A method, device and system for evaluating the remaining life of a power battery of an electric vehicle

Through a combination of Gaussian process-machine learning and transfer learning, a comprehensive evaluation model of multi-type power battery is constructed and the charging data is used for evaluation, which solves the universality of power battery evaluation methods and resource waste, and achieves high precision and low cost life expectancy prediction.

CN114859230BActive Publication Date: 2024-11-22CHEBAI ZHONGQI TECH (BEIJING) CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202210390899.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-14
Publication Date
2024-11-22
Estimated Expiration
2042-04-14

AI Technical Summary

Technical Problem

The existing residual life evaluation methods of power batteries lack universality and are difficult to apply to different types of power batteries. They rely on on-board sensor data and do not fully utilize charging behavior data, resulting in waste of resources and high evaluation costs.

Method used

The Gaussian process-machine learning method is used to construct the power battery life decay data set, combining the state of charge, charging rate and temperature, and through the superposition law and transfer learning method, a comprehensive evaluation model of multi-type power batteries is constructed, and the charging pile end data is used for Evaluate.

Benefits of technology

It improves the accuracy and universality of the power battery life evaluation model, reduces the evaluation cost, and directly predicts the charging data, which improves the convenience of data acquisition.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114859230B_ABST
    Figure CN114859230B_ABST
Patent Text Reader

Abstract

The present invention relates to a method, device and system for evaluating the remaining life of a power battery of an electric vehicle, and belongs to the field of power batteries. The method adopts a method based on Gaussian process-machine learning combined with transfer learning to evaluate the remaining life of the power battery. A multi-factor impact comprehensive life model of different types of power batteries is constructed to effectively improve the accuracy of the power battery life model and greatly improve the generalization ability of the life model. Based on the power battery charging data, the remaining life of the power battery is predicted directly from the charging pile end, rather than based on the power battery discharge end data measured by multiple groups of vehicle-mounted sensors, which greatly improves the convenience of data acquisition and reduces the data acquisition cost.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of power batteries, and in particular to a method, device and system for evaluating the remaining life of a power battery of an electric vehicle. Background Art

[0002] As energy and environmental issues become increasingly severe, it has long been a consensus to vigorously develop electric vehicles. As one of the core components of electric vehicles, a reasonable and accurate assessment of the remaining life of power batteries can provide a reference for the control and management of power batteries, thereby improving the overall economy of the vehicle. There are many types of automotive power batteries. The most widely used lithium-ion batteries alone include lithium cobalt oxide batteries, lithium manganese oxide batteries, lithium iron phosphate batteries, and many other types. The remaining life assessment models of different types of power batteries are determined by their inherent characteristics. At present, the remaining life of power batteries is usually evaluated by using a model-based or data-driven method for a specific power battery. The remaining life model has extremely poor generalization ability and is almost impossible to apply to other types of power batteries. The specificity of the remaining life assessment method determined by the inherent characteristics of different types of power batteries greatly limits the scope of application of the remaining life assessment method of power batteries, resulting in a large waste of resources. Based on the objective facts that there are many types of current automotive power batteries, the remaining life assessment model is complex to establish, and the model accuracy is low, how to migrate and apply a power battery remaining life assessment method to another different type of power battery, abandon specificity and have universality, and thus achieve the goal of significantly reducing the comprehensive cost of power battery remaining life assessment is a technical issue worthy of in-depth study.

[0003] Furthermore, the current evaluation of the remaining life of power batteries is mostly based on the state data of the power battery discharge end. In particular, for the use of electric vehicles, it is often based on the data of the power battery in the discharge state collected by many on-board sensors, and the remaining life is evaluated according to a specific remaining life evaluation model established for this type of power battery. The accuracy of this type of method is often greatly dependent on the accuracy of the on-board sensors, and only considers a single electric vehicle, and does not fully utilize the infrastructure such as smart charging piles in the current electric vehicle network system that is being vigorously developed. For the charging behavior that each electric vehicle must perform, a large amount of charging end data is not fully utilized, resulting in a huge waste of resources. Therefore, how to make full use of the large amount of data generated by the charging behavior of electric vehicles to evaluate the remaining life of power batteries is an industry pain point. Summary of the invention

[0004] The purpose of the present invention is to provide a method, device and system for evaluating the remaining life of a power battery of an electric vehicle, so as to improve the accuracy and universality of the life evaluation model, make full use of the data generated by the charging behavior of the electric vehicle, and reduce the comprehensive cost of the remaining life evaluation of the power battery.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for evaluating the remaining life of a power battery of an electric vehicle, the method comprising:

[0007] Constructing a life decay data set of a type of power battery; the life decay data set includes a health state, a life cycle number, and a discharge depth that can reflect the life decay trajectory of the power battery;

[0008] The Gaussian process-machine learning method is used to train the power battery life degradation law reflected by the life degradation data set to obtain a power battery life basic degradation assessment model; the power battery life basic degradation assessment model is a function of the health status of the power battery with respect to the life cycle number and the discharge depth;

[0009] Based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the power battery health status based on the state of charge, charging rate and temperature to construct a type of power battery life degradation comprehensive assessment model;

[0010] Based on a type of comprehensive evaluation model for power battery life degradation, a transfer learning method is used to construct multiple different types of power battery remaining life evaluation models;

[0011] When the electric vehicle starts charging, the current status data of the electric vehicle power battery is obtained from the charging pile; the status data includes the state of charge, charging rate and temperature;

[0012] According to the current electric vehicle power battery type identified by the charging pile, a power battery remaining life assessment model corresponding to the power battery type is obtained;

[0013] According to the status data, using the power battery remaining life assessment model corresponding to the power battery type and the relationship between the discharge depth and the state of charge, the number of life cycles when the health state of the power battery reaches the end of life decline is determined as the remaining service life of the current electric vehicle power battery.

[0014] Optionally, the step of constructing a life degradation data set for a type of power battery specifically includes:

[0015] Conduct accelerated aging experiments on multiple groups of power batteries of the same type;

[0016] Record the remaining capacity and life cycle number of each group of power batteries after completing each life cycle;

[0017] When the remaining capacity is less than or equal to the capacity threshold, it is determined that the power battery has reached the end of life decline and the accelerated aging test is stopped;

[0018] The health status SOH of the power battery after each life cycle is determined using the formula SOH = Q / Q0; where Q represents the remaining capacity of the power battery and Q0 represents the rated capacity of the power battery;

[0019] According to the relationship between the life cycle number and the discharge depth of the same type of power battery, determine the discharge depth corresponding to each recorded life cycle number;

[0020] The health status and life cycle number of all groups of power batteries after each life cycle, as well as the discharge depth corresponding to each life cycle number, are recorded to form a life degradation data set Ψ (SOH, Cycles, DOD) of the same type of power battery; where Ψ represents the data set, Cycles represents the life cycle number, and DOD represents the discharge depth.

[0021] Optionally, the process of the accelerated aging experiment is:

[0022] Each group of power batteries is charged from a state of charge SOC of 80% to a state of charge SOC of 100%, and the charging voltage of each group of power batteries is made to 3.6V and the charging current is made to C / 50 by a constant current-constant voltage charging current of 1C; wherein C represents the rate;

[0023] Each group of power batteries is discharged in a constant current-constant voltage discharge step from a discharge current of 4C to a discharge voltage of 2.0V, and the discharge current cutoff value is C / 50, completing one life cycle of the power battery;

[0024] Repeat the above steps to complete multiple life cycles of the power battery.

[0025] Optionally, the Gaussian process-machine learning method is used to train the power battery life degradation law reflected by the life degradation data set to obtain a power battery life basic degradation assessment model, specifically including:

[0026] The life decline data set Ψ(SOH, Cycles, DOD) is reconstructed as SOH=ψ(Cycles, DOD); wherein ψ represents a health state function;

[0027] Divide the reconstructed life decay dataset into a training dataset and a test dataset;

[0028] In each training step, the training data set is further divided into a sub-training data set and a sub-testing data set;

[0029] The Gaussian process-machine learning method is used to train the power battery life degradation law reflected by the training data set within the current training step, and the test data set is used to complete the test verification, correct the degradation law, and obtain the power battery life degradation law under the current training step;

[0030] The power battery life degradation law under the current training step is passed to the next training step, and when training is performed in the next training step, the step "using the Gaussian process-machine learning method to train the power battery life degradation law reflected by the training data set in the current training step, and using the test data set to complete the test verification, correct the degradation law, and obtain the power battery life degradation law under the current training step" is returned;

[0031] After completing the training within all training steps, the trained power battery life degradation law is obtained;

[0032] The test data set is used to verify the trained power battery life degradation law, correct the degradation law, obtain the power battery life degradation law after the test, and determine it as the power battery life basic degradation assessment model.

[0033] Optionally, based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the state of charge, charge rate and temperature for the health status of the power battery to construct a type of power battery life degradation comprehensive assessment model, which specifically includes:

[0034] The correction model of the state of charge to the health state of the power battery is established as follows: Wherein, f() represents the correction relationship between the state of charge SOC and the health state SOH of the power battery, A i represents the polynomial coefficient of the i-th term, n represents the value of the highest order term in the polynomial form of the relationship between the state of charge SOC and the state of health SOH, SOC i represents the i-th term of the state of charge SOC;

[0035] The correction model of charging rate to the health status of power battery is established as follows: Among them, g() represents the correction relationship between the charging rate Crate of the power battery and the health state SOH, a, b, c and d represent the first, second, third, fourth and fifth coefficients respectively;

[0036] The correction model of temperature on the health status of power batteries is established as follows: Where p() represents the correction relationship between the temperature T of the power battery and the health state SOH, R2, R1, R0 and s represent the sixth, seventh, eighth and ninth coefficients respectively, T j represents the jth term of temperature T, B j represents the j-th term T with respect to temperature T j The coefficient of j is in the range of 0≤j≤5. Specifically, B5, B4, B3, B2, B1, and B0 represent the tenth, eleventh, twelfth, thirteenth, fourteenth, and fifteenth coefficients respectively.

[0037] Based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the power battery health status based on the state of charge, charging rate and temperature to construct a type of power battery life degradation comprehensive assessment model. Among them, ζ() represents the comprehensive evaluation function of power battery life degradation, It represents the basic degradation assessment model of power battery life. Represents the basic degradation assessment function of power battery life.

[0038] Optionally, the method of using a transfer learning method to construct multiple different types of power battery remaining life assessment models based on one type of power battery life degradation comprehensive assessment model specifically includes:

[0039] Based on the construction process of one type of comprehensive evaluation model for power battery life degradation, two other different types of comprehensive evaluation models for power battery life degradation are constructed;

[0040] The three types of comprehensive evaluation models for power battery life degradation are divided into training sets and test sets;

[0041] Based on the training set, a transfer learning method is used to construct a comprehensive evaluation model for life degradation of power batteries of types other than the three types.

[0042] The test set is used to verify and adjust the comprehensive evaluation model for life degradation of power batteries of types other than the three types, so as to obtain multiple different types of power battery remaining life evaluation models.

[0043] Optionally, according to the state data, a power battery remaining life assessment model corresponding to the power battery type and a relationship between the discharge depth and the state of charge are used to determine the life cycle number when the health state of the power battery reaches the end of life decline, specifically including:

[0044] According to the state of charge in the state data, the discharge depth of the current electric vehicle power battery is determined by using the relationship between the discharge depth DOD and the state of charge SOC DOD=1-SOC;

[0045] Inputting the state data and the current discharge depth of the electric vehicle power battery into a power battery remaining life assessment model corresponding to the power battery type, and obtaining a relationship between the life cycle number and the health state of the current electric vehicle power battery;

[0046] The health status is set to drop to 80% as the end point of life decline, and the life cycle number of the current electric vehicle power battery is calculated using the relationship between the life cycle number and the health status.

[0047] A device for evaluating the remaining life of a power battery of an electric vehicle, the device comprising: a charging pile end and a cloud end;

[0048] The cloud is connected to the charging pile end, and the cloud is used to store multiple different types of power battery remaining life assessment models;

[0049] The charging pile end is used to obtain the status data of the electric vehicle during charging, and at the same time identify the type of power battery of the charged electric vehicle, obtain the power battery remaining life assessment model corresponding to the power battery type of the charged electric vehicle from the cloud, and use the power battery remaining life assessment method of the electric vehicle to calculate the remaining service life of the power battery of the charged electric vehicle, and then transmit the remaining service life to the power battery management system of the charged electric vehicle.

[0050] Optionally, the device further includes: a first communication unit and a second communication unit;

[0051] The first communication unit is used to complete the communication between the charging pile end and the charging electric vehicle;

[0052] The second communication unit is used to complete the communication between the charging pile and the cloud.

[0053] A system for evaluating the remaining life of a power battery of an electric vehicle, the system comprising:

[0054] A life decay data set construction module is used to construct a life decay data set of a type of power battery; the life decay data set includes a health state, a life cycle number and a discharge depth that can reflect the life decay trajectory of the power battery;

[0055] A basic degradation assessment model acquisition module is used to use a Gaussian process-machine learning method to train the power battery life degradation law reflected by the life degradation data set to obtain a power battery life basic degradation assessment model; the power battery life basic degradation assessment model is a function of the health status of the power battery with respect to the life cycle number and the discharge depth;

[0056] A module for constructing a comprehensive evaluation model for the decline of power battery life, which is used to construct a comprehensive evaluation model for the decline of power battery life based on the superposition law by combining the basic evaluation model for the decline of power battery life with the correction models for the health status of power battery by state of charge, charging rate and temperature;

[0057] A multi-type model building module is used to build multiple different types of power battery remaining life assessment models based on one type of power battery life degradation comprehensive assessment model using a transfer learning method;

[0058] The status data acquisition module is used to obtain the current status data of the electric vehicle power battery from the charging pile when the electric vehicle starts charging; the status data includes the state of charge, charging rate and temperature;

[0059] A matching module is used to obtain a power battery remaining life assessment model corresponding to the power battery type according to the current power battery type of the electric vehicle identified by the charging pile end;

[0060] The remaining service life calculation module is used to determine the number of life cycles when the health state of the power battery reaches the end of life decline based on the status data, using the power battery remaining service life assessment model corresponding to the power battery type and the relationship between the discharge depth and the state of charge, as the remaining service life of the current electric vehicle power battery.

[0061] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0062] The present invention discloses a method, device and system for evaluating the remaining life of a power battery of an electric vehicle. The method adopts a method based on a combination of Gaussian process-machine learning and transfer learning to evaluate the remaining life of the power battery. A multi-factor impact comprehensive life model of different types of power batteries is constructed to effectively improve the accuracy of the power battery life model while greatly improving the generalization ability of the life model. Based on the power battery charging data, the remaining life of the power battery is predicted directly from the charging pile end, rather than based on the power battery discharge end data measured by multiple groups of vehicle-mounted sensors, which greatly improves the convenience of data acquisition and reduces the data acquisition cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0064] Figure 1A flowchart of a method for evaluating the remaining life of a power battery of an electric vehicle provided in Embodiment 1 of the present invention;

[0065] Figure 2 This is a schematic diagram of the structure of a device for assessing the remaining life of a power battery of an electric vehicle provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] The purpose of the present invention is to provide a method, device and system for evaluating the remaining life of a power battery of an electric vehicle, so as to improve the accuracy and universality of the life evaluation model, make full use of the data generated by the charging behavior of the electric vehicle, and reduce the comprehensive cost of the remaining life evaluation of the power battery.

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] Embodiment 1

[0070] The present invention provides a method for evaluating the remaining life of a power battery of an electric vehicle. Figure 1 As shown, the method includes:

[0071] Step S1, constructing a life decay data set of a type of power battery; the life decay data set includes a health state, a life cycle number and a discharge depth that can reflect the life decay trajectory of the power battery.

[0072] Exemplarily, the steps for constructing a lifespan decline dataset are:

[0073] ① Accelerated aging experiment is conducted on multiple groups of power batteries of the same batch and model. The experimental process is set as follows: all batteries are charged from 80% SOC to 100% SOC, and are charged to 3.6V through a uniform constant current-constant voltage charging step with a charging current of 1C. The charging current cutoff value is C / 50 (the symbol "C" is used to represent the ratio of the charging and discharging current of the power battery, that is, the rate, and its value is the ratio of the charging and discharging current of the power battery to the rated capacity). Subsequently, each group of power batteries is discharged in a constant current-constant voltage discharge step with a discharge current of 4C to a discharge voltage of 2.0V. The discharge current cutoff value is C / 50, and the power battery completes a life cycle. The remaining capacity Q and the number of life cycles of the power battery are recorded. The health state SOH of the power battery is defined as the ratio of the remaining capacity Q of the power battery to its rated capacity Q0. When the remaining capacity Q drops to 80% of its rated capacity Q0, it can be considered that the power battery has reached the end of its life decline. A life degradation test data set Γ(SOH, Cycles) of this type of power battery is obtained offline, and the data set includes the health state SOH and life cycle number Cycles of the power battery.

[0074] ② Directly read the relationship between the number of life cycles Cycles and the depth of discharge DOD of this type of power battery when it leaves the factory, and combine it with the life degradation test data set of this type of power battery obtained in ① to construct a life degradation test data set Ψ(SOH, Cycles, DOD) containing the life degradation trajectory of the power battery.

[0075] ③ The life degradation test data set Ψ(SOH, Cycles, DOD) obtained in ② containing the life degradation trajectory of the power battery is used to complete the missing data points using the difference fitting method.

[0076] Step S2, using Gaussian process-machine learning method to train the power battery life degradation law reflected by the life degradation data set to obtain a basic power battery life degradation assessment model; the power battery life degradation assessment model is a function of the health status of the power battery with respect to the number of life cycles and the depth of discharge.

[0077] Exemplarily, the process of obtaining the basic battery life degradation assessment model is as follows:

[0078] ① Reconstruct the data set Ψ(SOH, Cycles, DOD) into the form of SOH=ψ(Cycles, DOD).

[0079] ② Divide the SOH=ψ(Cycles,DOD) data set obtained in ① into a training set and a test set according to the principle of 11:1. During the data training process, let the training step length of the kth step be Step(k). Within the training step length Step(k), the training set data is repeatedly divided into a sub-training set and a sub-test set according to the principle of 11:1. The Gaussian process-machine learning method is used to train the power battery life degradation law reflected in the sub-training set, and the sub-test set is used to complete the test verification, correct the degradation law, and pass the result of the power battery life degradation law obtained under the current training step length Step(k) to the next adjacent training step length, that is, the k+1th training step length Step(k+1). This cycle is repeated to achieve the purpose of training and testing at the same time, and to build a basic power battery life degradation assessment model.

[0080] Step S3, based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the health status of the power battery by the state of charge, charging rate and temperature, so as to construct a type of comprehensive power battery life degradation assessment model.

[0081] Exemplarily, the construction process of a type of power battery life degradation comprehensive assessment model is as follows:

[0082] Based on the influence of the battery state of charge SOC, charge / discharge rate Crate and temperature T on the health state SOH expressed in the empirical formula (as shown in the following formula), based on the superposition law, combined with the constructed power battery life degradation basic evaluation model, a power battery life degradation comprehensive evaluation model SOH=ζ(Cycles,DOD,SOC,Crate,T) is formed, which is specifically shown as follows:

[0083]

[0084] In the formula, SOC represents the state of charge of the power battery, A i represents the polynomial coefficient, ∑ represents the summation symbol, f() represents the correction relationship between the power battery state of charge SOC and the power battery state of health SOH, n represents the value of the highest order term in the polynomial form of the influence relationship between the power battery state of charge SOC and its health state SOH, A i and n can be determined by experience, SOC i Represents the i-th term of the state of charge SOC.

[0085]

[0086] In the formula, Crate represents the charge / discharge rate of the power battery, g() represents the correction relationship between the charge / discharge rate Crate of the power battery and its health state SOH, and the coefficient a, b, c and d can be determined empirically.

[0087]

[0088] Where, T represents the power battery temperature, coefficient B j , R2, R1, R0 and s can be determined by experience, and p() represents the correction relationship between the power battery temperature T and its health state SOH.

[0089] Based on the superposition law, combined with the constructed basic evaluation model of power battery life degradation The comprehensive evaluation model for power battery life degradation is constructed as follows: Denoted as SOH=ζ(Cycles,DOD,SOC,Crate,T).

[0090] Step S4, based on one type of comprehensive evaluation model for power battery life degradation, a transfer learning method is used to construct multiple different types of power battery remaining life evaluation models.

[0091] Exemplarily, the process of constructing the remaining life assessment model of various types of power batteries is as follows:

[0092] ① Based on the process of constructing a comprehensive evaluation model for power battery life degradation, two other common comprehensive evaluation models for power battery life degradation in the current market are constructed. The comprehensive evaluation models for the life degradation of three different types of power batteries are respectively denoted as SOH=ζ i (Cycles,DOD,SOC,Crate,T),i=1,2,3.

[0093] ② Based on the comprehensive evaluation model of life decline of three different types of power batteries constructed in ① SOH=ζ i (Cycles, DOD, SOC, Crate, T), i = 1, 2, 3, the life decline comprehensive evaluation model of three different types of power batteries is divided into training set and test set according to the 2:1 principle, and the transfer learning method is used to construct the remaining life evaluation model SOH = ζ for typical types of power batteries in the current market of electric vehicles. i (Cycles, DOD, SOC, Crate, T), i = 1, 2, 3..., N, making it generalizable and universal. N represents the number of power battery types, which is a positive integer.

[0094] Step S5, when the electric vehicle starts charging, the current status data of the electric vehicle power battery is obtained from the charging pile; the status data includes the state of charge, charging rate and temperature.

[0095] Step S6, according to the current electric vehicle power battery type identified by the charging pile end, a power battery remaining life assessment model corresponding to the power battery type is obtained.

[0096] Step S7, based on the status data, using the power battery remaining life assessment model corresponding to the power battery type and the relationship between the discharge depth and the state of charge, determine the number of life cycles when the health state of the power battery reaches the end of life decline, as the remaining service life of the current electric vehicle power battery.

[0097] Exemplarily, the step of determining the number of life cycles specifically includes:

[0098] ① According to the state of charge in the status data, the discharge depth of the current electric vehicle power battery is determined by using the relationship between the discharge depth DOD and the state of charge SOC DOD = 1-SOC;

[0099] ② Input the status data and the current discharge depth of the electric vehicle power battery into the power battery remaining life assessment model corresponding to the power battery type to obtain the relationship between the life cycle number and health status of the current electric vehicle power battery;

[0100] ③ Set the time when the health status drops to 80% as the end point of life decline, and use the relationship between the life cycle number and health status of the current electric vehicle power battery to calculate the life cycle number of the current electric vehicle power battery.

[0101] The present invention predicts the remaining life of a power battery based on a method that combines Gaussian process machine learning and transfer learning. The constructed comprehensive life model of power batteries with multiple factors has high precision and good generalization ability, that is, for different types of power batteries, the remaining life model still has good universality; based on the power battery charging data, the remaining life of the power battery is predicted directly from the charging pile end, rather than based on the power battery discharge end data measured by multiple groups of vehicle-mounted sensors, which greatly improves the convenience of data acquisition and reduces the cost of data acquisition.

[0102] Embodiment 2

[0103] The present invention also provides a device for evaluating the remaining life of a power battery of an electric vehicle, such as Figure 2 As shown, the device includes: a charging pile end and a cloud end.

[0104] The cloud is connected to the charging pile end, and the cloud is used to store various different types of power battery remaining life assessment models. The charging pile end is used to obtain the state data of the electric vehicle during charging, and at the same time identify the power battery type of the charging electric vehicle, obtain the power battery remaining life assessment model corresponding to the power battery type of the charging electric vehicle from the cloud, and use the aforementioned electric vehicle power battery remaining life assessment method to calculate the remaining service life of the power battery of the charging electric vehicle, so as to transmit the remaining service life to the power battery management system of the charging electric vehicle.

[0105] Exemplarily, the device further includes: a first communication unit and a second communication unit. The first communication unit is used to complete the communication between the charging pile end and the charging electric vehicle. The second communication unit is used to complete the communication between the charging pile end and the cloud.

[0106] The remaining life assessment device for power batteries is a vehicle-pile network system, and the functions of each module are as follows:

[0107] Charging pile end: obtain the status data of the electric vehicle during charging, including but not limited to the charging rate Crate, state of charge SoC, temperature T, etc.;

[0108] Electric vehicle side: Receive the remaining life evaluation value of the power battery from the charging pile side, and then provide a reference for the battery management system to manage and control the power battery, so that the power battery works in the most reasonable state;

[0109] Cloud: Store the established database of typical types of power battery remaining life assessment and share it with the charging pile end;

[0110] Communication unit: completes the communication between electric vehicles and charging piles, and between charging piles and the cloud, etc. Including but not limited to Bluetooth and wireless LAN.

[0111] Based on the various status data of the electric vehicle during charging mapped by the charging station pile, combined with the constructed power battery remaining life assessment model, a comprehensive assessment can be made on the remaining life of the power battery of the electric vehicle, and the data can be fed back to the control unit of the corresponding charging pile in the charging station. The charging pile communicates the predicted value of the remaining life of the power battery with the electric vehicle battery management system through the CAN bus, which is used to guide the battery status management and control during the actual operation of the electric vehicle. The device directly assesses the remaining life of the power battery from the charging data end, rather than based on the power battery discharge end data measured by multiple sets of on-board sensors, which greatly improves the convenience of data acquisition, reduces the cost of data acquisition, and provides a reference for the status management and control of the power battery during the actual operation of the electric vehicle.

[0112] Embodiment 3

[0113] The present invention also provides a system for evaluating the remaining life of a power battery of an electric vehicle, the system comprising:

[0114] The life decay dataset construction module is used to construct a life decay dataset of a type of power battery; the life decay dataset contains the health status, life cycle number and discharge depth that can reflect the life decay trajectory of the power battery.

[0115] The basic degradation assessment model acquisition module is used to use the Gaussian process-machine learning method to train the power battery life degradation law reflected by the life degradation data set to obtain the power battery life basic degradation assessment model; the power battery life basic degradation assessment model is a function of the health status of the power battery with respect to the number of life cycles and the depth of discharge.

[0116] The module for constructing a comprehensive degradation assessment model is used to construct a type of comprehensive degradation assessment model for power battery life based on the superposition law, combining the basic degradation assessment model for power battery life with the correction models for the health status of the power battery according to the state of charge, charging rate and temperature.

[0117] The multi-type model building module is used to build multiple different types of power battery remaining life assessment models based on one type of power battery life degradation comprehensive assessment model using transfer learning methods.

[0118] The status data acquisition module is used to obtain the current status data of the electric vehicle power battery from the charging pile when the electric vehicle starts charging; the status data includes the state of charge, charging rate and temperature.

[0119] The matching module is used to obtain a power battery remaining life assessment model corresponding to the power battery type according to the current electric vehicle power battery type identified by the charging pile end.

[0120] The remaining service life calculation module is used to determine the number of life cycles when the health state of the power battery reaches the end of life decline based on the status data, using the power battery remaining life assessment model corresponding to the power battery type and the relationship between the discharge depth and the state of charge, as the remaining service life of the current electric vehicle power battery.

[0121] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0122] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for evaluating the remaining life of a power battery of an electric vehicle, It is characterized in that The method comprises: Constructing a life decay data set of a type of power battery; the life decay data set includes a health state, a life cycle number, and a discharge depth that can reflect the life decay trajectory of the power battery; The Gaussian process-machine learning method is used to train the power battery life degradation law reflected by the life degradation data set to obtain a power battery life basic degradation assessment model; the power battery life basic degradation assessment model is a function of the health status of the power battery with respect to the life cycle number and the discharge depth; Based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the power battery health status based on the state of charge, charging rate and temperature to construct a type of power battery life degradation comprehensive assessment model; Based on a type of comprehensive evaluation model for power battery life degradation, a transfer learning method is used to construct multiple different types of power battery remaining life evaluation models; When the electric vehicle starts charging, the current status data of the electric vehicle power battery is obtained from the charging pile; the status data includes the state of charge, charging rate and temperature; According to the current electric vehicle power battery type identified by the charging pile, a power battery remaining life assessment model corresponding to the power battery type is obtained; According to the status data, using the power battery remaining life assessment model corresponding to the power battery type and the relationship between the discharge depth and the state of charge, the number of life cycles when the health state of the power battery reaches the end of life decline is determined as the remaining service life of the current electric vehicle power battery.

2. The method for evaluating the remaining life of a power battery of an electric vehicle according to claim 1, It is characterized in that The constructing of a life degradation data set for a type of power battery specifically includes: Conduct accelerated aging experiments on multiple groups of power batteries of the same type; Record the remaining capacity and life cycle number of each group of power batteries after completing each life cycle; When the remaining capacity is less than or equal to the capacity threshold, it is determined that the power battery has reached the end of life decline and the accelerated aging test is stopped; Using the formula SOH = Q / Q 0 , determine the health status SOH of the power battery after each life cycle; where Q represents the remaining capacity of the power battery, Q 0 Indicates the rated capacity of the power battery; According to the relationship between the life cycle number and the discharge depth of the same type of power battery, determine the discharge depth corresponding to each recorded life cycle number; The health status and life cycle number of all groups of power batteries after each life cycle, as well as the discharge depth corresponding to each life cycle number, are recorded to form a life degradation data set Ψ (SOH, Cycles, DOD) of the same type of power battery; where Ψ represents the data set, Cycles represents the life cycle number, and DOD represents the discharge depth.

3. The method for evaluating the remaining life of a power battery of an electric vehicle according to claim 2, It is characterized in that The process of the accelerated aging experiment is as follows: Each group of power batteries is charged from a state of charge SOC of 80% to a state of charge SOC of 100%, and the charging voltage of each group of power batteries is made to 3.6V and the charging current is made to C / 50 by a constant current-constant voltage charging current of 1C; wherein C represents the rate; Each group of power batteries is discharged in a constant current-constant voltage discharge step from a discharge current of 4C to a discharge voltage of 2.0V, and the discharge current cutoff value is C / 50, completing one life cycle of the power battery; Repeat the above steps to complete multiple life cycles of the power battery.

4. The method for evaluating the remaining life of a power battery of an electric vehicle according to claim 2, It is characterized in that The Gaussian process-machine learning method is used to train the power battery life degradation law reflected by the life degradation data set to obtain the power battery life basic degradation assessment model, specifically including: The life decline data set Ψ(SOH, Cycles, DOD) is reconstructed as SOH=ψ(Cycles, DOD); wherein ψ represents a health state function; Divide the reconstructed life decay dataset into a training dataset and a test dataset; In each training step, the training data set is further divided into a sub-training data set and a sub-testing data set; The Gaussian process-machine learning method is used within the current training step to train the power battery life degradation law reflected by the training data set, and the test data set is used to complete the test verification, correct the degradation law, and obtain the power battery life degradation law under the current training step; The power battery life degradation law under the current training step is passed to the next training step, and when training is performed in the next training step, the step "using the Gaussian process-machine learning method to train the power battery life degradation law reflected by the training data set in the current training step, and using the test data set to complete the test verification, correct the degradation law, and obtain the power battery life degradation law under the current training step" is returned; After completing the training within all training steps, the trained power battery life degradation law is obtained; The test data set is used to verify the trained power battery life degradation law, correct the degradation law, obtain the power battery life degradation law after the test, and determine it as the power battery life basic degradation assessment model.

5. The method for evaluating the remaining life of a power battery of an electric vehicle according to claim 1, It is characterized in that Based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the health status of the power battery by the state of charge, charging rate and temperature, to construct a type of power battery life degradation comprehensive assessment model, which specifically includes: The correction model of the state of charge to the health state of the power battery is established as follows: Wherein, f() represents the correction relationship between the state of charge SOC and the health state SOH of the power battery, A i represents the polynomial coefficient of the i-th term, n represents the value of the highest order term in the polynomial form of the relationship between the state of charge SOC and the state of health SOH, SOC i represents the i-th term of the state of charge SOC; The correction model of charging rate to the health status of power battery is established as follows: Among them, g() represents the correction relationship between the charging rate Crate of the power battery and the health state SOH, a, b, c and d represent the first, second, third, fourth and fifth coefficients respectively; The correction model of temperature on the health status of power batteries is established as follows: Where p() represents the correction relationship between the temperature T of the power battery and the health state SOH, R 2 , R 1 , R 0 and s represent the sixth, seventh, eighth and ninth coefficients respectively, T j represents the jth term of temperature T, B j represents the j-th term T with respect to temperature T j The coefficient of j is in the range of 0≤j≤5. 5 , B 4 , B 3 , B 2 , B 1 , B 0 denote the tenth, eleventh, twelfth, thirteenth, fourteenth, and fifteenth coefficients respectively; Based on the superposition law, the basic power battery life degradation assessment model is combined with the correction models of the power battery's health status according to the state of charge, charging rate and temperature, and a type of power battery life degradation comprehensive assessment model is constructed: SOH = θ(Cycles, DOD)·f(SOC)·g(Crate)·p(T) = ζ(Cycles, DOD, SOC, Crate, T); wherein ζ() represents the power battery life degradation comprehensive assessment function, SOH = θ(Cycles, DOD) represents the power battery life degradation model, and θ() represents the power battery life degradation function.

6. The method for evaluating the remaining life of a power battery of an electric vehicle according to claim 1, It is characterized in that The method uses a transfer learning method to construct multiple different types of power battery remaining life assessment models based on a type of power battery life degradation comprehensive assessment model, specifically including: Based on the construction process of one type of comprehensive evaluation model for power battery life degradation, two other different types of comprehensive evaluation models for power battery life degradation are constructed; The three types of comprehensive evaluation models for power battery life degradation are divided into training sets and test sets; Based on the training set, a transfer learning method is used to construct a comprehensive evaluation model for life degradation of power batteries of types other than the three types. The test set is used to verify and adjust the comprehensive evaluation model for life degradation of power batteries of types other than the three types, so as to obtain multiple different types of power battery remaining life evaluation models.

7. The method for evaluating the remaining life of a power battery of an electric vehicle according to claim 1, It is characterized in that According to the state data, the remaining life assessment model of the power battery corresponding to the power battery type and the relationship between the discharge depth and the state of charge are used to determine the life cycle number when the health state of the power battery reaches the end of the life decline, specifically including: According to the state of charge in the state data, the discharge depth of the current electric vehicle power battery is determined by using the relationship between the discharge depth DOD and the state of charge SOC DOD=1-SOC; Inputting the state data and the current discharge depth of the electric vehicle power battery into a power battery remaining life assessment model corresponding to the power battery type, and obtaining a relationship between the life cycle number and the health state of the current electric vehicle power battery; The health status is set to drop to 80% as the end point of life decline, and the life cycle number of the current electric vehicle power battery is calculated using the relationship between the life cycle number and the health status.

8. A device for evaluating the remaining life of a power battery of an electric vehicle, It is characterized in that The device comprises: a charging pile end and a cloud end; The cloud is connected to the charging pile end, and the cloud is used to store multiple different types of power battery remaining life assessment models; The charging pile end is used to obtain the status data of the electric vehicle when it is charging, and at the same time identify the type of power battery of the charged electric vehicle, obtain the power battery remaining life assessment model corresponding to the power battery type of the charged electric vehicle from the cloud, and use the power battery remaining life assessment method of the electric vehicle described in any one of claims 1-7 to calculate the remaining service life of the power battery of the charged electric vehicle, thereby transmitting the remaining service life to the power battery management system of the charged electric vehicle.

9. The remaining life evaluation device for a power battery of an electric vehicle according to claim 8, It is characterized in that The device further includes: a first communication unit and a second communication unit; The first communication unit is used to complete the communication between the charging pile end and the charging electric vehicle; The second communication unit is used to complete the communication between the charging pile and the cloud.

10. A system for evaluating the remaining life of a power battery of an electric vehicle. It is characterized in that The system comprises: A life decay data set construction module is used to construct a life decay data set of a type of power battery; the life decay data set includes a health state, a life cycle number and a discharge depth that can reflect the life decay trajectory of the power battery; A basic degradation assessment model acquisition module is used to use a Gaussian process-machine learning method to train the power battery life degradation law reflected by the life degradation data set to obtain a power battery life basic degradation assessment model; the power battery life basic degradation assessment model is a function of the health status of the power battery with respect to the life cycle number and the discharge depth; A module for constructing a comprehensive evaluation model for the decline of power battery life, which is used to construct a comprehensive evaluation model for the decline of power battery life based on the superposition law by combining the basic evaluation model for the decline of power battery life with the correction models for the health status of power battery by state of charge, charging rate and temperature; A multi-type model building module is used to build multiple different types of power battery remaining life assessment models based on one type of power battery life degradation comprehensive assessment model using a transfer learning method; The status data acquisition module is used to obtain the current status data of the electric vehicle power battery from the charging pile when the electric vehicle starts charging; the status data includes the state of charge, charging rate and temperature; A matching module is used to obtain a power battery remaining life assessment model corresponding to the power battery type according to the current power battery type of the electric vehicle identified by the charging pile end; The remaining service life calculation module is used to determine the number of life cycles when the health state of the power battery reaches the end of life decline based on the status data, using the power battery remaining service life assessment model corresponding to the power battery type and the relationship between the discharge depth and the state of charge, as the remaining service life of the current electric vehicle power battery.

Citation Information

Patent Citations

  • Battery health state estimation and residual life prediction method based on real vehicle data

    CN112540317A

  • Electric vehicle charging detection system and method

    CN113954682A

  • Charging control method and system of new energy vehicle, vehicle-mounted terminal and medium

    CN113997805A

  • Battery life predicting device and battery life predicting method

    US20090037145A1

  • Online determination method and system for state of health of power battery pack of electric vehicle

    WO2021185308A1