阅读说明：本技术 一种利用狼群算法的fbg非均匀温度重构方法 (FBG (fiber Bragg Grating) non-uniform temperature reconstruction method by utilizing wolf colony algorithm ) 是由 邹红波 陶娟 于 2020-04-14 设计创作，主要内容包括：一种利用狼群算法的FBG非均匀温度重构方法,首先将FBG传感器粘贴固定在铝合金板上,对铝合金板进行局部受热的方式以使FBG传感器受到一个非均匀温度,通过光谱仪显示非均匀温度下的FBG反射谱波形,再利用狼群算法WPA对FBG反射谱波形进行非均匀温度重构,最后通过PC机显示结果。通过传输矩阵法对非均匀温度分布下的反射光谱进行模拟,这些光谱作为重构时的目标光谱,适应度函数定义为目标光谱和重构光谱之间的误差,使用狼群算法WPA重构温度分布。本发明一种利用狼群算法的FBG非均匀温度重构方法,采用狼群算法WPA,依据非均匀温度下的FBG反射光谱,结合FBG反射光谱分析的传输矩阵法,解决FBG的非均匀温度重构问题。具有重构速度快,精度高等优点。(The utility model provides an utilize inhomogeneous temperature reconstruction method of FBG of wolf crowd algorithm, pastes the FBG sensor at first and fixes on aluminum alloy plate, carries out the mode that local is heated so that the FBG sensor receives an inhomogeneous temperature to aluminum alloy plate, shows FBG reflection spectrum waveform under the inhomogeneous temperature through the spectrum appearance, utilizes wolf crowd algorithm WPA to carry out inhomogeneous temperature reconstruction to FBG reflection spectrum waveform again, shows the result through PC at last. The reflection spectrums under the non-uniform temperature distribution are simulated through a transmission matrix method, the spectrums are used as target spectrums during reconstruction, a fitness function is defined as an error between the target spectrums and the reconstruction spectrums, and the temperature distribution is reconstructed by using a wolf pack algorithm WPA. The invention relates to a FBG non-uniform temperature reconstruction method by utilizing a wolf pack algorithm. The method has the advantages of high reconstruction speed, high precision and the like.)

1. A FBG non-uniform temperature reconstruction method using a wolf pack algorithm is characterized in that: the method comprises the steps of firstly applying non-uniform temperature on an FBG sensor (2), displaying an FBG reflection spectrum waveform under the non-uniform temperature through a spectrometer (3), then performing non-uniform temperature reconstruction on the FBG reflection spectrum waveform by using a wolf pack algorithm WPA (4), and finally displaying a result through a PC (5).

2. The FBG nonuniform temperature reconstruction method by using the wolf pack algorithm according to claim 1, characterized in that:

firstly, the FBG sensor (2) is adhered and fixed on the aluminum alloy plate (1), and the FBG sensor (2) is subjected to non-uniform temperature by locally heating the aluminum alloy plate (1).

3. The FBG nonuniform temperature reconstruction method by using the wolf pack algorithm according to claim 1, characterized in that: the reflection spectrums under the non-uniform temperature distribution are simulated through a transmission matrix method, the spectrums are used as target spectrums during reconstruction, a fitness function is defined as an error between the target spectrums and the reconstruction spectrums, and the temperature distribution is reconstructed by using a wolf pack algorithm WPA.

4. The FBG nonuniform temperature reconstruction method by using the wolf pack algorithm according to claim 1, characterized in that: the wolf pack algorithm WPA comprises the following steps:

step 1: carrying out initialization setting, wherein the set parameters comprise N artificial wolfs and X initial position_iAt the mostLarge number of wandering times T_maxAnd number of iterations k_maxStep size factor S, distance decision factor ω, wolf detection, update scale factors α and β;

step 2: updating the position of the wandering wolf according to the formula (1) until the fitness Y of the wolf is reached_i>Y_leadReplacing the wolf with the exploring wolf, proceeding to the next step, otherwise, continuing to wander until reaching the maximum times T_maxThen carrying out the next step;

in the formula, the wolf is visited to walk in h directions in total, and the step length of the walking isThe initial fitness of the wolf detection is Y_iDetecting wolf at current position x_idMoving to p (p is 1,2, …, h) directions;

and step 3: updating the position of the rushing wolf according to the formula (2) until the detection fitness Y of the wolf_i>Y_leadReplacing the wolf with the wolf, otherwise continuing to attack until the distance from the wolf is less than d_nearCarrying out the next step;

in the formula (I), the compound is shown in the specification,for the position of the wolf at the kth iteration,step size for wolf rushes;

and 4, step 4: performing containment attack on the prey according to the formula (3), and updating the wolf head;

in the formula (I), the compound is shown in the specification,the position of the prey is the position of the wolf,the step length of the wolf colony attacking the prey;

and 5: updating the whole wolf group;

step 6: repeating the evolution cycle: steps 2-5 are repeated until a stopping rule is met, typically a sufficiently small fitness value or a sufficiently large number of iterations.

5. The FBG nonuniform temperature reconstruction method by using the wolf pack algorithm according to claim 2, characterized in that: the transmission matrix method equally divides a non-uniform grating into M sections, each section is regarded as a uniform grating, and R is defined_iAnd S_iThe amplitudes of the forward and backward transmission modes of the i-th segment, respectively, through which the electromagnetic wave is described as:

in the formula, R_iAnd S_iRespectively representing the amplitudes of the forward and backward transmission modes, R, of the i-th section of the grating_i-1And S_i-1Respectively representing the amplitudes of the forward and backward transmission modes, F, of the i-1 th section of the grating_iThe transmission matrix of the ith section of grating is expressed as follows:

in the formula:k is AC coupling coefficient, sigma is DC self-coupling coefficient, Δ z is length of each segment, i represents ith segment lightA gate; the transmission mode of the entire grating is then described as:

in the formula, R₀And S₀The amplitudes of the forward and backward transmission modes of the initial section of the grating are respectively expressed, R and S respectively express the amplitudes of the forward and backward transmission modes of the whole grating, F is the transmission matrix of the whole grating, and the expression is as follows:

in the formula, F₁,F₂…F_MThe transmission matrix, f, respectively represented as the 1 st and 2 nd segments … Mth segment gratings₁₁,f₁₂,f₂₁,f₂₂Is an element in the transmission matrix F;

the reflection coefficient of FBG is:

when one FBG sensor is subjected to a non-uniform temperature T (z), the equivalent period Λ (z) of the FBG is described as:

Λ(z)＝Λ₀[1+(α+ξ)T(z)]0≤z≤L (9)；

wherein L is the length of the grating, Λ₀Is the period of the FBG at normal temperature, α is the thermo-optic coefficient, ξ is the thermal expansion coefficient, T (z) represents the non-uniform temperature to which the FBG sensor is subjected, z represents a point in the axial direction of the grating, and k and sigma in equation (5) are described as:

in the formula: n is_effIs an optical fiberΔ n is the index modulation depth, s is the fringe visibility, λ is the wavelength of the FBG;

for any given parameter L, Λ₀,n_effΔ n, s, α and T (z), from the above equation, the reflection spectrum of the FBG at non-uniform temperatures can be calculated.

Technical Field

The invention relates to the technical field of FBG sensor temperature measurement, in particular to a FBG non-uniform temperature reconstruction method by utilizing a wolf colony algorithm.

Background

When the FBG sensor is subjected to non-uniform strain or temperature, the FBG's reflection spectrum may be distorted or even peaky, in which case it is not feasible to measure only the Bragg wavelength shift to obtain strain or temperature data. The reflection spectrum of the FBG sensor under non-uniform strain or temperature can be easily measured by the optical spectrum analyzer, so that how to reconstruct the strain or temperature distribution from the reflection spectrum becomes an important inverse problem.

There are several approaches to this problem, the simplest one being based on intensity spectroscopy, but this approach is only applicable to monotonic strain or temperature profiles, i.e. increasing or decreasing profiles. Fourier transform methods are also feasible, which have the advantage of a short reconstruction time, but which are only suitable for weak gratings and require knowledge of both the magnitude and phase spectra of the fiber grating. In general, the measurement of the phase spectrum requires a more complex apparatus than the measurement of the amplitude spectrum. Thus, a reconstruction method that requires only a magnitude spectrum is advantageous over a method that requires a phase spectrum.

In recent years, some evolutionary algorithms have been used to solve the problem of how to reconstruct the strain or temperature distribution from the reflection spectrum, mainly including simulated annealing algorithms and genetic algorithms. These methods only require the magnitude spectrum of the fiber grating for reconstruction, and do not require the phase spectrum of the fiber grating, and thus have been widely used to solve this problem. However, the simulated annealing algorithm and the genetic algorithm have long running time, and parameters also need to be optimized to skip the local optimal solution, so that some improved algorithms mainly comprise a chaotic genetic algorithm, an adaptive simulated annealing algorithm and a simulated annealing evolution algorithm.

The chaotic genetic algorithm is complex, the program running time is long, and the algorithm convergence speed is slow; the self-adaptive simulated annealing algorithm is easy to fall into a local minimum value in a multivariable optimization problem, and meanwhile, the method is insufficient in the aspects of calculation aging and precision; the parameters of the simulated annealing evolution algorithm are difficult to control, and the optimal value can not be converged once, and the optimal value can be obtained by multiple attempts.

Disclosure of Invention

In order to solve the technical problem, the invention provides a method for reconstructing the non-uniform temperature of an FBG (fiber Bragg Grating) by using a wolf pack algorithm WPA (fiber Bragg Grating), and the problem of the non-uniform temperature reconstruction of the FBG is solved by combining a transmission matrix method of FBG reflection spectrum analysis according to the FBG reflection spectrum under the non-uniform temperature. The method has the advantages of high reconstruction speed, high precision and the like.

The technical scheme adopted by the invention is as follows:

the utility model provides an utilize inhomogeneous temperature reconstruction method of FBG of wolf crowd algorithm, pastes the FBG sensor at first and fixes on aluminum alloy plate, carries out the mode that local is heated so that the FBG sensor receives an inhomogeneous temperature to aluminum alloy plate, shows FBG reflection spectrum waveform under the inhomogeneous temperature through the spectrum appearance, utilizes wolf crowd algorithm WPA to carry out inhomogeneous temperature reconstruction to FBG reflection spectrum waveform again, shows the result through PC at last.

The reflection spectrums under the non-uniform temperature distribution are simulated through a transmission matrix method, the spectrums are used as target spectrums during reconstruction, a fitness function is defined as an error between the target spectrums and the reconstruction spectrums, and the temperature distribution is reconstructed by using a wolf pack algorithm WPA.

The invention discloses a FBG non-uniform temperature reconstruction method by utilizing a wolf pack algorithm, which has the following technical effects:

1) and simulating the FBG reflection spectrum under the actual non-uniform temperature by adopting a transmission matrix method, and reconstructing the non-uniform temperature distribution by using a wolf colony algorithm. The reconstruction method can be used for the non-uniform temperature measurement of an actual structure.

2) Compared with the traditional evolutionary algorithm, such as a simulated annealing algorithm and a genetic algorithm, the wolf colony algorithm has better convergence precision and global search capability, thereby having the advantages of high reconstruction speed, high precision and the like.

3) The wolf colony algorithm is used for reconstructing the nonuniform temperature of the FBG, the defect that the traditional FBG sensor can only measure the uniform temperature is overcome, and the temperature measurement field of the FBG sensor is further expanded by the uniform temperature and the nonuniform temperature.

Drawings

FIG. 1 is a flow chart of the reconstruction of the present invention;

FIG. 2 is a graph comparing the original temperature and the reconstructed temperature distribution.

Detailed Description

Principle analysis:

when the FBG sensor 2 is subjected to a uniform strain or temperature, the center wavelength thereof linearly drifts with the strain or temperature, in which case only the Bragg wavelength drift needs to be measured to obtain an external strain or temperature value. But when the FBG sensor 2 is subjected to non-uniform strain or temperature, the FBG's reflection spectrum may be distorted or even peaky, in which case it is not feasible to measure only the Bragg wavelength drift to obtain strain or temperature data. However, the reflection spectrum of the FBG sensor under non-uniform strain or temperature can be easily measured by the optical spectrum analyzer. The invention adopts a transmission matrix method to calculate the FBG reflection spectrum under the action of non-uniform temperature, simulates an actual spectrum, and reconstructs temperature distribution by using a wolf pack algorithm WPA.

The utility model provides an utilize inhomogeneous temperature reconstruction method of FBG of wolf crowd's algorithm, paste FBG sensor 2 and fix on aluminum alloy plate 1, carry out the mode that local is heated so that FBG sensor 2 receives an inhomogeneous temperature to aluminum alloy plate 1, show the FBG reflection spectrum waveform under the inhomogeneous temperature through spectrum appearance 3, reuse wolf crowd algorithm WPA4 to carry out inhomogeneous temperature reconstruction to FBG reflection spectrum waveform, show the result through PC 5 at last.

The aluminum alloy plate 1 is adopted, the material has good thermal conductivity and is sensitive to temperature, and plates made of other materials with good thermal conductivity can be adopted. The FBG sensor 2 is adhered and fixed on the aluminum alloy plate 1, and the FBG sensor 2 is subjected to non-uniform temperature by locally heating the aluminum alloy plate 1.

Firstly, calculating reflection spectrums under non-uniform temperature distribution by using a transmission matrix method, wherein the spectrums are used as target spectrums during reconstruction and simulated measurement spectrums; the fitness function is defined as the error between the target spectrum and the reconstructed spectrum. Then, the temperature distribution is reconstructed using the wolf pack algorithm WPA, and the average of the results of 20 runs is taken as the final result.

The wolf pack algorithm WPA comprises the following steps:

step 1: make an initialization setupThe set parameters include the number of artificial wolfs N and the initial position X_iMaximum number of walks T_maxAnd number of iterations k_maxStep size factor S, distance decision factor ω, wolf detection, update scale factors α and β;

step 2: updating the position of the wandering wolf according to the formula (1) until the fitness Y of the wolf is reached_i>Y_leadReplacing the wolf with the exploring wolf, proceeding to the next step, otherwise, continuing to wander until reaching the maximum times T_maxThen carrying out the next step;

in the formula, the wolf is visited to walk in h directions in total, and the step length of the walking isThe initial fitness of the wolf detection is Y_iDetecting wolf at current position x_idThen, the movement is performed in p (p is 1,2, …, h) directions.

And step 3: updating the position of the rushing wolf according to the formula (2) until the detection fitness Y of the wolf_i>Y_leadReplacing the wolf with the wolf, otherwise continuing to attack until the distance from the wolf is less than d_nearCarrying out the next step;

in the formula (I), the compound is shown in the specification,for the position of the wolf at the kth iteration,representing the stride length of a wolf burst.

And 4, step 4: performing containment attack on the prey according to the formula (3), and updating the wolf head;

in the formula (I), the compound is shown in the specification,the position of the prey is the position of the wolf,the step length when the wolf group attacks the prey.

And 5: updating the whole wolf group;

step 6: repeating the evolution cycle: steps 2-5 are repeated until a stopping rule is met, typically a sufficiently small fitness value or a sufficiently large number of iterations.

The transmission matrix method equally divides a non-uniform grating into M sections, each section is regarded as a uniform grating, and R is defined_iAnd S_iThe amplitudes of the forward and backward transmission modes of the i-th segment, respectively, through which the electromagnetic wave passes can be described as:

in the formula, R_iAnd S_iRespectively representing the amplitudes of the forward and backward transmission modes, R, of the i-th section of the grating_i-1And S_i-1Respectively representing the amplitudes of the forward and backward transmission modes, F, of the i-1 th section of the grating_iThe transmission matrix of the ith section of grating is expressed as follows:

in the formula:k is an alternating current coupling coefficient, sigma is a direct current self-coupling coefficient, Δ z is the length of each section, and i represents the ith section of the grating. Thus, the transmission mode of the entire grating can be described as:

in the formula, R₀And S₀The amplitudes of the forward and backward transmission modes of the initial section of the grating are respectively expressed, R and S respectively express the amplitudes of the forward and backward transmission modes of the whole grating, F is the transmission matrix of the whole grating, and the expression is as follows:

in the formula, F₁,F₂…F_MThe transmission matrix, f, respectively represented as the 1 st and 2 nd segments … Mth segment gratings₁₁,f₁₂,f₂₁,f₂₂Are elements in the transmission matrix F.

The reflection coefficient of FBG is:

when one FBG sensor is subjected to a non-uniform temperature T (z), the equivalent period Λ (z) of the FBG can be described as:

Λ(z)＝Λ₀[1+(α+ξ)T(z)]0≤z≤L (9)；

wherein L is the length of the grating, Λ₀Is the period of the FBG at normal temperature, α is the thermo-optic coefficient, ξ is the coefficient of thermal expansion, t (z) represents the non-uniform temperature experienced by the FBG sensor, and z represents a point in the axial direction of the grating.

Then k and σ in equation (5) can be described as:

in the formula: n is_effIs the mode effective index of the fiber, Δ n is the index modulation depth, s is the fringe visibility, and λ is the wavelength of the FBG.

For any given ginsengNumber L, Λ₀,n_effΔ n, s, α and t (z), the reflection spectrum of FBG at non-uniform temperature can be calculated from the above equation.

In order to test the effectiveness of the wolf pack algorithm for FBG non-uniform temperature reconstruction, reconstruction is performed by taking secondary temperature distribution as an example. The reconstruction flow diagram is shown in fig. 1, and fig. 2 shows the original and reconstructed temperature distributions. As can be seen from fig. 2, the temperature distribution obtained by using the WPA algorithm is better matched with the original temperature, which indicates that the reconstruction accuracy is higher.