Gas concentration detection system and denoising method

Abstract

The invention provides a gas concentration detection system, which comprises a laser module used for providing temperature, generating a modulation signal and outputting a light beam with a specific wavelength; the photoelectric conversion module is used for collecting a signal generated after the light beam passes through the gas absorption cell and is absorbed by the to-be-detected gas; the signal processing module is used for demodulating a second harmonic signal according to the signal collected by the photoelectric conversion module; aiming at the fact that a large amount of noise is introduced into a generated modulation signal in subsequent steps in gas concentration detection, a subtraction average optimizer algorithm is adopted to optimize a VMD parameter [K, alpha], the accuracy of signal decomposition is improved, a wavelet soft threshold method is combined to carry out noise reduction processing on a noise-containing modal component, and a pure modal component after noise reduction is reconstructed, so that the accuracy of signal decomposition is improved. And obtaining the second harmonic signal after noise reduction. According to the method, the subtraction average optimizer algorithm is introduced to optimize the VMD parameters, and the wavelet threshold denoising method is combined, so that the problems of inaccuracy and serious loss of signal details due to manual parameter selection in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of gas detection, and in particular relates to a gas concentration detection system and a denoising method. Background Art

[0002] Based on the requirements of healthy housing and safe production, in many cases, it is required to detect certain specific gases, such as indoor formaldehyde, coal mine methane and pipeline oxygen, etc. The importance of gas in industrial production and life is self-evident. Therefore, accurate monitoring and control of gas concentration not only improves production efficiency and product quality, but also helps reduce environmental pollution and ensure the safety of workers.

[0003] However, traditional gas detection methods mainly include catalytic combustion, gas chromatography and electrochemical methods, which have problems such as low detection accuracy, environmental pollution and complex operation, and the signal cannot be transmitted over long distances. In recent years, with the development of optical technology, there have been new progress in gas detection. TDLAS is a tunable diode laser absorption spectroscopy technology with high sensitivity and suitable for multi-component gases. It uses the Lambert-Beer law to continuously scan the spectrum of the measured gas to obtain the concentration of the measured gas. However, in the existing TDLAS gas detection technology, the modulated signal is demodulated in multiple complex steps. A large amount of noise will be introduced in the demodulation step, causing signal distortion and loss, and the amount of data processed is large, resulting in inaccurate detection accuracy. In this regard, in order to meet existing needs, we propose a gas concentration detection system and denoising method. Summary of the invention

[0004] In view of this, an object of the present invention is to provide a gas concentration detection system and a denoising method to solve the technical problems of low gas concentration detection accuracy, signal distortion and loss caused by a large amount of noise, etc. proposed in the above background technology.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A gas concentration detection system, comprising:

[0007] A laser module is used to provide temperature and generate modulation signals to output a light beam of a specific wavelength;

[0008] A photoelectric conversion module, used for receiving the signal generated by the absorption of the gas to be measured after the light beam passes through the gas absorption cell;

[0009] A signal processing module, used for demodulating the signal collected by the photoelectric conversion module to obtain a second harmonic signal;

[0010] The laser module includes a laser, a signal generator and a temperature control unit, the photoelectric conversion module includes a photodetector and a gas chamber, the signal processing module includes a phase-locked amplifier, a collection unit and a host computer, the signal generator is connected to the laser, the photodetector is connected to the phase-locked amplifier, and the phase-locked amplifier is connected to the collection unit.

[0011] Optionally, the signal generator is used to generate a corresponding carrier signal by controlling a power supply according to the absorption spectrum characteristics of the gas being measured, the laser is used to receive the carrier signal generated by the signal generator and output a light beam, and the temperature control unit is used to control the temperature of the laser.

[0012] Optionally, the gas chamber is used to store the gas to be detected, and the photoelectric detector is used to collect the light beam absorbed by the gas to be detected in the gas chamber, and convert the light signal into an electrical signal.

[0013] Optionally, the lock-in amplifier demodulates the electrical signal converted by the photodetector to obtain a second harmonic signal, and the acquisition unit is used to collect the second harmonic signal to the host computer for filtering processing.

[0014] Optionally, the signal generator generates a low-frequency sawtooth wave signal and a high-frequency sine wave signal.

[0015] A denoising method for a gas concentration detection system, characterized in that it comprises the following steps:

[0016] Step 1: acquiring the second harmonic signal containing noise through the acquisition unit;

[0017] Step 2: Optimize the parameters of the VMD algorithm using a subtraction mean optimizer algorithm to obtain the optimal parameters [K, α], and perform variational mode decomposition on the second harmonic signal according to the obtained optimal parameters;

[0018] Step 3, calculating the correlation coefficient between each modal component and the original second harmonic signal, and determining the effective component and noise component of each modal component;

[0019] Step 4: retain the obtained effective component and denoise the noise component using a wavelet threshold method;

[0020] Step 5: Reconstruct the denoised effective component and the effective component obtained in step 4 using a wavelet threshold method to obtain a denoised second harmonic signal.

[0021] Optionally, in step 2, the VMD algorithm decomposes the constrained variational model as follows:

[0022]

[0023] In the formula, u k is the collection of K modal components obtained by decomposition, that is, u k ={u1,u2,…,u k},ω k is the collection of the center frequencies of each modal component, that is, ω k ={ω1,ω2,…,ω k}, δ(t) is the unit time pulse function, is the partial derivative of the function with respect to time t, e -jωt is the correction index, where j is the imaginary unit and f(t) is the input signal function.

[0024] Optionally, when the subtraction mean optimizer algorithm is used to find the optimal solution in step 2, the fitness function is the minimum envelope entropy, and the calculation formula is expressed as follows:

[0025]

[0026] Where a(j) is the envelope signal of the modal component after Hilbert transform demodulation, p j is the normalized form of a(j), E p It is calculated based on information entropy, through E p Measure the effect of modal decomposition.

[0027] Optionally, the step 4 uses the wavelet threshold method as a wavelet soft threshold function, and the expression of the wavelet soft threshold function is:

[0028]

[0029] Where sgn() is the sign function, y i is the original noisy wavelet coefficient, and λ is the wavelet threshold.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention provides a denoising method for a gas concentration detection system. The VMD algorithm is improved by using a subtraction mean optimizer algorithm and a wavelet threshold is used for signal processing. The subtraction mean optimizer is used to calculate the subtraction mean of individuals to update the positions of group members in the search space, and a global optimal solution is quickly found. The subtraction mean optimizer algorithm is used to improve the decomposition ability of parameters in the VMD, and the problem of local optimal solution is effectively solved. The method has strong search ability, fast convergence, simplicity and high efficiency, avoids the decomposition signal error caused by manual selection of the decomposition layer number K and the penalty factor α, and increases the accuracy of the model. The method can also remove noise pollution in the frequency band, improve the signal-to-noise ratio, and improve the signal accuracy, and overcomes the problems of modal aliasing and residual noise after denoising and serious loss of signal details in traditional signal decomposition methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A schematic diagram of a gas concentration detection system provided by the present invention;

[0033] Figure 2 A flow chart of an algorithm for a denoising method of a gas concentration detection system provided by the present invention;

[0034] Figure 3 A schematic diagram of a curve showing how the fitness function value changes with the number of iterations when the subtraction mean optimizer algorithm is used to find the optimal solution provided by the present invention;

[0035] Figure 4 A schematic diagram of modal components obtained by decomposing a noisy second harmonic signal using the VMD algorithm with optimal parameters provided by the present invention;

[0036] Figure 5 A time domain schematic diagram of a noise-reduced second harmonic signal provided by the present invention;

[0037] In the figure, 1, laser module; 1-1, laser; 1-2, signal generator; 1-3, temperature control unit; 2, photoelectric conversion module; 2-1, photodetector; 2-2, gas chamber; 3, signal processing module; 3-1, phase-locked amplifier; 3-2, acquisition unit; 3-3, host computer. DETAILED DESCRIPTION

[0038] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, but the present invention is not limited to these embodiments. The present invention covers any substitution, modification, equivalent method and scheme made within the spirit and scope of the present invention.

[0039] In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without these detailed descriptions.

[0040] The present invention is described in more detail in the following paragraphs by way of example with reference to the accompanying drawings. It should be noted that the accompanying drawings are all simplified and not in exact proportion, and are only used for the purpose of conveniently and clearly assisting in explaining the embodiments of the present invention.

[0041] Please refer to Figure 1As shown, as an implementation method, the gas concentration detection system includes a laser module 1, which is used to provide temperature and generate a modulation signal, and output a light beam of a specific wavelength; a photoelectric conversion module 2, which is used to collect the signal generated after the light beam passes through the gas absorption cell and is absorbed by the gas to be measured; a signal processing module 3, which is used to demodulate the second harmonic signal according to the signal collected by the photoelectric conversion module; the laser module 1 includes a laser 1-1, a signal generator 1-2 and a temperature control unit 1-3, the photoelectric conversion module 2 includes a photodetector 2-1 and a gas chamber 2-2, the signal processing module 3 includes a phase-locked amplifier 3-1, a collection unit 3-2 and a host computer 3-3, the signal generator 1-2 is connected to the laser 1-1 through a signal transmission line, the signal generator 1-2 controls the power supply to generate a corresponding carrier signal according to the absorption spectrum characteristics of the gas to be measured, the laser 1-1 receives the modulation signal generated by the signal generator 1-2, and outputs a light beam, and the temperature control unit 1-3 controls the laser 1-1 within the operable temperature;

[0042] Specifically, in this embodiment, the wavelength of the laser signal output by the laser is modulated, and the modulation signal adopts a low-frequency sawtooth wave modulation signal and a high-frequency sine wave. The signal generator 1-2 modulates to generate a low-frequency 5Hz sawtooth wave signal and a high-frequency 10kHz sine wave signal. The modulated laser signal is:

[0043] I(t)=I0(1+k i sawtooth(t)+jsin(ωt)) (1)

[0044] Where I0 is the initial output signal of the laser 1-1, k i is the sawtooth wave modulation coefficient, sαwtooth(t) is the sawtooth wave signal, sin(ωt) is the superimposed high-frequency sine wave signal, and j is the sine wave modulation coefficient;

[0045] Furthermore, the gas chamber 2-2 is used to store the gas to be detected, the laser 1-1 outputs a light beam to the gas chamber 2-2, the gas to be detected absorbs the light beam, the photodetector 2-1 receives the light beam absorbed by the gas to be detected in the gas chamber 2-2, converts the absorbed light signal into a current signal through I / V, and transmits the current signal to the phase-locked amplifier 3-1 through a data transmission line. The phase-locked amplifier 3-1 demodulates the noisy second harmonic signal according to the electrical signal converted by the photodetector 2-1, and the acquisition unit 3-2 acquires the noisy second harmonic signal into the host computer 3-3 for filtering.

[0046] Please refer to Figures 2 to 5 As shown, as an embodiment, the present invention proposes a denoising method for a gas concentration detection system, comprising the following steps:

[0047] Step 1: acquiring a noisy second harmonic signal through the acquisition unit;

[0048] Step 2: Optimize the parameters of the VMD algorithm using a subtraction mean optimizer algorithm to obtain the optimal parameters [K, α], and perform variational mode decomposition on the second harmonic signal according to the obtained optimal parameters;

[0049] Step 3, calculating the correlation coefficient between each modal component and the original second harmonic signal, and determining the effective component and noise component of each modal component;

[0050] Step 4: retain the obtained effective component and denoise the noise component using a wavelet threshold method;

[0051] Step 5: Reconstruct the denoised effective component and the effective component obtained in step 4 using a wavelet threshold method to obtain a denoised second harmonic signal.

[0052] The specific implementation method of this method is described in detail below:

[0053] In a specific embodiment, in step 2, the VMD algorithm decomposes the constrained variational model as follows:

[0054]

[0055] In the formula, u k is the collection of K modal components obtained by decomposition, that is, u k ={u1,u2,…,u k},ω k is the collection of the center frequencies of each modal component, that is, ω k ={ω1,ω2,…,ω k}, δ(t) is the unit time pulse function, is the partial derivative of the function with respect to time t, e -jωt is the correction index, where j is the imaginary unit and f(t) is the input signal function;

[0056] The quadratic penalty factor α and the Lagrange multiplication operator λ(t) are introduced to transform the above constrained variational model into an unconstrained variational model:

[0057]

[0058] Continuously iteratively update the spectrum of each mode Center frequency and the Grange multiplication operator The update formula is as follows:

[0059]

[0060] In the formula, represents the spectrum of the kth mode after n+1 iterations, represents the center frequency of the kth mode after n+1 iterations, represents the Granger multiplication operator of the kth mode after n+1 iterations, where n is the number of iterations, τ is the maximum noise, and ω is the frequency;

[0061] When it is determined whether the convergence condition is met, the update is stopped. The convergence condition is expressed by the following formula:

[0062]

[0063] Where ε is the discriminant progress, and ε>0. When the condition is met, K modal components are output, otherwise continue to update;

[0064] Specifically, VMD is a variational mode decomposition method that can adaptively match the optimal center frequency of each mode with the priority, effectively separate the signal, and perform frequency domain segmentation on the signal to obtain effective signal decomposition components, thereby obtaining the optimal solution to the variational problem. The VMD algorithm requires manual selection of the decomposition layer k and the penalty factor α, and selection of appropriate K and α to ensure high-quality signal decomposition, avoid modal aliasing or over-decomposition due to the k value, resulting in false components, or signal loss or signal redundancy due to improper α selection. In order to avoid errors caused by relying solely on experience to set parameters, the subtraction average optimizer algorithm is used for adaptive optimization, thereby solving the parameter selection problem in the VMD algorithm.

[0065] In a specific embodiment, the specific optimization process of optimizing the quadratic penalty factor α and the decomposition parameter K of VMD using the subtraction mean optimizer algorithm in step 2 is as follows:

[0066] Step 2.1: Population initialization

[0067] Use a matrix to represent the population and continuously update the position of the population in the search space. The expression is:

[0068]

[0069] x i,j =lb j +r(ub j -lb j ) (9)

[0070] Where, i = 1, 2, ..., N, j = 1, 2, ..., m; x i，j For individuals, j To find the optimal lower bound, lb jTo find the upper bound, r is a random number in [0, 1]; the objective function can be evaluated by each population member, The function represents the evaluation value of the objective function:

[0071]

[0072] In the formula, is the vector representation of the evaluation value, FX i is the evaluation value of the objective function by the i-th population member;

[0073] In this embodiment, the optimization upper boundary of the penalty factor is set to 2800, and the optimization lower boundary of the penalty factor is set to 130; the optimization upper boundary of the modal decomposition layer number is set to 12, and the optimization lower boundary of the modal decomposition layer number is set to 1, and the population size of the algorithm is set to M=30, and the maximum number of iterations is 10.

[0074] Step 2.2: “-v” search

[0075] When performing individual updates, use the arithmetic mean of all populations obtained after the Nth iteration as the new position after each iteration, and use the position of the A population member minus the As the position of the B member, this method can well prevent the algorithm from falling into the problem of local optimal solution. This method is called the "-v" method:

[0076]

[0077] In the formula, is an m-dimensional vector whose value is a random number in the interval [1, 2]. F(A) and F(B) are and The value of the objective function;

[0078] Step 2.3: Search for the new location of the proxy

[0079] For each population member X i The position of each individual in the population is updated as follows:

[0080]

[0081] Where i = 1, 2, ..., N, X i is the position of the ith population member, is the new position of the ith population member, N is the total number of population members, m is a vector with normal distribution and its value range is [0,1];

[0082] Step 2.4: Generate a new position

[0083] If the result in Step 2.3 causes a change in the objective function value, then take the value obtained in Step 2.3 as the global optimal position of the entire population, and this optimal solution corresponds to the optimal combination of parameters [K, α] of VMD;

[0084] In a specific embodiment, when using the subtraction average optimizer algorithm to find the optimal solution in Step 2, the fitness function is the minimum envelope entropy. During the process of optimizing the parameters, the smaller the envelope entropy value, the higher the signal correlation degree, and vice versa. The calculation formula is expressed as:

[0085]

[0086] In the formula, a(j) is the envelope signal after the modal component undergoes Hilbert transform demodulation transformation, and p j is the normalized form of a(j), and E p is calculated according to the information entropy, and the modal decomposition effect is measured through E p

[0087] Specifically, in the parameter combination [K, α], select the minimum envelope entropy among the K components as the local minimum entropy. The component corresponding to this minimum envelope entropy value contains more pure second harmonic signal information. Use the local minimum entropy value as the fitness function for searching and updating the position, and find the parameter group [K, α] corresponding to the optimal component to complete the parameter optimization process.

[0088] Specifically, in this embodiment, after 10 iterations of using the subtraction average optimizer to update the VMD parameters, the best choice for the penalty factor α of VMD is: 872, and the best choice for the decomposition modal number K of VMD is 7;

[0089] In a specific embodiment, in Step 3, after obtaining the K modal components, to calculate the correlation between each modal component and the original signal, the Pearson correlation coefficient is used. Set the correlation coefficient greater than or equal to P as the effective modal component, where 0 < P < 1, and the expression is:

[0090]

[0091] In the formula, f is the original signal, is the mean value of the original signal, ω i represents the modal component obtained by VMD decomposition, represents the mean value of the modal component; in this embodiment, the value of P is 0.5;

[0092] ​Specifically, by calculating the correlation coefficient, the modal components with P less than 0.5 are removed, and the modal components with P greater than or equal to 0.5 are left, completing the preliminary denoising process of the noisy second harmonic signal, and obtaining the modal components without noise and the modal components with noise.

[0093] In a specific embodiment, in step 4, the wavelet threshold method is a nonlinear noise reduction algorithm, the key of which is to select an appropriate number of decomposition layers and a threshold function, and to perform noise reduction on the modal component components containing noise obtained in step 3 using a wavelet basis function according to a certain number of decomposition layers, thereby obtaining the modal component components after wavelet noise reduction.

[0094] Specifically, the wavelet basis function selected in this implementation case is the db10 wavelet function, and the number of decomposition layers is 8;

[0095] Specifically, the threshold function is mainly divided into two types: hard threshold and soft threshold. The hard threshold function is discontinuous at the threshold point. In this implementation case, the wavelet soft threshold function is used, and the expression is:

[0096]

[0097] Where sgn() is the sign function, y i is the original noisy wavelet coefficient, λ is the wavelet threshold;

[0098] In a specific embodiment, in step 5, the effective component denoised by wavelet threshold method and the effective component obtained in step 3 are superimposed and reconstructed according to certain weights to obtain a second harmonic signal after denoising.

[0099] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods or computer program products, etc. Therefore, the present invention may take the form of a complete software embodiment, a complete hardware embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0100] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0101] The above only describes the preferred embodiments of the present invention, but it should not be understood as limiting the claims. The present invention is not limited to the above embodiments, and its specific structure is allowed to be changed. In short, all changes made within the protection scope of the independent claims of the present invention are within the protection scope of the present invention.

Claims

1. A gas concentration detection system, characterized in that: include: A laser module (1), used to provide temperature and generate a modulation signal, and output a light beam of a specific wavelength; A photoelectric conversion module (2) is used to receive a signal generated by absorption of the gas to be measured after the light beam passes through the gas absorption cell; A signal processing module (3) is used to demodulate the signal collected by the photoelectric conversion module (2) to obtain a second harmonic signal; The laser module (1) comprises a laser (1-1), a signal generator (1-2) and a temperature control unit (1-3); the photoelectric conversion module (2) comprises a photoelectric detector (2-1) and a gas chamber (2-2); the signal processing module (3) comprises a lock-in amplifier (3-1), a collection unit (3-2) and a host computer (3-3); the signal generator (1-2) is connected to the laser (1-1); the photoelectric detector (2-1) is connected to the lock-in amplifier (3-1); and the lock-in amplifier (3-1) is connected to the collection unit (3-2).

2. The gas concentration detection system according to claim 1, characterized in that: The signal generator (1-2) is used to generate a corresponding carrier signal by controlling a power supply according to the absorption spectrum characteristics of the gas to be measured; the laser (1-1) is used to receive the carrier signal generated by the signal generator (1-2) and output a light beam; and the temperature control unit (1-3) is used to control the temperature of the laser (1-1).

3. The gas concentration detection system according to claim 1, characterized in that: The gas chamber (2-2) is used to store the gas to be detected, and the photoelectric detector (2-1) is used to collect the light beam absorbed by the gas to be detected passing through the gas chamber (2-2), and convert the light signal into an electrical signal.

4. The gas concentration detection system according to claim 1 or 3, characterized in that: The phase-locked amplifier (3-1) demodulates the electrical signal converted by the photodetector (2-1) to obtain a second harmonic signal, and the acquisition unit (3-2) is used to collect the second harmonic signal to the host computer (3-3) for filtering.

5. The gas concentration detection system according to claim 1, characterized in that: The signal generator (1-2) generates a low-frequency sawtooth wave signal and a high-frequency sine wave signal.

6. A denoising method for a gas concentration detection system according to any one of claims 1 to 5, characterized in that: The following steps are involved: Step 1: acquiring the second harmonic signal containing noise through the acquisition unit (3-2); Step 2: Optimize the parameters of the VMD algorithm using a subtraction mean optimizer algorithm to obtain the optimal parameters [K, α], and perform variational mode decomposition on the second harmonic signal according to the obtained optimal parameters; Step 3, calculating the correlation coefficient between each modal component and the original second harmonic signal, and determining the effective component and noise component of each modal component; Step 4: retain the obtained effective component and denoise the noise component using a wavelet threshold method; Step 5: Reconstruct the denoised effective component and the effective component obtained in step 4 using a wavelet threshold method to obtain a denoised second harmonic signal.

7. A denoising method for a gas concentration detection system according to claim 6, characterized in that: In step 2, the VMD algorithm decomposes the constrained variational model into: In the formula, u k is the collection of K modal components obtained by decomposition, that is, u k ={u1,u2,…,u k },ω k is the collection of the center frequencies of each modal component, that is, ω k ={ω1,ω2,…,ω k }, δ(t) is the unit time pulse function, is the partial derivative of the function with respect to time t, e -jωt is the correction index, where j is the imaginary unit and f(t) is the input signal function.

8. A denoising method for a gas concentration detection system according to claim 6, characterized in that: When the step 2 uses the subtraction mean optimizer algorithm to find the best solution, the fitness function is the minimum envelope entropy, and the calculation formula is expressed as: Where a(j) is the envelope signal of the modal component after Hilbert transform demodulation, p j is the normalized form of a(j), E p It is calculated based on information entropy, through E p Measure the effect of modal decomposition.

9. A denoising method for a gas concentration detection system according to claim 6, characterized in that: The step 4 adopts the wavelet threshold method as a wavelet soft threshold function, and the expression of the wavelet soft threshold function is: Where sgn() is the sign function, y i is the original noisy wavelet coefficient, and λ is the wavelet threshold.

Images