Lithium battery overcharge early warning system based on big data

Abstract

The invention discloses a lithium battery overcharge early warning system based on big data, and the system comprises a data collection module, a data processing module, a dynamic threshold generation module and a risk analysis module, and relates to the technical field of data processing. A local mean value and a standard deviation are calculated by using a sliding window method, a similar data cluster is constructed by using a dynamic time warping algorithm, and overcharge voltage and temperature thresholds are calculated in combination with a preset risk ratio, so that the limitation of a single Gaussian mixture model is avoided, and the accuracy of the thresholds is improved. Historical data are also acquired, a prediction model is constructed by using an autoregression integral moving average model, a slope is calculated to obtain a correction factor, an aging model constructed by machine learning is combined to update a threshold value, dynamic change and aging of the battery are combined, battery state change is adapted, and early warning reliability and timeliness are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a lithium battery overcharge warning system based on big data. Background Art

[0002] During the use of lithium batteries, overcharging is a problem that seriously affects battery performance and safety. Overcharging may cause the internal temperature of the battery to rise, the pressure to increase, and even lead to safety accidents such as combustion and explosion. Therefore, it is crucial to effectively warn against lithium battery overcharging.

[0003] Publication No. CN110148991B discloses a battery overcharge warning method and system based on big data. The method includes: determining the battery type; if it is a ternary lithium battery, calculating the state of charge value of the ternary lithium battery; determining whether a fault occurs based on the state value; extracting the highest single-cell voltage and the highest probe temperature of the ternary lithium battery; determining the overcharge threshold of the ternary lithium battery; dividing the fault level and giving a warning based on the overcharge threshold, the highest single-cell voltage, and the highest probe temperature. If it is a lithium iron phosphate battery, the same method is used to divide the faults of the lithium iron phosphate battery.

[0004] However, the above application still has the following problems: Publication No. CN110148991B generates overcharge voltage thresholds and overcharge temperature thresholds for different battery types by dynamically generating thresholds through a Gaussian mixture model. The above application does not fully consider the influence of different charging stages, different usage environments, and battery aging factors on the overcharge threshold. The Gaussian mixture model assumes that the data is composed of multiple Gaussian distributions mixed together, and this assumption will cause the model to be unable to accurately capture the true characteristics of the data, thereby generating overcharge thresholds that deviate from the actual situation and reducing the accuracy of the warning. Summary of the Invention

[0005] To solve the technical problems in the background art, the present invention proposes a lithium battery overcharge warning system based on big data.

[0006] The lithium battery overcharge warning system based on big data proposed by the present invention includes:

[0007] Data acquisition module: used to obtain the charging data during the charging process of the lithium battery in real time, record the corresponding charge and discharge cycle times and calendar time at the same time, and transmit the charging data, the corresponding charge and discharge cycle times, and the calendar time to the data processing module;

[0008] The charging data includes battery type, charging stage data, and usage environment data;

[0009] Data processing module: used to preprocess the charging data and transmit the preprocessed charging data to the dynamic threshold generation module;

[0010] Dynamic threshold generation module: used to generate overcharge voltage threshold and overcharge temperature threshold, and transmit the generated overcharge voltage threshold and overcharge temperature threshold to the risk analysis module;

[0011] Risk analysis module: used to generate an overcharge risk level model and trigger multi-level warnings according to the risk level output by the overcharge risk level model.

[0012] Preferably, the dynamic threshold generation module includes:

[0013] Classification and feature extraction unit: classifies historical charging data according to battery type, charging stage data, and usage environment data;

[0014] At the same time, the dynamic time warping algorithm is used to calculate the similarity between data and construct similar data clusters;

[0015] Initial threshold determination unit: obtains each similar data cluster in the classification and feature extraction unit, and calculates the threshold range of the overcharge voltage threshold and overcharge temperature threshold for each similar data cluster;

[0016] Threshold adjustment unit: used to update the threshold range of the overcharge voltage threshold and overcharge temperature threshold obtained in the initial threshold determination unit.

[0017] Preferably, in the initial threshold determination unit, the threshold range of the overcharge voltage threshold and overcharge temperature threshold for each similar data cluster is calculated as follows:

[0018] When calculating the overcharge voltage threshold or overcharge temperature threshold of a certain similar data cluster, let a certain similar data cluster X = {x1, x2,..., x n}, and calculate the local mean μ and local standard deviation σ of the voltage data or the local mean μ and local standard deviation σ of the temperature data in the similar data cluster X through the sliding window method;

[0019] x i is the specific value of the voltage data or temperature data, i = 1, 2,..., n;

[0020] Preset risk ratio p, for each data point x i , with μ as the center and σ as the scale parameter, calculate the probability density value f(x i );

[0021]

[0022] Sort all f(x i ) from small to large, and take the n×p-th probability density value as ∈;

[0023] ∈ represents the probability density value corresponding to the preset risk ratio p, representing a small probability density threshold;

[0024] Solve for the x in f(x i ) < ∈, that is, the overcharge threshold range; i

[0025] Let f(x i ) = ∈, that is:

[0026] Transform both sides of the equation:

[0027]

[0028] Then, take the natural logarithm of both sides. According to the property of logarithms, ln(exp(a)) = a, we can get:

[0029]

[0030] Multiply both sides of the equation by -2σ 2 , and we get:

[0031]

[0032] Take the square root of both sides of the equation to solve for x i , and we get:

[0033]

[0034] Then the lower threshold is:

[0035]

[0036] The upper threshold is:

[0037]

[0038] That is, the threshold range is

[0039] If the local mean μ and local standard deviation σ of the voltage data are used, the calculated value is the overcharge voltage threshold;

[0040] If the local mean μ and local standard deviation σ of the temperature data are used, the calculated value is the overcharge temperature threshold.

[0041] Preferably, in the threshold adjustment unit, update the threshold range of the overcharge voltage threshold and overcharge temperature threshold obtained in the initial threshold determination unit as follows:

[0042] Obtain the voltage data sequence and temperature data sequence in the historical charging data, and record the corresponding charge and discharge cycle times and calendar time simultaneously;

[0043] ​For voltage data and temperature data, an autoregressive integrated moving average model is used for modeling to construct a voltage prediction model and a temperature prediction model. The voltage prediction model is used to predict voltage values, and the temperature prediction model is used to predict temperature values;

[0044] The voltage prediction model and the temperature prediction model are used to predict the voltage and temperature in the future unit time, and a predicted voltage sequence and a predicted temperature sequence are obtained;

[0045] Calculate the slope ΔV of the predicted value of the predicted voltage sequence slope ;

[0046] Calculate the slope ΔT of the predicted value of the predicted temperature sequence slope ;

[0047] Calculate the voltage correction factor γV:

[0048]

[0049] where α is the slope sensitivity coefficient;

[0050] Calculate the temperature correction factor γT:

[0051]

[0052] where β is the slope sensitivity coefficient;

[0053] For the recorded charge and discharge cycle times and calendar time, a battery aging model is constructed through a machine learning algorithm to obtain the overcharge voltage threshold aging coefficient k V and the overcharge temperature threshold aging coefficient k T ;

[0054] Determine the overcharge voltage threshold and the overcharge temperature threshold obtained by the unit according to the initial threshold;

[0055] Let the overcharge voltage threshold be the overcharge temperature threshold be

[0056] In the future unit time, update the overcharge voltage threshold and the overcharge temperature threshold to:

[0057]

[0058] Preferably, the risk analysis module includes:

[0059] State of Charge calculation unit: Obtain charging data and calculate the battery SOC in real time based on the ampere-hour integration method and the open-circuit voltage method;

[0060] Fault level classification unit: Obtain the SOC obtained by the state of charge calculation unit, the overcharge voltage threshold and the overcharge temperature threshold obtained by the dynamic threshold generation module, and the real-time charging data obtained by the data acquisition module, and construct an overcharge risk level model by using the fuzzy logic algorithm;

[0061] Real-time warning unit: Trigger multi-level warnings through the risk level output by the overcharge risk level model in the fault level classification unit, and push them to the user terminal through the cloud platform.

[0062] Preferably, the data processing module includes:

[0063] Missing value filling subunit: Used to fill in missing data through the time series correlation analysis algorithm;

[0064] Outlier filtering subunit: Used to identify and correct outliers through numerical boundary analysis and dynamic standard deviation method.

[0065] Preferably, it further includes:

[0066] Visualization interface module: Used to display in real time the charging data during the lithium battery charging process, the overcharge voltage threshold and the overcharge temperature threshold generated by the dynamic threshold generation module, and the risk level output by the overcharge risk level model in the risk analysis module.

[0067] A lithium battery overcharge warning method based on big data, including the following steps:

[0068] S1. Obtain the charging data during the lithium battery charging process in real time, and record the corresponding charge and discharge cycle times and calendar time;

[0069] The charging data includes battery type, charging stage data, and usage environment data;

[0070] S2. Preprocess the charging data;

[0071] S3. Classify the historical charging data according to the battery type, charging stage data and usage environment data, calculate the local mean and standard deviation of the data by using the sliding window method, calculate the similarity between the data by using the dynamic time warping algorithm to construct similar data clusters, and obtain the threshold ranges of the overcharge voltage threshold and the overcharge temperature threshold for each similar data cluster;

[0072] By obtaining the historical voltage and temperature data sequences and the corresponding charge and discharge cycle times and calendar time, use the autoregressive integrated moving average model to construct a voltage prediction model and a temperature prediction model, predict the voltage sequence and the temperature sequence predicted in the future unit time, calculate the slopes of the predicted voltage sequence and the predicted temperature sequence to obtain the voltage correction factor and the temperature correction factor, and construct a battery aging model through machine learning algorithms to obtain the overcharge voltage threshold aging coefficient and the overcharge temperature threshold aging coefficient;

[0073] In the next unit of time, update the overcharge voltage threshold and overcharge temperature threshold to:

[0074] The updated threshold of the overcharge voltage threshold = the overcharge voltage threshold × the overcharge voltage threshold aging coefficient × the voltage correction factor;

[0075] The updated threshold of the overcharge temperature threshold = the overcharge temperature threshold × the overcharge temperature threshold aging coefficient × the temperature correction factor;

[0076] S4. Generate an overcharge risk level model, calculate the battery SOC in real time based on the ampere-hour integration method and the open-circuit voltage method, obtain the SOC, the overcharge voltage threshold and the overcharge temperature threshold obtained by the dynamic threshold generation module, and the real-time charging data. Construct an overcharge risk level model by using the fuzzy logic algorithm, trigger multi-level warnings through the risk level output by the overcharge risk level model, and push them to the user terminal through the cloud platform.

[0077] In the present invention, the proposed overcharge warning system for lithium batteries based on big data has the following beneficial technical effects:

[0078] 1. In this application, by classifying historical charging data according to battery type, charging stage data, and usage environment data, and comprehensively classifying data according to battery type, charging stage, and usage environment, it can more precisely reflect the charging characteristics of the battery under different conditions;

[0079] Adopt the dynamic time warping algorithm to construct similar data clusters. For each similar data cluster, use the sliding window method to calculate the local mean and standard deviation of the data. The sliding window method ensures that the overcharge voltage threshold and the overcharge temperature threshold can quickly respond to short-term changes in voltage or temperature;

[0080] By presetting the risk ratio, calculate the overcharge voltage threshold and the overcharge temperature threshold, avoiding the limitations of the single Gaussian mixture model, thereby improving the accuracy of the overcharge voltage threshold and the overcharge temperature threshold;

[0081] By obtaining the historical voltage and temperature data sequences and the corresponding charge and discharge cycle numbers and calendar times, use the autoregressive integrated moving average model to construct a voltage prediction model and a temperature prediction model to predict the predicted voltage sequence and the predicted temperature sequence in the next unit of time, which is used to reflect the long-term trends of voltage and temperature and avoid the one-sidedness of local statistics of the sliding window method;

[0082] The voltage correction factor and the temperature correction factor are obtained by calculating the slopes of the predicted voltage sequence and the predicted temperature sequence. The overcharge voltage threshold aging coefficient and the overcharge temperature threshold aging coefficient are obtained by constructing a battery aging model through a machine learning algorithm, and then the overcharge voltage threshold and the overcharge temperature threshold are updated. This process combines the dynamic changes of battery data and the impact of battery aging on the overcharge threshold, adapts to the change of battery performance over time, can adapt to the change of battery state in real time, and improves the reliability and timeliness of overcharge warning.

[0083] The risk analysis module of the present application calculates the battery SOC in real time based on the ampere-hour integration method and the open-circuit voltage method, combines the SOC, the dynamic threshold, and the real-time charging data, and constructs an overcharge risk level model by using a fuzzy logic algorithm, which can more comprehensively and accurately evaluate the overcharge risk of the battery and realize multi-level warning.

[0084] The additional aspects and advantages of the present invention will be given in part in the following description, will become apparent in part from the following description, or will be understood through the practice of the present invention. Brief Description of the Drawings

[0085] Figure 1 is a schematic block diagram of the system of the present invention;

[0086] Figure 2 is a flowchart of the method of the present invention. Detailed Description of the Embodiments

[0087] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, in which the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention and should not be construed as limiting the present invention.

[0088] As Figure 1 shown, a lithium battery overcharge warning system based on big data, a data acquisition module: for real-time obtaining the charging data during the charging process of the lithium battery, and simultaneously recording the corresponding charge and discharge cycle times and calendar time, and transmitting the charging data, the corresponding charge and discharge cycle times and calendar time to the data processing module;

[0089] The charging data includes battery type, charging stage data, and usage environment data;

[0090] The charging stage data includes voltage data, internal resistance data, and temperature data;

[0091] The data processing module: for preprocessing the charging data and transmitting the preprocessed charging data to the dynamic threshold generation module;

[0092] Dynamic threshold generation module: used to generate overcharge voltage threshold and overcharge temperature threshold, and transmit the generated overcharge voltage threshold and overcharge temperature threshold to the risk analysis module;

[0093] Risk analysis module: used to generate an overcharge risk level model and trigger multi-level warnings according to the risk level output by the overcharge risk level model.

[0094] The risk analysis module includes:

[0095] State of charge calculation unit: obtains charging data and calculates the battery SOC in real time based on the ampere-hour integration method and the open-circuit voltage method;

[0096] The battery SOC is the charging state of the battery, indicating the percentage of the current remaining available power of the battery in the total capacity;

[0097] Fault level classification unit: obtains the SOC obtained by the state of charge calculation unit, the overcharge voltage threshold and overcharge temperature threshold obtained by the dynamic threshold generation module, and the real-time charging data obtained by the data acquisition module, and constructs an overcharge risk level model by using the fuzzy logic algorithm;

[0098] Real-time warning unit: triggers multi-level warnings through the risk level output by the overcharge risk level model in the fault level classification unit, and pushes them to the user terminal through the cloud platform.

[0099] The fuzzy logic algorithm is a prior art and is a mathematical tool that simulates the fuzziness of human thinking and is used to process uncertain, inaccurate or fuzzy information;

[0100] The risk analysis module of this application calculates the battery SOC in real time based on the ampere-hour integration method and the open-circuit voltage method, combines the SOC, dynamic threshold and real-time charging data, and constructs an overcharge risk level model by using the fuzzy logic algorithm, which can more comprehensively and accurately evaluate the overcharge risk of the battery and realize multi-level warnings.

[0101] The dynamic threshold generation module includes:

[0102] Classification and feature extraction unit: classifies the historical charging data according to the battery type, charging stage data and usage environment data;

[0103] At the same time, the dynamic time warping algorithm is used to calculate the similarity between data and construct similar data clusters to handle the stretching and deformation problems of data on the time axis;

[0104] Initial threshold determination unit: obtains each similar data cluster in the classification and feature extraction unit, and calculates the threshold range of the overcharge voltage threshold and overcharge temperature threshold of each similar data cluster;

[0105] Threshold adjustment unit: used to update the threshold ranges of the overcharge voltage threshold and the overcharge temperature threshold obtained in the initial threshold determination unit;

[0106] In an alternative embodiment, in the initial threshold determination unit, the threshold ranges of the overcharge voltage threshold and the overcharge temperature threshold for each similar data cluster are calculated as follows:

[0107] When calculating the overcharge voltage threshold or the overcharge temperature threshold of a certain similar data cluster, let a certain similar data cluster X = {x1, x2,..., x n}, and the local mean μ and the local standard deviation σ of the voltage data or the local mean μ and the local standard deviation σ of the temperature data in the similar data cluster X are calculated by the sliding window method. The sliding window method can quickly capture the local change characteristics of the data;

[0108] x i is the specific value of the voltage data or the temperature data, i = 1, 2,..., n;

[0109] Because it is calculated by the sliding window method, it is the local mean μ and the local standard deviation σ;

[0110] A preset risk ratio p is set, and for each data point x i , with μ as the center and σ as the scale parameter, the probability density value f(x i ) is calculated;

[0111]

[0112] In an alternative embodiment, p ≤ 5%;

[0113] exp is the exponential function; the value of exp can be obtained by using a scientific calculator;

[0114] All f(x i ) are sorted from smallest to largest, and the n×p-th probability density value is taken as ∈;

[0115] ∈ represents the probability density value corresponding to the preset risk ratio p and represents a small probability density threshold;

[0116] Solve the range of values of x i for which f(x i ) < ∈, that is, the overcharge threshold range;

[0117] Let f(x i ) = ∈, that is:

[0118] Transform both sides of the equation: isolate the exponential part;

[0119]

[0120] Then, take the natural logarithm on both sides. According to the property of logarithms, ln(exp(a)) = a, we can get:

[0121]

[0122] Multiply both sides of the equation by -2σ 2 , and we get:

[0123]

[0124] Take the square root of both sides of the equation to solve for x i , and we get:

[0125]

[0126] Then the lower threshold is:

[0127]

[0128] The upper threshold is:

[0129]

[0130] That is, the threshold range is When the value of the data point x i exceeds this range, there is a risk of overcharging;

[0131] If the local mean μ and local standard deviation σ of the voltage data are used, the calculated value is the overcharge voltage threshold;

[0132] If the local mean μ and local standard deviation σ of the temperature data are used, the calculated value is the overcharge temperature threshold;

[0133] As an illustration, the overcharge voltage threshold and overcharge temperature threshold obtained in the above manner are the overcharge voltage threshold or overcharge temperature threshold of a certain similar data cluster, that is, the overcharge voltage threshold and overcharge temperature threshold generated for different similar data clusters are different;

[0134] In an optional embodiment, in the threshold adjustment unit, the threshold range of the overcharge voltage threshold and overcharge temperature threshold obtained in the initial threshold determination unit is updated as follows:

[0135] Obtain the voltage data sequence and temperature data sequence in the historical charging data, and record the corresponding charge and discharge cycle times and calendar time at the same time;

[0136] For voltage data and temperature data, an autoregressive integrated moving average model is used for modeling to construct a voltage prediction model and a temperature prediction model. The voltage prediction model is used to predict voltage values, and the temperature prediction model is used to predict temperature values;

[0137] The voltage prediction model and the temperature prediction model are used to predict the voltage and temperature in the future unit time, and a predicted voltage sequence and a predicted temperature sequence are obtained;

[0138] Calculate the slope ΔV of the predicted value of the predicted voltage sequence slope ;

[0139] Calculate the slope ΔT of the predicted value of the predicted temperature sequence slope ;

[0140] Calculate the voltage correction factor γV:

[0141]

[0142] where α is the slope sensitivity coefficient;

[0143] Calculate the temperature correction factor γT:

[0144]

[0145] where β is the slope sensitivity coefficient;

[0146] α and β can be obtained by fitting experimental data or by analyzing historical charging data through a machine learning algorithm;

[0147] For the recorded corresponding charge-discharge cycle times and calendar time, a battery aging model is constructed through a machine learning algorithm to obtain the overcharge voltage threshold aging coefficient kV and the overcharge temperature threshold aging coefficient k T ;

[0148] Determine the overcharge voltage threshold and the overcharge temperature threshold obtained by the unit according to the initial threshold;

[0149] Let the overcharge voltage threshold be The overcharge temperature threshold is

[0150] In the future unit time, update the overcharge voltage threshold and the overcharge temperature threshold to:

[0151]

[0152] In this application, by classifying historical charging data according to battery type, charging stage data, and usage environment data, and comprehensively classifying data according to battery type, charging stage, and usage environment, the charging characteristics of the battery under different conditions can be reflected more meticulously;

[0153] The dynamic time warping algorithm is used to construct similar data clusters. For each similar data cluster, the sliding window method is used to calculate the local mean and standard deviation of the data. The sliding window method ensures that the overcharge voltage threshold and the overcharge temperature threshold can quickly respond to short-term changes in voltage or temperature;

[0154] By presetting the risk ratio, the overcharge voltage threshold and the overcharge temperature threshold are calculated, avoiding the limitations of a single Gaussian mixture model, thereby improving the accuracy of the overcharge voltage threshold and the overcharge temperature threshold;

[0155] By obtaining the historical voltage and temperature data sequences and the corresponding charge and discharge cycle numbers and calendar times, the autoregressive integrated moving average model is used to construct a voltage prediction model and a temperature prediction model to predict the predicted voltage sequence and the predicted temperature sequence in the future unit time, which are used to reflect the long-term trends of voltage and temperature, avoiding the one-sidedness of local statistics of the sliding window method;

[0156] By calculating the slopes of the predicted voltage sequence and the predicted temperature sequence, the voltage correction factor and the temperature correction factor are obtained. By using a machine learning algorithm to construct a battery aging model, the overcharge voltage threshold aging coefficient and the overcharge temperature threshold aging coefficient are obtained, and then the overcharge voltage threshold and the overcharge temperature threshold are updated. This process combines the dynamic changes of battery data and the influence of battery aging on the overcharge threshold, adapts to the change of battery performance over time, can adapt to the change of battery state in real time, and improves the reliability and timeliness of overcharge warning.

[0157] In an optional embodiment, the data processing module includes:

[0158] Missing value filling subunit: used to fill in missing data through the time series correlation analysis algorithm;

[0159] The time series correlation analysis algorithm is a mathematical method used to reveal the correlation relationship between time series data;

[0160] Outlier filtering subunit: used to identify and correct outliers through numerical limit analysis and dynamic standard deviation method.

[0161] In an optional embodiment, it further includes: a visualization interface module: used to display in real time the charging data during the lithium battery charging process, the overcharge voltage threshold and the overcharge temperature threshold generated by the dynamic threshold generation module, and the risk level output by the overcharge risk level model in the risk analysis module.

[0162] As Figure 2 shown, a lithium battery overcharge warning method based on big data includes the following steps:

[0163] S1. Obtain the charging data during the lithium battery charging process in real time, and record the corresponding charge and discharge cycle numbers and calendar times at the same time;

[0164] The charging data includes battery type, charging stage data, and usage environment data;

[0165] S2. Preprocess the charging data;

[0166] S3. Classify the historical charging data according to the battery type, charging stage data, and usage environment data. Use the sliding window method to calculate the local mean and standard deviation of the data. Adopt the dynamic time warping algorithm to calculate the similarity between the data to construct similar data clusters, and obtain the threshold ranges of the overcharge voltage threshold and overcharge temperature threshold for each similar data cluster through calculation;

[0167] By obtaining the historical voltage and temperature data sequences and the corresponding charge and discharge cycle numbers and calendar times, use the autoregressive integrated moving average model to construct a voltage prediction model and a temperature prediction model, predict the voltage sequence and temperature sequence predicted in the future unit time, calculate the slopes of the predicted voltage sequence and predicted temperature sequence to obtain the voltage correction factor and temperature correction factor, and construct a battery aging model through machine learning algorithms to obtain the overcharge voltage threshold aging coefficient and overcharge temperature threshold aging coefficient;

[0168] In the future unit time, update the overcharge voltage threshold and overcharge temperature threshold as follows:

[0169] The updated overcharge voltage threshold = overcharge voltage threshold × overcharge voltage threshold aging coefficient × voltage correction factor;

[0170] The updated overcharge temperature threshold = overcharge temperature threshold × overcharge temperature threshold aging coefficient × temperature correction factor;

[0171] S4. Generate an overcharge risk level model. Based on the ampere-hour integration method and open-circuit voltage method, calculate the battery SOC in real time. Obtain the SOC, the overcharge voltage threshold and overcharge temperature threshold obtained by the dynamic threshold generation module, and the real-time charging data. Construct an overcharge risk level model by adopting the fuzzy logic algorithm, trigger multi-level early warnings through the risk level output by the overcharge risk level model, and push them to the user terminal through the cloud platform.

[0172] Meanwhile, the content not detailedly described in this specification belongs to the prior art well-known to those skilled in the art.

[0173] In the embodiments provided by the present invention, it should be understood that the disclosed system or method can be implemented in other ways. For example, the above-described invention embodiments are merely illustrative. For example, the division of modules is only a logical function division, and there may be other division methods in actual implementation.

[0174] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, and they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0175] In addition, in each embodiment of the present invention, the functional modules may be integrated into one processing module, or each module may exist physically alone, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0176] For those skilled in the field of operation and maintenance, it is obvious that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and can be implemented in other specific forms without departing from the basic characteristics of the present invention.

[0177] As mentioned above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent replacements or changes, and should be covered by the protection scope of the present invention.

Claims

1. A lithium battery overcharge warning system based on big data, characterized in that, Including: Data acquisition module: It is used to obtain the charging data during the charging process of the lithium battery in real time, record the corresponding charge-discharge cycle times and calendar time at the same time, and transmit the charging data, the corresponding charge-discharge cycle times and calendar time to the data processing module; The charging data includes battery type, charging stage data, and usage environment data; Data processing module: It is used to preprocess the charging data and transmit the preprocessed charging data to the dynamic threshold generation module; Dynamic threshold generation module: It is used to generate overcharge voltage thresholds and overcharge temperature thresholds, and transmit the generated overcharge voltage thresholds and overcharge temperature thresholds to the risk analysis module; Risk analysis module: It is used to generate an overcharge risk level model and trigger multi-level warnings according to the risk level output by the overcharge risk level model.

2. The lithium battery overcharge warning system based on big data according to claim 1, wherein The dynamic threshold generation module includes: Classification and feature extraction unit: Classify the historical charging data according to battery type, charging stage data, and usage environment data; At the same time, use the dynamic time warping algorithm to calculate the similarity between data and construct similar data clusters; Initial threshold determination unit: Obtain each similar data cluster in the classification and feature extraction unit, and calculate the threshold ranges of the overcharge voltage thresholds and overcharge temperature thresholds for each similar data cluster; Threshold adjustment unit: It is used to update the threshold ranges of the overcharge voltage thresholds and overcharge temperature thresholds obtained in the initial threshold determination unit.

3. The lithium battery overcharge warning system based on big data according to claim 2, wherein In the initial threshold determination unit, calculate the threshold ranges of the overcharge voltage thresholds and overcharge temperature thresholds for each similar data cluster as follows: When calculating the overcharge voltage threshold or overcharge temperature threshold of a certain similar data cluster, assume that a certain similar data cluster X = {x1, x2,..., x n}, and calculate the local mean μ and local standard deviation σ of the voltage data in the similar data cluster X, or the local mean μ and local standard deviation σ of the temperature data through the sliding window method; x i is the specific value of voltage data or temperature data, where i = 1, 2,..., n; Preset risk ratio p, for each data point x i , centered at μ and with σ as the scale parameter, calculate the probability density value f(x i ); Sort all f(x i ) in ascending order, and take the n×p-th probability density value as ∈; Solve for x where f(x i ) < ∈, that is, the overcharge threshold range; i The range of values of x is the overcharge threshold range; Let f(x i ) = ∈, that is: Transform both sides of the equation: Then, take the natural logarithm on both sides. According to the property of logarithms, ln(exp(a)) = a, we can get: Multiply both sides of the equation by -2σ 2 , we get: Solve for x by taking the square root of both sides of the equation i , resulting in: The lower threshold value is Upper limit threshold is: That is, the threshold range is, If the local mean μ and local standard deviation σ of the voltage data are used, the calculated value is the overcharge voltage threshold; If the local mean μ and local standard deviation σ of the temperature data are used, the calculated value is the overcharge temperature threshold.

4. The lithium battery overcharge warning system based on big data according to claim 3, characterized in that, In the threshold adjustment unit, update the threshold ranges of the overcharge voltage thresholds and overcharge temperature thresholds obtained in the initial threshold determination unit as follows: Obtain the voltage data sequence and temperature data sequence in the historical charging data, and record the corresponding charge-discharge cycle times and calendar time at the same time; For the voltage data and temperature data, use the autoregressive integrated moving average model for modeling to construct a voltage prediction model and a temperature prediction model. The voltage prediction model is used to predict the voltage value, and the temperature prediction model is used to predict the temperature value; Use the voltage prediction model and the temperature prediction model to predict the voltage and temperature in the future unit time to obtain the predicted voltage sequence and the predicted temperature sequence; Calculate the slope ΔV of the predicted value of the predicted voltage sequence slope ; Calculate the slope ΔT of the predicted value of the predicted temperature sequence slope ; Calculate the voltage correction factor γV: Where α is the slope sensitivity coefficient; Calculate the temperature correction factor γT: Where β is the slope sensitivity coefficient; For recording the corresponding charge and discharge cycle times and calendar time, a battery aging model is constructed through a machine learning algorithm to obtain the overcharge voltage threshold aging coefficient k V and the overcharge temperature threshold aging coefficient k T ; According to the overcharge voltage threshold and overcharge temperature threshold obtained by the initial threshold determination unit; Set the overcharge voltage threshold to Set the overcharge temperature threshold to In the future unit time, update the overcharge voltage threshold and overcharge temperature threshold to:

5. The overcharge warning system for lithium batteries based on big data according to claim 1, characterized in that, The risk analysis module includes: State of charge calculation unit: Obtain the charging data and calculate the battery SOC in real time based on the ampere-hour integration method and the open-circuit voltage method; Fault level classification unit: Obtain the SOC obtained by the state of charge calculation unit, the overcharge voltage threshold and overcharge temperature threshold obtained by the dynamic threshold generation module, and the real-time charging data obtained by the data acquisition module, and construct an overcharge risk level model by using the fuzzy logic algorithm; Real-time warning unit: Trigger multi-level warnings through the risk level output by the overcharge risk level model in the fault level classification unit, and push them to the user terminal through the cloud platform.

6. The lithium battery overcharge warning system based on big data according to claim 1, characterized in that, The data processing module includes: Missing value filling subunit: Used to fill in missing data through the time series correlation analysis algorithm; Outlier filtering subunit: Used to identify and correct outliers through numerical limit analysis and dynamic standard deviation method.

7. The overcharge warning system for lithium batteries based on big data according to claim 1, characterized in that, It also includes: Visualization interface module: Used to display in real time the charging data during the lithium battery charging process, the overcharge voltage threshold and overcharge temperature threshold generated by the dynamic threshold generation module, and the risk level output by the overcharge risk level model in the risk analysis module.

8. Using the big data-based lithium battery overcharge warning method according to any one of claims 1-7, characterized in that It includes the following steps: S1. Obtain the charging data during the lithium battery charging process in real time, and record the corresponding charge and discharge cycle times and calendar time at the same time; The charging data includes battery type, charging stage data, and usage environment data; S2. Preprocess the charging data; S3. Classify the historical charging data according to the battery type, charging stage data, and usage environment data, calculate the local mean and standard deviation of the data by using the sliding window method, calculate the similarity between the data by using the dynamic time warping algorithm to construct similar data clusters, and obtain the threshold ranges of the overcharge voltage threshold and overcharge temperature threshold for each similar data cluster through calculation; By obtaining the historical voltage and temperature data sequences and the corresponding charge and discharge cycle times and calendar time, use the autoregressive integrated moving average model to construct a voltage prediction model and a temperature prediction model, predict the voltage sequence and temperature sequence predicted in the future unit time, calculate the slopes of the predicted voltage sequence and predicted temperature sequence to obtain the voltage correction factor and temperature correction factor, and construct a battery aging model through machine learning algorithm to obtain the overcharge voltage threshold aging coefficient and overcharge temperature threshold aging coefficient; In the future unit time, update the overcharge voltage threshold and overcharge temperature threshold to: Updated threshold of overcharge voltage threshold = overcharge voltage threshold × overcharge voltage threshold aging coefficient × voltage correction factor; Updated threshold of overcharge temperature threshold = overcharge temperature threshold × overcharge temperature threshold aging coefficient × temperature correction factor; S4. Generate an overcharge risk level model, calculate the battery SOC in real time based on the ampere-hour integration method and open-circuit voltage method, obtain the SOC, the overcharge voltage threshold and overcharge temperature threshold obtained by the dynamic threshold generation module, and the real-time charging data, construct an overcharge risk level model by using the fuzzy logic algorithm, trigger multi-level warnings through the risk level output by the overcharge risk level model, and push them to the user terminal through the cloud platform.

Images