WO2022183817A1

WO2022183817A1 - Temperature consistency prediction method and apparatus, prediction device, and storage medium

Info

Publication number: WO2022183817A1
Application number: PCT/CN2021/141562
Authority: WO
Inventors: 付振; 梁小明; 王明月; 彭凯; 刘相超; 王文彬; 袁鲁峰; 邵天东; 丁聪敏; 孙建蕾
Original assignee: 中国第一汽车股份有限公司
Priority date: 2021-03-01
Filing date: 2021-12-27
Publication date: 2022-09-09
Also published as: CN113022378B; CN113022378A

Abstract

A temperature consistency prediction method and apparatus, a prediction device, and a storage medium. The method comprises: collecting vehicle state data, the vehicle state data comprising temperature data collected by a common temperature sensor between multiple cells in each set of module in multiple sets of modules and time sequence information; and predicting temperature consistency among the multiple sets of modules according to the vehicle state data by means of a prediction model. The apparatus, prediction device, and storage medium comprise programs executing the method. The phenomenon of battery temperature imbalance can be found in time by means of the method.

Description

Temperature consistency prediction method, device, prediction device and storage medium

This application claims the priority of a Chinese patent application with application number 202110227426.0 filed with the China Patent Office on March 1, 2021, the entire contents of which are incorporated herein by reference.

technical field

The embodiments of the present application relate to the technical field of vehicle monitoring, for example, to a temperature consistency prediction method, device, prediction device, and storage medium.

Background technique

With the rapid development of the Internet of Vehicles and new energy vehicles, many types of failures of new energy vehicles are caused by thermal runaway of power batteries, and the cause of thermal runaway is due to abnormal temperature, so the temperature monitoring of power batteries is particularly important. The power battery includes multiple single cells. The temperature consistency between different single cells is an important indicator to ensure the safety of the power battery. The temperature consistency indirectly reflects whether the power battery is in a normal working state. It is too late to make an alarm or adjust the voltage and current after monitoring the large temperature difference between different single cells, and the occurrence of faults cannot be avoided, which has a great potential safety hazard.

SUMMARY OF THE INVENTION

The present application provides a temperature consistency prediction method, device, prediction device and storage medium, so as to predict the temperature consistency between modules, so as to facilitate the timely detection of temperature imbalance, improve the safety of power batteries, and achieve fault isolation. .

An embodiment of the present application provides a method for predicting temperature consistency, including: collecting vehicle state data, where the vehicle state data includes a shared temperature between a plurality of single cells in each group of modules through a plurality of groups of modules Temperature data and time series information collected by sensors; temperature consistency among the multiple groups of modules is predicted according to the vehicle state data through a prediction model.

The method further includes: acquiring historical temperature data and historical time series information; and constructing the prediction model according to the historical temperature data and the historical time series information.

Using the prediction model to predict the temperature consistency between the multiple groups of modules according to the vehicle state data includes: using the prediction model and according to the time series information, sequentially calculating the temperature consistency of the multiple groups of modules in each The difference between the highest temperature and the lowest temperature at the moment; when the difference is greater than or equal to the first threshold, the count value of the counter is incremented by 1 on the basis of the count value at the previous moment at each moment, and then When the difference is less than the first threshold, the count value of the counter is cleared; when the count value of the counter reaches the second threshold, it is determined that the temperatures among the multiple groups of modules are inconsistent.

Using the prediction model to predict the temperature consistency among the multiple groups of modules according to the vehicle state data includes: extracting the temperature data using a set sliding window according to the time series information through the prediction model The characteristics of the temperature data include the temperature of the multiple groups of modules at different times; the temperature consistency between the multiple groups of modules is predicted according to the characteristics of the temperature data.

Predicting the temperature consistency among the multiple groups of modules according to the characteristics of the temperature data includes: classifying the characteristics of the temperature data based on the reinforcement learning AdaBoost algorithm to predict the temperature among the multiple groups of modules consistency.

The vehicle status data also includes at least one of the following: vehicle code, cell temperature, cell voltage, cell current, charging and discharging status, charging and discharging current, charging and discharging voltage, vehicle working status, cell voltage extreme value, Body current extreme value, cell temperature extreme value, battery management system (Battery Management System, BMS) alarm information.

The method further includes: preprocessing the vehicle state data; the preprocessing includes at least one of the following: data deduplication; unavailable value processing; abnormal data processing; alarm correction; current correction; voltage correction; temperature correction.

An embodiment of the present application provides a temperature consistency prediction device, including: a collection module configured to collect vehicle state data, where the vehicle state data includes passing through a plurality of single cells in each group of modules in a plurality of groups of modules The temperature data and time series information collected by the shared temperature sensor between them; the prediction module is configured to predict the temperature consistency among the multiple groups of modules according to the vehicle state data through a prediction model.

An embodiment of the present application provides a prediction device, including: one or more processors; a memory configured to store one or more programs; when the one or more programs are executed by the one or more processors, The one or more processors are caused to implement the temperature consistency prediction method as described in the first aspect.

Embodiments of the present application further provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, implements the temperature consistency prediction method described in the first aspect.

Description of drawings

1 is a flowchart of a method for predicting temperature consistency provided in Embodiment 1 of the present application;

FIG. 2 is a schematic diagram of deploying a temperature sensor according to Embodiment 1 of the present application;

3 is a flowchart of a temperature consistency prediction method provided in Embodiment 2 of the present application;

4 is a flowchart of a temperature consistency prediction method provided in Embodiment 3 of the present application;

5 is a schematic diagram of the realization of a temperature consistency prediction process provided in Embodiment 2 of the present application;

6 is a schematic structural diagram of a temperature consistency prediction device provided in Embodiment 4 of the present application;

FIG. 7 is a schematic diagram of a hardware structure of a prediction device according to Embodiment 5 of the present application.

Detailed ways

The present application will be described below with reference to the accompanying drawings and embodiments. It should be understood that the embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that, for the convenience of description, the drawings only show some but not all the structures related to the present application.

Before discussing the exemplary embodiments, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowchart depicts various steps as a sequential process, many of the steps may be performed in parallel, concurrently, or concurrently. Furthermore, the order of the various steps can be rearranged. The process may be terminated when multiple operations are completed, but may also have additional steps not included in the figures. The processing may correspond to a method, function, procedure, subroutine, subroutine, or the like.

It should be noted that concepts such as "first" and "second" mentioned in the embodiments of the present application are only used to distinguish different devices, modules, units or other objects, and are not used to limit these devices, modules, units or The order or interdependence of functions performed by other objects.

Example 1

FIG. 1 is a flowchart of a temperature consistency prediction method provided in Embodiment 1 of the present application. This embodiment can be applied to the situation of predicting the temperature consistency of the power battery module. The temperature consistency prediction method can be executed by a temperature consistency prediction device, and the temperature consistency prediction device can be implemented by means of software and/or hardware, and is integrated into a prediction device. Predictive devices include, but are not limited to, electronic devices such as desktop computers, trip computers, smart phones, Internet of Vehicles servers, and cloud servers.

As shown in Figure 1, the method includes the following steps.

S110. Collect vehicle state data, where the vehicle state data includes temperature data and time sequence information collected through a common temperature sensor between a plurality of single cells in each group of modules in a plurality of groups of modules.

Power batteries mainly refer to batteries that provide power for electric vehicles, usually including multiple single cells to provide high voltage for vehicles. Among them, multiple single cells can form a group of modules. Modularization and standardization of power batteries can improve heat dissipation efficiency and reduce the risk of thermal runaway. In this embodiment, a corresponding temperature sensor is deployed for each group of modules, and the temperature sensor is set to collect the temperature data of the corresponding module, for example, once a minute, and the collected temperature data is related to the time series information, including The time of the second collection and the frequency of data collection, etc. It should be noted that a group of modules includes at least two single cells. Usually, the temperature difference between multiple cells in a group of modules is small and can be approximately regarded as equal. Therefore, each group of cells has a small temperature difference. The single cells in the module can share the same temperature sensor, and the temperature data is collected in units of modules.

FIG. 2 is a schematic diagram of deploying a temperature sensor according to Embodiment 1 of the present application. As shown in Figure 2, two single cells form a group of modules, and two single cells share a temperature sensor. On this basis, the number of temperature sensors deployed can be reduced. Each temperature sensor is set to collect adjacent cells. temperature data of the battery, thereby reducing costs and improving the efficiency of data acquisition and temperature consistency prediction.

S120. Predict the temperature consistency among the multiple groups of modules according to the vehicle state data by using a prediction model.

The prediction model can be a machine learning model. After training with a large amount of sample data, the prediction model has learned the rules for determining the temperature consistency of multiple groups of modules. For the input temperature data and time series information, the prediction results can be output efficiently and accurately. In order to predict the temperature consistency between multiple groups of modules, it is convenient to detect abnormalities in time. The rule for judging the temperature consistency of multiple groups of modules is, for example, if the temperature difference (the difference between the maximum temperature and the minimum temperature) between the multiple groups of modules at a time is greater than the set threshold, then determine the temperature between the multiple groups of modules. If it does not meet the consistency, it means that the power battery has a fault or fault trend, and an alarm is required; it can also be that the temperature difference between multiple groups of modules is continuously greater than the set threshold within a set period, or the temperature data collected at multiple consecutive times are all Indicates that the temperature difference between multiple groups of modules is greater than the set threshold, then it is determined that the temperature between multiple groups of modules does not conform to the consistency; it can also be based on the temperature of multiple groups of modules at adjacent times or between adjacent time periods. Factors such as the degree of jumping, the standard deviation or variance of the temperature of multiple groups of modules, etc. determine whether the temperature among multiple groups of modules conforms to the consistency.

Optionally, the vehicle status data collected in real time can be collected through the Controller Area Network (CAN) bus and uploaded to the Internet of Vehicles platform by the Telematics BOX (T-BOX). State data is stored to a data storage system (data lake). The prediction equipment of the Internet of Vehicles platform uses the prediction model to predict the temperature consistency of the vehicle state data. The prediction result can be that the temperature between multiple groups of modules conforms to the temperature consistency or does not conform to the temperature consistency (for example, output "0" indicates that the consistency is met, and output "1" indicates that the consistency is not met), or it can be multiple groups of modules. The probability that the temperature between groups is consistent with or not consistent with temperature (eg the output is a value belonging to the interval [0,1]). In some embodiments, it is also possible to predict the time when the temperature of the multiple groups of modules will change from conforming to non-consistent in a set period of time in the future according to the change trend of the temperature of the multiple groups of modules, so as to implement predictive maintenance (Predictive Maintenance, PdM), which is convenient for making decisions in advance and avoiding risks.

In one embodiment, the vehicle status data may further include at least one of the following: vehicle codes, such as vehicle frame number, vehicle identification, identification number ((Identity document), ID), etc., used to distinguish different vehicles; monomer temperature , that is, the temperature of each single battery in the module; the single voltage, that is, the voltage of each single battery in the module; the single current, that is, the current of each single battery in the module; the charging and discharging state, including The charging state or discharging state of the power battery; the charging and discharging current, including the charging current or discharging current of the power battery; the charging and discharging voltage, including the charging voltage or discharging voltage of the power battery; the working state of the vehicle, including the flameout state, the driving state and the reversing state and the idle state, etc.; the extreme value of the single cell voltage, that is, the highest voltage and/or the lowest voltage of the single cell in the power battery or each group of modules; the extreme value of the single cell current, that is, in the power battery or in each group of modules The maximum current and/or the minimum current of the single battery; the extreme value of the single temperature, that is, the maximum temperature and/or the minimum temperature of the single battery in the power battery or in each group of modules; the battery management system (Battery Management System, BMS) alarm information, such as the alarm information of energy imbalance between single cells, and the abnormal alarm information of single cells or modules.

Based on the above vehicle status data, it can not only predict the consistency of module temperature and prevent failures, but also monitor the working parameters of the power battery in real time. When the temperature of multiple groups of modules does not meet the consistency, it can also quickly locate the abnormal position and analyze the fault. The reason is convenient for troubleshooting in time, so as to ensure the safety of the power battery. It should be noted that the voltage sensor, current sensor, etc. may be deployed separately for multiple single cells, which is convenient for accurately locating abnormal single cells and ensures the safety and reliability of the power battery. A temperature consistency prediction method provided in the first embodiment of the present application, by building an artificial intelligence (Artificial Intelligence, AI) model, according to the collected known data can predict the temperature consistency between multiple groups of modules, to a certain extent The safety hazards caused by the unbalanced temperature of the power battery cells are eliminated. The method combines artificial intelligence technology to achieve power battery failure prediction and health management (Prognostics and Health Management, PHM), can use known information to predict the time or probability of power battery failure, etc., can provide preventive maintenance and maintenance decisions for vehicles Reference, reduce maintenance costs and failure probability. By predicting the temperature consistency between modules from the time dimension and the dimension of different batteries, it is convenient to detect temperature imbalance in time, improve the safety of power batteries, and achieve fault isolation to a certain extent.

Embodiment 2

FIG. 3 is a flowchart of a temperature consistency prediction method provided in Embodiment 2 of the present application. This embodiment is described on the basis of the above-mentioned embodiment, and the process of temperature consistency prediction is described. It should be noted that, for technical details not described in this embodiment, reference may be made to any of the foregoing embodiments.

As shown in Figure 3, the method includes the following steps.

S210. Collect vehicle state data.

The vehicle state data includes temperature data and time series information of each group of modules, for example, includes temperature data of each group of modules within 20 minutes, and each minute corresponds to a group of temperature data. It should be noted that the vehicle status data can be collected and predicted in real time, that is, if the temperature data is collected once at the current moment, the prediction model can be based on the temperature data and time series information collected in the previous 19 minutes, combined with the temperature data at the current moment. And time series information for comprehensive prediction, the temperature data and time series information for the first 19 minutes can be downloaded from the cloud platform.

S220. Calculate the difference between the highest temperature and the lowest temperature of the multiple groups of modules at each moment in turn according to the time series information through the prediction model.

In this embodiment, the prediction model sequentially calculates the difference between the highest temperature and the lowest temperature among the temperatures of the multiple groups of modules at each moment according to the time series information, that is, the sequence of temperature data collection. At this moment, the temperature among the multiple groups of modules is relatively balanced, which is consistent with the temperature consistency; if the difference is large, it means that the prediction result at this moment shows that the temperature among the multiple groups of modules is relatively unbalanced and does not meet the temperature consistency. stability, or a tendency to have inconsistent temperatures, could easily lead to failures due to thermal runaway in the future.

S230. For each moment, if the difference is greater than or equal to the first threshold, execute S240; if the difference is smaller than the first threshold, execute S250.

S240. The count value of the counter is incremented by 1 on the basis of the count value at the previous moment at each moment.

S250, the count value of the counter is cleared.

The first threshold is used to determine whether the temperatures of the multiple groups of modules at one moment meet the temperature consistency. In some embodiments, if the difference between the highest temperature and the lowest temperature among the temperatures of the multiple groups of modules at one moment reaches a first threshold, it may be determined that the predicted results are that the temperatures of the multiple groups of modules are inconsistent. In this embodiment, in order to avoid misjudgment caused by instantaneous temperature instability and improve the reliability of prediction, the number of times when the temperature difference of multiple groups of modules reaches the first threshold is also recorded by a counter. Timing length) When the temperature difference of the multiple groups of modules continues to reach the first threshold, the final judgment is that the temperature of the multiple groups of modules is inconsistent.

S260. If the count value of the counter reaches the second threshold, execute S290; if the count value of the counter does not reach the second threshold, return to S230.

If the difference between the highest temperature and the lowest temperature at multiple consecutive times reaches the first threshold, the count value +1; if the difference between the highest temperature and the lowest temperature at a moment is less than the first threshold, the count value is cleared until the count value When the second threshold is reached, it is finally determined that the prediction result is that the temperatures of multiple groups of modules are inconsistent. By setting the second threshold, misjudgment caused by instantaneous temperature instability can be avoided, and the reliability of prediction can be improved.

S270. It is determined that the temperatures among the multiple groups of modules are inconsistent.

The following is another example. In the case of collecting vehicle state data in real time and predicting temperature consistency in real time, the implementation process of temperature consistency prediction includes: Step1: Define two variables: count (count value), and index[] (Temperature difference time series index container), initialized to 0 and [] respectively; Step2: Put the current real-time collected temperature data and the corresponding time series information into the temperature difference time series index container; Step3: Calculate the highest temperature of the temperature of multiple groups of modules at the current moment The difference between the minimum temperature and the minimum temperature, if the difference is greater than or equal to the first threshold, count+1; if the difference is less than the first threshold, the count is initialized to 0; Step4: Whether the count reaches the second threshold, if the count reaches the second threshold Threshold, the output prediction result is that the temperature between multiple groups of modules is inconsistent; if the count does not reach the second threshold, then index[]+1, continue to collect the temperature data collected at the next moment at the current moment and the corresponding The timing information is put into the temperature difference timing index container, and returns to Step3.

On the basis that the prediction result at one moment is that the temperature among the multiple groups of modules is inconsistent, the difference between the highest temperature and the lowest temperature of the temperature of the multiple groups of modules at any subsequent time is greater than or equal to the first threshold, which is regarded as a The temperature between groups of modules does not meet the consistency.

In one embodiment, the vehicle state data further includes at least one of the following: vehicle code, cell temperature, cell voltage, cell current, charging and discharging status, charging and discharging current, charging and discharging voltage, vehicle working status, cell voltage Extreme value, cell current extreme value, cell temperature extreme value, battery management system alarm information.

In one embodiment, the method further includes: preprocessing the vehicle state data; the preprocessing includes at least one of the following: data deduplication; unavailable value processing; abnormal data processing; alarm correction; current correction; voltage correction; temperature Correction.

The sensors in the vehicle collect vehicle state data and upload it to the prediction device. The prediction device performs at least one of the following preprocessing on the vehicle state data before building a prediction model or predicting temperature consistency to improve data quality: data deduplication, For example, to de-duplicate the data of the repeated time in the vehicle status data, the repeated time may be caused by the delay in uploading the timing information. When one time corresponds to multiple repeated temperature data, only one of all the repeated temperature data is retained; Use value (NaN value) processing to delete all the individual voltage data and all temperature measurement point data in the vehicle status data, the data at the time containing the NaN value; abnormal data processing, the vehicle status data may have a part of the time series. If the data is abnormal, for example, the month is not within 1-12, the time is not within 0:00-24:00, etc., these abnormal data can be deleted; alarm correction, the alarm (Mask) field in the vehicle status data is in decimal However, the alarm field set in the alarm field comparison table may be in binary form, for example, "0" indicates that a single battery is normal, "1" indicates that a single battery is abnormal and triggers an alarm, therefore, the vehicle can be The value of the alarm field in the status data is converted into a binary value; the current correction, due to the working process of the vehicle's controller and components, and the possible instability, jump or system error in the data transmission process, cause the monomer The extreme value of current and cell current may deviate from the measured value, so the current can be corrected, for example, the maximum value of cell current and the minimum value of cell current at each moment are recalculated, instead of uploading to the prediction The extreme value of the cell current of the device, or re-estimate the current at the middle time according to the current before and after the set time period; for voltage correction, the cell voltage and the extreme value of the cell voltage may deviate from the measured value , so the voltage is corrected, for example, the maximum value of the cell voltage and the minimum value of the cell voltage at each moment are recalculated to replace the extreme value of the cell voltage uploaded to the prediction device, or, according to the set time period The voltage at the time before and after the inner time is re-estimated the voltage at the middle time, etc.; temperature correction, the temperature of the cell and the extreme value of the cell temperature may deviate from the measured value, so correct the temperature, for example, recalculate the cell temperature at each time. The maximum value of the body temperature and the minimum value of the monomer temperature are used to replace the extreme value of the monomer temperature uploaded to the prediction device, or the temperature at the middle time is re-estimated according to the temperature at the time before and after the set time period; Processing may also include data combination, disassembly, extraction, and integration, among others. For example, the cell voltage and cell temperature in the vehicle status data are stored in the listcell and listtenp lists, and are stored in the form of strings in the list, but for subsequent actual use, it may be necessary to expand into a tabular data structure (DataFrame) , so it is necessary to disassemble and extract the listcell and listtemp; for another example, it is necessary to combine the data of multiple moments according to a specific time window, and also perform operations such as averaging and variance; another example, the processed The vehicle status data can be integrated into files in npy or pickle format, which occupies less memory, saves and reads quickly, and can improve data processing efficiency.

In addition, preprocessing can also include edge computing to reduce data transmission and storage pressure. Edge computing is performed on the data collected by the vehicle, that is, data is encoded, packaged, or compressed. For example, when there are many types of alarms, uploading one by one will put more pressure on the storage. By encoding the alarm information into binary form, cloud storage resources can be effectively saved; for example, due to the high frequency of data collection at the vehicle end, the millisecond level , the original data can be processed at low frequency to reduce the amount of data while ensuring the diversity of data.

In a temperature consistency prediction method provided in the second embodiment of the present application, the counter records the number of times when the difference between the highest temperature and the lowest temperature among the temperatures of multiple groups of modules reaches a first threshold, so as to avoid instantaneous temperature instability. Misjudgment and improving the reliability of predictions; by preprocessing the original data, the amount of data can be reduced, unnecessary calculations can be avoided, and data quality can be improved, thereby improving the accuracy and reliability of predictions, and facilitating the timely detection of temperature imbalances , thereby improving the safety of the power battery and achieving fault isolation.

Embodiment 3

FIG. 4 is a flowchart of a temperature consistency prediction method provided in Embodiment 3 of the present application. This embodiment is described on the basis of the above-mentioned embodiment, and describes the working principle of the prediction model. It should be noted that, for technical details not described in this embodiment, reference may be made to any of the foregoing embodiments.

As shown in Figure 4, the method includes the following steps.

S310 , acquiring historical temperature data and historical time series information.

The historical temperature data and the prediction results corresponding to the historical time series information are known, that is, the labels are known, so they can be used as sample data for the prediction model.

S320. Build a prediction model according to historical temperature data and historical time series information.

The historical temperature data and historical time series information are used as input, and the corresponding historical prediction results are used as output to train the prediction model, so that the prediction model can learn the law of predicting temperature consistency according to the temperature data and time series information.

In one embodiment, after the training of the prediction model is completed, the performance of the prediction model is also tested. For example, divide all sample data into training data set and test data set according to the ratio of 7:3, and use the sample data in the test data set to test the performance of the prediction model. The ratio of temperature consistency samples) data is 10:1, and the evaluation indicators of the test can be precision rate, recall rate, precision rate and/or receiver operating characteristic curve (Receiver Operating Characteristic Curve, ROC) and the coordinate axis enclosed Area (Area Under Curve, AUC). In some embodiments, the actual vehicle data can also be verified by a real vehicle when the test result is ideal, and the prediction model can be put into practical application only if the real vehicle verification effect is ideal.

S330. Collect vehicle state data.

S340. Extract the features of the temperature data by using the prediction model and according to the time series information using a set sliding window.

Since a single piece of data is prone to fluctuation and has low stability, in this embodiment, a sliding window is used to extract features in feature engineering. Exemplarily, the size of the sliding window spans M pieces of data, the time span is T, the frequency of data collection is different, the values of M and T can be set flexibly, and usually T is about 5 minutes. Exemplarily, the sliding window starts on the 0th to (M-1)th data, and the number of sliding steps is 1, then the 1st to the Mth data, the 2nd to the (M+1) ) pieces of data to extract features of temperature data, and so on, continuously extract features of temperature data.

Selecting M pieces of data for smoothing has higher reliability than extracting time-series features from a single piece of data, and can avoid accidental errors; and, due to the small amount of continuous negative sample data on vehicles with temperature differences, About 50 consecutive negative sample data can reflect the obvious temperature difference jump, so it is necessary to select as many negative sample data as possible in the sliding window to ensure the stability of the sample data. If the sliding window is too large, some negative samples will not be extracted, resulting in the loss of valuable features; if the sliding window is too small, the timing information will be incomplete.

S350. Classify the features of the temperature data based on the AdaBoost algorithm to predict the temperature consistency among multiple groups of modules.

The features of the extracted temperature data are classified, that is, whether it is a positive sample or a negative sample, so as to predict the temperature consistency. Considering the situation of temperature jump, the characteristic variable has nonlinear factors, this embodiment selects the AdaBoost algorithm of nonlinear classification as the classification model, so as to improve the accuracy of prediction. The AdaBoost algorithm can serially learn a series of weak learners from the training data, and linearly combine these weak learners into a strong learner.

FIG. 5 is a schematic implementation diagram of a temperature consistency prediction process provided in Embodiment 2 of the present application. As shown in Figure 5, online, the collected vehicle status data is input to the prediction model after edge computing and data preprocessing, and the prediction model can automatically output prediction results; offline, the downloaded historical data is used as sample data to train prediction Model, it is not necessary to do any processing if the temperature of multiple groups of modules is consistent, and if the temperature of multiple groups of modules is inconsistent, it is necessary to further extract these data for preprocessing, and input them into the classification model. The temperature and time series characteristics fully learn the prediction rules, continuously optimize the model parameters, and obtain the constructed prediction model. The constructed prediction model can also be tested with a part of the test data. If the prediction result is inaccurate, the problem can be backtracked and the model can be adjusted. The trained prediction model can be directly put into the application of online real-time prediction. It should be noted that after the data collected in real time is uploaded and saved to the cloud, it also becomes historical data, which can be used to train or update the prediction model. The above process completes the offline construction of the temperature inconsistency model, and predicts the real-time vehicle state data online, and the prediction result can be stored or called in the form of an interface. The temperature consistency prediction can be realized through the data drive of the Internet of Vehicles, and the monitoring and management of the vehicle by the server such as product R&D, quality assurance and after-sales can be driven by the real-time data stream interface call. The client is more transparent to the vehicle status and improves the user experience.

A temperature consistency prediction method provided in the third embodiment of the present application is described on the basis of the above-mentioned embodiment. By testing the prediction model and verifying the actual vehicle, the reliability of the prediction model is guaranteed; by adopting a set sliding window Extract the features of temperature data to improve the comprehensiveness and stability of the features; use the AdaBoost algorithm of nonlinear classification as a classification model to improve the accuracy of prediction, facilitate the timely detection of temperature imbalances, and then improve the safety of power batteries. Fault isolation.

Embodiment 4

FIG. 6 is a schematic structural diagram of a temperature consistency prediction apparatus provided in Embodiment 4 of the present application. The temperature consistency prediction device provided in this embodiment includes: a collection module 410, configured to collect vehicle state data, where the vehicle state data includes passing through the communication between a plurality of single cells in each group of modules in a plurality of groups of modules. The temperature data and time series information collected by the temperature sensor are shared; the prediction module 420 is configured to predict the temperature consistency among the multiple groups of modules according to the vehicle state data through a prediction model.

A temperature consistency prediction device provided in the fourth embodiment of the present application predicts the temperature consistency between modules from the time dimension and the dimension of different batteries, so as to facilitate the timely detection of temperature imbalance and improve the safety of power batteries. To a certain extent, fault isolation is achieved.

On the basis of the above-mentioned embodiment, it further includes: a historical data acquisition module, configured to acquire historical temperature data and historical time series information; and a construction module, configured to construct a prediction model according to the historical temperature data and historical time series information.

The prediction module 420 includes: a calculation unit, configured to calculate the difference between the highest temperature and the lowest temperature of multiple groups of modules at each moment in turn through the prediction model and according to the time sequence information; a counting unit, configured for each moment , if the difference is greater than or equal to the first threshold, the count value of the counter is incremented by 1 on the basis of the count value at the previous moment at each time, and if the difference is less than the first threshold, the count of the counter is The count value is cleared to zero; the determination unit is set to determine that the temperature among the multiple groups of modules does not conform to the consistency if the count value of the counter reaches the second threshold.

The prediction module 420 includes: a feature extraction unit, configured to extract the features of the temperature data through the prediction model and according to the time series information using a set sliding window, and the features of the temperature data include multiple groups of modules in the Temperatures at different times; the prediction unit is configured to predict the temperature consistency among multiple groups of modules according to the characteristics of the temperature data.

The prediction unit is set to: classify the features of the temperature data based on the AdaBoost algorithm, so as to predict the temperature consistency among the multiple groups of modules.

The vehicle status data also includes at least one of the following: vehicle code, cell temperature, cell voltage, cell current, charging and discharging status, charging and discharging current, charging and discharging voltage, vehicle working status, cell voltage extreme value, Body current extreme value, cell temperature extreme value, battery management system alarm information.

The device further includes: a preprocessing module configured to preprocess the vehicle state data; the preprocessing includes at least one of the following: data deduplication; unavailable value processing; abnormal data processing; alarm correction; current correction; Voltage correction; temperature correction.

The temperature consistency prediction apparatus provided in the fourth embodiment of the present application can be used to execute the temperature consistency prediction method provided in any of the above embodiments, and has corresponding functions.

Embodiment 4

FIG. 7 is a schematic diagram of a hardware structure of a prediction device according to Embodiment 5 of the present application. Predictive devices include, but are not limited to, electronic devices such as desktop computers, trip computers, smart phones, Internet of Vehicles servers, and cloud servers. As shown in FIG. 7 , the prediction device provided by the present application includes a memory 42, a processor 41, and a computer program stored in the memory and executable on the processor, and the processor 41 implements the above-mentioned temperature consistency when executing the program. method of prediction.

The prediction device may also include a memory 42; the number of processors 41 in the prediction device may be one or more, one processor 41 is taken as an example in FIG. 7 ; the memory 42 is configured to store one or more programs; the one or more Each program is executed by the one or more processors 41 , so that the one or more processors 41 implement the temperature consistency prediction method as described in the embodiments of the present application.

The prediction apparatus further includes: a communication device 43 , an input device 44 and an output device 45 .

The processor 41 , the memory 42 , the communication device 43 , the input device 44 and the output device 45 in the prediction device may be connected by a bus or in other ways, and the connection by a bus is taken as an example in FIG. 7 .

The input device 44 may be configured to receive input numerical or character information, and to generate key signal input related to user settings and function control of the predictive device. The output device 45 may include a display device such as a display screen.

The communication device 43 may include a receiver and a transmitter. The communication device 43 is configured to transmit and receive information according to the control of the processor 41 .

As a computer-readable storage medium, the memory 42 can be configured to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the temperature consistency prediction method described in the embodiments of the present application (for example, temperature consistency prediction method). acquisition module 410 and prediction module 320 in the device). The memory 42 may include a stored program area and a stored data area, wherein the stored program area may store an operating system, an application program required for at least one function; the stored data area may store data created according to the usage of the predicted device, and the like. Additionally, memory 42 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some instances, memory 42 may also include memory located remotely from processor 41, which may be connected to the prediction device through a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

On the basis of the above-mentioned embodiment, this embodiment also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by the temperature consistency prediction apparatus, the temperature consistency in any of the above-mentioned embodiments of the present application is realized. A prediction method, the method comprising: collecting vehicle state data, the vehicle state data including temperature data and time sequence information collected through a shared temperature sensor between a plurality of single cells in each group of modules in a plurality of groups of modules ; predicting the temperature consistency among the multiple groups of modules according to the vehicle state data through a prediction model.

A storage medium containing computer-executable instructions provided by the embodiments of the present application may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or any combination of the above. Examples (non-exhaustive list) of computer-readable storage media include: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (Read Only Memory) Memory, ROM), erasable programmable read only memory (Erasable Programmable Read Only Memory, EPROM), flash memory, optical fiber, portable compact disk read only memory (Compact Disc Read-Only Memory, CD-ROM), optical storage devices, Magnetic memory device, or any suitable combination of the above. A computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

A computer-readable signal medium may include a propagated data signal in baseband or as part of a carrier wave, with computer-readable program code embodied thereon. Such propagated data signals may take a variety of forms including, but not limited to, electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device .

The program code embodied on the computer readable medium can be transmitted by any suitable medium, including but not limited to: wireless, wire, optical fiber cable, radio frequency (RF), etc., or any suitable combination of the above.

Computer program code for carrying out the operations of the present application may be written in one or more programming languages, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional A procedural programming language, such as the "C" language or similar programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server. In situations involving a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or Wide Area Network (WAN), or may be connected to an external computer (eg, use an internet service provider to connect via the internet).

From the above description of the embodiments, those skilled in the art can understand that the present application can be implemented by means of software and general-purpose hardware. Based on this understanding, the technical solutions of the present application can be embodied in the form of software products, and the computer software products can be stored in a computer-readable storage medium, such as a floppy disk, ROM, RAM, flash memory (FLASH), hard disk or optical disk of a computer etc., including a plurality of instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the temperature consistency prediction method described in the various embodiments of the present application.

Claims

A temperature consistency prediction method, including:

collecting vehicle state data, where the vehicle state data includes temperature data and time sequence information collected through a shared temperature sensor between a plurality of single cells in each group of modules of the plurality of modules;

The temperature consistency among the multiple groups of modules is predicted according to the vehicle state data through a prediction model.
The method of claim 1, further comprising:

Obtain historical temperature data and historical time series information;

The prediction model is constructed according to the historical temperature data and the historical time series information.
The method according to claim 1, wherein predicting the temperature consistency among the plurality of groups of modules according to the vehicle state data by the prediction model comprises:

Calculate the difference between the highest temperature and the lowest temperature of the multiple groups of modules at each moment in turn according to the time sequence information through the prediction model;

In the case that the difference is greater than or equal to the first threshold, the count value of the counter is incremented by 1 on the basis of the count value at the previous moment at each time; when the difference is less than the first threshold In this case, the count value of the counter is cleared;

When the count value of the counter reaches the second threshold, it is determined that the temperatures among the plurality of groups of modules do not conform to the consistency.
The method according to claim 1, wherein predicting the temperature consistency among the plurality of groups of modules according to the vehicle state data by the prediction model comprises:

Through the prediction model, according to the time series information, a set sliding window is used to extract the characteristics of the temperature data, and the characteristics of the temperature data include the temperatures of the multiple groups of modules at different times;

The temperature consistency among the plurality of modules is predicted according to the characteristics of the temperature data.
The method according to claim 4, wherein predicting the temperature consistency among the plurality of groups of modules according to the characteristics of the temperature data comprises:

The features of the temperature data are classified based on the reinforcement learning AdaBoost algorithm to predict the temperature consistency among the multiple groups of modules.
The method according to claim 1, wherein the vehicle state data further comprises at least one of the following: vehicle code, cell temperature, cell voltage, cell current, charging and discharging status, charging and discharging current, charging and discharging voltage, Vehicle working status, cell voltage limit, cell current limit, cell temperature limit, battery management system alarm information.
The method according to claim 1, further comprising: preprocessing the vehicle state data; the preprocessing includes at least one of the following: data deduplication; unavailable value processing; abnormal data processing; alarm correction; current correction ; Voltage correction; Temperature correction.
A temperature consistency prediction device, comprising:

a collection module, configured to collect vehicle state data, where the vehicle state data includes temperature data and time sequence information collected through a shared temperature sensor between a plurality of single cells in each group of modules in the multiple groups of modules;

A prediction module, configured to predict the temperature consistency among the multiple groups of modules according to the vehicle state data through a prediction model.
A forecasting device comprising:

at least one processor;

a memory, arranged to store at least one program;

When the at least one program is executed by the at least one processor, the at least one processor implements the temperature consistency prediction method according to any one of claims 1-7.
A computer-readable storage medium on which a computer program is stored, wherein when the computer program is executed by a processor, the temperature consistency prediction method according to any one of claims 1-7 is implemented.