Lithium battery parallel arc detection method, device and storage medium

Abstract

The present invention discloses a lithium battery parallel arc detection method, device, and storage medium, belonging to the field of battery technology. The method comprises: obtaining current data from a target detection point in a lithium battery circuit, dividing the current data into multiple data sets according to a preset period; determining the time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics corresponding to the data set based on the current data in the data set, wherein the time-frequency combination characteristics include a combined time domain characteristic and frequency domain characteristic; inputting the time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics into a parallel arc detection model, and determining the parallel arc detection results for the corresponding time period of the data set based on the output of the parallel arc detection model. The present invention aims to improve the accuracy of parallel arc detection and enhance the safety of lithium battery systems.

Description

Technical Field

[0001] The present invention relates to the field of battery technology, and in particular to a method, device and storage medium for detecting parallel arcs in lithium batteries. Background Art

[0002] Thermal runaway is one of the most serious safety hazards in lithium-ion battery systems, and current anomalies caused by parallel arcs are a potential cause of thermal runaway in lithium-ion batteries. When a parallel arc occurs, it releases a large amount of heat and sparks, potentially igniting surrounding combustibles and causing a fire. Furthermore, because parallel arcs have more complex electrical characteristics and signal forms than series arcs, current arc detection methods generally rely on the characteristics of series arcs. This lacks accuracy for parallel arc detection, leading to safety hazards in lithium-ion batteries caused by thermal runaway due to parallel arcs. Summary of the Invention

[0003] The present invention provides a method, device and storage medium for detecting parallel arcs in lithium batteries to solve one or more technical problems existing in the prior art and at least provide a beneficial option or create conditions.

[0004] The present invention provides a method for detecting arcs in parallel with lithium batteries, the method comprising:

[0005] Acquire current data of a target detection point in a lithium battery circuit, and divide the current data into multiple data sets according to a preset period;

[0006] Determining, based on the current data of the data set, a time domain feature, a frequency domain feature, and a time-frequency combination feature corresponding to the data set, wherein the time-frequency combination feature includes the combined time domain feature and the frequency domain feature;

[0007] The time domain features, the frequency domain features and the time-frequency combination features are input into a parallel arc detection model, and a parallel arc detection result of a time period corresponding to the data set is determined according to an output result of the parallel arc detection model.

[0008] Optionally, the lithium battery parallel arc detection method further includes:

[0009] Acquiring current data of a preset detection point in an experimental platform, dividing the current data into a plurality of training data sets according to the preset period, the experimental platform being used to controllably form a parallel arc;

[0010] The preset model is trained using the training data set to obtain the parallel arc detection model.

[0011] Optionally, the parallel arc detection model is a decision tree model generated based on a CART algorithm, and the step of training a preset model using the training data set includes:

[0012] Dividing the training data set into a training set and a test set;

[0013] Optimizing the Gini index equation of the CART algorithm through the training set and the test set;

[0014] Training the preset model using the training set and the optimized Gini index equation;

[0015] The prediction accuracy of the preset model is evaluated using the test set, and when the prediction accuracy is greater than or equal to the preset accuracy, it is determined that the training of the preset model is completed.

[0016] Optionally, the step of optimizing the Gini index equation of the CART algorithm using the training set and the test set includes:

[0017] Calculating the Euclidean distance between each of the training sets and each of the test sets;

[0018] For any of the test sets, determining a corresponding training set having the smallest Euclidean distance with the test set as a target training set, and associating the test set with the target training set;

[0019] For any of the training sets, determining a weight parameter of the training set according to the number of the test sets associated with the training set;

[0020] The probability parameter of the Gini index equation is replaced by a preset parameter, where the preset parameter is the ratio of the weight parameter of each training set to the weight parameters of all training sets.

[0021] Optionally, the step of associating the test set with the target training set includes:

[0022] When the Euclidean distance between the test set and the target training set is less than or equal to a preset distance, associating the test set with the target training set;

[0023] When the Euclidean distance between the test set and the target training set is greater than the preset distance, the test set and the target training set are not associated.

[0024] Optionally, the step of determining the time domain features corresponding to the data set according to the current data of the data set includes:

[0025] Calculate the time domain mean, time domain median, time domain variance, time domain root mean square, time domain integral, time domain kurtosis, time domain entropy and time domain peak-to-peak value of the current data of the data set, wherein the time domain features include the time domain mean, the time domain median, the time domain variance, the time domain root mean square, the time domain integral, the time domain kurtosis, the time domain entropy and the time domain peak-to-peak value.

[0026] Optionally, the step of determining the frequency domain features corresponding to the data set according to the current data of the data set includes:

[0027] Performing Fourier transform on the current data of the data set to obtain frequency domain data corresponding to the data set;

[0028] Calculate the frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy and frequency domain peak-to-peak value of the frequency domain data, where the frequency domain features include the frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy and frequency domain peak-to-peak value.

[0029] Optionally, the step of determining the time-frequency combination features corresponding to the data set according to the current data of the data set includes:

[0030] Combining the time domain peak-to-peak value and the frequency domain entropy into a first feature;

[0031] combining the time domain average value and the frequency domain kurtosis as a second feature;

[0032] Combining the time domain median and the frequency domain integral into a third feature;

[0033] Combining the time domain root mean square and the frequency domain variance into a fourth feature;

[0034] Combining the time domain variance and the frequency domain root mean square as a fifth feature;

[0035] Combining the time domain integral and the frequency domain median as a sixth feature;

[0036] combining the time domain kurtosis and the frequency domain average value as a seventh feature;

[0037] Combining the time domain entropy and the frequency domain peak-to-peak value into an eighth feature;

[0038] The time-frequency combination features include the first feature, the second feature, the third feature, the fourth feature, the fifth feature, the sixth feature, the seventh feature and the eighth feature.

[0039] In addition, in order to achieve the above-mentioned purpose, the present application also proposes a lithium battery parallel arc detection device, which includes: a memory, a processor, and a lithium battery parallel arc detection program stored in the memory and runnable on the processor, and the lithium battery parallel arc detection program is configured to implement the steps of the lithium battery parallel arc detection method as described in any of the above items.

[0040] In addition, in order to achieve the above-mentioned purpose, the present application also proposes a storage medium, on which a lithium battery parallel arc detection program is stored. When the lithium battery parallel arc detection program is executed by a processor, the steps of the lithium battery parallel arc detection method as described in any of the above items are implemented.

[0041] The present invention has at least the following beneficial effects: by collecting current data from target detection points in a lithium battery circuit, the lithium battery circuit is monitored, the current data is divided into multiple data sets based on a preset period, the current data corresponding to each data set represents the operating status of the lithium battery circuit during that period, and the time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics obtained by combining the time domain characteristics and frequency domain characteristics are calculated based on the current data of the data set, thereby obtaining characteristic data that can reflect the operating status of the lithium battery circuit in the time domain and frequency domain, respectively. The time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics are then input into a parallel arc detection model, and the output of the model is used to determine whether a parallel arc occurs in the lithium battery circuit during the time period corresponding to the data set. Compared with current detection solutions based on series arc characteristics, the detection method that combines the circuit's time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics can more effectively target the electrical characteristics and signal form of parallel arcs, improve the accuracy of parallel arc detection, and enhance the safety of lithium battery systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are used to provide a further understanding of the technical solution of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the technical solution of the present invention and do not constitute a limitation to the technical solution of the present invention.

[0043] Figure 1 This is a flow chart of a first embodiment of a method for detecting arcs in parallel connected lithium batteries according to the present invention;

[0044] Figure 2 This is a flow chart of a second embodiment of a method for detecting arcs in parallel connected lithium batteries according to the present invention;

[0045] Figure 3 A schematic diagram of an experimental platform for a second embodiment of a parallel arc detection method for lithium batteries according to the present invention;

[0046] Figure 4 Schematic diagram of the training set and the test set in the second embodiment of the method for detecting arc in parallel of lithium batteries of the present invention;

[0047] Figure 5 A waveform diagram of time domain characteristics in the third embodiment of the method for detecting arcs in parallel with lithium batteries of the present invention;

[0048] Figure 6 Another waveform diagram of the time domain characteristics of the third embodiment of the parallel arc detection method for lithium batteries of the present invention;

[0049] Figure 7 This is a schematic diagram of the device structure involved in the operation of an embodiment of the lithium battery parallel arc detection device of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] It should be noted that although the system diagrams illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the system or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0052] The embodiment of the present invention provides a method for detecting arc in parallel of lithium batteries. Figure 1 , a first embodiment of the lithium battery parallel arc detection method of the present application is proposed. In this embodiment, the lithium battery parallel arc detection method includes:

[0053] Step S10: acquiring current data of a target detection point in a lithium battery circuit, and dividing the current data into a plurality of data sets according to a preset period.

[0054] Specifically, the target detection point is the current sampling resistor of the battery module in the lithium battery circuit, and the current data obtained is the current amplitude passing through the current sampling resistor. In other embodiments, the current data of the battery module can also be collected based on other components. The current data is sampled at a frequency of 250kHz. In this embodiment, the preset period is 0.8ms, so multiple data can be collected in each preset period. The current data is divided according to the preset period, and the current data collected in a single preset period forms the period, that is, the data set corresponding to the time period. In other embodiments, the sampling frequency of the current data and the preset period for dividing the data set can also be set to other parameters, but it is necessary to ensure that multiple data can be sampled in each preset period.

[0055] Step S20 , determining the time domain features, frequency domain features and time-frequency combination features corresponding to the data set according to the current data in the data set, where the time-frequency combination features include the combined time domain features and frequency domain features.

[0056] In order to more effectively target the electrical characteristics and signal forms of parallel arcs, this application analyzes them by combining time domain features, frequency domain features, and time-frequency combination features. The time domain features can understand the state of the current signal changing over time, the frequency domain features can understand the intensity of different frequency components in the current signal, and the time-frequency combination features are formed by combining the time domain features and the frequency domain features, which can reflect the correlation between individual time domain features and individual frequency domain features. For each data set, the time domain features, frequency domain features, and time-frequency combination features of the corresponding current data are calculated separately.

[0057] Step S30: input the time domain features, frequency domain features and time-frequency combination features into the parallel arc detection model, and determine the parallel arc detection results of the corresponding time period of the data set according to the output results of the parallel arc detection model.

[0058] The data sequence formed by time domain features, frequency domain features and time-frequency combined features is input into the parallel arc detection model. The parallel arc detection model is a classification model. It classifies the data sequence according to its input and outputs whether the data sequence reflects the characteristics of a parallel arc. Therefore, based on the output results of the parallel arc detection model, it can be determined whether a parallel arc occurs in the corresponding time period of the data set.

[0059] The embodiments of the present invention have at least the following beneficial effects: by collecting current data from target detection points in a lithium battery circuit, the lithium battery circuit is monitored, the current data is divided into multiple data sets according to a preset period, the current data corresponding to each data set represents the operating status of the lithium battery circuit within the period, and the time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics obtained by combining the time domain characteristics and frequency domain characteristics are calculated based on the current data of the data set, thereby obtaining characteristic data that can reflect the operating status of the lithium battery circuit in the time domain and frequency domain respectively. The time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics are then input into a parallel arc detection model, and the output of the model is used to determine whether a parallel arc occurs in the lithium battery circuit within the time period corresponding to the data set. Compared with the current detection scheme based on the characteristics of series arcs, the detection method that combines the time domain characteristics, frequency domain characteristics, and time-frequency combination characteristics of the circuit can more effectively target the electrical characteristics and signal forms of parallel arcs, improve the accuracy of parallel arc detection, and improve the safety of lithium battery systems.

[0060] Further, based on the above embodiment, refer to Figure 2 , a second embodiment of the lithium battery parallel arc detection method of the present application is proposed. In this embodiment, the lithium battery parallel arc detection method also includes:

[0061] Step S100: obtaining current data of a preset detection point in an experimental platform, dividing the current data into a plurality of training data sets according to a preset period, and the experimental platform is used to controllably form a parallel arc.

[0062] Step S110: training a preset model using a training data set to obtain a parallel arc detection model.

[0063] In order to detect parallel arcs through the parallel arc detection model, it is necessary to first train a model with reliable detection accuracy. Therefore, before applying the parallel arc detection model to the lithium battery circuit, an experimental platform for controllable formation of parallel arcs is built. Figure 3 The experimental platform includes a multifunctional DC source, a three-phase inverter for simulating power system loads, and an arc generator. The arc generator includes a stepper motor and arc rods that can adjust the gap through the stepper motor. The arc generator and the limiting resistor After connecting in series, they are connected in parallel with the multi-function DC power supply and the three-phase inverter. The preset detection point is the node next to the multi-function DC power supply, and the output current of the multi-function DC power supply is collected through an oscilloscope. As current data, the sampling frequency is the same as that of the above embodiment. Regarding the controllable formation method of parallel arc, the experimental platform can generate parallel arcs through various specifications, as shown in Table 1. For the various specifications shown in Table 1, any one can be selected to generate parallel arcs.

[0064]

[0065] Table 1

[0066] The collected current data is also divided into multiple training data sets according to the preset period, and the time domain features, frequency domain features and time-frequency combination features corresponding to each training data set are also calculated. The preset model is trained through multiple training data sets to obtain a parallel arc detection model.

[0067] By building an experimental platform for controllable parallel arc formation and training a parallel arc detection model based on the training data set collected by the experimental platform, the parallel arc detection model can be trained based on the accurate electrical characteristics and signal form of the parallel arc, thereby improving the accuracy of the model and thus improving the accuracy of parallel arc detection.

[0068] Furthermore, in this embodiment, the parallel arc detection model is a decision tree model generated based on the CART algorithm, and the steps of training the preset model using the training data set include:

[0069] Divide the training dataset into training and testing sets.

[0070] The Gini index equation of the CART algorithm is optimized using the training set and the test set.

[0071] Train the preset model using the training set and the optimized Gini index equation.

[0072] The prediction accuracy of the preset model is evaluated through the test set. When the prediction accuracy is greater than or equal to the preset accuracy, it is determined that the preset model training is completed.

[0073] The training data set is divided into a training set and a test set, where the number of training sets is larger than the number of test sets.

[0074] Commonly used generation algorithms for decision tree models include ID3, C4.5, and CART. CART can process both nominal attribute data sets and continuous data sets with classification labels to generate decision trees. Therefore, the CART algorithm is selected to generate the decision tree model of this embodiment. The CART algorithm uses the Gini index equation to construct a decision tree. The Gini index equation is specifically:

[0075] in, It represents the probability that rows share the same label, which is defined as a probability parameter in this embodiment.

[0076] The algorithm selects the optimal features to split the training set according to the Gini index equation. This process is recursively performed on the obtained subsets until the stopping criteria are met. The stopping criteria are, for example, the maximum depth of the split node or the minimum number of samples. The decision tree is then pruned to improve the performance of the decision tree by removing branches with lower accuracy, thus completing the construction of the decision tree.

[0077] In addition, to address the problem of CART instability, that is, the problem that slight modifications to the data can lead to significant changes in the final decision tree, this embodiment optimizes the Gini index equation through training sets and test sets before training the preset model to solve this problem and improve the accuracy of decision tree prediction.

[0078] When training the preset model, the prediction accuracy of the preset model is evaluated through the test set. The prediction accuracy is the ratio of the number of test sets predicted correctly to the total number of test sets. When the prediction accuracy is greater than or equal to the preset accuracy, for example, 95%, it is determined that the prediction accuracy of the preset model meets the requirements, the training is completed, and the parallel arc detection model is obtained.

[0079] A parallel arc detection model is obtained by generating a decision tree model based on the CART algorithm, so that the parallel arc detection model can effectively detect whether a parallel arc occurs in the lithium battery circuit based on the characteristics of the current data, thereby improving the accuracy of the parallel arc detection model and thus improving the accuracy of parallel arc detection.

[0080] Furthermore, in this embodiment, the step of optimizing the Gini index equation of the CART algorithm using the training set and the test set includes:

[0081] Compute the Euclidean distance between each training set and each test set.

[0082] For any test set, determine the corresponding training set with the smallest Euclidean distance to the test set as the target training set, and associate the test set and the target training set.

[0083] For any training set, the weight parameter of the training set is determined according to the number of test sets associated with the training set.

[0084] The probability parameter of the Gini index equation is replaced by a preset parameter, which is the ratio of the weight parameter of each training set to the weight parameters of all training sets.

[0085] To address the aforementioned CART instability issue, this application introduces a gradient-based single-sided sampling technique, ignoring most low-gradient data instances and using only the remaining samples to calculate information gain. Specifically, the Euclidean distance between each training set and each test set is calculated, with the size of the distance representing the similarity between the training and test sets.

[0086] refer to Figure 4 (a) state, for any test set ( Figure 4 The double circle in ), there are multiple training sets ( Figure 4 The single circle in the figure calculates the Euclidean distance with the test set, and selects a training set with the smallest Euclidean distance from multiple training sets as the target training set. The target training set has the highest similarity with the test set, and the test set and the target training set are associated. In this way, reference Figure 4 In the (b) state, each test set has a unique corresponding target training set, and at this time, for any training set, there may be zero or several test sets associated with it. Figure 4 (c) state, the weight parameter of the training set is determined according to the number of test sets associated with the training set. In this embodiment, the weight parameter is equal to the number of test sets associated with the corresponding training set. In other embodiments, the weight parameter can also be a parameter positively correlated with the number of associated test sets, which is not limited here. Based on the above steps, the weight parameter of each training set is determined, and the preset parameter is set to , where w is the weight parameter of a single training set, and s is the sum of the weight parameters of all training sets. After replacing the probability parameters with the preset parameters, the optimized Gini index equation is obtained:

[0087]

[0088] By determining the weight of each training set based on the Euclidean distance between the training set and the test set, the training process of the model takes into account the similarity between the training set and the test set. A training set with zero weight means that it does not closely correspond to any test set and can therefore be excluded from the training process. This allows the deletion of training data that may reduce the accuracy of the model, filters out irrelevant data noise, and improves the predictive ability of the decision tree, thereby improving the accuracy of the parallel arc detection model and thus improving the accuracy of parallel arc detection.

[0089] Furthermore, in this embodiment, the step of associating the test set with the target training set includes:

[0090] When the Euclidean distance between the test set and the target training set is less than or equal to the preset distance, the test set and the target training set are associated.

[0091] When the Euclidean distance between the test set and the target training set is greater than the preset distance, the test set and the target training set are not associated.

[0092] In order to avoid the situation where the similarity between the test set and the target training set is still insufficient at the minimum Euclidean distance, a preset distance is set to evaluate the Euclidean distance. When the Euclidean distance is less than or equal to the preset distance, the test set is determined to be similar to the target training set, and the test set and the target training set are associated; when the Euclidean distance is greater than the preset distance, it means that there is actually no training set similar to the test set in the training set, and the test set and the target training set are not associated.

[0093] By setting a preset distance to evaluate the Euclidean distance and then determining the correlation between the test set and the training set, the accuracy of the training set weight parameters is improved, thereby improving the accuracy of the parallel arc detection model and thus improving the accuracy of parallel arc detection.

[0094] Furthermore, based on the above embodiments, a third embodiment of the present invention is proposed for a method for detecting arcs in parallel in lithium batteries. In this embodiment, the step of determining the time domain characteristics corresponding to the data set based on the current data of the data set includes:

[0095] Calculate the time domain mean, time domain median, time domain variance, time domain root mean square, time domain integral, time domain kurtosis, time domain entropy and time domain peak-to-peak value of the current data of the data set. The time domain features include time domain mean, time domain median, time domain variance, time domain root mean square, time domain integral, time domain kurtosis, time domain entropy and time domain peak-to-peak value.

[0096] Specifically, the time domain average value is calculated by the following formula:

[0097]

[0098] The time domain median is calculated as follows:

[0099] If K is an even number,

[0100] If K is an odd number,

[0101] The time domain variance is calculated as follows:

[0102]

[0103] The time domain RMS is calculated as follows:

[0104]

[0105] The time domain integral is calculated as follows:

[0106]

[0107] The time domain kurtosis is calculated as follows:

[0108]

[0109] The time domain entropy is calculated by the following formula:

[0110]

[0111] The peak-to-peak value in the time domain is calculated using the following formula:

[0112]

[0113] In the above calculation formula, K is the number of current data in the data set. Based on the above calculation formula, eight time domain characteristics of the current data of the data set are calculated. Figure 5 and Figure 6 , Figure 5 Schematic diagram of the waveform of the time domain average value (a), time domain median (b), time domain root mean square (c), and time domain peak-to-peak value (d). Figure 6 Schematic diagram of the waveforms of time domain variance (a), time domain integral (b), time domain kurtosis (c), and time domain entropy (d), where Normal state indicates that the circuit is in a normal state, and Arcing state indicates that the circuit is in a state where a parallel arc occurs.

[0114] By calculating the eight time domain characteristics of the current data in the data set, the parallel arc detection model has sufficient data support analysis in the time domain, which is beneficial for the electrical characteristics and signal form of the parallel arc, improving the accuracy of parallel arc detection and improving the safety of the lithium battery system.

[0115] Furthermore, in this embodiment, the step of determining the frequency domain features corresponding to the data set according to the current data of the data set includes:

[0116] Perform Fourier transform on the current data of the data set to obtain the frequency domain data corresponding to the data set.

[0117] Calculate the frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy and frequency domain peak-to-peak value of frequency domain data. Frequency domain features include frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy and frequency domain peak-to-peak value.

[0118] By performing Fourier transform on the current data of the data set, it can be converted into the corresponding frequency domain data. Based on this, the frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy and frequency domain peak-to-peak value of the frequency domain data of the data set are calculated, a total of eight frequency domain features. The calculation formulas can refer to the calculation formulas of the above-mentioned time domain features and will not be repeated here.

[0119] By calculating the eight frequency domain characteristics of the current data in the data set, the parallel arc detection model has sufficient data support analysis in the frequency domain, which is beneficial for the electrical characteristics and signal form of the parallel arc, improving the accuracy of parallel arc detection and enhancing the safety of the lithium battery system.

[0120] Furthermore, in this embodiment, the step of determining the time-frequency combination features corresponding to the data set according to the current data of the data set includes:

[0121] The time domain peak-to-peak value and the frequency domain entropy are combined as the first feature.

[0122] The time domain mean and frequency domain kurtosis are combined as the second feature.

[0123] The time domain median and the frequency domain integral are combined into the third feature.

[0124] The time domain RMS and frequency domain variance are combined into the fourth feature.

[0125] The time domain variance and the frequency domain RMS are combined into the fifth feature.

[0126] The time domain integral and the frequency domain median are combined into the sixth feature.

[0127] The time domain kurtosis and the frequency domain mean are combined as the seventh feature.

[0128] The time domain entropy and the frequency domain peak-to-peak value are combined to form the eighth feature.

[0129] The time-frequency combination features include a first feature, a second feature, a third feature, a fourth feature, a fifth feature, a sixth feature, a seventh feature and an eighth feature.

[0130] In order to explore the correlation between time domain features and frequency domain features, the staff compared the combination effects of various time domain features and various frequency domain features through the control variable method, and finally determined the combination rules as described above. Specifically, the time domain peak-to-peak value is combined with the frequency domain entropy, the time domain average value is combined with the frequency domain kurtosis, the time domain median is combined with the frequency domain integral, the time domain root mean square is combined with the frequency domain variance, the time domain variance is combined with the frequency domain root mean square, the time domain integral is combined with the frequency domain median, the time domain kurtosis is combined with the frequency domain average value, and the time domain entropy is combined with the frequency domain peak-to-peak value, forming the first feature, the second feature, the third feature, the fourth feature, the fifth feature, the sixth feature, the seventh feature and the eighth feature, respectively, a total of eight time-frequency combination features.

[0131] By combining time domain features and frequency domain features to form eight time-frequency combination features, the parallel arc detection model can combine the correlation between time-frequency features and frequency domain features for data analysis, which is beneficial to the electrical characteristics and signal form of parallel arcs, improve the accuracy of parallel arc detection, and improve the safety of lithium battery systems.

[0132] In this embodiment, eight time domain features, eight frequency domain features and eight time-frequency combination features are combined to form a total of 24 input data sequences, which are input into the parallel arc detection model for data analysis, thereby combining the time domain characteristics, frequency domain characteristics and time-frequency correlation of the current data to realize the detection of parallel arcs.

[0133] On the other hand, reference Figure 7 , Figure 7 It is a schematic diagram of the device structure of the lithium battery parallel arc detection equipment.

[0134] A parallel arc detection device for lithium batteries is provided, comprising: a processor and a memory; wherein the memory is configured to store a computer-readable program. When the computer-readable program is executed by the processor, the processor implements the parallel arc detection method for lithium batteries as described in any one of the above technical solutions.

[0135] Those skilled in the art will appreciate that all or some of the steps and systems disclosed above can be implemented as software, firmware, hardware, or any combination thereof. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is well known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVDs) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. As is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0136] In addition, an embodiment of the present invention further provides a storage medium storing a lithium battery parallel arc detection program. When the lithium battery parallel arc detection program is executed by a processor, the relevant steps of any embodiment of the lithium battery parallel arc detection method described above are implemented.

[0137] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0138] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0139] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0140] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0141] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0142] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0143] Although the description of the present application has been quite detailed and specifically describes several embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, but should be considered to provide a broad possible interpretation of these claims by reference to the appended claims, taking into account the prior art, so as to effectively cover the intended scope of the present application. In addition, the above description of the present application is based on the embodiments foreseeable by the inventors, which is intended to provide a useful description, and those non-substantial changes to the present application that have not yet been foreseen may still represent equivalent changes to the present application.

Claims

1. A method for detecting arc in parallel of lithium batteries, characterized in that: The lithium battery parallel arc detection method includes: Acquire current data of a target detection point in a lithium battery circuit, and divide the current data into multiple data sets according to a preset period; Determining, based on the current data of the data set, time domain features, frequency domain features, and time-frequency combination features corresponding to the data set, the time domain features including time domain mean, time domain median, time domain variance, time domain root mean square, time domain integral, time domain kurtosis, time domain entropy, and time domain peak-to-peak value; the frequency domain features including frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy, and frequency domain peak-to-peak value; and the time-frequency combination features including the combined time domain features and the frequency domain features; inputting the time domain features, the frequency domain features, and the time-frequency combination features into a parallel arc detection model, and determining a parallel arc detection result for a time period corresponding to the data set according to an output result of the parallel arc detection model, wherein the parallel arc detection model is a decision tree model generated based on a CART algorithm; The step of determining the time-frequency combination features corresponding to the data set according to the current data of the data set includes: Combining the time domain peak-to-peak value and the frequency domain entropy into a first feature; combining the time domain average value and the frequency domain kurtosis as a second feature; Combining the time domain median and the frequency domain integral into a third feature; Combining the time domain root mean square and the frequency domain variance into a fourth feature; Combining the time domain variance and the frequency domain root mean square into a fifth feature; Combining the time domain integral and the frequency domain median as a sixth feature; combining the time domain kurtosis and the frequency domain average value as a seventh feature; The time domain entropy and the frequency domain peak-to-peak value are combined to form an eighth feature.

2. The lithium battery parallel arc detection method according to claim 1, characterized in that: The lithium battery parallel arc detection method further includes: Acquiring current data of a preset detection point in an experimental platform, dividing the current data into a plurality of training data sets according to the preset period, the experimental platform being used to controllably form a parallel arc; The preset model is trained using the training data set to obtain the parallel arc detection model.

3. The lithium battery parallel arc detection method according to claim 2, characterized in that: The step of training the preset model using the training data set includes: Dividing the training data set into a training set and a test set; Optimizing the Gini index equation of the CART algorithm through the training set and the test set; Training the preset model using the training set and the optimized Gini index equation; The prediction accuracy of the preset model is evaluated using the test set, and when the prediction accuracy is greater than or equal to the preset accuracy, it is determined that the training of the preset model is completed.

4. The lithium battery parallel arc detection method according to claim 3, characterized in that: The step of optimizing the Gini index equation of the CART algorithm through the training set and the test set includes: Calculating the Euclidean distance between each of the training sets and each of the test sets; For any of the test sets, determining a corresponding training set having the smallest Euclidean distance with the test set as a target training set, and associating the test set with the target training set; For any of the training sets, determining a weight parameter of the training set according to the number of the test sets associated with the training set; The probability parameter of the Gini index equation is replaced by a preset parameter, where the preset parameter is the ratio of the weight parameter of each training set to the weight parameters of all training sets.

5. The lithium battery parallel arc detection method according to claim 4, characterized in that: The step of associating the test set with the target training set includes: When the Euclidean distance between the test set and the target training set is less than or equal to a preset distance, associating the test set with the target training set; When the Euclidean distance between the test set and the target training set is greater than the preset distance, the test set and the target training set are not associated.

6. The lithium battery parallel arc detection method according to claim 1, characterized in that: The step of determining the time domain features corresponding to the data set according to the current data of the data set includes: The time domain mean, time domain median, time domain variance, time domain root mean square, time domain integral, time domain kurtosis, time domain entropy and time domain peak-to-peak value of the current data of the data set are calculated.

7. The lithium battery parallel arc detection method according to claim 6, characterized in that: The step of determining the frequency domain features corresponding to the data set according to the current data of the data set includes: Performing Fourier transform on the current data of the data set to obtain frequency domain data corresponding to the data set; The frequency domain mean, frequency domain median, frequency domain variance, frequency domain root mean square, frequency domain integral, frequency domain kurtosis, frequency domain entropy and frequency domain peak-to-peak value of the frequency domain data are calculated.

8. A lithium battery parallel arc detection device, characterized in that: The lithium battery parallel arc detection device includes: a memory, a processor, and a lithium battery parallel arc detection program stored in the memory and executable on the processor. The lithium battery parallel arc detection program is configured to implement the steps of the lithium battery parallel arc detection method according to any one of claims 1 to 7.

9. A storage medium, characterized in that: The storage medium stores a lithium battery parallel arc detection program, which, when executed by a processor, implements the steps of the lithium battery parallel arc detection method according to any one of claims 1 to 7.

Images