CN114355218A - Lithium ion battery charge state prediction method based on multi-feature quantity screening - Google Patents
Lithium ion battery charge state prediction method based on multi-feature quantity screening Download PDFInfo
- Publication number
- CN114355218A CN114355218A CN202210079471.0A CN202210079471A CN114355218A CN 114355218 A CN114355218 A CN 114355218A CN 202210079471 A CN202210079471 A CN 202210079471A CN 114355218 A CN114355218 A CN 114355218A
- Authority
- CN
- China
- Prior art keywords
- curve
- lithium ion
- ion battery
- charge
- prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Secondary Cells (AREA)
Abstract
The invention relates to prediction of a lithium ion battery charge state, in particular to a lithium ion battery charge state prediction method based on multi-feature quantity screening. A lithium ion battery state of charge prediction method based on multi-feature quantity screening comprises the step of collecting external feature parameters of a lithium ion battery for predicting the state of charge (SOC) of the battery in the discharging process of the lithium ion battery. The invention has the beneficial effects that: meanwhile, the charge state of the lithium ion battery is predicted by combining various characteristic quantities of battery charging and discharging; in order to reduce the sensor precision and the environmental influence in the data extraction, the calculation error is reduced by using the finite difference; in the data screening part, the optimal data combination is selected by grouping and screening a capacity increment curve, a temperature change curve, an equal-time voltage drop curve and an equal-time current drop curve; smoothing the characteristic curve using gaussian smoothing in the data fitting; the state of charge of the lithium ion battery is predicted by using the optimized neural network in SOC prediction.
Description
Technical Field
The invention relates to prediction of a lithium ion battery charge state, in particular to a lithium ion battery charge state prediction method based on multi-feature quantity screening.
Background
With the continuous exhaustion of fossil fuels and the resulting increasingly worsening environmental problems, lithium ion batteries are widely used in various fields as a sign of new energy. However, the capacity of a lithium ion battery gradually decreases after repeated charge and discharge cycles. The aging of the battery can affect the safety and reliability of the equipment and even cause serious accidents, such as automobile spontaneous combustion, mobile phone explosion and the like. All countries advocate the development of electric vehicle technology and the promotion of commercialization of electric vehicles. The power battery is the only power source of the pure electric vehicle, and the performance of the electric vehicle depends on the characteristics of the power battery to a great extent. Lithium ion batteries are favored by the industry due to the advantages of high density, long service life, low self-discharge rate and the like, currently occupy the main market of power batteries, and in order to ensure the normal operation of electric automobiles, the real-time monitoring and management of the state of the lithium ion batteries are important. Therefore, a Battery Management System (BMS) is becoming a research focus.
The battery management system is the brain of the power battery system and provides necessary conditions for ensuring the normal operation of the battery module. The indices predicted by the battery management system mainly include SOC (battery state of charge), capacity, energy state, and power state. Among them, the SOC and capacity of the battery are the two most critical indicators to be monitored by the battery management system. Because accurate SOC estimation is of great significance in preventing overcharge or overdischarge, improving battery energy utilization, and ensuring safety and stability of the battery system of the electric vehicle. However, SOC, which is an internal feature of a lithium ion battery, cannot be directly measured during the operation of an automobile, and can be predicted only from external parameters that can be directly measured, such as voltage, current, temperature, and internal resistance. Therefore, the lithium ion battery state of charge prediction method based on the surface temperature and the capacity increment is provided.
The existing battery SOC prediction methods include an ampere-hour integration method, an open-circuit voltage method, a model method, and a data-driven method. The ampere-hour integration method is the simplest SOC estimation method. The ampere-hour integration method integrates the current in the charging and discharging process in the time dimension to obtain the SOC of the battery, but the ampere-hour integration method has strong limitation. On one hand, the method needs to provide an accurate initial SOC value, otherwise, an initial error is easily introduced; on the other hand, the method belongs to an open-loop method, the feedback correction capability is lacked, and the accumulated error is larger and larger along with the increase of time.
The open-circuit voltage method establishes a relation between the open-circuit voltage and the SOC according to the battery off-line open-circuit voltage test data. Commonly used test methods mainly include a low current test method and a pulse discharge test method. The low current test method usually needs to wait for the battery to reach the equilibrium before the open circuit voltage of the battery at equilibrium can be measured, and usually takes a long time, so that the method is difficult to be applied to electric vehicles.
The model method can be classified into an SOC estimation method based on an electrochemical model and an SOC estimation method based on an equivalent circuit model according to a model modeling form. Electrochemical modeling models are too complex to be suitable for real-time measured battery prediction. The equivalent circuit needs to be accurately modeled, parameters of the equivalent circuit need to be identified, the accuracy of an equivalent circuit model with fixed parameters obtained by an offline parameter identification algorithm is usually not high, and real-time calculation also provides new requirements for calculation.
The data driving method is the mainstream method at present, and along with the continuous improvement of the performance of computer hardware, on one hand, the storage capacity is obviously improved, and a large amount of data can be stored; on the other hand, the computing power is obviously enhanced, and the development and the application of a machine learning algorithm are promoted. The increase in storage capacity of large-scale data and the advancement of machine learning algorithms have greatly facilitated the application of data-driven methods. The data driving method adopted in the SOC estimation process can be divided into three categories: fuzzy logic, support vector machines, neural networks. The neural network has stronger nonlinear modeling capability and is suitable for large-scale data processing. However, the conventional SOC estimation method is mainly directed to voltage and current characteristics, rarely considers temperature change in the battery degradation process, lacks uniform consideration on voltage, current and temperature, and also has a fluctuation influence caused by the accuracy of a measurement sensor and the change of the environment, and also has a requirement on the SOC prediction accuracy of the battery.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a lithium ion battery charge state prediction method based on multi-feature quantity screening.
The invention is realized by the following technical scheme:
a lithium ion battery state of charge prediction method based on multi-feature quantity screening comprises the following steps:
s1: in the discharging process of the lithium ion battery, acquiring external characteristic parameters of the lithium ion battery for predicting the SOC of the battery;
s2: according to the error of the collected data, determining the sampling time interval of the finite difference of the data, thereby determining the capacity increment and the surface temperature change;
s3: after a sampling interval is established, obtaining a capacity increment curve IC and a surface temperature change curve DT through a Gaussian smooth curve, and recording a voltage change curve and a current change curve during discharging;
s4: the following characteristic quantities are extracted from the IC and DT curves respectively: selecting the coordinates of the peak value and the valley value of the two curves, the distance between the peak and the peak, the distance between the peak and the valley and the area between the peak and the peak, calculating the equal-time voltage drop value for the voltage curve and recording the equal-time current drop value for the current curve;
s5: screening and grouping the characteristic quantities for prediction, comparing the prediction result with the algorithm calculation speed and precision, and selecting the optimal characteristic quantity group;
s6: and leading the optimal characteristic quantity of the battery under different battery SOC conditions into a neural network prediction model so as to obtain a prediction result, and analyzing the prediction result and the prediction error.
According to the above technical solution, preferably, in the step S1, in the running process of the electric vehicle, the external characteristic parameters of the battery collected in real time include voltage U, current I, temperature T, charge-discharge time T, and a true value of the state of charge SOC of the battery.
According to the above technical solution, in step S3, the calculation formula at any time k of the capacity increment curve and the surface temperature change curve is preferably as follows:
according to the above technical solution, preferably, in step S5, the feature quantities are subjected to screening and grouping:
a first group: a capacity increment curve, a temperature change curve, an equal-time voltage drop curve and an equal-time current drop curve;
second group: a capacity increment curve, a temperature change curve;
third group: a capacity increment curve, a temperature change curve, an equal time pressure drop curve;
and a fourth group: a capacity increment curve, a temperature change curve, and an equal-time current reduction curve;
and respectively predicting the SOC of the battery for the four groups of data.
According to the above technical solution, preferably, in step S5, the selected optimal feature quantity group includes: the integrated circuit comprises an IC curve two-peak height value, a voltage value corresponding to an IC curve two-peak, an IC curve 3.5V-3.9V curve integral value, a DT curve one-valley height value, a DT curve one-peak corresponding voltage value, a DT curve two-valley height value, a DT curve two-valley corresponding voltage value and a DT curve one-peak two-valley distance.
According to the above technical solution, preferably, the optimal feature quantity group selected in step S5 is divided into a training set and a test set.
According to the above technical solution, preferably, in the step S6, the neural network prediction model includes: the system comprises an input layer, a hidden layer and an output layer, wherein the input layer is a selected optimal characteristic quantity combination, and the input layer selects 9 nodes; the hidden layer is used for information transformation of an information processing end and selects 6 nodes; the output of the output layer is the state of charge (SOC) of the battery, and 1 node is selected.
According to the above technical solution, preferably, the process of constructing the neural network prediction model is as follows: and importing the training set into a neural network, enabling the neural network to complete optimization selection through an optimization algorithm, and predicting the test set by the optimized neural network.
According to the above technical solution, preferably, the analysis prediction result and the prediction error include a mean square error MSE, a root mean square error RMSE, a mean absolute error MAE, and a mean absolute percentage error MAPE.
According to the above technical solution, preferably, the specific calculation formula of the mean square error MSE is:
the specific calculation formula of the root mean square error RMSE is as follows:
the average absolute error MAE is specifically calculated by the following formula:
the mean absolute percent error MAPE is:
the invention has the beneficial effects that: meanwhile, the charge state of the lithium ion battery is predicted by combining various characteristic quantities of battery charging and discharging; in order to reduce the sensor precision and the environmental influence in the data extraction, the calculation error is reduced by using the finite difference; in the data screening part, the optimal data combination is selected by grouping and screening a capacity increment curve, a temperature change curve, an equal-time voltage drop curve and an equal-time current drop curve; smoothing the characteristic curve using gaussian smoothing in the data fitting; the state of charge of the lithium ion battery is predicted by using the optimized neural network in SOC prediction.
Drawings
Fig. 1 shows a battery IC curve for different fade periods for an embodiment of the present invention.
Fig. 2 shows a battery DT curve for different decay periods for an embodiment of the present invention.
FIG. 3 shows a prediction algorithm flow diagram of an embodiment of the present invention.
Fig. 4 shows a flow chart of a prediction method structure of an embodiment of the invention.
Fig. 5 shows a schematic flow diagram of a lithium ion battery state of charge prediction method of an embodiment of the invention.
Fig. 6 shows battery degradation test set results for an embodiment of the present invention.
Fig. 7 shows a fitness curve for an embodiment of the invention.
Fig. 8 illustrates a prediction error distribution of a battery degradation data set of an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood by those skilled in the art, the present invention will be further described in detail with reference to the accompanying drawings and preferred embodiments.
As shown in the figure, the method for predicting the state of charge of the vehicle-mounted lithium ion battery based on battery multi-feature quantity screening comprises the following steps:
s1: in the discharging process of the lithium ion battery, acquiring external characteristic parameters of the lithium ion battery for predicting the SOC of the battery;
s2: determining the sampling time interval of finite difference of data according to the error of the acquired data
Determining the capacity increment and the temperature change;
s3: after a sampling interval is established, processing data, obtaining a capacity increment curve IC and a surface temperature change curve DT through a Gaussian smooth curve, and recording a voltage change curve and a current change curve between discharges;
s4: respectively extracting characteristic quantities of the IC curve and the DT curve, selecting coordinates of peaks and valleys of the IC curve and the DT curve, selecting distances between peaks and peaks, distances between peaks and valleys and areas between peaks and peaks, calculating an equal-time voltage drop value for the voltage curve, and recording the equal-time current drop value for the current curve;
s5: screening the extracted characteristic quantity, grouping and predicting, comparing a prediction result with the algorithm calculation speed and precision, considering actual engineering conditions, and selecting an optimal characteristic quantity group on the premise of ensuring the prediction precision;
s6: and leading the optimal characteristic quantity of the battery under different SOC conditions into a neural network so as to obtain a prediction result, and analyzing the prediction result and the prediction error.
In step S1, the external characteristic parameters of the battery collected in real time during the charging and discharging process of the battery include voltage (U), current (I), temperature (T), charging and discharging time (T), and the true value of the SOC of the battery.
In step S2, since the accuracy of the temperature sensor is limited, if the direct calculation is easily affected by external noise, and a large deviation is generated, in order to avoid such problems, the temperature change rate is approximately obtained by using a finite difference method, which can be selected according to actual conditions, where L = 30.
In step S3, the calculation formula at any time k of the capacity increase curve and the surface temperature change curve is:
in step S4, the required characteristic amount is obtained by plotting a capacity increment curve and a temperature change curve for different SOCs. Wherein fig. 1 shows capacity increment curves of 500, 1000, 2000, 3000, 5000, 6000, 7000 and 8100 charge-discharge cycles, respectively. Two peak value coordinates of the curve, a valley value coordinate, a peak-peak distance value and a peak integral area value are extracted. Wherein fig. 2 shows temperature change curves of 500, 1000, 3000, 5000, 7000 and 8100 charge-discharge cycles, respectively. Extracting the first valley coordinate, the first peak coordinate, the second valley coordinate, the maximum coordinate, and each peak-peak and peak-valley distance value of the curve. The equal-time voltage drop value is determined as voltage change data corresponding to 60s, and the equal-time current drop value is determined as voltage change data corresponding to 60 s.
In step S5, the characteristic value data of the capacity increment curve, the temperature change curve, the equal-time voltage drop curve, and the equal-time current drop curve are sorted and grouped:
firstly, the method comprises the following steps: a capacity increment curve, a temperature change curve, an equal-time voltage drop curve and an equal-time current drop curve;
II, secondly: a capacity increment curve, a temperature change curve;
thirdly, the method comprises the following steps: a capacity increment curve, a temperature change curve, an equal time pressure drop curve;
fourthly, the method comprises the following steps: capacity increment curve, temperature change curve, equal time current reduction curve.
The four groups of data are respectively subjected to SOC prediction, and the calculation and optimization algorithm iteration time can be reduced as far as possible under the premise of ensuring the prediction precision by considering a capacity increment curve and a temperature change curve, so that the real-time SOC monitoring of the battery is facilitated, and the method has better engineering significance.
In step S5, the determined capacity increment curve shows that a peak gradually decreases with the cycle, and after 3000 times, the peak is basically unrecognizable, and a peak also decreases, but is still clearly visible, so the peak height value of the IC curve is selected, and the integral area of the curve between two peaks of 3.5V and 3.9V is selected to represent the characteristic quantity. The height value of a first valley of the temperature change curve is gradually reduced along with circulation, the peak point is gradually shifted to the right, the height value of a first valley of the DT curve is gradually reduced along with circulation and is shifted to the right, the second valley of the DT curve is increased along with circulation and is shifted to the left, the distance between the first peak point and the second valley point of the DT curve is gradually reduced, the voltage values of the highest point and the highest point of the right-side curve are quite disordered, the peak voltage values and the highest point voltage values are all lower than 0.6 through Pearson correlation coefficient calculation, and good linear correlation is lacked.
Through the above description, it is determined that parameters of selecting IC curve two-peak height value, voltage value corresponding to IC curve two-peak, IC curve 3.5V-3.9V curve integral value, DT curve one-valley height value, DT curve one-peak corresponding voltage value, DT curve two-valley height value, DT curve two-valley corresponding voltage value, and distance 9 between DT curve one-peak two-valley are used as characteristic quantities, and pearson related parameters between each parameter and SOC are calculated with reference to the battery deterioration data set as follows:
| characteristic parameter Numerical value | IC curve Peak height of two Degree of value | IC curve two peak Corresponding voltage Value of | IC curve 3.5V- 3.9V Curve product Score value | DT curve One valley height Degree of value | DT curve Peak height of Degree of value | DT curve 1 Peak to peak power Pressure value | DT curve Height of two valleys Degree of value | DT curve two Corresponding electricity in valley Pressure value | Peak of DT curve Distance between two valleys Separation device |
| Pearson's capsule Correlation system Number of | - 0.9485 | 0.9501 | 0.9959 | 0.6255 | 0.9465 | -0.8587 | -0.6495 | 0.8858 | 0.9675 |
The results show that the 9 parameters have strong correlation with the SOC of the battery, the SOC under the comprehensive judgment of 9 groups of data is more accurate, and the prediction of the SOC of the vehicle-mounted battery is facilitated.
In step S5, the training set and the test set are obtained by dividing the calculated characteristic parameter data set, and the training set and the test set are divided according to 85% and 15%, where the decline data set is three batteries of the same model and has 200 groups of data, where the training set is 170 groups and the test set is 30 groups. The neural network belongs to an artificial intelligence algorithm, has a three-layer structure and comprises: input layer, hidden layer, output layer, as shown in fig. 4. Wherein:
the input layer is different input characteristic parameter data set combinations in the step S4, and 9 nodes are input into the input layer because the input characteristic quantities are 9;
the hidden layer is used for information transformation at the information processing end, the middle layer can be designed into a single hidden layer or a multi-hidden layer structure according to the requirement of information change capability, and the output of the hidden layer is calculated by the following formula:
in the formula:
is an excitation function;
a threshold corresponding to the output layer;
、
respectively, the connection weight value between the hidden layer and the output value of the hidden layer. Wherein
Is a nonlinear transformation function-Sigmoid function (also called S-function). This hidden layer is 6.
The output layer is an SOC value of the lithium ion battery obtained through prediction, and the output value of each node of the output layer is as follows:
in the formula:
a threshold value for a neural network hidden layer node;
the weight value of the connection between the input layer and the hidden layer of the network. Since the output is the battery SOC, the output level is selected to be 1. To solve the problem that the performance of the neural network is weighted by the connection
、
And connection threshold
、
Shadow ofAnd (3) adopting a neural network for optimizing the weight to solve the problem of larger noise, and introducing an optimization algorithm to the initialization parameters to complete optimization selection. The flow of the network prediction algorithm for optimizing the parameters is shown in fig. 3. The lithium ion battery prediction model based on the battery temperature and the capacity change comprises the following specific steps:
(1) and respectively putting the multiple groups of training sets in the step S4 into the optimized neural networks.
(2) And training the neural network by using the training set until the requirements of training errors and training times are met.
(2) And operating the test set by using the trained neural network to obtain a predicted value.
(3) And carrying out error analysis on the SOC predicted values and the actual values of the test set of different lithium ion batteries.
(4) If the error is good, the trained model can be suitable for SOC prediction under the working condition of the battery, and the prediction result can further accord with the true value through the increase of the training set, so that the accuracy of predicting the SOC of the battery is greatly improved. Finally, a model is established by using a battery decline data set, the prediction result is shown in figure 6, the prediction error of 30 groups of results is shown in figure 8, the maximum error is 9%, and the average absolute error is about 3.6%.
Wherein, in the step S5, the performing precision analysis on the predicted value includes: mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percent Error (MAPE). The MSE can evaluate the change degree of data, and the smaller the MSE value is, the better the accuracy of the prediction model is. RMSE is the arithmetic square root of MSE. The MAE can better reflect the actual situation of predicted value errors, and the smaller the MAE value is, the better the accuracy of the prediction model is. The smaller the MAPE value, the better the accuracy of the prediction model. Wherein:
the specific calculation formula of the mean square error MSE is as follows:
the specific calculation formula of the root mean square error RMSE is as follows:
the average absolute error MAE is specifically calculated by the following formula:
the mean absolute percent error MAPE is:
the invention provides a lithium ion battery state of charge prediction method based on battery multi-feature quantity screening, which takes the practical problem that direct SOC prediction is difficult in the operation of an electric automobile lithium ion battery under the actual condition into consideration. The influence of different input parameter combinations on the SOC value is analyzed, data minimization is achieved on the premise that prediction accuracy is guaranteed, and calculation time is shortened. The method has the advantages that the optimized neural network model is utilized to realize the estimation of the state of charge of the battery, the model has high calculation speed and small error on the prediction result of the state of charge of the lithium ion battery, the requirements on the prediction result of the state of charge of the lithium ion battery under different running conditions are met, the usability is high, the capacity prediction model result is shown in figure 6, and the errors after prediction are as follows:
| class of error | Mean square error MSE | Root mean square error RMSE | Mean absolute error MAE | Mean absolute percent error MAPE |
| Error value | 0.0019564 | 0.044231 | 0.036792 | 9.3051% |
The method can be obtained through final prediction results and error analysis, and the lithium ion battery charge state prediction method based on multi-feature quantity screening has good prediction precision and introduces temperature change, thereby being beneficial to the grasp of the battery working state and preventing the occurrence of battery thermal runaway; according to the invention, a relatively accurate prediction model is successfully established through external characteristic quantity changes such as voltage, current, temperature, charging and discharging time and the like, and the prediction method has relatively good prediction precision for the prediction of the SOC of the lithium ion battery and has a certain application value.
The invention has the beneficial effects that: meanwhile, the charge state of the lithium ion battery is predicted by combining various characteristic quantities of battery charging and discharging; in order to reduce the sensor precision and the environmental influence in the data extraction, the calculation error is reduced by using the finite difference; in the data screening part, the optimal data combination is selected by grouping and screening a capacity increment curve, a temperature change curve, an equal-time voltage drop curve and an equal-time current drop curve; smoothing the characteristic curve using gaussian smoothing in the data fitting; the state of charge of the lithium ion battery is predicted by using the optimized neural network in SOC prediction.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A lithium ion battery state of charge prediction method based on multi-feature quantity screening is characterized by comprising the following steps:
s1: in the discharging process of the lithium ion battery, acquiring external characteristic parameters of the lithium ion battery for predicting the SOC of the battery;
s2: according to the error of the collected data, determining the sampling time interval of the finite difference of the data, thereby determining the capacity increment and the surface temperature change;
s3: after a sampling interval is established, obtaining a capacity increment curve IC and a surface temperature change curve DT through a Gaussian smooth curve, and recording a voltage change curve and a current change curve during discharging;
s4: the following characteristic quantities are extracted from the IC and DT curves respectively: selecting the coordinates of the peak value and the valley value of the two curves, the distance between the peak and the peak, the distance between the peak and the valley and the area between the peak and the peak, calculating the equal-time voltage drop value for the voltage curve and recording the equal-time current drop value for the current curve;
s5: screening and grouping the characteristic quantities for prediction, comparing the prediction result with the algorithm calculation speed and precision, and selecting the optimal characteristic quantity group;
s6: and leading the optimal characteristic quantity of the battery under different battery SOC conditions into a neural network prediction model so as to obtain a prediction result, and analyzing the prediction result and the prediction error.
2. The lithium ion battery state of charge prediction method based on multi-feature quantity screening according to claim 1, characterized in that: in the step S1, in the running process of the electric vehicle, the external characteristic parameters of the battery collected in real time include voltage U, current I, temperature T, charge-discharge time T, and the true value of the state of charge SOC of the battery.
3. The method according to claim 2, wherein in step S3, the calculation formula at any time k of the capacity increment curve and the surface temperature variation curve is:
4. the lithium ion battery state of charge prediction method based on multi-feature quantity screening according to claim 3, wherein in the step S5, the feature quantities are screened and grouped as follows:
a first group: a capacity increment curve, a temperature change curve, an equal-time voltage drop curve and an equal-time current drop curve;
second group: a capacity increment curve, a temperature change curve;
third group: a capacity increment curve, a temperature change curve, an equal time pressure drop curve;
and a fourth group: a capacity increment curve, a temperature change curve, and an equal-time current reduction curve;
and respectively predicting the SOC of the battery for the four groups of data.
5. The lithium ion battery state of charge prediction method based on multi-feature-quantity screening according to claim 4, wherein in the step S5, the selected optimal feature quantity group includes: the integrated circuit comprises an IC curve two-peak height value, a voltage value corresponding to an IC curve two-peak, an IC curve 3.5V-3.9V curve integral value, a DT curve one-valley height value, a DT curve one-peak corresponding voltage value, a DT curve two-valley height value, a DT curve two-valley corresponding voltage value and a DT curve one-peak two-valley distance.
6. The lithium ion battery state of charge prediction method based on multi-feature-quantity screening according to claim 5, wherein the optimal feature quantity group selected in the step S5 is divided into a training set and a test set.
7. The lithium ion battery state of charge prediction method based on multi-feature quantity screening according to claim 6, wherein in the step S6, the neural network prediction model comprises: the system comprises an input layer, a hidden layer and an output layer, wherein the input layer is a selected optimal characteristic quantity combination, and the input layer selects 9 nodes; the hidden layer is used for information transformation of an information processing end and selects 6 nodes; the output of the output layer is the state of charge (SOC) of the battery, and 1 node is selected.
8. The lithium ion battery state of charge prediction method based on multi-feature quantity screening according to claim 7, characterized in that the process of constructing the neural network prediction model is as follows: and importing the training set into a neural network, enabling the neural network to complete optimization selection through an optimization algorithm, and predicting the test set by the optimized neural network.
9. The lithium ion battery state of charge prediction method based on multi-feature quantity screening of claim 8, wherein the analysis prediction result and the prediction error comprise mean square error MSE, mean square error RMSE, mean absolute error MAE, and mean absolute percentage error MAPE.
10. The lithium ion battery state of charge prediction method based on multi-feature-quantity screening according to claim 9, wherein the specific calculation formula of the mean square error MSE is as follows:
the specific calculation formula of the root mean square error RMSE is as follows:
the average absolute error MAE is specifically calculated by the following formula:
the mean absolute percent error MAPE is:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210079471.0A CN114355218A (en) | 2022-01-24 | 2022-01-24 | Lithium ion battery charge state prediction method based on multi-feature quantity screening |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210079471.0A CN114355218A (en) | 2022-01-24 | 2022-01-24 | Lithium ion battery charge state prediction method based on multi-feature quantity screening |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114355218A true CN114355218A (en) | 2022-04-15 |
Family
ID=81092231
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210079471.0A Pending CN114355218A (en) | 2022-01-24 | 2022-01-24 | Lithium ion battery charge state prediction method based on multi-feature quantity screening |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114355218A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117783887A (en) * | 2024-02-28 | 2024-03-29 | 深圳市神通天下科技有限公司 | Lithium ion battery cell matching screening method |
| CN118395349A (en) * | 2024-06-24 | 2024-07-26 | 广东阿尔派电力科技股份有限公司 | Lithium battery energy loss prediction method and system based on big data analysis |
-
2022
- 2022-01-24 CN CN202210079471.0A patent/CN114355218A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117783887A (en) * | 2024-02-28 | 2024-03-29 | 深圳市神通天下科技有限公司 | Lithium ion battery cell matching screening method |
| CN117783887B (en) * | 2024-02-28 | 2024-05-14 | 深圳市神通天下科技有限公司 | Lithium ion battery cell matching screening method |
| CN118395349A (en) * | 2024-06-24 | 2024-07-26 | 广东阿尔派电力科技股份有限公司 | Lithium battery energy loss prediction method and system based on big data analysis |
| CN118395349B (en) * | 2024-06-24 | 2024-09-13 | 广东阿尔派电力科技股份有限公司 | Lithium battery energy loss prediction method and system based on big data analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111398833B (en) | Battery health state assessment method | |
| CN113805064B (en) | Lithium ion battery pack health state prediction method based on deep learning | |
| CN107741568B (en) | Lithium battery SOC estimation method based on state transition optimization RBF neural network | |
| CN112904219B (en) | Big data-based power battery health state prediction method | |
| CN114355218A (en) | Lithium ion battery charge state prediction method based on multi-feature quantity screening | |
| CN114818831B (en) | Bidirectional lithium ion battery fault detection method and system based on multi-source perception | |
| CN111191824B (en) | Power battery capacity attenuation prediction method and system | |
| CN114280490B (en) | Lithium ion battery state of charge estimation method and system | |
| CN116298934B (en) | Modeling method of prediction network for lithium battery health state estimation | |
| CN116774086B (en) | Lithium battery health state estimation method based on multi-sensor data fusion | |
| CN110568360A (en) | lithium battery aging diagnosis method based on fuzzy logic algorithm | |
| WO2024103563A1 (en) | Method for screening and sorting retired batteries | |
| CN110673037A (en) | Battery SOC estimation method and system based on improved simulated annealing algorithm | |
| Yao et al. | Fault identification of lithium-ion battery pack for electric vehicle based on ga optimized ELM neural network | |
| CN116400241A (en) | Battery health assessment method and device based on full life cycle operation data | |
| CN112287980A (en) | Power battery screening method based on typical feature vector | |
| CN115219913A (en) | Power battery full-life-cycle management system based on capacity increment method | |
| CN115730525A (en) | Rail transit UPS storage battery health state prediction method | |
| CN114646888A (en) | Assessment method and system for capacity attenuation of power battery | |
| CN114545275A (en) | Indirect prediction method for remaining service life of lithium ion battery | |
| CN117169743A (en) | Battery health state estimation method and device based on partial data and model fusion | |
| CN117129872A (en) | Self-adaptive estimation method for lithium ion battery health state | |
| CN116736172A (en) | Lithium battery pack health state evaluation method based on generation type countermeasure network | |
| CN116679208A (en) | Lithium battery residual life estimation method | |
| CN114545276A (en) | Power battery service life prediction method based on capacity test and Internet of vehicles big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |