Multi-component identification method based on catalytic luminescence and artificial neural network

Abstract

The invention relates to the technical field of analytical chemistry. The invention provides a multi-component identification method based on catalytic luminescence and an artificial neural network, and the method comprises the steps: carrying out the catalytic luminescence reaction through a single sensing unit and a plurality of training samples, collecting a corresponding catalytic luminescence response signal, and obtaining an initial data set; preprocessing the initial data set to obtain a preprocessed data set; constructing a component identification initial model based on a neural network, and training the component identification initial model through the preprocessing data set to obtain a component identification model; and collecting a catalytic luminescence response signal of a to-be-detected sample, performing pretreatment, and inputting the pretreated catalytic luminescence response signal of the to-be-detected sample into the component identification model to obtain a component identification result of the to-be-detected sample. The problems that most of existing catalytic luminescence technologies are limited to single-component detection, a multi-component system needs to be detected by a plurality of sensing units respectively, the efficiency is low, and simultaneous recognition of multiple components is difficult to achieve are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of analytical chemistry, and in particular to a multi-component identification method based on catalytic luminescence and artificial neural network. Background Art

[0002] In the field of chemical analysis, multi-component identification is a crucial technology, which is widely used in environmental monitoring, food safety, drug analysis, etc. Traditional multi-component identification methods usually rely on complex separation techniques and a combination of multiple sensors, which not only increases the complexity of experimental operations, but also increases costs, limiting its popularity in practical applications.

[0003] In recent years, catalytic luminescence (CTL) technology has gradually emerged in the field of analytical chemistry due to its advantages such as high sensitivity, high selectivity and simple operation. Catalytic luminescence is a process in which the substance to be tested reacts chemically with the luminescent reagent through the catalytic action of the catalyst, thereby stimulating a light signal. There is a certain relationship between the intensity of this light signal and the concentration of the substance to be tested, so it can be used for quantitative analysis. However, traditional catalytic luminescence technology is mostly limited to the detection of a single component. For multi-component systems, multiple sensing units are often required for separate detection, which is not only inefficient, but also difficult to achieve simultaneous identification of multiple components. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-component identification method based on catalytic luminescence and artificial neural network, aiming to solve the problem that the existing catalytic luminescence technology is mostly limited to single component detection, and multiple sensing units are required to detect the multi-component system separately, resulting in low efficiency and difficulty in achieving simultaneous identification of multiple components.

[0005] The present invention is achieved through the following technical solutions:

[0006] A multi-component identification method based on catalytic luminescence and artificial neural network comprises the following steps:

[0007] A single sensor unit is used to perform a catalytic luminescence reaction with several training samples, and a catalytic luminescence response signal corresponding to each training sample is collected to obtain an initial data set;

[0008] Preprocessing the initial data set to obtain a preprocessed data set;

[0009] An initial component recognition model is constructed based on a neural network, and the initial component recognition model is trained by preprocessing a data set to obtain a component recognition model;

[0010] The catalytic luminescence response signal of the sample to be detected is collected and preprocessed, and the preprocessed catalytic luminescence response signal of the sample to be detected is input into the component recognition model to obtain the component recognition result of the sample to be detected.

[0011] Optionally, the preparation process of the single sensing unit is:

[0012] La(NO3)3·6H2O, urea, Eu(NO3)3·6H2O and Tb(NO3)3·6H2O were added into a certain amount of ultrapure water in proportion, and stirred until completely dissolved to form a mixed solution;

[0013] The mixed solution is transferred to a high-pressure reactor to perform a hydrothermal reaction to generate a precursor precipitate;

[0014] After the hydrothermal reaction is completed, the high-pressure reactor is cooled, the generated precursor precipitate is collected, and the precursor precipitate is washed with ultrapure water and anhydrous ethanol respectively to remove residual impurities;

[0015] The washed precursor precipitate is vacuum dried at a preset temperature to obtain a precursor solid;

[0016] The precursor solid is placed in a tubular furnace, the heating rate is preset, and high-temperature calcination is performed to finally obtain Eu / Tb co-doped La2O2CO3 sensing material.

[0017] Optionally, the single sensing unit adopts Eu / Tb co-doped La2O2CO3 material; wherein the atomic ratio of the dosage of Eu and Tb to La is 0.5%-5% respectively.

[0018] Optionally, the specific process of performing catalytic luminescence reaction on a single sensing unit and several training samples and collecting catalytic luminescence response signals corresponding to each training sample to obtain the initial data set is as follows:

[0019] Uniformly coating a single sensing unit on a ceramic rod in a reaction chamber;

[0020] Each training sample is gasified and flows through the reaction chamber respectively under the driving of carrier gas, so that the training sample and the sensing material undergo catalytic luminescence reaction;

[0021] The catalytic luminescence response signal corresponding to each training sample is collected by a chemiluminescence instrument to form an initial data set containing the catalytic luminescence kinetic curves of each sample.

[0022] Optionally, the specific process of preprocessing the initial data set to obtain the preprocessed data set is:

[0023] The catalytic luminescence kinetic curves of each training sample in the initial data set are averaged and sampled to unify the data length of each curve;

[0024] The initial data set after average sampling is filtered, and the principal component analysis method is used to reduce the dimension of the filtered data, extract the characteristic parameters that have a key contribution to component identification, and obtain the characteristic vector corresponding to each training sample;

[0025] The feature vectors corresponding to each training sample are combined into a feature vector set as a preprocessing data set.

[0026] Optionally, the specific process of constructing the initial component identification model based on the neural network and training the initial component identification model by preprocessing the data set to obtain the component identification model is:

[0027] A back propagation neural network is selected as the model architecture to construct an initial component identification model; wherein the initial component identification model includes an input layer, a hidden layer, and an output layer;

[0028] The hyperparameters of the initial component identification model are set, and the initial component identification model is trained according to the preprocessed data set. The weights and biases in the network are iteratively adjusted through the back propagation algorithm with the goal of minimizing the prediction error until the preset number of training rounds is reached or the preset stopping condition is met, and the component identification model is obtained.

[0029] Optionally, during the training process, a leave-one-out validation method is used to evaluate the performance of the model, and the hyperparameters of the component identification model are adjusted according to the evaluation results to obtain the optimal component identification model.

[0030] Optionally, during the training process, the leave-one-out validation method is used to evaluate the performance of the model, and the hyperparameters of the component identification model are adjusted according to the evaluation results. The specific process of obtaining the optimal component identification model is as follows:

[0031] The leave-one-out validation method was used to divide the preprocessed data set into a training set and a validation set. The component recognition model was trained and validated several times, and the validation results were recorded each time.

[0032] According to the verification results, the average classification accuracy, average quantitative accuracy and average percentage error of non-zero true concentration of the model were calculated to evaluate the performance of the model and obtain the performance evaluation results;

[0033] According to the performance evaluation results, the structure, parameters and hyperparameters of the component identification model are adjusted to obtain the optimal component identification model;

[0034] The optimal component recognition model is applied to multi-component recognition of actual samples. By collecting the catalytic luminescence response signals of the samples to be tested, preprocessing them and inputting them into the model, the component recognition results of the samples to be tested are obtained, and the accuracy and reliability of the model are verified.

[0035] Optionally, during the training process of the initial component identification model, the tanh function is used as the activation function, and the model structure is optimized by adjusting the number of hidden layers and the number of neurons.

[0036] Optionally, the training sample is a mixture containing persistent organic pollutants.

[0037] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0038] Improved efficiency of multi-component identification: By using a single sensor unit combined with an advanced neural network model, simultaneous identification of multiple components is achieved, avoiding the tedious steps of separate detection by multiple sensor units in traditional methods, and significantly improving detection efficiency.

[0039] Reduced cost: Only one sensor unit is needed to complete the multi-component identification task. Compared with the configuration of multiple sensor units used in the traditional method, the cost of the present invention is effectively controlled, which is conducive to its wide application in environmental monitoring, food safety, drug analysis and other fields.

[0040] Enhanced recognition accuracy and reliability: By collecting the catalytic luminescence response signals of training samples and constructing a neural network model for training, the unique characteristics of different components in the catalytic luminescence reaction can be learned, thereby achieving accurate recognition of multiple components. This data-driven recognition method is more accurate and reliable than traditional methods.

[0041] Simplified experimental operation: No complicated separation technology and combination of multiple sensors are required. Multi-component identification can be completed through a single sensing unit and neural network model, which greatly simplifies the experimental operation process and reduces the difficulty of the experiment.

[0042] It has broad application prospects: Since the catalytic luminescence technology itself has the advantages of high sensitivity, high selectivity and simple operation, combined with the single sensing unit multi-component recognition method of the present invention, it will show broader application prospects in multiple fields such as environmental monitoring, food safety, and drug analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the process of a multi-component identification method based on catalytic luminescence and artificial neural network according to an embodiment of the present invention;

[0044] Figure 2Schematic diagram of the relationship between the rare earth ion doping ratio and the chemiluminescence intensity in three organophosphate flame retardants in the embodiment of the present invention;

[0045] Figure 3 Schematic diagram of the relationship between the catalytic luminescence response signal intensity and signal-to-noise ratio of TCP and the carrier gas flow rate in an embodiment of the present invention;

[0046] Figure 4 Schematic diagram of the relationship between the signal intensity and signal-to-noise ratio of TCP catalytic luminescence response and temperature in an embodiment of the present invention;

[0047] Figure 5 Schematic diagram of the relationship between time and signal intensity curves of different OPFRs in an embodiment of the present invention;

[0048] Figure 6 A schematic diagram of a classification accuracy curve of an initial component identification model inputting an unfiltered curve in an embodiment of the present invention;

[0049] Figure 7 A schematic diagram of a quantitative accuracy curve of an initial component identification model inputting an unfiltered curve in an embodiment of the present invention;

[0050] Figure 8 A curve diagram of the average percentage error MAPE of the component identification initial model inputting the unfiltered curve in an embodiment of the present invention;

[0051] Fig. 9 It is a schematic diagram of the structure of the combination of principal component analysis (PCA) and back propagation neural network (BP) in an embodiment of the present invention;

[0052] Figure 10-Figure 21 Schematic diagram of the relationship between the catalytic luminescence response signal intensity and concentration of the six OPFRs in the embodiment of the present invention;

[0053] Fig. 22 Schematic diagram of the curve relationship between the catalytic luminescence response signal of TBOEP and time in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following is a specific implementation method in conjunction with the drawings.

[0055] Reference Figure 1 , a multi-component identification method based on catalytic luminescence and artificial neural network, comprising the steps of:

[0056] Step 1: A catalytic luminescence reaction is performed between a single sensor unit and several training samples, and a catalytic luminescence response signal corresponding to each training sample is collected to obtain an initial data set.

[0057] In some embodiments, the preparation process of a single sensor unit is:

[0058] La(NO3)3·6H2O, urea, Eu(NO3)3·6H2O and Tb(NO3)3·6H2O were added into a certain amount of ultrapure water in proportion, and stirred until completely dissolved to form a mixed solution;

[0059] The mixed solution is transferred to a high-pressure reactor to perform a hydrothermal reaction to generate a precursor precipitate;

[0060] After the hydrothermal reaction is completed, the high-pressure reactor is cooled, the generated precursor precipitate is collected, and the precursor precipitate is washed with ultrapure water and anhydrous ethanol respectively to remove residual impurities;

[0061] The washed precursor precipitate is vacuum dried at a preset temperature to obtain a precursor solid;

[0062] The precursor solid is placed in a tubular furnace, the heating rate is preset, and high-temperature calcination is performed to finally obtain Eu / Tb co-doped La2O2CO3 sensing material.

[0063] In some embodiments, a single sensing unit uses Eu / Tb co-doped La2O2CO3 material, wherein the atomic ratio of the dosage of Eu and Tb to La is 0.5%-5%, respectively.

[0064] In some embodiments, a Eu / Tb co-doped La2O2CO3 sensing material is prepared by a hydrothermal synthesis method. 2.6g La(NO3)3·6H2O, 1.2g urea, Eu(NO3)3·6H2O and Tb(NO3)3·6H2O (the atomic ratio of Eu and Tb to La is 0.5%, 1%, 2%, 3%, 5%, respectively, and the products are respectively expressed as x% Eu / Tb co-doped La2O2CO3) are added to 30ml ultrapure water and stirred and dispersed at room temperature until completely dissolved. The dispersion is then transferred to a high-pressure reactor and subjected to a hydrothermal reaction in an oven at 170°C for 6h. After the reaction, the dispersion is cooled to room temperature and the resulting white precursor is collected by centrifugation, washed 3 times with ultrapure water and anhydrous ethanol, respectively, and vacuum dried at 60°C for 24h. Finally, the dried solid was calcined at 550°C in a tube furnace for 3h (atmosphere condition: air, heating rate: 10°C / min), and the obtained white solid was the Eu / Tb co-doped La2O2CO3 sensing material. In addition, the sensing material without Eu / Tb doping does not need to add Eu(NO3)3·6H2O and Tb(NO3)3·6H2O, and the rest of the steps are the same as the above steps.

[0065] In some embodiments, a specific process of obtaining an initial data set by performing a catalytic luminescence reaction on a single sensor unit and a plurality of training samples and collecting a catalytic luminescence response signal corresponding to each training sample is as follows:

[0066] Uniformly coating a single sensing unit on a ceramic rod in a reaction chamber;

[0067] Each training sample is gasified and flows through the reaction chamber respectively under the driving of carrier gas, so that the training sample and the sensing material undergo catalytic luminescence reaction;

[0068] The catalytic luminescence response signal corresponding to each training sample is collected by a chemiluminescence instrument to form an initial data set containing the catalytic luminescence kinetic curves of each sample.

[0069] In some embodiments, the training sample can be a mixture containing persistent organic pollutants, and the persistent organic pollutants can be organophosphate flame retardants (OPFRs). The CTL experiment of OPFRs is carried out on a self-built CTL device, which includes a BPCL-2-TGC ultra-weak chemiluminescence instrument, a self-developed quartz reaction tube, a carrier gas flow path, and a voltage controller for regulating the reaction temperature as main components. First, 30 mg of the pre-prepared optimized Eu / Tb co-doped La2O2CO3 sensing material is dispersed in ultrapure water, and then evenly applied to a ceramic rod, and the ceramic rod is placed in a quartz reaction tube, and the quartz reaction tube is connected to a heating device, and the quartz reaction tube is used as a reaction chamber. The OPFRs vapor to be tested flows through the reaction chamber under the drive of the carrier gas (dry air), and a CTL reaction occurs immediately. The catalytic luminescence response signal corresponding to the organophosphate flame retardant (OPFRs) is collected by the BPCL-2-TGC ultra-weak chemiluminescence instrument, with an integration time of 0.1s and an operating voltage of -0.82kV. The catalytic luminescence wavelength of OPFRs was obtained through a series of filters (460 nm, 490 nm, 520 nm, 555 nm, 575 nm, 620 nm, and 640 nm).

[0070] In order to achieve the ideal catalytic activity, the main factors affecting the CTL response intensity were investigated. First, the effect of different Eu / Tb dosages on the catalytic luminescence response signal of organophosphate flame retardants (OPFRs) was investigated, and three representative organophosphate flame retardants (OPFRs) including alkyl TEP (triethyl phosphate), chlorinated TCEP (tri(2-carboxyethyl)phosphine) and aryl-substituted TCP (tricalcium phosphate) were selected as the analysis objects. Figure 2 As shown in the figure, it can be observed that the catalytic luminescence response signal of the rare earth ion doped material is significantly higher than that of the undoped material, and when the Eu / Tb dosage is 1% of La, the catalytic luminescence response signal is the strongest, and the catalytic activity of the sensing material is the highest at this time. Therefore, the sensing material of La2O2CO3 co-doped with 1% Eu / Tb was selected. The experimental conditions such as the carrier gas flow rate and detection temperature of CTL were optimized. Taking TEP with the highest signal intensity as an example, Figure 3 As shown, the catalytic luminescence response signal intensity and signal-to-noise ratio (S / N) of the system under different flow rate conditions (100mL / min, 200mL / min, 300mL / min, 400mL / min and 500mL / min) were tested and calculated respectively. It can be found that the catalytic luminescence response signal intensity and signal-to-noise ratio first increased with the increase of flow rate and then decreased with the increase of flow rate. The catalytic luminescence response signal intensity and signal-to-noise ratio reached the highest at a carrier gas flow rate of 300mL / min. This is because the diffusion of TCP is relatively slow at a lower flow rate, and the reaction is not sufficient at an excessively high flow rate. Accordingly, a carrier gas flow rate of 300mL / min was selected as the optimal condition. Figure 4 As shown in the figure, from 193℃ to 273℃, the intensity of the catalytic luminescence response signal increases with the increase of the detection temperature, and the signal-to-noise ratio reaches the best at 263℃, so 263℃ is selected as the best detection temperature. In summary, all subsequent experiments were carried out at a sensing material dosage of 1% of La, a carrier gas flow rate of 300mL / min, and a detection temperature of 263℃. Six representative OPFRs, TEP, TNBP (tributyl phosphate), TBOEP (tri(2-butoxy)ethyl phosphate), TCEP, TCPP (tri(2-chloropropyl) phosphate) and TCP, were selected, and the single component of each compound and four multi-component mixtures of TCPP and TNBP, TCEP and TEP, TCEP and TCP, and TEP and TCP were used as training samples. The catalytic luminescence response signals of each training sample were detected at concentrations of 1.76 μg / mL, 3.52 μg / mL, 5.38 μg / mL, 7.04 μg / mL, and 8.8 μg / mL, and 10 parallel experiments were performed. Figure 5 As shown, the obtained time-signal intensity curves of different OPFRs are used as initial data to establish the initial data set.

[0071] Step 2: Preprocess the initial data set to obtain a preprocessed data set.

[0072] In some embodiments, the specific process is:

[0073] The catalytic luminescence kinetic curves of each training sample in the initial data set are averaged and sampled to unify the data length of each curve;

[0074] The initial data set after average sampling is filtered, and the principal component analysis method is used to reduce the dimension of the filtered data, extract the characteristic parameters that have a key contribution to component identification, and obtain the characteristic vector corresponding to each training sample;

[0075] The feature vectors corresponding to each training sample are combined into a feature vector set as a preprocessing data set.

[0076] According to the catalytic luminescence kinetic curves, it can be found that due to the inevitable noise in the measurement process of the catalytic luminescence response signal, the spectral vector of each training sample appears to have a lot of "burrs" after visualization. Assuming that the "burrs" caused by the noise do not contain the characteristic information we need, the catalytic luminescence response signal is filtered to obtain a more "smooth" curve, and the unfiltered curve and the filtered curve are used to establish the data set input component identification initial model and repeat the experiment 10 times for comparison. In the 10 experiments, the classification accuracy of the component identification initial model with the unfiltered curve as input (such as Figure 6 ) and quantitative accuracy (as shown in Figure 7 The average percentage error MAPE of the initial model of component identification of the input filtered curve (as shown in Figure 8 The average classification accuracy dropped to 95.41%, the average quantitative accuracy dropped to 84.44%, and the average percentage error MAPE rose to 0.13. This is enough to show that the noise in the catalytic luminescence response signal has little practical significance for the identification of OPFRs. The more essential information is still contained in the main waveform of the model, and the uniform transformation of the spectral shape during filtering is also acceptable. At the same time, the filtered catalytic luminescence kinetic curve is more conducive to the subsequent model steps focusing on more obvious waveform differences rather than noise.

[0077] The spectral vector of the catalytic luminescence kinetic curve of the training sample is processed by average sampling to make its length 64. The initial vector that is filtered but not reduced in dimension is still very long and of varying lengths. Due to the high dimensionality of the feature space, the distance between any two samples in the feature space will be extremely far, thus causing the dimensionality disaster. Therefore, the principal component analysis (PCA) and the back propagation neural network (BP) can be combined (such as Fig. 9 The problem can be solved by using the method shown in the figure. Its advantage is that PCA is a clear and reversible orthogonal transformation. Assuming that the original data is X, and the projection matrix composed of eigenvectors is P, the reduced-dimensional matrix Y can be calculated by Y=X*P, or the projection matrix P can be saved and the original matrix can be calculated by the reduced-dimensional matrix Y. At the same time, PCA transforms a set of possibly related variables into a set of linearly unrelated variables through an orthogonal transformation, which lays the foundation for the extracted feature parameters to be independent of each other, and retains the features with large differences between samples when mapping the sample set to low latitudes. Finally, the first 6 dimensions after dimensionality reduction are used as input features.

[0078] Step 3: construct an initial component recognition model based on a neural network, and train the initial component recognition model by preprocessing the data set to obtain a component recognition model.

[0079] In some embodiments, the initial component recognition model is constructed based on a neural network, and the initial component recognition model is trained by preprocessing a data set to obtain the specific process of the component recognition model:

[0080] The back propagation neural network was selected as the model architecture to construct the initial model for component identification; the initial model for component identification included an input layer, a hidden layer, and an output layer; the hyperparameters of the initial model for component identification were set, and the initial model for component identification was trained according to the preprocessed data set. The weights and biases in the network were iteratively adjusted by the back propagation algorithm to minimize the prediction error until the preset number of training rounds was reached or the preset stop condition was met to obtain the component identification model. The back propagation neural network was selected as the model architecture for the initial model for component identification because it is more suitable for the simple data set of OPFRs. The back propagation neural network was optimized according to the data set. After optimization, it contained 5 hidden layers in the middle, and the number of neurons was 48, 72, 54, 36, and 24, respectively. In order to avoid negative numbers after the data was standardized, the tanh function whose mapping range included all real numbers was selected as the activation function. The learning rate was set to 0.01, and it was reduced to 0.5 times the previous value every 500 rounds. The training was terminated after 1200 rounds and various indicators were recorded. The advantage of neural networks over other regression algorithms or ordinary machine learning methods is that they can easily predict the concentrations of multiple compounds and establish the relationship between different compounds. The key is that the distortion of the catalytic luminescence kinetic curve in a multi-component system may not be linear, which is why traditional methods have difficulty in dealing with multi-component identification. When multiple compounds coexist, spectral distortion occurs not only when the type of compound changes, but also when the concentration of the compound changes. In this case, the back propagation neural network will have a better fitting effect. In order to meet the actual requirements of multi-component identification and clearly express the test results, the constructed initial component identification model has 6 output layers, which can output the concentration test results of 6 compounds at the same time, to ensure that the initial component identification model can simultaneously evaluate and detect the concentrations of 6 OPFRs for each unknown sample, so that the final output result gives the number of components of the unknown sample and the qualitative and quantitative analysis results of each component at the same time.

[0081] In some embodiments, during the training process, the leave-one-out validation method is used to evaluate the performance of the model, and the hyperparameters of the component identification model are adjusted according to the evaluation results to obtain the optimal component identification model.

[0082] In some embodiments, during the training process, the leave-one-out validation method is used to evaluate the performance of the model, and the hyperparameters of the component identification model are adjusted according to the evaluation results. The specific process of obtaining the optimal component identification model is as follows:

[0083] The leave-one-out validation method was adopted to divide the preprocessed data set into a training set and a validation set. The component identification model was trained and validated several times, and the validation results of each time were recorded. According to the validation results, the average classification accuracy, average quantitative accuracy and average percentage error of the non-zero true concentration of the model were calculated to evaluate the performance of the model and obtain the performance evaluation results. Multiple data can better measure the prediction effect of the model. According to the performance evaluation results, the structure, parameters and hyperparameters of the component identification model were adjusted to obtain the optimal component identification model. The optimal component identification model was applied to multi-component identification of actual samples. The catalytic luminescence response signals of the samples to be tested were collected, preprocessed and input into the model to obtain the component identification results of the samples to be tested, and the accuracy and reliability of the model were verified.

[0084] According to the 6-layer output results given by the trained component recognition model, the simultaneous detection results of the 6 OPFRs (TEP, TNBP, TBOEP, TCEP, TCPP and TCP) of the training set and the validation set are plotted respectively. Through the component recognition model, 10 experiments of multi-component OPFRs were conducted to evaluate the versatility of the model. The classification accuracy, quantitative accuracy and mean percentage error (MAPE) results of the 10 experiments are shown in Table 1 below:

[0085] Table 1 Component recognition model 10 multi-component OPFRs experimental results

[0086]

[0087] As shown in Table 1, the average classification accuracy for multi-component identification of 6 OPFRs is 98.38%, and the classification accuracy of each training is 94.59% and above. The average quantitative accuracy is 94.42%, and the quantitative accuracy of each training is 90.20% and above. The average MAPE is 0.08, and the MAPE of each training is 0.10 and below. Based on this, it is verified that the constructed component recognition model can quickly and accurately realize multi-component recognition and quantitative analysis of 6 OPFRs based on the catalytic luminescence kinetic curve, including but not limited to quantitative analysis of concentration. This recognition strategy avoids the shortcomings of the existing methods in terms of insufficient multi-component recognition ability, and further reflects the great potential of the component recognition model for the recognition of multi-component OPFRs.

[0088] At the same time, the analytical performance of the component identification model was also investigated. Figures 10 to 21As shown in the figure, the catalytic luminescence response signal intensity of the six OPFRs showed a good linear relationship with different concentrations. By calculation, the detection limits of TBOEP, TEP, TCP, TNBP, TCPP and TCEP were 0.16μg / mL, 0.67μg / mL, 1.51μg / mL, 0.49μg / mL, 0.31μg / mL and 0.99μg / mL, respectively. In addition, the stability of the material is very important for gas sensing. The catalytic luminescence response signal of TBOEP on the surface of 1% Eu / Tb-doped La2O2CO3 was detected, as shown in the figure. Fig. 22 The results show that the sensing material exhibits excellent stability within 5 days, with an RSD (relative standard deviation) of less than 5.30%, which provides a strong guarantee for the long-term identification and detection of OPFRs in future practical applications.

[0089] Step 4: Collect the catalytic luminescence response signal of the sample to be detected, perform preprocessing, and input the preprocessed catalytic luminescence response signal of the sample to be detected into the component recognition model to obtain the component recognition result of the sample to be detected.

[0090] In some embodiments, according to actual needs, prepare the mixture sample to be tested, ensure that the number and state of the sample meet the experimental requirements; ensure that the state of the experimental device (such as the CTL device) is consistent with that during training, including the carrier gas flow rate, detection temperature, etc.; gasify the sample to be tested, and flow through a reaction chamber equipped with a single sensor unit under the drive of the carrier gas, so that the sample and the sensor material undergo a catalytic luminescence reaction; collect the catalytic luminescence response signal corresponding to the sample to be tested through a chemiluminescence instrument, and ensure that the parameters such as the integration time and the working voltage are consistent with those during training. The collected catalytic luminescence kinetic curve is averaged and sampled, and the data length of each curve is unified to ensure that it is consistent with the data format of the training set; the catalytic luminescence response signal is filtered using the same filtering method as during training to remove noise and obtain a smoother curve; the principal component analysis method is used to reduce the dimension of the filtered data, extract the characteristic parameters that have a key contribution to component identification, and obtain the characteristic vector corresponding to the sample to be tested. The feature vectors of the preprocessed samples to be tested are used as input data to ensure that the data format is consistent with that during training; the input data is input into the trained component identification model, and the model will automatically make predictions; the component identification results of the samples to be tested are read from the model output, including the qualitative and quantitative analysis results of each component. The component identification results predicted by the model are compared with other methods in the laboratory, such as standard detection methods or verified alternative methods, to verify the accuracy and reliability of the model; based on the comparison results, the classification accuracy, quantitative accuracy, and average percentage error of the model are evaluated to ensure that the performance of the model is stable and reliable in practical applications. The verified model is applied to the multi-component identification of actual samples, and the catalytic luminescence response signals of new samples to be tested are continuously collected and input into the model to achieve rapid and accurate identification of multi-component pollutants; based on the feedback and data accumulation in practical applications, the structure and parameters of the model are continuously optimized to improve the recognition accuracy and generalization ability of the model.

Claims

1. A multi-component identification method based on catalytic luminescence and artificial neural network, characterized in that: Includes steps: A single sensor unit is used to perform a catalytic luminescence reaction with several training samples, and a catalytic luminescence response signal corresponding to each training sample is collected to obtain an initial data set; Preprocessing the initial data set to obtain a preprocessed data set; An initial component recognition model is constructed based on a neural network, and the initial component recognition model is trained by preprocessing a data set to obtain a component recognition model; The catalytic luminescence response signal of the sample to be detected is collected and preprocessed, and the preprocessed catalytic luminescence response signal of the sample to be detected is input into the component recognition model to obtain the component recognition result of the sample to be detected.

2. The multi-component identification method based on catalytic luminescence and artificial neural network as claimed in claim 1, characterized in that: The preparation process of the single sensing unit is as follows: La(NO3)3·6H2O, urea, Eu(NO3)3·6H2O and Tb(NO3)3·6H2O were added into a certain amount of ultrapure water in proportion, and stirred until completely dissolved to form a mixed solution; The mixed solution is transferred to a high-pressure reactor to perform a hydrothermal reaction to generate a precursor precipitate; After the hydrothermal reaction is completed, the high-pressure reactor is cooled, the generated precursor precipitate is collected, and the precursor precipitate is washed with ultrapure water and anhydrous ethanol respectively to remove residual impurities; The washed precursor precipitate is vacuum dried at a preset temperature to obtain a precursor solid; The precursor solid is placed in a tube furnace, the heating rate is preset, and high-temperature calcination is performed to finally obtain Eu / Tb co-doped La2O2CO3 sensing material.

3. The multi-component identification method based on catalytic luminescence and artificial neural network as claimed in claim 1, characterized in that: The single sensing unit adopts Eu / Tb co-doped La2O2CO3 material, wherein the atomic ratio of the dosage of Eu and Tb to La is 0.5%-5% respectively.

4. The multi-component identification method based on catalytic luminescence and artificial neural network as claimed in claim 1, characterized in that: The specific process of performing catalytic luminescence reaction on a single sensor unit and several training samples and collecting catalytic luminescence response signals corresponding to each training sample to obtain the initial data set is as follows: Uniformly coating a single sensing unit on a ceramic rod in a reaction chamber; Each training sample is gasified and flows through the reaction chamber under the driving of carrier gas, so that the training sample and the sensing material undergo catalytic luminescence reaction; The catalytic luminescence response signal corresponding to each training sample is collected by a chemiluminescence instrument to form an initial data set containing the catalytic luminescence kinetic curves of each sample.

5. The multi-component identification method based on catalytic luminescence and artificial neural network as claimed in claim 4, characterized in that: The specific process of preprocessing the initial data set to obtain the preprocessed data set is as follows: The catalytic luminescence kinetic curves of each training sample in the initial data set are averaged and sampled to unify the data length of each curve; The initial data set after average sampling is filtered, and the principal component analysis method is used to reduce the dimension of the filtered data, extract the characteristic parameters that have a key contribution to component identification, and obtain the characteristic vector corresponding to each training sample; The feature vectors corresponding to each training sample are combined into a feature vector set as a preprocessing data set.

6. The multi-component identification method based on catalytic luminescence and artificial neural network as claimed in claim 1, characterized in that: The specific process of constructing the initial component recognition model based on the neural network and training the initial component recognition model by preprocessing the data set to obtain the component recognition model is as follows: A back propagation neural network is selected as the model architecture to construct an initial component identification model; wherein the initial component identification model includes an input layer, a hidden layer, and an output layer; The hyperparameters of the initial component identification model are set, and the initial component identification model is trained according to the preprocessed data set. The weights and biases in the network are iteratively adjusted through the back propagation algorithm with the goal of minimizing the prediction error until the preset number of training rounds is reached or the preset stopping condition is met, and the component identification model is obtained.

7. The multi-component identification method based on catalytic luminescence and artificial neural network according to claim 6, characterized in that: During the training process, the leave-one-out validation method is used to evaluate the performance of the model. The hyperparameters of the component recognition model are adjusted according to the evaluation results, and the optimal component recognition model is obtained.

8. The multi-component identification method based on catalytic luminescence and artificial neural network as claimed in claim 7, characterized in that: During the training process, the leave-one-out validation method is used to evaluate the performance of the model, and the hyperparameters of the component identification model are adjusted according to the evaluation results. The specific process of obtaining the optimal component identification model is as follows: The leave-one-out validation method was used to divide the preprocessed data set into a training set and a validation set. The component recognition model was trained and validated several times, and the validation results were recorded each time. According to the verification results, the average classification accuracy, average quantitative accuracy and average percentage error of non-zero true concentration of the model were calculated to evaluate the performance of the model and obtain the performance evaluation results; According to the performance evaluation results, the structure, parameters and hyperparameters of the component identification model are adjusted to obtain the optimal component identification model; The optimal component recognition model is applied to multi-component recognition of actual samples. By collecting the catalytic luminescence response signals of the samples to be tested, preprocessing them and inputting them into the model, the component recognition results of the samples to be tested are obtained, and the accuracy and reliability of the model are verified.

9. The multi-component identification method based on catalytic luminescence and artificial neural network according to claim 6, characterized in that: During the training process of the component identification initial model, the tanh function is used as the activation function, and the model structure is optimized by adjusting the number of hidden layers and the number of neurons.

10. The multi-component identification method based on catalytic luminescence and artificial neural network according to any one of claims 1 to 9, characterized in that: The training samples are mixtures containing persistent organic pollutants.

Images