Thermal runaway prevention and control method based on artificial intelligence, energy storage power supply, electronic equipment and medium

Abstract

The invention discloses a thermal runaway prevention and control method based on artificial intelligence, an energy storage power supply, electronic equipment and a medium, and is applied to the technical field of thermal runaway prevention and control of the energy storage power supply. The method comprises the steps that multi-mode collection data of an energy storage power supply is obtained, a fire fighting module is arranged in the energy storage power supply, and the fire fighting module is used for spraying out a fire extinguishing agent to extinguish fire after being triggered; inputting the multi-modal acquisition data to a preset thermal runaway detection model for thermal runaway detection to obtain a thermal runaway detection result; and based on the thermal runaway detection result, performing thermal runaway prompting and / or controlling the fire-fighting module to spray out a fire extinguishing agent. The thermal runaway can be timely and accurately detected in combination with multi-modal data collected by the energy storage power supply in real time and the artificial intelligence model, and the thermal runaway can be prompted based on the thermal runaway detection result and extinguished in combination with the preset fire-fighting module, so that the thermal runaway is prevented and controlled; and the thermal runaway risk of the energy storage power supply and the damage of thermal runaway to personal safety and economy are reduced.

Description

Technical Field

[0001] This application belongs to the field of thermal runaway prevention and control technology for energy storage power supplies, and particularly relates to an artificial intelligence-based thermal runaway prevention and control method, energy storage power supply, electronic equipment, and computer-readable storage medium. Background Technology

[0002] In recent years, Europe has been mired in the dual predicament of energy supply shortages and soaring prices, with many European countries facing severe power shortage risks. Energy price indices have repeatedly hit record highs, with peak electricity prices in some regions skyrocketing tenfold, significantly increasing the cost of electricity for residents. Against this backdrop, the EU is accelerating its energy independence strategy, seeking to diversify its natural gas imports on the one hand, and positioning distributed renewable energy as a core breakthrough for energy transition on the other.

[0003] Against the backdrop of a global push for sustainable energy development, photovoltaic (PV)-based home energy storage systems are rapidly emerging as a rising star in the distributed energy sector due to their unique advantages. Balcony PV energy storage systems, with their extremely low deployment threshold and rapid energy self-sufficiency capabilities, have become an innovative home energy storage solution and have been widely adopted in the European market. Balcony PV energy storage systems are plug-and-play clean energy solutions specifically designed for apartment balconies, small courtyards, and windowsills. They integrate PV power generation and energy storage / backup in a simple enclosure, empowering families to achieve energy self-sufficiency and electricity security. Their core function is to provide users with "self-consumption and surplus power fed into the grid." "Self-consumption" allows PV power to be supplied to household loads (such as refrigerators and lighting equipment), reducing the need for grid connection and saving on electricity bills. "Surplus power fed into the grid" allows excess power generated by PV to be fed back into the public grid, generating income for the user.

[0004] However, home energy storage systems, or balcony photovoltaic energy storage systems, can experience thermal runaway due to various external factors (such as high temperatures) or internal factors (such as poor heat dissipation and poor cell uniformity). Severe thermal runaway can even lead to fires. According to a survey by the European Photovoltaic Industry Association, nearly 80% of potential users have abandoned the purchase of balcony photovoltaic energy storage systems due to concerns about fire hazards. Therefore, the detection and prevention of thermal runaway is a crucial technical issue that urgently needs to be addressed to ensure the safety of energy storage systems. Summary of the Invention

[0005] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes an artificial intelligence-based thermal runaway prevention method, energy storage power supply, electronic equipment, and computer-readable storage medium. This method combines real-time multimodal data collected by the energy storage power supply with an artificial intelligence model to achieve timely and accurate detection of thermal runaway. Furthermore, it can provide alerts based on the thermal runaway detection results and, in conjunction with pre-set fire suppression modules, extinguish fires, thereby achieving the prevention and control of thermal runaway, reducing the risk of thermal runaway in energy storage power supplies and the damage caused by thermal runaway to personal safety and the economy.

[0006] Firstly, this application provides an artificial intelligence-based method for preventing thermal runaway, the method comprising:

[0007] The system acquires multimodal data from the energy storage power source, which includes a variety of different types and / or formats of data sets. The energy storage power source has a built-in fire suppression module, which is used to spray fire extinguishing agent to extinguish fires when triggered.

[0008] Input multimodal acquisition data into a preset thermal runaway detection model to perform thermal runaway detection and obtain thermal runaway detection results;

[0009] Based on the thermal runaway detection results, provide thermal runaway alerts and / or control the fire suppression module to spray extinguishing agents.

[0010] Secondly, this application provides an energy storage power source, comprising:

[0011] Fire suppression module, used to spray extinguishing agent;

[0012] Battery module; and

[0013] The controller is connected to the battery module and the fire protection module. The controller is used to implement thermal runaway prevention methods.

[0014] Thirdly, the electronic device provided in this application includes:

[0015] It includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, it implements the above-mentioned thermal runaway prevention and control method.

[0016] Fourthly, this application provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described thermal runaway prevention and control method.

[0017] The artificial intelligence-based thermal runaway prevention and control method, energy storage power supply, electronic device, and computer-readable storage medium provided in this application embodiment can acquire multimodal acquisition data of the energy storage power supply in real time. The multimodal acquisition data includes data of various types and / or different formats (such as temperature data, electrical parameter data, smoke concentration, and other acquisition data related to thermal runaway). By acquiring multimodal acquisition data, more data references can be provided for thermal runaway detection, which is beneficial to improving the accuracy of thermal runaway detection.

[0018] It is understandable that, given the diverse data types and formats involved in multimodal acquisition, conventional control logic would be excessively complex for thermal runaway detection. Therefore, using an AI-based thermal runaway detection model to process multimodal acquisition data not only facilitates the full utilization of this data but also reduces the complexity of the control logic, thereby improving the accuracy of thermal runaway detection while lowering its implementation cost.

[0019] Finally, based on accurate thermal runaway detection results, thermal runaway warnings or fire suppression measures are implemented. When there is no risk of thermal runaway, users can clearly perceive the situation, improving the safety and perceptibility of the energy storage power supply. When there is a risk of thermal runaway, timely warnings are provided, allowing users to take preventative measures. If thermal runaway has already occurred, fire suppression measures can be implemented directly, thus achieving thermal runaway control.

[0020] In this way, thermal runaway can be detected, prevented, and controlled, improving the safety and perceptibility of energy storage power sources, reducing the risk of thermal runaway and the damage to personal safety and the economy caused by thermal runaway.

[0021] Additional aspects and advantages of embodiments of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of embodiments of this application. Attached Figure Description

[0022] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0023] Figure 1 This is a schematic diagram illustrating an application scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0024] Figure 2 This is a first flowchart illustrating the thermal runaway prevention and control method provided in certain embodiments of this application;

[0025] Figure 3 This is a schematic diagram of the second process of the thermal runaway prevention and control method provided in some embodiments of this application;

[0026] Figure 4This is a schematic diagram of the structure of a thermal runaway detection model provided in some embodiments of this application;

[0027] Figure 5 This is a schematic diagram of the third process of the thermal runaway prevention and control method provided in some embodiments of this application;

[0028] Figure 6 This is a schematic diagram showing the connection between the controller and the fire-fighting module of the energy storage power supply provided in some embodiments of this application;

[0029] Figure 7 This is a schematic diagram of a first scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0030] Figure 8 This is a schematic diagram of a second scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0031] Figure 9 This is a schematic diagram of a third scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0032] Figure 10 This is a schematic diagram of the fourth process of the thermal runaway prevention and control method provided in some embodiments of this application;

[0033] Figure 11 This is a schematic diagram of a fourth scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0034] Figure 12 This is a schematic diagram of the fifth process of the thermal runaway prevention and control method provided in some embodiments of this application;

[0035] Figure 13 This is a schematic diagram of a fifth scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0036] Figure 14 This is a schematic diagram of a sixth scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0037] Figure 15 This is a schematic diagram of the seventh scenario of the thermal runaway prevention and control method provided in some embodiments of this application;

[0038] Figure 16 This is a schematic diagram of the sixth process of the thermal runaway prevention and control method provided in some embodiments of this application;

[0039] Figure 17 This is a schematic diagram of a thermal runaway prevention device provided in some embodiments of this application. Detailed Implementation

[0040] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0041] Please see Figure 1 , Figure 1 This is an application scenario diagram of an artificial intelligence-based thermal runaway prevention method provided in an embodiment of this application. The application scenario provided in this application includes an energy storage power supply 100 and an electronic device 200.

[0042] Artificial Intelligence (AI) refers to the theories, methods, technologies, and application systems that utilize digital computers or computers-controlled machines to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce new intelligent machines that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess perception, reasoning, and decision-making capabilities.

[0043] Machine Learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning.

[0044] Deep Learning (DL) is a branch of machine learning that attempts to perform high-level abstractions of data using multiple processing layers containing complex structures or multiple nonlinear transformations. Deep learning learns the inherent patterns and hierarchical representations of training sample data; the information gained during this learning process greatly aids in interpreting data such as text, images, and sound. The ultimate goal of deep learning is to enable machines to possess analytical and learning capabilities similar to humans, capable of recognizing data such as text, images, and sound. Deep learning is a complex machine learning algorithm, and its performance in speech and image recognition far surpasses previous related technologies.

[0045] Here, the energy storage power supply 200 refers to a device capable of storing power. It is generally equipped with a rechargeable battery. By storing a large amount of power in the battery within the energy storage power supply, the energy storage power supply can output the stored electrical energy when needed.

[0046] The energy storage power supply 200 may include a battery module 201 and a motherboard 202. The battery module 201 is used to store power, while the motherboard 202 can control and manage the battery module, as well as control other components of the energy storage power supply.

[0047] There are many types of energy storage power supplies, which can be classified according to application scenarios:

[0048] (1) Portable energy storage: It is generally a small energy storage power supply, using lithium-ion batteries, etc. It is easy to carry and used in outdoor camping, emergency charging and other scenarios. It can power mobile phones, computers, lighting equipment, etc.

[0049] (2) Home energy storage: Used in homes to store solar power or electricity generated during off-peak hours of the power grid for use by home electrical equipment, achieving the purpose of peak shaving and valley filling, saving electricity costs, etc.

[0050] (3) Industrial and commercial energy storage: Used in factories, data centers, shopping malls and other places, it can be used for load regulation, demand-side management, power quality improvement, etc., to help users reduce electricity costs and improve power supply reliability.

[0051] (4) Grid energy storage: It is widely used in power systems to regulate the peak-valley difference of the power grid, smooth the fluctuations of renewable energy generation, and improve the stability and reliability of the power grid. Common types include large lithium-ion battery energy storage power stations, flow battery energy storage power stations, and pumped storage power stations.

[0052] In order to adapt to the increasingly diverse power consumption scenarios, portable energy storage power supplies have emerged. Portable energy storage power supplies, also known as portable lithium-ion battery energy storage power supplies or outdoor power supplies, usually refer to backup or emergency power supplies weighing no more than 18 kg. They use lithium-ion batteries as energy storage components and have AC or DC input charging interfaces as well as AC or DC output interfaces.

[0053] Balcony photovoltaic energy storage power supply typically refers to a small photovoltaic power generation and energy storage power supply installed in external spaces such as family balconies, garages, and courtyards. The energy storage power supply includes an inverter that can perform DC / DC conversion and DC / AC conversion, thereby realizing AC-DC conversion for AC charging and discharging, as well as DC charging and discharging.

[0054] Balcony energy storage is typically considered household energy storage. However, to enable dual-use for both home and outdoor applications, portable energy storage power supplies can be used for home energy storage (e.g., multiple portable energy storage power supplies operating in parallel). When outdoor power is needed, these portable energy storage power supplies can be taken out as portable energy storage power supplies, thus enabling cross-domain applications and expanding the application scenarios of portable energy storage.

[0055] In an optional embodiment, the energy storage power supply 100 further includes a fire suppression module 203. The fire suppression module 203 is used to be triggered in the event of thermal runaway of the energy storage power supply 100, thereby spraying a fire extinguishing agent to extinguish the fire.

[0056] In one optional embodiment, the extinguishing agent may be at least one selected from aerosol, heptafluoropropane, carbon dioxide, and inert gas. The aerosol not only isolates oxygen but also encapsulates flammable materials in the air, further enhancing the extinguishing effect.

[0057] The energy storage power supply 100 can communicate with the electronic device 200 to cooperate with the electronic device 200 to implement the thermal runaway detection method of this application.

[0058] Optionally, the electronic device 200 includes at least one of a terminal and a server.

[0059] The terminal may include, but is not limited to: smartphones (such as Android phones, iOS phones, etc.), tablet computers, laptops, desktop computers, smart speakers, smartwatches, portable personal computers, mobile internet devices (MIDs), smart voice interaction devices, smart home appliances, vehicle terminals, aircraft, wearable devices, etc., but this application embodiment does not limit the scope of the terminal.

[0060] The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. This application does not limit this.

[0061] The artificial intelligence-based thermal runaway prevention method of this application can be implemented by an energy storage power source or electronic equipment, or by an energy storage power source in conjunction with electronic equipment, without any limitation.

[0062] Based on the above-described scenarios, this application provides an artificial intelligence-based method for preventing thermal runaway. The method is described in detail below:

[0063] Please see Figure 2 This application provides an artificial intelligence-based thermal runaway prevention and control method, which is implemented by steps 011 to 013, as described in detail below.

[0064] Step 011: Acquire multi-modal acquisition data of the energy storage power source.

[0065] The multimodal acquisition data includes various data sets of different types and / or formats, all of which are related to thermal runaway. These modes can describe the impact of thermal runaway from different dimensions, providing more comprehensive and richer information through fusion and complementarity.

[0066] For example, multimodal acquisition data includes various types of acquisition data (such as temperature data related to thermal runaway, battery electrical parameters, smoke concentration data, etc.); or multimodal acquisition data includes acquisition data in various formats (such as smoke concentration data recorded as text, smoke concentration data recorded as images, etc.); or multimodal acquisition data includes acquisition data of various types and formats (such as temperature data recorded as text, battery electrical parameters, smoke concentration data, etc., smoke concentration data recorded as images, etc.).

[0067] In one optional embodiment, the multimodal acquisition data includes at least one of temperature data, smoke data, battery electrical parameter data, electrochemical gas data, and air pressure data.

[0068] Please refer to the following again. Figure 1 The energy storage power supply 200 also includes a temperature sensing wire 204, which is installed in the battery module 201. The temperature data includes battery temperature data and fuse data, and the fuse data includes whether the temperature sensing wire 204 has melted.

[0069] Among them, the temperature sensing wire 204 is a thermistor. When the temperature reaches a preset threshold (an empirical value) and remains there for a certain period of time (an empirical value, such as 10 seconds), the temperature sensing wire 204 will disconnect, and thus be detected by the battery management system of the energy storage power supply 200 (such as deployed on the motherboard 202); or, the temperature sensing wire 204 is connected to the fire-fighting module 203. The mechanical energy generated after the temperature sensing wire 204 melts will trigger the combustion of the initiator in the fire-fighting module 203, thereby causing the fire-fighting module 203 to spray out the extinguishing agent.

[0070] The energy storage power supply 200 can be provided with a temperature sensing line 204 on the surface of the battery module 201. The temperature sensing line 204 can pass through the area corresponding to each cell of the battery module 200, so that in the event of thermal runaway in any cell, the temperature sensing line 204 can be melted in time, and the generated melting data (such as whether it has melted) can provide an important basis for subsequent thermal runaway detection.

[0071] Battery module 201 typically controls charging and discharging based on battery temperature to prevent thermal runaway caused by overheating of the cells during charging and discharging. Multiple temperature sensors are generally installed within battery module 201 to collect the temperature of the battery cells. By monitoring battery temperature data (such as the temperature of each individual cell), overheated cells can be detected early, and thermal runaway warnings can be issued before thermal runaway occurs, thus facilitating its prevention.

[0072] The energy storage power supply 200 is also equipped with a smoke sensor 205, which is used to collect smoke data.

[0073] It is understandable that when thermal runaway begins, the internal reaction of the battery cell will generate a large amount of gas, resulting in smoke. Collecting smoke data through a smoke sensor can provide strong data support for thermal runaway detection and improve the accuracy of thermal runaway detection.

[0074] The energy storage power supply 200 is also equipped with an electrochemical sensor 206, which is used to collect electrochemical gas data, including the type and concentration of thermal runaway marker gas.

[0075] Similarly, when thermal runaway begins, the internal reaction of the battery cell generates a large amount of gas (including combustible gases such as hydrogen fluoride). Electrochemical sensors can collect the types and concentrations of various combustible gases, providing strong data support for thermal runaway detection and improving the accuracy of thermal runaway detection.

[0076] The energy storage power supply 200 also includes a pressure sensor. The battery module includes multiple cells, and each cell has a pressure sensor inside. The pressure sensor is used to collect pressure data, including the internal pressure of the cell.

[0077] When thermal runaway begins, the internal reaction of the battery cell generates a large amount of gas, causing the gas pressure inside the cell to gradually increase. Therefore, by using gas pressure sensors inside the battery cell, the gas pressure inside each cell can be collected, providing strong data support for thermal runaway detection and improving the accuracy of thermal runaway detection.

[0078] Step 012: Input multimodal acquisition data into the preset thermal runaway detection model to perform thermal runaway detection and obtain thermal runaway detection results.

[0079] Among them, thermal runaway detection models can be based on artificial intelligence neural network models.

[0080] It is understandable that, given the diverse data types and formats involved in multimodal acquisition, conventional control logic would be excessively complex for thermal runaway detection. Using an AI-based thermal runaway detection model to process multimodal acquisition data not only facilitates the full utilization of this data but also reduces the complexity of the control logic, thereby improving the accuracy of thermal runaway detection while lowering its implementation cost.

[0081] In one alternative embodiment, the thermal runaway detection model can be trained using multimodal training data.

[0082] Among them, multimodal training data can be generated based on the historical multimodal acquisition data of each energy storage power source.

[0083] Multimodal training data can be used to form training samples for training thermal runaway detection models. The training of thermal runaway detection models can be completed using these training samples. After the thermal runaway detection model is trained, it can be deployed to energy storage power supplies and / or electronic devices to achieve real-time thermal runaway detection results by inputting real-time multimodal acquisition data into the thermal runaway detection model.

[0084] For example, thermal runaway detection results are used to characterize the thermal runaway status of energy storage power supplies. These results may include whether thermal runaway has occurred. Alternatively, they may include whether thermal runaway has not occurred, is suspected to be thermal runaway, or has occurred.

[0085] Please see Figure 3 and Figure 4 In an optional embodiment, the thermal runaway detection model includes a first feature extraction module and a second feature extraction module. The first feature extraction module includes a convolutional neural network, and the second feature extraction module includes a long short-term memory network. Step 012 includes:

[0086] Step 0121: Input multimodal acquisition data into the first feature extraction module and the second feature extraction module respectively, so as to extract the first feature vector through the convolutional neural network and extract the second feature vector through the long short-term memory network;

[0087] Step 0122: Concatenate the first and second feature vectors to generate the third feature vector;

[0088] Step 0123: Perform thermal runaway detection based on the third feature vector to obtain the thermal runaway detection result.

[0089] Convolutional neural networks (CNNs) are a type of neural network specifically designed to process data with a grid-like topological structure. Their core function is to extract spatial features through convolution operations, and they are widely used in fields such as image recognition and video analysis.

[0090] Among them, Long Short-Term Memory (LSTM) is an improved version of Recurrent Neural Network (RNN). It solves the long-term dependency problem of traditional RNN through gating mechanism (early information is easily lost in long sequences) and is suitable for processing time-series data.

[0091] It is understandable that multimodal acquisition data includes instantaneous data (such as battery temperature data, thermal sensing wire meltdown data, etc.) as well as data that gradually accumulates over time (such as the rate of temperature change, smoke concentration, combustible gas concentration, air pressure data, etc.).

[0092] For instantaneous data, features can be effectively extracted using convolutional neural networks, which form the first feature vector. For data that accumulates over time, features can be extracted using long short-term memory networks, which form the second feature vector.

[0093] Therefore, by using an appropriate feature extraction network to extract features from data with different characteristics, the accuracy of feature extraction can be maximized.

[0094] Then, by concatenating the first and second feature vectors, the third feature vector of all extracted features is obtained.

[0095] Finally, accurate thermal runaway detection is performed based on the accurate third feature vector, thus obtaining accurate thermal runaway detection results.

[0096] Please refer to it again. Figure 3 and Figure 4 In an optional embodiment, the thermal runaway detection model further includes an attention module and a fully connected classification module, and step 0123 includes:

[0097] Step 01231: Assign different weights to different types of data in the third feature vector through the attention module to obtain the fourth feature vector, wherein the weight of the first type of data is greater than the weight of the second type of data, and the first type of data includes at least data related to the concentration of hydrogen fluoride.

[0098] Step 01232: Process the fourth feature vector through the fully connected classification module to obtain the thermal runaway detection result.

[0099] The attention module is a model component that simulates human attention. By calculating the weights of the input data, it allows the model to focus on key information, suppress irrelevant information, and improve processing efficiency and accuracy.

[0100] Among them, the fully connected classification module is a neural network structure in which all neurons in the layers are interconnected, and finally outputs the classification result. It is one of the most basic classification models.

[0101] After obtaining the third feature vector, in order to further improve the model's detection accuracy, weights can be assigned to different types of data based on the impact of different types of collected data on thermal runaway judgment. The greater the weight assigned to a type of data, the more the model will focus on that type of data based on the attention mechanism of the attention module, and the greater the impact of that type of data on the thermal runaway detection results.

[0102] For example, the weight of the first type of data can be set higher than that of the second type of data. The first and second types of data are empirical values, representing data that has a greater impact on accurately determining thermal runaway in real-world scenarios. For instance, the first type of data may include at least data related to hydrogen fluoride concentration, or key data for determining thermal runaway such as hydrogen fluoride concentration and melting point. By assigning higher weights to the first type of data, the model can focus more on the key data, thereby improving the model's detection accuracy.

[0103] After processing the third feature vector through the attention module, a fourth feature vector that is more focused on the key data is obtained. Finally, the fourth feature vector is processed through the fully connected classification module to output the thermal runaway detection result.

[0104] In one alternative embodiment, the thermal runaway detection model includes a lightweight edge model or an anomaly detection model.

[0105] Lightweight edge models refer to low-power, low-computational-complexity deep learning models specifically designed for edge computing devices (such as mobile phones, IoT terminals, and embedded devices). They can be deployed on users' electronic devices, thus avoiding excessive computational resource consumption by thermal runaway detection models.

[0106] Among them, anomaly detection models are algorithms used to identify data that do not conform to expected patterns (i.e., "anomalies" or "outliers"), and are widely used in fields such as fraud detection, industrial fault diagnosis, and medical monitoring.

[0107] It is understandable that the probability of thermal runaway in energy storage power sources is extremely low, resulting in very little sample data on thermal runaway. However, anomaly detection models can address the problem of imbalanced samples by using only normal samples (i.e., samples that have not experienced thermal runaway or may experience thermal runaway) for training, thus obtaining a thermal runaway detection model with satisfactory accuracy.

[0108] Step 013: Based on the thermal runaway detection results, provide thermal runaway alerts and / or control the fire suppression module to spray extinguishing agents.

[0109] After obtaining accurate and timely thermal runaway detection results from the thermal runaway detection model, thermal runaway prevention and control of energy storage power supplies can be achieved based on the thermal runaway detection results.

[0110] For example, thermal runaway warnings can be issued based on thermal runaway detection results. When there is no risk of thermal runaway, users can clearly perceive this, improving the safety and perceptibility of the energy storage power supply. When there is a risk of thermal runaway, timely warnings can be issued, allowing users to take preventative measures. If thermal runaway has already occurred, warnings can promptly inform users so they can take appropriate fire-fighting measures, thereby preventing fires caused by thermal runaway of the energy storage power supply.

[0111] For example, the fire suppression module can be controlled to spray extinguishing agents based on thermal runaway detection results. In the event of thermal runaway, the fire suppression module can be controlled to spray extinguishing agents to control the thermal runaway, thereby eliminating it and preventing a fire.

[0112] Please see Figure 5 In an optional embodiment, the thermal runaway detection result includes the thermal runaway level. If the thermal runaway level reaches a preset level, it is determined that the energy storage power supply has thermally runaway. Step 013 includes:

[0113] Step 0131: Based on the thermal runaway level, issue a thermal runaway warning, and when the thermal runaway level reaches the preset level, control the fire-fighting module to spray extinguishing agent.

[0114] The thermal runaway level is a numerical value used to assess the risk of thermal runaway; the higher the risk of thermal runaway, the higher the thermal runaway level.

[0115] For example, when there is no risk of thermal runaway, the thermal runaway level is the lowest; when there is a risk of thermal runaway, the thermal runaway level is the highest, reaching the preset level; and when thermal runaway is likely to occur, the thermal runaway level is positively correlated with the probability of thermal runaway occurring.

[0116] In this way, by using thermal runaway levels, the risk of thermal runaway is quantified, allowing users to intuitively perceive the risk of thermal runaway. Furthermore, when the thermal runaway level reaches a preset level, the fire-fighting module is controlled to spray extinguishing agent to achieve thermal runaway treatment and eliminate thermal runaway.

[0117] Please see Figure 6 In one optional embodiment, the energy storage power supply also includes a controller (such as motherboard 202), which is connected to the electric initiator of the fire-fighting module. The controller can send an electric start signal to the electric initiator to activate the extinguishing agent in the fire-fighting module, thereby causing the extinguishing agent to be sprayed out.

[0118] In an optional embodiment, step 013 or step 0131 includes: providing a thermal runaway warning based on the thermal runaway level, including:

[0119] A thermal runaway warning is issued that matches the thermal runaway level, and there is a positive correlation between the thermal runaway level and the warning intensity.

[0120] It is understandable that the higher the risk of thermal runaway, the more timely the warning information needs to be perceived by the user. Therefore, thermal runaway warnings can be issued based on the thermal runaway level; the higher the thermal runaway level, the stronger the warning, ensuring that the user is aware of the thermal runaway situation in a timely manner without unduly disturbing the user.

[0121] Taking thermal runaway levels as an example, with N levels as the preset level and M as the current thermal runaway level:

[0122] (1) Please refer to Figure 7 When M=0, a Level 0 thermal runaway warning is displayed on the display screen of the energy storage power supply and / or on the status interface of the energy storage power supply displayed on the electronic devices associated with the energy storage power supply. A Level 0 thermal runaway warning includes no risk of thermal runaway.

[0123] This can enhance users' perception of the safety of energy storage power supplies, increase their confidence in their safety, and help increase the sales of energy storage power supplies.

[0124] (2) When M is less than N, an M-level thermal runaway warning is issued through the first warning device of the energy storage power supply and / or the second warning device of the electronic device associated with the energy storage power supply.

[0125] In one optional embodiment, the first prompting device includes at least one of a display screen of an energy storage power source, a speaker, a vibration motor, and an indicator light, and the second prompting device includes at least one of a display screen of an electronic device, a speaker, a vibration motor, and an indicator light.

[0126] It is understandable that when M is not 0 and is less than N, it indicates that the energy storage power supply may be at risk of thermal runaway. In this case, in addition to the visual prompts on the display screen of the energy storage power supply, other prompts such as audio prompts and vibration prompts can be added to increase the intensity of the prompts.

[0127] In one alternative embodiment, the number of components used for prompting by the first and second prompting devices and / or the prompting intensity are both positively correlated with the thermal runaway level.

[0128] For example, the higher the thermal runaway level, the stronger the alerting intensity of the components used by the first and second alerting devices; the lower the thermal runaway level, the weaker the alerting intensity of the components used by the first and second alerting devices.

[0129] If N is 3 and the thermal runaway level is 0, only the displays of the energy storage power supply and electronic devices will issue thermal runaway warnings; when the thermal runaway level is 1, the displays and speakers of the energy storage power supply, as well as the displays and speakers of the electronic devices, will all issue thermal runaway warnings; when the thermal runaway level is 2, the displays, speakers, and vibration motors of the energy storage power supply, as well as the displays, speakers, and vibration motors of the electronic devices, will all issue thermal runaway warnings; when the thermal runaway level is 3, the displays, speakers, vibration motors, and indicator lights of the energy storage power supply, as well as the displays, speakers, vibration motors, and indicator lights of the electronic devices, will all issue thermal runaway warnings.

[0130] For example, the higher the thermal runaway level, the more components the first and second warning devices use for warning; the lower the thermal runaway level, the fewer components the first and second warning devices use for warning.

[0131] If both the first and second notification devices include a display screen and a speaker, the higher the thermal runaway level, the brighter the thermal runaway notification displayed on the display screen and the louder the notification sound from the speaker; or, the more eye-catching the way the display screen displays the thermal runaway notification (such as status bar notifications, lock screen notifications, notification bar notifications, and pop-up notifications, the more eye-catching they become).

[0132] For example, the higher the thermal runaway level, the more components and intensity of the first and second warning devices are used for warning; the lower the thermal runaway level, the fewer components and intensity of the first and second warning devices are used for warning.

[0133] In one optional embodiment, the M-level thermal runaway indication includes the thermal runaway probability and guidance information for preventing thermal runaway. The thermal runaway probability and the value of M are positively correlated, and the guidance information for preventing thermal runaway is determined based on the thermal runaway probability.

[0134] It's understandable that there's a positive correlation between thermal runaway level and thermal runaway probability. By specifying the thermal runaway level, the corresponding probability can be quickly determined and displayed to the user through a thermal runaway warning, allowing them to intuitively see the probability. Compared to displaying the thermal runaway level, where users might not be familiar with the highest level and therefore cannot intuitively understand the risk, displaying the maximum thermal runaway probability—which requires no active learning from the user—provides a more direct understanding of the thermal runaway risk.

[0135] In addition to the probability of thermal runaway, some users who lack knowledge of thermal runaway prevention need to be given certain thermal runaway guidance so that they know how to investigate the possible causes of thermal runaway in energy storage power supplies based on the current thermal runaway risk, or how to deal with factors that may cause thermal runaway, thereby achieving thermal runaway prevention.

[0136] For thermal runaway warnings of different thermal runaway levels, the guidance information for preventing thermal runaway can be different. Compared to thermal runaway guidance information that includes guidance information for all thermal runaway levels, this allows users to achieve thermal runaway prevention more quickly and in a more targeted manner.

[0137] For example, if N is 5 and M is 1, 2, 3, or 4, the probability of thermal runaway in a Level 1 thermal runaway warning is 20%, 40%, 60%, and 80%. Figure 8 As shown, a Level 1 thermal runaway warning is displayed, with a thermal runaway probability of 20%, and the guidance message is "Please turn off the energy storage power supply to cool down!".

[0138] (3) Please refer to Figure 9 When M reaches N, an N-level thermal runaway warning is issued through the first warning device of the energy storage power supply and / or the second warning device of the electronic device associated with the energy storage power supply. The N-level thermal runaway warning includes information on the thermal runaway of the energy storage power supply and guidance on handling the thermal runaway.

[0139] When M reaches N, meaning the energy storage power supply has experienced thermal runaway, it is essential to ensure that users are aware of the thermal runaway warning. Therefore, the highest level N thermal runaway warning can be issued through a first and a second warning device, providing the thermal runaway warning in the most easily perceptible way and with the strongest possible intensity.

[0140] Furthermore, since thermal runaway has already occurred, there is no longer a problem of thermal runaway prevention. Therefore, the Level N thermal runaway warning also needs to include guidance information on handling thermal runaway, instructing users on how to deal with it (such as contact information for the fire department, fire extinguishing measures, etc.), so as to help users handle thermal runaway in a timely and effective manner.

[0141] For example, such as Figure 9 As shown, the N-level thermal runaway warning is "The energy storage power supply has experienced thermal runaway. Please contact firefighters or use fire extinguishers to put out the fire immediately!"

[0142] In one alternative embodiment, the electronic devices associated with the energy storage power source may include electronic devices owned by the user of the energy storage power source and electronic devices owned by the user of fire-related platforms (such as electronic devices associated with the fire brigade's official WeChat account).

[0143] In this way, thermal runaway alerts can be sent not only to users of energy storage power sources, but also to fire brigades, which is beneficial for fire control.

[0144] It's understandable that, to avoid disrupting fire-related platforms, thermal runaway information could only be pushed to users' electronic devices when the risk reaches a preset level (i.e., thermal runaway occurs). This way, in the event of thermal runaway, the fire department can receive fire-related information immediately, facilitating timely response and preventing greater losses.

[0145] The thermal runaway prevention and control method based on artificial intelligence provided in this application embodiment can acquire multimodal acquisition data of energy storage power supply in real time. The multimodal acquisition data includes data of various types and / or different formats (such as temperature data, electrical parameter data, smoke concentration and other acquisition data related to thermal runaway). By acquiring multimodal acquisition data, more data references can be provided for thermal runaway detection, which is conducive to improving the accuracy of thermal runaway detection.

[0146] It is understandable that, given the diverse data types and formats involved in multimodal acquisition, conventional control logic would be excessively complex for thermal runaway detection. Therefore, using an AI-based thermal runaway detection model to process multimodal acquisition data not only facilitates the full utilization of this data but also reduces the complexity of the control logic, thereby improving the accuracy of thermal runaway detection while lowering its implementation cost.

[0147] Finally, based on accurate thermal runaway detection results, thermal runaway warnings or fire suppression measures are implemented. When there is no risk of thermal runaway, users can clearly perceive the situation, improving the safety and perceptibility of the energy storage power supply. When there is a risk of thermal runaway, timely warnings are provided, allowing users to take preventative measures. If thermal runaway has already occurred, fire suppression measures can be implemented directly, thus achieving thermal runaway control.

[0148] In this way, thermal runaway can be detected, prevented, and controlled, improving the safety and perceptibility of energy storage power sources, reducing the risk of thermal runaway and the damage to personal safety and the economy caused by thermal runaway.

[0149] Please see Figure 10 In some embodiments, the thermal runaway prevention and control method further includes:

[0150] Step 014: After the preset time for the fire extinguishing module to spray the extinguishing agent, obtain the thermal runaway detection results again;

[0151] Step 015: If the thermal runaway detection result indicates no risk of thermal runaway, issue a message indicating successful thermal runaway handling.

[0152] The preset duration is an empirical value, which can be determined based on the time it takes for the extinguishing agent sprayed by the fire-fighting module to fill the entire energy storage power source, obtained from experiments conducted on the fire-fighting module before delivery.

[0153] It is understandable that, in order to achieve full-process user awareness of thermal runaway prevention and control, after a thermal runaway occurs in the energy storage power supply and is extinguished in time by the fire suppression module, only some of the battery cells may be damaged, while other battery cells and other components of the energy storage power supply may still function normally. Therefore, the thermal runaway detection result can be obtained again after a preset time when the fire suppression module has sprayed the extinguishing agent. If the thermal runaway has been eliminated, the thermal runaway detection result at this time becomes "no risk of thermal runaway," and a prompt message indicating successful thermal runaway handling can be issued through the first and / or second prompt devices (e.g., ...). Figure 11 The message "Thermal runaway has been successfully resolved. Please proceed to confirm!" allows users to know immediately that the thermal runaway has been eliminated, improving the user experience and further enhancing users' awareness of the safety of energy storage power supplies.

[0154] Please see Figure 12 In some embodiments, the thermal runaway prevention and control method further includes:

[0155] Step 0161: Obtain historical temperature data for each cell in the energy storage power supply and assign a number to each cell;

[0156] Step 0162: Determine the temperature anomaly information of each cell based on historical temperature data. The temperature anomaly information includes the number of times the temperature rises abnormally and / or the number of times thermal runaway occurs.

[0157] Step 0163: Count the number of abnormal temperature rises and thermal runaways of different cell numbers, and generate an anomaly statistics chart;

[0158] Step 0164: If the energy storage power supply does not experience thermal runaway, a thermal runaway early warning will be issued based on the anomaly statistics chart to prompt the user to inspect the target cells whose temperature rises abnormally more than a preset number of times in the statistics chart.

[0159] It is understandable that thermal runaway in battery cells is caused by the cells being subjected to improper operating conditions for an extended period of time. Therefore, monitoring the historical status data (such as historical temperature data) of each battery cell is extremely important for preventing thermal runaway.

[0160] Historical temperature data can be used to determine whether each cell has experienced temperature anomalies, including abnormal temperature rises and thermal runaway.

[0161] During the thermal runaway prevention phase, temperature anomalies only include abnormal temperature increases; however, after a cell experiences thermal runaway, information about the cell that first experienced thermal runaway can be obtained. Therefore, by statistically analyzing temperature anomaly information from historical temperature data, an anomaly statistics chart containing the number of temperature anomalies and thermal runaways for each cell can be obtained.

[0162] like Figure 13The anomaly statistics chart shown depicts the battery module, including cells 1 through 6. The number of temperature anomalies is S1, and the number of thermal runaways is S2. The chart clearly shows the number of temperature anomalies and thermal runaways for each cell.

[0163] During the thermal runaway prevention phase, the higher the number of abnormal temperature rises, the closer the cell is to thermal runaway. Therefore, an anomaly statistics chart can be sent to the user to prompt the user to send the target cell with more than a preset number of abnormal temperature rises (an empirical value) in the anomaly statistics chart for repair (e.g., if the preset number is 5, the target cell is cell 5). This will prevent thermal runaway and reduce the probability of thermal runaway.

[0164] Please refer to it again. Figure 12 In some embodiments, the thermal runaway prevention and control method further includes:

[0165] Step 0165: Based on the anomaly statistics charts corresponding to each energy storage power source, identify the defective cells.

[0166] It is understandable that by obtaining the anomaly statistics charts for each user's energy storage power supply, the total number of abnormal temperature rises and the total number of thermal runaways for different cell numbers can be recalculated to obtain a master statistics chart. From this master statistics chart, it can be determined whether the battery module contained defective cells at the time of manufacture. If the total number of abnormal temperature rises for any cell exceeds a first predetermined number and / or the total number of thermal runaways exceeds a second predetermined number, that cell is identified as a defective cell. This helps employees identify production defects in the battery module, thereby controlling the quality of the battery module at the source of production and further reducing the risk of thermal runaway.

[0167] like Figure 14 The overall statistics chart shown includes cells 1 to 6 in the battery module. The number of temperature anomalies is S1, and the number of thermal runaways is S2. The overall statistics chart clearly shows the number of temperature anomalies and thermal runaways for each cell, thus identifying defective cells (e.g., cell 6, with more than 3 thermal runaways, is a defective cell).

[0168] Please refer to it again. Figure 12 In some embodiments, the thermal runaway prevention and control method further includes:

[0169] Step 017: Every preset monitoring cycle, generate a health status report of the energy storage power supply and push it to the electronic devices associated with the energy storage power supply;

[0170] The preset monitoring period is a pre-defined value, such as a user-defined setting or the factory default period. For example, the preset monitoring period can be 1 day, 1 week, 1 month, etc.

[0171] The health status report includes at least one of the following: thermal runaway risk trend information of the energy storage power supply during the corresponding monitoring period; health status information of each cell of the energy storage power supply; self-test results of the fire protection module; and health status of relevant sensors used for thermal runaway detection.

[0172] Within the current monitoring period, based on the thermal runaway detection results (specifically, the thermal runaway level) collected at different times, the thermal runaway risk at different times can be determined (e.g., displayed as thermal runaway probability), thereby obtaining thermal runaway risk trend information (e.g., trend chart).

[0173] The number of temperature anomalies and thermal runaways for each cell in an energy storage power supply can be used to characterize the health status of each cell (e.g., displayed as a health rating (values ​​ranging from 0% to 100%)). For example, both the number of temperature anomalies and the number of thermal runaways are positively correlated with the health rating.

[0174] The self-test result of the fire protection module can indicate whether the fire protection module can work normally. If the self-test result is normal, the fire protection module can be triggered normally and spray the extinguishing agent; if the self-test result is abnormal, the fire protection module may not be triggered normally.

[0175] Sensors used for thermal runaway detection may include temperature sensors, smoke sensors, electrochemical sensors, and pressure sensors. The health status of these sensors refers to whether they are functioning correctly, and this health status can be categorized as normal or abnormal.

[0176] Please see Figure 15 In one example, this is the health status report S3 pushed to the user's electronic device at the end of the current monitoring period.

[0177] As can be seen from the thermal runaway risk trend chart S4 in the health status report S3, there has been no thermal runaway risk during the current monitoring period.

[0178] In the health status report S3, it can be seen from the health status information diagram S5 of each cell that the health status of each cell is greater than 95% during the current monitoring period, and there is basically no risk of thermal runaway.

[0179] In the health status report S3, the self-test results S6 of the fire protection module show that the fire protection module is operating normally and there is no risk of thermal runaway during the current monitoring period.

[0180] In the health status report S3, the health status S7 of each sensor shows that during the current monitoring period, each sensor is operating normally and there is no risk of thermal runaway.

[0181] Please see Figure 16In some embodiments, the thermal runaway prevention and control method further includes:

[0182] Step 021: Obtain the sample set, which includes samples of different thermal runaway levels and multi-modal acquisition data of the energy storage power source;

[0183] Step 022: Input samples into the initial model to obtain thermal runaway training results, which include the thermal runaway level;

[0184] Step 023: Calculate the loss value based on the thermal runaway training results and the corresponding thermal runaway level of the sample;

[0185] Step 024: Adjust the initial model based on the loss value to obtain a thermal runaway detection model that has been trained to convergence.

[0186] Specifically, the thermal runaway model needs to be trained in advance before it can be used normally. The training process is as follows:

[0187] (1) First, obtain the sample set, which includes samples of different thermal runaway levels and multi-modal acquisition data of energy storage power supply.

[0188] (2) After obtaining the sample set, model training can begin. Input the samples into the selected initial model in sequence to obtain the thermal runaway training results, which include the thermal runaway training level.

[0189] For example, the initial model may include a convolutional neural network combined with a long short-term memory network; or a convolutional neural network, a long short-term memory network combined with an attention model; or a lightweight edge model or an anomaly detection model.

[0190] (3) By comparing the thermal runaway training level in the thermal runaway training results with the thermal runaway level corresponding to the sample, the difference between the actual output of the model and the true value can be determined, i.e., the model loss value.

[0191] (4) Finally, based on the loss values ​​corresponding to different samples, the initial model is continuously trained so that the thermal runaway training level output by the initial model is closer to the thermal runaway level corresponding to the sample, until the initial model converges (e.g., the detection accuracy is greater than the preset accuracy), and the thermal runaway detection model can be obtained.

[0192] In an optional embodiment, step 023 includes:

[0193] Step 0231: Based on the proportion of samples at different thermal runaway levels, determine the weighting value corresponding to each sample at different thermal runaway levels. The weighting value corresponding to any thermal runaway level is negatively correlated with the proportion of samples at any thermal runaway level.

[0194] Step 0232: Calculate the loss value based on the thermal runaway training results, the thermal runaway level and weighting value corresponding to the sample.

[0195] Specifically, in cases of imbalanced samples with different thermal runaway levels, to ensure model performance, the influence of samples with a lower proportion on model training can be increased to avoid model overfitting and thus guarantee the detection accuracy of events with a lower proportion of thermal runaway levels.

[0196] If a corresponding weighting value can be set according to the sample ratio (the lower the sample ratio, the larger the weighting value), when calculating the loss value of each sample, the loss value is determined based on the weighting value corresponding to the sample ratio, the thermal runaway training results, and the thermal runaway level corresponding to the sample.

[0197] In one alternative embodiment, the proportions of samples with different thermal runaway levels are balanced.

[0198] In this way, by using balanced samples, overfitting of the thermal runaway detection model can be avoided, thereby improving the detection accuracy of thermal runaway events at various thermal runaway levels.

[0199] In one alternative embodiment, when the proportion of samples with different thermal runaway levels is unbalanced, a generative adversarial network is used to expand the sample size so that the proportion of samples with different thermal runaway levels is balanced.

[0200] Generative Adversarial Networks (GANs) are generative models that learn data distributions through adversarial training mechanisms. They were proposed by Ian Goodfellow in 2014. Their core idea originates from two-player zero-sum games in game theory. Through the adversarial interaction between a generator and a discriminator, the generator gradually learns to generate samples that closely resemble real-world data.

[0201] If the initial model is an anomaly detection model, the samples in the sample set do not need to be balanced. Only positive samples (such as samples that have not thermally run away or may thermally run away) need to be processed by a generative adversarial network to generate adversarial samples (i.e., negative samples, such as samples that have thermally run away).

[0202] In this way, the imbalance of samples in thermal runaway can be addressed, ensuring the performance of the thermal runaway detection model.

[0203] This application also provides a thermal runaway prevention device 300 for performing the steps described above in the artificial intelligence-based thermal runaway prevention method. Please refer to... Figure 17 , Figure 17This is a schematic diagram of a thermal runaway prevention device 300 provided in an embodiment of this application. The thermal runaway prevention device 300 includes:

[0204] The acquisition module 301 is used to acquire multimodal acquisition data of the energy storage power supply. The multimodal acquisition data includes a variety of different types and / or formats of data sets. The energy storage power supply is equipped with a fire-fighting module, which is used to spray fire extinguishing agent to extinguish fires.

[0205] The detection module 302 is used to input multimodal acquisition data into a preset thermal runaway detection model to perform thermal runaway detection and obtain thermal runaway detection results;

[0206] The prevention and control module 303 is used to provide thermal runaway warnings and / or control the fire-fighting module to spray extinguishing agents based on the thermal runaway detection results.

[0207] It should be noted that the specific details of each module unit in the above-mentioned thermal runaway prevention and control device 300 have been described in detail in the embodiments of the above-mentioned thermal runaway prevention and control method, and will not be repeated here.

[0208] In this application embodiment, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0209] In some embodiments, the thermal runaway prevention device in this application can be implemented in hardware, such as an energy storage power supply or a component in the energy storage power supply, such as an integrated circuit or a chip; the thermal runaway prevention device can also be implemented in software, such as as a terminal or an application installed in an energy storage power supply.

[0210] In some embodiments, the energy storage power supply includes a fire suppression module, a battery module, and a controller (such as the aforementioned motherboard or battery management system). The fire suppression module is used to spray extinguishing agent; the controller is connected to the battery module and the fire suppression module, and the controller is used to implement the various processes described above in the embodiments of the AI-based thermal runaway prevention and control method, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0211] In some embodiments, the electronic device includes a processor and a memory. The memory stores a computer program that can run on the processor. When executed by the processor, the program implements the various processes described above in the embodiments of the AI-based thermal runaway prevention method and achieves the same technical effects. To avoid repetition, it will not be described again here.

[0212] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described embodiments of the artificial intelligence-based thermal runaway prevention and control method, and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0213] The processor can be the processor in the energy storage power supply of the above embodiments. The computer-readable storage medium can be a computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc.

[0214] Computer-readable media can include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media include RAM, ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other solid-state storage technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape cassettes, magnetic tape, disk storage, or other magnetic storage devices. Of course, those skilled in the art will recognize that computer storage media are not limited to the above-mentioned types.

[0215] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned thermal runaway prevention method. The processor may be the processor in the energy storage power supply described in the above embodiments. When executed by the processor, the computer program implements the various processes of the embodiments of the artificial intelligence-based thermal runaway prevention method and achieves the same technical effects; therefore, to avoid repetition, further details are omitted here.

[0216] It is understood that in the specific implementation of this application, data related to user identity or characteristics is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0217] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for preventing thermal runaway based on artificial intelligence, characterized in that, include: The energy storage power supply acquires multimodal acquisition data, which includes a variety of different types and / or formats of data sets. The energy storage power supply has a built-in fire-fighting module, which is used to spray fire extinguishing agent to extinguish fires when triggered. The multimodal acquisition data is input into a preset thermal runaway detection model to perform thermal runaway detection and obtain thermal runaway detection results; Based on the thermal runaway detection results, a thermal runaway warning is issued and / or the fire-fighting module is controlled to spray the extinguishing agent.

2. The thermal runaway prevention and control method according to claim 1, characterized in that, The thermal runaway detection result includes a thermal runaway level, which is positively correlated with the thermal runaway risk of the energy storage power source. If the thermal runaway level reaches a preset level, it is determined that the energy storage power source has experienced thermal runaway. Based on the thermal runaway detection result, issuing a thermal runaway warning or controlling the fire suppression module to spray the extinguishing agent includes: Based on the thermal runaway level, a thermal runaway warning is issued, and if the thermal runaway level reaches the preset level, the fire-fighting module is controlled to spray the extinguishing agent.

3. The thermal runaway prevention and control method according to claim 2, characterized in that, The process of providing thermal runaway alerts based on the thermal runaway level includes: A thermal runaway alert is issued that matches the thermal runaway level, and the intensity of the thermal runaway alert is positively correlated with the thermal runaway level.

4. The thermal runaway prevention and control method according to claim 3, characterized in that, The thermal runaway level includes N levels, the preset level is N, the current thermal runaway level is M, and issuing a thermal runaway alert matching the thermal runaway level includes: When M=0, a Level 0 thermal runaway warning is displayed on the display screen of the energy storage power supply and / or on the status interface of the energy storage power supply displayed on the electronic device associated with the energy storage power supply. The Level 0 thermal runaway warning includes no risk of thermal runaway. When M is less than N, an M-level thermal runaway warning is issued through the first warning device of the energy storage power supply and / or the second warning device of the electronic device associated with the energy storage power supply. The M-level thermal runaway warning includes the thermal runaway probability and guidance information for preventing thermal runaway. The thermal runaway probability is positively correlated with the value of M, and the guidance information for preventing thermal runaway is determined based on the thermal runaway probability. When M reaches N, an N-level thermal runaway warning is issued through the first warning device of the energy storage power supply and / or the second warning device of the electronic device associated with the energy storage power supply. The N-level thermal runaway warning includes information on the thermal runaway of the energy storage power supply and guidance on handling the thermal runaway.

5. The thermal runaway prevention and control method according to claim 4, characterized in that, The first alerting device includes at least one of the following: a display screen, a speaker, a vibration motor, and an indicator light of the energy storage power supply. The second alerting device includes at least one of the following: a display screen, a speaker, a vibration motor, and an indicator light of the electronic device. The number of components used for alerting by the first alerting device and the second alerting device and / or the alerting intensity are positively correlated with the thermal runaway level.

6. The thermal runaway prevention and control method according to claim 1, characterized in that, The thermal runaway detection result includes a thermal runaway level, which is positively correlated with the thermal runaway risk of the energy storage power source. The method further includes: Every preset monitoring cycle, a health status report of the energy storage power supply is generated and pushed to the electronic devices associated with the energy storage power supply; The health status report includes at least one of the following: thermal runaway risk trend information of the energy storage power supply during the corresponding monitoring period; health status information of each cell of the energy storage power supply; self-test results of the fire protection module; and health status of relevant sensors used for thermal runaway detection.

7. The thermal runaway prevention and control method according to claim 1, characterized in that, Also includes: After a preset time has elapsed since the fire-fighting module sprayed the extinguishing agent, the thermal runaway detection result is acquired again. If the thermal runaway detection result indicates no risk of thermal runaway, a message indicating successful thermal runaway handling will be issued.

8. The thermal runaway prevention and control method according to claim 1, characterized in that, The thermal runaway detection model includes a first feature extraction module and a second feature extraction module. The first feature extraction module includes a convolutional neural network, and the second feature extraction module includes a long short-term memory network. The process of inputting the multimodal acquisition data into the preset thermal runaway detection model to perform thermal runaway detection and obtain thermal runaway detection results includes: The multimodal acquisition data is input to the first feature extraction module and the second feature extraction module respectively, so as to extract the first feature vector through the convolutional neural network and extract the second feature vector through the long short-term memory network; The first and second feature vectors are concatenated to generate the third feature vector; Thermal runaway detection is performed based on the third feature vector to obtain the obtained thermal runaway detection result.

9. The thermal runaway prevention and control method according to claim 8, characterized in that, The thermal runaway detection model further includes an attention module and a fully connected classification module. The thermal runaway detection based on the third feature vector to obtain the thermal runaway detection result includes: The attention module assigns different weights to different types of data in the third feature vector to obtain a fourth feature vector, wherein the weight of the first type of data is greater than the weight of the second type of data, and the first type of data includes at least data related to the concentration of hydrogen fluoride. The fourth feature vector is processed by a fully connected classification module to obtain the thermal runaway detection result.

10. The thermal runaway prevention and control method according to claim 1, characterized in that, The thermal runaway detection model includes a lightweight edge model or an anomaly detection model.

11. The thermal runaway prevention and control method according to claim 1, characterized in that, The multimodal acquisition data includes at least one of temperature data, smoke data, battery electrical parameter data, electrochemical gas data, and air pressure data.

12. The thermal runaway prevention and control method according to claim 11, characterized in that, The energy storage power supply also includes a temperature sensing wire and a battery module. The temperature sensing wire is installed in the battery module. The temperature data includes battery temperature data and fuse failure data. The fuse failure data includes whether the temperature sensing wire has melted.

13. The thermal runaway prevention and control method according to claim 11, characterized in that, The energy storage power supply is also equipped with a smoke sensor, which is used to collect smoke data.

14. The thermal runaway prevention and control method according to claim 11, characterized in that, The energy storage power supply is also equipped with an electrochemical sensor, which is used to collect electrochemical gas data, including the type and concentration of thermal runaway marker gas.

15. The thermal runaway prevention and control method according to claim 1, characterized in that, The energy storage power supply also includes a battery module and a pressure sensor. The battery module includes multiple battery cells, and the pressure sensor is installed inside each battery cell. The pressure sensor is used to collect pressure data, including the internal pressure of the battery cell.

16. The thermal runaway prevention and control method according to claim 1, characterized in that, The energy storage power supply also includes a controller, which is connected to the electric initiator of the fire-fighting module. Controlling the fire-fighting module to spray the extinguishing agent includes: An electric start signal is sent to the electric initiator to activate the extinguishing agent in the fire-fighting module, causing the extinguishing agent to be sprayed out.

17. The thermal runaway prevention and control method according to claim 1, characterized in that, Also includes: Acquire historical temperature data for each cell in the energy storage power supply and assign a number to each cell; Based on the historical temperature data, temperature anomaly information for each cell is determined, including the number of abnormal temperature rises and / or the number of thermal runaways. The number of abnormal temperature rises and the number of thermal runaways for different cell numbers are counted, and an anomaly statistics chart is generated. In the absence of thermal runaway in the energy storage power supply, a thermal runaway early warning is issued based on the abnormal statistics chart. The thermal runaway early warning includes maintenance suggestions for each cell.

18. The thermal runaway prevention and control method according to claim 17, characterized in that, Also includes: Based on the anomaly statistics charts corresponding to each of the energy storage power sources, defective cells are identified.

19. The thermal runaway prevention and control method according to claim 1, characterized in that, Also includes: Acquire a sample set, which includes samples of different thermal runaway levels, and the samples include multimodal acquisition data of the energy storage power source; Input the sample into the initial model to obtain the thermal runaway training result, which includes the thermal runaway training level; Based on the thermal runaway training results and the thermal runaway level corresponding to the sample, the loss value is calculated; The initial model is adjusted based on the loss value to obtain the thermal runaway detection model that has been trained to convergence.

20. The thermal runaway prevention and control method according to claim 19, characterized in that, The proportions of samples with different thermal runaway levels are balanced.

21. The thermal runaway prevention and control method according to claim 19, characterized in that, The acquisition of the sample set includes: When the proportion of samples with different thermal runaway levels is unbalanced, a generative adversarial network is used to expand the sample size so that the proportion of samples with different thermal runaway levels is balanced.

22. The thermal runaway prevention and control method according to claim 19, characterized in that, The calculation of the loss value based on the thermal runaway training results and the thermal runaway level corresponding to the sample includes: Based on the proportion of samples at different thermal runaway levels, weighted values ​​corresponding to samples at different thermal runaway levels are determined respectively. The weighted value corresponding to any thermal runaway level is negatively correlated with the proportion of samples at that thermal runaway level. The loss value is calculated based on the thermal runaway training results, the thermal runaway level corresponding to the sample, and the weighting value.

23. An energy storage power source, characterized in that, include: Firefighting module, the firefighting module being used to spray fire extinguishing agent; Battery module; and A controller, which is connected to the battery module and the fire protection module, is used to implement the thermal runaway prevention and control method as described in any one of claims 1-22.

24. An electronic device, characterized in that, include: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the thermal runaway prevention method as described in any one of claims 1-22.

25. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the thermal runaway prevention method as described in any one of claims 1-22.

Images