Quality testing method of ternary positive electrode material, medium and terminal equipment

Abstract

The invention relates to a quality testing method of a ternary positive electrode material, a medium and terminal equipment, and the method comprises the steps: firstly, knowing the initial state of the material, and laying a foundation for subsequent changes; secondly, charging and discharging data and a structure change speed of a cycle period are collected, and a subsequent change direction and a process speed are established; the method comprises the following steps of: firstly, predicting charge-discharge data of a next cycle period, and finally, continuously carrying out iterative prediction on new prediction data until charge-discharge data after a final cycle number is obtained, so that the capacity retention ratio after a final preset period can be determined, and whether the quality of the ternary positive electrode material reaches the standard or not is judged according to the capacity retention ratio. On one hand, various properties of the ternary positive electrode material are comprehensively and systematically analyzed, and comprehensiveness is achieved; on the other hand, the experiment frequency can be effectively reduced, the labor intensity of workers and the time required for quality detection are reduced, the quality detection efficiency is improved, and the service life and customer satisfaction after delivery are improved.

Description

Technical Field

[0001] The present invention relates to the field of quality testing, and in particular to a quality testing method, medium and terminal equipment for a ternary positive electrode material. Background Art

[0002] Positive electrode materials, negative electrode materials, electrolytes and separators are the four main materials of power batteries, among which positive electrode materials account for 40% of the total cost and directly determine the energy density of the battery cell. Ternary positive electrode materials have gradually become the mainstream positive electrode materials in the market due to their higher energy density and lower cost. The gradual popularization of pure electric vehicles will lead to a strong market demand for high energy density ternary positive electrode materials. LiNi 0.5 Co 0.5 Mn 0.3 As a typical representative of ternary positive electrode materials for lithium-ion batteries, O2 has become a strong competitor for positive electrode materials for lithium-ion batteries due to its good safety, low cost, and high capacity. 0.5 Co 0.5 Mn 0.3 The application of O2 ternary materials in power batteries, energy storage and small high-energy-density lithium-ion batteries for electronic products requires quality testing and evaluation before leaving the factory. One of the most important performance indicators is its cycle performance, which refers to the capacity retention rate of the battery after a certain number of cycles.

[0003] In the prior art, it is usually necessary to conduct many cycles of testing, usually more than 200 times. Not only is the testing process complicated, time-consuming and costly, but the testing process itself consumes the product, greatly compromising its factory quality. How to improve its quality testing efficiency is a major difficulty in this field.

[0004] The Chinese patent with announcement number CN115338149B first conducts a small number of charge and discharge cycle tests on the battery, and then analyzes the X-ray powder diffraction spectra of the ternary layered positive electrode material before and after the cycle to obtain the change rate of lithium-nickel mixing and the thickness change of the rock salt phase before and after the cycle, and quickly judges the cycle life of the lithium battery ternary layered positive electrode material. To a certain extent, the speed of cycle life judgment has been improved. However, the structure of the ternary positive electrode material is complex, and its cycle life and quality are not determined solely by structural changes such as the lithium-nickel mixing change rate and the rock salt phase thickness change. Prediction and judgment based on only these two indicators is rather one-sided, which may cause certain false detections and missed detections.

[0005] Therefore, how to improve its quality testing methods and take into account both testing efficiency and testing accuracy is a technical problem that needs to be urgently solved in this field. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a quality testing method for a ternary positive electrode material, which comprises: S1: Obtaining initial data of power-on and initial data of structure; initial data of power-on includes: any one or more of first efficiency, charge-discharge capacity at different powers, and first charge-discharge capacity; initial data of structure includes: any one or more of specific surface area, tap density, compaction density, and content of each component; S2: Obtain the charge and discharge curves of N cycles; S3: Divide the N cycles into M stages, and obtain the structural change data after each stage cycle; N and M are integers greater than 1, and N is greater than M; the structural change data include: any one or more of particle size, morphology, distribution, surface roughness, lithium-nickel mixing data, layered structure damage, and crack conditions; S4: predicting the charge and discharge curve of the next cycle according to the initial data of power-on, the initial data of structure, the charge and discharge curve of N cycles and the structural change data after each stage cycle; S5: Determine the number of cycle cycles, iterate step S4, predict the charge and discharge curve for the next period of time, determine the capacity retention rate after the set number of cycle cycles, and obtain the quality test result.

[0007] Furthermore, step S4 includes: constructing and training a prediction model with the initial data of power-off, the initial data of the structure, the charge and discharge curves of N cycles and the structural change data after each stage cycle as input, and the charge and discharge curve after the next cycle as output, extracting relevant features, constructing the intricate connection between the input and output, and obtaining the prediction result.

[0008] Furthermore, the prediction model constructed includes: The input module includes two channels. The first channel is used to input the initial data of the power-on, the initial data of the structure, and the data of the structural change after each stage cycle. The second channel is used to input the charge and discharge curve of N cycles. The feature extraction module includes: a first feature extraction unit connected to the first channel, used to analyze the three-dimensional matrix information and extract data features according to the initial data of power-on, the initial data of the structure and the data of the structural change after the cycle; a second feature extraction unit connected to the second channel, used to analyze the capacity retention rate, the charging medium voltage and the discharging medium voltage according to the charge and discharge curve of N cycles, and extract the curve features; The prediction module is connected to the feature extraction module and is used to predict the charge and discharge curve of the next cycle according to the data features and curve features.

[0009] Furthermore, the first feature extraction unit includes multiple convolutional layers and pooling layers; The convolution layer is used to implement forward propagation and extract multi-scale features of the three-dimensional matrix. The number of input channels is 3, the number of output channels is 64, the convolution kernel size is 3, the step size is 1, the padding is 1, the expansion is 1, the number of groups is 1, and an offset is added; the pooling layer is used to return multi-scale features.

[0010] Furthermore, the prediction module includes two fully connected layers and an activation function located between the two fully connected layers; the first fully connected layer is used to fuse data features and curve features and merge features in the feature dimension; and after the activation function, the prediction result is output by the second fully connected layer.

[0011] Furthermore, the steps of training the prediction model include: Obtain the initial data of the charge-discharge of several ternary cathode materials; initial data of the structure; charge-discharge curves of 50 cycles; and structural change data after 5, 12, and 20 cycles; Input the initial data of power-off, initial data of structure, charge-discharge curves of the first 20 cycles, and structural change data after 5, 12, and 20 cycles of each ternary cathode material into the prediction model, and output the charge-discharge curve of the next cycle; The charge and discharge data of the last 30 cycles were used as a validation data set, and regression verification was performed on the charge and discharge curves predicted by the model to modify various parameters of the prediction model and obtain a trained prediction model.

[0012] Furthermore, when the N cycles are divided into M stages, the division is performed in a manner of increasing intervals.

[0013] Furthermore, the initial data of power-on deduction includes: the first efficiency and the charge and discharge capacity at different rates; Initial structural data, including Ni, Co, Mn, impurity content and compaction density; Structural change data, including: any one or more of particle size, morphology, distribution, surface roughness, lithium-nickel mixing data, whether the layered structure is destroyed, and whether microcracks appear.

[0014] On the other hand, the present invention further provides a computer storage medium storing executable program code; the executable program code is used to execute any of the above-mentioned quality testing methods.

[0015] On the other hand, the present invention further provides a terminal device, comprising a memory and a processor; the memory stores a program code executable by the processor; the program code is used to execute any of the above-mentioned quality testing methods.

[0016] The present invention provides a quality testing method, medium and terminal equipment for a ternary positive electrode material. First, the initial data of charge and structure are collected to understand the initial state of the material, so as to lay a foundation for subsequent changes; secondly, the charge and discharge data of the cycle period and the speed of structural change are collected to understand the change of the material, so as to lay a foundation for the direction and process speed of subsequent changes; then, according to the data obtained in the first step, the charge and discharge data of the next cycle period are predicted, and finally, the new predicted data is continuously iterated and predicted until a preset iteration period is reached, and the charge and discharge data after the final number of cycles is obtained, so as to determine the capacity retention rate after the final preset period, and judge whether the quality of the ternary positive electrode material meets the standard. Compared with the prior art, on the one hand, it comprehensively and systematically analyzes the various properties of the ternary cathode material by combining the initial data of power deduction, the initial data of structure, the charge-discharge curve of N cycles and the structural change data after each stage cycle, and has comprehensiveness; on the other hand, it predicts the charge-discharge curve of the next cycle through the previous data, that is, predicts its performance in the future cycle, so as to evaluate the quality of its material, which can effectively reduce the number of experiments, reduce the labor intensity and the time required for quality inspection, improve the efficiency of quality inspection, and at the same time reduce the performance consumption of the product in the quality inspection process, improve its service life after leaving the factory, and improve customer satisfaction. Therefore, the quality testing method of the ternary cathode material of the present invention can be compatible with the requirements of test efficiency and test accuracy, reduce the time and cost required for testing, and avoid phenomena such as missed detection and false detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flow chart of an embodiment of a method for testing the quality of a ternary cathode material of the present invention; Figure 2 It is a schematic diagram of the charge and discharge curve results of one cycle; Figure 3 Schematic diagram of the predicted charge and discharge curves for N cycles. DETAILED DESCRIPTION

[0018] 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.

[0019] It should be noted that if the embodiments of the present invention involve directional indications, such as up, down, left, right, front, back, etc., then the directional indication is only used to explain the relative position relationship, movement status, etc. between the components in a certain specific posture. If the specific posture changes, the directional indication will also change accordingly. In addition, if the embodiments of the present invention involve descriptions of "first, second", "S1, S2", "step one, step two", etc., then such descriptions are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of indicated technical features or indicating the execution order of the method, etc. Those skilled in the art can understand that anything that does not violate the gist of the invention under the technical concept of the invention should be included in the protection scope of the present invention.

[0020] like Figure 1 As shown, the present invention provides a quality testing method for a ternary positive electrode material, comprising: S1: Get the initial data of power-on and structure; S2: Obtain the charge and discharge curves of N cycles; S3: Divide the N cycles into M stages, and obtain the structural change data after each stage cycle; N and M are integers, and N is greater than M; S4: predicting the charge and discharge curve of the next cycle according to the initial data of power-on, the initial data of structure, the charge and discharge curve of N cycles and the structural change data after each stage cycle; S5: Determine the number of cycle cycles, iterate step S4, predict the charge and discharge curve for the next period of time, determine the capacity retention rate after the set number of cycle cycles, and obtain the quality test result.

[0021] In this embodiment, a quality testing method for the ternary positive electrode material of the present invention is provided. First, in step S1, initial data of charge and structure are collected to understand the initial state of the material to lay the foundation for subsequent changes; secondly, in steps S2-S3, charge and discharge data of the cycle period and the speed of structural change are collected to understand the change of the material to lay the foundation for the direction and process speed of subsequent changes; then in step S4, the charge and discharge data of the next cycle period are predicted based on the data obtained in the above steps; and finally, in step S5, the new predicted data is continuously iterated to step S4 until a preset iteration period is reached, and the charge and discharge data after the final number of cycles is obtained, so as to determine the capacity retention rate after the final preset period, and judge whether the quality of the ternary positive electrode material is up to standard. Compared with the prior art, on the one hand, it comprehensively analyzes the various properties of the ternary positive electrode material by combining the initial data of deduction, the initial data of structure, the charge-discharge curve of N cycles and the structural change data after each stage cycle, avoiding one-sidedness and having comprehensiveness; on the other hand, it predicts the charge-discharge curve of the next cycle through the previous data, that is, predicts its performance in the future cycle, thereby evaluating the quality of its material, which can effectively reduce the number of experiments. For example, in the prior art, if it is necessary to understand the capacity retention rate after 1000 cycles, it is necessary to perform 1000 cycles to understand its final charge-discharge curve, which is not only time-consuming but also cannot be done for each product, otherwise it will greatly reduce the service life of the product after leaving the factory. By adopting the method of the present invention, only the charge-discharge curve and structural change of the first dozens of times can be tested to make subsequent unlimited charge-discharge curve predictions to obtain its final capacity retention rate. Therefore, the present invention can reduce the labor intensity and the time required for quality inspection, improve the efficiency of quality inspection, and at the same time reduce the performance consumption of the product in the quality inspection process, improve its service life after leaving the factory, and improve customer satisfaction. The quality testing method of the ternary positive electrode material of the present invention can meet the requirements of testing efficiency and testing accuracy, reduce the time and cost required for testing, and avoid phenomena such as missed detection and false detection.

[0022] S1: Get the initial data of power-on and structure; Specifically, "button battery" refers to button-type full battery, which is a form of battery testing used in laboratories. It is mainly composed of positive electrode shell, negative electrode shell, electrode, diaphragm, gasket, shrapnel and electrolyte; it is used to evaluate the electrochemical properties of battery materials, such as capacity, cycle life, etc.

[0023] As shown in Table 1, the initial data after power-on may include but is not limited to: first efficiency, i.e., the ratio of the discharge capacity of the battery in the first charge and discharge cycle to the theoretical capacity; charge and discharge capacity at different rates, such as 0.2C; 0.5C; 1C; first charge and discharge capacity, etc.

[0024] Table 1: Schematic table of initial data of ternary positive electrode materials

[0025] Initial structural data, which can be obtained by X-ray diffraction (XRD), scanning electron microscopy (SEM) and other techniques to obtain initial structural information such as crystal structure and morphology of the material, can include but is not limited to: specific surface area, tap density, compaction density, content of each component, such as moisture, residual alkali, Li content, Ni / Co / Mn content, impurities, etc., any one or more, as shown in Tables 2 and 3.

[0026] Table 2: Schematic diagram of initial structural data of ternary cathode materials

[0027] Table 3: Schematic diagram of the initial structural data of ternary cathode materials

[0028] S2: Obtain the charge and discharge curves of N cycles; Specifically, it is possible to perform a charge-discharge cycle test at 25 degrees for a period of time, and record the charge-discharge curve after each cycle. For example, taking 20 charge-discharge cycles as an example, the charge-discharge curve is recorded for each cycle, that is, the horizontal axis is the specific capacity and the vertical axis is the voltage. Figure 2 shown.

[0029] S3: Divide the N cycles into M stages, and obtain the structural change data after each stage cycle; Specifically, taking the above-mentioned 20 cycles as an example, they can be divided into 4 stages on average. After 5, 10, 15 and 20 cycles respectively, high-resolution surface morphology images of the battery material are obtained by scanning electron microscopy, and the structural change data are observed and obtained, which may include but are not limited to: particle size, morphology, distribution, surface roughness, lithium-nickel mixing data, layered structure damage, the appearance of microcracks or cracks, etc. Examples: 1. Particle size, initial state: the average particle size is 5 μm, and the distribution is uniform; after 10 cycles: the particle size increases to 8 μm, and some particles agglomerate; more preferably, the agglomeration ratio can also be statistically analyzed; 2. Morphology, initial state: the particle surface is smooth and has a regular polyhedral shape; after cycling: corrosion, depressions or holes appear on the particle surface, and the morphology becomes irregular; 3. Distribution, initial state: the particles are evenly distributed in the electrode without obvious agglomeration; after cycling: the particle distribution is uneven, and local agglomeration or voids appear; 4. Surface roughness, initial state: the surface roughness (Ra) is 50 nm; after cycling: the surface roughness increases to 200 nm, and the surface becomes rough; 5. Lithium-nickel mixing data, initial state: the lithium-nickel mixing degree is low, and the proportion of nickel ions in the lithium layer is 2%; after cycling: the lithium-nickel mixing degree increases, and the proportion of nickel ions in the lithium layer increases to 10%; 6. Destruction of layered structure, initial state: the material has a complete layered structure, and the (003) peak is obvious in the XRD spectrum; after cycling: the layered structure is partially destroyed, the (003) peak intensity weakens, and a spinel phase or a rock salt phase appears; 7. The appearance of microcracks or cracks, initial state: the particles are intact and have no cracks; after cycling: microcracks appear inside and on the surface of the particles, and the crack width is 100 nm, etc.

[0030] More preferably, since the structural change characteristics of the ternary positive electrode material are more obvious in the early stage and the subsequent changes are smaller, it is possible to divide the N cycles into M stages in an increasing interval. For example, taking the above 20 cycles as an example, it is divided into 3 stages, and the structural change data are observed and obtained after 5, 12, and 20 cycles, that is, the structural change data are obtained at intervals of 5, 7, and 8 times, respectively.

[0031] It is worth noting that the number of cycles, specific intervals, etc., can be set arbitrarily by those skilled in the art. The present invention is subsequently described using historical cycle = 20 times and predicted cycle = 50 times as an example, but is not limited thereto. In the actual detection process, the charge and discharge curves after hundreds or even thousands of cycles can be predicted.

[0032] S4: predicting the charge and discharge curve of the next cycle according to the initial data of power-on, the initial data of structure, the charge and discharge curve of N cycles and the structural change data after each stage cycle; Specifically, step S4 may optionally, but not limited to, use machine learning, statistics, etc. to make predictions based on historical data. Preferably, a neural network model may be used to construct and train a prediction model with initial data of power-off, initial data of structure, charge-discharge curves of N cycles, and structural change data after each stage cycle as inputs, and the charge-discharge curve after the next cycle as output, extract relevant features, and construct an intricate connection between input and output to obtain a prediction result. More specifically, step S4 may optionally include two steps: model construction and training.

[0033] S41: Model construction, any model of neural network can be used.

[0034] Preferably, the prediction model comprises: A: Input module, including 2 channels; A1: The first channel is used to input the initial data of the charge, the initial data of the structure, and the data of the structural changes after each cycle. These data can be organized into a three-dimensional matrix, which contains the initial state of the material and the structural change data after the cycle.

[0035] A2: The second channel is used to input the charge and discharge curves of N cycles. These data can be organized into time series data to record the performance changes of the battery during the cycle process.

[0036] B: Feature extraction module, including: B1: A first feature extraction unit, connected to the first channel, for analyzing the three-dimensional matrix information and extracting data features according to the initial data of power-on, the initial data of the structure and the data of the structural change after the cycle.

[0037] Specifically, considering that the initial data of power-on, the initial data of structure, and the data of structural change after the cycle are all local features, the first feature extraction unit can optionally use a convolutional neural network (CNN) to extract features of the initial data and structural change data using its powerful local feature extraction capabilities. The example can be designed based on PyTorch, including multiple layers of convolutional layers and pooling layers; the example convolutional layer is used to implement forward propagation and extract multi-scale features of a three-dimensional matrix; taking a convolutional layer as an example, the optional input channel number is 3, the output channel number is 64, the convolution kernel size is 3, the step size is 1, the padding is 1, the expansion is 1, the number of groups is 1, and a bias is added; the pooling layer is used to return multi-scale features.

[0038] B2: a second feature extraction unit connected to the second channel, used to analyze the capacity retention rate, charging medium voltage, and discharging medium voltage according to the charge and discharge curve of N cycles, and extract curve features; Specifically, considering that the charge and discharge curve of N cycles is charge and discharge cycle data within a period of time, it has time series characteristics. The second feature extraction unit preferably uses a recurrent neural network RNN ​​and its variants, such as a long short-term memory network LSTM and a gated recurrent unit GRU. Preferably, the construction of the second feature extraction unit includes: First, the input data of the second channel needs to be preprocessed into a shape suitable for LSTM input: # Assume that input_data2 is the input data of the second channel, and its shape is (batch_size, sequence_length, input_size) # input_size is the number of features per time step features = FeatureExtractionUnit2(input_size=10, hidden_size=64) #Adjust input_size and hidden_size according to the actual number of features extracted_features = features(input_data2) Then, the second feature extraction unit is constructed based on LSTM: import torch import torch.nn as nn class FeatureExtractionUnit2(nn.Module): def __init__(self, input_size, hidden_size, num_layers=1): super(FeatureExtractionUnit2, self).__init__() self.lstm = nn.LSTM(input_size, hidden_size, num_layers,batch_first=True) def forward(self, x): # The shape of x is (batch_size, sequence_length, input_size) # Initialize hidden state and cell state h0=torch.zeros(self.lstm.num_layers,x.size(0), self.lstm.hidden_size).to(x.device) c0=torch.zeros(self.lstm.num_layers,x.size(0), self.lstm.hidden_size).to(x.device) # Forward propagation LSTM out, _ = self.lstm(x, (h0, c0)) # You can choose to return the output of the last time step or all outputs return out[:, -1, :] # Returns the output of the last time step C: prediction module, connected with the feature extraction module, used to predict the charge and discharge curve of the next cycle according to the data characteristics and curve characteristics; Specifically, the prediction module includes two fully connected layers and an activation function between the two fully connected layers; the first fully connected layer is used to fuse data features and curve features and merge features in the feature dimension; and after the activation function, the second fully connected layer outputs the prediction result. Specifically: import torch import torch.nn as nn class PredictionUnit(nn.Module): def __init__(self, feature_dim1, feature_dim2, output_dim): super(PredictionUnit, self).__init__() # Assume that feature_dim1 and feature_dim2 are the feature dimensions output by the two feature extraction modules respectively # output_dim is the dimension of the prediction result. The prediction is the charge and discharge curve, which is the number of points on the curve self.fc1 = nn.Linear(feature_dim1 + feature_dim2, 128) # The first fully connected layer self.relu = nn.ReLU() self.fc2 = nn.Linear(128, output_dim) # Final output layer def forward(self, x1, x2): # x1 and x2 are the outputs of the two feature extraction modules combined_features = torch.cat((x1, x2), dim=1) # Merge features on the feature dimension x = self.fc1(combined_features) # Pass through the first fully connected layer x = self.relu(x) # activation function x = self.fc2(x) # Pass the second fully connected layer and output the prediction result return x S42: Training phase: Obtain the initial data of the charge-discharge of several ternary cathode materials; initial data of the structure; charge-discharge curves of 50 cycles; and structural change data after 5, 12, and 20 cycles; Input the initial data of power-off, initial structural data, charge-discharge curves of the first 20 cycles, and structural change data after 5, 12, and 20 cycles of each ternary positive electrode material into the prediction model constructed in step S41, output the charge-discharge curve of the next cycle, use the charge-discharge data of the last 30 cycles as the verification data set, perform regression verification on the charge-discharge curve predicted by the model, and modify various parameters of the prediction model to obtain the trained prediction model. When the subsequent specific test is performed, the initial data of power-off, initial structural data, charge-discharge curves of N cycles, and structural change data after each stage of the cycle of the current ternary positive electrode material can be collected to predict and output the charge-discharge curve of the next cycle through the trained prediction model, such as Figure 3 , as shown in Table 4 and Table 5.

[0039] Table 4 Charge and discharge data of several cycles 1

[0040] Table 5 Charge and discharge data of several cycles 2

[0041] S5: Determine the number of cycles, iterate step S4, predict the charge and discharge curves for the next period of time, determine the capacity retention rate after the set number of cycles, and obtain the quality test results. Specifically, the discharge specific capacity is obtained through the predicted charge and discharge curves, and its ratio to the initial value is calculated to determine the capacity retention rate after the set number of cycles, as shown in Table 6.

[0042] Table 6: Prediction results of capacity retention rate

[0043] Specifically, the number of cycle cycles can be determined according to the actual quality requirements, such as the specific field to which the ternary positive electrode material is applicable. For general factory requirements, the number of cycle cycles can be set to 200 times; for high-standard industries, the number of cycle cycles can be determined to be hundreds or even thousands of times, and step S4 is iterated until the set cycle period is reached, and the final capacity retention rate is calculated. If the predicted capacity retention rate is higher than a certain threshold, such as 80%, it is considered that the material has good cycle stability and is of qualified quality. If the predicted capacity retention rate is lower than the threshold, it is considered that the material does not have good cycle stability and is of unqualified quality. It can be eliminated or the trial requirements can be lowered. The material can also be further modified or optimized according to the quality inspection pass rate of batch products to improve its cycle stability and electrochemical performance.

[0044] On the other hand, the present invention further provides a computer storage medium storing executable program code; the executable program code is used to execute any of the above-mentioned quality testing methods.

[0045] On the other hand, the present invention further provides a terminal device, comprising a memory and a processor; the memory stores a program code executable by the processor; the program code is used to execute any of the above-mentioned quality testing methods.

[0046] Exemplarily, the program code may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the program code in the terminal device.

[0047] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the terminal device may also include an input / output device, a network access device, a bus, etc.

[0048] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0049] The memory may be an internal storage unit of the terminal device, such as a hard disk or a memory. The memory may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal device. Further, the memory may also include both an internal storage unit of the terminal device and an external storage device. The memory is used to store the program code and other programs and data required by the terminal device. The memory may also be used to temporarily store data that has been output or is to be output.

[0050] The above-mentioned computer storage medium and terminal equipment are created based on the above-mentioned quality testing method, and their technical functions and beneficial effects are no longer repeated here. The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the various technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0051] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A quality testing method for a ternary cathode material, characterized in that: include: S1: Get the initial data of power-on and structure; Initial data after power-off, including: any one or more of the first efficiency, charge and discharge capacity at different powers, and first charge and discharge capacity; initial structural data, including: any one or more of the specific surface area, tap density, compaction density, and content of each component; S2: Obtain the charge and discharge curves of N cycles; S3: Divide the N cycles into M stages, and obtain the structural change data after each stage cycle; N and M are integers greater than 1, and N is greater than M; the structural change data include: any one or more of particle size, morphology, distribution, surface roughness, lithium-nickel mixing data, layered structure damage, and crack conditions; S4: predicting the charge and discharge curve of the next cycle according to the initial data of power-on, the initial data of structure, the charge and discharge curve of N cycles and the structural change data after each stage cycle; S5: Determine the number of cycle cycles, iterate step S4, predict the charge and discharge curve for the next period of time, determine the capacity retention rate after the set number of cycle cycles, and obtain the quality test result.

2. The quality testing method according to claim 1, characterized in that: Step S4 includes: constructing and training a prediction model with the initial data of power-off, the initial data of structure, the charge-discharge curves of N cycles and the structural change data after each stage cycle as input, and the charge-discharge curve after the next cycle as output, extracting relevant features, constructing a complex connection between input and output, and obtaining a prediction result.

3. The quality testing method according to claim 2, characterized in that: The prediction model constructed includes: The input module includes two channels. The first channel is used to input the initial data of the power-on, the initial data of the structure, and the data of the structural change after each stage cycle. The second channel is used to input the charge and discharge curve of N cycles. The feature extraction module includes: a first feature extraction unit connected to the first channel, used to analyze the three-dimensional matrix information and extract data features according to the initial data of power-on, the initial data of the structure and the data of the structural change after the cycle; a second feature extraction unit connected to the second channel, used to analyze the capacity retention rate, the charging medium voltage and the discharging medium voltage according to the charge and discharge curve of N cycles, and extract the curve features; The prediction module is connected to the feature extraction module and is used to predict the charge and discharge curve of the next cycle according to the data features and curve features.

4. The quality testing method according to claim 3, characterized in that: A first feature extraction unit, including multiple convolutional layers and pooling layers; The convolution layer is used to implement forward propagation and extract multi-scale features of the three-dimensional matrix. The number of input channels is 3, the number of output channels is 64, the convolution kernel size is 3, the step size is 1, the padding is 1, the expansion is 1, the number of groups is 1, and an offset is added; the pooling layer is used to return multi-scale features.

5. The quality testing method according to claim 3, characterized in that: The prediction module includes two fully connected layers and an activation function between the two fully connected layers; the first fully connected layer is used to fuse data features and curve features and merge features in the feature dimension; After the activation function, the second fully connected layer outputs the prediction result.

6. The quality testing method according to claim 2, characterized in that: The steps to train a prediction model include: Obtain the initial data of the charge-discharge of several ternary cathode materials; initial data of the structure; charge-discharge curves of 50 cycles; and structural change data after 5, 12, and 20 cycles; Input the initial data of power-off, initial data of structure, charge-discharge curves of the first 20 cycles, and structural change data after 5, 12, and 20 cycles of each ternary cathode material into the prediction model, and output the charge-discharge curve of the next cycle; The charge and discharge data of the last 30 cycles were used as a validation data set, and regression verification was performed on the charge and discharge curves predicted by the model to modify various parameters of the prediction model and obtain a trained prediction model.

7. The quality testing method according to claim 1, characterized in that: When dividing the N cycles into M stages, the division is performed in a manner of increasing intervals.

8. The quality testing method according to any one of claims 1 to 7, characterized in that: Initial data after power-off, including: initial efficiency and charge and discharge capacity at different rates; Initial structural data, including Ni, Co, Mn, impurity content and compaction density; Structural change data, including: any one or more of particle size, morphology, distribution, surface roughness, lithium-nickel mixing data, whether the layered structure is destroyed, and whether microcracks appear.

9. A computer storage medium, characterized in that An executable program code is stored; the executable program code is used to execute the quality testing method described in any one of claims 1-8.

10. A terminal device, characterized in that: It comprises a memory and a processor; the memory stores program codes executable by the processor; the program codes are used to execute the quality testing method according to any one of claims 1 to 8.

Images