RC algorithm-assisted optical fiber partial discharge detection system and high-speed signal inquiry method

Abstract

The invention is suitable for the field of optical fiber sensing technology and power equipment detection, and provides an RC algorithm-assisted optical fiber partial discharge detection system and a high-speed signal inquiry method. The system comprises a mode-locked laser, a dispersion compensation optical fiber, an erbium-doped optical fiber amplifier, a first optical fiber coupler, a single-mode optical fiber, an adjustable optical fiber delay line, a second optical fiber coupler, a sensing unit, a photoelectric detector, an oscilloscope and a PC terminal, and the sensing unit is a micro-nano optical fiber coupler and comprises an input end, an output end and a ZnO composite graphene coating area. The PC terminal inquires a signal through an RC algorithm, the RC algorithm is composed of an input layer, a storage layer and an output layer, the sensing method is that the wavelength of an ultra-short optical pulse is mapped to time, a non-equilibrium dispersion MZI structure carries out chirp frequency coding on the broadened optical pulse, the RC algorithm optimizes a data processing method, and the inquiry speed is further improved by five times. In addition, through combination of mapping from wavelength to time, chirp frequency coding and an RC algorithm, high sensing system stability and high demodulation speed are realized.

Description

Technical Field

[0001] The present invention relates to the field of optical fiber sensing technology and power equipment detection, and in particular to an optical fiber partial discharge detection system assisted by an RC algorithm and a high-speed signal inquiry method. Background Art

[0002] The application of online condition monitoring technology in power equipment mainly includes dielectric spectrum analysis, dissolved gas analysis, frequency response analysis, partial discharge detection and insulation resistance measurement. Among them, the partial discharge detection system technology based on optical fiber sensing has important development significance due to its high accuracy, short response time and anti-electromagnetic interference.

[0003] Traditional partial discharge fiber optic sensors lack precise calibration. In actual applications, the sensor performance will be affected by other environmental parameters such as temperature, electromagnetic interference level, and refractive index of the measured environment. It is difficult to achieve high-precision spectral demodulation and high-speed signal interrogation. The current demand for high-precision and ultra-fast demodulation of partial discharge sensors is difficult to meet.

[0004] The wavelength-to-time (WTT) mapping method of ultrashort optical pulses enables ultrafast real-time interrogation of wavelength-modulated optical fiber sensors. The use of an unbalanced Mach-Zehnder interferometer (MZI) in chirped microwave frequency encoding (CMFE) can eliminate the interference caused by random delays, and the data processing method is optimized using the RC algorithm to further increase the interrogation speed. Therefore, an optical fiber partial discharge detection system assisted by an RC algorithm and a high-speed signal interrogation method have broad development space in the field of sensing. Summary of the invention

[0005] In view of the defects of the prior art and the need for improvement, the present invention provides an RC algorithm-assisted optical fiber partial discharge detection system, the purpose of which is to achieve high sensing system stability and high demodulation speed by coating the surface of a high-coupling rate and high-sensitivity micro-nano optical fiber coupler with a ZnO composite graphene film, combining wavelength-to-time mapping, chirp frequency encoding and RC algorithm, so as to prepare an ultra-high precision ultra-fast optical fiber partial discharge sensing system.

[0006] Another object of the present invention is to provide a high-speed signal interrogation method for optical fiber partial discharge detection assisted by an RC algorithm.

[0007] The technical solution adopted for the first purpose of the present invention is as follows:

[0008] An RC algorithm-assisted optical fiber partial discharge detection system and a high-speed signal interrogation method, characterized in that the system is composed of a mode-locked laser (1), a dispersion-compensating optical fiber (2), an erbium-doped optical fiber amplifier (3), an optical fiber coupler No. 1 (4), a single-mode optical fiber (5), an adjustable optical fiber delay line (6), an optical fiber coupler No. 2 (7), a sensor unit (8), a photodetector (9), an oscilloscope (10), and a PC terminal (11);

[0009] The sensing unit (8) is a micro-nano optical fiber coupler, which is composed of an input end (8-1), an output end (8-2), and a ZnO composite graphene coating area (8-3); the PC terminal (11) inquires about the signal through an RC algorithm, and the RC algorithm is composed of an input layer (11-1), a storage layer (11-2), and an output layer (11-3);

[0010] In the RC algorithm-assisted optical fiber partial discharge detection system, a laser pulse provided by a mode-locked laser (1) is transmitted to an erbium-doped optical fiber amplifier (3) through a dispersion compensation optical fiber (2) for amplification. The amplified optical pulse is further divided into two branches in equal proportion by an optical fiber coupler No. 1 (4), wherein the upper branch is connected to a single-mode optical fiber (5) and the lower branch is connected to an adjustable optical fiber delay line (6). The two optical pulses are then coupled by an optical fiber coupler No. 2 (7) and transmitted to a sensing unit (8) through an input end (8-1). The sensing optical signal generated by the sensing unit (8) is converted into an electrical signal by a photoelectric detector (9) through an output end (8-2), and time domain sampling is performed by an oscilloscope (10). Finally, the signal is calculated and processed by a PC terminal (11).

[0011] The RC algorithm-assisted optical fiber partial discharge detection system is characterized in that the mode-locked laser (1) outputs a femtosecond pulse optical signal with a central wavelength of 1550nm, a 3dB spectral bandwidth of 16nm, a pulse width of 800fs, and a repetition frequency of 50MHz.

[0012] The RC algorithm-assisted optical fiber partial discharge detection system is characterized in that the optical fiber coupler No. 1 (4), the single-mode optical fiber (5), the adjustable optical fiber delay line (6), and the optical fiber coupler No. 2 (7) constitute an unbalanced dispersion MZI (Mach-Zehnder interferometer) structure.

[0013] The RC algorithm-assisted optical fiber partial discharge detection system is characterized in that: the sensing unit (8) is a micro-nano optical fiber coupler, and the total length of the waist cone region of the micro-nano optical fiber coupler is 42.38 mm, the diameter of the coupling waist region is 3.67 μm, and the high-sensitivity sensing of the refractive index of the coupling region to the external environment is achieved through the large-scale evanescent field effect; the ZnO composite graphene coating area (8-3) is a ZnO composite graphene film coated on the surface of the coupling waist region, and the wide bandgap and high exciton binding energy characteristics of ZnO have selective characteristics for the ultraviolet spectrum. By doping graphene, the sensitivity of the sensing unit (8) to the ultraviolet light band of partial discharge radiation is further improved.

[0014] The technical solution adopted for the second purpose of the present invention is as follows:

[0015] A high-speed signal interrogation method for partial discharge detection of optical fiber assisted by an RC algorithm, the method being used for the RC algorithm-assisted partial discharge detection system of optical fiber, characterized in that an ultrashort optical pulse generated by a mode-locked laser (1) passes through a dispersion compensation optical fiber (2), and its waveform in the time domain is broadened, an unbalanced dispersion MZI structure performs chirp frequency encoding on the broadened optical pulse, and then a sensor unit (8) loads the generated sensor signal onto the optical pulse, a photodetector (9) receives the carrier optical pulse and converts it into an electrical signal, which is transmitted to an oscilloscope (10) for presentation, and finally the signal is interrogated by an RC algorithm through a PC terminal (11).

[0016] The RC algorithm-assisted high-speed signal interrogation method for optical fiber partial discharge detection is characterized in that the wavelength is mapped to time through a mode-locked laser (1) and a dispersion-compensating optical fiber (2), and the wavelength interrogation speed and resolution are increased to 100 kHz and 0.2 pm respectively by using the photon time stretching method.

[0017] The invention discloses a high-speed signal interrogation method for optical fiber partial discharge detection assisted by an RC algorithm, characterized in that the chirp frequency coding is realized by an unbalanced dispersion MZI structure, specifically by adjusting the adjustable optical fiber delay line (6) so that the upper and lower branches of the MZI maintain time domain consistency at the input end of the optical fiber coupler No. 2 (7). At this time, the optical pulse is loaded with chirp frequency information due to the total dispersion difference brought by the single-mode optical fiber (5) in the MZI. Changing the chirp rate of the chirp microwave signal can be achieved by adjusting the length of the single-mode optical fiber (5). The central frequency position of the chirp signal is affected by the delay modulation of the two optical paths of the MZI. The chirp frequency coding eliminates the instantaneous microwave frequency affected by random time shift.

[0018] The RC algorithm-assisted high-speed signal interrogation method for optical fiber partial discharge detection is characterized in that: the pulse frequency interrogation method using the RC algorithm in the PC terminal (11) uses a randomly generated sparsely connected internal weight matrix to obtain a neuron library, the input signal enters from the input layer (11-1), and the signal is transmitted to the storage layer (11-2) after activation function calculation. The storage layer (11-2) constitutes a nonlinear dynamic system, whose internal states are periodically connected to each other, and the output signal is composed only of a linear combination of weighted internal variables. The storage layer (11-2) then processes the signal and transmits it to the output layer (11-3), and the feedback matrix then transmits the output signal from the output layer (11-3) to the storage layer (11-2), and finally the output layer (11-3) outputs the optimal solution. The RC algorithm increases the signal processing speed by 5 times.

[0019] In general, the RC algorithm-assisted optical fiber partial discharge detection system and high-speed signal interrogation method proposed in the present invention have the following beneficial effects compared with the prior art:

[0020] 1. The micro-nano fiber coupler with ZnO composite graphene film coated on the surface of the coupling waist region has the advantages of high sensitivity, high selectivity, and high stability, providing the system with higher ultraviolet detection sensitivity.

[0021] 2. The photon time stretching method improves the speed and resolution of wavelength demodulation technology. The unambiguous dynamic wavelength detection method based on chirped microwave frequency coding effectively solves the demodulation deviation caused by random time delay in the environment, and realizes fast and accurate wavelength interrogation in complex situations.

[0022] 3. The introduction of RC algorithm increases the frequency evaluation speed by 5 times, which has great application advantages in the field of dynamic signal demodulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The schematic diagram of an RC algorithm-assisted optical fiber partial discharge detection system according to an embodiment of the present invention.

[0024] Figure 2 4 is a structural diagram of a sensor unit according to an embodiment of the present invention.

[0025] Figure 3 4 is a structural diagram of the RC algorithm of an embodiment of the present invention.

[0026] Figure 4 The frequency response of the sensor is measured for pulses at different center wavelengths.

[0027] Figure 5 RC results for sensing pulses with center wavelengths between 1546nm and 1550nm.

[0028] In the figure: 1—mode-locked laser, 2—dispersion-compensating fiber, 3—erbium-doped fiber amplifier, 4—fiber coupler No. 1, 5—single-mode fiber, 6—adjustable fiber delay line, 7—fiber coupler No. 2, 8—sensing unit, 8-1—input end, 8-2—output end, 8-3—ZnO composite graphene coating area, 9—photodetector, 10—oscilloscope, 11—PC terminal, 11-1—input layer, 11-2—storage layer, 11-3—output layer. DETAILED DESCRIPTION

[0029] The specific implementation manner of the present invention is further described below in conjunction with the drawings of the specification, but the embodiments and protection scope of the present invention are not limited thereto, and all substantially the same implementation manners of the present invention belong to the protection scope of the present invention.

[0030] like Figure 1 As shown, the present invention provides an RC algorithm-assisted optical fiber partial discharge detection system, which includes a mode-locked laser (1), a dispersion-compensating optical fiber (2), an erbium-doped optical fiber amplifier (3), an optical fiber coupler No. 1 (4), a single-mode optical fiber (5), an adjustable optical fiber delay line (6), an optical fiber coupler No. 2 (7), a sensor unit (8), a photodetector (9), an oscilloscope (10), and a PC terminal (11).

[0031] The structure of the sensor unit (8) is as follows: Figure 2 As shown, the structure includes: an input end (8-1), an output end (8-2), and a ZnO composite graphene coating area (8-3), wherein the composite graphene coating area (8-3) is the sensing body, and the sensing unit (8) is made of two standard single-mode optical fibers with part of the coating removed and cross-fused tapers, and the total length of the waist taper area of ​​the micro-nano optical fiber coupler is 42.38 mm, and the diameter of the coupling waist area is 3.67 μm. The coupling area is highly sensitive to the refractive index of the external environment through the large-scale evanescent field effect; the ZnO composite graphene film coated on the surface of the coupling waist area, the wide bandgap and high exciton binding energy characteristics of ZnO have selective characteristics for the ultraviolet spectrum, and the sensitivity of the sensing unit (8) to the ultraviolet light band of partial discharge radiation is further improved by doping graphene. The RC algorithm structure is as follows Figure 3 As shown, it consists of an input layer (11-1), a storage layer (11-2), and an output layer (11-3).

[0032] Furthermore, the optical fiber coupler No. 1 (4), the single-mode optical fiber (5), the adjustable optical fiber delay line (6), and the optical fiber coupler No. 2 (7) are connected in sequence to form an unbalanced dispersion MZI (Mach-Zehnder interferometer) structure.

[0033] In this embodiment, the mode-locked laser (1) outputs a femtosecond pulse optical signal with a central wavelength of 1550 nm, a 3dB spectral bandwidth of 16 nm, a pulse width of 800 fs, and a repetition frequency of 50 MHz, which matches the detected spectral band.

[0034] The following is an explanation of the RC algorithm-assisted high-speed signal interrogation method for optical fiber partial discharge detection:

[0035] The ultrashort optical pulse generated by the mode-locked laser (1) passes through the dispersion compensation optical fiber (2), and its waveform in the time domain is broadened. The unbalanced dispersion MZI structure performs chirp frequency encoding on the broadened optical pulse, and then the sensor unit (8) loads the generated sensing signal onto the optical pulse. The photodetector (9) receives the carrier optical pulse and converts it into an electrical signal, which is transmitted to the oscilloscope (10) for presentation, and finally the signal is interrogated through the RC algorithm via the PC terminal (11).

[0036] Furthermore, chirp frequency encoding is implemented by an unbalanced dispersion MZI structure. Specifically, the upper and lower branches of the MZI are kept consistent in time domain at the input end of the optical fiber coupler No. 2 (7) by adjusting the adjustable optical fiber delay line (6). At this time, the optical pulse is loaded with chirp frequency information due to the total dispersion difference brought by the single-mode optical fiber (5) in the MZI. The chirp rate of the chirp microwave signal can be changed by adjusting the length of the single-mode optical fiber (5). The central frequency position of the chirp signal is affected by the delay modulation of the two optical paths of the MZI. The chirp frequency encoding eliminates the instantaneous microwave frequency affected by random time shift.

[0037] Furthermore, the RC algorithm structure is as follows Figure 3 As shown, it consists of an input layer (11-1), a storage layer (11-2), and an output layer (11-3).

[0038] Furthermore, the input signal enters from the input layer (11-1), and is transmitted to the storage layer (11-2) after the activation function calculation. The storage layer (11-2) constitutes a nonlinear dynamic system, whose internal states are periodically connected to each other, and the output signal is composed only of a linear combination of weighted internal variables. The storage layer (11-2) then processes the signal and transmits it to the output layer (11-3). The feedback matrix then transmits the output signal from the output layer (11-3) to the storage layer (11-2), and finally the output layer (11-3) outputs the optimal solution.

[0039] Working principle:

[0040] Working process: The laser pulse provided by the mode-locked laser (1) is transmitted to the erbium-doped fiber amplifier (3) through the dispersion compensation fiber (2) for amplification. The amplified optical pulse is further divided into two branches in equal proportion by the optical fiber coupler No. 1 (4). The upper branch is connected to the single-mode optical fiber (5) and the lower branch is connected to the adjustable optical fiber delay line (6). The two optical pulses are then coupled by the optical fiber coupler No. 2 (7) and transmitted to the sensor unit (8) through the input end (8-1). The sensing light signal generated by the sensor unit (8) is converted into an electrical signal by the photoelectric detector (9) through the output end (8-2). The oscilloscope (10) performs time domain sampling. Finally, the PC terminal (11) performs calculation processing on the signal.

[0041] Measurement principle: The present invention uses a photon time stretching system based on the principle of real-time Fourier transform to achieve stable demodulation of the signal. Signal chirp frequency encoding technology is also used to reduce the negative impact of additional delay on the accuracy of system signal demodulation, thereby ensuring the stability and accuracy of the detection system. The RC algorithm for pulse frequency solution is used to further improve the signal processing speed.

[0042] The micro-nano fiber coupler uses the strong evanescent field effect to make the coupling coefficient closely related to the wavelength and the ambient refractive index, thus achieving high-precision refractive index detection. According to the coupled mode theory, the micro-nano fibers in the MFC work together, and the interaction between the odd mode and the even mode affects the transmission characteristics. Assuming that the light with power Pin enters the input port, the power at the output port can be expressed as:

[0043] P out =P in sin 2 (CZ) (1)

[0044] Where Z is the coupling length of the MFC and C is the coupling coefficient of the coupling region, which can be expressed as:

[0045]

[0046] Where V is the normalized frequency, λ is the wavelength of the incident light, a is the fiber radius in the waist region, and n1 and n2 are the refractive indices of the cladding and the environment, respectively. Under the action of ultraviolet radiation, the ZnO material efficiently absorbs light energy, promotes the surge of electron-hole pairs, significantly changes the refractive index of ZnO, further causes changes in the coupling coefficient, and causes the transmission spectrum to drift, thus achieving high-sensitivity ultraviolet detection.

[0047] The RC algorithm includes an input layer (11-1), a storage layer (11-2), and an output layer (11-3), and uses a randomly generated sparsely connected internal weight matrix to obtain a neuron library to project the input into a high-dimensional nonlinear representation. The typical reservoir update equation can be described by equations (3) and (4):

[0048]

[0049] The storage layer (11-2) constitutes a nonlinear dynamic system whose internal states are periodically connected to each other, but the output consists only of a linear combination of internal variables weighted according to equation (5). This unique property makes the RC algorithm remarkably efficient in complex time-varying classification tasks.

[0050] y(n)=W out x(n) (5)

[0051] First, the frequency of the entire chirp signal is calculated using the short-time Fourier transform (STFT) method. The frequency results of the pulse signal with a central wavelength ranging from 1546nm to 1550nm (with a step size of 1nm) are calculated and evaluated using the RC algorithm. By comparison, it is found that the frequency obtained by the RC algorithm is consistent with that of the STFT method, and the linear response R 2 The value reaches 0.979. For 100,000 sampling points, the frequency evaluation time of the RC algorithm is about 0.28 seconds, while under the same calculation conditions, the average time of the STFT method is about 1.42 seconds. Therefore, RC shows higher efficiency and the frequency evaluation speed is increased by 5 times.

Claims

1. An RC algorithm-assisted optical fiber partial discharge detection system and a high-speed signal interrogation method, characterized in that: The system is composed of a mode-locked laser (1), a dispersion-compensating optical fiber (2), an erbium-doped optical fiber amplifier (3), an optical fiber coupler No. 1 (4), a single-mode optical fiber (5), an adjustable optical fiber delay line (6), an optical fiber coupler No. 2 (7), a sensor unit (8), a photodetector (9), an oscilloscope (10), and a PC terminal (11); The sensing unit (8) is a micro-nano optical fiber coupler, which is composed of an input end (8-1), an output end (8-2), and a ZnO composite graphene coating area (8-3); the PC terminal (11) inquires about the signal through an RC algorithm, and the RC algorithm is composed of an input layer (11-1), a storage layer (11-2), and an output layer (11-3); In the RC algorithm-assisted optical fiber partial discharge detection system, a laser pulse provided by a mode-locked laser (1) is transmitted to an erbium-doped optical fiber amplifier (3) through a dispersion compensation optical fiber (2) for amplification. The amplified optical pulse is further divided into two branches in equal proportion by an optical fiber coupler No. 1 (4), wherein the upper branch is connected to a single-mode optical fiber (5) and the lower branch is connected to an adjustable optical fiber delay line (6). The two optical pulses are then coupled by an optical fiber coupler No. 2 (7) and transmitted to a sensing unit (8) through an input end (8-1). The sensing optical signal generated by the sensing unit (8) is converted into an electrical signal by a photoelectric detector (9) through an output end (8-2), and time domain sampling is performed by an oscilloscope (10). Finally, the signal is calculated and processed by a PC terminal (11).

2. The RC algorithm-assisted optical fiber partial discharge detection system according to claim 1, characterized in that: The mode-locked laser (1) outputs a femtosecond pulse light signal with a central wavelength of 1550nm, a 3dB spectrum bandwidth of 16nm, a pulse width of 800fs, and a repetition frequency of 50MHz.

3. The RC algorithm-assisted optical fiber partial discharge detection system according to claim 1, characterized in that: The optical fiber coupler No. 1 (4), the single-mode optical fiber (5), the adjustable optical fiber delay line (6), and the optical fiber coupler No. 2 (7) constitute an unbalanced dispersion MZI (Mach-Zehnder interferometer) structure.

4. The RC algorithm-assisted optical fiber partial discharge detection system according to claim 1, characterized in that: The sensing unit (8) is a micro-nano optical fiber coupler, and the total length of the waist cone region of the micro-nano optical fiber coupler is 42.38 mm, and the diameter of the coupling waist region is 3.67 μm. The coupling region realizes high-sensitivity sensing of the refractive index of the external environment through the large-scale evanescent field effect; the ZnO composite graphene coating region (8-3) is a ZnO composite graphene film coated on the surface of the coupling waist region. The wide bandgap and high exciton binding energy characteristics of ZnO have selective characteristics for the ultraviolet spectrum. By doping graphene, the sensitivity of the sensing unit (8) to the ultraviolet light band of partial discharge radiation is further improved.

5. A high-speed signal interrogation method for optical fiber partial discharge detection assisted by an RC algorithm, the method being used in an optical fiber partial discharge detection system assisted by an RC algorithm according to claim 1, characterized in that: The ultrashort optical pulse generated by the mode-locked laser (1) passes through the dispersion compensation optical fiber (2), and its waveform in the time domain is broadened. The unbalanced dispersion MZI structure performs chirp frequency encoding on the broadened optical pulse, and then the sensor unit (8) loads the generated sensing signal onto the optical pulse. The photodetector (9) receives the carrier optical pulse and converts it into an electrical signal, which is transmitted to the oscilloscope (10) for presentation, and finally the signal is interrogated through the RC algorithm via the PC terminal (11).

6. The RC algorithm-assisted high-speed signal interrogation method for optical fiber partial discharge detection according to claim 5, characterized in that: The wavelength is mapped to time through a mode-locked laser (1) and a dispersion-compensating optical fiber (2), and the wavelength interrogation speed and resolution are increased to 100 kHz and 0.2 pm respectively by using this photon time stretching method.

7. The RC algorithm-assisted high-speed signal interrogation method for optical fiber partial discharge detection according to claim 5, characterized in that: The chirp frequency encoding is realized by an unbalanced dispersion MZI structure. Specifically, the upper and lower branches of the MZI are kept consistent in the time domain at the input end of the optical fiber coupler No. 2 (7) by adjusting the adjustable optical fiber delay line (6). At this time, the optical pulse is loaded with chirp frequency information due to the total dispersion difference brought by the single-mode optical fiber (5) in the MZI. The chirp rate of the chirp microwave signal can be changed by adjusting the length of the single-mode optical fiber (5). The central frequency position of the chirp signal is affected by the delay modulation of the two optical paths of the MZI. The chirp frequency encoding eliminates the instantaneous microwave frequency affected by random time shift.

8. The RC algorithm-assisted high-speed signal interrogation method for optical fiber partial discharge detection according to claim 5, characterized in that: In the PC terminal (11), the pulse frequency query method using the RC algorithm first uses a randomly generated internal weight matrix to obtain a neuron library, an input signal enters from the input layer (11-1), and after the activation function is calculated, the signal is transmitted to the storage layer (11-2), the storage layer (11-2) constitutes a nonlinear dynamic system, and its internal states are periodically connected to each other, the storage layer (11-2) processes the signal and transmits it to the output layer (11-3), the feedback matrix in the output layer (11-3) transmits the signal back to the storage layer (11-2), and finally the optimal solution output by the output layer (11-3) is composed of only a linear combination of weighted internal variables, and the signal processing speed is increased by 5 times through the RC algorithm.

Images