Monitoring method and device for dual-point near-infrared spectroscopy signal separation, storage medium

Abstract

The application provides a kind of double-measuring-point near-infrared spectrum signal separation monitoring method, it is characterized in that: including: setting two measuring points on industrial production line, respectively collecting near-infrared spectrum signal of feed end and discharge end;Two near-infrared spectrum signals of the two measuring points are combined into one, form mixed spectrum signal, and are guided to the entrance of slit of near-infrared spectrometer, the collected mixed spectrum signal is carried out smoothing processing;The mixed spectrum signal is separated mathematically using constrained independent component analysis method, and the spectral characteristics of feed end and discharge end are extracted;The spectral characteristics of feed end and discharge end are combined, respectively construct the quantitative analysis model of feed end and discharge end, and the component prediction is carried out through the real-time collected spectral data.The scheme makes the optical signals of multiple measuring points enter the spectrometer at the same time by optimizing the optical fiber design, and then the signal separation is realized by using mathematical method, so that the near-infrared prediction model of different measuring points is established respectively.

Description

Technical Field

[0001] This invention relates to near-infrared spectroscopy intelligent analysis technology, and more particularly to a monitoring method, device, and storage medium for separating near-infrared spectral signals at two measurement points. Background Technology

[0002] Near-infrared spectroscopy is widely used in process monitoring due to its rapid and non-destructive characteristics. However, because near-infrared spectrometers are relatively expensive, in industrial applications where multiple measurement points need to be measured simultaneously, using a single spectrometer to monitor multiple points is the best solution to reduce equipment investment costs. Traditional methods typically rely on fiber optic multiplexers to acquire spectral data from different measurement points through mechanical switching devices or controlled timing. This reliance on movable parts (such as mechanical switching devices) not only increases system complexity and maintenance costs but may also cause measurement timing delays, making it unsuitable for high-frequency continuous monitoring.

[0003] Meanwhile, losses and deviations during optical path switching may reduce data consistency between different measurement points, thus affecting modeling accuracy and long-term stability. In many monitoring processes, the spectral signals at different measurement points (such as the feed and discharge ends) may exhibit significant differences. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a monitoring method for separating near-infrared spectral signals at two measurement points, comprising:

[0005] Two measuring points are set up on the industrial production line to collect near-infrared spectral signals at the feed end and the discharge end, respectively;

[0006] The two near-infrared spectral signals from the two measurement points are combined into one to form a mixed spectral signal, which is then guided to the slit entrance of the near-infrared spectrometer. The acquired mixed spectral signal is then smoothed.

[0007] The mixed spectral signals were mathematically separated using constrained independent component analysis to extract spectral features from the feed and discharge ends.

[0008] By combining the spectral characteristics of the feed end and the discharge end, quantitative analysis models for the feed end and the discharge end are constructed respectively, and component prediction is performed using real-time acquired spectral data.

[0009] In some embodiments, the near-infrared spectral signals from the two measurement points are guided to a beam splitter via an optical fiber, and the beam splitter merges the two near-infrared spectral signals into one, forming a mixed spectral signal.

[0010] In some embodiments, the two near-infrared spectral signals are arranged symmetrically with the beam splitter at a 45° angle.

[0011] In some embodiments, the constrained independent component analysis method includes the following steps:

[0012] Assume that the near-infrared spectral signal X consists of the spectrum X1 at the feed end of measuring point 1 and the spectrum X2 at the discharge end of measuring point 2, that is:

[0013] X(λ,t)=a(t)S1(λ)+b(t)S2(λ)+∈(λ)

[0014] Where: S1 and S2 correspond to the standard spectra of the materials at the feed end and the discharge end, respectively; a(t) and b(t) are time-varying mixing coefficients, respectively; ∈ is the noise term, which is first removed by SG or smoothing filter;

[0015] An arbitrary standard spectrum R(λ) at the feed end is introduced as a constraint to establish the objective function, J(w) = non-Gaussianity measure - γ·constraint term, where γ is the constraint weight, and the value range is 0 to 1;

[0016] The objective function is solved using Newton's iteration method, and the spectral signals at the feed end and the discharge end are separated.

[0017] In some embodiments, an arbitrary standard spectrum R(λ) at the feed end is introduced as a constraint to establish the objective function as follows:

[0018] maxJ(w)=E{G(w T X)}-ρ·Corr(w T X,R)

[0019] Where G is a nonlinear function measuring non-Gaussianity, and ρ is a constraint weight with a value ranging from 0.1 to 0.9. In some embodiments, the solution is obtained using Newton's iteration method.

[0020]

[0021] Iterate n times to separate in,

[0022] In some embodiments, the constrained independent component analysis method further includes the following steps:

[0023] The separated signals are correlated using either Euclidean distance or similarity coefficient, with the similarity coefficient being the preferred choice.

[0024]

[0025] The value of δ is generally greater than 0.95. Depending on the signal quality, both signals can be selected to satisfy the verification, or either one can be satisfied.

[0026] If the verification condition is not met within n iterations, then an adaptive hybrid matrix update is triggered:

[0027] [a,b] T =(S T S) -1 S T X

[0028] Where S = [S1, S2] is the calibration spectral matrix composed of the standard spectra of the inlet and outlet.

[0029] In some embodiments, the quantitative analysis model is constructed using methods such as multiple linear regression, principal component regression, partial least squares regression, support vector regression, or deep learning.

[0030] The present invention also provides a monitoring device for separating near-infrared spectral signals at two measurement points, comprising:

[0031] The data acquisition module sets up two measuring points on the industrial production line to collect near-infrared spectral signals from the feed end and the discharge end, respectively.

[0032] The preprocessing module merges the two near-infrared spectral signals from the two measurement points into one, forming a mixed spectral signal, and guides it to the slit entrance of the near-infrared spectrometer to smooth the acquired mixed spectral signal.

[0033] The feature extraction module uses constrained independent component analysis to mathematically separate the mixed spectral signals and extract the spectral features of the feed end and the discharge end;

[0034] The prediction module combines the spectral characteristics of the feed end and the discharge end to construct quantitative analysis models for the feed end and the discharge end respectively, and performs component prediction based on the real-time collected spectral data.

[0035] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0036] As can be seen from the above solutions, the advantages of the present invention are:

[0037] The monitoring method based on dual-measurement-point near-infrared spectral signal separation provided by this invention is suitable for simultaneous in-situ monitoring of signals at multiple measurement points with different characteristics during the production process. The technical solution provided by this invention optimizes the optical fiber design so that the optical signals from multiple measurement points enter the spectrometer simultaneously, and then uses mathematical methods to achieve signal separation, thereby establishing near-infrared prediction models for different measurement points. Attached Figure Description

[0038] Figure 1 As shown in the embodiments of the present invention Figure 1A schematic diagram of the overall process of the monitoring method for separating near-infrared spectral signals at two measurement points, as shown in an embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of the light mixing device shown in an embodiment of the present invention;

[0040] Figure 3 This is a flowchart illustrating the steps of the constrained independent component analysis method according to an embodiment of the present invention;

[0041] Figure 4 The PLS regression model diagram is shown in the embodiment of the present invention, which uses SG derivative and MSC for spectral preprocessing of straw, and CARS for feature engineering.

[0042] Figure 5 As shown in the embodiments of the present invention, the straw is preprocessed using SG derivative and SNV, and feature engineering is performed using CARS to establish a PLS regression model diagram.

[0043] Figure 6 This is a schematic diagram of the overall structure of the monitoring device for separating near-infrared spectral signals at two measurement points, as shown in an embodiment of the present invention.

[0044] In the attached figures, the following labels are used:

[0045] 300- Monitoring device for separating near-infrared spectral signals at dual measurement points;

[0046] 310 - Acquisition Module;

[0047] 320 - Preprocessing module;

[0048] 330 - Feature Extraction Module;

[0049] 340 - Prediction Module;

[0050] S1-S4: Steps. Detailed Implementation

[0051] To make the above features and effects of the present invention clearer and easier to understand, specific embodiments are described below, and detailed descriptions are provided in conjunction with the accompanying drawings.

[0052] See Figure 1 An embodiment of the present invention provides a monitoring method for separating near-infrared spectral signals at two measurement points, comprising:

[0053] S1. Set up two measuring points on the industrial production line to collect near-infrared spectral signals at the feed end and the discharge end, respectively;

[0054] S2. The two near-infrared spectral signals from the two measuring points are combined into one to form a mixed spectral signal, which is then guided to the slit entrance of the near-infrared spectrometer. The acquired mixed spectral signal is then smoothed. Specifically, the near-infrared spectral signals from the two measuring points are guided to a beam splitter via an optical fiber. The beam splitter combines the two near-infrared spectral signals into one to form a mixed spectral signal.

[0055] S3. Use constrained independent component analysis to mathematically separate the mixed spectral signals and extract the spectral features of the feed end and the discharge end;

[0056] S4. Combining the spectral characteristics of the feed end and the discharge end, quantitative analysis models are constructed for the feed end and the discharge end respectively, and component prediction is performed using real-time acquired spectral data.

[0057] This embodiment employs a dual-probe structure, positioned at the inlet and outlet ends respectively. Near-infrared light signals from each measurement point are collected via optical fiber. To reduce equipment size and optimize the way light signals enter the spectrometer, see [reference needed]. Figure 2 This is an optical mixing device. Specifically, a beam splitter is used to combine two optical signals into one, which is then guided to the spectrometer via a spatial optical path or optical fiber, and finally to the slit entrance of the spectrometer. In this embodiment, the beam splitter is a non-polarized beam splitter cube with a certain splitting ratio, and its operating wavelength covers the 900-2500nm near-infrared region. The two near-infrared spectral signals are symmetrically arranged with the beam splitter at a 45° angle to optimize the combining efficiency of the optical signals.

[0058] Furthermore, since the spectral signals from different measuring points are mixed and then enter the same spectrometer, the directly measured spectral data are a mixture of light signals from the feed and discharge ends. Therefore, in the data processing stage, this invention first smooths the spectral signals to reduce the impact of noise on subsequent analysis. Then, constrained independent component analysis (cICA) is used to mathematically separate the spectral signals. cICA, based on the principle of independent component analysis (ICA) and combined with known prior information, unmixes the signals, effectively improving the separation accuracy of spectral signals from different measuring points and ensuring the accurate extraction of spectral data from each measuring point. For details, see [link to documentation]. Figure 3 The constrained independent component analysis method includes the following steps:

[0059] Assume that the near-infrared spectral signal X consists of the spectrum X1 at the feed end of measuring point 1 and the spectrum X2 at the discharge end of measuring point 2, that is:

[0060] X(λ,t)=a(t)S1(λ)+b(t)S2(λ)+∈(λ)

[0061] Where: S1 and S2 correspond to the standard spectra of the materials at the feed end and the discharge end, respectively; a(t) and b(t) are time-varying mixing coefficients, respectively; ∈ is the noise term, which is first removed by SG or smoothing filter;

[0062] An arbitrary standard spectrum R(λ) at the feed end is introduced as a constraint to establish the objective function, J(w) = non-Gaussianity measure - γ·constraint term, where γ is the constraint weight, and the value range is 0 to 1;

[0063] The objective function is solved using Newton's iteration method, and the spectral signals at the feed and discharge ends are separated.

[0064] The objective function is established by introducing any standard spectrum R(λ) at the feed end as a constraint, as follows:

[0065] maxJ(w)=E{G(w T X)}-ρ·Corr(w T X,R)

[0066] Where G is a nonlinear function that measures non-Gaussianity, and ρ is a constraint weight with a value of 0.1 to 0.9.

[0067] The solution is obtained using Newton's iteration method:

[0068]

[0069] Iterate n times to separate in,

[0070] Furthermore, the correlation of the separated signals is verified using Euclidean distance or similarity coefficient, with the similarity coefficient being the preferred choice.

[0071]

[0072] The value of δ is generally greater than 0.95. Depending on the signal quality, both signals can be selected to satisfy the verification, or either one can be satisfied.

[0073] If the verification condition is not met within n iterations, then an adaptive hybrid matrix update is triggered:

[0074] [a,b] T =(S T S) -1 S T X

[0075] Where S = [S1, S2] is the calibration spectral matrix composed of the standard spectra of the inlet and outlet.

[0076] The quantitative analysis model construction method described in this embodiment includes multiple linear regression, principal component regression, partial least squares regression, support vector regression, or deep learning methods.

[0077] The standard model establishment and prediction in this embodiment of the invention are as follows:

[0078] First, absorbance calculation:

[0079] Using the blank spectrum collected at any measurement point as the reference I0, the absorbance is calculated according to the Lambert-Beer law:

[0080] A = -log 10 (I / I0)

[0081] The reference can be fixed or a blank spectrum collected at regular intervals. The blank spectrum can be collected by the light source and detector through a direct beam, or it can be the spectrum of a standard diffuse reflectance white board (which can be a ceramic board, a polytetrafluoroethylene white board, a gold-plated board, etc.).

[0082] Furthermore, a quantitative analysis model is constructed:

[0083] Using the separated spectra and the corresponding reference values ​​of the materials (usually property data obtained by standard laboratory methods), a quantitative analysis model can be established through spectral preprocessing, feature engineering, and regression methods (multivariate linear regression, principal component regression, partial least squares regression, support vector regression, univariate linear regression, etc.). Alternatively, a quantitative analysis model can be established end-to-end through deep learning.

[0084] Finally, a prediction is made:

[0085] The spectra of materials with unknown properties are collected, and after signal separation and absorbance calculation, they are subjected to the same spectral preprocessing and feature engineering, and then input into the quantitative analysis model to obtain the required quantitative analysis data.

[0086] The monitoring method based on dual-point near-infrared spectral signal separation provided in this embodiment is applicable to scenarios with ≥2 measurement points. It achieves the merging and separation of multiple optical signals through cascaded beam splitters or multi-channel beam splitting devices. The spectral type can be near-infrared, Raman, or ultraviolet. The reference spectrum can be a fixed reference or a periodically acquired blank spectrum. The blank spectrum acquisition method includes through-beam acquisition by the light source and detector or acquisition using a standard diffuse reflection white plate. The standard diffuse reflection white plate can be a ceramic plate, a polytetrafluoroethylene white plate, or a gold-plated plate. The fiber core diameter is 200-600 μm, and the numerical aperture is 0.22-0.39 to ensure optical signal coupling efficiency.

[0087] The technical solution of this invention will be described in detail below, taking online monitoring of the biochar production process as an example:

[0088] (1) Monitoring content: fixed carbon of straw and calorific value of biochar

[0089] (2) Experimental setup

[0090] Two measuring points were set up on the biochar production line, one on the conveyor belt of the feeder at the inlet and the other on the pipe at the outlet. Signals from the two probes were transmitted to an optical signal mixing device via optical fiber (400 μm core diameter). The beam splitter used in the mixing device had a splitting ratio of 30:70, with the inlet measuring point at point 30 and the outlet measuring point at point 70. A near-infrared spectrometer was used to collect the combined signal; the mixing device was connected to the spectrometer via a 600 μm core diameter optical fiber. A reference plate was installed at the inlet measuring point and moved to the lower end of the measuring port by a motor.

[0091] (3) Data Acquisition

[0092] Spectra were acquired every 5 minutes, and the reference signal was acquired every 8 hours. During the modeling phase, the time points of the sampling were recorded, and three spectra before and after the sampling were captured to build the calibration model.

[0093] (4) Signal Processing

[0094] The original signal was smoothed using the SG method, with a window width of 13, a fitting order of 2, and a derivative degree of 0. The inlet and outlet signals were separated using the cICA method.

[0095] (5) Model building

[0096] Calculate the absorbance using the separated spectrum and the latest updated reference spectrum, see [link / reference]. Figure 4 The straw samples underwent spectral preprocessing using SG derivatives and MSC, and feature engineering was performed using CARS to establish a PLS regression model; see reference. Figure 5 The straw was preprocessed using SG derivative and SNV, and feature engineering was performed using CARS to establish a PLS regression model.

[0097] (6) Online monitoring

[0098] The spectrum is acquired in real time and separated using cICA. The absorbance is calculated, the spectrum is preprocessed, the characteristic bands are selected, and the data are input into the model to obtain the predicted values.

[0099] Further, see Figure 6 Another embodiment of the present invention provides a monitoring device 300 for dual-point near-infrared spectral signal separation, comprising:

[0100] The acquisition module 310 sets up two measuring points on the industrial production line to collect near-infrared spectral signals at the feed end and the discharge end, respectively.

[0101] The preprocessing module 320 merges the two near-infrared spectral signals from the two measurement points into one, forming a mixed spectral signal, and guides it to the slit entrance of the near-infrared spectrometer to smooth the acquired mixed spectral signal.

[0102] The feature extraction module 330 uses constrained independent component analysis to mathematically separate the mixed spectral signal and extract the spectral features of the feed end and the discharge end;

[0103] The prediction module 340 combines the spectral characteristics of the feed end and the discharge end to construct quantitative analysis models for the feed end and the discharge end respectively, and performs component prediction through real-time acquired spectral data.

[0104] This device embodiment can be implemented in conjunction with the implementation methods described above. The relevant technical details mentioned in the implementation methods described above remain valid in the implementation methods of this device embodiment, and will not be repeated here to avoid repetition.

[0105] Another embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the online monitoring method based on dual-point near-infrared spectral signal separation in the above embodiments.

[0106] Furthermore, it should be understood that the storage medium in the device embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0107] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0108] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A monitoring method for separating near-infrared spectral signals at two measurement points, characterized in that: include: Two measuring points are set up on the industrial production line to collect near-infrared spectral signals at the feed end and the discharge end, respectively; The two near-infrared spectral signals from the two measurement points are combined into one to form a mixed spectral signal, which is then guided to the slit entrance of the near-infrared spectrometer. The acquired mixed spectral signal is then smoothed. The mixed spectral signals were mathematically separated using constrained independent component analysis to extract spectral features from the feed and discharge ends. By combining the spectral characteristics of the feed end and the discharge end, quantitative analysis models for the feed end and the discharge end are constructed respectively, and component prediction is performed using real-time acquired spectral data; The constrained independent component analysis method includes the following steps: Assume that the near-infrared spectral signal X consists of the spectrum X1 at the feed end of measuring point 1 and the spectrum X2 at the discharge end of measuring point 2, that is: Where: S1 and S2 correspond to the standard spectra of the materials at the feed and discharge ends, respectively; a(t) and b(t) are the time-varying mixing coefficients, respectively. It's a noise term; first, use SG or smoothing filters to remove the noise term. Remove; An arbitrary standard spectrum R(λ) at the feed end is introduced as a constraint to establish the objective function, J(w) = non-Gaussianity measure − γ ⋅ constraint term, where γ is the constraint weight, and the value range is 0~1; The objective function is solved using Newton's iteration method, and the spectral signals at the feed end and the discharge end are separated. .

2. The monitoring method for separating near-infrared spectral signals at dual measurement points according to claim 1, characterized in that: The near-infrared spectral signals from the two measurement points are guided to a beam splitter via an optical fiber. The beam splitter then merges the two near-infrared spectral signals into a single, mixed spectral signal.

3. The monitoring method for separating near-infrared spectral signals at dual measurement points according to claim 2, characterized in that: The two near-infrared spectral signals are arranged symmetrically with the beam splitter at a 45° angle.

4. The monitoring method for separating near-infrared spectral signals at dual measurement points according to claim 1, characterized in that: Introducing any standard spectrum R(λ) from the feed end as a constraint, the objective function is established as follows: Among them, For nonlinear functions, measure non-Gaussianity. The constraint weights are set to a value between 0.1 and 0.

9.

5. The monitoring method for separating near-infrared spectral signals at dual measurement points according to claim 1, characterized in that: in, Solve using Newton's iteration method: Iterate n times to separate ,in, , .

6. The monitoring method for separating near-infrared spectral signals at dual measurement points according to claim 5, characterized in that: The constrained independent component analysis method further includes the following steps: The separated signals are correlated using Euclidean distance or similarity coefficient, where the similarity coefficient is: If the value is greater than 0.95, depending on the signal quality, you can choose to satisfy the verification for both signals, or you can choose to satisfy either one. If the verification condition is not met within n iterations, then an adaptive hybrid matrix update is triggered: Where S=[S1,S2] is the calibration spectral matrix composed of the standard spectra of the inlet and outlet.

7. The monitoring method for separating near-infrared spectral signals at dual measurement points according to claim 1, characterized in that: The methods for constructing the quantitative analysis model include multiple linear regression, principal component regression, partial least squares regression, support vector regression, or deep learning methods.

8. A monitoring device for separating near-infrared spectral signals at dual measurement points, characterized in that: The monitoring device, employing the monitoring method as described in any one of claims 1-7, comprises: The data acquisition module sets up two measuring points on the industrial production line to collect near-infrared spectral signals from the feed end and the discharge end, respectively. The preprocessing module merges the two near-infrared spectral signals from the two measurement points into one, forming a mixed spectral signal, and guides it to the slit entrance of the near-infrared spectrometer to smooth the acquired mixed spectral signal. The feature extraction module uses constrained independent component analysis to mathematically separate the mixed spectral signals and extract the spectral features of the feed end and the discharge end; The prediction module combines the spectral characteristics of the feed end and the discharge end to construct quantitative analysis models for the feed end and the discharge end respectively, and performs component prediction based on the real-time collected spectral data.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the method described in any one of claims 1-7.

Images