Sulfide all-solid-state battery production safety control method, device and electronic equipment

Abstract

The application provides a sulfide all-solid-state battery production safety control method, device and electronic equipment, and belongs to the technical field of all-solid-state battery production. The method comprises the following steps: acquiring energy efficiency state data, mechanical state data and hydrogen sulfide concentration gradient parameters of a sulfide all-solid-state battery production line; extracting temperature time domain features and voltage time domain features of the energy efficiency state data, and extracting vibration frequency domain features of the mechanical state data; inputting the temperature time domain features, the voltage time domain features, the vibration frequency domain features and the hydrogen sulfide concentration gradient parameters into a risk assessment model to obtain a hydrogen sulfide leakage risk assessment result output by the risk assessment model; wherein the risk assessment model is obtained by training based on temperature time domain feature samples, voltage time domain feature samples, vibration frequency domain feature samples and hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels. The application fuses multi-dimensional features to predict hydrogen sulfide leakage risk, and can realize accurate risk prediction.

Description

Technical Field

[0001] This invention relates to the field of all-solid-state battery production technology, and in particular to a method, apparatus and electronic device for safety control in the production of sulfide all-solid-state batteries. Background Technology

[0002] Sulfide-based all-solid-state batteries have attracted much attention due to their high energy density and safety. However, during the production process, sulfide electrolytes readily react with moisture in the air to generate highly toxic hydrogen sulfide gas. Hydrogen sulfide has strong neurotoxicity, and its occupational exposure limit is extremely low. It not only threatens the lives of personnel on site but may also corrode production equipment.

[0003] Current production line safety control primarily relies on a single gas sensor to detect hydrogen sulfide concentration and uses safety devices for negative pressure control to suppress leaks. However, this approach suffers from significant lag, typically triggering alarms and emergency responses only after a leak occurs and the concentration reaches a threshold. Furthermore, existing technologies struggle to capture early warning signs of leaks, failing to meet the dynamic safety requirements under complex and multi-condition operating conditions, and hindering accurate risk prediction and proactive defense. Summary of the Invention

[0004] This invention provides a method, apparatus, and electronic device for safety control in the production of sulfide all-solid-state batteries, which addresses the shortcomings of existing safety control methods for sulfide all-solid-state battery production, which struggle to achieve accurate risk prediction.

[0005] This invention provides a method for safety control in the production of sulfide all-solid-state batteries, comprising:

[0006] Acquire energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and obtain hydrogen sulfide concentration gradient parameters in the production environment;

[0007] Extract the time-domain features of temperature and voltage from the energy efficiency status data, and extract the frequency-domain features of vibration from the mechanical status data;

[0008] The temperature time-domain characteristics, voltage time-domain characteristics, vibration frequency-domain characteristics, and hydrogen sulfide concentration gradient parameters are input into the risk assessment model to obtain the hydrogen sulfide leakage risk assessment result output by the risk assessment model.

[0009] The risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels.

[0010] In some embodiments, the risk assessment model includes a thermal runaway risk assessment layer, a mechanical failure risk assessment layer, and a comprehensive risk assessment layer;

[0011] The thermal runaway risk assessment layer is used to: assess the thermal runaway risk based on the temperature time-domain characteristics and the voltage time-domain characteristics, and obtain the thermal runaway risk probability;

[0012] The mechanical failure risk assessment layer is used to: assess the mechanical failure risk based on the vibration frequency domain characteristics, and obtain the mechanical failure risk probability;

[0013] The comprehensive risk assessment layer is used to comprehensively assess the risk of hydrogen sulfide leakage based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, and to obtain the risk assessment result of hydrogen sulfide leakage.

[0014] In some embodiments, assessing the thermal runaway risk based on the temperature time-domain characteristics and the voltage time-domain characteristics to obtain the thermal runaway risk probability includes:

[0015] Determine the weights of the temperature time-domain features and the voltage time-domain features;

[0016] Based on the weights of the temperature time-domain features and the voltage time-domain features, the temperature time-domain features and the voltage time-domain features are fused to obtain the fused thermal features;

[0017] Based on the fusion thermal characteristics, the thermal runaway risk of the sulfide all-solid-state battery production line is assessed, and the probability of the thermal runaway risk is obtained.

[0018] In some embodiments, the step of comprehensively assessing the hydrogen sulfide leakage risk based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter to obtain the hydrogen sulfide leakage risk assessment result includes:

[0019] The thermal runaway risk probability is normalized to obtain a normalized thermal runaway risk probability; the mechanical failure risk probability is normalized to obtain a normalized mechanical failure risk probability; and the hydrogen sulfide concentration gradient parameter is normalized to obtain a normalized hydrogen sulfide concentration gradient parameter.

[0020] The first weight of the normalized thermal runaway risk probability, the second weight of the normalized mechanical failure risk probability, and the third weight of the normalized hydrogen sulfide concentration gradient parameter are determined.

[0021] Based on the first weight, the second weight, and the third weight, the normalized thermal runaway risk probability, the normalized mechanical failure risk probability, and the normalized hydrogen sulfide concentration gradient parameter are integrated to obtain the hydrogen sulfide leakage risk assessment result.

[0022] In some embodiments, extracting the temperature time-domain and voltage time-domain features of the energy efficiency status data, and extracting the vibration frequency-domain features of the mechanical status data, includes:

[0023] The energy efficiency status data is preprocessed to obtain preprocessed energy efficiency status data. Features are extracted from the preprocessed energy efficiency status data to obtain the temperature time-domain features and the voltage time-domain features.

[0024] The mechanical state data is preprocessed to obtain preprocessed mechanical state data, and features are extracted from the preprocessed mechanical state data to obtain the vibration frequency domain features.

[0025] In some embodiments, after obtaining the hydrogen sulfide leakage risk assessment result output by the risk assessment model, the method further includes:

[0026] If a risk of hydrogen sulfide leakage is determined, an early warning message is generated to issue an alarm based on the hydrogen sulfide leakage risk assessment results, and / or a safety control command is generated.

[0027] In some embodiments, after obtaining the hydrogen sulfide leakage risk assessment result output by the risk assessment model, the method further includes:

[0028] Obtain user feedback information;

[0029] Based on the user feedback information, the parameters of the risk assessment model are optimized.

[0030] In some embodiments, the risk assessment model is trained based on the following steps:

[0031] Acquire energy efficiency status data samples and mechanical status data samples of the sulfide all-solid-state battery production line, and acquire hydrogen sulfide concentration gradient parameter samples in the production environment.

[0032] Extract the temperature time-domain feature samples and voltage time-domain feature samples corresponding to the energy efficiency status data samples, and extract the vibration frequency-domain feature samples corresponding to the mechanical status data samples;

[0033] Determine the corresponding label for the hydrogen sulfide leakage risk assessment result;

[0034] Using the temperature time-domain feature samples, the voltage time-domain feature samples, the vibration frequency-domain feature samples, and the hydrogen sulfide concentration gradient parameter samples as training samples, and using the hydrogen sulfide leakage risk assessment result labels as sample labels, an initial risk assessment model is trained. After training, the risk assessment model is obtained.

[0035] This invention also provides a safety control device for the production of sulfide all-solid-state batteries, comprising:

[0036] The acquisition unit is used to acquire energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and to acquire hydrogen sulfide concentration gradient parameters in the production environment.

[0037] The feature extraction unit is used to extract the temperature time-domain features and voltage time-domain features of the energy efficiency status data, and to extract the vibration frequency-domain features of the mechanical status data.

[0038] The prediction unit is used to input the temperature time-domain characteristics, the voltage time-domain characteristics, the vibration frequency-domain characteristics, and the hydrogen sulfide concentration gradient parameters into the risk assessment model to obtain the hydrogen sulfide leakage risk assessment result output by the risk assessment model.

[0039] The risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels.

[0040] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the safety control method for the production of sulfide all-solid-state batteries as described above.

[0041] The present invention provides a method, apparatus, and electronic device for safety control in the production of sulfide all-solid-state batteries. This method acquires energy efficiency status data, mechanical status data, and hydrogen sulfide concentration gradient parameters from the sulfide all-solid-state battery production line; extracts the time-domain and voltage characteristics of the energy efficiency status data, and extracts the vibration frequency domain characteristics of the mechanical status data; inputs these characteristics into a risk assessment model, which then integrates the multi-dimensional features to conduct a risk assessment, resulting in a hydrogen sulfide leakage risk assessment. This enables accurate risk prediction. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 This is a schematic flowchart of the safety control method for the production of sulfide all-solid-state batteries provided in this embodiment of the invention.

[0044] Figure 2This is a flowchart illustrating the training process of the risk assessment model provided in this embodiment of the invention.

[0045] Figure 3 This is a schematic diagram of the structure of the sulfide all-solid-state battery production safety control device provided in an embodiment of the present invention.

[0046] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0048] The terms "first," "second," etc., used in this invention are used to distinguish similar objects, not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of the invention can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, in this invention, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0049] This invention provides a safety control system for the production of sulfide all-solid-state batteries. The system includes: a physical protection subsystem, a risk control subsystem, an intelligent monitoring subsystem, and a control center subsystem.

[0050] The physical protection subsystem forms the basic barrier of the production environment, used to physically isolate hydrogen sulfide leaks. Specifically, this subsystem includes a sealed enclosure with an observation window and a silicone sealing ring. In addition, the sealed enclosure is equipped with a movable door, which employs a three-layer independent sealing structure: a fluororubber O-ring, a polyurethane filling layer, and an epoxy resin infusion layer.

[0051] The ventilation control subsystem regulates airflow and pressure within the microenvironment, actively intervening in leak diffusion. This subsystem consists of a distributed array of fan filter units (FFUs). It is equipped with an air handling unit that employs a combined process of refrigeration dehumidification and molecular sieve adsorption to maintain a stable low dew point temperature, thus inhibiting the chemical reaction of sulfides. The exhaust system is designed with a dual-loop structure, including process exhaust ducts and emergency exhaust ducts.

[0052] The intelligent monitoring subsystem is used to collect multi-dimensional status data. This subsystem includes, but is not limited to, electrochemical sensors, temperature sensors, voltage sensors, mechanical vibration sensors, and differential pressure sensors. The electrochemical sensor is used to monitor sulfide concentration, and the differential pressure sensor is used to monitor wind pressure. The temperature sensor, voltage sensor, and mechanical vibration sensor are used to monitor the temperature, voltage, and mechanical vibration of the production equipment, respectively.

[0053] The control center subsystem integrates an edge computing module, which deploys a risk prediction model to predict the risk of hydrogen sulfide leakage based on multi-dimensional features. The control center subsystem includes a main control unit, which controls relevant actuators based on the risk assessment results predicted by the model to prevent hydrogen sulfide leakage.

[0054] This invention provides a safety control method for the production of sulfide all-solid-state batteries. This method is mainly applied to the manufacturing process of sulfide all-solid-state batteries, particularly in critical processes prone to hydrogen sulfide leakage, such as sulfide electrolyte treatment, electrode coating, and rolling. This invention aims to solve the alarm lag problem caused by relying on single gas concentration detection in existing technologies, and achieves accurate risk prediction and proactive defense by integrating multi-dimensional physical quantities.

[0055] Figure 1 This is a schematic flowchart illustrating the safety control method for the production of sulfide all-solid-state batteries provided in an embodiment of the present invention. Figure 1 As shown, a method for safety control in the production of sulfide all-solid-state batteries is provided, including the following steps: step 110, step 120, and step 130. This method's steps are merely one possible implementation of the present invention.

[0056] Step 110: Obtain the energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and obtain the hydrogen sulfide concentration gradient parameters in the production environment.

[0057] The sulfide all-solid-state battery production line covers a complete equipment system from raw material mixing to cell assembly, especially including equipment that easily generates heat or mechanical stress, such as coating machines and rolling mills.

[0058] Energy efficiency status data refers to data reflecting the energy conversion efficiency and thermodynamic state of production equipment during operation. In this embodiment, energy efficiency status data may specifically include, but is not limited to, real-time temperature data of key components of the equipment, and voltage, current, or power data of the equipment drive unit. For example, real-time temperature data of the rollers and battery materials can be collected by miniature thermocouples deployed in the roller pressing area of ​​the roller press, and voltage data of the drive motor of the roller press can be collected by voltage sensors.

[0059] Mechanical condition data refers to data reflecting the physical structural integrity, operational stability, and sealing performance of production equipment. In this embodiment, mechanical condition data mainly refers to the equipment's vibration data. Specifically, vibration acceleration signals can be collected using acceleration sensors installed in the equipment's bearing housing, motor housing, or sealed connections.

[0060] The hydrogen sulfide concentration gradient parameter refers to the rate of change of hydrogen sulfide gas concentration over time or space within the production environment, particularly the microenvironment enclosure. Real-time concentration values ​​can be obtained through a network of electrochemical sensors deployed at different heights on the production line or near critical leak points, and the rate of change can be calculated. It should be noted that although this embodiment emphasizes the fusion of multi-dimensional data, the hydrogen sulfide concentration itself remains an important benchmark reference.

[0061] Step 120: Extract the time-domain features of temperature and voltage from the energy efficiency status data, and extract the frequency-domain features of vibration from the mechanical status data.

[0062] In this step, the collected raw data needs to be processed by feature engineering to extract feature indicators that are strongly correlated with leakage risk.

[0063] Temperature time-domain characteristics refer to statistical quantities that characterize the trend of temperature change over time. For example, the temperature slope, maximum temperature difference, or temperature variance within a set time window can be extracted. Its physical meaning is that if a device experiences thermal runaway due to increased friction or a localized short circuit in battery materials, it often first manifests as an abnormally rapid rise in temperature. Voltage time-domain characteristics refer to statistical quantities that characterize voltage stability over time. For example, voltage ripple amplitude and voltage fluctuation rate can be extracted. Its physical meaning is that when a device experiences uneven mechanical load or electrical faults, the driving voltage often exhibits abnormal fluctuations.

[0064] Vibration frequency domain features refer to the features extracted after converting time-domain vibration signals into frequency-domain signals through signal transformation. For example, a Fast Fourier Transform (FFT) can be performed on mechanical state data to extract vibration amplitude or spectral energy in specific frequency ranges, such as bearing fault characteristic frequencies and meshing frequencies. Its physical significance lies in the fact that seal failure is often accompanied by wear or loosening of mechanical components; these mechanical faults exhibit significant fingerprint characteristics in the frequency domain, such as enhanced high-frequency vibration.

[0065] Understandably, by extracting the time-domain features of temperature and voltage from energy efficiency status data, and the frequency-domain features of vibration from mechanical status data, early signals of complex failure modes can be keenly captured, laying a data foundation for accurate predictions in subsequent risk assessment models.

[0066] In some embodiments, step 120 extracts the temperature time-domain features and voltage time-domain features of the energy efficiency status data, and extracts the vibration frequency-domain features of the mechanical status data, including:

[0067] Step 121: Preprocess the energy efficiency status data to obtain preprocessed energy efficiency status data. Extract features from the preprocessed energy efficiency status data to obtain temperature time-domain features and voltage time-domain features.

[0068] Because raw data collected from the production site may contain issues such as noise interference, inconsistent dimensions, and reference drift, preprocessing of the energy efficiency status data is necessary as the first step.

[0069] Optionally, a sliding time window is set, for example, with a window length of 5 minutes and a sliding step of 1 second, to divide the continuous time series data into a series of data segments; the original data within the window is standardized to eliminate baseline differences under different devices or different operating conditions, and preprocessed energy efficiency status data is obtained.

[0070] Optionally, for temperature data, the linear regression coefficients (i.e., the temperature slope) within the window can be calculated, or the maximum temperature difference between thermocouples at different locations within the same workstation can be calculated. The temperature slope is used to identify trends of excessively rapid temperature rise. For example, under normal operating conditions, temperature changes are gradual, while precursors to thermal runaway often manifest as an abnormally sharp increase in the temperature slope. The maximum temperature difference is used to identify localized overheating phenomena.

[0071] In this embodiment, the voltage amplitude index extracted after transformation is classified as one of the voltage time-domain feature descriptions, which reflects the intensity of voltage fluctuations within a specific time window.

[0072] Step 122: Preprocess the mechanical state data to obtain preprocessed mechanical state data, and extract features from the preprocessed mechanical state data to obtain vibration frequency domain features.

[0073] Similar to energy efficiency status data, mechanical status data also requires preprocessing. The preprocessing process also includes sliding window truncation and standardization to remove high-frequency noise interference, resulting in preprocessed mechanical status data.

[0074] Optionally, a Fast Fourier Transform (FFT) is performed on the preprocessed time-domain vibration signal to convert it from the time domain to the frequency domain. The amplitude values ​​at the inherent characteristic frequencies of key mechanical components are extracted. For example, the amplitude values ​​at the inherent characteristic frequencies of rotating components such as bearings and gears are extracted. The root mean square value or the sum of the spectral energy of the acceleration data within the window is calculated as a comprehensive indicator to assess the severity of the overall mechanical vibration.

[0075] Step 130: Input the time-domain characteristics of temperature, the time-domain characteristics of voltage, the frequency-domain characteristics of vibration, and the hydrogen sulfide concentration gradient parameters into the risk assessment model to obtain the hydrogen sulfide leakage risk assessment results output by the risk assessment model.

[0076] The risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels.

[0077] The core of this step lies in using a risk assessment model to perform nonlinear fusion analysis on the aforementioned multi-source heterogeneous characteristics.

[0078] This risk assessment model is a predictive model built on machine learning or deep learning algorithms. It preferably adopts a neural network structure that can process time series data, such as Long Short-Term Memory (LSTM) network, Gated Recurrent Unit (GRU), etc.

[0079] The model's input layer receives the aforementioned time-domain features of temperature, time-domain features of voltage, frequency-domain features of vibration, and hydrogen sulfide concentration gradient parameters. The model's output layer outputs a probability value or risk level characterizing the likelihood of a hydrogen sulfide leak occurring within a future period, i.e., the hydrogen sulfide leak risk assessment result.

[0080] In the specific training process of the risk assessment model, historical production data can be collected. Historical data from the period preceding the leak event is used as positive samples, while data from normal operation is used as negative samples. Labels can be binary labels or continuous risk indices. Through training, the model can learn the complex nonlinear mapping relationship between abnormal temperature increases, increased voltage ripple, enhanced vibration at specific frequencies, and hydrogen sulfide leaks.

[0081] In this embodiment of the invention, instead of relying solely on hydrogen sulfide gas concentration for post-event alarms, the energy efficiency and mechanical status of the equipment are innovatively introduced as early warning signals. Since mechanical failure and thermal runaway are often precursors to hydrogen sulfide leaks, and changes in these physical quantities often precede the diffusion of gas concentration, this embodiment of the invention can achieve accurate pre-emptive prediction of hydrogen sulfide leak risks. This multi-dimensional feature fusion assessment method effectively avoids false alarms caused by environmental interference from a single sensor, significantly improving safety and emergency response speed in the all-solid-state battery production process.

[0082] In some embodiments, the risk assessment model includes a thermal runaway risk assessment layer, a mechanical failure risk assessment layer, and a comprehensive risk assessment layer;

[0083] The thermal runaway risk assessment layer is used to: assess the thermal runaway risk based on the time-domain characteristics of temperature and the time-domain characteristics of voltage, and obtain the probability of thermal runaway risk;

[0084] The mechanical failure risk assessment layer is used to: assess the mechanical failure risk based on vibration frequency domain characteristics and obtain the mechanical failure risk probability;

[0085] The comprehensive risk assessment layer is used to comprehensively assess the risk of hydrogen sulfide leakage based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, and to obtain the hydrogen sulfide leakage risk assessment result.

[0086] The thermal runaway risk assessment layer and the mechanical failure risk assessment layer can employ LSTM or a one-dimensional convolutional neural network. The thermal runaway risk assessment layer is primarily used to capture early signs of thermal runaway caused by battery material defects or equipment overload. The mechanical failure risk assessment layer is used to identify mechanical failures that may lead to seal failure.

[0087] For example, when the temperature slope in the roll forming area is detected to increase sharply from 0.1℃ / min to 0.8℃ / min, accompanied by an increase in voltage ripple amplitude, the thermal runaway risk assessment layer can identify this abnormal thermo-electric coupling mode and output a high probability of thermal runaway risk. This probability directly reflects the likelihood of a thermal runaway event occurring in the near future.

[0088] For example, if a significant increase in amplitude is detected at the characteristic frequency of bearing failure, it indicates that the bearing is worn or has changed clearance. This can easily lead to uneven stress on the sealing gasket at the connection, causing it to fail. The mechanical failure risk assessment layer then outputs the probability of mechanical failure risk based on this.

[0089] The comprehensive risk assessment layer, consisting of a fully connected layer, is the final decision-making unit of the model. It is used to integrate the various sub-risks with environmental parameters, capture the synergistic effect of multiple factors, and output the final hydrogen sulfide leakage risk assessment result.

[0090] Through this hierarchical assessment mechanism, the embodiments of the present invention can not only output the final risk assessment results, but also use the output of the intermediate layer to help determine whether the source of risk is overheating or mechanical failure, thereby guiding maintenance personnel to conduct targeted troubleshooting.

[0091] In some embodiments, the risk of thermal runaway is assessed based on temperature time-domain characteristics and voltage time-domain characteristics to obtain the probability of thermal runaway risk, including:

[0092] Determine the weights of the time-domain features of temperature and the time-domain features of voltage;

[0093] Based on the weights of temperature time-domain features and voltage time-domain features, the temperature time-domain features and voltage time-domain features are fused to obtain fused thermal features;

[0094] Based on the thermal characteristics of fusion, the thermal runaway risk of a sulfide all-solid-state battery production line is assessed, and the probability of thermal runaway risk is obtained.

[0095] Specifically, when assessing the risk of thermal runaway, although both temperature changes and voltage fluctuations point to potential thermal failures, their contributions differ at different stages of the failure. For example, in the early stages of a micro-short circuit inside the battery, voltage ripple may appear earlier than temperature rise; while in the material decomposition stage, the temperature slope contributes more. Therefore, the weights can be determined through attention mechanisms or pre-defined expert rules.

[0096] For example, a fully connected layer network can dynamically output weights based on the current eigenvalue magnitude. When extremely severe voltage ripple is detected, higher weights are automatically assigned to the voltage characteristics to highlight the inducing effect of electrical faults on thermal runaway.

[0097] Optionally, the temperature time-domain features are normalized to obtain normalized temperature time-domain features; the voltage time-domain features are normalized to obtain normalized voltage time-domain features.

[0098] The weights of the normalized temperature time-domain features and the normalized voltage time-domain features are determined. Based on the weights of the normalized temperature time-domain features and the normalized voltage time-domain features, the normalized temperature time-domain features and the normalized voltage time-domain features are fused to obtain the fused thermal features.

[0099] Optionally, the fused thermal features are input into the classifier or regressor of the thermal runaway assessment layer. The thermal runaway assessment layer calculates the probability value that the current state belongs to the precursor of thermal runaway, i.e., the probability of thermal runaway risk, based on the distribution location of the fused thermal features in the potential space.

[0100] Through this weighted fusion mechanism, this embodiment can effectively handle the signal competition problem between multi-source data, ensuring that even when a single feature is not obvious but the combined features are abnormal, it can still keenly capture the early signals of thermal runaway.

[0101] In some embodiments, the risk of hydrogen sulfide leakage is comprehensively assessed based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, resulting in a hydrogen sulfide leakage risk assessment result, including:

[0102] The probability of thermal runaway risk is normalized to obtain the normalized probability of thermal runaway risk; the probability of mechanical failure risk is normalized to obtain the normalized probability of mechanical failure risk; the hydrogen sulfide concentration gradient parameter is normalized to obtain the normalized hydrogen sulfide concentration gradient parameter.

[0103] The first weight of the normalized thermal runaway risk probability, the second weight of the normalized mechanical failure risk probability, and the third weight of the normalized hydrogen sulfide concentration gradient parameter are determined.

[0104] Based on the first, second, and third weights, the normalized thermal runaway risk probability, the normalized mechanical failure risk probability, and the normalized hydrogen sulfide concentration gradient parameter are integrated to obtain the hydrogen sulfide leakage risk assessment result.

[0105] Optionally, the first, second, and third weights can be determined based on a dynamic weighting strategy or a collaborative enhancement strategy. For example, an attention mechanism can be used to dynamically calculate the first, second, and third weights. Another example is that when both the normalized probability of thermal runaway and the normalized probability of mechanical failure are higher than preset thresholds, the first and second weights can be increased using a nonlinear function to reflect the multiplicative effect brought about by multiphysics coupling.

[0106] Through this comprehensive evaluation mechanism, the embodiments of the present invention have achieved a leap from single-parameter alarm to multi-dimensional proactive risk management, which not only avoids false alarms caused by fluctuations of a single sensor, but also ensures that the highest level of safety response can be triggered in a timely manner under complex fault conditions.

[0107] In some embodiments, after obtaining the hydrogen sulfide leakage risk assessment results output by the risk assessment model, the method further includes:

[0108] If a risk of hydrogen sulfide leakage is identified, an early warning message is generated based on the risk assessment results, and / or a safety control command is generated.

[0109] Optionally, safety control commands include, but are not limited to, raw material blocking commands, shutdown commands, ventilation commands, chemical neutralization commands, and inert gas injection commands.

[0110] Optionally, the hydrogen sulfide leakage risk assessment result includes the probability of hydrogen sulfide leakage risk. If the probability of hydrogen sulfide leakage risk is less than a first preset threshold, it is determined to be a safe state. If the probability of hydrogen sulfide leakage risk is greater than or equal to the first preset threshold and less than a second preset threshold, it is determined to be a warning state, generating a warning message for alarm, and / or generating a safety control command to automatically increase the exhaust volume. If the probability of hydrogen sulfide leakage risk is greater than or equal to the second preset threshold, it is determined to be a high-risk state, generating a warning message for alarm, and / or generating a safety control command to immediately trigger emergency responses such as shutdown, valve shut-off, and activation of emergency exhaust. The first preset threshold is less than the second preset threshold.

[0111] Through the above implementation methods, the embodiments of the present invention can not only inform users of risks, but also transform the prediction results into active control actions of the equipment, intervene in the early stage of an accident, and significantly reduce the safety hazards in the production of sulfide all-solid-state batteries.

[0112] In some embodiments, after obtaining the hydrogen sulfide leakage risk assessment results output by the risk assessment model, the method further includes:

[0113] Obtain user feedback information;

[0114] Based on user feedback, the parameters of the risk assessment model were optimized.

[0115] User feedback information mainly comes from confirmation operations by on-site operators or safety experts, and specific scenarios include false alarm feedback, missed alarm feedback, and confirmation of new operating conditions.

[0116] Specifically, the collected historical data with user feedback tags, along with some representative historical samples, are combined to form a fine-tuning dataset. The fine-tuning dataset is then used to backpropagate and train the current risk assessment model, enabling the model to learn new failure modes or correct misjudgments of specific interference signals.

[0117] By introducing user feedback and online learning mechanisms, this invention addresses the concept drift problem commonly encountered in industrial applications of deep learning models. This means the system can continuously evolve as the production line operates, automatically adapting to the effects of equipment wear, seasonal environmental changes, and process fine-tuning, maintaining a high level of prediction accuracy and significantly reducing downtime losses due to false alarms and safety risks due to missed alarms.

[0118] Figure 2 This is a flowchart illustrating the training process of the risk assessment model provided in an embodiment of the present invention. Figure 2 As shown, in some embodiments, the risk assessment model is trained based on the following steps:

[0119] Step 210: Obtain energy efficiency status data samples and mechanical status data samples of the sulfide all-solid-state battery production line, and obtain hydrogen sulfide concentration gradient parameter samples in the production environment.

[0120] Optionally, the energy efficiency status data sample includes temperature data samples and pressure data samples.

[0121] Step 220: Extract the temperature time-domain feature samples and voltage time-domain feature samples corresponding to the energy efficiency status data samples, and extract the vibration frequency-domain feature samples corresponding to the mechanical status data samples.

[0122] Optionally, the temperature time-domain feature samples include at least the temperature slope and the maximum temperature difference, and the vibration frequency-domain feature samples include at least the characteristic frequency amplitude and the total vibration value.

[0123] Step 230: Determine the corresponding hydrogen sulfide leakage risk assessment result label;

[0124] Step 240: Using temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, and hydrogen sulfide concentration gradient parameter samples as training samples, and using the hydrogen sulfide leakage risk assessment result labels as sample labels, train the initial risk assessment model. After training, the risk assessment model is obtained.

[0125] Optionally, the initial risk assessment model includes an initial thermal runaway risk assessment layer, an initial mechanical failure risk assessment layer, and an initial comprehensive risk assessment layer.

[0126] Optionally, determine the probability labels for thermal runaway risk and mechanical failure risk; train an initial thermal runaway risk assessment layer using temperature time-domain feature samples and voltage time-domain feature samples as training samples and thermal runaway risk probability labels as sample labels; train an initial mechanical failure risk assessment layer using vibration frequency-domain feature samples as training samples and mechanical failure risk probability labels as sample labels.

[0127] The training method of this invention enables the model to learn complex nonlinear mapping relationships between features of different dimensions. For example, the model can learn that at a specific temperature slope, even if the vibration amplitude does not exceed the standard, it may indicate an extremely high risk of seal failure, thereby significantly improving the accuracy and robustness of risk assessment.

[0128] The following describes the sulfide all-solid-state battery production safety control device provided in the embodiments of the present invention. The sulfide all-solid-state battery production safety control device described below can be referred to in correspondence with the sulfide all-solid-state battery production safety control method described above.

[0129] Figure 3 This is a schematic diagram of the structure of the safety control device for the production of sulfide all-solid-state batteries provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the sulfide all-solid-state battery production safety control device 300 includes:

[0130] The acquisition unit 310 is used to acquire the energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and to acquire the hydrogen sulfide concentration gradient parameters in the production environment.

[0131] The feature extraction unit 320 is used to extract the temperature time-domain features and voltage time-domain features of the energy efficiency status data, and to extract the vibration frequency-domain features of the mechanical status data.

[0132] The prediction unit 330 is used to input the time-domain characteristics of temperature, the time-domain characteristics of voltage, the frequency-domain characteristics of vibration, and the hydrogen sulfide concentration gradient parameters into the risk assessment model to obtain the hydrogen sulfide leakage risk assessment results output by the risk assessment model.

[0133] The risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels.

[0134] Optionally, the risk assessment model includes a thermal runaway risk assessment layer, a mechanical failure risk assessment layer, and a comprehensive risk assessment layer;

[0135] The thermal runaway risk assessment layer is used to: assess the thermal runaway risk based on the time-domain characteristics of temperature and the time-domain characteristics of voltage, and obtain the probability of thermal runaway risk;

[0136] The mechanical failure risk assessment layer is used to: assess the mechanical failure risk based on vibration frequency domain characteristics and obtain the mechanical failure risk probability;

[0137] The comprehensive risk assessment layer is used to comprehensively assess the risk of hydrogen sulfide leakage based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, and to obtain the hydrogen sulfide leakage risk assessment result.

[0138] Optionally, based on the time-domain characteristics of temperature and the time-domain characteristics of voltage, the risk of thermal runaway is assessed to obtain the probability of thermal runaway risk, including:

[0139] Determine the weights of the time-domain features of temperature and the time-domain features of voltage;

[0140] Based on the weights of temperature time-domain features and voltage time-domain features, the temperature time-domain features and voltage time-domain features are fused to obtain fused thermal features;

[0141] Based on the thermal characteristics of fusion, the thermal runaway risk of a sulfide all-solid-state battery production line is assessed, and the probability of thermal runaway risk is obtained.

[0142] Optionally, based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameters, a comprehensive assessment of hydrogen sulfide leakage risk is obtained, including:

[0143] The probability of thermal runaway risk is normalized to obtain the normalized probability of thermal runaway risk; the probability of mechanical failure risk is normalized to obtain the normalized probability of mechanical failure risk; the hydrogen sulfide concentration gradient parameter is normalized to obtain the normalized hydrogen sulfide concentration gradient parameter.

[0144] The first weight of the normalized thermal runaway risk probability, the second weight of the normalized mechanical failure risk probability, and the third weight of the normalized hydrogen sulfide concentration gradient parameter are determined.

[0145] Based on the first, second, and third weights, the normalized thermal runaway risk probability, the normalized mechanical failure risk probability, and the normalized hydrogen sulfide concentration gradient parameter are integrated to obtain the hydrogen sulfide leakage risk assessment result.

[0146] Optionally, the time-domain features of temperature and voltage of the energy efficiency status data are extracted, and the frequency-domain features of vibration of the mechanical status data are extracted, including:

[0147] The energy efficiency status data is preprocessed to obtain preprocessed energy efficiency status data. Feature extraction is performed on the preprocessed energy efficiency status data to obtain temperature time-domain features and voltage time-domain features.

[0148] The mechanical state data is preprocessed to obtain preprocessed mechanical state data. Feature extraction is then performed on the preprocessed mechanical state data to obtain vibration frequency domain features.

[0149] Optionally, the sulfide all-solid-state battery production safety control device 300 also includes:

[0150] The generation unit is used to generate early warning information for alarm purposes based on the hydrogen sulfide leakage risk assessment results when a hydrogen sulfide leakage risk is determined to exist, and / or generate safety control instructions.

[0151] Optionally, the sulfide all-solid-state battery production safety control device 300 also includes:

[0152] Feedback unit, used to obtain user feedback information;

[0153] The optimization unit is used to optimize the parameters of the risk assessment model based on user feedback.

[0154] Optionally, the risk assessment model is trained based on the following steps:

[0155] Acquire energy efficiency status data samples and mechanical status data samples of the sulfide all-solid-state battery production line, and acquire hydrogen sulfide concentration gradient parameter samples in the production environment.

[0156] Extract the time-domain features of temperature and voltage corresponding to the energy efficiency status data samples, and extract the vibration frequency-domain features corresponding to the mechanical status data samples;

[0157] Determine the corresponding label for the hydrogen sulfide leakage risk assessment result;

[0158] The initial risk assessment model was trained using temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, and hydrogen sulfide concentration gradient parameter samples, with the hydrogen sulfide leakage risk assessment result labels as sample labels. After training, the risk assessment model was obtained.

[0159] It should be noted that the sulfide all-solid-state battery production safety control device provided in this embodiment of the invention can realize all the method steps implemented in the above-mentioned sulfide all-solid-state battery production safety control method embodiment, and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0160] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440. The processor 410, communication interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a safety control method for the production of sulfide all-solid-state batteries. This method includes: acquiring energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and acquiring hydrogen sulfide concentration gradient parameters within the production environment; extracting temperature time-domain and voltage time-domain features from the energy efficiency status data, and extracting vibration frequency-domain features from the mechanical status data; inputting the temperature time-domain features, voltage time-domain features, vibration frequency-domain features, and hydrogen sulfide concentration gradient parameters into a risk assessment model to obtain a hydrogen sulfide leakage risk assessment result output by the risk assessment model; wherein the risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels.

[0161] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0162] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0163] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for safety control in the production of sulfide all-solid-state batteries, characterized in that, include: Acquire energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and obtain hydrogen sulfide concentration gradient parameters in the production environment; Extract the time-domain features of temperature and voltage from the energy efficiency status data, and extract the frequency-domain features of vibration from the mechanical status data; The temperature time-domain characteristics, voltage time-domain characteristics, vibration frequency-domain characteristics, and hydrogen sulfide concentration gradient parameters are input into the risk assessment model to obtain the hydrogen sulfide leakage risk assessment result output by the risk assessment model. The risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels. The risk assessment model includes a thermal runaway risk assessment layer, a mechanical failure risk assessment layer, and a comprehensive risk assessment layer. The thermal runaway risk assessment layer is used to: assess the thermal runaway risk based on the temperature time-domain characteristics and the voltage time-domain characteristics, and obtain the thermal runaway risk probability; The mechanical failure risk assessment layer is used to: assess the mechanical failure risk based on the vibration frequency domain characteristics, and obtain the mechanical failure risk probability; The comprehensive risk assessment layer is used to comprehensively assess the risk of hydrogen sulfide leakage based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, and to obtain the risk assessment result of hydrogen sulfide leakage.

2. The method for safety control in the production of sulfide all-solid-state batteries according to claim 1, characterized in that, The process of assessing the risk of thermal runaway based on the time-domain characteristics of temperature and the time-domain characteristics of voltage, and obtaining the probability of thermal runaway risk, includes: Determine the weights of the temperature time-domain features and the voltage time-domain features; Based on the weights of the temperature time-domain features and the voltage time-domain features, the temperature time-domain features and the voltage time-domain features are fused to obtain the fused thermal features; Based on the aforementioned thermal characteristics, the thermal runaway risk of the sulfide all-solid-state battery production line is assessed, and the probability of the thermal runaway risk is obtained.

3. The method for safety control in the production of sulfide all-solid-state batteries according to claim 1, characterized in that, The hydrogen sulfide leakage risk is comprehensively assessed based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, resulting in a hydrogen sulfide leakage risk assessment result, including: The thermal runaway risk probability is normalized to obtain a normalized thermal runaway risk probability; the mechanical failure risk probability is normalized to obtain a normalized mechanical failure risk probability; and the hydrogen sulfide concentration gradient parameter is normalized to obtain a normalized hydrogen sulfide concentration gradient parameter. The first weight of the normalized thermal runaway risk probability, the second weight of the normalized mechanical failure risk probability, and the third weight of the normalized hydrogen sulfide concentration gradient parameter are determined. Based on the first weight, the second weight, and the third weight, the normalized thermal runaway risk probability, the normalized mechanical failure risk probability, and the normalized hydrogen sulfide concentration gradient parameter are integrated to obtain the hydrogen sulfide leakage risk assessment result.

4. The method for safety control in the production of sulfide all-solid-state batteries according to claim 1, characterized in that, The extraction of temperature and voltage time-domain features from the energy efficiency status data, and the extraction of vibration frequency-domain features from the mechanical status data, include: The energy efficiency status data is preprocessed to obtain preprocessed energy efficiency status data. Features are extracted from the preprocessed energy efficiency status data to obtain the temperature time-domain features and the voltage time-domain features. The mechanical state data is preprocessed to obtain preprocessed mechanical state data, and features are extracted from the preprocessed mechanical state data to obtain the vibration frequency domain features.

5. The method for safety control in the production of sulfide all-solid-state batteries according to claim 1, characterized in that, After obtaining the hydrogen sulfide leakage risk assessment result output by the risk assessment model, the process also includes: If a risk of hydrogen sulfide leakage is determined, an early warning message is generated to issue an alarm based on the hydrogen sulfide leakage risk assessment results, and / or a safety control command is generated.

6. The method for safety control in the production of sulfide all-solid-state batteries according to claim 1, characterized in that, After obtaining the hydrogen sulfide leakage risk assessment result output by the risk assessment model, the process also includes: Obtain user feedback information; Based on the user feedback information, the parameters of the risk assessment model are optimized.

7. The method for safety control in the production of sulfide all-solid-state batteries according to claim 1, characterized in that, The risk assessment model is trained based on the following steps: Acquire energy efficiency status data samples and mechanical status data samples of the sulfide all-solid-state battery production line, and acquire hydrogen sulfide concentration gradient parameter samples in the production environment. Extract the temperature time-domain feature samples and voltage time-domain feature samples corresponding to the energy efficiency status data samples, and extract the vibration frequency-domain feature samples corresponding to the mechanical status data samples; Determine the corresponding label for the hydrogen sulfide leakage risk assessment result; Using the temperature time-domain feature samples, the voltage time-domain feature samples, the vibration frequency-domain feature samples, and the hydrogen sulfide concentration gradient parameter samples as training samples, and using the hydrogen sulfide leakage risk assessment result labels as sample labels, an initial risk assessment model is trained. After training, the risk assessment model is obtained.

8. A safety control device for the production of sulfide all-solid-state batteries, characterized in that, include: The acquisition unit is used to acquire energy efficiency status data and mechanical status data of the sulfide all-solid-state battery production line, and to acquire hydrogen sulfide concentration gradient parameters in the production environment. The feature extraction unit is used to extract the temperature time-domain features and voltage time-domain features of the energy efficiency status data, and to extract the vibration frequency-domain features of the mechanical status data. The prediction unit is used to input the temperature time-domain characteristics, the voltage time-domain characteristics, the vibration frequency-domain characteristics, and the hydrogen sulfide concentration gradient parameters into the risk assessment model to obtain the hydrogen sulfide leakage risk assessment result output by the risk assessment model. The risk assessment model is trained based on temperature time-domain feature samples, voltage time-domain feature samples, vibration frequency-domain feature samples, hydrogen sulfide concentration gradient parameter samples, and hydrogen sulfide leakage risk assessment result labels. The risk assessment model includes a thermal runaway risk assessment layer, a mechanical failure risk assessment layer, and a comprehensive risk assessment layer. The thermal runaway risk assessment layer is used to: assess the thermal runaway risk based on the temperature time-domain characteristics and the voltage time-domain characteristics, and obtain the thermal runaway risk probability; The mechanical failure risk assessment layer is used to: assess the mechanical failure risk based on the vibration frequency domain characteristics, and obtain the mechanical failure risk probability; The comprehensive risk assessment layer is used to comprehensively assess the risk of hydrogen sulfide leakage based on the probability of thermal runaway, the probability of mechanical failure, and the hydrogen sulfide concentration gradient parameter, and to obtain the risk assessment result of hydrogen sulfide leakage.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the sulfide all-solid-state battery production safety control method as described in any one of claims 1 to 7.

Images