A thermal runaway monitoring system for a retired lithium battery detection process based on an intelligent internet of things

Abstract

The application provides a system for thermal runaway monitoring of a retired lithium battery detection process based on an intelligent Internet of Things. A sensor network for lithium battery thermal runaway monitoring is added to the system architecture. In terms of algorithm technology, the internal data of the battery and the external data sensed by the sensor during the retired lithium battery detection process are fully mined and fused for modeling. The method has the characteristics of real-time, intelligence and low cost, and is particularly suitable for factories that simultaneously detect large quantities of retired lithium batteries. The core module is composed of data acquisition, feature fusion, modeling and early warning. After the multi-source data is collected, it is processed and integrated in real time by means of an embedded device, and then uploaded to a distributed cloud management platform for in-depth analysis. The cloud dynamically models and captures time sequence features through multi-modal fusion algorithms, sliding window method, long short-term memory network (LSTM) and decision tree algorithm, and realizes accurate thermal runaway early warning and hierarchical response.

Description

Technical Field

[0001] This invention relates to the technical field of power battery testing, and in particular to a thermal runaway monitoring system and a real-time anomaly early warning method for the testing process of retired lithium batteries based on the Internet of Things. Background Technology

[0002] With the escalating global energy crisis and heightened environmental awareness, lithium-ion batteries, due to their high energy density, long lifespan, and environmentally friendly characteristics, are widely used in electric vehicles, consumer electronics, and energy storage. However, as the production and use of lithium batteries continue to increase, the number of retired lithium batteries is also experiencing explosive growth. Research shows that improper handling of retired lithium batteries can not only pollute the environment but also lead to resource waste and potential safety hazards. Therefore, the testing and reuse of retired lithium batteries has become a key focus of current scientific research and industry.

[0003] Although retired lithium batteries have reduced capacity to the point where they cannot meet the needs of primary applications, they still possess a certain energy storage capacity and can be used in low-energy storage devices or remanufactured. Furthermore, retired lithium batteries have significant resource recycling value, especially the precious metals they contain, such as lithium, cobalt, and nickel. However, thermal runaway is a major safety challenge in the testing, classification, and reuse of retired lithium batteries. Thermal runaway refers to the rapid rise in temperature of a lithium battery due to internal or external factors (such as overcharging, over-discharging, short circuits, and mechanical damage), triggering a series of uncontrollable chain reactions (such as electrolyte decomposition and cathode material oxidation), which may ultimately lead to fire or even explosion. This phenomenon not only threatens personal and property safety but also poses a significant risk to factory operations. Therefore, developing efficient and intelligent monitoring systems to address the thermal runaway risk during the testing of retired lithium batteries is crucial for ensuring safety and improving testing efficiency.

[0004] Currently, most lithium battery testing plants still rely on manual testing, supplemented by some basic temperature monitoring devices. However, this approach has the following drawbacks: (1) Insufficient real-time performance: Manual testing often has a lag and cannot respond quickly in the early stages of thermal runaway, thus missing the best intervention opportunity; (2) Limited coverage: Traditional temperature monitoring devices cannot fully cover every corner of the testing plant, and are prone to overlooking potential hidden dangers; (3) Low level of intelligence: Traditional systems lack the ability to intelligently analyze massive amounts of testing data and cannot predict and warn of thermal runaway through data mining and modeling; (4) Poor scalability: The integration of existing equipment with the plant process is relatively rigid and difficult to adapt to the needs of different testing scales or complex scenarios.

[0005] The rapid development of intelligent Internet of Things (AIoT) technology has provided a new opportunity to solve the above problems. By combining sensors, smart devices, data communication technology and distributed cloud management platforms, the Internet of Things can realize comprehensive perception, real-time monitoring and intelligent control of lithium battery testing plants. In thermal runaway monitoring, the Internet of Things technology can provide the following advantages: (1) Multi-dimensional data acquisition: Multi-dimensional data of the battery during the testing process can be obtained through temperature sensors, voltage sensors, gas sensors (such as detecting the gas generated when electrolyte leaks); (2) Real-time transmission and storage: The Internet of Things terminal devices transmit the collected data to the distributed cloud management platform through wireless communication technology to ensure the real-time nature and availability of information; (3) Intelligent analysis and decision-making: The collected data can be modeled and analyzed using big data and artificial intelligence technologies to achieve prediction and early warning of thermal runaway; (4) Automated response: By linking with the factory automation system, when the risk of thermal runaway is detected, the system can automatically trigger cooling, power outage or alarm measures to avoid the escalation of the accident; (5) Visual management: The Internet of Things platform provides a visual interface for factory managers to display the factory's operating status and risk distribution, and provides decision support.

[0006] Based on the above background, this patent proposes a thermal runaway monitoring system for the detection process of retired lithium batteries based on the intelligent Internet of Things (IoT). This system combines IoT technology and artificial intelligence algorithms to construct a complete monitoring link from data acquisition to intelligent early warning. In its specific implementation, it utilizes a distributed sensor network to achieve comprehensive monitoring of the detection area; edge computing nodes perform preliminary data processing and filtering; a distributed cloud management platform is used for in-depth analysis and modeling of large-scale data; and a predictive model provides early warning of battery thermal runaway. Furthermore, the system integrates automated control equipment, enabling rapid response in the event of a risk, ensuring the safety of the factory and personnel.

[0007] This system focuses on intelligent manufacturing and large-scale automation. Its innovation lies in the application of intelligent IoT in thermal runaway monitoring of decommissioned lithium battery testing plants, the overall system design and feasibility verification, and the design and implementation of related intelligent algorithms. The research and implementation of this patent not only solves the safety hazards of decommissioned lithium battery testing plants but also provides a reference for the deep application of IoT technology in industrial scenarios, possessing significant social and economic value. Summary of the Invention

[0008] This invention provides a thermal runaway monitoring system for the testing process of retired lithium batteries based on the Internet of Things (IoT). It solves the problem that existing retired lithium battery testing processes generally lack control over the risk of thermal runaway. Compared with other similar products, this invention combines the advantages of IoT and artificial intelligence, and has the advantages of strong real-time performance, wide coverage, and high level of intelligence. It has significant practical value for solving the safety hazards of thermal runaway in the testing process of retired lithium batteries, and is particularly suitable for factories that conduct large-scale testing of retired lithium batteries.

[0009] This invention can be achieved through the following technical solutions:

[0010] A thermal runaway monitoring system for the testing process of retired lithium batteries based on the Internet of Things (IoT) is characterized by the following: the retired lithium battery testing circuit and the battery's internal data acquisition circuit are integrated into a single circuit. An embedded device is connected to multiple battery packs under test via loads and switches, controlling these battery packs to perform testing operations. The battery packs are connected to the testing circuit through a Battery Management System (BMS), which transmits data generated during the discharge process to the embedded device. The embedded device parses the received messages and uploads the parsed data to a distributed cloud management platform.

[0011] The distributed cloud management platform processes and analyzes the received experimental data. After verifying the reliability of the data through the master-slave verification method of sensor data, it further extracts the internal characteristics of the battery pack.

[0012] Furthermore, the system constructs a signal sensing layer through a smart Internet of Things (IoT) based on multiple sensor devices, used to collect key external information of multiple power battery packs under test in real time. The collected data is aggregated to the IoT data aggregation layer via embedded devices, which are responsible for integrating the sensor data and uploading the results to a distributed cloud management platform.

[0013] The IoT is built upon a long-range, low-power radio frequency communication module, specifically implemented using the following modules: sensor nodes, a wireless gateway, embedded devices, and a distributed cloud management platform. Each sensor node has a built-in communication module for collecting external environmental information (such as temperature, humidity, gas concentration, tilt angle, etc.) of the battery pack under test and sending the data to the wireless gateway. The wireless gateway, acting as a relay device, receives signals from multiple sensor nodes, performs preliminary filtering and integration, and then transmits the data to the embedded device. The embedded device further processes and aggregates the data, and then uploads the integrated data to the distributed cloud management platform via a wide area network.

[0014] The distributed cloud management platform analyzes the received data, verifies the data's reliability using a master-slave verification method for sensor data, further extracts the external characteristics of the battery pack, and combines them with internal characteristics to make a comprehensive judgment, thereby achieving remote monitoring and real-time early warning of thermal runaway in retired lithium batteries.

[0015] From the perspective of mathematical summary and practical implementation, the internal characteristics of the battery generated in the retired battery pack testing experiment and the external characteristics sensed by the external Internet of Things during the experiment can be jointly represented as follows:

[0016] X 联合 =[X 内部 X 外部 ]

[0017] Preliminary anomaly detection results were obtained through dynamic modeling using a Long Short-Term Memory (LSTM) network.

[0018] h t =LSTM(X) 联合 )

[0019] Further, through secondary modeling using decision trees, a risk assessment and response mechanism is formed, yielding the final result:

[0020]

[0021] This solution integrates the advantages of IoT data acquisition, LSTM prediction, and decision tree classification to provide retired lithium battery testing plants with efficient thermal runaway monitoring capabilities, ensuring safety and intelligent management.

[0022] Beneficial effects

[0023] 1. This system can realize data acquisition, transmission, and intelligent analysis and processing. Compared with conventional EIS testing equipment or other customized anomaly monitoring equipment, it is not only simpler in experimental procedures, but also lighter, lower in cost, and has a wider range of applications.

[0024] 2. This system can achieve real-time monitoring of the environment and equipment status through sensor networks, providing accurate and timely data. It can also automatically collect and integrate data from multiple sources and of multiple types, providing a rich foundation for subsequent analysis.

[0025] 3. This system possesses intelligent data analysis and decision-making capabilities. Leveraging edge computing, cloud computing, and artificial intelligence, the IoT system can process and intelligently analyze collected data in real time. Furthermore, through artificial intelligence algorithms, the system can autonomously identify patterns, predict trends, and even make direct decisions, improving efficiency and response speed.

[0026] 4. This system has remote communication capabilities, which can upload experimental data to a distributed cloud management platform. Based on the diagnostic functions, it enables traceability of diagnostic results and allows for effective prediction and early warning. Attached Figure Description

[0027] Figure 1 : A schematic diagram of the overall structure of the present invention.

[0028] Figure 2 : A schematic diagram of the structure of the Internet of Things system of the present invention.

[0029] Figure 3 : A schematic diagram showing the connection between the battery module of this invention and various sensing sensors in the Internet of Things.

[0030] Figure 4 : A schematic diagram of the core modules of this invention.

[0031] Figure 5 : A schematic diagram of the system early warning and response algorithm of the present invention. Detailed Implementation

[0032] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings and preferred embodiments.

[0033] like Figure 1 As shown, this invention provides a thermal runaway monitoring system for the testing process of retired lithium batteries based on the intelligent Internet of Things (IoT). The retired lithium battery testing circuit and the battery's internal data acquisition circuit are the same circuit, including an embedded device. This device performs testing experiments on the battery pack through a load and a switch. The battery pack under test is connected to a Battery Management System (BMS), which transmits data to the embedded device in the form of messages during the testing process. The embedded device parses the received messages and uploads the data to a distributed cloud management platform. The distributed cloud management platform processes and analyzes the experimental data to extract the internal characteristics of the battery pack. Simultaneously, the system integrates an intelligent IoT based on multiple sensors, collecting key external data of the battery pack under test at the signal sensing layer and integrating the sensor data through the aggregation layer of the embedded device. After being uploaded to the distributed cloud management platform, the cloud can generate the external characteristics of the battery pack. By comprehensively analyzing the internal and external characteristics, the distributed cloud management platform can determine in real time whether the battery pack under test has a risk of thermal runaway and provide early warnings.

[0034] The data generated in the testing experiments of retired lithium batteries described in this invention will be further analyzed as internal battery data. The testing experiments include: controlling the battery pack to discharge at a fixed current using a programmable current source; recording voltage changes, discharge capacity, and other important electrochemical parameters generated during the discharge process. The internal sensors of the retired lithium batteries, which collect battery electrochemical characteristics, are also important sources of information, including current and voltage sensors that can record the battery's operating current I(t) and terminal voltage V(t) in real time. The data analysis sources involved in this invention do not include the analysis of DC internal resistance using electrochemical impedance spectroscopy, as this method is costly and generally not cost-effective in large-scale industrial settings.

[0035] like Figure 2 As shown, this invention utilizes an intelligent IoT architecture, combining external sensor data, internal battery characteristic data, and a joint modeling method to construct a system for high-precision thermal runaway monitoring of retired lithium batteries. Four external sensors are deployed to sense changes in the external state of the retired lithium battery during the detection process, and the data is transmitted via a long-distance, low-power radio frequency communication module.

[0036] like Figure 3 As shown, each battery pack under test is connected to a set of sensors (gas sensor, tilt sensor, humidity sensor, and temperature sensor) to monitor changes in the external environment and state of the battery pack. The tilt sensor detects the tilt and vibration of the battery, assessing whether the battery has deformed and whether there is a risk of thermal runaway; the other three sensors are used to monitor the temperature, humidity, and gas composition of the battery surface. The data is uploaded to the central processing unit for analysis in real time via an IoT terminal 5G module.

[0037] like Figure 4 As shown, this invention includes processes such as data acquisition, feature extraction and fusion, modeling, and early warning. Data preprocessing and feature extraction include time-series data preprocessing, feature extraction, and fusion. Time-series data preprocessing uses the sliding window method to extract statistical features (such as mean and standard deviation) for each time period during constant current discharge. Noise signals such as voltage and temperature are processed using low-pass filtering to eliminate interference data. Feature extraction and fusion include internal features of current I(t) and voltage V(t), and external features such as temperature T(t), humidity H(t), gas concentration G(t), and tilt angle θ(t). The joint feature matrix is ​​represented as follows:

[0038] X=[I(t), V(t), T(t), H(t), G(t), θ(t)]

[0039] The data modeling in this invention includes joint LSTM modeling and decision tree-assisted modeling. The loss function used in the joint LSTM modeling is cross-entropy loss, defined as follows:

[0040]

[0041] Where yi represents the real label. This is the predicted value. The LSTM model captures the dynamic changes of multidimensional time series features to predict the risk of thermal runaway. The time step of the input layer is defined as T, and the feature dimension is d.

[0042]

[0043] The LSTM cell state update formula is obtained as follows:

[0044] f t =σ(W f ·[h t-1 x t ]+b f (Forgotten Gate)

[0045] i t =σ(W i .[h t-1 x t ]+b i (Input Gate)

[0046]

[0047]

[0048] o t =σ(W o ·[h t-1 x t ]+b o (Output Gate)

[0049] h t =o t ⊙tanh(c t (Output status)

[0050] A fully connected layer is used in the output layer to predict the probability of thermal runaway.

[0051] LSTM models provide time-series prediction capabilities, while decision tree models are used to explain the main causes of thermal runaway at a specific time. Specific decision tree-assisted modeling methods include feature partitioning and outputting the thermal runaway level. In feature partitioning, at each decision node, feature x is selected. j And a threshold s, partitioned according to information gain, Q(t), R(t) are the partitioned subsets, and each leaf node t L Corresponding to a specific thermal runaway risk level Y t Low, medium, and high are used to output the thermal runaway level:

[0052]

[0053] Entropy is defined as follows:

[0054]

[0055] like Figure 5 As shown, the system early warning and response algorithm in this invention includes a multimodal fusion and early warning mechanism. The multimodal fusion algorithm combines the thermal runaway probability P(y=1|X) output by the LSTM with the risk level Y of the decision tree model. t By integrating these elements, a comprehensive evaluation index can be formed:

[0056]

[0057] Where α is the weighting function. The early warning mechanism is implemented through a threshold, and different corresponding measures are set according to the threshold of S: (1) S < 0.5: normal operation; (2) 0.5 ≤ S < 0.8: early warning, prompting factory managers to check the relevant areas; (3) S ≥ 0.8: high risk, triggering automatic response (such as power outage, starting cooling device).

Claims

1. A thermal runaway monitoring system for the detection process of decommissioned lithium batteries based on the Internet of Things, characterized in that: The system architecture incorporates a sensor network for monitoring thermal runaway in lithium batteries. Algorithmically, it fully leverages internal battery data and external sensor data collected during the testing process of retired lithium batteries, integrating and modeling them. Specifically, the retired lithium battery testing circuit and the internal battery data acquisition circuit are the same circuit. An embedded device connects to multiple battery packs under test via loads and switches, simultaneously controlling these packs for testing. Each battery pack is connected to a Battery Management System (BMS), which transmits data during the testing process to the embedded device. The embedded device parses the data and uploads it to a distributed cloud management platform. The distributed cloud management platform analyzes and processes the received experimental data to generate internal battery pack characteristics F. internal =f(voltage, current); A smart Internet of Things (IoT) built based on various sensor devices is used for the collection of external data of the battery. The signal sensing layer of the IoT simultaneously collects key external information of multiple power battery packs under test. The data aggregation layer of the IoT is integrated into an embedded device. The embedded device aggregates the sensor information received and uploads it to a distributed cloud management platform to generate the external features F of the battery pack. external =f′(temperature, inclination, gas, humidity), the distributed cloud management platform uses an artificial intelligence fusion algorithm to analyze the internal characteristics F of the battery pack. internal and external features of the battery pack F external After comprehensive analysis, it can determine in real time whether the power battery pack under test has thermal runaway and issue timely warnings.

2. The thermal runaway monitoring system for the detection process of decommissioned lithium batteries based on the intelligent Internet of Things as described in claim 1, characterized in that: The Internet of Things (IoT) is constructed based on long-distance, low-power radio frequency communication modules, featuring low power consumption, wide area network, low cost, open standards, and easy deployment. The communication structure of the IoT mainly involves the physical layer (PHY) and the data link layer (MAC). The physical layer uses spread spectrum modulation and can increase the communication range through the spread spectrum factor. The data link layer defines three device categories: Class A, Class B, and Class C, each with different power consumption and communication methods.

3. The thermal runaway monitoring system for the detection process of decommissioned lithium batteries based on the intelligent Internet of Things as described in claim 2, characterized in that: The IoT based on sensor devices uses sensors that monitor the physical state of batteries as signal sensing units to monitor abnormal conditions of the power battery pack under test in real time, preventing the power battery pack under test from catching fire and causing accidents during charge and discharge experiments. The sensors include a temperature sensor, a tilt sensor, a gas sensor, and a humidity sensor connected to the power battery pack under test. The temperature sensor senses the surface temperature of the power battery pack under test. The tilt sensor calibrates and monitors the surface deformation of the power battery pack under test through a horizontal grid. The gas sensor monitors the concentration of carbon monoxide gas, which is mainly released in the event of thermal runaway of the power battery pack. The humidity sensor indirectly measures the surface humidity of the power battery pack through a moisture-sensitive material film. Each set of sensors includes a temperature sensor, a tilt sensor, a gas sensor, and a humidity sensor, and is connected to the corresponding power battery pack under test in sequence. It monitors the data from each power battery pack under test during the experiment and then transmits the obtained data back to the data aggregation layer.

4. A thermal runaway monitoring system for the detection process of retired lithium batteries based on the intelligent Internet of Things as described in claim 1, the core requirement of which is to achieve real-time monitoring and early warning of the status of all battery packs while performing basic testing on retired lithium battery packs, so as to prevent on-site disasters. Its intelligent features include: automated calculation of sensor positions, validity detection of sensor data, and intelligent detection algorithms for anomalies in the tested power battery packs based on artificial intelligence methods, wherein... When conducting experiments on a large number of battery packs under test, the sensors and the battery packs they monitor are usually physically close. Utilizing the sensor's own wireless interferometric positioning technology, automatic positioning with centimeter-level accuracy can be achieved. This location information is crucial for quickly identifying potential hazards.

5. The thermal runaway monitoring system for the detection process of decommissioned lithium batteries based on the intelligent Internet of Things as described in claim 4, characterized in that: The algorithm for validating sensor data uses a master-slave verification method. For sensors deployed in the same group of batteries for testing, when examining the authenticity of data from one sensor, data from other sensors is used for comparison. If the data comparison result is less than a given threshold and matches the results of most sensors, it indicates that the data is reliable. The communication status is checked by checking whether the communication life information contained in the 8th byte of 186040F3 in the BMS system information service communication protocol is continuous.

6. The thermal runaway monitoring system for the detection process of decommissioned lithium batteries based on intelligent Internet of Things as described in claim 4, characterized in that: The intelligent detection algorithm for anomalies in the tested power battery pack is based on a fusion model of deep learning and machine learning algorithms. It uses a Long Short-Term Memory (LSTM) network model to model the dynamic changes in features inside and outside the battery during the detection process, capturing minute anomalies before thermal runaway. Further, a decision tree-based secondary modeling process is used to form a risk assessment and response mechanism, reducing the false alarm rate.

7. The thermal runaway monitoring system for the detection process of decommissioned lithium batteries based on the intelligent Internet of Things as described in claim 1, characterized in that: The diagnostic report covers the detection of system anomalies, measurement of DC internal resistance of individual cells and assessment of self-discharge rate, as well as analysis and detection of internal resistance consistency.

Images