Intelligent data analysis technology of lithium battery fire hazard detector

Abstract

The invention provides an intelligent data analysis technology of a lithium battery fire hazard detector, relates to the technical field of lithium battery safety early warning, predicts the occurrence of thermal runaway by detecting an extremely trace amount of characteristic gas separated out from a lithium battery, and provides an intelligent data analysis technology of a lithium battery fire hazard detector through a large number of experimental tests on different brands of battery cells. According to the technology, a lithium battery thermal runaway first-stage characteristic model database is established, an intelligent data analysis algorithm is adopted, detected gas data can be analyzed in real time and compared with the characteristic model database, and therefore whether the lithium battery is in a thermal runaway early stage or not is rapidly judged; the intelligent data analysis algorithm needs to consider various complex conditions and factors, such as ambient temperature, battery aging degree, charge and discharge states and the like, is integrated with a battery management system (BMS), immediately gives an alarm to the BMS once detecting the hidden danger of thermal runaway of the lithium battery, and suggests to take corresponding treatment measures, so that the safety of the lithium battery is ensured. And the efficiency and the automation level of lithium battery fire hazard treatment are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery safety early warning, and in particular to an intelligent data analysis technology for a lithium battery fire hazard detector. Background Art

[0002] There are three stages of battery thermal runaway. Stage one is the reversible stage. The lithium battery's own protection mechanism is activated, and the electrolyte gas is released through the battery's built-in pressure relief valve. At this time, if the lithium battery is intervened or isolated, there is more than a 90% chance of blocking the occurrence of thermal runaway. Stages two and three are irreversible stages. In stage two, the lithium battery's self-protection mechanism can no longer maintain the stable operation of the battery itself, and a large amount of smoke and heat are generated. At this time, the thermal runaway will be irreversible. In stage three, thermal runaway begins to spread rapidly between lithium batteries, causing huge losses and safety hazards.

[0003] The prior art has the following technical problems when used:

[0004] Problem 1: Early warnings for lithium batteries generally only warn of the second and third irreversible stages, but fail to warn of the early reversible stages, resulting in direct damage to most lithium battery fires.

[0005] Problem 2: The detection range and types of lithium batteries are insufficient, resulting in inaccurate fire warning data, which greatly reduces the accuracy of detection and warning;

[0006] Problem three: There is no way to provide timely and rapid feedback on lithium battery fire warning signals, resulting in untimely receipt of warning information, causing the lithium battery fire to be scrapped due to untimely repairs. Summary of the invention

[0007] Technical issues solved

[0008] In view of the shortcomings of the prior art, the present invention provides an intelligent data analysis technology for a lithium battery fire hazard detector, which solves the following problems:

[0009] 1. Failure to provide early warning for lithium batteries, resulting in irreversible damage to lithium batteries due to fire;

[0010] 2. The detection range and data parameter range of lithium battery fire are insufficient, resulting in reduced detection accuracy;

[0011] 3. The lithium battery fire warning signal cannot be received in time, resulting in untimely maintenance.

[0012] Technical Solution

[0013] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent data analysis technology for a lithium battery fire hazard detector, the intelligent data analysis technology comprising the following methods:

[0014] Sp1: Installation of sensor array hardware equipment, using PID principle gas detectors to monitor characteristic gases released by lithium batteries, such as CO, H2, and VOCs, and installing integrated temperature sensors and voltage and current monitoring modules to form a sensor array to obtain gas, temperature, voltage, and current data of lithium batteries in real time;

[0015] Sp2: Installation of the main controller, which uses a high-performance microcontroller and is responsible for receiving sensor data, executing data analysis algorithms, and sending control instructions;

[0016] Sp3: Data acquisition and preprocessing, the sensor array collects the characteristic gas, temperature, voltage and current data of the lithium battery in real time, and the data is transmitted to the main controller through the sensor interface. The main controller filters, amplifies and digitizes the received raw data;

[0017] Sp4: Intelligent data analysis: Through a large number of experimental tests on batteries of different brands, a characteristic model database of the first stage of thermal runaway of lithium batteries is established. The collected data is analyzed in real time using a machine learning algorithm. The algorithm compares the real-time data with the data in the characteristic model database to determine whether the lithium battery is in the early stage of thermal runaway.

[0018] Sp5: Early warning mechanism. According to the data in the characteristic model database, the early warning threshold is set. When the real-time data exceeds the threshold, the early warning mechanism is triggered, and an alarm is issued to the on-duty personnel through various means such as sound and light. At the same time, an alarm signal is sent to the BMS through the BMS interface, and corresponding treatment measures are recommended;

[0019] Sp6: System optimization learning. The system can automatically adjust the thresholds and parameters of detection alarms according to monitoring data and operating status, and can automatically detect the fault status of key components such as sensors and main controllers. When a fault is detected, the fault information is sent through the BMS interface for timely maintenance.

[0020] Preferably, the sensor array communicates with the main controller via wired or wireless means, and the main controller has one or more external design interfaces of UART, I2C, SPI for communicating with other modules.

[0021] Preferably, the system optimization learning supports remote maintenance functions, and remote configuration, data query and troubleshooting operations can be performed through remote communication interfaces such as Ethernet and WiFi.

[0022] Preferably, the database in the intelligent data analysis contains data on characteristic gases released by lithium batteries of various brands and models in the early stages of thermal runaway.

[0023] Preferably, in the data collection and preprocessing, the data is cleaned to remove abnormal values, missing values ​​or duplicate data, the cleaned data is standardized, and the data is converted to a uniform scale using a standardized formula.

[0024] Preferably, the data is subjected to feature extraction before the intelligent data analysis, including features based on gas sensors and features based on physical parameters, specifically including the following contents:

[0025] Gas sensor-based features: Extract key features from the data obtained by the gas sensor, such as gas concentration and rate of change;

[0026] Features based on physical parameters: Extract features such as rate of change and peak value from physical parameters such as temperature, voltage, and current.

[0027] Preferably, in the intelligent data analysis, a suitable machine learning algorithm, such as logistic regression or random forest, is selected according to the characteristics of the problem and the data features, and the selected algorithm is trained using historical data and labels to obtain a model that can identify potential fire hazards of lithium batteries. The performance of the model is evaluated by cross-validation and other methods to ensure the effectiveness and accuracy of the model.

[0028] Preferably, the early warning mechanism uses a trained model to make real-time predictions on new lithium battery data to determine whether it is in the early stages of thermal runaway. Based on the prediction results, if it is determined that the lithium battery has a fire hazard, the early warning mechanism is triggered.

[0029] Beneficial Effects

[0030] The present invention provides an intelligent data analysis technology for a lithium battery fire hazard detector. It has the following beneficial effects:

[0031] 1. The present invention predicts the occurrence of thermal runaway by detecting extremely small amounts of characteristic gases precipitated in lithium batteries, so as to ensure that these characteristic gases can be accurately identified in the early stages of thermal runaway. Through a large number of experimental tests on battery cells of different brands, this technology has established a characteristic model database for the first stage of thermal runaway of lithium batteries. This database contains data on characteristic gases released by lithium batteries of various brands and models in the early stages of thermal runaway, providing an important reference for subsequent very early detection.

[0032] 2. The present invention adopts an intelligent data analysis algorithm, which can analyze the detected gas data in real time and compare it with the characteristic model database, so as to quickly determine whether the lithium battery is in the early stage of thermal runaway. This intelligent data analysis algorithm needs to take into account various complex situations and factors, such as ambient temperature, battery aging, charge and discharge status, etc., to ensure the accuracy and reliability of the analysis.

[0033] 3. The technology used in the present invention can be integrated with the battery management system BMS. Once the potential risk of thermal runaway of lithium batteries is detected, an alarm will be immediately sent to the BMS, and corresponding treatment measures will be recommended. This integration method can greatly improve the efficiency and automation level of lithium battery fire hazard treatment. It has a high degree of real-time performance, can monitor the status of lithium batteries in real time, and immediately issue an early warning when a thermal runaway risk is found. This early warning mechanism can timely discover and deal with potential problems of lithium batteries, avoiding further development of thermal runaway and causing greater losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flow chart of the intelligent data analysis technology monitoring method of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Specific embodiment one:

[0037] like Figure 1 As shown, an intelligent data analysis technology for a lithium battery fire hazard detector includes the following methods:

[0038] Sp1: Installation of sensor array hardware equipment, using PID principle gas detectors to monitor characteristic gases released by lithium batteries, such as CO, H2, and VOCs, and installing integrated temperature sensors and voltage and current monitoring modules to form a sensor array to obtain gas, temperature, voltage, and current data of lithium batteries in real time;

[0039] Sp2: Installation of the main controller, which uses a high-performance microcontroller and is responsible for receiving sensor data, executing data analysis algorithms, and sending control instructions;

[0040] Sp3: Data acquisition and preprocessing, the sensor array collects the characteristic gas, temperature, voltage and current data of the lithium battery in real time, and the data is transmitted to the main controller through the sensor interface. The main controller filters, amplifies and digitizes the received raw data;

[0041] Sp4: Intelligent data analysis. Through a large number of experimental tests on batteries of different brands, a characteristic model database of the first stage of thermal runaway of lithium batteries is established. The database in the intelligent data analysis contains data on characteristic gases released by lithium batteries of various brands and models in the early stage of thermal runaway. The occurrence of thermal runaway is predicted by detecting extremely small amounts of characteristic gases precipitated in lithium batteries to ensure that these characteristic gases can be accurately identified in the early stage of thermal runaway. Through a large number of experimental tests on batteries of different brands, this technology has established a characteristic model database of the first stage of thermal runaway of lithium batteries. This database contains data on characteristic gases released by lithium batteries of various brands and models in the early stage of thermal runaway, which provides an important reference for subsequent very early detection. The machine learning algorithm is used to analyze the collected data in real time. The algorithm compares the real-time data with the data in the characteristic model database to determine whether the lithium battery is in the early stage of thermal runaway. The intelligent data analysis algorithm is used to analyze the detected gas data in real time and compare it with the characteristic model database, so as to quickly determine whether the lithium battery is in the early stage of thermal runaway. This intelligent data analysis algorithm needs to take into account various complex situations and factors, such as ambient temperature, battery aging, charge and discharge status, etc., to ensure the accuracy and reliability of the analysis;

[0042] Sp5: Early warning mechanism. According to the data in the characteristic model database, the early warning threshold is set. When the real-time data exceeds the threshold, the early warning mechanism is triggered, and an alarm is issued to the on-duty personnel through various means such as sound and light. At the same time, an alarm signal is sent to the BMS through the BMS interface, and appropriate treatment measures are recommended. The adopted technology can be integrated with the battery management system BMS. Once the hidden danger of thermal runaway of the lithium battery is detected, an alarm will be immediately issued to the BMS, and appropriate treatment measures will be recommended. This integration method can greatly improve the efficiency and automation level of lithium battery fire hazard handling;

[0043] Sp6: System optimization learning, the system can automatically adjust the threshold and parameters of detection alarm according to monitoring data and operating status, can automatically detect the fault status of key components such as sensors and main controllers, and send fault information through the BMS interface when a fault is detected for timely maintenance. It has a high degree of real-time performance, can monitor the status of lithium batteries in real time, and immediately issue an early warning when a thermal runaway hazard is found. This early warning mechanism can timely discover and deal with potential problems of lithium batteries, avoiding further development of thermal runaway and causing greater losses. Specific embodiment 2:

[0045] like Figure 1 As shown, according to the content of the specific embodiment 1, the following contents are further disclosed:

[0046] The sensor array communicates with the main controller via wired or wireless means. The main controller has one or more external design interfaces such as UART, I2C, and SPI for communicating with other modules. The lithium battery fire hazard detector is usually designed to be directly installed near the lithium battery pack of the energy storage station to ensure that the status of the lithium battery can be monitored in real time and accurately. The sensor array integrates high-precision gas sensors, temperature sensors, voltage and current monitoring modules, etc., which are used to collect various data of the lithium battery in real time. The sensor array is connected to the battery management system (BMS) via wired or wireless means, and uses high-precision sensors to monitor the key parameters of the lithium battery such as temperature, voltage, current, and the precipitated characteristic gases in real time to realize real-time transmission and sharing of data.

[0047] During system optimization and learning, remote maintenance functions are supported. Remote configuration, data query and troubleshooting operations can be performed through remote communication interfaces such as Ethernet and WiFi. The system can be linked with fire-fighting equipment, security monitoring and other systems to achieve information sharing and collaborative operations to ensure that fire accidents are handled in a timely and effective manner. During actual operation, the system can automatically adjust the detection alarm thresholds and parameters based on monitoring data and operating status to improve the accuracy and reliability of the system. At the same time, the system can also record and analyze operating data to provide strong support for optimizing the safety management of energy storage power stations. Specific embodiment three:

[0049] like Figure 1 As shown, according to the content of the specific embodiment 1, the following contents are further disclosed:

[0050] Data preprocessing:

[0051] Data cleaning: remove outliers, missing values ​​or duplicate data.

[0052] Data standardization: converting data to a uniform scale for subsequent processing;

[0053] Feature extraction:

[0054] Gas sensor-based features: Extract key features from the data obtained by the gas sensor, such as gas concentration and rate of change;

[0055] Features based on physical parameters: Extract features such as rate of change and peak value from physical parameters such as temperature, voltage, and current.

[0056] In intelligent data analysis, appropriate machine learning algorithms, such as logistic regression and random forest, are selected according to the characteristics of the problem and data features. The selected algorithms are trained using historical data and labels to obtain a model that can identify potential fire hazards of lithium batteries. The performance of the model is evaluated through cross-validation and other methods to ensure the effectiveness and accuracy of the model.

[0057] The early warning mechanism uses a trained model to make real-time predictions on new lithium battery data to determine whether it is in the early stages of thermal runaway. Based on the prediction results, if the lithium battery is judged to have a fire hazard, the early warning mechanism is triggered. Specific embodiment four:

[0059] like Figure 1 As shown, according to the content of the specific embodiment three, the following contents are further disclosed:

[0060] For the bamboo slips learning algorithm model, a logistic regression model is used. The logistic regression model is used for binary classification problems to map the input to the category probability, which is convenient for early fire warning based on the lithium battery detection data combined with the internal data of the database. The specific formula is as follows:

[0061] Logical functions:

[0062]

[0063] in:

[0064] p(y=1|x) is the probability that Y is 1 given input X;

[0065] β0, β1…β n are model parameters;

[0066] Loss function:

[0067]

[0068] in:

[0069] m is the sample size;

[0070] y (i) is the true label;

[0071] p (i) is the probability predicted by the model;

[0072] The loss function is minimized through optimization algorithms such as gradient descent to find the optimal model parameters.

[0073] The machine learning ensemble algorithm model uses random forest and decision tree-based ensemble learning algorithms to improve performance by constructing multiple decision trees and integrating their prediction results.

[0074] Main steps:

[0075] Randomly sample multiple subsets from the original dataset;

[0076] Build a decision tree for each subset;

[0077] The prediction results of all decision trees are integrated (such as taking the average or mode) to obtain the final prediction result.

[0078] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprising a reference structure" do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0079] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent data analysis technology for a lithium battery fire hazard detector, characterized in that: The intelligent data analysis technology includes the following methods: Sp1: Installation of sensor array hardware equipment, using PID principle gas detectors to monitor characteristic gases released by lithium batteries, such as CO, H2, and VOCs, and installing integrated temperature sensors and voltage and current monitoring modules to form a sensor array to obtain gas, temperature, voltage, and current data of lithium batteries in real time; Sp2: Installation of the main controller, which uses a high-performance microcontroller and is responsible for receiving sensor data, executing data analysis algorithms, and sending control instructions; Sp3: Data acquisition and preprocessing, the sensor array collects the characteristic gas, temperature, voltage and current data of the lithium battery in real time, and the data is transmitted to the main controller through the sensor interface. The main controller filters, amplifies and digitizes the received raw data; Sp4: Intelligent data analysis: Through a large number of experimental tests on batteries of different brands, a characteristic model database of the first stage of thermal runaway of lithium batteries is established. The collected data is analyzed in real time using a machine learning algorithm. The algorithm compares the real-time data with the data in the characteristic model database to determine whether the lithium battery is in the early stage of thermal runaway. Sp5: Early warning mechanism. According to the data in the characteristic model database, the early warning threshold is set. When the real-time data exceeds the threshold, the early warning mechanism is triggered, and an alarm is issued to the on-duty personnel through various means such as sound and light. At the same time, an alarm signal is sent to the BMS through the BMS interface, and corresponding treatment measures are recommended; Sp6: System optimization learning. The system can automatically adjust the thresholds and parameters of detection alarms according to monitoring data and operating status, and can automatically detect the fault status of key components such as sensors and main controllers. When a fault is detected, the fault information is sent through the BMS interface for timely maintenance.

2. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1 is characterized in that: The sensor array communicates with the main controller in a wired or wireless manner. The main controller has one or more external design interfaces among UART, I2C, and SPI for communicating with other modules.

3. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1 is characterized in that: The system optimization learning supports remote maintenance functions, and remote configuration, data query and troubleshooting operations can be performed through remote communication interfaces such as Ethernet and WiFi.

4. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1, characterized in that: The database in the intelligent data analysis contains data on characteristic gases released by lithium batteries of various brands and models in the early stages of thermal runaway.

5. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1, characterized in that: In the data collection and preprocessing, the data is cleaned to remove abnormal values, missing values ​​or duplicate data, the cleaned data is standardized, and the data is converted into a uniform scale using a standardized formula.

6. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1, characterized in that: Before the intelligent data analysis, the data is subjected to feature extraction, including features based on gas sensors and features based on physical parameters. Includes the following: Gas sensor-based features: Extract key features from the data obtained by the gas sensor, such as gas concentration and rate of change; Features based on physical parameters: Extract features such as rate of change and peak value from physical parameters such as temperature, voltage, and current.

7. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1, characterized in that: In the intelligent data analysis, a suitable machine learning algorithm, such as logistic regression and random forest, is selected according to the characteristics of the problem and the data features. The selected algorithm is trained using historical data and labels to obtain a model that can identify potential fire hazards of lithium batteries. The performance of the model is evaluated through cross-validation and other methods to ensure the effectiveness and accuracy of the model.

8. The intelligent data analysis technology for a lithium battery fire hazard detector according to claim 1, characterized in that: The early warning mechanism uses a trained model to make real-time predictions on new lithium battery data to determine whether it is in the early stages of thermal runaway. Based on the prediction results, if it is determined that the lithium battery has a fire hazard, the early warning mechanism is triggered.

Images