Gas sensor, preparation method, detection method and thermal runaway monitoring method

Abstract

The invention relates to a MOFs-based gas sensor, a preparation method, a detection method and a thermal runaway monitoring method, the MOFs-based gas sensor comprises a plurality of sensing units, and each sensing unit comprises a silicon substrate layer; the nano structure is positioned on the surface of the silicon substrate layer; the MOFs gas sensitive layer covers the surface of the nano structure, and is used for adsorbing target gas and carrying out chemical reaction with the adsorbed target gas through an MOFs material so as to cause the change of an electric signal; the interdigital electrode is positioned on the surface of the silicon substrate layer and is used for amplifying the change of the electric signal based on electric field distribution; a packaging layer, wherein the packaging layer covers the sensing unit; wherein the packaging layer is provided with a porous structure allowing gas to permeate. The gas sensor provided by the invention has the characteristics of high sensitivity and high response speed, and can effectively detect the gas type and concentration.

Description

Technical Field

[0001] The present application relates to the technical field of gas sensors, and in particular to a MOFs-based gas sensor, a preparation method, a detection method, and a thermal runaway monitoring method. Background Art

[0002] With the rapid development of new energy vehicles, lithium batteries, with their high energy density and long cycle life, have become an important energy source for electric vehicles, energy storage systems, and other fields. However, lithium batteries are prone to thermal runaway under conditions such as overcharging and short circuiting, rapidly releasing gases such as CO, H2, CH4, and C2H4. Changes in the concentration of these gases can serve as early warning signals of thermal runaway. Therefore, rapid and accurate detection of these gases is crucial to improving lithium battery safety.

[0003] Currently, sensors for gas detection primarily include infrared gas sensors, electrochemical gas sensors, and semiconductor gas sensors. Infrared gas sensors primarily detect gas molecules based on their absorption characteristics at specific infrared wavelengths. However, these sensors rely on optical systems, which make the sensors bulky and expensive, making them difficult to integrate with lithium battery packs. Furthermore, they lack sensitivity to low-concentration gases. Electrochemical gas sensors primarily measure gas concentrations based on electrochemical reactions. However, their electrolytes are prone to volatilization and their electrodes are prone to passivation, resulting in a short service life (typically less than two years), making them inadequate for long-term monitoring of battery gas leaks. Semiconductor gas sensors require high temperatures of 200°C to 400°C to fully exploit their sensing properties. This not only increases energy consumption but can also interfere with the battery's thermal management system, adding additional risks. Furthermore, existing gas sensors suffer from several other issues, including slow response speeds (>10 seconds) and insufficient simultaneous multi-gas detection capabilities. However, the thermal runaway process of lithium batteries is sudden and spreads rapidly. Existing gas sensors are unable to capture sudden changes in gas concentration in a timely manner, which can easily lead to delayed warnings and cause immeasurable losses. Summary of the Invention

[0004] In view of the above problems in the prior art, the present application provides a MOFs-based gas sensor, preparation method, detection method and thermal runaway monitoring method. The gas sensor has the advantages of small size, long service life, high sensitivity, fast response speed, and support for simultaneous detection of multiple gases.

[0005] To achieve the above-mentioned objectives, the first aspect of the present application provides a MOFs-based gas sensor, comprising: a plurality of sensing units, each sensing unit comprising: a silicon base layer; a nanostructure, the nanostructure being located on the surface of the silicon base layer; a MOFs gas-sensitive layer, the MOFs gas-sensitive layer covering the surface of the nanostructure, for adsorbing the target gas and reacting chemically with the adsorbed target gas through the MOFs material to cause a change in the electrical signal; interdigitated electrodes, the interdigitated electrodes being located on the surface of the silicon base layer, for amplifying the change in the electrical signal based on the electric field distribution; an encapsulation layer, the encapsulation layer covering the sensing unit; wherein the encapsulation layer has a porous structure that allows gas to pass through.

[0006] As described above, in the presence of a target gas, the MOFs material in the MOFs gas-sensitive layer reacts with the target gas, causing changes in the electrical signal at the interdigital electrodes to achieve gas detection. This gas sensor has a small size, long service life, high sensitivity, and fast response speed. Because it includes multiple sensing units, each made of a different MOFs material, it can support the simultaneous detection of multiple gases.

[0007] As an implementation of this aspect, the gas-sensitive layers of the multiple sensing units are made of different MOFs materials doped with different metal oxides.

[0008] As an implementation of this aspect, the multiple sensing units include: a first sensing unit, a second sensing unit, a third sensing unit, a fourth sensing unit, a fifth sensing unit and / or a sixth sensing unit; the gas-sensitive layer of the first sensing unit is made of ZIF-8 doped SnO2; the gas-sensitive layer of the second sensing unit is made of HKUST-1 doped ZnO; the gas-sensitive layer of the third sensing unit is made of MOF-74 (Ni) doped WO3; the gas-sensitive layer of the fourth sensing unit is made of UiO-66 doped TiO2; the gas-sensitive layer of the fifth sensing unit is made of MIL-101 (Cr) doped Fe2O3; and the gas-sensitive layer of the sixth sensing unit is made of ZIF-67 doped Co3O4.

[0009] From the above, different sensing units are made of different materials, so that multiple gas detections can be achieved simultaneously.

[0010] As an implementation of this aspect, the nanostructure includes: a nanoarray composed of multiple nanopillars, a nanoarray composed of multiple nanowires, or a nanoarray composed of multiple nanoholes; wherein, in the nanoarray, the ratio of the height to the width of each nanomonomer exceeds a first threshold.

[0011] From the above, by setting a nanostructure with a larger aspect ratio, the surface area of the MOFs gas-sensitive layer can be increased, thereby improving the adsorption capacity.

[0012] As an implementation method of this aspect, it also includes: a detection device, which is connected to the interdigital electrode; the detection device is used to determine the concentration and type of the target gas based on the change of the electrical signal; the detection device includes: a signal acquisition unit, which is used to acquire electrical signals of multiple data channels based on a preset sampling frequency to generate time series data sets of multiple data channels, wherein different data channels correspond to different sensing units, and different data channels correspond to different time series data sets; a signal preprocessing unit, which is used to perform signal denoising and normalization processing on the time series data sets of the multiple data channels to obtain multiple standardized time series data sets; a feature extraction unit, which is used to extract the feature sequences of each of the multiple standardized time series data sets based on a convolutional neural network; a gas identification unit, which is used to determine the target gas type probability and target gas concentration prediction value corresponding to the feature sequence based on a long short-term memory network, and determine the gas type with the highest probability as the detected gas type, and use the gas concentration prediction value as the detected gas concentration; an early warning unit, which is used to issue early warning signals of different levels according to the concentration and type of the target gas.

[0013] From the above, the detection accuracy and detection speed can be improved through the above detection method.

[0014] A second aspect of the present application provides a method for preparing a gas sensor, comprising: making a gold mask on a silicon substrate layer by photolithography technology, and etching other areas on the silicon substrate layer by metal-assisted etching technology to obtain a nanostructure, wherein the other areas include areas not covered by the gold mask; in-situ synthesizing an initial MOFs gas-sensitive layer by MOFs material and metal oxide on the nanostructure by layer-by-layer deposition, and high-temperature activation of the initial MOFs gas-sensitive layer in an inert gas environment to obtain a final MOFs gas-sensitive layer; making interdigitated electrodes on the silicon substrate layer by electron beam evaporation technology and photolithography technology; and covering the surface of the sensing unit with a porous structure allowing gas to pass as an encapsulation layer by plasma-enhanced chemical vapor deposition technology.

[0015] As an implementation of this aspect, it also includes: the MOFs material is made by a solvent hot melt method.

[0016] As an implementation of this aspect, the interdigital electrodes are connected to a detection device by extracting electrode signals through a gold wire bonding process.

[0017] The beneficial effects of this aspect can also be found in the description of the beneficial effects of each part of the first aspect above.

[0018] The third aspect of the present application provides a method for gas detection using the sensor described in any one of the first aspects above, comprising: collecting electrical signals of multiple data channels based on a preset sampling frequency to generate time series data sets of multiple data channels, wherein different data channels correspond to different sensing units, and different data channels correspond to different time series data sets; performing signal denoising and normalization processing on the time series data sets of the multiple data channels to obtain multiple standardized time series data sets; extracting the characteristic sequences of each of the multiple standardized time series data sets based on a convolutional neural network, and determining the gas type probability and gas concentration prediction value corresponding to the characteristic sequence based on a long short-term memory network; determining the gas type with the highest probability as the detected gas type, and using the gas concentration prediction value as the detected gas concentration.

[0019] The beneficial effects of this aspect can also be found in the description of the beneficial effects of each part of the first aspect above.

[0020] The fourth aspect of the present application provides a method for monitoring thermal runaway of a lithium battery, the method comprising: using the MOFs-based gas sensor described in any one of the first aspects above to monitor the type and concentration of gas released by the lithium battery, and judging whether thermal runaway occurs in the lithium battery based on the type and concentration of gas.

[0021] The beneficial effects of this aspect can also be found in the description of the beneficial effects of each part of the first aspect above. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The following further illustrates the various technical features of the present application and the relationships between them with reference to the accompanying drawings. The accompanying drawings are exemplary, and some technical features are not shown in actual proportion. In addition, some drawings may omit technical features that are commonly used in the technical field to which the present application belongs and are not essential for understanding and implementing the present application, or additional technical features that are not essential for understanding and implementing the present application may be shown. In other words, the combination of the various technical features shown in the accompanying drawings is not intended to limit the present application. In addition, throughout the present application, the same figure numbers refer to the same content. The specific description of the drawings is as follows:

[0023] Figure 1a A schematic diagram of the three-dimensional structure of a MOFs-based gas sensor provided in an embodiment of the present application;

[0024] Figure 1b An enlarged view of the three-dimensional structure of a MOFs-based gas sensor provided in an embodiment of the present application;

[0025] Figure 2An exemplary diagram of a gas sensor based on MOFs provided in an embodiment of the present application including multiple sensing units;

[0026] Figure 3a A schematic diagram of the preparation process of the nanostructure provided in the embodiments of the present application;

[0027] Figure 3b Schematic diagram of the preparation process of the MOFs gas-sensitive layer provided in the embodiment of the present application;

[0028] Figure 4 Flowchart of the gas detection method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The technical solution provided by this application is further described below with reference to the accompanying drawings and examples. It should be understood that the sensor structure and business scenarios provided in the examples of this application are mainly for illustrating possible implementation methods of the technical solution of this application and should not be interpreted as the sole limitation of the technical solution of this application. It is known to those skilled in the art that with the evolution of sensor structures and the emergence of new business scenarios, the technical solution provided by this application is also applicable to similar technical problems.

[0030] It should be understood that the embodiments of this application provide a MOFs-based gas sensor solution. Because these technical solutions solve the same or similar problems, some repetitions may not be repeated in the following specific embodiments. However, these specific embodiments should be considered as cross-references and can be combined with each other.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art in the art of this application. In the event of any inconsistency, the meaning described in this specification or the meaning derived from the contents recorded in this specification shall prevail. In addition, the terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit this application.

[0032] The MOFs-based gas sensor solution provided in this application is mainly used in the field of monitoring thermal runaway of lithium batteries. During the thermal runaway process of lithium batteries, gases such as CO, H2, CH4, and C2H4 will be released rapidly. Their concentration changes earlier than the temperature changes, which is a key signal for early warning of thermal runaway. Therefore, this application integrates a MOFs-based gas sensor into a lithium battery module or battery management system. Different sensing units in the sensor are used to detect different types of gases. This can not only achieve simultaneous detection of multiple gases, but also improve detection sensitivity and response speed, thereby providing real-time and accurate monitoring of lithium battery thermal runaway.

[0033] The embodiment of the present application provides a gas sensor based on MOFs. Figure 1a 、 Figure 1b and Figure 2 The gas sensor is introduced in detail. Figure 1a is a schematic diagram of the three-dimensional structure of the gas sensor. Figure 1b This is an enlarged view of the three-dimensional structure of the gas sensor. Figure 2 A gas sensor includes multiple sensing units.

[0034] like Figure 1a 、 Figure 1b and Figure 2 As shown, the gas sensor includes a plurality of sensing units (eg Figure 2 The first sensing unit a to the sixth sensing unit f) are shown, and each sensing unit includes a silicon base layer 110, a nanostructure 120, a MOFs gas sensitive layer 130, an interdigitated electrode 140 and a packaging layer (the packaging layer is not shown in the figure in order to show the structure of each layer of the sensor). In this embodiment, the gas sensor also includes a detection device (not shown). Among them, the multiple sensing units can share the same silicon base layer 110 and the same packaging layer; of course, the multiple sensing units can also be respectively provided with independent silicon base layers 110 and independent packaging layers. In addition, the multiple sensing units can share the same detection device, or an independent detection device can be connected to each sensing unit. The various structures of the gas sensor are described in detail below.

[0035] The silicon base layer 110 is used to provide support for each sensing unit.

[0036] Nanostructure 120 is located on the surface of silicon substrate layer 110. Nanostructure 120 can be a nanoarray structure composed of multiple nanopillars, a nanowire, or a nanopore. In this embodiment, since nanostructure 120 serves as a carrier for MOF gas-sensing layer 130, in order to increase the surface area and enhance the adsorption capacity of gas molecules, the ratio of the height to width of each nanomonomer (a nanomonomer refers to a single nanopillar, a single nanowire, or a single nanopore in a nanoarray) is set to exceed a first threshold. For example, the first threshold can range from 10:1 to 50:1, preferably 20:1, i.e., the ratio of the height to width of each nanomonomer is set to 20:1. Specifically, the width (also referred to as the diameter) of each nanomonomer can be set to the order of hundreds of nanometers (e.g., approximately 100 nanometers), and the height of each nanomonomer can be set to the order of micrometers (e.g., approximately 2 μm). The spacing between different nanomonomers in the same nanoarray can also be set to between 100 nm and 500 nm. In this embodiment, the spacing is set to 200 nm.

[0037] The MOFs gas-sensing layer 130 covers the surface of the nanostructure 120 and is used to adsorb the target gas and react chemically with the adsorbed target gas through the MOFs material, thereby causing changes in electrical signals (such as resistivity, current, inductance, etc.). In this embodiment, the gas-sensing layers of different sensing units in the sensor are made of different MOFs materials doped with different metal oxides. For example:

[0038] (1) For the first sensing unit a: its gas-sensitive layer is made of ZIF-8 doped with SnO2, and can be used to adsorb (detect) gases such as CO and H2.

[0039] (2) For the second sensing unit b: its gas-sensitive layer is made of HKUST-1 doped with ZnO, which can be used to adsorb gases such as H2 and CH4.

[0040] (3) For the third sensing unit c: its gas-sensitive layer is made of MOF-74 (Ni) doped with WO3, which can be used to adsorb gases such as CH4 and H2.

[0041] (4) For the fourth sensing unit d: its gas-sensitive layer is made of UiO-66 doped with TiO2, which can be used to adsorb gases such as C2H4 and CO.

[0042] (5) For the fifth sensing unit e: its gas-sensitive layer is made of MIL-101 (Cr) doped with Fe2O3, which can be used to adsorb gases such as C2H4 and CH4;

[0043] (6) For the sixth sensor f: its gas-sensitive layer is made of ZIF-67 doped with Co3O4, and can be used to adsorb gases such as H2 and CO.

[0044] The interdigitated electrodes 140 are located on the surface of the silicon substrate layer 110 and are used to amplify changes in electrical signals based on their own electric field distribution, thereby improving detection sensitivity. The electrode width of the interdigitated electrodes 140 can be set to between 3μm and 10μm, and the spacing between the two electrodes can be set to between 2μm and 8μm. In this embodiment, the electrode width of the interdigitated electrodes 140 is set to 5μm, the electrode spacing is set to 3μm, the number of electrode pairs is 6, and the electrode material is platinum (Pt).

[0045] The encapsulation layer covers the sensing unit. To allow gas to pass through the encapsulation layer and contact the MOFs gas-sensing layer 130, in this embodiment, the encapsulation layer is selected to have a porous structure that allows gas to pass through, such as a porous SiO2 film (e.g., 500nm thick, 0.8μm pore size). This encapsulation layer can protect the MOFs gas-sensing layer 130, while preventing dust interference while allowing gas to pass through. In this embodiment, the electrode signal of the interdigitated electrode 140 can be extracted through a gold wire bonding process.

[0046] The detection device is connected to the interdigital electrodes 140 and is used to determine the concentration and type of the target gas based on the detected electrical signal changes. The detection device includes a signal acquisition unit, a signal preprocessing unit, a feature extraction unit, a gas identification unit, and an early warning unit. Specifically: the signal acquisition unit is used to collect electrical signals from multiple data channels based on a preset sampling frequency to generate time series data sets for multiple data channels, wherein different data channels correspond to different sensing units, and different data channels correspond to different time series data sets; the signal preprocessing unit is used to perform signal denoising and normalization processing on the time series data sets of the multiple data channels of the sensing units to obtain multiple standardized time series data sets; the feature extraction unit is used to extract the feature sequences of the multiple standardized time series data sets of the sensing units based on a convolutional neural network; the gas identification unit is used to determine the target gas type probability and target gas concentration prediction value corresponding to the sensor unit feature sequence based on a long short-term memory network, and determine the gas type with the highest probability of the sensing unit as the detected gas type, and use the predicted value of the gas concentration of the sensing unit as the detected gas concentration; the early warning unit is used to issue early warning signals of different levels according to the concentration and type of the target gas in the sensing unit. For the specific gas detection process, please refer to the detailed description of the gas detection part below.

[0047] In the gas sensor provided by the embodiments of this application, when a target gas is present, the MOFs material in the MOFs gas-sensing layer reacts with the target gas, causing changes in the electrical signal of the interdigital electrodes, enabling rapid gas detection. Furthermore, the sensor of this embodiment includes multiple sensing units, each made of a different MOFs material, thus enabling simultaneous detection of multiple gases.

[0048] Next, combine Figure 3a and Figure 3b The preparation method of the gas sensor is introduced below. The preparation method includes the following steps:

[0049] A: Preparation of nanostructures 120: Figure 3a As shown, a gold mask is made on the silicon base layer 110 by photolithography technology, and metal assisted etching technology (MacEtch) is used to etch other areas on the silicon base layer 110 to obtain nanostructures 120, and the other areas include areas not covered by the gold mask.

[0050] B: Preparation of MOFs gas-sensitive layer 130: Figure 3b As shown, MOFs material is first prepared by solvent hot melting method, and MOFs material and metal oxide are in situ synthesized into an initial MOFs gas-sensitive layer on the nanostructure 130 by layer-by-layer deposition method, for example Figure 3bAs shown, a layer of metal oxide is first sprayed onto the nanostructure 120 to form a metal center site layer, and then a layer of MOFs material is sprayed onto the nanostructure 120 to form a ligand layer. After repeatedly forming an initial MOFs gas-sensing layer, the initial MOFs gas-sensing layer is subjected to high-temperature activation in an inert gas environment (e.g., nitrogen N2, argon Ar, helium He) to obtain a final stable MOFs gas-sensing layer 130. The high-temperature activation temperature can be 300°C to 500°C, and the activation time can be 30 minutes to 60 minutes. In this embodiment, the temperature is 400°C and the activation time is 45 minutes.

[0051] C: Fabrication of interdigital electrodes 140: A 100 nm platinum (Pt) thin film is deposited on the surface of the silicon substrate 110, and then fabricated using electron beam evaporation and photolithography. Furthermore, gold wire bonding is required to connect interdigital electrodes 140 to the detection device by extracting the electrode signal.

[0052] D: Preparation of Encapsulation Layer: A porous structure with gas permeability is deposited on the surface of the sensor element by plasma-enhanced chemical vapor deposition (PECVD). In this embodiment, the encapsulation layer can be, for example, a porous SiO2 film (500 nm thick, 0.8 μm pore size).

[0053] Next, combine Figure 4 The following describes a method for gas detection using the gas sensor (i.e., the detection principle). The method mainly includes three steps: S410 data acquisition and preprocessing, S420 feature extraction, and S430 time series analysis. Specifically:

[0054] In the S410 data acquisition and preprocessing step: the sensor is placed in the environment to be tested. When the target gas exists in the environment to be tested, the MOFs gas-sensitive layer of the sensor contacts the target gas and produces a chemical reaction, thereby causing the electrical signal of the interdigital electrode to change. The interdigital electrode amplifies the electrical signal change through its own electric field distribution. Then, the detection device connected to the interdigital electrode synchronously collects the electrical signals of multiple data channels based on a preset sampling frequency to generate a time series data set of multiple data channels, wherein different data channels correspond to different sensing units, and different data channels correspond to different time series data sets.

[0055] As an implementation method, the collected real-time data (ie, time series data sets of multiple data channels) can be stored for subsequent use.

[0056] Then, the time series data sets of multiple data channels are subjected to signal denoising and normalization processing respectively to obtain multiple standardized time series data sets.

[0057] In the feature extraction step S420: feature sequences of respective multiple standardized time series data sets are extracted based on a convolutional neural network (CNN).

[0058] In the S430 time series analysis step: based on the long short-term memory network (LSTM), the gas type probability and gas concentration prediction value corresponding to the feature sequence are determined, and the gas type with the highest probability is determined as the detected gas type, and the gas concentration prediction value is used as the detected gas concentration.

[0059] In order to better understand the detection algorithm, a specific example is given below to illustrate.

[0060] First, a gas detection model was constructed based on a convolutional neural network (CNN) and a long short-term memory (LSTM) model. The gas detection model was trained using a training dataset consisting of sensor response data (or response patterns) for different gases at different concentrations.

[0061] Next, we will introduce the process of detecting gas types and gas concentrations using this gas detection model:

[0062] S410 data acquisition and preprocessing:

[0063] (1) Data collection:

[0064] Input: resistance change rate signal of each sensor unit (ΔR / R0, where R0 is the initial resistance value and ΔR is the real-time resistance change);

[0065] Acquisition tool: 16-bit analog-to-digital converter (ADC), sampling frequency set to 100 Hz, input voltage range 0-5 V;

[0066] Output: Generate a time series data set including each sensor unit. Each sensor unit collects 100 data points per second, and a total of 600 data points are collected in 6 seconds. Therefore, the 6 sensor units have 6*600 data points.

[0067] (2) Signal denoising:

[0068] Method: Wavelet transform, the specific parameters are as follows: wavelet basis function: Daubechies wavelet; decomposition level: 3 layers; denoising threshold setting: Where σ is the standard deviation of noise, N is the number of sampling points (N = 600 for each sensor unit corresponding to the data set);

[0069] Output: Denoised smoothed time series that preserves gas response characteristics and suppresses ambient noise.

[0070] (3) Normalization:

[0071] Method: Use Min-Max normalization to linearly map the denoised signal to the interval [0,1];

[0072] Output: Standardized time series set, eliminating dimensional differences between sensor units and improving convergence efficiency.

[0073] S420 feature extraction:

[0074] (1) Data reorganization:

[0075] Input: Normalized time series of multi-sensor units (number of data points: 6*600);

[0076] Method: The normalized time series set of multiple sensor units is converted into a two-dimensional matrix with the dimension of the number of sensor units * the number of sampling points of each sensor unit (e.g. 6*600);

[0077] Output: 2D tensor, suitable for input format of convolutional neural network (CNN).

[0078] (2) Convolutional Neural Network (CNN) Feature Extraction:

[0079] Network architecture: 3-layer convolutional neural network, the specific parameters are as follows:

[0080] First convolutional layer: 32 3×1 convolution kernels, stride 1, activation function ReLU;

[0081] First pooling layer: 2×1 max pooling, stride 2;

[0082] Second convolutional layer: 64 3×1 convolution kernels, stride 1, activation function ReLU;

[0083] Second pooling layer: 2×1 maximum pooling, stride 2;

[0084] The third convolutional layer: 128 3×1 convolution kernels, stride 1, activation function ReLU;

[0085] Output: After convolution and pooling operations, a three-dimensional spatial feature map (dimensions: 1,128,146) is generated.

[0086] (3) Feature sequence tiling processing:

[0087] Method: The 3D feature map (1,128,146) output by CNN is tiled by time step, and the dimension is (1,146,128).

[0088] Output: A three-dimensional tensor (dimensions: 1,146,128), suitable for time series modeling requirements of long short-term memory (LSTM) networks.

[0089] S430 Time Series Analysis (LSTM):

[0090] (1) LSTM network architecture:

[0091] A two-layer LSTM network with 64 units per layer was used. The activation function used was the tanh function for the hidden state and the sigmoid function for the gating unit. Regularization was performed with a dropout ratio of 0.2 to prevent overfitting.

[0092] Input data: feature sequence extracted by CNN;

[0093] Output: time series hidden state sequence (dimension: 1,64);

[0094] (2) Fully connected layer:

[0095] Input: Time series hidden state sequence of LSTM output (dimension: 1, 64);

[0096] Output layer: The number of neurons is equal to the number of gas categories + the number of concentration levels;

[0097] Activation function: Softmax is used to output the probability distribution of gas types; activation function Linear is used to output the predicted value of gas concentration;

[0098] Output results: gas type probability distribution and gas concentration prediction value.

[0099] (3) Classification and regression:

[0100] For gas types: output the category with the highest probability as the gas type;

[0101] For gas concentration: Output the linear predicted gas concentration as gas concentration.

[0102] In this embodiment, a sliding window technique may be used to update the input data, for example, updating the input data every 0.5 seconds, thereby increasing the real-time nature of the data and ensuring the timeliness of the detection.

[0103] In the above example, the detection algorithm of S410 to S430 improves the speed of gas recognition through the CNN-LSTM hybrid model.

[0104] The present application also provides a method for detecting thermal runaway in a lithium battery, comprising integrating the aforementioned sensor with a battery management system, utilizing the sensor to monitor the type and concentration of gases released by the lithium battery in real time, and determining whether thermal runaway has occurred in the lithium battery based on the type and concentration of the gases. For example, when the concentration of a gas detected exceeds a preset concentration (e.g., CO > 50 ppm and / or H2 > 100 ppm), a graded alarm is automatically triggered.

[0105] Note that the above are only preferred embodiments of the present application and the technical principles employed. Those skilled in the art will understand that the present application is not limited to the specific embodiments herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the scope of protection of the present application, all of which fall within the scope of protection of the present application.

Claims

1. A gas sensor based on MOFs, characterized in that: include: Multiple sensing units, each sensing unit includes: Silicon base layer; A nanostructure, wherein the nanostructure is located on the surface of the silicon base layer; A MOFs gas-sensitive layer, the MOFs gas-sensitive layer covering the surface of the nanostructure, for adsorbing target gas and chemically reacting with the adsorbed target gas through the MOFs material to cause a change in the electrical signal; interdigital electrodes, the interdigital electrodes being located on a surface of the silicon substrate layer and configured to amplify changes in the electrical signal based on an electric field distribution; A packaging layer covers the sensing unit; wherein the packaging layer has a porous structure that allows gas to pass through.

2. The gas sensor according to claim 1, wherein The gas-sensitive layers of the multiple sensing units are made of different MOFs materials doped with different metal oxides.

3. The gas sensor according to claim 2, characterized in that The plurality of sensing units include: a first sensing unit, a second sensing unit, a third sensing unit, a fourth sensing unit, a fifth sensing unit and / or a sixth sensing unit; The gas-sensitive layer of the first sensing unit is made of ZIF-8 doped with SnO2; The gas-sensitive layer of the second sensing unit is made of HKUST-1 doped with ZnO; The gas-sensitive layer of the third sensing unit is made of MOF-74 (Ni) doped with WO3; The gas-sensitive layer of the fourth sensing unit is made of UiO-66 doped with TiO2; The gas-sensitive layer of the fifth sensing unit is made of MIL-101 (Cr) doped with Fe2O3; The gas-sensitive layer of the sixth sensing unit is made of ZIF-67 doped with Co3O4.

4. The gas sensor according to claim 1, wherein The nanostructure includes: a nanoarray composed of a plurality of nanopillars, a nanoarray composed of a plurality of nanowires, or a nanoarray composed of a plurality of nanopores; Wherein, in the nanoarray, a ratio of the height to the width of each nanomonomer exceeds a first threshold.

5. The gas sensor according to claim 1, wherein Also includes: a detection device connected to the interdigital electrodes; The detection device is used to determine the concentration and type of the target gas based on the change of the electrical signal; The detection device comprises: A signal acquisition unit, configured to acquire electrical signals from a plurality of data channels based on a preset sampling frequency to generate a time series data set of the plurality of data channels, wherein different data channels correspond to different sensing units, and different data channels correspond to different time series data sets; a signal preprocessing unit, configured to perform signal denoising and normalization processing on the time series data sets of the multiple data channels to obtain multiple standardized time series data sets; A feature extraction unit, configured to extract a feature sequence of each of the plurality of standardized time series data sets based on a convolutional neural network; a gas identification unit, configured to determine, based on a long short-term memory network, a target gas species probability and a target gas concentration prediction value corresponding to the feature sequence, and determine the gas species with the highest probability as the detected gas species, and use the gas concentration prediction value as the detected gas concentration; The early warning unit is used to issue early warning signals of different levels according to the concentration and type of the target gas.

6. A method for preparing a gas sensor according to any one of claims 1 to 5, characterized in that: include: Making a gold mask on a silicon substrate layer by photolithography, and etching other areas on the silicon substrate layer by metal-assisted etching to obtain nanostructures, wherein the other areas include areas not covered by the gold mask; In situ synthesizing an initial MOFs gas-sensing layer by a layer-by-layer deposition method of MOFs materials and metal oxides on the nanostructure, and performing high-temperature activation on the initial MOFs gas-sensing layer in an inert gas environment to obtain a final MOFs gas-sensing layer; Fabricating interdigital electrodes on the silicon substrate layer by electron beam evaporation technology and photolithography technology; The surface of the sensing unit is covered with a porous structure that allows gas to pass through as a packaging layer by plasma enhanced chemical vapor deposition technology.

7. The method according to claim 6, characterized in that The MOFs material is prepared by a solvent hot melting method.

8. The method according to claim 6, characterized in that The interdigital electrodes are connected to a detection device by extracting electrode signals through a gold wire bonding process.

9. A method for gas detection using the sensor according to any one of claims 1 to 5, characterized in that: include: Collecting electrical signals of multiple data channels based on a preset sampling frequency to generate time series data sets of multiple data channels, wherein different data channels correspond to different sensing units, and different data channels correspond to different time series data sets; performing signal denoising and normalization processing on the time series data sets of the multiple data channels respectively to obtain multiple standardized time series data sets; Extracting characteristic sequences of each of the plurality of standardized time series data sets based on a convolutional neural network, and determining gas type probabilities and gas concentration prediction values corresponding to the characteristic sequences based on a long short-term memory network; The gas type with the highest probability is determined as the detected gas type, and the predicted gas concentration value is used as the detected gas concentration.

10. A method for detecting thermal runaway of a lithium battery, characterized in that: The method comprises: The MOFs-based gas sensor according to any one of claims 1 to 5 is used to monitor the type and concentration of gas released by the lithium battery, and to determine whether thermal runaway occurs in the lithium battery based on the type and concentration of gas.

Images