A power battery abnormality early warning method and early warning device

Abstract

The application provides a power battery abnormality early warning method and early warning device, and belongs to the technical field of power battery early warning. The power battery abnormality early warning method comprises the following steps: collecting real-time flow data and offline batch data of a power battery itself and related data; calculating an abnormal risk value of the power battery according to the real-time flow data and the offline batch data; deriving an early warning level according to the abnormal risk value; and performing corresponding early warning according to the early warning level. The power battery abnormality early warning method can more comprehensively and accurately evaluate the abnormality of the power battery.

Description

Technical Field

[0001] The present invention relates to the technical field of power battery early warning, and in particular to a power battery abnormality early warning method and early warning device. Background Art

[0002] With the increasing popularity of new energy electric vehicles, the importance of early warning of battery anomalies is gradually increasing. In this context, it is necessary to develop a battery anomaly warning system. By analyzing and mining battery-related data, it can warn of potential battery anomalies and promptly notify relevant personnel. This can provide early warning of new energy power battery anomalies, ensuring the safety of relevant personnel and brand trust.

[0003] The new energy power battery anomaly warning system collects a large amount of historical power battery data to evaluate abnormal power battery conditions. In most scenarios, battery status changes are linear. However, in some situations (such as high or low temperatures), battery status can undergo sudden changes. Analysis of offline power battery data alone cannot reflect these changes, and therefore cannot accurately and comprehensively assess power battery anomalies. Summary of the Invention

[0004] An object of the first aspect of the present invention is to provide a power battery abnormality early warning method that can more comprehensively and accurately assess the abnormal conditions of the power battery.

[0005] A further object of the present invention is to make the obtained short-term abnormal risk value more accurate.

[0006] Another object of the present invention is to improve the early warning experience.

[0007] An object of the second aspect of the present invention is to provide a power battery abnormality warning device for implementing the above-mentioned power battery abnormality warning method.

[0008] In particular, the present invention provides a power battery abnormality early warning method, comprising:

[0009] Collect the power battery itself and related real-time streaming data as well as offline batch data;

[0010] Calculating an abnormal risk value of the power battery according to the real-time stream data and the offline batch data respectively;

[0011] Determine a warning level based on the abnormal risk value;

[0012] A corresponding warning is issued according to the warning level.

[0013] Optionally, the step of respectively calculating the abnormal risk value of the power battery according to the real-time stream data and the offline batch data includes:

[0014] constructing feature values ​​according to the real-time streaming data and the offline batch data respectively;

[0015] The feature values ​​constructed using the offline batch data are used to train corresponding anomaly assessment models according to the types of parameters of the power battery, and the anomaly assessment models are used to output a model anomaly risk value;

[0016] Establishing an expert rule module corresponding to each of the anomaly assessment models, wherein the expert rule module is used to output a rule anomaly risk value according to preset rules;

[0017] Inputting the real-time stream data into the anomaly assessment model and the expert rule module respectively to obtain a first model anomaly risk value and a first rule anomaly risk value of the real-time stream data;

[0018] The offline batch data is input into the anomaly assessment model and the expert rule module respectively to obtain a second model anomaly risk value and a second rule anomaly risk value of the offline batch data.

[0019] Optionally, the step of deriving a warning level according to the abnormal risk value includes:

[0020] Assigning corresponding weighting coefficients to the first model abnormal risk value and the first rule abnormal risk value to obtain a real-time abnormal risk value;

[0021] Assigning corresponding weighting coefficients to the second model abnormal risk value and the second rule abnormal risk value to obtain an offline abnormal risk value;

[0022] Calculating a short-term abnormal risk value and a long-term abnormal risk value of the power battery according to the real-time abnormal risk value and the offline abnormal risk value;

[0023] A corresponding warning level is derived according to the short-term abnormal risk value and the long-term abnormal risk value.

[0024] Optionally, the step of deriving the short-term abnormal risk value and the long-term abnormal risk value of the power battery according to the real-time abnormal risk value and the offline abnormal risk value includes:

[0025] The offline abnormal risk value is used as a reference for the real-time abnormal risk value to obtain the short-term abnormal risk value.

[0026] Optionally, the step of issuing a corresponding warning according to the warning level includes:

[0027] When the warning level is greater than the safety level threshold, determining whether the vehicle is in a driving state;

[0028] If so, play the warning notification via the vehicle’s voice broadcast;

[0029] Otherwise, make an early warning notification by making a phone call;

[0030] When the warning level is less than or equal to the safety level threshold, a warning is issued through in-station information.

[0031] Optionally, before the step of issuing a corresponding warning according to the warning level, the step includes:

[0032] Record the interval time of the warning levels of the same level;

[0033] The warning levels whose interval time is less than the time threshold are merged.

[0034] Optionally, sources of the real-time streaming data and the offline batch data include power batteries of vehicles and / or charging stations.

[0035] Optionally, after the step of collecting the power battery itself and related real-time streaming data and offline batch data, the following step is further included:

[0036] The real-time stream data and the offline batch data are cached in each message partition using a distributed message queue service, and each message partition is used to store data of a power battery.

[0037] Optionally, before the step of respectively calculating the abnormal risk value of the power battery according to the real-time stream data and the offline batch data, the step further includes:

[0038] Determining whether the real-time stream data and the offline batch data carry an alarm flag;

[0039] If so, issue an early warning notification directly.

[0040] In particular, the present invention also provides a power battery abnormality warning device, including a controller, the controller including a memory and a processor, the memory storing a control program, and the control program, when executed by the processor, is used to implement any of the power battery abnormality warning methods described above.

[0041] According to one embodiment of the present invention, the abnormal risk situation of the power battery is evaluated respectively through the real-time streaming data and offline batch data of the power battery, that is, the abnormal risk value of the power battery is calculated respectively, and then the abnormal risk value obtained according to the real-time streaming data and the abnormal risk value obtained according to the offline batch data are combined to determine the warning level, and then a warning is issued according to the warning level. This warning method that combines the real-time streaming data and offline batch data of the power battery can not only issue risk warnings for short-term sudden changes in battery status, but also mine abnormal risks from long-term continuous changing data and issue warnings, and can more comprehensively and accurately evaluate the abnormal situation of the power battery.

[0042] According to one embodiment of the present invention, an abnormality assessment model and an expert rule module are established. By inputting the characteristic values ​​constructed by real-time streaming data and offline batch data into the above two modules respectively, corresponding abnormal risk values ​​can be obtained respectively, so as to use multiple evaluation rules to evaluate the abnormal conditions of the power battery reflected by the real-time streaming data and offline batch data respectively, so as to obtain more abnormal risk values ​​for the subsequent determination of the warning level, making the warning more accurate and comprehensive.

[0043] According to one embodiment of the present invention, the offline abnormal risk value is used as a reference for the real-time abnormal risk value to derive a short-term abnormal risk value, so that the obtained short-term abnormal risk value has higher accuracy.

[0044] According to one embodiment of the present invention, after real-time streaming data and offline batch data are collected, it is checked whether they carry an alarm flag. If they carry an alarm flag, an early warning notification is immediately issued, thereby improving the timeliness of the early warning.

[0045] Furthermore, merging warning levels with smaller intervals can control fatigue and avoid playing the same warning content too frequently, thereby improving the warning experience.

[0046] Furthermore, when the warning level is high, the user can be notified directly by phone or by in-vehicle voice broadcast according to the driving status, which can effectively notify the driver while ensuring driving safety.

[0047] Based on the following detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings, those skilled in the art will become more aware of the above and other objects, advantages and features of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Hereinafter, some specific embodiments of the present invention will be described in detail in an exemplary and non-limiting manner with reference to the accompanying drawings. The same reference numerals in the accompanying drawings indicate the same or similar components or parts. It should be understood by those skilled in the art that these drawings are not necessarily drawn to scale. In the accompanying drawings:

[0049] Figure 1This is a flowchart of a power battery abnormality warning method according to one embodiment of the present invention;

[0050] Figure 2 This is a flowchart of the steps of calculating an abnormality risk value in a power battery abnormality warning method according to one embodiment of the present invention;

[0051] Figure 3 is a flowchart of a power battery abnormality warning method according to another embodiment of the present invention;

[0052] Figure 4 yes Figure 3 Schematic diagram of the power battery abnormality warning method in.

[0053] Reference numerals:

[0054] 10-Unified data acquisition system, 20-Distributed message queuing service, 30-Earning system, 31-Stream and batch integrated processing module, 311-Data processing module, 312-Feature construction module, 32-Offline batch processing module, 321-Model training module, 322-Anomaly assessment model, 323-Expert rule module, 324-First comprehensive decision module, 33-Online stream processing module, 331-Second comprehensive decision module, 34-Fusion module. DETAILED DESCRIPTION

[0055] In the description of the present embodiment, reference to the terms "one embodiment," "some embodiments," "illustrative embodiments," "example," "specific example," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the exemplary expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples.

[0056] Unless otherwise defined, all terms (including technical terms and scientific terms) used in the description of this embodiment have the same meaning as commonly understood by ordinary technicians in the technical field to which this application belongs.

[0057] Figure 1 FIG. 1 is a flow chart of a power battery abnormality warning method according to an embodiment of the present invention. Figure 1 As shown, in one embodiment, the power battery abnormality warning method of the present invention includes:

[0058] Step S100, collecting the power battery itself and related real-time streaming data and offline batch data;

[0059] Step S200, calculating the abnormal risk value of the power battery based on the real-time stream data and the offline batch data;

[0060] Step S300, determining a warning level based on the abnormal risk value;

[0061] Step S400: Issue a corresponding warning according to the warning level.

[0062] In step S100, the real-time streaming data and offline batch data of the power battery itself may include data such as battery voltage, battery temperature, and battery capacity. The real-time streaming data and offline batch data related to the power battery may include data such as ambient temperature, ambient air pressure, and ambient humidity. Real-time streaming data is characterized by high timeliness and can reflect short-term changes in the power battery. It is data within a smaller time window, such as a few hours or minutes. Offline batch data is a large amount of static data with significant latency, corresponding to a larger time window, such as several days.

[0063] In step S200 , some algorithms, models, formulas, etc. may be used to calculate the abnormal risk value.

[0064] This embodiment evaluates the abnormal risk of the power battery through the real-time streaming data and offline batch data of the power battery, that is, calculates the abnormal risk value of the power battery respectively, and then determines the warning level by combining the abnormal risk value obtained from the real-time streaming data and the abnormal risk value obtained from the offline batch data. Then, a warning is issued according to the warning level. This warning method that combines the real-time streaming data and offline batch data of the power battery can not only issue risk warnings for short-term sudden changes in battery status, but also mine abnormal risks from long-term continuous changing data and issue warnings, which can more comprehensively and accurately evaluate the abnormal conditions of the power battery.

[0065] Figure 2 FIG. 1 is a flow chart of the steps of calculating abnormal risk value in a power battery abnormality warning method according to an embodiment of the present invention. Figure 2 As shown, in one embodiment, step S200 includes:

[0066] Step S210, constructing feature values ​​based on real-time streaming data and offline batch data respectively;

[0067] Step S220: using the feature values ​​constructed from the offline batch data to train corresponding anomaly assessment models according to the types of power battery parameters, the anomaly assessment models are used to output a model anomaly risk value;

[0068] Step S230: establishing an expert rule module corresponding to each anomaly assessment model, wherein the expert rule module is used to output a rule anomaly risk value according to a preset rule;

[0069] Step S240: input the real-time stream data into the anomaly assessment model and the expert rule module respectively to obtain a first model anomaly risk value and a first rule anomaly risk value of the real-time stream data;

[0070] Step S250 : inputting the offline batch data into the anomaly assessment model and the expert rule module respectively to obtain the second model anomaly risk value and the second rule anomaly risk value of the offline batch data.

[0071] The feature values ​​in step S210 include time series features, statistical features, and short-term stable difference features. Time series features are constructed based on data collected from power batteries with the same battery code (sorted by data collection time to form a data queue), including features such as battery voltage difference and battery temperature difference. Statistical features are constructed by counting the number of uploaded data for power batteries with the same battery code to determine their service lifecycle (i.e., effective operating time, as the state of batteries with longer service will change differently than those with less service), the time range from when the battery was shipped to the current state, and the number of charge cycles. Construction of short-term stable difference features: Taking battery capacity (which can also include battery voltage and battery temperature) as an example, in most cases, the battery capacity decay difference is relatively stable over a short, continuous period. Therefore, the online real-time battery capacity is subtracted from the average battery capacity over a long historical period to form a short-term stable difference feature for battery capacity. The average battery capacity over a long historical period is generated by batch processing.

[0072] The abnormality assessment model in step S220 can be obtained by a machine through model training using a large amount of offline batch data, and can be some battery abnormality assessment models commonly used in the prior art, and is not limited here.

[0073] The expert rule module in step S230 is a module that sets some clear rules, such as some formulas or judgment rules.

[0074] In this embodiment, an abnormality assessment model and an expert rule module are established. By inputting the characteristic values ​​constructed by real-time streaming data and offline batch data into the above two modules respectively, corresponding abnormal risk values ​​can be obtained respectively, so that the abnormal conditions of the power battery reflected by the real-time streaming data and offline batch data can be evaluated respectively using multiple evaluation rules to obtain more abnormal risk values ​​for the subsequent determination of the warning level, making the warning more accurate and comprehensive.

[0075] In a further embodiment, step S300 includes:

[0076] Assign corresponding weighting coefficients to the abnormal risk value of the first model and the abnormal risk value of the first rule to obtain a real-time abnormal risk value;

[0077] Assign corresponding weighting coefficients to the abnormal risk value of the second model and the abnormal risk value of the second rule to obtain an offline abnormal risk value;

[0078] Calculate the short-term and long-term abnormal risk values ​​of the power battery based on the real-time abnormal risk value and the offline abnormal risk value;

[0079] The corresponding warning level is derived based on the short-term abnormal risk value and the long-term abnormal risk value.

[0080] For example, the real-time anomaly risk value = A1 * first model anomaly risk value + B1 * first rule anomaly risk value; the offline anomaly risk value = A2 * first model anomaly risk value + B2 * first rule anomaly risk value. A1 + B1 = 1, A2 + B2 = 1, and A1, B1, A2, and B2 are weighting coefficients. The values ​​of A1, B1, A2, and B2 can be set based on the specific situation, for example, all set to 0.5. All of these risk values ​​can be expressed as percentages.

[0081] In one embodiment, the step of deriving the short-term abnormal risk value and the long-term abnormal risk value of the power battery according to the real-time abnormal risk value and the offline abnormal risk value includes:

[0082] The offline abnormal risk value is used as a reference for the real-time abnormal risk value to obtain the short-term abnormal risk value.

[0083] Assuming the real-time abnormal risk value of the power battery capacity is between a first threshold and a second threshold (the second threshold is greater than the first threshold), while the offline abnormal risk value is greater than the second threshold, the short-term abnormal risk value can be determined to be greater than the second threshold, corresponding to a more severe warning level. Alternatively, a weighted calculation can be performed on the real-time abnormal risk value and the offline abnormal risk value, with the offline abnormal risk value weighted with a smaller coefficient.

[0084] The long-term abnormal risk value can directly correspond to the offline abnormal risk value, or it can be corrected with reference to the real-time abnormal risk value. The specific method is similar to that of the short-term abnormal risk value and is not limited here.

[0085] In this embodiment, a real-time abnormal risk value is first obtained by weighting the two abnormal risk values ​​obtained by the real-time stream data input abnormality assessment model and the expert rule module, and an offline abnormal risk value is obtained by weighting the two abnormal risk values ​​obtained by the offline batch data input abnormality assessment model and the expert rule module. Then, the short-term abnormal risk value and the long-term abnormal risk value of the power battery are derived based on the real-time abnormal risk value and the offline abnormal risk value. Finally, the corresponding warning level is derived based on the short-term abnormal risk value and the long-term abnormal risk value. The resulting warning level can reflect both the long-term risk and the short-term mutation risk of the power battery, so that the warning is timely and comprehensive.

[0086] Furthermore, the offline abnormal risk value is used as a reference for the real-time abnormal risk value to derive a short-term abnormal risk value, so that the obtained short-term abnormal risk value has higher accuracy.

[0087] In one embodiment, sources of real-time streaming data and offline batch data include power batteries of vehicles and / or charging stations.

[0088] The abnormality warning method of this embodiment can not only warn the power battery of the vehicle, but also access the data of the charging station to warn the power battery of the charging station.

[0089] Figure 3 is a flowchart of a power battery abnormality warning method according to another embodiment of the present invention. Figure 4 yes Figure 3 Schematic diagram of the power battery abnormality warning method in another embodiment. Figure 3 As shown, after step S100, the following steps are included:

[0090] Step S150: Utilize the distributed message queue service to cache the real-time stream data and offline batch data in each message partition, where each message partition is used to store data of one power battery;

[0091] Step S160, determining whether the real-time stream data and offline batch data carry an alarm flag, if so, proceeding to step S170, otherwise proceeding to step S180;

[0092] Step S170, directly issuing an early warning notification;

[0093] Step S180: performing outlier processing on the real-time streaming data and offline batch data.

[0094] The steps before step S400 include:

[0095] Step S350, recording the interval time of warning levels of the same level;

[0096] Step S360: Merge the warning levels whose interval time is less than the time threshold.

[0097] Step S400 includes:

[0098] Step S410, determine whether the warning level is greater than the safety level threshold, if so, proceed to step S420, otherwise proceed to step S450;

[0099] Step S420: When the warning level is greater than the safety level threshold, determine whether the vehicle is in a driving state. If so, proceed to step S430; otherwise, proceed to step S440.

[0100] Step S430, playing the warning notification via in-vehicle voice playback;

[0101] Step S440, making a warning notification by making a phone call;

[0102] Step S450: When the warning level is less than or equal to the safety level threshold, a warning is issued through in-station information.

[0103] The distributed message queue service in step S150 is a common middleware used in large-scale distributed systems. Message queues primarily address issues such as application coupling, asynchronous messaging, and traffic clipping, offering high performance, high availability, scalability, and eventual consistency. Once data is sent to the distributed message queue service, subsequent services can process it based on their own processing capabilities. Furthermore, data with the same battery code will be placed in the same message partition, facilitating sequential processing of the same battery by subsequent services.

[0104] like Figure 4 As shown, step S100 can utilize the unified data acquisition system 10 to collect data, and connect to various external data sources in different docking methods (such as socket, http, message queue, etc.), including parsing different protocols (the parsed data will be unified into a structured format), exception checking, etc., and then partition the data according to the battery code in the data and send it to different message partitions of the distributed message queue service 20.

[0105] Step S180 can be performed by Figure 4The data processing module 311 of the integrated stream-batch processing module 31 performs outlier processing on the data, including null value filling and smoothing of outliers at a specific moment. The data processing module 311 can also determine whether the data carries an alarm flag (in the GBT 32960 protocol, data uploaded via the vehicle-mounted terminal will contain alarm data). The integrated stream-batch processing module 31 also includes a feature construction module 312 for constructing the aforementioned feature values. The integrated stream-batch processing module 31 has the ability to access and process real-time streaming data and offline data. The functions of the integrated stream-batch processing module 31 can be implemented based on the Apache Flink component (Apache Flink is an open source stream processing framework that can process finite data streams and infinite data, that is, it can process bounded and unbounded data streams. Unbounded data streams are truly streaming data, so Flink supports stream computing. Bounded data streams are batch data, so it also supports batch processing.). Real-time streaming data and offline batch data are processed using the same set of code to reduce maintenance costs. The aforementioned data processing module 311 and feature construction module 312 can be abstracted into the same set of code. The only difference between processing real-time streaming data and offline batch data is the size of the processing window. For example, the processing window for real-time streaming data is 1 hour, while the processing window for offline batch data is 3 days.

[0106] In step 450, the user may be notified via information within the APP / Web site.

[0107] In this embodiment, after real-time stream data and offline batch data are collected, it is checked whether they carry an alarm flag. If they carry an alarm flag, an early warning notification is immediately issued, thereby improving the timeliness of the early warning.

[0108] Furthermore, merging warning levels with smaller intervals can control fatigue and avoid playing the same warning content too frequently, thereby improving the warning experience.

[0109] Furthermore, when the warning level is high, the user can be notified directly by phone or by in-vehicle voice broadcast according to the driving status, which can effectively notify the driver while ensuring driving safety.

[0110] like Figure 4As shown, the downstream of the distributed message queue service 20 is an early warning system 30. This early warning system 30 includes the aforementioned integrated stream-batch processing module 31, an offline batch processing module 32 connected to the integrated stream-batch processing module 31, and an online stream processing module 33. The offline batch processing module 32 includes a model training module 321 to obtain an anomaly assessment model 322. Both the offline batch processing module 32 and the online stream processing module 33 include an anomaly assessment model 322 and an expert rule module 323. The offline batch processing module 32 includes a first comprehensive decision module 324, while the online stream processing module 33 includes a second comprehensive decision module 331. The second comprehensive decision module 331 is used to assign corresponding weighting coefficients to the first model anomaly risk value and the first rule anomaly risk value to obtain a real-time anomaly risk value. The first comprehensive decision module 324 is used to assign corresponding weighting coefficients to the second model anomaly risk value and the second rule anomaly risk value to obtain an offline anomaly risk value. Finally, the fusion module 34 uses the real-time anomaly risk value and the offline anomaly risk value to determine the short-term and long-term anomaly risk values ​​of the power battery, and outputs an early warning level and an early warning notification.

[0111] The present invention also provides a power battery abnormality warning device, including a controller, the controller including a memory and a processor, the memory storing a control program, and the control program, when executed by the processor, is used to implement the power battery abnormality warning method in any of the above embodiments.

[0112] The processor can be a central processing unit (CPU) or a digital processing unit, etc. The processor sends and receives data through a communication interface. The memory is used to store the program executed by the processor. The memory is any medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, or it can be a combination of multiple memories. The above-mentioned computer program can be downloaded from a computer-readable storage medium to a corresponding computing / processing device or downloaded to a computer or external storage device via a network (such as the Internet, a local area network, a wide area network and / or a wireless network).

[0113] At this point, those skilled in the art will recognize that, although a number of exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications consistent with the principles of the present invention may be directly determined or derived from the disclosure of the present invention without departing from the spirit and scope of the present invention. Therefore, the scope of the present invention should be understood and deemed to cover all such other variations or modifications.

Claims

1. A power battery abnormality early warning method, characterized in that: include: Collect the power battery itself and related real-time streaming data as well as offline batch data; Calculating an abnormal risk value of the power battery according to the real-time stream data and the offline batch data respectively; Determine a warning level based on the abnormal risk value; Issue corresponding warnings according to the warning levels; The step of respectively calculating the abnormal risk value of the power battery according to the real-time stream data and the offline batch data includes: constructing feature values ​​according to the real-time streaming data and the offline batch data respectively; The feature values ​​constructed using the offline batch data are used to train corresponding anomaly assessment models according to the types of parameters of the power battery, and the anomaly assessment models are used to output a model anomaly risk value; Establishing an expert rule module corresponding to each of the anomaly assessment models, wherein the expert rule module is used to output a rule anomaly risk value according to preset rules; Inputting the real-time stream data into the anomaly assessment model and the expert rule module respectively to obtain a first model anomaly risk value and a first rule anomaly risk value of the real-time stream data; The offline batch data is input into the anomaly assessment model and the expert rule module respectively to obtain a second model anomaly risk value and a second rule anomaly risk value of the offline batch data.

2. The power battery abnormality warning method according to claim 1, characterized in that: The steps of deriving a warning level according to the abnormal risk value include: Assigning corresponding weighting coefficients to the first model abnormal risk value and the first rule abnormal risk value to obtain a real-time abnormal risk value; Assigning corresponding weighting coefficients to the second model abnormal risk value and the second rule abnormal risk value to obtain an offline abnormal risk value; Calculating a short-term abnormal risk value and a long-term abnormal risk value of the power battery according to the real-time abnormal risk value and the offline abnormal risk value; A corresponding warning level is derived according to the short-term abnormal risk value and the long-term abnormal risk value.

3. The power battery abnormality warning method according to claim 2, characterized in that: The step of obtaining the short-term abnormal risk value and the long-term abnormal risk value of the power battery according to the real-time abnormal risk value and the offline abnormal risk value includes: The offline abnormal risk value is used as a reference for the real-time abnormal risk value to obtain the short-term abnormal risk value.

4. The power battery abnormality warning method according to claim 1, characterized in that: The steps of issuing corresponding warnings according to the warning level include: When the warning level is greater than the safety level threshold, determining whether the vehicle is in a driving state; If so, play the warning notification via the vehicle’s voice broadcast; Otherwise, make an early warning notification by making a phone call; When the warning level is less than or equal to the safety level threshold, a warning is issued through in-station information.

5. The power battery abnormality warning method according to claim 1, characterized in that: The step of issuing a corresponding warning according to the warning level includes: Record the interval time of the warning levels of the same level; The warning levels whose interval time is less than the time threshold are merged.

6. The power battery abnormality warning method according to claim 1, characterized in that: Sources of the real-time streaming data and the offline batch data include power batteries of vehicles and / or charging stations.

7. The power battery abnormality warning method according to claim 6, characterized in that: The steps of collecting the power battery itself and related real-time streaming data and offline batch data also include: The real-time stream data and the offline batch data are cached in each message partition using a distributed message queue service, and each message partition is used to store data of a power battery.

8. The power battery abnormality warning method according to any one of claims 1 to 7, characterized in that: Before the step of respectively calculating the abnormal risk value of the power battery according to the real-time stream data and the offline batch data, the method further includes: Determining whether the real-time stream data and the offline batch data carry an alarm flag; If so, issue an early warning notification directly.

9. A power battery abnormality warning device, characterized in that: The controller includes a memory and a processor. The memory stores a control program. When the control program is executed by the processor, it is used to implement the power battery abnormality warning method according to any one of claims 1 to 8.

Images