Wearable / attachable ammonia colorimetric sensor, preparation method and leakage risk monitoring application of wearable / attachable ammonia colorimetric sensor

Abstract

The invention belongs to the technical field of production control, and particularly relates to a wearable / attachable ammonia colorimetric sensor, a preparation method and leakage risk monitoring application thereof. The invention discloses an adhesive and stretchable ammonia colorimetric sensor. The sensor is deposited on a non-woven fabric through electrostatic spinning of polylactic acid (PLA) based on bromocresol green (BCG). The feature color data is extracted by combining a deep learning semantic segmentation algorithm, and the data is converted into a CIE L * a * b * color space, so that colors are analyzed and compared more accurately. The conversion from color data to gas concentration is realized by adopting a deep neural network, and the system has a low-power-consumption, wearable / attachable and visual ammonia gas leakage real-time monitoring function.

Description

Technical Field

[0001] The present invention belongs to the technical field of production control, and particularly relates to a wearable / attachable ammonia colorimetric sensor, a preparation method thereof, and an application for monitoring ammonia leakage risk. Background Art

[0002] With the development of the times, the industrialization process is booming, but at the same time, a large number of industrial toxic gases and chemicals have been produced, which have had a great impact on human health, the ecosystem, and the overall air quality, and more attention needs to be paid to chemical safety. Ammonia is a harmful gas and also an important marker, which is usually widely used and detected in the production of pharmaceuticals, polymers, foods, fertilizers, explosives, cleaning products, and textiles. Research shows that when the concentration of ammonia in the air reaches 15 - 28%, it will become flammable. Ammonia (<1 ppm) can irritate the eyes, throat, and nose. When the concentration of gaseous ammonia reaches 25 ppm, in addition to causing dangerous damage to the respiratory system, it can also cause skin and eye burns, and the human tolerance limit is about 700 ppm. Short-term exposure to an environment with a concentration higher than this may lead to difficulty in breathing and even death. The National Institute for Occupational Safety and Health (NIOSH) in the United States has set the permissible exposure limit of NH3 at 25 ppm. Therefore, there is an urgent need to develop a method with high sensitivity, accuracy, and selectivity that can simply and quickly perform on-site real-time detection of NH3.

[0003] Currently, a variety of technologies have emerged for ammonia detection, including semiconductor sensors, electrochemical sensors, liquid chromatography, gas chromatography, and optical sensors. Although these "traditional" gas sensors have been widely used in the industrial field, they have deficiencies in many aspects, including high energy consumption, high cost, unstable performance, and low detection range and response sensitivity. These limitations result in poor performance in terms of mobility and stability, and in most cases, they are not suitable for on-site real-time monitoring and analysis. In addition, these sensors usually also require trained technical personnel to operate and maintain.

[0004] Therefore, developing an early warning system that can meet the real-time gas monitoring requirements in industrial sites is of great significance for the fields of safe production and production control technology. Summary of the Invention

[0005] In view of this, in order to meet the real-time gas monitoring requirements in industrial sites, the present invention has developed a wearable / attachable ammonia colorimetric sensor based on the combination of deep learning algorithms and optical sensors, as well as a low-power, visual long-term real-time monitoring and early warning system for ammonia leakage and its usage method.

[0006] To achieve the above object, the first object of the present invention is to provide a preparation method of a wearable / attachable ammonia colorimetric sensor, adopting the following technical scheme:

[0007] A preparation method of a wearable / attachable ammonia colorimetric sensor, depositing electrospun poly(lactic acid) of bromocresol green on a non-woven fabric to obtain the wearable / attachable ammonia colorimetric sensor.

[0008] It is worth noting that, based on its inherent advantages, including real-time monitoring ability, high sensitivity and cost-effectiveness, optical sensors show great application prospects in the field of NH3 detection. Colorimetric sensors belong to a type of optical sensors and are considered excellent products for on-site real-time detection of NH3. Because of its low power consumption, long-term stability, high selectivity, low detection limit, fast response, it can provide the most intuitive visual response, greatly reducing manpower consumption and detection time.

[0009] Compared with other colorimetric sensors, bromocresol green used in the present invention has higher sensitivity, and a highly sensitive and high specific surface area NH3 colorimetric sensor can be simply prepared in one step by blending electrospinning solution. Traditional sensor preparation methods usually require complex chemical reactions, post-treatment and multi-step operations, while the introduction of one-step electrospinning technology integrates all processes into one step, greatly simplifying the production process. Compared with traditional sensor production methods, the "one-step electrospinning" disclosed in the present invention can reduce processes, simplify operations, reduce labor and time costs, and at the same time reduce the dependence on expensive raw materials, contributing to large-scale production and commercial application.

[0010] Furthermore, the specific steps of the preparation method of the wearable / attachable ammonia colorimetric sensor include:

[0011] (1) Prepare a poly(lactic acid) (PLA) solution using a mixed solvent A, then weigh an appropriate amount of bromocresol green (BCG) powder and dissolve it in the poly(lactic acid) solution to obtain an electrospinning solution for standby;

[0012] (2) Paste the non-woven fabric on the receiving plate of the electrospinning machine and spin to obtain a wearable / attachable ammonia colorimetric sensor with a nanofiber film.

[0013] Even further, in the step (1), the mixed solvent A is N,N-dimethylformamide (DMF) and tetrahydrofuran (THF) with a volume ratio of 1:1, the concentration of the poly(lactic acid) solution is 6 wt% - 10 wt%, and the concentration of bromocresol green in the electrospinning solution is 1 wt% - 4 wt%.

[0014] Further, in the step (2), the liquid addition rate of electrospinning is 0.06 mm / min, the distance between the syringe and the collection device is set to 20 cm, and nanofibers are collected under a negative voltage of -2 kV and a positive voltage of 14 kV, while maintaining the environmental humidity at 55% and the temperature at 22°C.

[0015] It should be noted that compared with the existing NH3 sensors, the present invention adopts a one-step electrospinning technology, and the preparation process is simple. Based on the adhesive characteristics, this sensor can be applied to various complex environments for monitoring. The thin and flexible structure enables it to be integrated into wearable devices, providing new solutions for personal safety and environmental monitoring. The simplified preparation process reduces the production cost and improves the market competitiveness of the sensor.

[0016] The second object of the present invention is to provide a wearable / attachable ammonia colorimetric sensor obtained by the above preparation method.

[0017] A wearable / attachable ammonia colorimetric sensor, the sensor comprising a colloid, a fabric, and nanofibers, and one side of the fabric is attached with the colloid, and the other side is attached with the nanofibers.

[0018] It should be noted that for the wearable / attachable ammonia colorimetric sensor disclosed in the present invention, according to the usage form, from the inside to the outside, the colloid is the first layer of the sensor, and its main function is to provide good adhesiveness. The viscosity of the colloid enables the sensor to firmly adhere to the fabric or skin, ensuring its long-term stable operation without easy detachment. This is particularly important for wearable or attachable sensors, especially in dynamic movement or usage environments.

[0019] The middle fabric is the basic layer in the sensor structure, endowing the sensor with good flexibility and wearability. The fabric provides a stable basic structure for the entire sensor, enabling the colloid and nanofibers to be effectively maintained in their predetermined positions. Its fiber structure can enhance the durability and strength of the sensor, preventing deformation or damage under external forces.

[0020] The outermost nanofiber layer is the core functional part of the sensor, responsible for the detection of ammonia. Due to its unique microstructure, nanofibers have a very high specific surface area. This means that more ammonia molecules can react with the sensor surface, thereby enhancing the sensitivity of the sensor to ammonia. The high specific surface area is particularly important for the detection of low-concentration ammonia because it can significantly improve the gas adsorption efficiency and thus lower the minimum sensitivity threshold for detection. The diameter of the nanofibers of the sensor disclosed in the present invention ranges from 10 to 500 nm, providing an ideal network structure that is neither too sparse nor too dense. Considering that overly thin fibers may cause the material to be too fragile and affect the durability of the sensor; while overly thick fibers may reduce the specific surface area and the gas molecule contact efficiency, affecting the sensitivity and response speed of the sensor. The nanofiber structure disclosed in the present invention helps the sensor to maintain high sensitivity while maintaining fast response and good stability.

[0021] Furthermore, the nanofibers are bromocresol green / polylactic acid electrospun films, and the nanofiber film has a three-dimensional network structure with a diameter of 10 - 500 nm; the fabric is a non-woven medical breathable tape.

[0022] It should be noted that in the present invention, a bromocresol green (BCG) / polylactic acid (PLA) nanofiber membrane is deposited on a stretchable and adhesive non-woven fabric by electrospinning to form a wearable / attachable ammonia colorimetric sensor. This sensor can visually observe obvious color changes at 1 ppm - 110 ppm. As the ammonia concentration increases, the sensor changes from yellow to blue. Moreover, the sensor exhibits excellent tensile properties and can be stably used for more than 6 weeks in various complex environments.

[0023] The third object of the present invention is to provide an ammonia leakage warning system based on the above-mentioned wearable / attachable ammonia colorimetric sensor.

[0024] An ammonia leakage warning system includes the above-mentioned wearable / attachable ammonia colorimetric sensor, a camera image module, a sensor color information extraction module, a gas concentration calculation module, and a gas concentration change line graph module. Among them, the sensor color information extraction module and the gas concentration calculation module are based on deep learning algorithms and through real-time analysis of images, realizing the monitoring of ammonia concentration.

[0025] It should be noted that the ammonia leakage warning system disclosed in the present invention combines a wearable / attachable ammonia colorimetric sensor, a camera image module, a sensor color information extraction module, a gas concentration calculation module, and a gas concentration change line graph module, and monitors the ammonia concentration through a deep learning algorithm. The wearable / attachable ammonia colorimetric sensor increases the flexible applicability of the system technically by controlling the nanofiber diameter (10 - 500 nm), improving the colloid adhesion strength, and ensuring that the color change range has high sensitivity and response speed. The camera image module adopts high-resolution and high-frame-rate technologies and has an adaptive lighting function to ensure effective capture of color changes in various light environments. The sensor color information extraction module realizes high-precision extraction and fast response of sensor color changes through precise color recognition algorithms and low-latency processing technologies. The gas concentration calculation module performs high-precision fitting analysis through a deep learning algorithm and has a self-calibration mechanism to ensure the accuracy of concentration calculation. The gas concentration change line graph module can be updated in real time and presents dynamic data in a user-friendly interface manner, supporting the backtracking and statistical analysis of historical data. The integrated innovation of this system not only improves the intelligent level and response speed of ammonia monitoring but also enhances the reliability of the system. It is an important technological advancement in improving the safety monitoring efficiency in various environments and provides strong data support for risk assessment and decision-making in practical applications.

[0026] Thus, the ammonia colorimetric sensor is used to capture the gas and convert the gas concentration signal into a color signal. The sensor realizes a wide NH3 detection range of 0 - 110 ppm and can be reused. The camera image module obtains the sensor information under the shooting of an industrial camera and ensures clear images at 4K resolution. The sensor color information extraction module accurately extracts the sensor color part from a complex environment based on the DeepLabv3+ deep learning algorithm, with an accuracy rate > 98.4%, showing a huge improvement compared with other algorithms. The gas concentration calculation module outputs the gas concentration value based on the CIE Lab and RGB color information obtained by the multi-layer perceptron (MLP) deep learning algorithm, realizing the conversion from color information to gas concentration, and the accuracy rate > 99.2%. The gas concentration change line graph module is an intuitive visualization tool. Through a clear line graph form, it helps users quickly understand the rising, falling, or stable trend of gas concentration, facilitates real-time monitoring of the dynamic changes of gas concentration values, and can update data according to the set fixed refresh frequency, thus providing strong support for real-time monitoring and early warning.

[0027] Furthermore, the wearable / attachable ammonia colorimetric sensor can be arranged on the terminal of a human wearable device, the surface of industrial production pipelines and devices, or the surface of other objects exposed to the production environment.

[0028] It should be noted that the ammonia colorimetric sensor disclosed in the present invention has good durability, excellent adhesion and tensile properties, and has broad application prospects in the field of human wearables. It can be easily pasted on parts such as the chest and arm of the human body for detecting ammonia. When the human body moves, such as when the arm bends, the ammonia sensing patch can still firmly adhere to the clothing without affecting its detection function. At the same time, due to the stretchability and adhesiveness of the ammonia colorimetric sensor, it is easy to adhere to the surface of an object. Considering that ammonia leakage in chemical production mostly occurs in key parts such as pipelines, the ammonia colorimetric sensor can be arranged on the surface of industrial production pipelines and devices. Thus, the ammonia colorimetric sensor can not only provide quantitative information such as ammonia concentration, but also identify the location of ammonia leakage, thereby providing strong technical support and theoretical basis for the safety monitoring of chemical plants.

[0029] The fourth object of the present invention is to provide a method for warning and monitoring ammonia leakage using the above system.

[0030] A method for warning and monitoring ammonia leakage, which performs warning and monitoring of ammonia leakage through the warning system as described above. The method includes:

[0031] Obtain the color data of the ammonia colorimetric sensor;

[0032] Adopt an image segmentation algorithm to extract the color information of the sensor from the environmental image. The extracted image information is converted into RGB three-channel values, and then converted into the CIE Lab space to calculate the color difference (ΔE) value;

[0033] According to a preset neural network model, calculate the current concentration of NH3 based on the input RGB values and ΔE values;

[0034] Match the corresponding preset warning strategy according to the NH3 concentration information, and perform warning and monitoring of ammonia leakage at the terminal according to the preset warning strategy.

[0035] It should be noted that in order to realize the real-time monitoring of the NH3 concentration in the environment, first, an image segmentation algorithm is used to extract the color information of the sensor from the environmental image to provide basic data for subsequent color analysis. Subsequently, the extracted image information is converted into RGB three-channel values, and then converted into the CIE Lab space to calculate the color difference (ΔE) value, which is a key index for evaluating the change of NH3 concentration. Based on the RGB values and ΔE values, they are input into a pre-trained neural network model, and this model can calculate the current concentration of NH3.

[0036] Before gas concentration detection, obtaining the sensor color data is a crucial step, and this process involves extracting feature data from images. In this invention, after training the sensor color change process with the semantic segmentation DeeplabV3+ model, information extraction of the sensor can be performed in different NH3 concentration environments. DeeplabV3+ maintains good extraction performance under different lighting conditions, thereby improving the accuracy of concentration detection.

[0037] The ammonia concentration detection model based on the multi-layer feedforward neural network (MLP) has significantly improved the detection accuracy with the introduction of deep learning technology, showing strong generalization ability and stability. The model optimizes the weights and biases in the network through the backpropagation algorithm, relying on the gradient descent method to minimize the mean square error (MSE) loss function on the training set, thereby achieving precise modeling and prediction of ammonia concentration. In the specific design, the model uses RGB color values and ΔE color difference values as the input features of the neural network. After these inputs undergo non-linear transformation in the hidden layer, the mapping relationship between data features and NH3 concentration is captured. During the training stage, the network parameters are continuously adjusted through iteration, enabling the model to effectively fit the complex non-linear relationship between gas concentration and input features. This method not only fully utilizes the advantages of deep learning in feature extraction and pattern recognition but also provides an effective solution for high-precision ammonia concentration detection in practical applications.

[0038] To enhance the generalization ability of the model and effectively reduce the risk of overfitting, the Dropout technique is introduced as a regularization means. In this invention, Dropout uses the probability value p (i.e., the dropout rate) generated by the Bernoulli distribution to randomly remove some neurons in the neural network during the training stage to achieve an effect similar to model ensemble. In addition, the Dropout layer is alternately distributed between the hidden layer and the output layer of the network to further improve the robustness of the model.

[0039] Furthermore, the ammonia leakage warning and monitoring method further includes ensuring the timeliness of concentration data by presetting the sampling frequency.

[0040] It should be noted that DeepLabV3+ is a deep learning model based on convolutional neural network (CNN), which focuses on semantic segmentation of images. It extracts high-resolution spatial features through dilated convolution and encoder-decoder structure, and is particularly suitable for image segmentation tasks in complex environments. MLP (Multi-Layer Perceptron) is a classic neural network architecture, usually used to process tabular data or simple image classification tasks. MLP processes input data through multiple fully connected layers, generates non-linear mappings through activation functions in each layer, and finally outputs prediction results. The preset neural network model can perform inference on new input data (such as new environmental images or sensor data) through the trained parameters (weights and biases) to generate outputs. This process does not require retraining, and only requires input data for the model to output results. Combining image segmentation, color feature extraction and concentration prediction, the preset warning strategy can send warning signals in time when the environment changes. For example, when the ammonia concentration exceeds the preset threshold, the system can automatically activate the alarm mechanism and take emergency measures in advance. Thus, the ammonia leakage warning system and monitoring method disclosed in the present invention can achieve real-time gas concentration monitoring. The sampling frequency can be manually fixed to ensure the timeliness of concentration data. When the set safety threshold is exceeded, the system automatically triggers an alarm and issues a warning through methods such as window reminder and information sending.

[0041] Compared with the prior art, the present invention discloses a highly sensitive stretchable and adhesive ammonia colorimetric sensor and a real-time monitoring ammonia leakage warning system for quantitative ammonia detection and early warning of ammonia leakage. In the presence of NH3, the color of the ammonia colorimetric sensor changes from yellow to blue, which is due to the deprotonation reaction of the BCG dye on the nanofibers. It can also be restored to the initial color through the protonation reaction for repeated use. By extracting the RGB values of the color image and calculating the ΔE response value, it is convenient for quantitative analysis and accuracy improvement. In addition, due to its stretchability and adhesiveness, it can be flexibly placed and is suitable for real-time monitoring of ammonia leakage in various complex environments in chemical plants. The ammonia colorimetric sensor is applicable to the field of production control technology, providing higher sensitivity and unique color changes. In addition, this sensing system also has the potential for air quality monitoring in homes and offices, especially in places where the NH3 concentration is low but may still pose a threat to health. The sensing system realizes real-time data transmission and analysis by combining with intelligent devices, facilitating users to keep track of the dynamic changes of ammonia concentration at any time. The system not only has high-efficiency NH3 detection ability, but also has intelligence, convenience and economy, providing a convenient and efficient solution for on-site applications. Brief Description of the Drawings

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0043] Figure 1 It is a schematic diagram of the preparation process of the colorimetric sensors in Example 1 and Comparative Example 1 of the present invention.

[0044] Figure 2 It is the surface morphology and cross-sectional morphology of the four colorimetric sensors in Experimental Example 1 of the present invention.

[0045] Figure 3 It is a schematic diagram of the ammonia leakage warning system and algorithm process in Example 2 of the present invention.

[0046] Figure 4 It is a schematic diagram of simulating actual application in Example 3 of the present invention.

[0047] Figure 5 It is the physical diagram of the color change and ΔE response values of the four colorimetric sensors in Experimental Example 2 of the present invention.

[0048] Figure 6 It is the change degree, ΔE response curve, and linear fitting curve of ΔE and gas concentration of the ammonia colorimetric sensor in the range of 0 - 140 ppm NH3 in Experimental Example 2 of the present invention.

[0049] Figure 7 It is the experimental results of the anti-interference performance and stability of the ammonia colorimetric sensor in Experimental Example 3 of the present invention.

[0050] Figure 8 It is the tensile test results and wearable application examples of the ammonia colorimetric sensor in Experimental Example 4 of the present invention. Specific Embodiments

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0052] As used herein, the term "embodiment" does not necessarily mean that any embodiment described as "exemplary" is superior or better than other embodiments. For the performance index tests in the embodiments of this application, unless otherwise specified, conventional test methods in the art are used. It should be understood that the terms described in this application are only used to describe specific embodiments and are not intended to limit the content disclosed in this application.

[0053] Unless otherwise specified, the technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; other test methods and technical means not specifically noted in this application refer to the experimental methods and technical means commonly used by those of ordinary skill in the art.

[0054] To better illustrate the content of this application, numerous specific details are given in the following specific embodiments. Those skilled in the art should understand that this application can still be implemented without certain specific details. In the embodiments, some methods, means, instruments, equipment, etc. well-known to those skilled in the art are not described in detail in order to highlight the gist of this application.

[0055] On the premise of no conflict, the technical features disclosed in the embodiments of this application can be combined arbitrarily, and the obtained technical solutions belong to the content disclosed in the embodiments of this application.

[0056] The present invention discloses a wearable / attachable ammonia colorimetric sensor, a preparation method thereof, and its application in leakage risk monitoring, belonging to the technical field of production control. The present invention discloses an adhesive and stretchable ammonia colorimetric sensor, which is deposited on non-woven fabric by electrospinning of poly(lactic acid) (PLA) based on bromocresol green (BCG). By combining a deep learning semantic segmentation algorithm, the extraction of characteristic color data is realized, and the data is converted into the CIE L*a*b* color space for more accurate color analysis and comparison. A deep neural network is used to realize the conversion from color data to gas concentration, with the functions of low power consumption, wearable / attachable, and real-time visual monitoring of ammonia leakage.

[0057] To better understand the present invention, the following embodiments are used to further specifically elaborate on the present invention, but it should not be construed as a limitation of the present invention. For some non-essential improvements and adjustments made by those skilled in the art based on the above-mentioned invention content, they are also considered to fall within the protection scope of the present invention.

[0058] The experimental materials involved in the present invention:

[0059] Bromocresol Green (BCG, indicator grade), Polylactic acid (PLA, Mw ~ 60,000), N,N-Dimethylformamide (DMF, AR grade), Tetrahydrofuran (THF, ACS grade), Absolute ethanol (C2H6O9, ≥99.5%), all of which were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Non-woven fabric, filter paper, and melt-blown fabric were purchased from Zhejiang Jinsanfa Group Co., Ltd.

[0060] Example 1

[0061] First, add DMF and 4.8 mL of THF (DMF:THF = 1:1) into a 50 mL beaker. Weigh 8 wt% of PLA powder and slowly pour it into the beaker. Stir at room temperature for 5 hours until the PLA powder is completely dissolved. Weigh 2 wt% of BCG powder and slowly pour it into the beaker. Stir at room temperature for 1 hour until the BCG powder is completely dissolved. Let it stand at room temperature for 30 minutes to prepare the electrospinning solution. Aspirate the resulting electrospinning solution into a disposable syringe and perform electrospinning using a 20-gauge metal needle. The electrospinning equipment consists of an injection system, a collection device, and a high-voltage power supply. Set the liquid addition rate to 0.06 mm / min, the distance between the syringe and the collection device to 20 cm, and collect the nanofibers under a negative voltage of -2 kV and a positive voltage of 14 kV, maintaining an environmental humidity of 55% and a temperature of 22°C. Paste the non-woven fabric on the receiving plate of the electrospinning machine. After spinning for 5 h, a colorimetric sensor with a 0.6 mm nanofiber film is obtained, as Figure 1 (a) shows.

[0062] Example 2

[0063] An ammonia leakage warning system includes an ammonia colorimetric sensor, a camera image module, a sensor color information extraction module, a gas concentration calculation module, and a gas concentration change line graph module. In this example, a Logitech c1000s camera was used. Combining with a color analysis algorithm, the RGB characteristic information of the sensor sensing area (10 mm × 10 mm) was extracted and analyzed in real time to detect the current NH3 gas concentration and display it on the intelligent terminal. The distance between the sensor and the camera was maintained at 10 cm to ensure the clarity of the image and the accuracy of the color, as Figure 3 (a) shows.

[0064] During use, first, obtain the color data of the ammonia colorimetric sensor; then, use an image segmentation algorithm to extract the color information of the sensor from the environmental image. The extracted image information is converted into RGB three-channel values and then converted to the CIE Lab color space to calculate the color difference ΔE value; next, according to a preset neural network model, calculate the current NH3 concentration based on the input RGB values and ΔE value; finally, match the corresponding preset warning strategy according to the NH3 concentration information, and perform ammonia leakage monitoring and warning on the terminal according to the preset warning strategy.

[0065] As the NH3 concentration increases, the color of the sensor surface changes from yellow to blue. To simplify the image analysis process, the CIE DE2000 color difference calculation formula based on CIE L*a*b* is applied in the present invention. The RGB color space is converted to the XYZ color space and calculated according to Equation (1), and then the XYZ color space is converted to the L*a*b color space to obtain the L*a*b value expressions of all image colors, which are calculated according to Equations (2) and (3). Finally, call the function of the CIE DE2000 color difference calculation formula to obtain the color difference (ΔE). ΔE is defined as the color response of the colorimetric sensor, and the change of the ΔE value is directly related to the change of the NH3 concentration, thus providing a quantitative method for the detection of NH3.

[0066]

[0067] To achieve real-time monitoring of the gas concentration change in the gas chamber, a colorimetric sensing system monitoring interface is developed based on the Qt platform. This interface can continuously call algorithm analysis and display the change of gas concentration in real time in the form of a line chart. Through this real-time monitoring, the change of ammonia concentration can be effectively detected, and data support can be provided for subsequent optimization and application.

[0068] During use, first use an image segmentation algorithm to extract the color information of the sensor from the environmental image to provide basic data for subsequent color analysis. Subsequently, the extracted image information is converted into RGB three-channel values and then converted to the CIE Lab color space to calculate the color difference (ΔE) value, which is a key index for evaluating the change of NH3 concentration. Based on the RGB values and the ΔE value, input them into a pre-trained neural network model, which can calculate the current NH3 concentration. As shown in Figure 3 (b), the picture algorithm flow chart realizes the detection of gas concentration.

[0069] The neural network model is trained on data and generates a model through the DeepLabv3+ semantic segmentation algorithm and the multi-layer perceptron (MLP) architecture. DeepLabv3+ is a semantic segmentation algorithm based on convolutional neural networks (CNNs), and its core innovation lies in how to improve the accuracy and efficiency of image segmentation. This algorithm effectively processes high-resolution information in images and expands its receptive field by introducing dilated convolution and an encoder-decoder structure, so as to capture more extensive context information. DeepLabv3+ uses the Xception convolutional neural network structure (depthwise separable convolution) in the encoder part to reduce the amount of computation and the number of parameters while maintaining efficient feature extraction capabilities. The decoder part is used to restore the high-dimensional semantic features extracted by the encoder to a higher resolution, thereby enhancing the details and accuracy of image segmentation. Through this encoder-decoder structure, DeepLabv3+ can effectively perform refined segmentation on images in complex scenes, that is, accurately extract optical sensors from complex backgrounds. The multi-layer perceptron (MLP) is a classic feedforward neural network structure, consisting of an input layer, multiple hidden layers, and an output layer. Each layer of the MLP transmits information through connections between neurons, and the neurons in each layer perform non-linear transformations through activation functions. The MLP shows strong capabilities in processing high-dimensional data and non-linear relationships, and is especially suitable for classification tasks and modeling of complex features. Therefore, it is used to construct a gas concentration output model, and the current gas concentration value is predicted by comparison in the neural network through color information.

[0070] The preset warning strategy is to use the DeepLabv3+ model in combination with optical sensors for analysis in ambient gas concentration detection. First, the optical sensor collects ambient images, capturing the interaction between the gas and the optical signal. Subsequently, the DeepLabv3+ model performs semantic segmentation on the image, accurately separating the sensor area from the complex background. Then, color values are extracted from the segmented gas area and converted from the RGB color space to the CIE Lab color space, which is closer to human visual perception and can more accurately describe color changes. Finally, the converted CIE Lab color values and RGB values are input into the multi-layer perceptron (MLP) model for analysis and prediction. The MLP can quickly and accurately output the gas concentration in the current environment by learning the complex non-linear relationship between the color values and the gas concentration. This method combines the image segmentation ability of deep learning and the prediction ability of machine learning, and has advantages such as non-contact detection, high flexibility, and strong adaptability. It is especially suitable for multi-gas detection in complex environments and shows broad application prospects.

[0071] Example 3

[0072] Practical applications of ammonia leakage warning system and its detection method:

[0073] As Figure 4 shown in (a), when gas leakage occurs in an industrial site, it causes the color change of the ammonia sensing patch. The camera transmits its color information in real time, and the concentration is detected and displayed through algorithm processing on the PC side. At the same time, the concentration data is uploaded through cloud services to achieve the remote monitoring function via the mobile phone App. The monitoring data during the experiment will be displayed in the software in real time and transmitted to the App. When the concentration exceeds the threshold, a red warning sign will be immediately displayed.

[0074] A section of PVC pipe through which NH3 is being introduced was selected as the experimental object, and an elliptical hole was precisely made on it to simulate the possible pipeline breakage in actual production. As Figure 4 shown in (b), in the experiment, an ammonia sensing patch was placed near the hole. Due to its stretchability and adhesiveness, it was easily attached to the pipeline surface. When NH3 flowed through the pipeline, ammonia leaked from the hole, and the ammonia sensing patch quickly turned blue. The leakage orientation could be clearly seen during the color change process. By identifying the current color through the colorimetric system, the ammonia concentration could be detected. The experimental results show that the ammonia sensing patch can not only provide quantitative information such as ammonia concentration, but also identify the location of ammonia leakage, thus providing strong technical support and theoretical basis for the safety monitoring of chemical plants.

[0075] To further prove the beneficial effects of the present invention for better understanding of the present invention, the following comparative examples and experimental examples further clarify the properties and application performance of the ammonia colorimetric sensor disclosed by the present invention, its preparation method, the ammonia leakage warning system using it and the monitoring method, but it should not be construed as a limitation of the present invention. For other determination methods and properties obtained by those skilled in the art through other measurement experiments based on the above-mentioned invention content and the applications based on the above-mentioned properties, they are also considered to fall within the protection scope of the present invention.

[0076] Comparative Example 1

[0077] First, 20 mL of absolute ethanol was added to a 50 mL beaker, 40 mg of BCG powder was weighed and slowly added to the beaker, and stirred at room temperature for 15 minutes until completely dissolved to form a BCG solution. Then, a 10 mm x 10 mm piece of meltblown cloth, non-woven fabric, and filter paper were each sheared, and 5 μL of the BCG solution was respectively pipetted and dropped onto the surfaces of the meltblown cloth, non-woven fabric, and filter paper. Finally, the three moist sensors were taken out after drying at 60 °C for 3 minutes to form a colorimetric sensor, as Figure 1 shown in (b).

[0078] Experimental Example 1

[0079] The surface morphology and structural cross-sectional morphology of the nanofiber film were characterized by scanning electron microscopy (SEM, Hitachi S-4800, Japan) for the four colorimetric sensors prepared in Example 1 and Comparative Example 1.

[0080] Figure 2 (a) shows the surface, cross-section and sample images of BCG / PLA electrospun films. The images show that the addition of BCG did not significantly affect the uniformity and diameter of the fibers, but caused more rough structures to appear on the fiber surface, which may be related to the distribution of BCG on the fiber surface. In addition, no obvious agglomeration or agglomeration was observed on the fiber surface, indicating that BCG had good dispersion in PLA. From the SEM image of the bonding cross-section between the BCG / PLA electrospun film and the non-woven fabric, it can be observed that a good bonding interface was formed between the fibers without obvious gaps or separation. The smooth and uniform surface of the fiber shows the affinity between PLA and the non-woven fabric, which may be due to the addition of bromocresol green to enhance the compatibility of the composite film. In particular, the structure of the contact area between the fiber and the non-woven fabric shows continuity and consistency, indicating good mechanical adhesion between them. This tightly fitted structure enhances the stability and durability between the film and the non-woven fabric, improving the overall performance of the film in practical applications. Figure 2 (b) shows the SEM characterization image and sample image of the meltblown cloth. On the surface of the meltblown cloth substrate, BCG is evenly distributed on the fiber surface, forming small particles or a film layer covering the fiber surface without obvious aggregation, showing good coating uniformity. Figure 2 (c) shows the SEM characterization image of the nonwoven substrate and the sample image. On the nonwoven substrate, BCG is more dispersed and sometimes local aggregation occurs, especially in the cross-sectional area where the fibers are interwoven. This may be due to the presence of more micropores and irregular structures on the surface of the nonwoven fabric, which makes BCG more likely to aggregate and form a thicker coating in these places. Figure 2 (d) shows the SEM characterization image of the filter paper substrate and the sample image. On the filter paper substrate, most of the BCG is attached to the fiber intersection or surface, forming a rough coating, and the particles in some areas may be larger, forming a small block structure, and the surface is not as smooth as the meltblown fabric substrate.

[0081] Experimental Example 2

[0082] The ammonia response characteristics of four different substrates of NH3 sensors are compared: electrospun film ammonia sensor, meltblown fabric, non-woven fabric and filter paper substrate ammonia sensor. These sensors were tested under controlled environmental conditions, i.e., relative humidity was 65±2% and temperature was maintained at 26°C. During the test, NH3 concentrations ranging from 0ppm to 100ppm were injected into the gas chamber to simulate the response characteristics under different NH3 concentrations.

[0083] The results show that the electrospun film ammonia sensor exhibits a more uniform color change when in contact with ammonia ( Figure 5 a), which provides a more accurate basis for subsequent color data extraction and concentration detection. Through the quantitative analysis of ΔE, the electrospun film shows excellent linearity and a wide detection range in the NH3 concentration range of 0 - 100 ppm, as shown in Figure 5 (b). Therefore, the electrospun film ammonia sensor has better color change uniformity, linearity, and detection range, has high sensitivity and good selectivity in NH3 concentration detection, and is of great significance for NH3 detection in the fields of environmental monitoring and industrial safety.

[0084] Furthermore, under the same conditions, the color response of the ammonia colorimetric sensor of the present invention was tested in the NH3 concentration range of 0 - 140 ppm. The ammonia sensing patch showed an obvious color change from yellow to blue ( Figure 6 b). The color change pictures of the ammonia sensing patch at 0 - 140 ppm concentration were converted into ΔE values ( Figure 6 c). When ΔE < 1.0, the color difference is hardly visible; when ΔE is 1 - 2, the color difference can be seen by careful observation; when ΔE is 2 - 10, the color difference can be seen at a glance; when ΔE > 10, the colors are different. It can be seen from the capture by the camera of this system that the ammonia sensing patch shows a color change at 1 ppm concentration and the ΔE value is greater than 1.5. After testing, the maximum response limit of this sensing patch is about 110 ppm, and the color does not change when NH3 is continuously added. And the ΔE response values at 0 - 110 ppm concentration were linearly fitted ( Figure 6 d), and it is concluded that this ammonia sensing patch has good linearity to better identify the current NH3 concentration value.

[0085] Experimental Example 3

[0086] Test the anti-interference performance and stability of the ammonia colorimetric sensor disclosed in the present invention.

[0087] Figure 7 (a) When 10 ppm NH3 and 10 ppm of NO2, O2, SO2, CH4, H2, C4H 10 , CO, and H2S exist in the gas chamber simultaneously, it is observed that ΔE maintains the response value at a normal concentration of only 10 ppm NH3, proving its good selectivity. Figure 7 (b) When 10 ppm NO2, O2, SO2, CH4, H2, C4H 10 , CO, and H2S are injected into the gas chamber in sequence, no obvious color change occurs. When 10 ppm NH3 is injected, the color immediately changes, further verifying its selectivity.

[0088] Long-term use of the sensor for real-time monitoring in industrial leaks is very important. Therefore, the stability of the ammonia sensing patch was tested over 6 weeks ( Figure 7 c). Each time 10 ppm NH3 was injected, and there was no significant difference in the ΔE response value, thus proving its long-term stability. The ammonia sensing patch can restore its initial color based on the protonation reaction. Therefore, the recovery speed at 60 °C and 30 °C was tested. As Figure 7 (d) shows, the ammonia sensing patch has a faster recovery speed at 60 °C and can recover within 20 minutes, while the natural recovery is slower at 30 °C. Based on the recovery characteristics of the ammonia sensing patch, the color change of the sensor at 10 ppm was tested three times in a row ( Figure 7 e). ΔE remained consistent, thus verifying its multiple reuse. This ammonia sensing patch has an extremely fast reaction to NH3. The present invention tested the dynamic change curves of the ammonia sensing patch at NH3 concentrations of 20, 40, 60, 80, and 100 ppm respectively ( Figure 7 f). It can be observed that it basically tends to a stable state at 15 s at different concentrations, providing faster and more obvious detection for ammonia leakage in chemical plants. The present invention tested the effect of different normal temperatures on the color change of the ammonia sensing patch. In an environment of 20 °C - 35 °C, 10 ppm ammonia was injected into the gas chamber for a color change reaction, and ΔE still remained at the same level. As Figure 7 (g) shows, it can be concluded that different normal temperatures do not affect the color change degree of this sensor at the same concentration of NH3. The present invention explored the effect of humidity on the color change of the ammonia sensing patch. The experimental results show that in a high-humidity environment, the same concentration of NH3 will cause a greater ΔE response change ( Figure 7 h) and a more significant color change ( Figure 7 i). This phenomenon can be attributed to the deprotonation process in the reaction of BCG with ammonia, where water molecules act as proton transfer media, promoting the transfer of protons. Specifically, when water molecules are absorbed into the pad, they promote the reaction between NH3 and H2O to generate NH 4+ and OH - , thus forming an alkaline environment on the film surface. Therefore, under high RH conditions, the generation amounts of NH 4+ and OH - on the film surface increase, resulting in a more obvious color change.

[0089] Experimental Example 4

[0090] To evaluate the tensile properties of the ammonia colorimetric sensing patch, tensile tests were carried out. These tests were performed on a tensile testing machine (TM2101-T5) at a test speed of 10 mm / min, a temperature of 22 °C, and the test object was a 50 mm × 15 mm ammonia colorimetric sensor.

[0091] Figure 8 As shown in (a), the strong stretchability of the ammonia sensing patch was tested. The BCG / PLA electrospun film was evenly distributed on the surface, and no obvious breakage or tearing of the film occurred during the stretching process. Moreover, the sensor was tested for 50 consecutive cycles of stretching by 17 mm. As Figure 8 shown in (b), fine wrinkles appeared on the surface of the sensor, and the length was stretched, but there was still no breakage or tearing on the surface, proving its durability. After testing, the maximum tensile force of the ammonia sensing patch at break can reach 43.12 kPa ( Figure 8 c). And it can be seen from the loading / unloading curves of 50 consecutive cycles under tensile strain that the tensile stress and dissipated energy decrease successively ( Figure 8 d), and the curves of the loading / unloading process basically coincide with the increase in the number of cycles. In addition, an obvious hysteresis loop was observed in the first cycle, which is mainly because the fibers of the non-woven fabric are arranged disorderly, and during the stretching process, the mutual slip and rearrangement between the fibers will occur. This slip and rearrangement will consume part of the externally applied energy and cannot be fully recovered during unloading, resulting in the formation of the hysteresis loop. In the subsequent loading / unloading curves, the hysteresis loop is limited, indicating the excellent elasticity of the ammonia sensing patch.

[0092] Due to its excellent adhesion and stretching properties, the ammonia sensing patch has broad application prospects in the field of human wearables. As Figure 8 shown in (e), the patch can be easily pasted on parts of the human body such as the chest and arm for detecting ammonia. When the human body moves, such as when the arm bends, the ammonia sensing patch can still firmly adhere to the clothing without affecting its detection function. This is due to its excellent mechanical properties and adhesion characteristics, which can stretch freely within the range of natural human movement and will not be damaged or lose its function due to stretching. It has great application potential in the field of human wearables and can provide effective technical support for health monitoring and environmental safety.

[0093] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for preparing a wearable / adhesive ammonia colorimetric sensor, characterized in that: The wearable / attachable ammonia colorimetric sensor is obtained by electrospinning polylactic acid with bromocresol green onto a non-woven fabric.

2. The preparation method according to claim 1, characterized in that: The specific steps include: (1) preparing a polylactic acid solution using a mixed solvent A, then weighing an appropriate amount of bromocresol green powder and dissolving it in the polylactic acid solution to obtain an electrospinning solution for later use; (2) The non-woven fabric is pasted on a receiving plate of an electrospinning machine to spin a wearable / adhesive ammonia colorimetric sensor made of a nanofiber film.

3. The preparation method according to claim 2, characterized in that: In the step (1), the mixed solvent A is N,N-dimethylformamide and tetrahydrofuran in a volume ratio of 1:1, the concentration of the polylactic acid solution is 6wt%-10wt%, and the concentration of bromocresol green in the electrospinning solution is 1wt%-4wt%.

4. The preparation method according to claim 2, characterized in that: In the step (2), the liquid addition rate of electrospinning is 0.06 mm / min, the distance between the syringe and the collecting device is set to 20 cm, and the nanofibers are collected at a negative voltage of -2 kV and a positive voltage of 14 kV, and the ambient humidity is maintained at 55% and the temperature is 22°C.

5. The wearable / adhesive ammonia colorimetric sensor prepared by the method according to any one of claims 1 to 4, characterized in that: The sensor comprises colloid, fabric and nanofiber, wherein the colloid is attached to one side of the fabric and the nanofiber is attached to the other side of the fabric.

6. The wearable / adhesive ammonia colorimetric sensor according to claim 5, characterized in that: The nanofiber is a bromocresol green / polylactic acid electrostatic spinning film, and the nanofiber film presents a three-dimensional network structure with a diameter of 10-500nm; the fabric is a non-woven medical breathable tape.

7. An ammonia leak warning system, characterized in that: It includes a wearable / attachable ammonia colorimetric sensor as described in any one of claims 5-6, a camera image module, a sensor color information extraction module, a gas concentration calculation module and a gas concentration change line graph module, wherein the sensor color information extraction module and the gas concentration calculation module realize the monitoring of ammonia concentration through real-time analysis of images based on a deep learning algorithm.

8. The ammonia leakage warning system according to claim 7, characterized in that: The wearable / attachable ammonia colorimetric sensor can be deployed on a wearable device terminal of a human body, an industrial production pipeline and device surface, or other object surfaces exposed to a production environment.

9. An ammonia leakage early warning monitoring method, characterized in that: Ammonia gas leakage early warning monitoring is performed by using the early warning system according to any one of claims 7 to 8, the method comprising: Get the color data of the ammonia colorimetric sensor; The image segmentation algorithm is used to extract the color information of the sensor from the environmental image. The extracted image information is converted into RGB three-channel values, and then converted into CIE Lab space to calculate the color difference ΔE value; According to the preset neural network model, the current NH3 concentration is calculated based on the input RGB value and ΔE value; The corresponding preset warning strategy is matched according to the NH3 concentration information, and ammonia leakage monitoring and warning are performed at the terminal according to the preset warning strategy.

10. The ammonia leakage early warning monitoring method according to claim 9, characterized in that: It also includes pre-setting the sampling frequency to ensure the timeliness of concentration data.

Images