Safety detection method and device for lithium battery

Abstract

The application discloses a kind of safety detection method and device of lithium battery.The method comprises: by a plurality of gas sensors, target gas data of target lithium battery is collected at target time;Determine the gas risk index corresponding to the target lithium battery based on the target gas data collected by a plurality of gas sensors;In the case where the gas risk index is greater than the target risk index threshold, generate the target prompt information corresponding to the target lithium battery, and display the target prompt information.Improve the accuracy of lithium battery safety detection.

Description

Technical Field

[0001] This invention relates to the field of lithium-ion battery testing technology, and in particular to a method and apparatus for safety testing of lithium batteries. Background Technology

[0002] With the rapid development of the economy and society, energy storage power stations, as an important part of the power system, have received increasing attention for their safety and reliability. Therefore, improving the sensitivity and accuracy of fire detection systems for energy storage power stations is particularly important.

[0003] Furthermore, lithium-ion batteries in energy storage power stations need to operate stably for extended periods. However, due to natural aging, overcharging and discharging, and the influence of external environmental factors, thermal runaway may occur when lithium-ion batteries operate for extended periods or are subjected to external impacts. Thermal runaway in lithium-ion batteries not only leads to severe performance degradation but can also cause serious consequences such as fires, resulting in significant property damage and safety hazards. However, existing battery safety detection methods often suffer from problems such as untimely alarms, low sensitivity, false alarms, and missed alarms, failing to guarantee the safety and stability of energy storage power station operation. Summary of the Invention

[0004] This invention provides a method and apparatus for the safety detection of lithium batteries, which solves the problems of untimely alarms, low sensitivity, false alarms and missed alarms that often exist in existing battery safety detection methods, which cannot guarantee the safety and stability of energy storage power station operation.

[0005] According to one aspect of the present invention, a safety testing method for lithium batteries is provided, the method comprising:

[0006] Data on the target gas emitted from the target lithium battery at the target time is collected using multiple gas sensors.

[0007] Based on the target gas data collected by multiple gas sensors, a gas risk index corresponding to the target lithium battery is determined;

[0008] If the gas risk index is greater than the target risk index threshold, a target prompt message corresponding to the target lithium battery is generated and displayed.

[0009] According to another aspect of the present invention, a safety detection device for a lithium battery is provided, the device comprising:

[0010] The gas data acquisition module is used to collect target gas data emitted by the target lithium battery at the target time through multiple gas sensors.

[0011] The risk indicator determination module is used to determine the gas risk indicator corresponding to the target lithium battery based on the target gas data collected by multiple gas sensors.

[0012] The information prompting module is used to generate and display target prompting information corresponding to the target lithium battery when the gas risk index is greater than the target risk index threshold.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] At least one processor; and

[0015] A memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the lithium battery safety detection method according to any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the lithium battery safety detection method according to any embodiment of the present invention.

[0018] The technical solution of this invention collects target gas data emitted by a target lithium battery at a target time using multiple gas sensors; it performs high-precision synchronous acquisition of the target gas data released by the target lithium battery at the target time. Based on the target gas data collected by multiple gas sensors, a gas risk index corresponding to the target lithium battery is determined; the gas risk index corresponding to the target lithium battery is accurately determined; finally, if the gas risk index is greater than the target risk index threshold, a target prompt message corresponding to the target lithium battery is generated and displayed. This solves the problems of untimely alarms, low sensitivity, false alarms, and missed alarms that often exist in existing battery safety detection methods, which cannot guarantee the safety and stability of energy storage power station operation. It achieves the beneficial effect of improving the accuracy of lithium battery safety detection and ensuring the safety and stability of the energy storage power station where the lithium battery is located.

[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a safety testing method for lithium batteries according to Embodiment 1 of the present invention;

[0022] Figure 2 This is a flowchart of a safety testing method for lithium batteries according to Embodiment 2 of the present invention;

[0023] Figure 3 This is a schematic diagram of the structure of a lithium battery safety detection device according to Embodiment 3 of the present invention;

[0024] Figure 4 This is a schematic diagram of the structure of an electronic device that implements the safety detection method for lithium batteries according to embodiments of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] Example 1

[0028] Figure 1This is a flowchart of a lithium battery safety testing method provided in Embodiment 1 of the present invention. This embodiment is applicable to the safety testing of lithium batteries. The method can be executed by a lithium battery safety testing device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0029] S110: Collect target gas data emitted by the target lithium battery at the target time using multiple gas sensors.

[0030] The target gas data can be understood as the gas data released during the thermal runaway of lithium ions. The target gas data includes multi-gas composition data and gas concentration data.

[0031] Specifically, based on a preset sampling frequency and sampling time, multiple gas sensors are used to synchronously collect real-time data on various target gases released during the thermal runaway of lithium ions.

[0032] Optionally, after collecting target gas data emitted by the target lithium battery at a target time through multiple gas sensors, the method further includes: preprocessing the target gas data to obtain preprocessed target gas data, wherein the preprocessing includes smoothing and noise reduction.

[0033] Specifically, the target gas data is smoothed using a sliding window averaging method to reduce random fluctuations or noise, resulting in a smoother data curve. Noise reduction processing is then applied to remove noise components from the target gas data, improving the signal-to-noise ratio. This yields the smoothed and noise-reduced target gas data.

[0034] For example, the sliding window averaging method can smooth target gas data, reducing noise and fluctuations. For a window of length N, it can be implemented using the following formula:

[0035]

[0036] Where x(t) represents the target gas data at time t. The target gas data is the smoothed output after being averaged by a sliding window, where N is the sliding window.

[0037] S120. Determine the gas risk index corresponding to the target lithium battery based on the target gas data collected by multiple gas sensors.

[0038] Among them, the gas risk index can be understood as the gas risk value, which is a parameter used to assess gas risk.

[0039] Specifically, the gas risk value corresponding to the target lithium battery is determined by comprehensively analyzing the target gas data collected by multiple gas sensors.

[0040] Optionally, determining the gas risk index corresponding to the target lithium battery based on the gas data collected by multiple gas sensors includes:

[0041] The sensor weight corresponding to each gas sensor is determined, and the gas data collected by the gas sensor is weighted according to the sensor weight. The weighted gas data corresponding to the gas sensor is then summed to obtain the gas risk index corresponding to the target lithium battery.

[0042] Specifically, data from multiple sensors can be combined and a comprehensive judgment achieved through weighted summation. The gas data collected by the gas sensors is weighted according to the sensor weights corresponding to the gas sensors using the following formula, and the weighted gas data corresponding to the gas sensors is then calculated as follows:

[0043]

[0044] Where R(t) is the gas risk index, x i (t) represents the target gas data collected by the i-th gas sensor, ω i The weight of the i-th gas sensor.

[0045] S130. If the gas risk index is greater than the target risk index threshold, generate target prompt information corresponding to the target lithium battery and display the target prompt information.

[0046] The target warning information can be understood as risk warning information for the target lithium battery.

[0047] Specifically, the gas risk index is compared with a set target risk index threshold. If the gas risk index exceeds the target risk index threshold, a target prompt message corresponding to the target lithium battery is generated and displayed on the terminal in a preset prompt mode. The preset prompt mode can be preset based on experience, such as at least one of text prompts, light prompts, and voice prompts. Optionally, if the gas risk index exceeds the target risk index threshold, a risk level corresponding to the gas risk index is determined based on a preset risk level threshold, and different risk prompt messages are generated based on different risk levels. The preset risk level threshold can be preset based on experience, and this embodiment does not limit it. For example, if the gas risk index exceeds the target risk index threshold, the target lithium battery with risk is identified, along with the gas components exceeding the target risk index and their current concentration. Preliminary response suggestions are provided, such as immediate evacuation, stopping related operations, and activating the ventilation system. The time of generating the prompt message is recorded for subsequent tracking and recording.

[0048] Optionally, before generating the target prompt information corresponding to the target lithium battery, the method further includes: determining an initial risk index threshold corresponding to each gas sensor, determining a risk weight corresponding to each initial risk index threshold, and performing a weighted summation of the multiple risk weights to obtain a target risk index threshold.

[0049] The target risk index threshold is obtained by weighted summation of the initial risk index thresholds of each gas sensor after adaptive threshold adjustment.

[0050] Specifically, adaptive threshold adjustment can dynamically adjust the anomaly detection threshold based on the changing trend of the target gas data, which can be achieved using the following formula:

[0051] TH(t+1)=TH(t)+ΔTH;

[0052] Where TH is the threshold, ΔTH is the change in the threshold, and ΔTH is determined by the rate of change of the current data.

[0053] ΔTH=-α×r(t);

[0054] Where α is the sensitivity parameter, and r(t) is the rate of change of the target gas data:

[0055] r(t) = x(t) - x(t-1);

[0056] It is worth noting that the adaptive threshold adjustment can be achieved by calculating the difference r(t) between the target gas data and the historical gas data of the previous moment, and adjusting the threshold according to the magnitude of the difference. If the rate of change is positive, the threshold is decreased to increase sensitivity; if the rate of change is negative, the threshold is increased to reduce false alarms.

[0057] The technical solution of this invention collects target gas data emitted by a target lithium battery at a target time using multiple gas sensors; it performs high-precision synchronous acquisition of the target gas data released by the target lithium battery at the target time. Based on the target gas data collected by multiple gas sensors, a gas risk index corresponding to the target lithium battery is determined; the gas risk index corresponding to the target lithium battery is accurately determined; finally, if the gas risk index is greater than the target risk index threshold, a target prompt message corresponding to the target lithium battery is generated and displayed. This solves the problems of untimely alarms, low sensitivity, false alarms, and missed alarms that often exist in existing battery safety detection methods, which cannot guarantee the safety and stability of energy storage power station operation. It achieves the beneficial effect of improving the accuracy of lithium battery safety detection and ensuring the safety and stability of the energy storage power station where the lithium battery is located.

[0058] Example 2

[0059] Figure 2 This is a flowchart of a safety testing method for lithium batteries provided in Embodiment 2 of the present invention. This embodiment is a further optimization of the above embodiment. Optionally, the method further includes: determining first concentration prediction data corresponding to the target gas data based on a multivariate regression analysis method; determining second concentration prediction data corresponding to the target gas data based on a gas concentration prediction model, wherein the gas concentration prediction model is obtained by training a pre-constructed decision tree model; determining target concentration prediction data for the target lithium battery based on the first concentration prediction data and the second concentration prediction data; and updating the target risk index threshold based on the target concentration prediction data to obtain an updated target risk index threshold.

[0060] like Figure 2 As shown, the method includes:

[0061] S210: Collect target gas data emitted by the target lithium battery at the target time using multiple gas sensors.

[0062] The target gas data is dimensionality reduced, and the gas composition data is extracted from the dimensionality-reduced target gas data through principal component analysis.

[0063] For example, construct a data matrix X, and calculate the covariance matrix ∑ in the data matrix:

[0064]

[0065] Performing eigenvalue decomposition on the covariance matrix ∑ yields an eigenvector matrix W, where each column of W is an eigenvector corresponding to an eigenvalue, satisfying:

[0066] ∑W=λW;

[0067] Here, λ is a diagonal matrix of eigenvalues, where the elements are the eigenvalues ​​of the covariance matrix ∑. Next, the original data is projected onto the eigenvector matrix W to obtain the gas composition data:

[0068] Z = XW;

[0069] Where Z represents the gas composition data, and X represents the original data matrix.

[0070] S220. Determine the gas risk index corresponding to the target lithium battery based on the target gas data collected by multiple gas sensors.

[0071] S230. If the gas risk index is greater than the target risk index threshold, generate target prompt information corresponding to the target lithium battery and display the target prompt information.

[0072] S240. Based on multivariate regression analysis, determine the first concentration prediction data corresponding to the target gas data, and based on the gas concentration prediction model, determine the second concentration prediction data corresponding to the target gas data, wherein the gas concentration prediction model is obtained by training a pre-constructed decision tree model.

[0073] The first concentration prediction data can be understood as the concentration prediction data of the regression model, and the second concentration prediction data can be understood as the concentration prediction data of the decision tree model.

[0074] Specifically, through multivariate regression analysis, based on gas composition data from multiple sensors, the changing trend of gas concentration data is predicted.

[0075] Optionally, the target gas data includes gas composition data, and the first concentration prediction data corresponding to the target gas data is determined by the following formula:

[0076] γ=β0+β1x1+β2x2+···+β n x n ;

[0077] Where γ represents the first concentration prediction data, x1, x2, ..., x n This represents gas composition data, β1, β2, ..., β n This represents the regression coefficient.

[0078] The regression coefficients were estimated from historical gas data using methods such as the least squares method. The least squares function is:

[0079]

[0080] Among them, y i The i-th historical gas data, is the predicted value of the i-th historical gas data, m is the number of historical gas data, and S is the sum of squared residuals.

[0081] Optionally, determining the second concentration prediction data corresponding to the target gas data based on a pre-trained concentration prediction model includes:

[0082] The target gas data is input into a pre-trained gas concentration prediction model, and the second gas concentration prediction data corresponding to the target gas data is determined based on the model output.

[0083] Optionally, before inputting the gas data to be detected into the pre-trained gas concentration prediction model, the method further includes: constructing a dataset based on a preset number of sample gas data, and dividing the dataset into a training set and a test set; training the constructed initial decision tree model based on the training set, calculating the entropy corresponding to the training set at the root node of the initial decision tree, and determining the gas data to be split and the split point based on at least one entropy; dividing the training set into multiple training subsets according to the gas data to be split and the split point, and repeatedly selecting a split point for each training subset to split into new subsets; and testing the decision tree model that has stopped splitting based on the test set when a preset stopping condition is met to obtain a model performance index, and determining the decision tree model that has stopped splitting as the gas concentration prediction model when the model performance index meets the preset model performance index.

[0084] Specifically, a decision tree model is trained using gas composition data obtained from principal component analysis alongside multivariate regression analysis to supplement the regression model and handle more complex relationships in the data. The decision tree model classifies the data using a series of splitting rules, selecting splitting nodes based on data characteristics, constructing a binary tree, and ultimately outputting the classification results. A training set is constructed using sample gas data, and the dataset is divided into training and test sets. The decision tree model is trained based on the training set. The sample gas data from the training set is placed at the root node, and a decision tree is constructed. At the root node, the entropy, conditional entropy, or Gini index corresponding to the training set is calculated, and the feature with the largest information gain or the largest decrease in the Gini index is selected for splitting. Based on the selected features and splitting points, the dataset is divided into several subsets, and the splitting point is repeatedly selected for each subset to create new subsets. Splitting stops when all samples in a node belong to the same category, the maximum tree depth is reached, or the information gain is less than a set threshold.

[0085] At each node, the splitting feature can be selected according to the following formula:

[0086]

[0087] Where S is the training set, S i It is the subset after splitting.

[0088] Optionally, the entropy corresponding to the training set can be calculated using the following formula:

[0089]

[0090] Where H(D) represents the entropy of the training set D, p k This represents the probability that the data belongs to the k-th data category in the training set, where K represents the total number of data categories.

[0091] For example, given feature A, the conditional entropy H(D|A) of feature A can be obtained by the following formula:

[0092]

[0093] D v It is a subset of data partitioned according to the value v of feature A, |D v | represents subset D v The size of |D| is the size of the original dataset.

[0094] Furthermore, the information gain can be obtained from the following formula:

[0095] IG(D,A)=H(D)-H(D|A);

[0096] IG(D,A) represents the information gain of feature A on dataset D, which means that the feature with the maximum information gain can be selected as the splitting feature of the current node.

[0097] S250. Based on the first concentration prediction data and the second concentration prediction data, determine the target concentration prediction data for the target lithium battery, and update the target risk index threshold based on the target concentration prediction data to obtain the updated target risk index threshold.

[0098] Among them, the target concentration prediction data can be used to interpret the predicted trend of gas concentration changes.

[0099] Specifically, predicting future trends in gas concentration can help optimize thresholds. For example, by predicting the trend of gas concentration changes, if the predicted data indicates that the gas concentration will steadily increase, the system will automatically lower the threshold to increase the sensitivity to gas concentration changes. If the prediction indicates that the concentration change tends to stabilize or decrease, the threshold will be appropriately raised to avoid false alarms.

[0100] In this embodiment of the invention, the target concentration prediction data for the target lithium battery is determined by combining the first concentration prediction data and the second concentration prediction data. This dual verification method enhances the reliability of the prediction results. The updated target risk index threshold is based on more accurate concentration prediction data, thus more accurately reflecting the risk status of the lithium battery and providing more effective guidance for risk management and control.

[0101] The technical solution of this invention determines first concentration prediction data corresponding to the target gas data based on multivariate regression analysis, and determines second concentration prediction data corresponding to the target gas data based on a gas concentration prediction model. The gas concentration prediction model is obtained by training a pre-constructed decision tree model. By comprehensively considering the influence of multiple variables on the target gas concentration, the first concentration prediction data is predicted more accurately. This method can capture the complex relationships between variables and improve prediction accuracy. Then, based on the first and second concentration prediction data, target concentration prediction data corresponding to the target lithium battery is determined. The target risk index threshold is updated based on the target concentration prediction data to obtain the updated target risk index threshold. The gas concentration prediction model can further refine the prediction, utilizing the branching structure of the decision tree to make accurate predictions based on different conditions. This helps to capture more details and further improve the accuracy of the prediction.

[0102] As an optional example of Embodiment 1 of the present invention, the lithium battery safety detection system of this embodiment specifically includes a main control module, a multi-parameter gas detection module, and a local prompting module.

[0103] The main control module connects to the multi-parameter gas detection module through its built-in ADC (Analog to Digital Converter) interface. Based on the target gas data collected by multiple gas sensors, it determines the gas risk index corresponding to the target lithium battery. If the gas risk index is greater than the target risk index threshold, it generates target prompt information and issues a warning based on the target prompt information through the local prompt module.

[0104] The multi-parameter gas detection module collects target gas data emitted from the target lithium battery at the target time using multiple gas sensors.

[0105] The main control module processes the collected target gas data locally. The algorithms for local processing can be sliding window averaging, adaptive threshold adjustment, and multi-parameter collaborative judgment.

[0106] For example, the sliding window averaging method can smooth target gas data, reducing noise and fluctuations. For a window of length N, it can be implemented using the following formula:

[0107]

[0108] Where x(t) represents the target gas data at time t. The target gas data is the smoothed output after being averaged by a sliding window, where N is the sliding window.

[0109] Specifically, adaptive threshold adjustment can dynamically adjust the anomaly detection threshold based on the changing trend of the target gas data, which can be achieved using the following formula:

[0110] TH(t+1)=TH(t)+ΔTH;

[0111] Where TH is the threshold, ΔTH is the change in the threshold, and ΔTH is determined by the rate of change of the current data.

[0112] ΔTH=-α×r(t);

[0113] Where α is the sensitivity parameter, and r(t) is the rate of change of the target gas data:

[0114] r(t) = x(t) - x(t-1);

[0115] It is worth noting that the adaptive threshold adjustment can be achieved by calculating the difference r(t) between the target gas data and the historical gas data of the previous moment, and adjusting the threshold according to the magnitude of the difference. If the rate of change is positive, the threshold is decreased to increase sensitivity; if the rate of change is negative, the threshold is increased to reduce false alarms.

[0116] Specifically, data from multiple sensors can be combined and a comprehensive judgment achieved through weighted summation. The gas data collected by the gas sensors is weighted according to the sensor weights corresponding to the gas sensors using the following formula, and the weighted gas data corresponding to the gas sensors is then calculated as follows:

[0117]

[0118] Where R(t) is the gas risk index, x i (t) represents the target gas data collected by the i-th gas sensor, ω i The weight of the i-th gas sensor.

[0119] Assign a weight ω to each sensor in the multi-parameter gas detection module. i The system calculates the weighted sum R(t) and then determines whether the weighted sum exceeds the comprehensive threshold. If it does, an alarm is triggered through the local alarm module to manually intervene in the lithium-ion battery fault.

[0120] For example, in a lithium-ion battery monitoring scenario for an energy storage power station, N lithium-ion battery thermal runaway gas detection modules include lithium-ion battery thermal runaway gas detection module 1, lithium-ion battery thermal runaway gas detection module 2, ..., lithium-ion battery thermal runaway gas detection module N-1, lithium-ion battery thermal runaway gas detection module N, etc., where N is greater than or equal to 4. These N lithium-ion battery thermal runaway gas detection modules need to be placed at different monitoring points within the energy storage power station to enable distributed monitoring of the status of all lithium-ion batteries in the power station.

[0121] For example, the target gas data received from multiple gas sensors is subjected to dimensionality reduction processing, and the main features in the data are extracted through principal component analysis.

[0122] For example, construct a data matrix X, and calculate the covariance matrix ∑ in the data matrix:

[0123]

[0124] Performing eigenvalue decomposition on the covariance matrix ∑ yields an eigenvector matrix W, where each column of W is an eigenvector corresponding to an eigenvalue, satisfying:

[0125] ∑W=λW;

[0126] Here, λ is a diagonal matrix of eigenvalues, where the elements are the eigenvalues ​​of the covariance matrix ∑. Next, the original data is projected onto the eigenvector matrix W to obtain the gas composition data:

[0127] Z = XW;

[0128] Where Z represents the gas composition data, and X represents the original data matrix.

[0129] Optionally, a first concentration prediction data corresponding to the target gas data is determined based on a multivariate regression analysis method, and a second concentration prediction data corresponding to the target gas data is determined based on a gas concentration prediction model, wherein the gas concentration prediction model is obtained by training a pre-constructed decision tree model; target concentration prediction data corresponding to the target lithium battery is determined based on the first concentration prediction data and the second concentration prediction data, and the target risk index threshold is updated based on the target concentration prediction data to obtain the updated target risk index threshold.

[0130] The target gas data includes gas composition data, and the first concentration prediction data corresponding to the target gas data is determined by the following formula:

[0131] γ=β0+β1x1+β2x2+···+β n x n ;

[0132] Where γ represents the first concentration prediction data, x1, x2, ..., x n This represents gas composition data, β1, β2, ..., β n This represents the regression coefficient.

[0133] The regression coefficients were estimated from historical gas data using methods such as the least squares method. The least squares function is:

[0134]

[0135] Among them, y i The i-th historical gas data, is the predicted value of the i-th historical gas data, m is the number of historical gas data, and S is the sum of squared residuals.

[0136] Specifically, a decision tree model is trained using gas composition data obtained from principal component analysis alongside multivariate regression analysis to supplement the regression model and handle more complex relationships in the data. The decision tree model classifies the data using a series of splitting rules, selecting splitting nodes based on data characteristics, constructing a binary tree, and ultimately outputting the classification results. A training set is constructed using sample gas data, and the dataset is divided into training and test sets. The decision tree model is trained based on the training set. The sample gas data from the training set is placed at the root node, and a decision tree is constructed. At the root node, the entropy, conditional entropy, or Gini index corresponding to the training set is calculated, and the feature with the largest information gain or the largest decrease in the Gini index is selected for splitting. Based on the selected features and splitting points, the dataset is divided into several subsets, and the splitting point is repeatedly selected for each subset to create new subsets. Splitting stops when all samples in a node belong to the same category, the maximum tree depth is reached, or the information gain is less than a set threshold.

[0137] At each node, the splitting feature can be selected according to the following formula:

[0138]

[0139] Where S is the training set, S i It is the subset after splitting.

[0140] Optionally, the entropy corresponding to the training set can be calculated using the following formula:

[0141]

[0142] Where H(D) represents the entropy of the training set D, p k This represents the probability that the data belongs to the k-th data category in the training set, where K represents the total number of data categories.

[0143] For example, given feature A, the conditional entropy H(D|A) of feature A can be obtained by the following formula:

[0144]

[0145] D v It is a subset of data partitioned according to the value v of feature A, |D v | represents subset D v The size of |D| is the size of the original dataset.

[0146] Furthermore, the information gain can be obtained from the following formula:

[0147] IG(D,A)=H(D)-H(D|A);

[0148] IG(D,A) represents the information gain of feature A on dataset D, which means that the feature with the maximum information gain can be selected as the splitting feature of the current node.

[0149] The technical solution of this invention monitors the release of trace amounts of gas generated in the early stages before thermal runaway of lithium-ion batteries, thereby predicting and monitoring potential battery failures as early as possible and ensuring the safe operation of energy storage power stations.

[0150] Example 3

[0151] Figure 3 This is a schematic diagram of a lithium battery safety detection device provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes: a gas data acquisition module 310, a risk indicator determination module 320, and an information prompting module 330.

[0152] The gas data acquisition module 310 is used to collect target gas data emitted by the target lithium battery at a target time through multiple gas sensors; the risk index determination module 320 is used to determine the gas risk index corresponding to the target lithium battery based on the target gas data collected by multiple gas sensors; and the information prompting module 330 is used to generate target prompt information corresponding to the target lithium battery and display the target prompt information when the gas risk index is greater than the target risk index threshold.

[0153] The technical solution of this invention collects target gas data emitted by a target lithium battery at a target time using multiple gas sensors; it performs high-precision synchronous acquisition of the target gas data released by the target lithium battery at the target time. Based on the target gas data collected by multiple gas sensors, a gas risk index corresponding to the target lithium battery is determined; the gas risk index corresponding to the target lithium battery is accurately determined; finally, if the gas risk index is greater than the target risk index threshold, a target prompt message corresponding to the target lithium battery is generated and displayed. This solves the problems of untimely alarms, low sensitivity, false alarms, and missed alarms that often exist in existing battery safety detection methods, which cannot guarantee the safety and stability of energy storage power station operation. It achieves the beneficial effect of improving the accuracy of lithium battery safety detection and ensuring the safety and stability of the energy storage power station where the lithium battery is located.

[0154] Optionally, the risk indicator determination module is specifically used for:

[0155] The sensor weight corresponding to each gas sensor is determined, and the gas data collected by the gas sensor is weighted according to the sensor weight. The weighted gas data corresponding to the gas sensor is then summed to obtain the gas risk index corresponding to the target lithium battery.

[0156] Optionally, the device further includes a risk weight determination module and a target risk threshold determination module.

[0157] The risk weight determination module is used to determine the initial risk index threshold corresponding to each gas sensor and the risk weight corresponding to each initial risk index threshold before generating the target prompt information corresponding to the target lithium battery.

[0158] The target risk threshold determination module is used to perform a weighted summation of multiple risk weights to obtain the target risk index threshold.

[0159] Optionally, the device may further include a concentration prediction module and a threshold adjustment module.

[0160] The concentration prediction module is used to determine the first concentration prediction data corresponding to the target gas data based on the multivariate regression analysis method, and to determine the second concentration prediction data corresponding to the target gas data based on the gas concentration prediction model, wherein the gas concentration prediction model is obtained by training a pre-constructed decision tree model.

[0161] The threshold adjustment module is used to determine the target concentration prediction data of the target lithium battery based on the first concentration prediction data and the second concentration prediction data, and update the target risk index threshold based on the target concentration prediction data to obtain the updated target risk index threshold.

[0162] Optionally, the target gas data includes gas composition data; correspondingly, the concentration prediction module is specifically used to: determine the first concentration prediction data corresponding to the target gas data using the following formula:

[0163] γ=β0+β1x1+β2x2+···+β n x n ;

[0164] Where γ represents the first concentration prediction data, x1, x2, ..., x n This represents gas composition data, β1, β2, ..., β n This represents the regression coefficient.

[0165] Optionally, the concentration prediction module is specifically used for:

[0166] The target gas data is input into a pre-trained gas concentration prediction model, and the second gas concentration prediction data corresponding to the target gas data is determined based on the model output.

[0167] Optionally, the device further includes a dataset partitioning module, a model training module, a subset splitting module, and a model determination module.

[0168] The dataset partitioning module is used to construct a dataset based on a preset number of sample gas data and divide the dataset into a training set and a test set before inputting the gas data to be detected into the pre-trained gas concentration prediction model.

[0169] The model training module is used to train the constructed initial decision tree model based on the training set. At the root node of the initial decision tree, the entropy corresponding to the training set is calculated, and the gas data to be split and the split point are determined based on at least one entropy.

[0170] The subset splitting module is used to divide the training set into multiple training subsets based on the gas data to be split and the splitting point, and to repeatedly select the splitting point for each training subset to split a new subset again.

[0171] The model determination module is used to test the decision tree model that has stopped splitting based on the test set when a preset stopping condition is met, so as to obtain the model performance index. If the model performance index meets the preset model performance index, the decision tree model that has stopped splitting is determined as the gas concentration prediction model.

[0172] Optionally, the model training module is specifically used for:

[0173] The entropy corresponding to the training set is calculated using the following formula:

[0174]

[0175] Where H(D) represents the entropy of the training set D, p k This represents the probability that the data belongs to the k-th data category in the training set, where K represents the total number of data categories.

[0176] Optionally, the device may further include a preprocessing module.

[0177] The preprocessing module is used to preprocess the target gas data after the target gas data emitted by the target lithium battery at the target time is collected by multiple gas sensors to obtain preprocessed target gas data. The preprocessing includes smoothing and noise reduction.

[0178] The lithium battery safety testing device provided in this embodiment of the invention can execute the lithium battery safety testing method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0179] Example 4

[0180] Figure 4 A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0181] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0182] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0183] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the method of lithium battery safety detection.

[0184] In some embodiments, the method for lithium battery safety detection can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the method for lithium battery safety detection described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the method for lithium battery safety detection by any other suitable means (e.g., by means of firmware).

[0185] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0186] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0187] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0188] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0189] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0190] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0191] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and no limitation is imposed herein.

[0192] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A safety detection method of a lithium battery, characterized by, The method comprises: acquiring target gas data emitted by a target lithium battery at a target time through a plurality of gas sensors; determining a gas risk index corresponding to the target lithium battery based on the target gas data acquired by the plurality of gas sensors; generating target prompt information corresponding to the target lithium battery and displaying the target prompt information when the gas risk index is greater than a target risk index threshold value; The method further comprises: determining first concentration prediction data corresponding to the target gas data based on a multivariate regression analysis method, and determining second concentration prediction data corresponding to the target gas data based on a gas concentration prediction model, wherein the gas concentration prediction model is obtained by training a pre-constructed decision tree model; determining target concentration prediction data of the target lithium battery based on the first concentration prediction data and the second concentration prediction data, and updating the target risk index threshold value based on the target concentration prediction data to obtain an updated target risk index threshold value; The determination of the second concentration prediction data corresponding to the target gas data based on the pre-trained concentration prediction model comprises: inputting the target gas data into a pre-trained gas concentration prediction model, and determining the second gas concentration prediction data corresponding to the target gas data based on the model output result; The entropy corresponding to the training set is calculated by the following formula: ； wherein, denotes the entropy of the training set , denotes the probability of a data class belonging to the th class in the training set, denotes the total number of data classes.

2. The method of claim 1, wherein, The determination of the gas risk index corresponding to the target lithium battery based on the gas data acquired by the plurality of gas sensors comprises: determining a sensor weight corresponding to each gas sensor respectively, weighting the gas data acquired by the gas sensor according to the sensor weight corresponding to the gas sensor, and summing the weighted gas data corresponding to the gas sensor to obtain the gas risk index corresponding to the target lithium battery.

3. The method of claim 1, wherein, Before the generation of the target prompt information corresponding to the target lithium battery, the method further comprises: determining an initial risk index threshold value corresponding to each gas sensor respectively, and determining a risk weight corresponding to each initial risk index threshold value; weighting and summing a plurality of risk weights to obtain a target risk index threshold value.

4. The method of claim 1, wherein, The target gas data comprises gas component data, and the first concentration prediction data corresponding to the target gas data is determined by the following formula: ； wherein, represents first concentration prediction data, represents gas composition data, represents regression coefficients.

5. The method of claim 1, wherein, Before inputting the target gas data into the pre-trained gas concentration prediction model, the method further comprises: constructing a data set based on a preset number of sample gas data, and dividing the data set into a training set and a test set; training an initial decision tree model based on the training set, calculating the entropy corresponding to the training set at the root node of the initial decision tree, determining to-be-split gas data and a split point based on at least one entropy at the root node of the initial decision tree; dividing the training set into a plurality of training subsets according to the to-be-split gas data and the split point, and repeatedly selecting a split point for each training subset to split out a new subset again; In the case of meeting the preset stop condition, the decision tree model with stopped splitting is tested based on the test set to obtain a model performance index, and in the case of the model performance index meeting a preset model performance index, the decision tree model with stopped splitting is determined as the gas concentration prediction model.

6. The method of claim 1, wherein, After the target gas data emitted by the target lithium battery at the target time is collected by the plurality of gas sensors, the method further includes: The target gas data is preprocessed to obtain preprocessed target gas data, wherein the preprocessing includes smoothing and noise reduction.

7. A safety detection device for a lithium battery for implementing the method according to any one of claims 1 to 6, characterized in that, It includes: A gas data acquisition module is configured to collect target gas data emitted by a target lithium battery at a target time through a plurality of gas sensors. A risk index determination module is configured to determine a gas risk index corresponding to the target lithium battery based on the target gas data collected by the plurality of gas sensors. An information prompting module is configured to generate target prompt information corresponding to the target lithium battery and display the target prompt information in the case that the gas risk index is greater than a target risk index threshold.

Images