Battery insurance fatigue failure prediction method and device, vehicle, medium and product

Abstract

The embodiment of the invention provides a battery insurance fatigue failure prediction method and device, a vehicle, a medium and a product. According to the method, based on operation data, the total future mileage of the vehicle in future time and current stress distribution characteristics corresponding to the total future mileage are determined. The probability of insurance fatigue failure of the battery in a future time period is estimated according to the data, and the battery insurance is maintained or warned in advance, so that the probability of occurrence of adverse events such as user complaints and traffic accidents is effectively reduced.

Description

Technical Field

[0001] The present application relates to the field of new energy vehicle technology, and in particular to a method, device, vehicle, medium and product for predicting fatigue failure of battery insurance. Background Art

[0002] With the rapid development of battery-powered new energy vehicles, battery system safety is paramount. Batteries may experience a variety of abnormal conditions during operation, one of the most dangerous being an external short circuit. When an external short circuit occurs, a large current is generated in a short period of time, potentially causing thermal runaway, fire, or even explosion, resulting in significant loss of life and property.

[0003] Currently, passive fuses are widely used in battery systems as safety protection components. When an external short circuit in a battery generates excessive current, the fuse in the passive fuse quickly melts due to the thermal effect of the current, disconnecting the high-voltage circuit and placing the battery system in a safe state, preventing any potential hazards.

[0004] When the battery is installed in a new energy vehicle, the risk of passive insurance failure will gradually accumulate as the vehicle runs, and the failure process is hidden. There is currently a lack of effective failure probability prediction methods. Summary of the Invention

[0005] The battery insurance fatigue failure prediction method, equipment, vehicle, medium and product provided in the embodiments of the present application are used to predict the probability of battery insurance fatigue failure and provide data support for battery insurance failure analysis.

[0006] In a first aspect, an embodiment of the present application provides a method for predicting fatigue failure of a battery fuse, comprising:

[0007] Obtaining, based on the vehicle operation data, a future total mileage of a battery of the vehicle at a future time, and a current stress distribution characteristic corresponding to the future total mileage;

[0008] Based on the future total mileage and the current stress distribution characteristics, the probability of insurance fatigue failure of the battery at the future time is predicted.

[0009] In a possible implementation, obtaining the current stress distribution characteristics corresponding to the future total mileage of the vehicle based on the vehicle operation data includes:

[0010] Based on the operating data of the vehicle, obtaining the interval of the current stress load intensity corresponding to the mileage of the vehicle; different intervals correspond to different operating conditions of the vehicle;

[0011] The current stress distribution characteristics are determined based on the current stress load intensity of each of the sections corresponding to the traveled mileage and the future total mileage.

[0012] In a possible implementation, determining the current stress distribution characteristics based on the current stress load intensity of each of the intervals corresponding to the traveled mileage and the future total mileage includes:

[0013] Obtaining the number of current stress load intensities in each of the intervals corresponding to the traveled mileage that exceeds a current stress load intensity threshold corresponding to the interval;

[0014] Obtaining the number of each of the intervals corresponding to the future total mileage based on the number of each of the intervals corresponding to the traveled mileage and the mileage difference between the traveled mileage and the future total mileage;

[0015] The current stress distribution characteristic is determined based on the number of each of the intervals corresponding to the future total mileage.

[0016] In a possible implementation, obtaining the future total mileage of the vehicle at a future time based on the vehicle operation data includes:

[0017] Determining an average mileage of the vehicle per unit time based on the total mileage traveled by the vehicle;

[0018] Obtaining an estimated mileage from the current time to the future time based on the future time, the current time, and the average mileage;

[0019] The future total mileage is determined based on the total mileage and the estimated mileage.

[0020] In one possible implementation, the predicting the probability of the battery's fatigue failure at the future time based on the future total mileage and the current stress distribution characteristics includes:

[0021] Based on the future total mileage, the current stress distribution characteristics, and the insurance fatigue failure prediction model, the insurance fatigue failure probability of the battery at the target time is predicted. The insurance fatigue failure prediction model is obtained by pre-training a neural network model using sample data, and the sample data includes: the total mileage, current stress distribution characteristics, and insurance fatigue failure probability of at least one sample vehicle.

[0022] In a possible implementation, after predicting the probability of the battery's fatigue failure at the future time, the method further includes:

[0023] When the probability of insurance fatigue failure of the battery at the future time is greater than a preset probability threshold, outputting warning information of insurance fatigue failure of the vehicle's battery.

[0024] In one possible implementation, the method further includes:

[0025] receiving a battery insurance fatigue failure prediction request from the vehicle, the request including identification information of the vehicle;

[0026] Based on the identification information of the vehicle, operating data of the vehicle is obtained from a database.

[0027] In a second aspect, an embodiment of the present application provides a battery insurance fatigue failure prediction device, comprising:

[0028] an acquisition module, configured to acquire, based on the vehicle's operating data, a total future mileage of the vehicle's battery at a future time, and a current stress distribution characteristic corresponding to the total future mileage;

[0029] A prediction module is used to predict the probability of insurance fatigue failure of the battery at the future time based on the future total mileage and the current stress distribution characteristics.

[0030] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory, a processor;

[0031] The memory stores computer-executable instructions;

[0032] The processor executes the computer-executable instructions stored in the memory, so that the processor executes any possible implementation of the first aspect above.

[0033] In a fourth aspect, an embodiment of the present application provides a vehicle, characterized in that the vehicle includes the electronic device as described in the third aspect.

[0034] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement any possible implementation method of the first aspect above.

[0035] In a sixth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any possible implementation method of the first aspect above.

[0036] The battery insurance fatigue failure prediction method, equipment, vehicle, medium, and product provided in the embodiments of the present application are that the greater the intensity of the current stress load and the more times it lasts, the faster the fatigue damage accumulation rate of the passive insurance. The total mileage of the vehicle reflects the operating conditions of the vehicle. Different operating conditions correspond to different stress loads. Therefore, mileage can also indirectly reflect the stress load conditions experienced. Therefore, based on vehicle operation data, the total mileage after the end of the future time period and the current stress distribution characteristics corresponding to the total mileage, the probability of insurance fatigue failure in the battery in the future time period can be predicted, providing data support for battery insurance failure analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0038] Figure 1 A schematic diagram of the circuit structure of a passive insurance is provided;

[0039] Figure 2 Schematic diagram of the process of predicting battery insurance fatigue failure provided in this application Figure 1 ;

[0040] Figure 3 Schematic diagram of the process of predicting battery insurance fatigue failure provided in this application Figure 2 ;

[0041] Figure 4 A schematic diagram of a neural network model provided for this application;

[0042] Figure 5 A schematic diagram of the structure of a server provided for this application;

[0043] Figure 6 A schematic diagram of the structure of the battery insurance fatigue failure prediction device provided in this application;

[0044] Figure 7 This is a schematic diagram of the structure of the electronic device provided in this application.

[0045] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0046] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0047] First, let’s explain the terms involved in this application:

[0048] Current stress refers to the physical or chemical forces exerted on internal battery materials (electrodes, separators, electrolytes) and external circuit components (wires, connectors, passive fuses) during battery operation due to changes in current magnitude, frequency, and waveform. Excessive current stress can cause material aging, structural damage, or heat accumulation, impacting battery performance and safety.

[0049] With the development of new energy vehicles, the safety of battery systems, as the power source of these vehicles, has attracted considerable attention. During operation, battery systems may encounter a variety of abnormal conditions, of which external short circuits are particularly dangerous. When a short circuit occurs, an extremely large current is generated in a very short period of time, which can easily lead to thermal runaway, resulting in fire or even explosion, posing a serious threat to life and property.

[0050] Passive insurance (such as fuses) is a key component of battery system safety protection and has been widely used in battery systems. Taking the passive insurance as an example, Figure 1 A schematic diagram of the circuit structure of a passive insurance is provided. Figure 1 As shown in the figure, the current flow between the battery and the load forms a high-voltage circuit. A current sensor is used to measure the current value during battery operation. When the battery experiences an external short circuit, that is, a connection is formed between the positive and negative terminals of the battery due to some external reason (such as accidental contact between wires or contact with a metal object), this connection causes the current to increase dramatically, far exceeding the normal tolerance of the battery system. At this point, the fuse will quickly melt due to the thermal effect of the current, disconnecting the high-voltage circuit and placing the battery system in a safe state. This prevents dangerous situations such as thermal runaway and fire, effectively ensuring the safety of the battery system.

[0051] However, when batteries are installed in new energy vehicles, the passive fuses are subjected to long-term current stress during operation. Current stress refers to the physical force exerted on the passive fuse by the continuous flow of current. This force can cause structural changes within the passive fuse, such as dislocations and distortions. Long-term accumulation can lead to cumulative material damage. Sustained current stress can cause fatigue failure in the passive fuse, which can prematurely fracture the metal conductors within the passive fuse, severing the battery system's high-voltage circuit, leading to charging interruptions or an inability to power the new energy vehicle.

[0052] In response to the above problems, the battery insurance fatigue failure prediction method provided in this application determines that the insurance failure is related to the current stress load intensity it bears through the fatigue failure principle of the battery insurance. Therefore, the failure probability of the battery insurance can be accurately predicted through data related to the current stress load intensity, providing data support for the analysis of battery insurance fatigue failure.

[0053] The execution subject of this application may be an electronic device, which may be an electronic device on a vehicle, or a processing device in a battery management system of a battery, such as a battery management control module (BMC), or a service end of the battery, such as a server, a server cluster, a computing platform, a cloud computing platform, etc., or a processing unit of an electrical device (such as a vehicle) using the battery, and this application does not limit this.

[0054] The embodiments of this specification can be implemented through an application, website, or applet with battery insurance prediction capabilities, or through the provision of a separate prediction device, or through the management platform of a vehicle maintenance center, without limitation. For example, a website that deploys an insurance fatigue failure prediction model can implement fatigue failure prediction for battery insurance.

[0055] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0056] Figure 2 Schematic diagram of the process of predicting battery insurance fatigue failure provided in this application Figure 1 ,like Figure 2 As shown, the method includes:

[0057] S201. Based on the vehicle's operating data, obtain the total mileage of the vehicle's battery in the future and current stress distribution characteristics corresponding to the total mileage in the future.

[0058] The current stress distribution characteristic refers to the distribution pattern of the current stress in the battery under different operating conditions. This current stress distribution characteristic can reflect the stress characteristics of the battery under different current stress load intensity ranges. The current stress load intensity is the degree of stress exerted on the materials and structures in the battery by the current load borne by the battery during vehicle operation. The division of the current stress load intensity range can be based on the power consumption of the vehicle during use, for example, the range can be divided based on the vehicle's operating conditions.

[0059] Taking operating conditions as an example, vehicle operating data may include data that can identify the vehicle's operating condition and the corresponding current stress load intensity. Data that can identify the vehicle's operating condition may include vehicle speed, acceleration, displacement, and engine speed.

[0060] The data related to the battery charging and discharging data in the above-mentioned vehicle operating data may be collected by the battery management system of the battery system, for example. The data related to the vehicle itself in the above-mentioned operating data may be collected by sensors installed on the vehicle, etc., and there is no limitation on this.

[0061] The above-mentioned operating data may also include the mileage that the vehicle has traveled. Accordingly, the total mileage of the vehicle in the future can be estimated by analyzing the vehicle's average daily mileage, weekly driving cycle patterns, or monthly / annual driving patterns through historical operating data, capturing user travel habits, and realizing the estimation of future mileage.

[0062] The current stress distribution characteristics corresponding to the future total mileage are estimated based on the current stress load intensity indicated in the operating data and the mileage inference.

[0063] It should be understood that the above-mentioned operating data may be data for a period of time before the current moment of the vehicle, or data from the first operation of the vehicle to the current moment.

[0064] In embodiments of the present application, the content of the present application can be executed upon user triggering. That is, when the user actively operates, the fatigue failure probability of the battery fuse in a future time period is predicted based on vehicle operating data. Alternatively, the prediction of the fatigue failure probability of the battery fuse can be automatically triggered and executed based on a pre-set time or period.

[0065] In an embodiment, based on the vehicle's operating data, the intervals of current stress load intensity corresponding to the vehicle's mileage can be obtained; then, based on the current stress load intensity of each interval corresponding to the mileage traveled and the total mileage in the future, the current stress distribution characteristics are determined.

[0066] Among them, different intervals correspond to different operating conditions of the vehicle.

[0067] In the analysis scenario of vehicle operation data, it is first necessary to classify the vehicle according to the set operating condition category based on the actual operating status of the vehicle. There are many types of vehicle operating conditions, such as starting, acceleration, constant speed, climbing, deceleration, braking, etc. Each condition corresponds to a specific vehicle operating state and electrical system working mode. That is, under different working conditions, the vehicle's operating state (such as speed, acceleration, driving direction, etc.) and the working mode of the electrical system (such as battery, motor, controller, etc.) will be different, and the current value output by the battery will also change accordingly. The mapping relationship between the vehicle operating condition and the interval has been pre-set. Therefore, according to the divided intervals, the current stress load intensity of each interval (that is, the current size and its comprehensive indicators such as action time and frequency) can be counted, and the distribution characteristics of the current stress can be analyzed accordingly.

[0068] For example, taking the acceleration interval as an example, a large amount of vehicle operation data is sorted and analyzed. By comprehensively considering multiple parameters such as the vehicle's speed change rate, acceleration value, throttle opening, etc., when these parameters all meet the judgment conditions of the acceleration condition, the corresponding time period can be marked as the acceleration interval.

[0069] For example, taking the climbing section as an example, the time period when the vehicle is in a climbing state can be identified based on the vehicle's slope sensor data, engine speed, vehicle load and other information, and the corresponding time period can be marked as a climbing section.

[0070] It can be understood that after determining multiple intervals and the time periods corresponding to the intervals, the current stress load intensity values ​​collected in these time periods can be further extracted. During the extraction, the current value of each frame can be extracted according to time. For example, one frame can be one second or one millisecond.

[0071] For example, during the acceleration range, the battery outputs a larger current because the motor needs to provide greater torque to drive the vehicle. This increases the current stress load intensity accordingly. By summing up all the current stress load intensity values ​​collected during the acceleration range, we form the current stress load intensity set corresponding to the acceleration range.

[0072] Based on the current stress load intensities in these intervals and the total future mileage, the current stress distribution characteristics can be further determined. This current stress distribution characteristic is a statistical description of the current stress load intensities, intuitively reflecting the changes and concentration trends of current stress load intensities under different operating conditions. The current stress load intensities for all intervals can be aggregated to obtain the current stress quantity for each interval. Alternatively, a series of data can be classified, counted, and analyzed, integrating and summarizing current stress data from different intensity ranges to obtain the current stress quantity.

[0073] Because there's a linear relationship between current stress and vehicle mileage, the current stress load increases with mileage. Therefore, the current stress load for the future total mileage can be predicted based on the current stress load. Based on the mileage data, these loads can reflect the changing characteristics of current stress corresponding to the user's driving behavior during vehicle operation. Based on this, the current stress load can be used to accurately predict the current stress distribution characteristics for the future total mileage of the vehicle.

[0074] For example, the amount of current stress per unit mileage can be determined based on the amount of current stress and the vehicle's mileage. The unit mileage can be, for example, 1000 kilometers. Based on the amount of current stress per unit mileage and the future total mileage, the current stress distribution characteristics for the future total mileage can be determined.

[0075] Another possible implementation method is that the vehicle's total mileage in the future can be determined by the total mileage traveled and the average mileage per unit time. That is, the average mileage can be determined by the total mileage and the time used for the total mileage, and the future total mileage can be obtained based on the average mileage and the future time that needs to be estimated.

[0076] You can also use a fixed multiple of the mileage from a known historical period to linearly extrapolate and add it to the current total mileage to predict mileage at a more distant point in the future. For example, if you know the mileage for the past 30 days, you can use twice the 30-day mileage to represent the mileage 60 days from now. In other words, the total mileage for the next 60 days is equal to the sum of the current total mileage and twice the 30-day mileage.

[0077] S202: Based on the current stress distribution characteristics and the total mileage in the future, predict the probability of insurance fatigue failure of the battery in the future.

[0078] Understandably, the current stress distribution reflects the intensity and unevenness of the electrochemical reaction within the battery, and also reflects the proportion of operating conditions used by the user in the past. The future total mileage reflects the intensity of battery usage. A longer future total mileage means the battery will experience more charge and discharge cycles, and will be subjected to longer periods of current stress, which increases the likelihood of battery fuse fatigue failure.

[0079] In one possible implementation, the present application can pre-establish a mapping relationship between current stress distribution characteristics, future total mileage, and insurance fatigue failure probability. This mapping relationship can then be used to determine the battery's insurance fatigue failure probability in the future. This mapping relationship can be constructed empirically or obtained through offline calibration based on relevant data from multiple sample vehicles, without limitation.

[0080] In another possible implementation, the present application may be pre-set with a fuse fatigue failure prediction model, so that the probability of fuse fatigue failure of the battery in the future can be predicted by the model.

[0081] The aforementioned insurance fatigue failure prediction model may be any model trained based on sample data, such as a neural network model. The sample data herein may include the total mileage, current stress distribution characteristics, and insurance fatigue failure probability of at least one sample vehicle.

[0082] As you can see, the fuse fatigue failure prediction model is the core of the entire prediction process. It is a complex algorithm trained using machine learning techniques, specifically neural networks, based on a large amount of sample data. This sample data covers the total mileage, current and stress distribution characteristics, and actual observed fuse fatigue failure states of different vehicles under different operating conditions. By learning the inherent patterns and regularities in this sample data, the neural network model is able to capture the correlation between current and stress distribution characteristics, total mileage, and fuse fatigue failure states.

[0083] In another possible implementation, the current stress distribution characteristics and the total future mileage can also be sent to other devices that can realize the prediction, such as the vehicle's service end, which can be a server located in the cloud, or a roadside device, or a user's terminal device connected to the vehicle. These devices can, for example, use the aforementioned method to predict the probability of battery insurance fatigue failure.

[0084] The battery insurance fatigue failure prediction method provided in the embodiments of the present application obtains the operating data of the vehicle to be predicted and specifies the future time period to be predicted. Based on the acquired operating data and the specified future time period, the method analyzes and determines the vehicle's current stress distribution characteristics and the total mileage data expected in the future. Subsequently, based on these characteristic data, the probability of the battery insurance fatigue failure within a specific future time period is estimated. This probability can be used to predict the possibility of battery insurance failure in advance, thereby effectively reducing the user complaint rate and the probability of potential risk events such as traffic accidents.

[0085] In some embodiments, after predicting the probability of battery failure due to fatigue in the future, the relationship between the probability of battery failure due to fatigue in the future and a preset probability may be determined, and different prompt information may be output.

[0086] As you can understand, the preset probability is determined based on a comprehensive consideration of historical failure data, usage conditions, environmental factors, and battery safety standards. It represents a risk threshold, used to measure whether the risk of battery fatigue failure has reached a level that requires action. For example, the preset probability could be 70%.

[0087] Comparing the predicted insurance fatigue failure probability with the preset probability essentially evaluates whether the battery's current and future safety performance meets expected requirements. The input of prompt information depends on the relationship between the insurance fatigue failure probability and the preset probability. The prompt information can be displayed on the vehicle's display panel, on the display interface of the entity executing this application, on the display panel of the prediction device, or on the device display screen of an application, website, or mini-program that performs the battery insurance prediction function. This application does not limit this.

[0088] For example, if the probability of the battery's insurance fatigue failure in the future is greater than a preset probability threshold, the vehicle's battery insurance fatigue failure warning information is output.

[0089] Understandably, if the predicted probability is higher than the preset probability, it indicates that the battery may face a high risk of fuse fatigue failure in the future, and its safety performance has exceeded the acceptable range, requiring timely action. In this case, a warning message can be output, indicating the current battery failure probability, and advising the user to replace the battery fuse in a timely manner. This prevents the battery from experiencing fuse fatigue failure during actual use.

[0090] Furthermore, the output information can be further refined and graded based on the degree of difference between the predicted probability and the preset probability. For example, when the predicted probability is only slightly higher than the preset probability, the user may only need to be advised to make minor usage adjustments or strengthen monitoring; however, when the predicted probability is significantly higher than the preset probability, the user may need to be urgently notified to take more stringent measures, or even consider immediately disabling the battery to avoid potential danger. This hierarchical output method helps users make more reasonable and effective decisions based on actual circumstances.

[0091] If the probability of battery insurance fatigue failure of multiple vehicles is predicted at the same time, the graded output method can also put vehicles with high risk levels at the front of the display interface, and can color-code them as warnings, using eye-catching colors to remind users that the corresponding vehicles need immediate maintenance.

[0092] For example, if the predicted probability is lower than the preset probability, this generally means that the battery's future risk of fatigue failure is within an acceptable range, and the battery's health is relatively good. It can continue to be used as planned without taking additional maintenance measures or adjusting usage strategies. In this case, a positive message can be output, informing the user that the battery is stable and the risk is controllable.

[0093] The above description is about how to use the insurance fatigue failure prediction model to predict the fatigue failure probability of the battery insurance. Furthermore, in some embodiments, the insurance fatigue failure prediction model is obtained by training the neural network model. Figure 3 Schematic diagram of the process of predicting battery insurance fatigue failure provided in this application Figure 2 ,like Figure 3 As shown in Figure 2, the training steps specifically include:

[0094] S301: Obtain sample data.

[0095] The sample data includes: total mileage, current stress distribution characteristics, and insurance fatigue failure status of at least one sample vehicle.

[0096] As you can understand, vehicle operating data is typically collected in real time through onboard sensors, global positioning systems, and the vehicle's electronic control unit (ECU), and stored centrally in a high-performance database. First, the database can be used to obtain the current stress distribution, stress duration, and fuse fatigue failure status of multiple vehicles.

[0097] After acquiring sample data, the continuous current stress is discretized through binning, and the number of battery stress load intensities in different classification intervals is counted. It is important to note that the classification interval division criteria during training are the same as the interval division criteria during prediction in the aforementioned embodiment. This can divide the input model data into the same standard scale, accelerating model learning and improving the model's predictive performance.

[0098] There are n intervals divided, and the total number of samples is k. During training, the samples can be divided into training samples and test samples based on the number of samples. The training samples are used for model learning and parameter adjustment, and the test samples are used for model evaluation. For example, 70% of the total number of samples can be used as training samples, and the remaining 30% can be used as test samples.

[0099] S302: Establish a neural network model based on sample data and task type.

[0100] As you can understand, the number of neural network layers and the number of neurons in each layer are defined based on the number of current intervals n in the input data and the number of outputs, which in turn determines the activation function. The number of neurons in the input layer is equal to the number of input features (the number of current intervals n). The number of hidden layers and the number of neurons in each layer are set based on the complexity of the problem, typically determined through experimentation or experience. The number of neurons in the output layer is equal to the number of outputs.

[0101] For example, we can set the hidden layer to have two layers, with m neurons in hidden layer 1 and i neurons in hidden layer 2. Since the model output is whether the battery fuse fatigue failure has occurred, which is a binary classification problem, we can choose the S-type function, also known as the Sigmoid function, as the activation function.

[0102] S303: Initialize the weight and bias of each neuron in the hidden layer.

[0103] As you can understand, weights determine how much the input signal affects the neuron's output, while biases are used to adjust the neuron's activation threshold. These initial values ​​are usually set to small random values. Figure 4 A schematic diagram of a neural network model provided in this application, such as Figure 4 As shown, the weight is W and the bias is b. The number of neurons in the input layer is n, the number of neurons in the hidden layer 1 is m, the number of neurons in the hidden layer 2 is i, and the final output layer is one neuron. The calculation formula is:

[0104] The weight W0 and bias b0 from the input layer to the hidden layer 1:

[0105]

[0106] It should be noted that "0" appears as a superscript in the calculation formula to clearly identify the weight matrix of this level in mathematical symbols, that is, it indicates that this is the weight matrix and configuration matrix of the first hidden layer, and the subscript is usually used to indicate the specific elements in the matrix.

[0107] For example, represents the weight from the nth neuron in the input layer to the mth neuron in the hidden layer 1.

[0108] The weight W1 and bias b1 from hidden layer 1 to hidden layer 2:

[0109]

[0110] The weight W2 and bias b2 from hidden layer 2 to the output layer:

[0111]

[0112] S304: Input training samples and calculate layer by layer to obtain output results.

[0113] It can be understood that layer-by-layer, also known as forward propagation, specifically refers to inputting the training sample (Input) into the neural network model established in steps S302-S303, and calculating the output of each layer according to the weight and bias of the current layer until the final output (Output) is obtained.

[0114] Input layer to hidden layer 1: A 0 =Sigmoid(W 0 Input+b 0 );

[0115] Hidden layer 1 to hidden layer 2: A 1 =Sigmoid(W 1 A 0 +b 1 );

[0116] Hidden layer 2 to output layer: Output = Sigmoid (W 2 A 1 +b 2 ).

[0117] In the above calculation formula, A 0 represents the output of hidden layer 1, A 1 represents the output of hidden layer 2.

[0118] S305: Determine the output layer error based on the output result, and transfer it back to the network layer to determine the error of each network layer.

[0119] It can be understood that the error δ of each network layer is calculated in reverse order by comparing the output value and the expected data Y error. The expected data Y is the value indicating whether the battery fuse in the sample data has failed. The value corresponding to the fuse failure can be 1, and the value corresponding to the fuse not failing can be 0. The specific calculation formula includes:

[0120] Output layer error: δ 2 =Output-Y;

[0121] Hidden layer 2 error: δ 1 =(W 2 ) T ⊙δ 2 Sigmoid′(W 1 A 0 +b 1 );

[0122] Hidden layer 1 error: δ 0 =(W 1 ) T δ 1 ⊙Sigmoid′(W 0 Input+b 0 ).

[0123] S306: Update and adjust the weights and bias of each network layer based on the error and learning rate.

[0124] As you can understand, the weights and biases of different layers are updated based on the error calculated by backpropagation, where η is the learning rate. By adjusting the parameters, the model's loss on the training data is reduced and the training data is better fitted. The specific calculation formula includes:

[0125] Weights and biases from the input layer to hidden layer 1:

[0126] W0=W0-η(Input) T δ 0 ; b0=b0-ηδ 0 ;

[0127] Weights and biases from hidden layer 1 to hidden layer 2:

[0128] W1=W1-η(A 0 ) T δ 1 ; b1=b1-ηδ 1 ;

[0129] Weights and biases from hidden layer 2 to the output layer:

[0130] W2=W2-η(A 1 ) T δ 0; b2=b2-ηδ 2 .

[0131] S307: Perform batch iterative training according to a preset number of training times and determine the accuracy.

[0132] As can be appreciated, steps S304 to S306 are repeated until the preset number of training cycles is reached. Once the preset number of training cycles is reached, accuracy is verified using test samples. Through multiple iterations, the model continuously learns patterns in the data, gradually improving performance. Accuracy can be determined by the number of correct predictions and the total number of test samples.

[0133] S308: Determine an insurance fatigue failure prediction model based on the accuracy.

[0134] As will be appreciated, when the accuracy meets the preset accuracy, the model is determined as the insurance fatigue failure prediction model. The preset accuracy is determined based on the required prediction precision; for example, the preset accuracy can be 98%. If the accuracy does not meet the preset accuracy, the model hyperparameters need to be re-adjusted, and steps S303 to S307 are repeated until the accuracy meets the preset accuracy. Hyperparameters refer to data such as the number of neurons and the number of neural network layers set in step S302.

[0135] It should be noted that the data used to train the insurance fatigue failure prediction model all comes from the big data end. The data on the big data end will be updated with the operating status of multiple vehicles and the failure of battery insurance. Therefore, a time period can be set to train and iterate the insurance fatigue failure prediction model according to the time period.

[0136] The above process describes the specific steps for training an insurance fatigue failure prediction model using a neural network model. This training process is only one approach proposed in this solution. In practical applications, the above training process can also be used to train other network models based on specific needs to obtain different insurance fatigue failure prediction models. This application does not limit this practice.

[0137] The battery fuse fatigue failure prediction method provided in the present application utilizes big data to record current stress distribution characteristics and fuse fatigue failure states. This method trains a neural network model to generate a model and parameters that meet the desired accuracy requirements. The trained fuse fatigue failure prediction model is then used to predict the fatigue failure probability of the battery fuse in the vehicle being predicted. This method integrates and organizes multi-source data to establish a correlation model between failure states and operating conditions, improving prediction accuracy and achieving precise prediction of the failure probability of battery fuse components.

[0138] The following is a specific example of a method for predicting fatigue failure of a battery insurance. In the following example, the execution subject of this application may be a server located in the cloud, which may also be called a vehicle service end or a battery service end.

[0139] Figure 5 A schematic diagram of the structure of a server provided for this application, such as Figure 5 As shown, the server is deployed with two functional modules: a training module and a prediction module. The training module can use the aforementioned method to train a fatigue failure prediction model for the vehicle battery and provide it to the prediction module, so that the prediction module can predict fatigue failure of the vehicle battery based on the fatigue failure prediction model.

[0140] It should be understood that the above functional modules are merely functional divisions based on the method of this application. In specific implementations, the above actions are all implemented by the server.

[0141] The following describes how the estimation module predicts insurance fatigue failure:

[0142] The estimation module can periodically receive and store the vehicle's operating data.

[0143] When the estimation module receives a battery insurance fatigue failure prediction request from a vehicle, the estimation module may obtain the vehicle's operating data from a database according to the vehicle's identification information included in the request.

[0144] As you can see, vehicle operating data is a crucial basis for evaluating vehicle performance, predicting potential failures, and optimizing maintenance strategies. This data is typically collected in real time through onboard sensors, global positioning systems, and the vehicle's internal ECU, and then centrally stored in a high-performance database.

[0145] When a battery insurance fatigue failure prediction request is received for a specific vehicle, it carries the vehicle's identification information, such as the Vehicle Identification Number (VIN) or a specific registration number. Based on this identification, the corresponding vehicle can be quickly located and relevant operating data from the database can be extracted, including but not limited to key parameters such as the number of battery charge and discharge cycles, current load intensity, ambient temperature records, mileage, and driving conditions.

[0146] It should be understood that the above-mentioned battery insurance fatigue failure prediction request can be triggered by a user on the vehicle side, or by the vehicle's operation and maintenance personnel or monitoring personnel, or by the battery's operation and maintenance personnel or monitoring personnel, etc.

[0147] Optionally, in some embodiments, the request may further include an estimated future time, or the estimated future time may be a preset default value.

[0148] After acquiring the vehicle operation data, the estimation module may obtain the number of current stress load intensities in each interval corresponding to the traveled mileage that exceeds a current stress load intensity threshold corresponding to the interval.

[0149] As you can understand, during vehicle operation, the battery, acting as a power source, continuously supplies current to power components such as the motor. The magnitude of this current directly reflects the battery's discharge capacity and the motor's power requirements. Within its normal operating range, the impact of current on the battery is manageable, and the battery can stably provide the required power. However, when the current exceeds the battery's designed normal operating range, current stress occurs. This stress can damage battery components, accelerating battery aging, causing capacity loss, and triggering thermal runaway.

[0150] A battery's current handling capacity is dynamically affected by factors such as temperature, state of charge, and aging. Different operating conditions (such as rapid acceleration, hill climbing, and fast charging) require different current demands and generate different amounts of heat. This necessitates setting different current stress load intensity thresholds to ensure safety, extend battery life, and maintain balanced performance. Within the ranges corresponding to different operating conditions, the corresponding current stress load intensity thresholds are different.

[0151] For each interval, the current stress load intensity values ​​at multiple moments in that interval can be compared with the corresponding current stress load intensity threshold. When the current stress load intensity threshold is exceeded, a count can be performed, or the current stress load intensity value at that moment can be marked or extracted. After comparing the current stress load intensity values ​​at all moments, the total count value, or the marked or extracted current stress load intensity values ​​for that interval, can be used to calculate the number of current stress load intensity values ​​that exceeded the current stress load intensity threshold in each interval, i.e., the current stress quantity. Because the current stress quantity is correlated with mileage, the current stress quantity per unit mileage can be determined.

[0152] For example, assume the vehicle has five zones, each with a mileage of 1000 kilometers. After sorting and analyzing the operating data, the current stress values ​​for each zone are (50, 80, 60, 40, 20). Furthermore, the operating data is from the past month, indicating a mileage of 2000 kilometers. Therefore, the current stress value per mileage is half of the monthly current stress value, corresponding to (25, 40, 30, 20, 10).

[0153] The total mileage in the future is then determined: based on the total mileage the vehicle has traveled, the average mileage of the vehicle per unit time is determined.

[0154] The unit time may be, for example, a week or a month, which is not limited.

[0155] It can be understood that the total mileage traveled by the vehicle can be obtained through the on-board recording equipment, and the average mileage can be obtained by analyzing the total mileage and the time taken by the vehicle to complete the total mileage. This processing method can evaluate the overall usage intensity of the vehicle and is not easily affected by short-term fluctuations.

[0156] The average mileage can also be determined by analyzing a specific recent time period. For example, the mileage of the last 30 days or 90 days can be selected to obtain the average mileage. This processing method can monitor the recent frequency of vehicle use and better predict the mileage in the future.

[0157] Get the estimated distance from the current time to the future time based on the future time, the current time, and the average distance.

[0158] Based on the total mileage and estimated mileage, determine the total future mileage.

[0159] As you can understand, the future time period refers to the predicted battery insurance failure timeframe. This timeframe is typically based on user needs. For example, the probability of a vehicle's battery insurance failure due to fatigue can be predicted within the next 30 days. Multiplying the average mileage per unit time by the future time period yields the estimated mileage from the current time to the future. Note that the unit standards for the unit time and the future time period must be consistent when multiplying. Adding the estimated mileage to the current total mileage yields the total future mileage.

[0160] For example, assuming that the unit time is weeks, the total mileage of the vehicle is 50,000 kilometers, and the vehicle travels an average of 500 kilometers per week, the total mileage for the next 60 days is predicted.

[0161] First, convert 60 days into unit time, that is, 60÷7=8.57 weeks.

[0162] The estimated mileage from the current time to the future time is: 8.57×500=4375 kilometers.

[0163] The total mileage in the future will be 50,000+4,375=54,375 kilometers.

[0164] Based on the number of intervals corresponding to the mileage already traveled and the mileage difference between the mileage already traveled and the future total mileage, the number of intervals corresponding to the future total mileage is obtained. Based on the number of intervals corresponding to the future total mileage, the current stress distribution characteristics are determined.

[0165] It is understandable that in the process of determining the total mileage in the future, the estimated mileage in the future time period is also determined. The estimated mileage is the mileage difference between the mileage already traveled and the total mileage in the future. According to the current stress quantity per unit mileage and the estimated mileage, the current stress quantity in the future time period can be determined. The current stress quantity in the future time period is combined with the current stress quantity obtained from the vehicle operation data to obtain the current stress distribution characteristics.

[0166] For example, using the above example, the current stress value corresponding to the estimated mileage (4375 kilometers) is 4.375 times the unit mileage (1000 kilometers). Therefore, the current stress value for the estimated mileage is (109.175.131.88.44). Because the current stress value is an integer, decimals can be rounded. The resulting current stress distribution characteristic can be the sum of the current stress values ​​corresponding to the estimated mileage and the mileage traveled, which corresponds to (159.255.191.128.64).

[0167] Finally, the future total mileage and current stress distribution characteristics are input into the insurance fatigue failure prediction model to obtain the vehicle's battery insurance fatigue failure probability.

[0168] Among them, the insurance fatigue failure prediction model is based on Figure 3 The training module is obtained by the training method shown.

[0169] Comparing the obtained battery insurance fatigue failure probability with the preset probability can trigger the early warning mechanism and output the corresponding early warning information.

[0170] Figure 6 This is a schematic diagram of the structure of the battery insurance fatigue failure prediction device provided in this application, such as Figure 6 As shown, the battery insurance fatigue failure prediction device 40 provided in this embodiment includes:

[0171] An acquisition module 401 is configured to acquire, based on the vehicle's operating data, a total future mileage of the vehicle's battery at a future time, and a current stress distribution characteristic corresponding to the total future mileage;

[0172] The prediction module 402 is configured to predict the probability of insurance fatigue failure of the battery at the future time based on the future total mileage and the current stress distribution characteristics.

[0173] In one possible implementation, the acquisition module 401 is specifically used to obtain, based on the operating data of the vehicle, an interval of current stress load intensity corresponding to the mileage traveled by the vehicle; different intervals correspond to different operating conditions of the vehicle; and based on the current stress load intensity of each of the intervals corresponding to the mileage traveled, and the total mileage in the future, determine the current stress distribution characteristics.

[0174] In one possible implementation, the acquisition module 401 is specifically used to obtain the number of current stress load intensities in each of the intervals corresponding to the mileage traveled that exceeds the current stress load intensity threshold corresponding to the interval; based on the number of each of the intervals corresponding to the mileage traveled and the mileage difference between the mileage traveled and the future total mileage, obtain the number of each of the intervals corresponding to the future total mileage; based on the number of each of the intervals corresponding to the future total mileage, determine the current stress distribution characteristics.

[0175] In one possible implementation, the acquisition module 401 is specifically used to determine the average mileage of the vehicle per unit time based on the total mileage traveled by the vehicle; obtain the estimated mileage from the current time to the future time based on the future time, the current time, and the average mileage; and determine the future total mileage based on the total mileage and the estimated mileage.

[0176] In one possible implementation, the prediction module 402 is specifically used to predict the probability of insurance fatigue failure of the battery at the target time based on the future total mileage, the current stress distribution characteristics, and the insurance fatigue failure prediction model. The insurance fatigue failure prediction model is obtained by pre-training a neural network model using sample data, and the sample data includes: the total mileage, current stress distribution characteristics, and insurance fatigue failure status of at least one sample vehicle.

[0177] In a possible implementation, the device further includes: an output module 403;

[0178] The output module 403 is configured to output the vehicle's battery insurance fatigue failure warning information when the probability of the battery insurance fatigue failure at the future time is greater than a preset probability threshold.

[0179] In a possible implementation, the apparatus further includes: a receiving module 404;

[0180] The receiving module 404 is configured to receive a battery insurance fatigue failure prediction request from the vehicle, the request including identification information of the vehicle; and obtain operating data of the vehicle from a database based on the identification information of the vehicle.

[0181] The battery insurance fatigue failure prediction device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar, and are not described in detail in this embodiment.

[0182] Figure 7 This is a schematic diagram of the structure of the electronic device provided in this application. Figure 7 As shown, the electronic device 50 provided in this embodiment includes: at least one processor 501 and a memory 502. Optionally, the device 50 further includes a communication component 503. The processor 501, the memory 502 and the communication component 503 are connected via a bus.

[0183] In a specific implementation process, at least one processor 501 executes the computer-executable instructions stored in the memory 502, so that the at least one processor 501 performs the above method.

[0184] The specific implementation process of the processor 501 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0185] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules in the processor.

[0186] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (NVM), such as at least one disk memory.

[0187] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be classified into address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0188] The present application also provides a vehicle, characterized in that the vehicle includes the electronic device as described above.

[0189] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0190] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0191] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0192] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0193] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, either through an interface, electrical, mechanical, or other means.

[0194] Units described as separate components may or may not be physically separate, and 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 these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0195] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0196] If the function is implemented in the form of 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, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0197] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0198] Finally, it should be noted that those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The present invention is not limited to the precise structure described above and illustrated in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof. The scope of the present invention is limited solely by the appended claims.

Claims

1. A method for predicting battery insurance fatigue failure, characterized in that: include: Obtaining, based on the vehicle operation data, a future total mileage of a battery of the vehicle at a future time, and a current stress distribution characteristic corresponding to the future total mileage; Based on the future total mileage and the current stress distribution characteristics, the probability of insurance fatigue failure of the battery at the future time is predicted.

2. The method according to claim 1, characterized in that The obtaining, based on the vehicle operation data, current stress distribution characteristics corresponding to the future total mileage of the vehicle includes: Based on the operating data of the vehicle, obtaining the interval of the current stress load intensity corresponding to the mileage of the vehicle; different intervals correspond to different operating conditions of the vehicle; The current stress distribution characteristics are determined based on the current stress load intensity of each of the sections corresponding to the traveled mileage and the future total mileage.

3. The method according to claim 2, characterized in that The determining of the current stress distribution characteristics based on the current stress load intensity of each of the intervals corresponding to the traveled mileage and the future total mileage includes: Obtaining the number of current stress load intensities in each of the intervals corresponding to the traveled mileage that exceeds a current stress load intensity threshold corresponding to the interval; Obtaining the number of each of the intervals corresponding to the future total mileage based on the number of each of the intervals corresponding to the traveled mileage and the mileage difference between the traveled mileage and the future total mileage; The current stress distribution characteristic is determined based on the number of each of the intervals corresponding to the future total mileage.

4. The method according to claim 1, wherein The obtaining of the total mileage of the vehicle in the future based on the vehicle operation data includes: Determining an average mileage of the vehicle per unit time based on the total mileage traveled by the vehicle; Obtaining an estimated mileage from the current time to the future time based on the future time, the current time, and the average mileage; The future total mileage is determined based on the total mileage and the estimated mileage.

5. The method according to any one of claims 1 to 4, characterized in that The predicting of the probability of the battery's insurance fatigue failure at the future time based on the future total mileage and the current stress distribution characteristics includes: Based on the future total mileage, the current stress distribution characteristics, and the insurance fatigue failure prediction model, the insurance fatigue failure probability of the battery at the target time is predicted. The insurance fatigue failure prediction model is obtained by pre-training a neural network model using sample data, and the sample data includes: the total mileage, current stress distribution characteristics, and insurance fatigue failure probability of at least one sample vehicle.

6. The method according to any one of claims 1 to 4, characterized in that After predicting the probability of the battery's fatigue failure at the future time, the method further includes: When the probability of insurance fatigue failure of the battery at the future time is greater than a preset probability threshold, outputting warning information of insurance fatigue failure of the vehicle's battery.

7. The method according to any one of claims 1 to 4, characterized in that The method further comprises: receiving a battery insurance fatigue failure prediction request from the vehicle, the request including identification information of the vehicle; Based on the identification information of the vehicle, operating data of the vehicle is obtained from a database.

8. A battery insurance fatigue failure prediction device, characterized in that: include: an acquisition module, configured to acquire, based on the vehicle's operating data, a total future mileage of the vehicle's battery at a future time, and a current stress distribution characteristic corresponding to the total future mileage; A prediction module is used to predict the probability of insurance fatigue failure of the battery at the future time based on the future total mileage and the current stress distribution characteristics.

9. An electronic device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 7.

10. A vehicle, characterized in that: The vehicle includes the electronic device according to claim 9.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

12. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when executed by a processor.

Images