Real-time residual capacity and residual mileage prediction method and system for lead-acid battery

Abstract

This invention relates to a method and system for predicting the real-time remaining capacity and range of a lead-acid battery. It continuously collects the battery's current, voltage, and temperature signals; inputs the temperature signal into a first neural network model to obtain a temperature calibration coefficient; and inputs the voltage signal into a second neural network model to obtain a voltage calibration coefficient. Based on predefined criteria, it calculates the current remaining capacity using either a first strategy or a second strategy. The first strategy calculates based on the temperature calibration coefficient, a first standard total capacity, and the current; the second strategy calculates based on the voltage calibration coefficient, a second standard total capacity, and the current. This invention provides accurate calculation results with low hardware requirements. The resulting range prediction is accurate and efficient. Furthermore, this invention integrates a local battery management module with a cloud management module to achieve data synchronization and calculation result correction. It can periodically correct calculation errors, update network parameters, and automatically upgrade the complete calculation model.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to a method and system for predicting the real-time remaining power and remaining range of a lead-acid battery. Background Technology

[0002] For electric bicycles and motorcycles, the battery is the sole source of power, making battery data crucial for the intelligent systems of these two-wheeled vehicles. During operation, users are most concerned with how much power has been consumed and how much range remains. This requires the vehicle control system to predict the remaining range based on the current driving conditions. However, remaining battery capacity, as an internal characteristic of the battery pack, cannot be directly measured. The relationship between remaining battery capacity and other battery parameters, such as voltage, current, and internal resistance, exhibits a high degree of non-linearity, making accurate estimation of remaining capacity difficult to achieve.

[0003] In existing technologies, the battery capacity prediction for lead-acid battery electric vehicles primarily uses a voltage conversion algorithm, directly calculating the battery capacity from the voltage to provide users with an approximate display. This method is inaccurate. First, errors in the raw data collection lead to inaccurate battery capacity readings. Second, during riding, battery voltage fluctuates with load changes, causing fluctuations in battery capacity information. For example, turning the accelerator lever immediately reduces the displayed battery capacity, making it difficult to improve the accuracy of subsequent range estimations. Furthermore, traditional prediction methods cannot accurately record the battery's charging and discharging status or assess battery aging. As lead-acid batteries age, their capacity decreases significantly. Using them with the charging and discharging procedures of a new battery poses a significant safety hazard, potentially leading to bulging or even fire. To address these issues, advanced solutions, such as those for lithium batteries, typically use an analog front-end (AFE) combined with an estimator (e.g., a Kalman filter) for battery capacity estimation. However, this method is costly in hardware and unsuitable for two-wheeled vehicle systems, thus limiting its application. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for predicting the real-time remaining power and range of lead-acid batteries, with the aim of improving the accuracy of power prediction.

[0005] The technical solution adopted in this invention is as follows:

[0006] This application provides a method for predicting the real-time remaining power of a lead-acid battery, including:

[0007] Continuously collect the battery's current, voltage, and temperature signals;

[0008] The temperature signal is input into the first neural network model to obtain the temperature calibration coefficient. The temperature calibration coefficient is used to characterize the ratio of the total battery capacity corresponding to the current battery temperature to the first standard total battery capacity corresponding to the standard temperature.

[0009] The voltage signal is input into the second neural network model to obtain the voltage calibration coefficient. The voltage calibration coefficient is used to characterize the ratio of the total battery capacity corresponding to the current battery voltage to the second standard total capacity corresponding to the battery under the standard voltage.

[0010] Based on the established judgment criteria, the current remaining power is calculated using either a first strategy or a second strategy. The first strategy calculates the remaining power based on the temperature calibration coefficient, the first standard total capacity, and the current. The second strategy calculates the remaining power based on the voltage calibration coefficient, the second standard total capacity, and the current.

[0011] The further technical solution is as follows:

[0012] The determination criteria are as follows:

[0013] The first strategy is used to calculate the current remaining power until: the battery current is detected to be continuously zero for a set time period, and the measured voltage signal is a static open-circuit voltage signal. The difference between the total battery capacity calibrated by the voltage calibration coefficient at the current moment and the power measured by current integration is calculated. If the difference is greater than the set value, the second strategy is used to calculate the current remaining power.

[0014] The first strategy and the second strategy also include:

[0015] The current remaining power is corrected by a vehicle self-loss calibration coefficient to obtain the corrected current remaining power. The vehicle self-loss calibration coefficient is used to characterize the power loss caused by the increase of vehicle self-loss over time.

[0016] The method for constructing the first neural network model includes:

[0017] A standard temperature and a first standard total capacity corresponding to the standard temperature are set. Temperature signals of the battery under various operating conditions and total battery capacity signals corresponding to the temperature signals are collected as a sample set. The ratio of the total battery capacity in the sample set to the first standard total capacity is calculated as a temperature calibration coefficient. A corresponding temperature calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning and testing, a first neural network model is obtained to establish the mapping relationship between temperature and temperature calibration coefficient.

[0018] The method for constructing the second neural network model includes:

[0019] A standard battery voltage and a second standard total capacity corresponding to the standard voltage are set. Battery voltage signals and corresponding total battery capacity signals under various operating conditions are collected as a sample set. The ratio of the total battery capacity in the sample set to the second standard total capacity is calculated as a voltage calibration coefficient. A corresponding voltage calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning, and testing, a second neural network model is obtained to establish the mapping relationship between voltage and voltage calibration coefficient.

[0020] The various operating conditions include non-driving conditions when the battery is charging, and driving and non-driving conditions when the battery is discharging.

[0021] The aforementioned method for predicting the real-time remaining power of a lead-acid battery further includes:

[0022] The remaining battery power calculated locally is uploaded to the cloud server to obtain the average value of multiple local calculation results. The local calculation results are compared with the average value output by the cloud server to evaluate the error value. When the error value is greater than a set value, the cloud data is synchronized to the local server to correct the local calculation results.

[0023] The aforementioned method for predicting the real-time remaining power of a lead-acid battery further includes:

[0024] The system continuously acquires the current battery output power and determines whether the output power exceeds the standard design power. If it does, an alarm message is sent to the vehicle's instrument panel.

[0025] In addition, it records the number of battery cycles after each power consumption and sends a replacement reminder when the number of battery cycles reaches a set threshold.

[0026] This application also provides a lead-acid battery management system, including a local battery management module and a cloud management module, wherein the local battery management module includes:

[0027] The front-end acquisition module is used to continuously acquire the battery's current, voltage, and temperature signals.

[0028] The signal processing module is used to input the temperature signal into a first neural network model to obtain a temperature calibration coefficient, which is used to characterize the ratio of the total battery capacity corresponding to the current battery temperature to the first standard total capacity corresponding to the battery at a standard temperature; and to input the voltage signal into a second neural network model to obtain a voltage calibration coefficient, which is used to characterize the ratio of the total battery capacity corresponding to the current battery voltage to the second standard total capacity corresponding to the battery at a standard voltage.

[0029] The calculation module is used to calculate the current remaining power according to the set judgment criteria, using a first strategy or a second strategy. The first strategy is to calculate based on the temperature calibration coefficient, the first standard total capacity and the current, and the second strategy is to calculate based on the voltage calibration coefficient, the second standard total capacity and the current.

[0030] The calculation module is also used to correct the current remaining power using a vehicle self-loss calibration coefficient to obtain the corrected current remaining power. The vehicle self-loss calibration coefficient is used to characterize the power loss caused by the increase of vehicle self-loss over time.

[0031] The cloud management module is used to connect with the local battery management module, process local data to obtain an average value, and correct the local calculation results.

[0032] This application also provides a method for predicting the real-time remaining range of a vehicle based on the aforementioned lead-acid battery real-time remaining power prediction method, comprising:

[0033] Collect the vehicle's mileage and power consumption from the previous moment to the current moment, and obtain the current power consumption per kilometer based on the mileage and power consumption.

[0034] The remaining driving range is calculated based on the current power consumption per kilometer and the current remaining power.

[0035] The method for constructing the first neural network model includes:

[0036] A standard temperature and a first standard total capacity corresponding to the standard temperature are set. Temperature signals of the battery under various operating conditions and total battery capacity signals corresponding to the temperature signals are collected as a sample set. The ratio of the total battery capacity in the sample set to the first standard total capacity is calculated as a temperature calibration coefficient. A corresponding temperature calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning and testing, a first neural network model is obtained to establish the mapping relationship between temperature and temperature calibration coefficient.

[0037] The method for constructing the second neural network model includes:

[0038] A standard battery voltage and a second standard total capacity corresponding to the standard voltage are set. Battery voltage signals and corresponding total battery capacity signals under various operating conditions are collected as a sample set. The ratio of the total battery capacity in the sample set to the second standard total capacity is calculated as a voltage calibration coefficient. A corresponding voltage calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning, and testing, a second neural network model is obtained to establish the mapping relationship between voltage and voltage calibration coefficient.

[0039] The further technical solution is as follows:

[0040] The remaining mileage calculated locally is uploaded to the cloud server to obtain the average of multiple local calculation results. The local calculation results are compared with the average value output by the cloud server to evaluate the error value. When the error value is greater than a set value, the cloud data is synchronized to the local server to correct the local calculation results.

[0041] The beneficial effects of this invention are as follows:

[0042] This invention constructs a neural network model to establish a correspondence between acquired signals and calibration coefficients. By selecting the appropriate calibration coefficients based on corresponding judgment criteria and combining this with the current integral term, the remaining battery power is calculated. The calculation results are accurate and have low hardware requirements. The resulting mileage prediction is accurate and highly efficient.

[0043] This invention combines local and cloud modules to achieve data synchronization and correction of local calculation results. Based on the correction results, network parameters can be updated, and the calculation model can be automatically upgraded.

[0044] This invention enables accurate assessment of battery cycle count through power prediction, thereby predicting and alerting on battery life and improving safety.

[0045] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating the real-time remaining power prediction method for lead-acid batteries according to an embodiment of the present invention.

[0047] Figure 2 This is a schematic diagram of the lead-acid battery management system according to an embodiment of the present invention. Detailed Implementation

[0048] The specific embodiments of the present invention are described below with reference to the accompanying drawings.

[0049] See Figure 1 This embodiment provides a method for predicting the real-time remaining power of a lead-acid battery, including:

[0050] Continuously collect the battery's current, voltage, and temperature signals;

[0051] The temperature signal is input into the first neural network model to obtain the temperature calibration coefficient. The temperature calibration coefficient is used to characterize the ratio of the total battery capacity corresponding to the current battery temperature to the first standard total battery capacity corresponding to the standard temperature.

[0052] The voltage signal is input into the second neural network model to obtain the voltage calibration coefficient. The voltage calibration coefficient is used to characterize the ratio of the total battery capacity corresponding to the current battery voltage to the second standard total capacity corresponding to the battery under the standard voltage.

[0053] Based on the established judgment criteria, the current remaining power is calculated using either a first strategy or a second strategy. The first strategy calculates the remaining power based on the temperature calibration coefficient, the first standard total capacity, and the current. The second strategy calculates the remaining power based on the voltage calibration coefficient, the second standard total capacity, and the current.

[0054] The determination criteria are as follows:

[0055] The first strategy is used to calculate the current remaining power until: the battery current is detected to be continuously zero for a set time period, and the measured voltage signal is a static open-circuit voltage signal. The difference between the total battery capacity calibrated by the voltage calibration coefficient at the current moment and the power measured by current integration is calculated. If the difference is greater than the set value, the second strategy is used to calculate the current remaining power.

[0056] The specific calculation process for the first strategy is as follows:

[0057] Calculate the integral of the current with respect to time to obtain the integral term; , For current, p For power, t For time,

[0058] Total capacity of the first standard C A With temperature calibration coefficient A Multiply to obtain the corrected capacity:

[0059] Divide the integral term by the corrected total capacity to obtain the remaining power.

[0060] The specific calculation process for the second strategy is as follows:

[0061] Calculate the integral of the current with respect to time to obtain the integral term; , For current, p For power, t For time,

[0062] The second standard total capacity C B With voltage calibration coefficient B Multiply to obtain the corrected capacity:

[0063] Divide the integral term by the corrected total capacity to obtain the remaining power.

[0064] The first strategy and the second strategy also include:

[0065] The current remaining power is corrected by a vehicle self-loss calibration coefficient to obtain the corrected current remaining power. The vehicle self-loss calibration coefficient is used to characterize the power loss caused by the increase of vehicle self-loss over time.

[0066] Some components in the vehicle itself, which are constantly powered, consume a tiny amount of current during standby operation. This consumption is so minute that current sensors cannot accurately detect it. After confirming the vehicle's status, this minute power consumption can be measured using precision laboratory instruments and calibrated into the algorithm. The vehicle's self-loss calibration coefficient is accumulated over time. As it increases, the remaining battery power calculated using either the first or second strategy is multiplied by the self-loss calibration coefficient, effectively deducting the self-loss power consumption to obtain the corrected remaining battery power.

[0067] The remaining power estimation method in this embodiment also includes recording the number of battery cycles after the power is consumed, and issuing a replacement reminder when the number of battery cycles reaches a set threshold.

[0068] In the above embodiments, the method for constructing the first neural network model includes:

[0069] Set the standard temperature of the battery and the first standard total capacity corresponding to the standard temperature. C A The battery temperature signals and the corresponding total battery capacity signals under various operating conditions are collected as a sample set, and the total battery capacity in the sample set is calculated. C Total capacity of the first standard C A The ratio of these values ​​is used as the temperature calibration coefficient. A The total battery capacity signal in the sample set is labeled with a corresponding temperature calibration coefficient. The sample set is then input into a neural network. After network training, learning, and testing, a first neural network model is obtained to establish the mapping relationship between temperature and temperature calibration coefficient.

[0070] In the above embodiments, the method for constructing the second neural network model includes:

[0071] Set the standard voltage of the battery and the second standard total capacity corresponding to the standard voltage. C B The battery voltage signals and the corresponding total battery capacity signals under various operating conditions are collected as a sample set, and the total battery capacity in the sample set is calculated. C Total capacity of the second standard C B The ratio is used as the voltage calibration coefficient. BThe total battery capacity signal in the sample set is labeled with a corresponding voltage calibration coefficient. The sample set is then input into the neural network. After network training, learning, and testing, a second neural network model is obtained to establish the mapping relationship between voltage and voltage calibration coefficient.

[0072] The various operating conditions include non-driving conditions where the battery is charging, and driving and non-driving conditions where the battery is discharging. By sampling samples under different operating conditions, the model can be made more accurate, enabling prediction of vehicle battery level under various conditions in real-world applications.

[0073] The real-time remaining power prediction method for lead-acid batteries in the above embodiments further includes:

[0074] The remaining power calculated locally is uploaded to the cloud server to obtain the average value of multiple local calculation results. The local calculation results are compared with the average value output by the cloud server to evaluate the error value. When the error value is greater than a set value, the cloud data is synchronized to the local server to correct the local calculation results.

[0075] Specifically, the cloud server carries a corresponding neural network model. The input of this neural network model is the calculation result uploaded locally, and the output is the average of multiple local calculation results. It is built through a large amount of learning and training.

[0076] The real-time remaining power prediction method for lead-acid batteries in the above embodiments further includes:

[0077] The system continuously acquires the current battery output power and determines whether the output power exceeds the standard design power. If it does, an alarm message is sent to the vehicle's instrument panel.

[0078] Based on the lead-acid battery real-time remaining power prediction method of the above embodiments, this embodiment also provides a vehicle real-time remaining range estimation method, including:

[0079] Collect the vehicle's mileage and power consumption from the previous moment to the current moment, and obtain the current power consumption per kilometer based on the mileage and power consumption.

[0080] The remaining driving range is calculated based on the current power consumption per kilometer and the current remaining power.

[0081] Also includes:

[0082] The remaining mileage calculated locally is uploaded to the cloud server to obtain the average of multiple local calculation results. The local calculation results are compared with the average value output by the cloud server to evaluate the error value. When the error value is greater than a set value (for example, the difference between the local calculation result and the average value output by the cloud server exceeds 5%), the cloud data is synchronized to the local machine to correct the local calculation results.

[0083] Specifically, the cloud server carries the corresponding neural network model. The input of this neural network model is the calculation result, and the output is the average of multiple local calculation results. It is built through a large amount of learning and training.

[0084] See Figure 2 This embodiment also provides a lead-acid battery management system, including a local battery management module and a cloud management module. The local battery management module includes:

[0085] The front-end acquisition module is used to continuously acquire the battery's current, voltage, and temperature signals.

[0086] The signal processing module is used to input the temperature signal into a first neural network model to obtain a temperature calibration coefficient, which is used to characterize the ratio of the total battery capacity corresponding to the current battery temperature to the first standard total capacity corresponding to the battery at a standard temperature; and to input the voltage signal into a second neural network model to obtain a voltage calibration coefficient, which is used to characterize the ratio of the total battery capacity corresponding to the current battery voltage to the second standard total capacity corresponding to the battery at a standard voltage.

[0087] The calculation module is used to calculate the current remaining power according to the set judgment criteria, using a first strategy or a second strategy. The first strategy is to calculate based on the temperature calibration coefficient, the first standard total capacity and the current, and the second strategy is to calculate based on the voltage calibration coefficient, the second standard total capacity and the current.

[0088] The calculation module is also used to correct the current remaining power using a vehicle self-loss calibration coefficient to obtain the corrected current remaining power. The vehicle self-loss calibration coefficient is used to characterize the power loss caused by the increase of vehicle self-loss over time.

[0089] The cloud management module is used to connect with the local battery management module to synchronize data and correct the local calculation results.

[0090] This application employs current, voltage, and temperature sensors to sample three types of data during battery use in front-end data acquisition. A neural network model is then used by the MCU for estimation, effectively solving the problems of inaccurate lead-acid battery charge, mileage, and cycle life. Furthermore, this application utilizes a local offline management model that works in conjunction with cloud-based mechanisms for self-iteration, ensuring the accuracy of lead-acid battery charge and lifespan management, resolving subsequent upgrade issues, and facilitating users' access to real-time vehicle information, thus reducing range anxiety. Simultaneously, data-driven reminders of lead-acid battery lifespan reduce the occurrence of fires.

[0091] It will be understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the real-time state of charge of a lead-acid battery, characterized in that, include: Continuously collect the battery's current, voltage, and temperature signals; The temperature signal is input into the first neural network model to obtain the temperature calibration coefficient. The temperature calibration coefficient is used to characterize the ratio of the total battery capacity corresponding to the current battery temperature to the first standard total battery capacity corresponding to the standard temperature. The voltage signal is input into the second neural network model to obtain the voltage calibration coefficient. The voltage calibration coefficient is used to characterize the ratio of the total battery capacity corresponding to the current battery voltage to the second standard total capacity corresponding to the battery under the standard voltage. According to the established judgment criteria, the current remaining power is calculated using either the first strategy or the second strategy. The first strategy is to calculate based on the temperature calibration coefficient, the first standard total capacity, and the current. The second strategy is to calculate based on the voltage calibration coefficient, the second standard total capacity, and the current. The method for constructing the first neural network model includes: A standard temperature and a first standard total capacity corresponding to the standard temperature are set. Temperature signals of the battery under various operating conditions and total battery capacity signals corresponding to the temperature signals are collected as a sample set. The ratio of the total battery capacity in the sample set to the first standard total capacity is calculated as a temperature calibration coefficient. A corresponding temperature calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning and testing, a first neural network model is obtained to establish the mapping relationship between temperature and temperature calibration coefficient. The method for constructing the second neural network model includes: A standard battery voltage and a second standard total capacity corresponding to the standard voltage are set. Battery voltage signals and corresponding total battery capacity signals under various operating conditions are collected as a sample set. The ratio of the total battery capacity in the sample set to the second standard total capacity is calculated as a voltage calibration coefficient. A corresponding voltage calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning, and testing, a second neural network model is obtained to establish the mapping relationship between voltage and voltage calibration coefficient.

2. The lead-acid battery real-time state-of-charge prediction method of claim 1, wherein, The determination criteria are as follows: The first strategy is used to calculate the current remaining power until: the battery current is detected to be continuously zero for a set time period, and the measured voltage signal is a static open-circuit voltage signal. The difference between the total battery capacity calibrated by the voltage calibration coefficient at the current moment and the power measured by current integration is calculated. If the difference is greater than the set value, the second strategy is used to calculate the current remaining power.

3. The lead-acid battery real-time state-of-charge prediction method of claim 1, wherein, The first strategy and the second strategy also include: The current remaining power is corrected by a vehicle self-loss calibration coefficient to obtain the corrected current remaining power. The vehicle self-loss calibration coefficient is used to characterize the power loss caused by the increase of vehicle self-loss over time.

4. The lead-acid battery real-time state-of-charge prediction method of claim 1, wherein, The various operating conditions include non-driving conditions when the battery is charging, and driving and non-driving conditions when the battery is discharging.

5. The method for predicting the real-time remaining power of a lead-acid battery according to claim 1, characterized in that, Also includes: The remaining battery power calculated locally is uploaded to the cloud server to obtain the average value of multiple local calculation results. The local calculation results are compared with the average value output by the cloud server to evaluate the error value. When the error value is greater than a set value, the cloud data is synchronized to the local server to correct the local calculation results.

6. The method for predicting the real-time remaining power of a lead-acid battery according to claim 1, characterized in that, Also includes: The system continuously acquires the current battery output power and determines whether the output power exceeds the standard design power. If it does, an alarm message is sent to the vehicle's instrument panel. In addition, it records the number of battery cycles after each power consumption and sends a replacement reminder when the number of battery cycles reaches a set threshold.

7. A lead-acid battery management system, characterized in that, It includes a local battery management module and a cloud management module, wherein the local battery management module includes: The front-end acquisition module is used to continuously acquire the battery's current, voltage, and temperature signals. The signal processing module is used to input the temperature signal into a first neural network model to obtain a temperature calibration coefficient, which is used to characterize the ratio of the total battery capacity corresponding to the current battery temperature to the first standard total capacity corresponding to the battery at a standard temperature; and to input the voltage signal into a second neural network model to obtain a voltage calibration coefficient, which is used to characterize the ratio of the total battery capacity corresponding to the current battery voltage to the second standard total capacity corresponding to the battery at a standard voltage. The calculation module is used to calculate the current remaining power according to the set judgment criteria, using a first strategy or a second strategy. The first strategy is to calculate based on the temperature calibration coefficient, the first standard total capacity and the current, and the second strategy is to calculate based on the voltage calibration coefficient, the second standard total capacity and the current. The calculation module is also used to correct the current remaining power using a vehicle self-loss calibration coefficient to obtain the corrected current remaining power. The vehicle self-loss calibration coefficient is used to characterize the power loss caused by the increase of vehicle self-loss over time. The cloud management module is used to connect with the local battery management module, process local data to obtain an average value, and correct the local calculation results; The method for constructing the first neural network model includes: A standard temperature and a first standard total capacity corresponding to the standard temperature are set. Temperature signals of the battery under various operating conditions and total battery capacity signals corresponding to the temperature signals are collected as a sample set. The ratio of the total battery capacity in the sample set to the first standard total capacity is calculated as a temperature calibration coefficient. A corresponding temperature calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning and testing, a first neural network model is obtained to establish the mapping relationship between temperature and temperature calibration coefficient. The method for constructing the second neural network model includes: A standard battery voltage and a second standard total capacity corresponding to the standard voltage are set. Battery voltage signals and corresponding total battery capacity signals under various operating conditions are collected as a sample set. The ratio of the total battery capacity in the sample set to the second standard total capacity is calculated as a voltage calibration coefficient. A corresponding voltage calibration coefficient label is added to the total battery capacity signal in the sample set. The sample set is input into a neural network. After network training, learning, and testing, a second neural network model is obtained to establish the mapping relationship between voltage and voltage calibration coefficient.

8. A method for predicting the real-time remaining range of a vehicle according to the lead-acid battery real-time remaining power prediction method according to any one of claims 1-6, characterized in that, include: Collect the vehicle's mileage and power consumption from the previous moment to the current moment, and obtain the current power consumption per kilometer based on the mileage and power consumption. The remaining driving range is calculated based on the current power consumption per kilometer and the current remaining power.

9. The method for predicting real-time remaining mileage of a vehicle according to claim 8, characterized in that, The remaining mileage calculated locally is uploaded to the cloud server to obtain the average of multiple local calculation results. The local calculation results are compared with the average value output by the cloud server to evaluate the error value. When the error value is greater than a set value, the cloud data is synchronized to the local server to correct the local calculation results.

Images