Optical fiber ultrasonic sensor demodulation system based on LMM-EF

Abstract

The invention relates to the technical field of optical fiber ultrasonic sensing, and discloses an optical fiber ultrasonic sensor demodulation system based on LMM-EF. The demodulation system jointly uses a broadband light source and a dense wavelength division multiplexer to realize wavelength division multiplexing of two paths of narrowband optical signals, optimizes an ellipse fitting process through a Levenberg-Marquardt method, introduces an N-order unit matrix I and a damping coefficient alpha, updates the value of alpha according to the approximation degree of a Jacobian matrix and an elliptic equation, and finally achieves the purpose of narrowband optical signal demodulation. And the stability of the second derivative is improved to greatly reduce the occurrence of ill-conditioned problems in the fitting process, so that the fitting iteration mode is between the Newton method and the gradient descent method. Ellipse fitting of Lissajous figures of two interference signals is realized by minimizing the sum of squares of errors between data points and functions, the fitting precision of an ellipse fitting algorithm to short arc signals is improved, and the problem that a traditional dual-wavelength demodulation system based on the ellipse fitting algorithm is not suitable for weak sound signal solution is solved. And the detection range of the optical fiber F-P ultrasonic sensing technology is expanded.

Description

Technical Field

[0001] The invention relates to the technical field of optical fiber ultrasonic sensing, and in particular to an optical fiber ultrasonic sensor demodulation system based on LMM-EF. Background Art

[0002] The diaphragm-type fiber Fabry-Perot (FP) ultrasonic sensor system has high sensitivity and simple structure. It is a hot topic in the current research of partial discharge ultrasonic signal detection. It includes two parts: fiber FP ultrasonic sensor and demodulation system. When the external ultrasonic signal acts on the sensing diaphragm, it will cause physical quantities such as the FP cavity length or the refractive index of the medium in the cavity to change, further causing the reflected light intensity and reflected spectrum of the interference light signal to change. The relationship between these changes and the measured signal can be found through the demodulation algorithm to achieve accurate measurement of the ultrasonic signal. Different demodulation systems focus on different information and need to be selected according to the actual application. At present, the commonly used demodulation systems are mainly divided into intensity demodulation and phase demodulation. Among them, the dual-wavelength demodulation technology based on ellipse fitting (EF) and differential cross multiplication (DCM) algorithm is simple and has fast data processing speed, and is widely used in the field of fiber FP sensor demodulation. However, since ultrasonic signals attenuate quickly in the air, the acoustic signals acting on the air-coupled sensor are usually small, and the Lissajous figures of the two-way interference signals are often short arc signals rather than closed ellipses. The traditional EF algorithm based on the least squares fitting method (LSM) has low fitting accuracy for short arc signals.

[0003] Therefore, it is necessary to provide a fiber optic ultrasonic sensor demodulation system based on LMM-EF to achieve high-precision demodulation of weak ultrasonic signals. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a fiber optic ultrasonic sensor demodulation system based on LMM-EF.

[0005] The LMM-EF-based fiber optic ultrasonic sensor demodulation system provided by the present invention comprises a broadband light source, a fiber optic circulator, a fiber optic FP ultrasonic sensor, a dense wavelength division multiplexer, a photoelectric detector, a photoelectric detector, an ADC data acquisition card and a signal processing unit.

[0006] Preferably, the wavelength range of the flat broadband optical signal emitted by the broadband light source is 1535nm-1565nm, and the average power is about 30mW.

[0007] Preferably, the fiber optic circulator has three ports: port 1, port 2 and port 3. The continuous broadband optical signal emitted by the broadband light source enters the fiber optic circulator through port 1, is injected into the optical fiber FP ultrasonic sensor through port 2, and interferes in the FP cavity. The interference optical signal returns to the fiber optic circulator from port 2 and enters the dense wavelength division multiplexer through port 3.

[0008] Preferably, the dense wavelength division multiplexer can separate broadband optical signals into two narrowband optical signals of different wavelengths, the isolation between channels is greater than 40dB, the channel bandwidth does not exceed 0.4nm, and the wavelengths of optical signals output by channel 1 and channel 2 are 1549.32nm and 1557.36nm respectively.

[0009] Preferably, when an acoustic signal with a relatively low sound pressure acts on the optical fiber FP ultrasonic sensor, the Lissajous figure of the two-path interference signal returned is a short arc signal and cannot form a complete ellipse.

[0010] Preferably, the photodetector and the photodetector have both high bandwidth and low noise characteristics, and the conversion gain is adjustable, the 3dB bandwidth can reach up to 510kHz, and the noise equivalent power is 17.6pW / Hz 1 / 2 The detectable light wavelength range is 800~1800nm.

[0011] Preferably, the ADC data acquisition card is used to convert the photodetector and the analog electrical signal output by the photodetector into a digital electrical signal. The ADC data acquisition card processes data in dual channels in parallel, uses USB2.0 communication, has a quantization bit number of 16 bits, and an adjustable sampling rate of each channel up to 1MHz, with an input range of ±10V.

[0012] Preferably, the system optimizes the ellipse fitting process by the Levenberg-Marquardt method to solve the problem of large parameter fitting errors of the traditional ellipse fitting algorithm for short arc signals.

[0013] Compared with the related art, the optical fiber ultrasonic sensor demodulation system based on LMM-EF provided by the present invention has the following beneficial effects:

[0014] The present invention provides a fiber optic ultrasonic sensor demodulation system based on LMM-EF:

[0015] 1. The present invention uses a broadband light source and a dense wavelength division multiplexer to realize wavelength division multiplexing of two narrowband optical signals, avoiding the use of two narrowband light sources and couplers, and greatly reducing the hardware complexity of the dual-wavelength demodulation system;

[0016] 2. The two narrow-band interference optical signals in the present invention are wavelength-division multiplexed, and the two signals are processed synchronously during the photoelectric conversion and analog-to-digital conversion process, thereby overcoming the defect that the time-division multiplexing system cannot realize high-frequency ultrasonic signal demodulation due to the influence of time slot division;

[0017] 3. The present invention corrects the problem of large fitting error of short arc signal by traditional LSM-EF algorithm through LMM-EF algorithm, reduces the error between the ellipse parameters obtained by fitting and the actual parameters, and further improves the demodulation accuracy of the dual-wavelength demodulation system for tiny sound signals;

[0018] 4. The present invention overcomes the limitation that the traditional multi-wavelength demodulation system requires the phase difference of each interference signal to remain orthogonal or anti-phase, and realizes high-precision demodulation of optical fiber FP ultrasonic sensors with various cavity lengths without changing the hardware equipment of the demodulation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is the overall flow chart of the demodulation algorithm;

[0020] Figure 2 It is the flow chart of LMM-EF algorithm;

[0021] Figure 3 To simulate the comparison of fitting errors of LMM-EF algorithm and LSM-EF algorithm under different arc length conditions;

[0022] Figure 4 It is a schematic diagram of the dual-wavelength demodulation system of the optical fiber FP ultrasonic sensor;

[0023] Figure 5 The reflection spectra of three fiber FP ultrasonic sensors with different initial cavity lengths;

[0024] Figure 6 This is the demodulation result diagram of the optical fiber FP ultrasonic sensor with a cavity length of 143.92μm under the action of a 40kHz ultrasonic signal;

[0025] Figure 7 This is the demodulation result diagram of the optical fiber FP ultrasonic sensor with a cavity length of 194.43μm under the action of a 40kHz ultrasonic signal;

[0026] Figure 8 This is the demodulation result diagram of the optical fiber FP ultrasonic sensor with a cavity length of 239.37μm under the action of a 40kHz ultrasonic signal;

[0027] Fig. 9 The Lissajous figure formed by two interference signals under different sound pressures and the time domain diagram of the demodulation results.

[0028] Numbers in the figure: 1. Broadband light source; 2. Fiber optic circulator; 3. Fiber optic FP ultrasonic sensor; 4. Dense wavelength division multiplexer; 5. Photodetector; 6. Photodetector; 7. ADC data acquisition card; 8. Signal processing unit. DETAILED DESCRIPTION

[0029] In order to better understand the technical content of the present invention, specific embodiments are provided below, and the present invention is further described in conjunction with the accompanying drawings.

[0030] Figure 1 For the demodulation algorithm flow chart, the reference sound pressure is introduced before the application of the acoustic sensing system to construct the system Lissajous figure to determine the ellipse parameters and introduce them into the subsequent signal processing as prior information.

[0031] Figure 2 This is the flow chart of the LMM-EF algorithm. By introducing the N-order unit matrix I and the damping coefficient α, the stability of the second-order derivative is improved to greatly reduce the occurrence of ill-conditioned problems in the signal processing process. The ellipse parameters are fitted, and the DC component and AC amplitude of the two interference signals can be obtained through the relationship between the ellipse parameters and the interference fringe parameters.

[0032] Figure 3 In order to simulate the comparison of fitting errors of LMM-EF algorithm and LSM-EF algorithm under different arc length conditions, the fitting errors of the two algorithms increase with the decrease of the amplitude (C) of the measured signal. When the C value is greater than 2rad, the fitting results of LMM-EF and LSM-EF are similar and the fitting error is small. When the C value is less than 2rad, the fitting accuracy of LSM-EF algorithm decreases greatly with the decrease of C value, while the fitting result of LMM-EF algorithm is relatively stable, and the fitting error is greatly reduced compared with LSM-EF.

[0033] Figure 4 In the schematic diagram of the dual-wavelength demodulation system of the optical fiber FP ultrasonic sensor shown: the demodulation system includes a broadband light source 1, an optical fiber circulator 2, an optical fiber FP ultrasonic sensor 3, a dense wavelength division multiplexer 4, a photodetector 5, a photodetector 6, an ADC data acquisition card 7 and a signal processing unit 8.

[0034] The broadband light signal emitted by the broadband light source 1 has an average power of 30 mW and a wavelength range of 1535 nm to 1565 nm.

[0035] The optical fiber FP ultrasonic sensor 3 is a diaphragm-type optical fiber Fabry-Perot (FP) interference sensor. When the external sound signal acts on the sensing diaphragm of the optical fiber FP ultrasonic sensor 3, the sound-light conversion is completed. The FP cavity length is controlled by adjusting the distance from the end face of the single-mode optical fiber to the sensing diaphragm, which is used for performance verification of the demodulation system.

[0036] The dense wavelength division multiplexer 4 filters the broadband interference optical signal returned by the optical fiber FP ultrasonic sensor 3 into two narrowband optical signals with central wavelengths of 1549.32 nm and 1557.36 nm respectively.

[0037] Figure 5Figure 2 shows the reflection spectra of three fiber FP ultrasonic sensors with different initial cavity lengths. The FP cavity lengths are 143.92μm, 194.43μm and 239.37μm, respectively. The interference fringe contrasts of the three sensors are different. Among them, the sensor with an initial cavity length of 143.92μm has the largest interference fringe contrast, which is about 36.97dB.

[0038] Figure 6 Figure 2 is the demodulation result of the fiber FP ultrasonic sensor with a cavity length of 143.92 μm under the action of a 40 kHz ultrasonic signal, where (a) is the time domain waveform of the two interference signals; (b) is the Lissajous figure synthesized by the two interference signals; (c) is the time domain waveform of the demodulated phase signal to be measured; (d) is the frequency domain waveform of the demodulated signal. Figure 3 (b) It can be seen that the phases of the two interference signals are not orthogonal.

[0039] Figure 7 Figure 2 is the demodulation result of the fiber FP ultrasonic sensor with a cavity length of 194.43 μm under the action of a 40 kHz ultrasonic signal, where (a) is the time domain waveform of the two interference signals; (b) is the Lissajous figure synthesized by the two interference signals; (c) is the time domain waveform of the demodulated phase signal to be measured; (d) is the frequency domain waveform of the demodulated signal. Figure 7 (b) It can be seen that the phase difference between the two interference signals is 0.59π, which is approximately orthogonal, and the eccentricity of the fitted elliptical signal is small.

[0040] Figure 8 Figure 2 is the demodulation result of the fiber FP ultrasonic sensor with a cavity length of 239.37 μm under the action of a 40 kHz ultrasonic signal, where (a) is the time domain waveform of the two interference signals; (b) is the Lissajous figure synthesized by the two interference signals; (c) is the time domain waveform of the demodulated phase signal to be measured; (d) is the frequency domain waveform of the demodulated signal. Figure 7 (b) It can be seen that the phase difference between the two signals is 1.19π, the phase deviates from orthogonality, and the eccentricity of the elliptical signal obtained by fitting is large.

[0041] Fig. 9 The Lissajous figure composed of two interference signals under different sound pressures and the time domain diagram of the demodulation results show that the tiny changes in cavity length caused by weak sound signals are also restored with high fidelity by the sensing system.

[0042] The demodulation principle of the improved algorithm is:

[0043] 1) After the light intensity signals corresponding to different wavelengths are converted into photoelectric signals, the output electrical signals can be expressed as:

[0044]

[0045] In the formula, A1 and A2 are the DC components of the two interference signals, and B1 and B2 are the AC amplitudes of the two interference signals. A1, A2, B1, B2 are different, λ1 and λ2 are the central wavelengths of the two narrowband optical signals, L is the initial cavity length, ΔL is the change in cavity length under the action of sound pressure, and n is the refractive index of the medium in the FP cavity, which is generally a constant of 1 for air.

[0046] 2) The general formula of the ellipse equation can be expressed as:

[0047] x 2 +Exy+Fy 2 +Gx+Hy+M=0

[0048] 3) The LMM algorithm formula can be expressed as:

[0049]

[0050] Where x = (E, F, G, H, M) T , F (k) =(F1(x (k) ),F2(x (k) ),……,F N (x (k) )) T , F is the general formula of the elliptic equation, k is the current number of iterations, A k is the Jacobian matrix of F.

[0051] The implementation process of the EF algorithm based on LMM is as follows:

[0052] (1) Select the initial value: initial guess x (1) , the maximum number of iterations k max , damping coefficient α and growth factor γ (γ>1);

[0053] (2) Calculate the Jacobian matrix based on the measured data points and construct the incremental equation;

[0054] (3) Solve the incremental equation to obtain x (k+1) ;

[0055] (4) For k and k max If k ≥ k max , then output x (k+1) , the algorithm ends. Otherwise, set k = k + 1 and go to step (5)

[0056] (5) Calculate F(x (k+1) ), and with F(x (k) )Compare;

[0057] (6) Update the damping coefficient. If F(x (k+1))>F(x (k) ), set α=γ·α, otherwise, α=α / γ.

[0058] After resetting α, return to step (3).

[0059] The corresponding relationship between the interference fringe parameters A1, A2, B1, B2 and the phase difference Δδ and the ellipse fitting parameters is:

[0060]

[0061] By using the DC component and fringe contrast obtained based on LMM-EF as prior information in subsequent signal processing and combining it with the DCM algorithm, high-precision demodulation of tiny sound signals by the optical fiber FP ultrasonic sensor can be achieved.

[0062] 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 principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A fiber optic ultrasonic sensor demodulation system based on LMM-EF, characterized in that: The demodulation system comprises a broadband light source (1), an optical fiber circulator (2), an optical fiber FP ultrasonic sensor (3), a dense wavelength division multiplexer (4), a photoelectric detector (5), a photoelectric detector (6), an ADC data acquisition card (7) and a signal processing unit (8).

2. The optical fiber ultrasonic sensor demodulation system based on LMM-EF according to claim 1, characterized in that: The broadband light source (1) emits a broadband optical signal with a power of 30 mW and a wavelength range of 1535 nm to 1565 nm.

3. The optical fiber ultrasonic sensor demodulation system based on LMM-EF according to claim 1, characterized in that: The optical fiber circulator (2) has three ports in total, wherein one port is used for a continuous broadband optical signal emitted by a broadband light source (1) to enter the optical fiber circulator (2) through this port, one port is used for injecting an optical fiber FP ultrasonic sensor (3), and another port is used for an interference optical signal to enter a dense wavelength division multiplexer (4) through this port.

4. The LMM-EF-based fiber optic ultrasonic sensor demodulation system according to claim 1, characterized in that: The dense wavelength division multiplexer (4) filters the broadband interference light signal returned by the optical fiber FP ultrasonic sensor (3) into two narrowband interference light signals, the isolation between the channels is greater than 40 dB, the channel bandwidth does not exceed 0.4 nm, and the central wavelengths of the narrowband light signals are 1549.32 nm and 1557.36 nm respectively.

5. The LMM-EF-based fiber optic ultrasonic sensor demodulation system according to claim 1, characterized in that: When a sound signal with a relatively low sound pressure acts on the optical fiber FP ultrasonic sensor (3), the two-path interference signal Lissajous figure shown in the return is a short arc signal.

6. The LMM-EF-based fiber optic ultrasonic sensor demodulation system according to claim 1, characterized in that: The photodetector (5) and the photodetector (6) have both high bandwidth and low noise characteristics, and their conversion gain is adjustable, the 3dB bandwidth can reach up to 510kHz, and the noise equivalent power is 17.6pW / Hz. 1 / 2 The detectable light wavelength range is 800~1800nm.

7. The LMM-EF-based fiber optic ultrasonic sensor demodulation system according to claim 1, characterized in that: The ADC data acquisition card (7) is used to convert the analog electrical signals output by the photodetector (5) and the photodetector (6) into digital electrical signals. The ADC data acquisition card (7) processes data in parallel via two channels, adopts USB2.0 communication, has a quantization bit number of 16 bits, and has an adjustable sampling rate of each channel up to 1 MHz. The input range is ±10V.

8. The LMM-EF-based fiber optic ultrasonic sensor demodulation system according to claim 1, characterized in that: The system uses the Levenberg-Marquardt method to optimize the ellipse fitting process.

Images