Optical fiber sound wave sensing on-line monitoring system for internal state of petrochemical reactor

Abstract

The invention discloses an optical fiber sound wave sensing on-line monitoring system for the internal state of a petrochemical reactor. The system comprises a sensing optical fiber module, a demodulation module, a sound wave signal reconstruction module, a feature selection extraction module and a monitoring classification module. The sensing optical fiber module is used for collecting optical signals carrying sound wave information in the gas-liquid transportation process and sending the optical signals to the demodulation module. The demodulation module is used for converting the optical signal into an electric signal and demodulating a phase signal which is in a linear relation with the sound wave signal; the sound wave signal reconstruction module is used for decomposing the phase signal into a plurality of intrinsic mode components, and selecting components with high correlation for superposition to obtain a de-noised and reconstructed sound wave signal; the feature selection and extraction module is used for selecting features with discriminability from the denoised and reconstructed sound wave signals to form a feature matrix; and the monitoring classification module monitors the internal state of the petrochemical reactor based on the feature matrix. According to the invention, real-time monitoring of the internal state of the petrochemical reactor is realized, and an accurate basis is provided for maintenance of the petrochemical reactor.

Description

Technical Field

[0001] The invention relates to an optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor. Background Art

[0002] In the process of oil refining, the internal temperature and pressure of the reactor, which is a key equipment, directly determine the catalytic efficiency of the catalyst, thereby affecting the efficiency and quality of oil output. During the production and operation cycle, the gas and liquid inside the petrochemical reactor are prone to problems such as overtemperature and overpressure, which poses a major safety hazard. At present, domestic and foreign petrochemical companies mainly use various mechanical instruments and electrical sensor systems to monitor the internal status of petrochemical reactors. The monitoring method of mechanical instruments mainly relies on manual inspection reading, but manual inspection has the disadvantages of being time-consuming and labor-intensive, with a small monitoring range and easy to be negligent. The monitoring method of the electrical sensor system has natural defects in explosion-proof and anti-electromagnetic interference. When this type of sensor system is subject to electromagnetic interference, it will cause the system signal-to-noise ratio to decrease, which reduces the accuracy of monitoring classification. Therefore, this type of sensor system has great limitations in the use of the petrochemical industry. Summary of the invention

[0003] The purpose of the present invention is to provide an optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, so as to solve the technical problems of realizing monitoring of the internal state of a petrochemical reactor, providing an accurate basis for the maintenance of the petrochemical reactor, and ensuring the personal safety of workers and equipment safety.

[0004] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical solution:

[0005] An optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, comprising a sensing optical fiber, a demodulation module, an acoustic wave signal reconstruction module, a feature selection and extraction module and a monitoring and classification module;

[0006] The sensing optical fiber is used to collect the optical signal carrying the acoustic wave information during the gas-liquid transportation process and send it to the demodulation module;

[0007] The demodulation module is used to convert the optical signal into an electrical signal and demodulate a phase signal that is linearly related to the acoustic wave signal;

[0008] The acoustic wave signal reconstruction module is used to decompose the phase signal into a plurality of eigenmode components, and calculate the correlation coefficients between the optical signal and the plurality of eigenmode components. The components with large correlation coefficients have high correlation, and the components with high correlation are selected for superposition to obtain a denoised and reconstructed acoustic wave signal;

[0009] The feature selection and extraction module is used to select discriminative features from the denoised and reconstructed sound wave signal to form a feature matrix; by calculating the KL distance between the original signal and the various features of the denoised and reconstructed sound wave signal, that is, measuring the difference between two probability distributions in the same event space, the larger the KL distance, the greater the feature difference, that is, the discriminative feature.

[0010] The monitoring classification module monitors the internal state of the petrochemical reactor based on the characteristic matrix.

[0011] According to the above scheme, in the on-line monitoring system of the internal state of a petrochemical reactor using optical fiber acoustic wave sensing, the sensing optical fiber is an acoustic wave sensing optical fiber encapsulated in an optical cable and engraved with an identical weak grating array.

[0012] According to the above scheme, in the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the acoustic wave sensing optical fiber is spirally wound around the outer walls of all pipelines along the pipeline axis to achieve distributed online monitoring.

[0013] According to the above scheme, the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the demodulation module is based on optical time domain reflection technology The distributed acoustic sensor demodulation unit (DAS) includes a light source, an acousto-optic modulator (AOM), an erbium-doped fiber amplifier (EDFA), a circulator, a Michelson interferometer, a Faraday rotator, a 3×3 coupler, a photodetector (PD) and a high-speed acquisition and processing module.

[0014] According to the above scheme, in the on-line monitoring system of the internal state of a petrochemical reactor using optical fiber acoustic wave sensors, the optical signal emitted by the light source is modulated by an acousto-optic modulator (AOM) and amplified by an erbium-doped fiber amplifier (EDFA) to form a pulsed light that enters the sensing optical fiber.

[0015] According to the above scheme, in the above-mentioned on-line monitoring system of the internal state of a petrochemical reactor by optical fiber acoustic wave sensing, the light reflected from two adjacent gratings in the above-mentioned sensing optical fiber weak grating array passes through the two arms of the Michelson interferometer and is reflected by the Faraday rotation mirror to the 3×3 coupler to cause interference. The spacing between the two arms of the Michelson interferometer is the same as the spacing between two adjacent gratings in the weak grating array.

[0016] According to the above scheme, in the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the 3×3 coupler outputs 3 optical signals with a phase difference of 2π / 3, which are converted into electrical signals by 3 photodetectors (PDs).

[0017] According to the above scheme, in the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the high-speed acquisition and processing module acquires three signals and uses an inverse tangent algorithm to calculate a phase signal that is linearly related to the acoustic wave signal.

[0018] According to the above scheme, in the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the acoustic wave signal reconstruction module includes a signal decomposition module and a signal superposition module.

[0019] According to the above scheme, in the on-line monitoring system of the internal state of a petrochemical reactor using optical fiber acoustic wave sensing, the signal decomposition module preprocesses the phase signal, including pre-emphasis, framing and windowing. The preprocessed signal is decomposed into a number of intrinsic mode components by Complete EEMD with Adaptive Noise.

[0020] According to the above scheme, in the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the signal superposition module uses a correlation coefficient to analyze the correlation between the intrinsic mode components and the preprocessed signals, and selects the intrinsic mode components with high correlation for superposition, thereby removing high-frequency noise and obtaining a denoised and reconstructed acoustic wave signal.

[0021] According to the above scheme, in the optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, the feature selection and extraction module uses the Kullback-Leibler (KL) distance method to exclude irrelevant or redundant features from the time domain features, frequency domain features and spectral density index of the denoised and reconstructed acoustic wave signal, selects discriminative features, and combines them with Mel-frequency cepstral coefficients (MFCC) to form a feature matrix.

[0022] According to the above scheme, in the on-line monitoring system of the internal state of a petrochemical reactor using optical fiber acoustic wave sensing, the monitoring and classification module uses a random forest classifier in machine learning to monitor and classify pipeline leakage based on a normalized feature matrix.

[0023] The present invention has the following advantages and positive effects:

[0024] 1. The present invention adopts an acoustic wave sensing optical fiber installed in an optical cable and engraved with a fully identical weak grating array as a sensor, which has the characteristics of anti-electromagnetic interference, corrosion resistance, high temperature resistance, and intrinsic explosion-proof, and realizes the collection of acoustic wave signals of the internal state of a petrochemical reactor in a harsh environment.

[0025] 2. The present invention adopts optical time domain reflection technology The distributed acoustic wave sensor demodulation unit demodulates the acoustic wave signal, and the demodulation unit has high positioning accuracy and fast response speed.

[0026] 3. The present invention adopts Complete EEMD with Adaptive Noise and correlation coefficient analysis method to reconstruct the sound wave signal, removes high-frequency noise, and improves the signal-to-noise ratio of the system.

[0027] 4. The present invention uses the Kullback-Leibler (KL) distance method to exclude irrelevant or redundant acoustic signal features and select discriminative acoustic signal features, so that the random forest classifier can more easily identify acoustic signal features, thereby improving the accuracy of system classification.

[0028] 5. The present invention successfully solves the problems of excessive pressure or temperature in the existing petrochemical reactor internal state monitoring method, realizes real-time monitoring of the internal state of the petrochemical reactor, provides an accurate basis for the maintenance of the petrochemical reactor, and avoids major safety accidents caused by overload of the petrochemical reactor. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic diagram of an optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor proposed in an embodiment of the present invention;

[0030] Figure 2 This is a method based on optical time domain reflection technology proposed in the embodiment of the present invention. Schematic diagram of a distributed acoustic sensor demodulation unit (DAS);

[0031] Figure 3 A schematic diagram of a sound wave signal reconstruction process based on Complete EEMD with Adaptive Noise proposed in an embodiment of the present invention;

[0032] Figure 4 The figure is a schematic diagram of a feature selection and extraction process based on the Kullback-Leibler (KL) distance method proposed in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0034] Figure 1 The present invention provides an on-line monitoring system for the internal state of a petrochemical reactor using an optical fiber acoustic wave sensor. Figure 1 As shown, it specifically includes the following modules: sensing optical fiber, demodulation module, acoustic signal reconstruction module, feature selection and extraction module and monitoring classification module.

[0035] In the embodiment of the present invention, the sensing optical fiber is used to collect the optical signal of the acoustic wave information inside the petrochemical reactor, as follows:

[0036] The sensing fiber uses an acoustic sensing fiber engraved with an identical weak grating array, which can greatly increase the number of multiplexed sensors in the system and realize real-time monitoring of the internal state of the petrochemical reactor.

[0037] In the embodiment of the present invention, the demodulation module is used to convert the optical signal into an electrical signal and demodulate a phase signal that is linearly related to the acoustic wave signal, as follows:

[0038] like Figure 2 As shown, the light source in the demodulation module emits an optical signal, which is modulated into an optical pulse signal by an acousto-optic modulator (AOM), amplified by an erbium-doped fiber amplifier (EDFA), and then enters the sensing fiber through a circulator 1. The light reflected back by two adjacent gratings in the sensing fiber weak grating array is amplified by an erbium-doped fiber amplifier (EDFA), and then connected to a bandpass filter to filter out the noise of the erbium-doped fiber amplifier (EDFA). It is reflected to the 3×3 coupler through the two arms of the Michelson interferometer and the Faraday rotator to interfere, where the spacing between the two arms of the Michelson interferometer is the same as the spacing between the two adjacent gratings in the weak grating array. The 3×3 coupler outputs three optical signals with a phase difference of 2π / 3, which are converted into electrical signals by three photodetectors (PD). The high-speed acquisition and processing module collects the three electrical signals and uses the inverse tangent algorithm to calculate the phase signal that is linearly related to the acoustic wave signal.

[0039] In the embodiment of the present invention, the acoustic wave signal reconstruction module is used to decompose the phase signal into a plurality of intrinsic mode components, select components with high correlation for superposition, and obtain a denoised and reconstructed acoustic wave signal, as follows:

[0040] The acoustic signal reconstruction module is divided into a signal decomposition module and a signal superposition module. Figure 3 As shown in the figure, the signal decomposition module first needs to pre-emphasize, frame and window the phase signal transmitted by the demodulation module. The preprocessed signal is decomposed into several intrinsic mode components through Complete EEMD with Adaptive Noise. The signal superposition module analyzes the correlation between several intrinsic mode components and the preprocessed signal through the correlation coefficient, selects the intrinsic mode components with high correlation for superposition, thereby removing high-frequency noise and obtaining a denoised and reconstructed sound wave signal.

[0041] In the embodiment of the present invention, the feature selection and extraction module is used to select discriminative features from the denoised and reconstructed sound wave signal to form a feature matrix, which is as follows:

[0042] like Figure 4As shown in the figure, the Kullback-Leibler (KL) distance measures the difference between one probability distribution function and another probability distribution function, which can reduce the feature dimension. A Kullback-Leibler (KL) distance d of 0 indicates that the distributions of the two features are the same, or that the two features are completely redundant. The larger the distance, the stronger the discriminative power of the feature. The Kullback-Leibler (KL) distance method is used to exclude irrelevant or redundant features from the time domain features, frequency domain features, and spectral density index of the denoised and reconstructed acoustic wave signal, select discriminative features, and combine them with the Mel Frequency Cepstral Coefficient (MFCC) to form a feature matrix, thereby improving the recognition accuracy of the classification.

[0043] In the embodiment of the present invention, the monitoring classification module uses a random forest classifier in machine learning to monitor the internal state of the petrochemical reactor based on a normalized feature matrix, and displays the monitoring results on a host computer in real time.

[0044] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. An optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor, comprising a sensing optical fiber, a demodulation module, an acoustic wave signal reconstruction module, a feature selection and extraction module and a monitoring and classification module, characterized in that: The sensing optical fiber is used to collect the optical signal carrying the acoustic wave information during the gas-liquid transportation process and send it to the demodulation module; The demodulation module is used to convert the optical signal into an electrical signal and demodulate a phase signal that is linearly related to the acoustic wave signal; The acoustic wave signal reconstruction module is used to decompose the phase signal into a plurality of eigenmode components, and calculate the correlation coefficients between the optical signal and the plurality of eigenmode components. The components with large correlation coefficients have high correlation, and the components with high correlation are selected for superposition to obtain a denoised and reconstructed acoustic wave signal; The feature selection and extraction module is used to select discriminative features from the denoised and reconstructed sound wave signal to form a feature matrix; by calculating the KL distance between the original signal and the various features of the denoised and reconstructed sound wave signal, that is, measuring the difference between two probability distributions in the same event space, the larger the KL distance, the greater the feature difference, that is, the discriminative feature. The monitoring classification module monitors the internal state of the petrochemical reactor based on the characteristic matrix.

2. According to claim 1, a fiber optic acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor is characterized in that: The sensing optical fiber is encapsulated in an optical cable and is engraved with an identical weak grating array.

3. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 2 is characterized in that: The acoustic wave sensing optical fiber is spirally wound on all the outer walls of the pipelines along the axial direction of the pipelines to realize distributed online monitoring.

4. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 1 is characterized in that: The demodulation module is based on optical time domain reflection technology The distributed acoustic wave sensor demodulation unit DAS includes a light source, an acousto-optic modulator AOM, an erbium-doped fiber amplifier EDFA, a circulator, a Michelson interferometer, a Faraday rotator, a 3×3 coupler, a photodetector PD and a high-speed acquisition and processing module.

5. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 4 is characterized by: The optical signal emitted by the light source is modulated by an acousto-optic modulator (AOM) and amplified by an erbium-doped fiber amplifier (EDFA) to form a pulse light that enters the sensing optical fiber; the light reflected by two adjacent gratings in the identical weak grating array of the sensing optical fiber passes through two arms of a Michelson interferometer and is reflected by a Faraday rotation mirror to a 3×3 coupler to cause interference; wherein the spacing between the two arms of the Michelson interferometer is the same as the spacing between two adjacent gratings in the weak grating array; the 3×3 coupler outputs three optical signals with a phase difference of 2π / 3, which are converted into electrical signals by three photoelectric detectors (PD); the high-speed acquisition and processing module acquires the three signals and uses an inverse tangent algorithm to calculate a phase signal that is linearly related to the acoustic wave signal.

6. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 1 is characterized by: The acoustic wave signal reconstruction module comprises a signal decomposition module and a signal superposition module.

7. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 6 is characterized by: The signal decomposition module preprocesses the phase signal, including pre-emphasis, framing and windowing; the preprocessed signal is decomposed into a number of intrinsic mode components through Complete EEMD with Adaptive Noise.

8. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 6 is characterized by: The signal superposition module uses the correlation coefficient to analyze the correlation between the intrinsic mode components and the preprocessed signal, and selects the intrinsic mode components with high correlation for superposition, thereby removing high-frequency noise and obtaining a denoised and reconstructed sound wave signal.

9. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 1 is characterized by: The feature selection and extraction module uses the Kullback-Leibler distance method to exclude irrelevant or redundant features from the time domain features, frequency domain features and spectral density index of the denoised and reconstructed sound wave signal, selects discriminative features, and combines them with Mel-frequency cepstral coefficients (MFCC) to form a feature matrix.

10. The optical fiber acoustic wave sensing online monitoring system for the internal state of a petrochemical reactor according to claim 1, characterized in that: The monitoring and classification module monitors and classifies pipeline leakages using a random forest classifier in machine learning based on a normalized feature matrix.

Images