start and end of the corresponding time period is small, indicating that the fluctuation of the remaining power within the corresponding time period is normal. Therefore, it can be determined that the remaining power did not change within the corresponding time period. Thus, the label for this set of sample data can be determined as "no change".

[0155] In this way, the label of each set of sample data can be automatically determined through the above operations, without the need for manual intervention to label each set of sample data. This saves manpower, improves the efficiency of label determination for the multiple sets of sample data, and reduces the production cost of the initial training dataset.

[0156] (2) For any set of sample data in the multiple sets of sample data, perform feature extraction on the set of sample data to obtain the features of the set of sample data.

[0157] In this case, by determining the characteristics of each set of sample data in the multiple sets of sample data, the data dimensionality of the multiple sets of sample data can be reduced, thereby reducing the computational load of the power jump prediction model and improving the training efficiency of the power jump prediction model.

[0158] Optionally, the multiple feature attributes of the multiple sets of sample data may include continuous feature attributes and discrete feature attributes. In this case, different features can be extracted from the sample data according to the different feature attribute types.

[0159] Continuous feature attributes refer to features whose values ​​are continuous within a given time period. Discrete feature attributes refer to features whose values ​​are discrete within a given time period. For example, if the multiple feature attributes are vehicle speed, mileage, remaining battery power, health status, and temperature, then the continuous feature attributes are vehicle speed, mileage, remaining battery power, and health status. The discrete feature attribute is temperature.

[0160] Specifically, step (2) can be performed as follows: for any continuous type feature attribute in this set of sample data, determine the mean, maximum, minimum and variance of this continuous type feature attribute in this set of sample data; for any discrete type feature attribute in this set of sample data, determine the median of this discrete type feature attribute in this set of sample data.

[0161] In this case, by using different feature extraction methods for continuous and discrete feature attributes within the corresponding time period, different features can be extracted from the sample data for different types of feature attributes, thus obtaining more accurate features.

[0162] For example, this sample data set contains historical vehicle data collected at three specific moments within a given time period. This data includes vehicle speed (40, 45, 50 km / h), mileage (300, 310, 320 km / h), remaining battery power (90%, 89%, 88%), health status (90%, 90%, 89%), and temperature (26°C, 26°C, 28°C) collected at these three moments. The continuous attribute vehicle speed has an average of 45 km / h, a maximum of 50 km / h, a minimum of 40 km / h, and a variance of 16.7. The continuous attribute mileage has an average of 310 km / h, a maximum of 320 km / h, a minimum of 300 km / h, and a variance of 66.7. The continuous attribute remaining battery power has an average of 89%, a maximum of 90%, a minimum of 88%, and a variance of 0.7. The continuous attribute health status has an average of 89.7%, a maximum of 90%, a minimum of 89%, and a variance of 0.2. The discrete attribute temperature has a median of 26°C.

[0163] (3) The features of this set of sample data are concatenated with the sample data to obtain the second training dataset.

[0164] Optionally, computer devices can use CONCAT to stitch together the features of this set of sample data.

[0165] Thus, by performing step (3) above on each of the multiple sets of sample data, a second training dataset can be obtained. The second training dataset includes the multiple sets of sample data, the features of the multiple sets of sample data, and the labels of the multiple sets of sample data.

[0166] (4) Based on the second training dataset, train the power jump prediction model.

[0167] Optionally, if the power jump prediction model is a decision tree model, step (4) can be performed as follows: For the root node in the first level of n levels, input the second training dataset into the root node, determine the information gain of multiple feature attributes under the root node, and divide the second training dataset based on the feature attribute with the largest information gain, so as to obtain multiple leaf nodes based on the root node split; For each leaf node in the i-th level of n levels, determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node based on the feature attribute with the largest information gain, so as to obtain multiple leaf nodes based on the current leaf node split. Split to form multiple leaf nodes in the (i+1)th layer, where i is an integer greater than or equal to 2 and less than or equal to n; let i = i+1, for each leaf node in the i-th layer of the n layers, determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node by the feature attribute with the largest information gain, and split to form multiple leaf nodes in the (i+1)th layer based on the current leaf node, until the predicted jump probability is determined based on the dataset contained in the split leaf nodes; adjust the parameters of the power jump prediction model based on the difference between the predicted jump probability and the label.

[0168] A decision tree model is a tree-structured model consisting of several nodes at n levels, connected by branches. Each node includes a root node and leaf nodes. The node at the top level (level 1) is called the root node, and the remaining nodes are called leaf nodes. The generation of a decision tree model is a supervised learning process. It involves providing several sets of sample data, each with its corresponding true classification result (label). That is, given the classification results, a decision tree model, i.e., the power consumption jump prediction model, is obtained by learning the classification results of these sample data. Optionally, this decision tree model can be based on LightGBM (LightGradient Boosting Machine).

[0169] Specifically, the training process in step (4) above is essentially a process of continuously dividing the dataset into optimal data subsets using the criteria that satisfy feature selection. For each division of the dataset, it is hoped that the optimal data subset will be obtained, so that the final generated decision tree model is optimal, that is, the trained power jump prediction model is the most accurate.

[0170] The root node of a decision tree model contains the initial dataset, which is the second training dataset. The datasets contained in the remaining leaf nodes are obtained by partitioning the datasets contained in their parent leaf nodes according to the corresponding feature attributes.

[0171] Information gain represents the difference in information entropy before and after partitioning a dataset by a certain feature attribute. The information gain corresponding to the feature attribute can be calculated using the following formula (2).

[0172]

[0173] Where D is the second training dataset, a i Let G(D, a) be the i-th feature attribute in the dataset contained in the current node. i ) represents the information gain corresponding to the i-th feature attribute, and Ent(D) represents the information entropy of the root node of the decision tree model. Let |D| represent the number of samples contained in the dataset divided by the k-th leaf node under the i-th feature attribute, and let |D| represent the total number of samples contained in the second training dataset. Let represent the information entropy of the k-th leaf node under the i-th feature attribute, and K represent the total number of leaf nodes under the i-th feature attribute.

[0174] As shown above, the information gain corresponding to a feature attribute represents the difference in information entropy before and after splitting the dataset based on that feature attribute. Generally, the information entropy Ent(D) of the root node in a decision tree model is constant; however, the information entropy after splitting the dataset varies. The smaller the value, the greater the information gain of the feature attribute, indicating that using this feature attribute to partition the dataset is more effective.

[0175] In this case, when partitioning the dataset, the information gain feature attribute among multiple feature attributes is always used to partition the dataset. This can reduce the uncertainty of data partitioning and thus improve the efficiency of model training.

[0176] The training process of step (4) above will now be illustrated with an example.

[0177] For example, multiple features include vehicle speed, mileage, health status, and temperature. The decision tree is set to have 4 layers.

[0178] For the root node (the node at the first level), the second training dataset is input into the root node, and the information gain corresponding to vehicle speed, mileage, health status, and temperature is determined respectively. Assuming that the information gain corresponding to mileage is the largest, the second training dataset can be divided into two subsets based on mileage. For example, it can be divided into a subset where the mileage is greater than a preset mileage and a subset where the mileage is less than or equal to a preset mileage. It is worth noting that these two subsets also include sample data corresponding to the other three feature attributes: vehicle speed, health status, and temperature.

[0179] For the two leaf nodes at the second level, i.e., the two leaf nodes obtained after splitting the root node, for the first leaf node (containing the data subset with a mileage greater than a preset mileage), determine the information gain corresponding to vehicle speed, health status, and temperature respectively. Assuming that the information gain corresponding to vehicle speed is the largest, then the data subset contained in the first leaf node is divided by vehicle speed, resulting in two data subsets: a subset of data with a mileage greater than the preset mileage and a vehicle speed greater than a preset speed threshold, and a subset of data with a mileage greater than the preset mileage and a vehicle speed less than or equal to the preset speed threshold. For the second leaf node (containing the data subset with a mileage less than or equal to the preset mileage), determine the information gain corresponding to vehicle speed, health status, and temperature respectively. Assuming that the information gain corresponding to health status is the largest, then the data subset contained in the second leaf node is divided by health status, resulting in two data subsets: a subset of data with a mileage less than or equal to the preset mileage and a data subset of data with a mileage less than or equal to the preset mileage and a health status less than or equal to the preset mileage and a health status not meeting the preset health status.

[0180] Similarly, the above process is performed on each leaf node in the third level, allowing for optimal dataset partitioning of the data subset under each leaf node to obtain leaf nodes in the fourth level. For each leaf node in the fourth level, the predicted transition probability can be determined based on the dataset contained in the current leaf node.

[0181] The predicted jump probability is the probability that the remaining charge of the power battery will jump during the model training process.

[0182] Subsequently, during the training of the power jump prediction model, each time a predicted jump probability is output, the predicted jump probability is compared with the corresponding label to obtain the difference between the predicted jump probability and the corresponding label. That is, the difference between the predicted value and the actual value of the power jump prediction model is obtained. The parameters of the power jump prediction model are then updated based on this difference.

[0183] It is worth noting that during the training of this battery level jump prediction model, the detailed operation of step (4) above needs to be continuously iterated. That is, the parameters of the battery level jump prediction model need to be continuously updated based on the difference between the predicted value and the actual value in order to train the battery level jump prediction model. Subsequently, by using this battery level jump prediction model, it is possible to accurately predict whether the remaining battery level will jump after a preset time.

[0184] Optionally, during the training of the battery level jump prediction model, feature importance can also be output, that is, the degree of importance of the multiple feature attributes to the battery level jump prediction model. This feature importance can be used as the basis for updating the parameters of the battery level jump prediction model during iterative training. For example, the higher the feature importance of a feature attribute, the higher the parameters of the leaf node corresponding to that feature attribute can be considered, thereby improving the training efficiency of the battery level jump prediction model.

[0185] Optionally, the trained decision tree model (battery jump prediction model) can also be pruned based on feature importance.

[0186] Because decision trees are complex trees generated by fully considering all data points, they continuously split nodes during the learning process to classify sample data as correctly as possible. This can lead to an excessive number of branches in the tree, making it very large. Overly large decision trees can lead to overfitting, and the more complex the decision tree, the higher the degree of overfitting. Therefore, to avoid overfitting, the decision model can be pruned.

[0187] Furthermore, after training the battery charge jump prediction model, a validation dataset can be obtained and input into the battery charge jump prediction model. The battery charge jump prediction model can then predict the sample data in the validation dataset to obtain the validation probability that the remaining charge of the power battery will jump.

[0188] The validation dataset includes multiple sets of sample data and their corresponding labels.

[0189] Furthermore, the computer device can also evaluate the power jump prediction model based on the verification probability of the jump obtained from multiple sets of sample data in the verification dataset and the labels corresponding to the multiple sets of sample data.

[0190] Specifically, the evaluation of the power consumption jump prediction model includes: determining the AUC (Area Under Curve) and average error of the validation dataset on the power consumption jump prediction model; and determining that the training of the power consumption jump prediction model is complete when the AUC value is greater than a preset evaluation threshold and the average error is less than a preset error threshold.

[0191] AUC refers to the area under the ROC (receiver operating characteristic curve) curve and the coordinate axis. It is used to measure the prediction accuracy of the electricity prediction model on the sample data in the validation dataset. The ROC curve is a tool for analyzing the performance of binary classification. The horizontal axis of the ROC curve is the false positive rate (the probability that a sample is predicted as positive but is not actually positive), and the vertical axis is the true positive rate (the probability that a sample is predicted as positive and is also actually positive).

[0192] In this case, when the AUC value is greater than the preset evaluation threshold and the average error is less than the prediction error threshold, it indicates that the power jump prediction model has a high prediction accuracy for the sample data in the validation dataset, that is, the prediction of the power jump prediction model is relatively accurate, indicating that the power jump prediction model can perform the prediction task, and the training of the power jump model can be completed.

[0193] Optionally, if the AUC value is less than or equal to a preset evaluation threshold, or if the average error is greater than or equal to a preset error threshold, it indicates that the prediction accuracy of the power consumption jump prediction model is low. In this case, the hyperparameters of the power consumption jump prediction model can be updated to retrain the model.

[0194] For example, the learning rate of the battery level jump prediction model can be updated, and then the model can be trained based on the updated learning rate.

[0195] The following example illustrates the overall training process of this power fluctuation prediction model. For example, Figure 4 See the flowchart for training a power level jump prediction model. Figure 4 The training process of this power jump prediction model includes the following steps 401-406.

[0196] Step 401: Data extraction.

[0197] During each of the vehicle's journeys, driving parameters, battery status, environmental parameters, and other data are continuously collected at multiple times to obtain historical vehicle data.

[0198] Step 402: Data partitioning.

[0199] The historical vehicle data is divided into multiple sets of sample data by using the target time period as the time interval.

[0200] Step 403: Feature Engineering.

[0201] Feature extraction was performed on each of the multiple sets of sample data to obtain the features of the multiple sets of sample data.

[0202] Step 404: Tag definition.

[0203] For each set of sample data in the multiple sets of sample data, determine the label corresponding to each set of sample data, that is, determine whether the remaining power level changes under each set of sample data.

[0204] Step 405: Model training.

[0205] Based on the multiple sets of sample data, the features of the multiple sets of sample data, and the labels of the multiple sets of sample data, the power fluctuation prediction model is trained.

[0206] Step 406: Parameter tuning.

[0207] Based on the difference between the predicted probability of a power jump and the corresponding label output by the power jump prediction model during training, the parameters of the power jump prediction model are updated.

[0208] Thus, through the above step 202, it is possible to predict whether the remaining power of the power battery will change after a preset time.

[0209] Step 203: When the remaining power of the power battery changes drastically after a preset time, the computer device sends a warning message to the vehicle. This warning message is used to instruct the vehicle to correct its remaining power.

[0210] This warning message is used to alert the user that the remaining charge level of the vehicle's power battery will change after a preset time, requiring the remaining charge level of the power battery to be corrected.

[0211] Optionally, the computer device can send the warning information to the vehicle's T-BOX via the TSP service, and then the T-BOX can forward the warning information to the vehicle's BMS, so that the BMS can know that the vehicle's remaining power needs to be corrected, and then the BMS can correct the remaining power of the power battery.

[0212] If the remaining battery charge level fluctuates after a preset time, it indicates that the battery charge calculation is inaccurate after that time. Since the vehicle's remaining range is determined based on its remaining battery charge, an inaccurate calculation of the remaining battery charge will also lead to an inaccurate calculation of the remaining range. Therefore, a warning message needs to be sent to the vehicle to instruct it to correct the remaining battery charge level.

[0213] In this situation, by sending a warning message to the vehicle when it is determined that the remaining battery power will change after a preset time, the vehicle can promptly correct the remaining battery power, thereby ensuring that the remaining range of the vehicle can be accurately determined and displayed, thus improving the user experience.

[0214] Optionally, step 203 may involve: when the remaining charge of the power battery changes abruptly after a preset time, determining the charge-discharge efficiency correction coefficient of the power battery based on the driving parameters, the health status, and the environmental parameters; and sending a warning message carrying the charge-discharge efficiency correction coefficient to the vehicle to instruct the vehicle to correct the remaining charge of the power battery based on the charge-discharge efficiency correction coefficient.

[0215] The charge / discharge efficiency coefficient refers to the energy consumption inside a power battery during the charging and discharging process. It is affected by various factors such as temperature and electrolyte; that is, the charge / discharge efficiency coefficient will change when the temperature and electrolyte conditions change.

[0216] The charge / discharge efficiency correction factor refers to a factor used to correct the charge / discharge efficiency coefficient. In this embodiment, the charge / discharge efficiency correction factor can be the actual charge / discharge efficiency coefficient of the power battery. Since the BMS may not accurately calculate the remaining capacity of the power battery, the computer device can determine the charge / discharge efficiency correction factor to correct the charge / discharge efficiency coefficient used in calculating the remaining capacity.

[0217] In this case, the computer equipment calculates the charging and discharging efficiency correction coefficient of the power battery, so that the energy consumption of the power battery during the charging and discharging process can be obtained more accurately, and then a more accurate remaining power battery capacity can be calculated.

[0218] The operation of determining the charge and discharge efficiency correction coefficient of the power battery based on the driving parameters, the battery state, and the environmental parameters can be as follows: obtain the charge and discharge efficiency correction coefficients corresponding to the driving parameters, the battery state, and the environmental parameters from the target mapping table.

[0219] The target mapping table is used to indicate the actual charge and discharge efficiency coefficient of a power battery under different driving parameters, battery states, and environmental parameters. The target mapping table can be calibrated in advance by technicians.

[0220] For example, technicians can conduct controlled experiments, setting different driving parameters, battery states, and environmental parameters to calculate the actual charge-discharge efficiency coefficient of the power battery under these conditions. This will allow them to calibrate a more accurate charge-discharge efficiency coefficient for the power battery.

[0221] For example, the driving parameters include mileage, battery status includes the health status of the power battery, and environmental parameters include the temperature of the vehicle's surroundings. Table 1 is a target mapping table. Table 1 includes multiple mileages, multiple health statuses, and multiple temperatures. Different mileages, health statuses, and temperatures correspond to different actual charge / discharge efficiency coefficients for the power battery. For example, if the vehicle's current mileage is 300 km, its current health status is 95%, and the current ambient temperature is 26°C, then the actual charge / discharge efficiency coefficient of the power battery can be obtained from Table 1, which is equivalent to obtaining a corresponding charge / discharge efficiency correction coefficient of 0.8.

[0222] Table 1

[0223]

[0224]

[0225] The embodiments in this application are merely illustrative examples of the target mapping table described in Table 1 above, and do not constitute a limitation on this application.

[0226] Thus, with a target mapping table pre-stored by technicians in the computer equipment, when the remaining battery power changes, the actual charge / discharge efficiency coefficient under the given driving parameters, battery state, and environmental parameters is retrieved from the target mapping table. This allows for a more accurate charge / discharge efficiency correction coefficient, which in turn enables the vehicle to determine a more accurate remaining battery power.

[0227] Furthermore, after the vehicle receives a warning message carrying the charge-discharge efficiency correction coefficient, the vehicle can correct the remaining charge of the power battery based on the charge-discharge efficiency correction coefficient.

[0228] Optionally, the vehicle's BMS can adjust the remaining charge of the power battery based on the charge / discharge efficiency correction coefficient.

[0229] After the vehicle's BMS receives a warning message carrying a charge / discharge efficiency correction coefficient, it can first obtain the instantaneous current of the power battery, and then correct the remaining power of the power battery based on the instantaneous current of the power battery, the rated capacity of the power battery, and the charge / discharge efficiency correction coefficient.

[0230] Specifically, the vehicle's BMS can correct the remaining power of the power battery based on the instantaneous current of the power battery, the remaining power of the power battery at the previous moment, the rated capacity of the power battery, and the charge and discharge efficiency correction coefficient, using the following formula (3).

[0231]

[0232] Where SOC is the remaining charge of the power battery after correction, and SOC0 is the remaining charge of the power battery at the previous moment. E The rated capacity of the power battery is given by η, where η is the charge / discharge efficiency correction coefficient, and I(t) is the instantaneous current of the power battery.

[0233] In this way, the BMS can correct the remaining power of the power battery, that is, calculate a more accurate remaining power of the power battery, thereby ensuring a more accurate remaining range in the future.

[0234] Furthermore, after correcting for the remaining charge of the power battery, the vehicle's remaining range can be determined based on the corrected remaining charge. This allows for a more accurate determination of the vehicle's remaining range, thereby improving the accuracy of the remaining range determination.

[0235] Specifically, the vehicle can determine its remaining range based on the remaining charge of the modified power battery using the following formula (4).

[0236]

[0237] Among them, S 剩 S represents the vehicle's remaining mileage. 总 State of Charge (SOC) is the vehicle's range on a full charge, meaning the distance the vehicle can travel when the battery is fully charged. 总 This refers to the total charge of the power battery.

[0238] For example, if the vehicle's full-charge range is 500 km, the total charge of the power battery is 100%, and the remaining charge of the power battery after correction is 80%, then the remaining range of the vehicle calculated by the above formula (4) is 400 km.

[0239] Furthermore, after determining the vehicle's remaining mileage, the remaining mileage can also be displayed.

[0240] Optionally, the vehicle can display the remaining mileage on the dashboard, or on the HUD (Head-Up Display), or on other display devices in the vehicle. This application embodiment does not limit this.

[0241] In this way, after obtaining the accurate remaining mileage, displaying the remaining mileage of the vehicle allows users to accurately know how much mileage the vehicle can still travel, thereby improving the user experience.

[0242] In this embodiment, the computer device first acquires the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's environment—that is, it acquires parameters that can affect the remaining charge of the power battery. Then, based on these driving parameters, battery status, and environmental parameters, it determines whether the remaining charge of the vehicle's power battery will change after a preset time period. Finally, if the remaining charge of the power battery changes after the preset time period, a warning message is sent to the vehicle—a charge change warning—so that the vehicle can subsequently correct the remaining charge. Thus, by sending a warning message to the vehicle when it is determined that the remaining charge will change after a preset time period, the vehicle can promptly correct the remaining charge, ensuring accurate determination and display of the vehicle's remaining range, thereby improving the user experience.

[0243] Figure 5 This is a schematic diagram of a power battery charge level jump warning device provided in an embodiment of this application. The power battery charge level jump warning device can be implemented as part or all of a computer device by software, hardware, or a combination of both. This computer device can be described below. Figure 6 The computer equipment shown. See also Figure 5 The device includes: a first acquisition module 501, a determination module 502, and an early warning module 503.

[0244] The first acquisition module 501 is used to acquire the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's environment.

[0245] The determination module 502 is used to determine, based on the driving parameters, the battery status and the environmental parameters, whether the remaining charge of the vehicle's power battery changes after a preset time.

[0246] The warning module 503 is used to send a warning message to the vehicle when the remaining power of the power battery changes abruptly after a preset time. The warning message is used to instruct the vehicle to correct the remaining power of the battery.

[0247] Optionally, the early warning module 503 is used for:

[0248] If the remaining charge of the power battery changes abruptly after a preset time, the charging and discharging efficiency correction coefficient of the power battery is determined based on the driving parameters, the battery state, and the environmental parameters.

[0249] Send a warning message to the vehicle carrying the charge / discharge efficiency correction factor to instruct the vehicle to correct the remaining charge of the power battery based on the charge / discharge efficiency correction factor.

[0250] Optionally, the determining module 502 is used for:

[0251] The current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature are input into the battery power jump prediction model. The model processes the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and outputs the probability that the remaining battery power will jump after a preset time.

[0252] If the probability is greater than or equal to a preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will change after a preset time.

[0253] If the probability is less than a preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will not change after a preset time.

[0254] Optionally, the power jump prediction model includes multiple leaf nodes, and the determination module 502 is used for:

[0255] The multiple leaf nodes are used to make decisions and judgments on the current vehicle speed, current mileage, current remaining battery power, current health status, and current temperature, and output the probability that the remaining battery power will change after a preset time.

[0256] Optionally, the device further includes:

[0257] The second acquisition module is used to acquire the first training dataset, which includes multiple sets of sample data and the labels of the multiple sets of sample data. Each set of sample data includes the values ​​of multiple feature attributes, including historical driving parameters, historical battery status, and historical environmental parameters. The label is used to indicate whether the remaining power of the power battery changes within the target time period under this set of sample data.

[0258] The feature extraction module is used to extract features from any one set of sample data in the multiple sets of sample data to obtain the features of the sample data.

[0259] The concatenation module is used to concatenate the features of this set of sample data with this set of sample data to obtain a second training dataset. The second training dataset includes the multiple sets of sample data, the features of the multiple sets of sample data, and the labels of the multiple sets of sample data.

[0260] The training module is used to train the power jump prediction model based on the second training dataset.

[0261] Optionally, the multiple feature attributes include continuous and discrete feature attributes, and the feature extraction module is used for:

[0262] For any continuous feature attribute in this set of sample data, determine the mean, maximum, minimum and variance of the continuous feature attribute in this set of sample data;

[0263] For any discrete feature attribute in this set of sample data, determine the median corresponding to the discrete feature attribute in this set of sample data.

[0264] Optionally, the training module is used for:

[0265] For the root node in the first level of n levels, the second training dataset is input into the root node, the information gain of multiple feature attributes under the root node is determined, and the second training dataset is divided based on the feature attribute with the largest information gain, so as to obtain multiple leaf nodes based on the root node split.

[0266] For each leaf node in the i-th level of the n levels, determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node by the feature attribute with the largest information gain, and form multiple leaf nodes in the (i+1)-th level based on the current leaf node, where i is an integer greater than or equal to 2 and less than or equal to n.

[0267] Let i = i+1, and perform the following steps for each leaf node in the i-th level of the n levels: determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node by the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the i+1-th level, until the predicted jump probability is determined based on the dataset contained in the split leaf nodes.

[0268] Based on the difference between the predicted jump probability and the label, the parameters of the power jump prediction model are adjusted.

[0269] Optionally, the second acquisition module is used for:

[0270] Acquire historical vehicle data, which includes historical driving parameters, historical battery status, and historical environmental parameters;

[0271] The historical vehicle data is divided into multiple sets of sample data by using the target time period as the time interval. Each set of sample data includes the historical driving parameters, historical battery status, and historical environmental parameters for the corresponding time period.

[0272] For any set of sample data from the multiple sets of sample data, the label of the sample data is determined based on the historical battery status of the corresponding time period.

[0273] Optionally, the second acquisition module is used for:

[0274] Determine the difference between the historical remaining electricity at the start time and the historical remaining electricity at the end time of the corresponding time period;

[0275] If the power difference is greater than or equal to a preset power threshold, the label of the sample data is determined to be a jump.

[0276] If the difference in battery power is less than the preset battery power threshold, the label of the sample data is determined to be "no change".

[0277] In this embodiment, the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's environment are first acquired; that is, parameters that can affect the remaining charge of the power battery are acquired. Then, based on these driving parameters, battery status, and environmental parameters, it is determined whether the remaining charge of the vehicle's power battery will change after a preset time period. Finally, if the remaining charge of the power battery changes after the preset time period, a warning message is sent to the vehicle, i.e., a charge change warning is given, allowing the vehicle to subsequently correct the remaining charge. Thus, by sending a warning message to the vehicle when it is determined that the remaining charge will change after a preset time period, the vehicle can promptly correct the remaining charge, ensuring accurate determination and display of the vehicle's remaining range, thereby improving the user experience.

[0278] It should be noted that the above-described power battery charge jump warning device provides an example of the division of the above functional modules when it warns of a jump in the remaining power of the power battery. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0279] The functional units and modules in the above embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of the embodiments of this application.

[0280] The power battery power jump warning device and the power battery power jump warning method provided in the above embodiments belong to the same concept. The specific working process and technical effects of the units and modules in the above embodiments can be found in the method embodiments section, and will not be repeated here.

[0281] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 6 As shown, the computer device 6 includes a processor 60, a memory 61, and a computer program 62 stored in the memory 61 and executable on the processor 60. When the processor 60 executes the computer program 62, it implements the steps in the power battery charge jump warning method in the above embodiment.

[0282] Computer device 6 can be a server, which can be a standalone server or a server cluster composed of multiple standalone servers. This application does not limit the specific type of computer device 6. Those skilled in the art will understand that... Figure 6 The computer device 6 is merely an example and does not constitute a limitation on the computer device 6. It may include more or fewer components than shown in the figure, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0283] Processor 60 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0284] In some embodiments, memory 61 may be an internal storage unit of the computer device 6, such as a hard disk or RAM of the computer device 6. In other embodiments, memory 61 may be an external storage device of the computer device 6, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD) card, flash card, etc., provided on the computer device 6. Furthermore, memory 61 may include both internal storage units and external storage devices of the computer device 6. Memory 61 is used to store the operating system, applications, boot loader, data, and other programs. Memory 61 can also be used to temporarily store data that has been output or will be output.

[0285] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the various method embodiments described above.

[0286] This application provides a computer program product that, when run on a computer, causes the computer to perform the steps described in the various method embodiments above.

[0287] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above method embodiments of this application can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographing device / terminal device, a recording medium, a computer memory, ROM (Read-Only Memory), RAM (Random Access Memory), CD-ROM (Compact Disc Read-Only Memory), magnetic tape, floppy disk, and optical data storage devices. The computer-readable storage medium mentioned in this application can be a non-volatile storage medium; in other words, it can be a non-transient storage medium.

[0288] It should be understood that all or part of the steps of the above embodiments can be implemented by software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented in whole or in part as a computer program product. The computer program product includes one or more computer instructions. The computer instructions can be stored in the above-described computer-readable storage medium.

[0289] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0290] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0291] In the embodiments provided in this application, it should be understood that the disclosed apparatus / computer devices and methods can be implemented in other ways. For example, the apparatus / computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0292] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0293] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for early warning of charge level fluctuations in a power battery, characterized in that, The method includes: Obtain the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's surroundings; Based on the driving parameters, the battery status, and the environmental parameters, determine whether the remaining charge of the vehicle's power battery changes abruptly after a preset time. If the remaining charge of the power battery changes abruptly after a preset period of time, a warning message is sent to the vehicle, and the warning message is used to instruct the vehicle to correct the remaining charge. Based on the driving parameters, the battery status, and the environmental parameters, determining whether the remaining charge of the vehicle's power battery changes abruptly after a preset time period includes: The current vehicle speed and current mileage, the current remaining charge and current health status of the power battery, and the current temperature of the environment in which the vehicle is located are input into the power jump prediction model. The power jump prediction model processes the current vehicle speed, current mileage, current remaining charge, current health status, and current temperature, and outputs the probability that the remaining charge of the power battery will jump after a preset time. If the probability is greater than or equal to a preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will change after a preset time.

2. The method as described in claim 1, characterized in that, When the remaining charge of the power battery changes abruptly after a preset period of time, a warning message is sent to the vehicle, including: If the remaining charge of the power battery changes abruptly after a preset time, a correction coefficient for the charging and discharging efficiency of the power battery is determined based on the driving parameters, the battery state, and the environmental parameters. Send a warning message carrying the charge / discharge efficiency correction coefficient to the vehicle to instruct the vehicle to correct the remaining charge of the power battery based on the charge / discharge efficiency correction coefficient.

3. The method as described in claim 1, characterized in that, The method further includes: If the probability is less than the preset probability threshold, it is determined that the remaining charge of the vehicle's power battery will not change after a preset time period.

4. The method as described in claim 3, characterized in that, The training method for the power level jump prediction model includes: Obtain a first training dataset, which includes multiple sets of sample data and labels for the multiple sets of sample data. Each set of sample data includes the values ​​of multiple feature attributes, including historical driving parameters, historical battery status, and historical environmental parameters. The labels are used to indicate whether the remaining power capacity of the power battery changes within a target time period under the sample data. For any set of sample data among the multiple sets of sample data, feature extraction is performed on the sample data to obtain the features of the sample data; The features of the sample data are concatenated with the sample data to obtain a second training dataset, which includes the multiple sets of sample data, the features of the multiple sets of sample data, and the labels of the multiple sets of sample data. The power fluctuation prediction model is trained based on the second training dataset.

5. The method as described in claim 4, characterized in that, The multiple feature attributes include continuous and discrete feature attributes. The feature extraction of the sample data to obtain the features of the sample data includes: For any continuous type feature attribute in the sample data, determine the mean, maximum, minimum and variance of the continuous type feature attribute in the sample data; For any discrete feature attribute in the sample data, determine the median corresponding to the discrete feature attribute in the sample data.

6. The method as described in claim 4, characterized in that, The step of training the battery level jump prediction model based on the second training dataset includes: For the root node in the first level of n levels, the second training dataset is input into the root node, the information gain of multiple feature attributes under the root node is determined, and the second training dataset is divided based on the feature attribute with the largest information gain, so as to obtain multiple leaf nodes based on the root node. For each leaf node in the i-th level of the n levels, determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node by the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the (i+1)-th level, where i is an integer greater than or equal to 2 and less than or equal to n. Let i = i+1, and execute the following steps for each leaf node in the i-th level of the n levels: determine the information gain of multiple feature attributes under each leaf node, divide the dataset contained in the current leaf node with the feature attribute with the largest information gain, and split the current leaf node to form multiple leaf nodes in the i+1-th level, until the predicted jump probability is determined based on the dataset contained in the split leaf nodes. Based on the difference between the predicted jump probability and the tag, the parameters of the power jump prediction model are adjusted.

7. The method as described in claim 4, characterized in that, The process of obtaining the first training dataset includes: Acquire historical vehicle data, which includes historical driving parameters, historical battery status, and historical environmental parameters; The historical vehicle data is divided into multiple sets of sample data by using the target time period as the time interval. Each set of sample data includes historical driving parameters, historical battery status, and historical environmental parameters for the corresponding time period. For any set of sample data from the multiple sets of sample data, the label of the sample data is determined based on the historical battery status of the corresponding time period.

8. The method as described in claim 7, characterized in that, The historical battery status of the corresponding time period in the sample data includes the historical remaining battery power within the corresponding time period. Determining the label of the sample data based on the historical battery status of the corresponding time period includes: Determine the difference between the historical remaining electricity at the start time and the historical remaining electricity at the end time of the corresponding time period; If the power difference is greater than or equal to a preset power threshold, the label of the sample data is determined to be a jump. If the power difference is less than the preset power threshold, the label of the sample data is determined to be "no change".

9. A power battery charge level jump warning device, characterized in that, The device includes: The first acquisition module is used to acquire the vehicle's current driving parameters, battery status, and environmental parameters of the vehicle's environment. The determination module is used to determine, based on the driving parameters, the battery status, and the environmental parameters, whether the remaining charge of the vehicle's power battery changes abruptly after a preset time. The warning module is used to send a warning message to the vehicle when the remaining power of the power battery changes abruptly after a preset time. The warning message is used to instruct the vehicle to correct the remaining power of the battery. The determining module is specifically used to: input the vehicle's current speed and current mileage, the current remaining charge and current health status of the power battery, and the current temperature of the vehicle's environment into the power battery jump prediction model; process the current speed, current mileage, current remaining charge, current health status, and current temperature through the power battery jump prediction model; and output the probability that the remaining charge of the power battery will jump after a preset time; if the probability is greater than or equal to a preset probability threshold, determine that the remaining charge of the vehicle's power battery will jump after a preset time.

10. A computer device, characterized in that, The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the method as described in any one of claims 1 to 8.

Images