Lithium battery SOC prediction method based on attention mechanism gating circulation unit network
By adopting the method of gating the recurrent unit network based on attention mechanism in lithium battery SOC prediction, the problem that existing models cannot reflect the actual load behavior of the battery instantly is solved, and a higher precision SOC prediction is achieved.
Patent Information
- Application Number
- CN202411585650.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-08
- Publication Date
- 2025-03-11
AI Technical Summary
Due to static training, the existing lithium battery SOC prediction model cannot instantly reflect the actual load behavior of the battery, resulting in the failure of the model and the accurate prediction of SOC.
The lithium battery SOC prediction method based on the attention mechanism gated cyclic unit network is adopted. By collecting and pre-processing the lithium battery operation data under different operating conditions, an attention mechanism gated cyclic unit network SOC prediction model is established, real-time SOC prediction is carried out and compared with other models.
The accuracy of lithium battery SOC prediction is improved, and the actual state of the battery can be more accurately reflected, solving the problem of existing model failure.
Smart Images
Figure CN119598280A_ABST
Abstract
Description
Technical Field
[0001] The present invention mainly relates to the technical field of lithium batteries, and specifically to a lithium battery SOC prediction method based on an attention mechanism gated recurrent unit network. Background Art
[0002] The battery state of charge (SOC) is a very important parameter indicator during the operation of electric vehicles. It is an important basis for judging the remaining battery power, preventing the battery from overcharging or over-discharging, and improving battery performance. It is also one of the key technologies that the battery management system needs to solve.
[0003] In actual situations, the internal state of the battery may be significantly different from the training model, because the battery training model used is static, and the actual load behavior of the battery does not change like the standard operating conditions. That is, the data used to train the model cannot instantly reflect the internal state of the battery under current conditions, and thus the training model will lose its effectiveness. Summary of the invention
[0004] The present invention mainly provides a lithium battery SOC prediction method based on an attention mechanism gated recurrent unit network to solve the technical problems raised in the above background technology.
[0005] The technical solution adopted by the present invention to solve the above technical problems is:
[0006] The lithium battery SOC prediction method based on the attention mechanism gated recurrent unit network includes the following steps:
[0007] S1: The stage of collecting lithium battery operation data under different working conditions: using voltage and current sensors to collect the voltage, current and power parameters of the battery operation, and using temperature sensors to collect the temperature of the battery operation;
[0008] S2: Preprocessing operation phase of battery data collection: Programming to remove abnormal values of the original data collected by the battery and extract characteristic data voltage, current, power and temperature of the battery SOC prediction task;
[0009] S3: Establish a SOC prediction model based on the attention mechanism gated recurrent unit network: divide the data set into a training set and a test set, and perform normalization. Input the preprocessed battery variables, including voltage, current, power, and temperature data, into the network input layer for training; build a SOC prediction model by referencing the attention mechanism to the gated recurrent unit network;
[0010] S4: Comparative analysis of SOC prediction experiments based on attention mechanism gated recurrent unit network model: To analyze the impact of the number of hidden layers and the number of hidden layer neurons on the SOC prediction results, adjust the network parameters to optimize the network weights and biases, adjust the network hyperparameters and select the network hidden layer settings when the loss function is minimized, complete the SOC real-time prediction, and compare the experimental analysis with other models.
[0011] Furthermore, in S1, the lithium battery operation under different working conditions includes three working conditions: DST, FUDS and US06; the time interval for recording and storing data in the acquisition phase is 1 second.
[0012] Furthermore, in S3, the data normalization processing method is a Z-Score normalization method, and its calculation method is:
[0013]
[0014] Among them, x is the original sample, is the standardized sample, μ is the sample mean, and σ is the sample standard deviation.
[0015] Furthermore, in S3, the attention mechanism refers to a gated recurrent unit network including an input layer, a hidden layer, a fully connected layer and an output layer. The input layer receives the training set data, the hidden layer contains multiple recurrent neural network neurons with element attention gates, and the fully connected layer further scales and transforms the output of the hidden layer and maps it to the output layer. The output layer calculates the information of the fully connected layer and gives the output result of the network, that is, the prediction result of SOC.
[0016] Furthermore, in S3, the performance of the SOC prediction model is determined by the measurement results and quantified using specific evaluation indicators, such as using the mean square error MSE as the loss function of the proposed model, and using the mean absolute error MAE and the root mean square error RMSE to evaluate the performance of the model, and the calculation formula is as follows:
[0017]
[0018]
[0019]
[0020]
[0021] Among them, SOC t is the reference SOC value at time t, is the SOC prediction value at time t output by the model, and K is the time length.
[0022] Furthermore, in S4, other models include any neuron structure of a recurrent neural network, such as RNN, LSTM or GRU.
[0023] Compared with the prior art, the present invention has the following beneficial effects:
[0024] In the collection phase of lithium battery operation data under different working conditions, the present invention collects the operation data of lithium batteries under different working conditions, including parameters such as voltage, current, power, and the temperature at which the battery operates. Data preprocessing is performed to remove outliers and extract key feature data required for the battery state prediction task, such as voltage, current, power, and temperature. A gated recurrent unit network based on the attention mechanism is established as a SOC prediction model. The model is compared with other models in experimental analysis to evaluate its performance, and compared with other models in experimental analysis. The accuracy of lithium battery SOC prediction is improved.
[0025] The present invention will be explained in detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 It is a flow chart of the present invention.
[0027] In the figure: 1. Base; 2. Electric push rod; 3. Support plate; 4. Display screen; 5. Eye height tracker; 6. Eye movement tracker. DETAILED DESCRIPTION
[0028] To facilitate understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings, but the present invention can be implemented in different forms and is not limited to the embodiments described in the text. On the contrary, these embodiments are provided to make the content disclosed in the present invention more thorough and comprehensive.
[0029] It should be noted that when an element is referred to as being "fixed on" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used in this article are for illustrative purposes only.
[0030] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.
[0031] For examples, please refer to the attached Figure 1 , a lithium battery SOC prediction method based on an attention mechanism gated recurrent unit network includes the following steps:
[0032] S1: The stage of collecting lithium battery operation data under different working conditions: using voltage and current sensors to collect the voltage, current and power parameters of the battery operation, and using temperature sensors to collect the temperature of the battery operation;
[0033] S2: Preprocessing operation phase of battery data collection: Programming to remove abnormal values of the original data collected by the battery and extract characteristic data voltage, current, power and temperature of the battery SOC prediction task;
[0034] S3: Establish a SOC prediction model based on the attention mechanism gated recurrent unit network: divide the data set into a training set and a test set, and perform normalization. Input the preprocessed battery variables, including voltage, current, power, and temperature data, into the network input layer for training; build a SOC prediction model by referencing the attention mechanism to the gated recurrent unit network;
[0035] S4: Comparative analysis of SOC prediction experiments based on attention mechanism gated recurrent unit network model: To analyze the impact of the number of hidden layers and the number of hidden layer neurons on the SOC prediction results, adjust the network parameters to optimize the network weights and biases, adjust the network hyperparameters and select the network hidden layer settings when the loss function is minimized, complete the SOC real-time prediction, and compare the experimental analysis with other models.
[0036] Furthermore, in S1, the lithium battery operation under different working conditions includes three working conditions: DST, FUDS and US06; the time interval for recording and storing data in the acquisition phase is 1 second.
[0037] Furthermore, in S3, the data normalization processing method is a Z-Score normalization method, and its calculation method is:
[0038]
[0039] Among them, x is the original sample, is the standardized sample, μ is the sample mean, and σ is the sample standard deviation.
[0040] Furthermore, in S3, the attention mechanism refers to a gated recurrent unit network including an input layer, a hidden layer, a fully connected layer and an output layer. The input layer receives the training set data, the hidden layer contains multiple recurrent neural network neurons with element attention gates, and the fully connected layer further scales and transforms the output of the hidden layer and maps it to the output layer. The output layer calculates the information of the fully connected layer and gives the output result of the network, that is, the prediction result of SOC.
[0041] Furthermore, in S3, the performance of the SOC prediction model is determined by the measurement results and quantified using specific evaluation indicators, such as using the mean square error MSE as the loss function of the proposed model, and using the mean absolute error MAE and the root mean square error RMSE to evaluate the performance of the model, and the calculation formula is as follows:
[0042]
[0043]
[0044]
[0045]
[0046] Among them, SOC t is the reference SOC value at time t, is the SOC prediction value at time t output by the model, and K is the time length.
[0047] Furthermore, in S4, other models include any neuron structure of a recurrent neural network, such as RNN, LSTM or GRU.
[0048] The above is an exemplary description of the present invention in combination with the accompanying drawings. It is obvious that the specific implementation of the present invention is not limited to the above-mentioned method. As long as such non-substantial improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.
Claims
1. A lithium battery SOC prediction method based on an attention mechanism gated recurrent unit network, characterized by: The following steps are involved: S1: The stage of collecting lithium battery operation data under different working conditions: using voltage and current sensors to collect the voltage, current and power parameters of the battery operation, and using temperature sensors to collect the temperature of the battery operation; S2: Preprocessing operation phase of battery data collection: Programming to remove abnormal values of the original data collected by the battery and extract characteristic data voltage, current, power and temperature of the battery SOC prediction task; S3: Establish a SOC prediction model based on the attention mechanism gated recurrent unit network: divide the data set into a training set and a test set, and perform normalization. Input the preprocessed battery variables, including voltage, current, power, and temperature data, into the network input layer for training; build a SOC prediction model by referencing the attention mechanism to the gated recurrent unit network; S4: Comparative analysis of SOC prediction experiments based on attention mechanism gated recurrent unit network model: To analyze the impact of the number of hidden layers and the number of hidden layer neurons on the SOC prediction results, adjust the network parameters to optimize the network weights and biases, adjust the network hyperparameters and select the network hidden layer settings when the loss function is minimized, complete the SOC real-time prediction, and compare the experimental analysis with other models.
2. The lithium battery SOC prediction method based on the attention mechanism gated recurrent unit network according to claim 1 is characterized in that: In S1, the lithium battery operation under different working conditions includes three working conditions: DST, FUDS and US06; the time interval for recording and storing data in the acquisition stage is 1 second.
3. The lithium battery SOC prediction method based on the attention mechanism gated recurrent unit network according to claim 1 is characterized in that: In S3, the data normalization processing method is the Z-Score normalization method, and its calculation method is: Among them, x is the original sample, is the standardized sample, μ is the sample mean, and σ is the sample standard deviation.
4. The lithium battery SOC prediction method based on the attention mechanism gated recurrent unit network according to claim 1 is characterized in that: In S3, the attention mechanism refers to a gated recurrent unit network including an input layer, a hidden layer, a fully connected layer and an output layer. The input layer receives the training set data, the hidden layer contains multiple recurrent neural network neurons with element attention gates, and the fully connected layer further scales and transforms the output of the hidden layer and maps it to the output layer. The output layer calculates the information of the fully connected layer and gives the output result of the network, that is, the prediction result of SOC.
5. The lithium battery SOC prediction method based on the attention mechanism gated recurrent unit network according to claim 1 is characterized in that: In S3, the performance of the SOC prediction model is determined by the measurement results and quantified using specific evaluation indicators, such as using the mean square error MSE as the loss function of the proposed model, and using the mean absolute error MAE and the root mean square error RMSE to evaluate the performance of the model. The calculation formula is as follows: Among them, SOC t is the reference SOC value at time t, is the SOC prediction value at time t output by the model, and K is the time length.
6. The lithium battery SOC prediction method based on the attention mechanism gated recurrent unit network according to claim 1 is characterized in that: In S4, other models include any recurrent neural network neuron structure.