A method for predicting lithium-ion battery capacity based on internal mechanism analysis
By collecting external characteristic parameters and internal mechanism changes data of lithium-ion batteries, the capacity decay process is divided into multiple stages, and a secondary prediction model is established, which solves the problem of low accuracy of lithium battery capacity prediction in low temperature environments in the prior art, and achieves fast and accurate capacity prediction.
Patent Information
- Application Number
- CN202210069751.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-21
- Publication Date
- 2025-04-29
- Estimated Expiration
- 2042-01-21
AI Technical Summary
The existing lithium battery capacity prediction methods have low accuracy in low temperature environments and large calculations, making it difficult to quickly and accurately predict lithium battery capacity.
By collecting external characteristic parameters and internal mechanism changes data during the discharge process of lithium-ion batteries, the capacity decay process is divided into multiple stages, a data set based on internal mechanism analysis and a primary capacity prediction model are established, and a secondary prediction model is then built to achieve accurate prediction.
It realizes fast and accurate prediction of lithium-ion battery capacity in low temperature environments, with fast calculation speed and small errors.
Smart Images
Figure CN114415034B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence, and in particular to a lithium-ion battery capacity prediction method based on internal mechanism analysis. Background Art
[0002] As environmental pollution becomes more and more serious, more and more automobile manufacturers are beginning to develop new energy vehicles. It has become a future development trend for electric vehicles to replace fuel vehicles. Lithium-ion batteries have the advantages of long cycle life, strong driving ability, high specific energy, short charging time, and high green environmental performance. They have become the main power components of electric vehicles. Studies have shown that the operating temperature of the battery has an important influence on the service life. The performance of the battery declines significantly in a low temperature environment, which will seriously affect the practicality of the battery in cold areas.
[0003] The existing lithium battery capacity prediction methods include equivalent model methods and artificial intelligence methods. The equivalent model method requires the establishment of a complex battery equivalent circuit model based on chemical reactions, and the accuracy of the predicted data depends on the accuracy of the model. It is very difficult to make accurate predictions in practice. Moreover, different batteries require the establishment of different equivalent circuit models, and the scope of application is narrow. Artificial intelligence methods are the current mainstream methods, including neural networks, support vector machines and other methods. On the one hand, these methods require a large amount of data, a large amount of calculation, and rely on the initial value of the data. On the other hand, most methods that use artificial intelligence algorithms to predict the capacity of lithium-ion batteries select the battery voltage or current as the input characteristic parameter. Such methods meet the requirements of rapid prediction in actual situations, but the accuracy is poor. Therefore, how to quickly and accurately predict the capacity of lithium batteries has become a problem that needs to be solved urgently by technicians in this field. Summary of the invention
[0004] The technical problem to be solved by the present invention is to overcome the deficiencies in the prior art and provide a method for predicting the capacity of a lithium battery at low temperature based on internal mechanism analysis.
[0005] The present invention is implemented by the following technical scheme: a lithium ion battery capacity prediction method based on internal mechanism analysis, comprising the following steps:
[0006] S1: Collect external characteristic parameters of lithium-ion batteries during discharge; disassemble lithium-ion batteries at different cycles during discharge, and simultaneously collect internal mechanism change data of lithium-ion batteries;
[0007] S2: According to the capacity change rate of lithium-ion batteries during discharge and the changes in the internal reaction mechanism of the battery, the capacity decay process of lithium-ion batteries is divided into multiple stages;
[0008] S3: Based on the external characteristic parameters and internal mechanism change data obtained in step S1, a data set of different external characteristic parameters and internal mechanism change data is established, and a first-level capacity prediction model is established based on the data set and the multiple stages divided in step S2;
[0009] S4: collecting data of multiple groups of lithium-ion batteries by repeating step S1 to obtain a training set and a test set containing external characteristic parameters and internal mechanism change data;
[0010] S5: importing the training set of step S4 into the primary capacity prediction model for training, and then establishing a secondary prediction model for accurately estimating the capacity of the lithium-ion battery;
[0011] S6: According to the secondary prediction model established in step S5, the test set is imported to obtain the predicted value, and the actual value is compared to perform accuracy analysis on the predicted value.
[0012] According to the above technical solution, preferably, in step S1, during the discharge process, the external characteristic parameters collected in real time include voltage (u), current (i), temperature (T), internal resistance (R) and battery capacity; the internal mechanism change data include the porosity and thickness of the battery cathode electrode.
[0013] According to the above technical solution, preferably, in step S2, the capacity decay process of the lithium-ion battery at low temperature is divided into three stages, including: 0-30 cycles: early stage of capacity decay of the lithium-ion battery; 31-70 cycles: middle stage of capacity decay of the lithium-ion battery; 71-100 cycles: late stage of capacity decay of the lithium-ion battery.
[0014] According to the above technical solution, preferably, in step S4, the training set is a data set containing external characteristic parameters and internal mechanism change data, and the test set is an average value.
[0015] According to the above technical solution, preferably, the step of establishing the secondary prediction model includes the following sub-steps:
[0016] (1) putting the multiple sets of data sets of the training set in step S4 into parallel primary capacity prediction models for training, thereby obtaining multiple different parallel secondary prediction models;
[0017] (2) Put the test sets with different data set inputs into the parallel secondary prediction model for testing to obtain the average capacity output values of different lithium-ion batteries;
[0018] (3) According to the three stages of step S2, an error analysis is performed between the average capacity output values of the lithium-ion battery under different decay modes and the true values of the test set.
[0019] According to the above technical solution, preferably, in step S5, the secondary prediction model includes a three-layer structure, including an input layer, an intermediate layer, and an output layer in sequence; wherein:
[0020] The input layer imports the dataset combination of the training set in step S4;
[0021] The middle layer is the corresponding mathematical relationship connecting the input layer and the output layer in the secondary prediction model;
[0022] The output layer is the predicted capacity of the lithium-ion battery.
[0023] The beneficial effects of the present invention are as follows: the present invention takes into account the practical problems of difficulty and low accuracy in predicting the capacity of lithium-ion batteries during operation of lithium-ion batteries in actual situations, and proposes a lithium-ion battery capacity prediction method based on internal mechanism analysis, which analyzes the influence of different input parameter combinations on the capacity prediction of lithium-ion batteries under different cycle periods, and uses a secondary prediction model to realize the estimation of the state of charge of the lithium-ion battery. The present invention has fast calculation speed and small error for the capacity prediction result of the lithium-ion battery. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 A schematic diagram of a process according to an embodiment of the present invention is shown;
[0025] Figure 2 A schematic diagram of a process of performing prediction at different stages using different data sets according to an embodiment of the present invention is shown;
[0026] Figure 3 A schematic diagram of the algorithm structure of a first-level capacity prediction model according to an embodiment of the present invention is shown;
[0027] Figure 4 A comparison of the porosity of the battery pole pieces of lithium-ion batteries at the 0-30 cycle stage is shown;
[0028] Figure 5 A comparison of the porosity of the battery pole pieces of lithium-ion batteries at the 31-70 cycle stage is shown;
[0029] Figure 6 A comparison of the porosity of the battery pole pieces of lithium-ion batteries at the 71-100 cycle stage is shown;
[0030] Figure 7 A three-dimensional photograph of the thickness of a battery pole piece of a lithium-ion battery is shown;
[0031] Figure 8 The three-dimensional pole piece thickness data diagram of the lithium-ion battery in the 0-30 cycle stage is shown;
[0032] Fig. 9The three-dimensional pole piece thickness data diagram of the lithium-ion battery in the 31-70 cycle stage is shown;
[0033] Fig.10 The three-dimensional pole piece thickness data diagram of the lithium-ion battery in the 71-100 cycle stage is shown;
[0034] Fig.11 A prediction result diagram according to an embodiment of the present invention is shown, in which the smooth curve is a prediction diagram; DETAILED DESCRIPTION
[0035] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and the best embodiments.
[0036] As shown in the figure, the present invention proposes a lithium-ion battery capacity prediction method based on internal mechanism analysis. The method divides the capacity decay process of the lithium-ion battery into three stages in combination with the internal mechanism changes during the capacity decay process of the lithium-ion battery. At the same time, the method considers the combination of multiple external parameters such as voltage, current, temperature, and internal resistance, and adopts different parameters in different stages, so as to better solve the problem of battery capacity prediction and evaluation.
[0037] A lithium-ion battery capacity prediction method based on internal mechanism analysis comprises the following steps:
[0038] S1: Collect external characteristic parameters of lithium-ion batteries during discharge; disassemble lithium-ion batteries at different cycles during discharge, and simultaneously collect internal mechanism change data of lithium-ion batteries;
[0039] S2: According to the capacity change rate of the lithium-ion battery during the discharge process and the changes in the internal reaction mechanism of the battery, the capacity decay process of the lithium-ion battery is divided into multiple stages. This embodiment is divided into three stages;
[0040] S3: Based on the external characteristic parameters and internal mechanism change data obtained in step S1, a data set of different external characteristic parameters and internal mechanism change data is established, and a primary capacity prediction model is established based on the data set and the three stages divided in step S2;
[0041] S4: collecting data of multiple groups of lithium-ion batteries by repeating step S1 to obtain a training set and a test set containing external characteristic parameters and internal mechanism change data;
[0042] S5: importing the training set of step S4 into the primary capacity prediction model for training, and then establishing a secondary prediction model for accurately estimating the capacity of the lithium-ion battery;
[0043] S6: According to the secondary prediction model established in step S5, the test set is imported to obtain the predicted value, compared with the actual value, and then the accuracy analysis of the predicted value is performed to determine the optimal solution and obtain the prediction result.
[0044] According to the above embodiment, preferably, in step S1, during the discharge process, the external characteristic parameters of the battery are collected in real time, including voltage (u), current (i), temperature (T), internal resistance (R) and battery capacity; the internal mechanism change data include the porosity and thickness of the battery cathode electrode.
[0045] According to the above embodiment, preferably, in step S2, the capacity decay process of the lithium-ion battery at low temperature is divided into three stages, including:
[0046] 0-30 cycles: early stage of lithium-ion battery capacity decay; 31-70 cycles: middle stage of lithium-ion battery capacity decay; 71-100 cycles: late stage of lithium-ion battery capacity decay.
[0047] According to the above embodiment, preferably, in step S3, the data set includes multiple groups of sample sets, which are set to five groups in this embodiment, including:
[0048] The first group, voltage, current, temperature;
[0049] The second group, voltage, current, internal resistance;
[0050] The third group, voltage, temperature, internal resistance;
[0051] Group 4: current, temperature, and internal resistance;
[0052] Group 5: voltage, current, temperature, and internal resistance.
[0053] And the process of building a first-level capacity forecasting model includes the following sub-steps:
[0054] (1) Determine the lithium-ion battery samples by random sampling;
[0055] (2) Construct a decision tree;
[0056] (3) Formation of the first-level capacity prediction model and execution of the algorithm;
[0057] The algorithm of the first-level capacity prediction model is as follows: Figure 8 As shown, including the following:
[0058] Each cell of the first-level capacity prediction model uses a gate structure to add or delete information. The gate structure selectively allows information to pass through by multiplying the activation function sigmoid with a stationary point. The output range of the activation function sigmoid is (0, 1), which represents the weight of the information to be passed. The first-level capacity prediction model uses the forget gate, input gate and output gate designed in this way to realize information screening and protection.
[0059] The forget gate is responsible for deciding what information the cells in the first-level capacity prediction model discard by reading h t-1 and x t Output a value f between 0 and 1 t Give each cell C t-1 , and f t The expression is:
[0060] f t =σ(W f [h t-1 ,x t ]+b f )
[0061] Where: h t-1 is the output state of the previous cell; x t is the current cell input state (voltage, current, temperature, internal resistance); W f is the weight matrix of the forget gate; b f is the bias value of the forget gate; σ is the sigmoid activation function.
[0062] The input gate is responsible for deciding how much new information to allow into the cells of the first-level capacity prediction model. First, the i output by the σ layer t Determine the information that needs to be updated, the vector generated by the tanh layer Content that is optional for updating;
[0063] i t =σ(W i [h t-1 ,x t ]+b i )
[0064]
[0065] Where: h t-1 is the output state of the previous cell; x t is the input state of the current cell; W i is the weight matrix of the input gate; b i is the bias value of the input gate; W f is the weight matrix of the forget gate; b fis the bias value of the forget gate; σ is the sigmoid activation function; W C is the weight matrix of the cell state; b C is the bias value of the cell state;
[0066] After the previous steps, the cell state can be updated:
[0067]
[0068] Where: C t is the cell state;
[0069] The output gate is responsible for determining the final output value. First, the output value o is passed through a σ layer. t To determine the part of the cell state output, then, through the tanh layer processing, get a value between -1 and 1, and multiply it with the output of the σ layer to determine the output h t .
[0070] o t =σ(W o [h t-1 ,x t ]+b o )
[0071] h t =o t tanh(C t )
[0072] Where h t-1 is the output state of the previous cell; x t is the input state of the current cell; W f is the weight matrix of the output gate; b o is the bias value of the output gate; C t-1 The cell state of the current cell.
[0073] According to the above embodiment, preferably, in step S4, the training set is a data set including external characteristic parameters and internal mechanism variation data, and the test set is an average value;
[0074] According to the above embodiment, preferably, the step of establishing the secondary prediction model includes the following sub-steps:
[0075] (1) putting the multiple sets of data sets of the training set in step S4 into parallel primary capacity prediction models for training, thereby obtaining multiple different parallel secondary prediction models;
[0076] (2) Put the test sets with different data set inputs into the parallel secondary prediction model for testing to obtain the average capacity output values of different lithium-ion batteries;
[0077] (3) According to the three stages of step S2, an error analysis is performed between the average capacity output values of the lithium-ion battery under different decay modes and the true values of the test set.
[0078] According to the above embodiment, preferably, in step S5, the secondary prediction model includes a three-layer structure, including an input layer, an intermediate layer, and an output layer in sequence; wherein:
[0079] The input layer imports the dataset combination of the training set in step S4;
[0080] The middle layer is the corresponding mathematical relationship connecting the input layer and the output layer in the secondary prediction model;
[0081] The output layer is the predicted capacity of the lithium-ion battery.
[0082] According to the above embodiment, preferably, in step S6, the accuracy analysis of the predicted capacity is an error analysis:
[0083]
[0084] Beneficial effect: The present invention takes into account the practical problems of difficulty and low accuracy in capacity prediction of lithium-ion batteries during operation of lithium-ion batteries in actual situations, and proposes a lithium-ion battery capacity prediction method based on internal mechanism analysis, which analyzes the influence of different input parameter combinations on the capacity prediction of lithium-ion batteries under different cycle periods, and uses a secondary prediction model to realize the estimation of the state of charge of the lithium-ion battery. The present invention has fast calculation speed and small error for the capacity prediction result of the lithium-ion battery.
[0085] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A lithium-ion battery capacity prediction method based on internal mechanism analysis, characterized in that: The following steps are involved: S1: Collect external characteristic parameters of lithium-ion batteries during discharge; disassemble lithium-ion batteries at different cycles during discharge, and simultaneously collect internal mechanism change data of lithium-ion batteries; S2: According to the capacity change rate of lithium-ion batteries during discharge and the changes in the internal reaction mechanism of the battery, the capacity decay process of lithium-ion batteries is divided into multiple stages; S3: Based on the external characteristic parameters and internal mechanism change data obtained in step S1, a data set of different external characteristic parameters and internal mechanism change data is established, and a first-level capacity prediction model is established based on the data set and the multiple stages divided in step S2; S4: collecting data of multiple groups of lithium-ion batteries by repeating step S1 to obtain a training set and a test set containing external characteristic parameters and internal mechanism change data; S5: importing the training set of step S4 into the primary capacity prediction model for training, and then establishing a secondary prediction model for accurately estimating the capacity of the lithium-ion battery; S6: According to the secondary prediction model established in step S5, the test set is imported to obtain the predicted value, and the actual value is compared to perform accuracy analysis on the predicted value.
2. The method for predicting lithium-ion battery capacity based on internal mechanism analysis according to claim 1, characterized in that: In step S1, during the discharge process, the external characteristic parameters collected in real time include voltage (u), current (i), temperature (T), internal resistance (R) and battery capacity; the internal mechanism change data include the porosity and thickness of the battery cathode electrode.
3. The method for predicting lithium-ion battery capacity based on internal mechanism analysis according to claim 1, characterized in that: In step S2, the capacity decay process of the lithium-ion battery at low temperature is divided into three stages, including: 0-30 cycles: the initial stage of capacity decay of the lithium-ion battery; 31-70 cycles: the middle stage of capacity decay of the lithium-ion battery; 71-100 cycles: the late stage of capacity decay of the lithium-ion battery.
4. The method for predicting lithium-ion battery capacity based on internal mechanism analysis according to claim 1, characterized in that: In step S4, the training set is a data set containing external characteristic parameters and internal mechanism variation data, while the test set is the average value.
5. The method for predicting lithium-ion battery capacity based on internal mechanism analysis according to claim 4, characterized in that: The steps for establishing the secondary prediction model include the following sub-steps: (1) putting the multiple sets of data sets of the training set in step S4 into parallel primary capacity prediction models for training, thereby obtaining multiple different parallel secondary prediction models; (2) Put the test sets with different data set inputs into the parallel secondary prediction model for testing to obtain the average capacity output values of different lithium-ion batteries; (3) According to the three stages of step S2, an error analysis is performed between the average capacity output values of the lithium-ion battery under different decay modes and the true values of the test set.
6. The method for predicting lithium-ion battery capacity based on internal mechanism analysis according to claim 5, characterized in that: In step S5, the secondary prediction model includes a three-layer structure, including an input layer, an intermediate layer, and an output layer in sequence; wherein: The input layer imports the dataset of the training set in step S4; The middle layer is the corresponding mathematical relationship connecting the input layer and the output layer in the secondary prediction model; The output layer is the predicted capacity of the lithium-ion battery.