Anti-aliasing filtering method and device for a distributed optical fiber acoustic sensing system

Abstract

The present application relates to the technical field of optical fiber sensing, and discloses an anti-aliasing filtering method and device for a distributed optical fiber acoustic sensing system, which dynamically adjusts the cut-off frequency and type of the anti-aliasing filter by analyzing the frequency band characteristics of environmental noise in real time, and combines frequency band energy detection to close invalid sampling channels, thereby significantly suppressing aliasing noise and reducing data volume. A device fault feature library matching mechanism is further introduced to trigger an emergency anti-aliasing mode when a sudden fault noise occurs, thereby improving fault detection accuracy. The method can be widely applied to oil and gas pipeline monitoring, industrial equipment health diagnosis and other fields, and solves the performance bottleneck of traditional fixed filters under wide frequency dynamic noise.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing technology, and in particular to an anti-aliasing filtering method and apparatus for a distributed fiber optic acoustic sensing system. Background Technology

[0002] Distributed fiber optic acoustic sensing (DAS) systems can detect vibration and acoustic signals over long distances with high spatial resolution by analyzing the phase changes of Rayleigh scattered light in optical fibers. They are widely used in oil and gas pipeline monitoring, industrial equipment health diagnostics, and urban underground space safety early warning systems. However, aliasing noise caused by broadband noise (such as mechanical vibration and electromagnetic interference) in complex environments severely limits system performance. The main solutions and limitations of existing technologies are as follows:

[0003] 1. Fixed anti-aliasing filter

[0004] Traditional DAS systems use low-pass filters with fixed cutoff frequencies (such as Butterworth filters) to suppress high-frequency noise (CN112511193A). However, in industrial scenarios, noise frequency bands change dynamically (e.g., the noise upper limit suddenly increases from 5kHz to 15kHz when equipment fails). Fixed filters cannot adapt dynamically, causing high-frequency noise to alias into the low-frequency effective signal band, resulting in a deterioration of the signal-to-noise ratio (SNR).

[0005] 2. Oversampling and digital post-filtering

[0006] The aliasing problem can be mitigated by increasing the sampling rate to four times the Nyquist frequency (e.g., 40 kHz / s) and combining it with digital downsampling (US20220065931A1). However, oversampling leads to a surge in data volume, generating hundreds of megabytes per second in long-distance DAS systems (e.g., 50 km fiber optic cables), exceeding transmission and storage capabilities. Furthermore, digital post-filtering cannot eliminate the aliasing distortion already introduced during the analog-to-digital conversion (ADC) stage.

[0007] 3. Frequency-band filter bank technology

[0008] A multi-channel bandpass filter (JP2021002573A) is used to divide the signal frequency band, and each sub-band is downsampled independently. However, the existing scheme has a fixed frequency band division, and when noise is concentrated in some frequency bands, it still samples the entire frequency band, resulting in a data redundancy rate as high as 60% (theoretical calculation). In addition, the hardware complexity and power consumption of the filter bank limit its engineering practicality.

[0009] In summary, existing technologies lack the ability to perceive dynamic noise environments in real time, and it is difficult to balance anti-aliasing strategies with resource efficiency, resulting in insufficient reliability of DAS systems in complex industrial scenarios. Summary of the Invention

[0010] The purpose of this invention is to propose an anti-aliasing filtering method and device for distributed fiber optic acoustic sensing systems, which solves the problem that existing technologies lack real-time sensing capabilities for dynamic noise environments and that anti-aliasing strategies and resource efficiency are difficult to balance, resulting in insufficient reliability of DAS systems in complex industrial scenarios.

[0011] Specifically, the present invention provides an anti-aliasing filtering method for a distributed fiber optic acoustic sensing system, comprising the following steps:

[0012] S1. Perform short-time high-speed sampling on the raw signal acquired by the distributed fiber optic acoustic sensing system, and extract the frequency band boundary of the current ambient noise through time-frequency analysis. , and transient noise markers;

[0013] S2, based on the noise frequency band boundary Dynamically calculate the cutoff frequency of the anti-aliasing filter and based on Select the filter type based on the frequency band range;

[0014] S3. The signal is divided into multiple sub-bands by analog filter bank, energy detection is performed on the signal of each sub-band, the sampling channel of the sub-band with energy below the preset threshold is closed, and the active sub-band is converted from analog to digital at sub-sampling rate to obtain the digital signal of the active sub-band.

[0015] S4. Perform digital interpolation and time-domain synthesis on the digital signal of the active sub-band to generate a full-band reconstructed signal;

[0016] S5. Perform similarity matching between the full-band reconstructed signal and the pre-stored equipment fault feature library. If the matching degree exceeds the threshold, trigger the anti-aliasing emergency mode, increase the sampling rate and enable the digital notch filter to prevent aliasing.

[0017] An anti-aliasing filter for a distributed fiber optic acoustic sensing system includes:

[0018] Noise sensing module: Includes an oversampling ADC unit, a short-time Fourier transform processor, and a transient noise detection unit. It is used to perform short-time high-speed sampling of the raw signals acquired by the distributed fiber optic acoustic sensing system, and extract the frequency band boundaries of the current ambient noise through time-frequency analysis. , and transient noise markers;

[0019] Dynamic filter module: Includes an analog low-pass filter with adjustable cutoff frequency and a filter type switching submodule, used to adjust the filter type according to the noise frequency band boundary. Dynamically calculate the cutoff frequency of the anti-aliasing filter and based on Select the filter type based on the frequency band range;

[0020] Frequency band processing module: includes a multi-channel bandpass filter bank, a sub-band energy detection unit and an independent ADC array, used to divide the signal into multiple sub-bands through the analog filter bank, perform energy detection on the signal of each sub-band, close the sampling channel of the sub-band with energy below a preset threshold, and perform analog-to-digital conversion on the active sub-band at a sub-sampling rate to obtain the digital signal of the active sub-band;

[0021] Signal reconstruction module: Includes a digital interpolator and a time-domain synthesizer, used to output a full-band reconstructed signal, and to perform digital interpolation and time-domain synthesis on the digital signals of the active sub-bands to generate a full-band reconstructed signal;

[0022] Fault Response Module: Includes a device fault feature library memory, a similarity matching processor, and an emergency mode controller. It is used to perform similarity matching between the full-band reconstructed signal and the pre-stored device fault feature library. If the matching degree exceeds the threshold, it triggers the anti-aliasing emergency mode, increases the sampling rate, and enables a digital notch filter to prevent aliasing.

[0023] The beneficial effects provided by this invention are: it can effectively suppress aliasing noise; theoretical calculations show that the dynamic cutoff frequency and filter switching strategy can reduce aliasing noise power to less than 20% of the traditional solution, and improve the signal-to-noise ratio by ≥12dB; it can improve resource efficiency; frequency band dynamic start-stop and data compression (differential coding + Huffman) reduce storage requirements by 40%-60%, and is suitable for long-distance DAS systems; it can improve the success rate of fault detection; in the case of sudden noise, the emergency mode can reduce the false alarm rate by ≥35% (Monte Carlo simulation verification). Attached Figure Description

[0024] Figure 1 This is a simplified flowchart of the method of the present invention;

[0025] Figure 2 This is a block diagram of the segmented processing module structure;

[0026] Figure 3 This is a schematic diagram of the fault response module. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0028] Before formally describing the present invention, a general description of the solution of the present invention will be given first to facilitate understanding.

[0029] Please refer to Figure 1 The present invention provides an anti-aliasing filtering method for a distributed fiber optic acoustic sensing system, comprising:

[0030] S1. Perform short-time high-speed sampling on the raw signal acquired by the distributed fiber optic acoustic sensing system, and extract the frequency band boundary of the current ambient noise through time-frequency analysis. , and transient noise markers;

[0031] S2, based on the noise frequency band boundary Dynamically calculate the cutoff frequency of the anti-aliasing filter and based on Select the filter type based on the frequency band range;

[0032] It should be noted that in step S2: when When, choose a 4th-order Butterworth filter; when When the filter is switched to a 6th-order elliptic filter.

[0033] S3. The signal is divided into multiple sub-bands by analog filter bank, energy detection is performed on the signal of each sub-band, the sampling channel of the sub-band with energy below the preset threshold is closed, and the active sub-band is converted from analog to digital at sub-sampling rate to obtain the digital signal of the active sub-band.

[0034] The analog filter bank mentioned in step S3 is a multi-channel bandpass filter, which divides the signal into... Sub-band ( The bandwidth of each sub-band is .

[0035] The subsampling rate described in step S3 satisfies Furthermore, the sampling rates of different sub-bands are independent of each other.

[0036] The energy detection threshold in step S3 is calculated as follows: ,in For background noise energy, k This is a preset coefficient, typically between 1.5 and 3.

[0037] S4. Perform digital interpolation and time-domain synthesis on the digital signal of the active sub-band to generate a full-band reconstructed signal;

[0038] S5. Perform similarity matching between the full-band reconstructed signal and the pre-stored equipment fault feature library. If the matching degree exceeds the threshold, trigger the anti-aliasing emergency mode, increase the sampling rate and enable the digital notch filter to prevent aliasing.

[0039] Step S5 is as follows:

[0040] S51. The equipment fault feature database stores at least the typical fault spectrum features of bearing wear, gear breakage, and pipeline leakage.

[0041] S52. The similarity matching adopts a normalized cross-correlation algorithm or cosine similarity calculation.

[0042] S53. If the similarity exceeds the threshold, the anti-aliasing emergency mode is triggered.

[0043] Specifically, the anti-aliasing emergency mode is as follows:

[0044] Instantly increase the sampling rate to The duration is 1.2-2 times the length of the burst noise window;

[0045] In the fault characteristic frequency band Internally enabled digital notch filter to suppress bandwidth The characteristic frequency is ±5%-10%.

[0046] An anti-aliasing filter for a distributed fiber optic acoustic sensing system includes:

[0047] Noise sensing module: Includes an oversampling ADC unit, a short-time Fourier transform processor, and a transient noise detection unit. It is used to perform short-time high-speed sampling of the raw signals acquired by the distributed fiber optic acoustic sensing system, and extract the frequency band boundaries of the current ambient noise through time-frequency analysis. , and transient noise markers;

[0048] Dynamic filter module: Includes an analog low-pass filter with adjustable cutoff frequency and a filter type switching submodule, used to adjust the filter type according to the noise frequency band boundary. Dynamically calculate the cutoff frequency of the anti-aliasing filter and based on Select the filter type based on the frequency band range;

[0049] Frequency band processing module: Includes a multi-channel bandpass filter bank, sub-band energy detection unit, and independent ADC array. It is used to divide the signal into multiple sub-bands using the analog filter bank, perform energy detection on the signal of each sub-band, close the sampling channels of sub-bands with energy below a preset threshold, and perform analog-to-digital conversion on the active sub-bands at a sub-sampling rate to obtain the digital signal of the active sub-band; see details for reference. Figure 2 , Figure 2 This is a block diagram of the segmented processing module;

[0050] Signal reconstruction module: Includes a digital interpolator and a time-domain synthesizer, used to output a full-band reconstructed signal, and to perform digital interpolation and time-domain synthesis on the digital signals of the active sub-bands to generate a full-band reconstructed signal;

[0051] Fault Response Module: This module includes a device fault feature library memory, a similarity matching processor, and an emergency mode controller. It performs similarity matching between the full-band reconstructed signal and the pre-stored device fault feature library. If the matching degree exceeds a threshold, an anti-aliasing emergency mode is triggered, increasing the sampling rate and enabling a digital notch filter to combat aliasing. (See details...) Figure 3 , Figure 3 This is a schematic diagram of the fault response module;

[0052] Specifically, the noise sensing module uses an FPGA-integrated STFT processor (Xilinx Zynq series).

[0053] The dynamic filter module uses a switched capacitor filter or a digitally controlled active filter, specifically a digitally controlled switched capacitor filter (LTC1562), which supports cutoff frequency adjustment from 1kHz to 20kHz.

[0054] In the frequency band processing module, the sampling rate of each sub-band ADC is dynamically configured by the FPGA; specifically, the frequency band ADC array is configured in multi-channel synchronous sampling mode (TI ADS127L01).

[0055] In this invention, differential coding and Huffman compression can also be performed on the sampling data of the active sub-band, with a compression rate of ≥30%; the ADC power consumption mode can be dynamically switched according to the noise activity, including full-speed mode, low-power mode and sleep mode.

[0056] The present invention provides two embodiments as follows:

[0057] Example 1: Bearing Wear Monitoring in Industrial Equipment

[0058] Scenario Description: In the monitoring of centrifugal pump units in chemical plants, the DAS system is deployed on the optical fiber laid on the surface of the pump body. It is necessary to extract the characteristic frequency of the bearing inner ring defect (theoretical calculated value 3.2kHz±200Hz) from the broadband background noise (motor fundamental frequency 500Hz, gear meshing harmonics 2-8kHz).

[0059] Implementation steps of this invention:

[0060] 1. Noise perception:

[0061] The oversampling ADC samples the original signal at 40 kS / s, and STFT analysis shows the upper limit of the noise bandwidth. The transient detection module marks periodic impacts (0.1s interval, 5ms duration).

[0062] 2. Dynamic filtering:

[0063] Calculate the cutoff frequency ,because Switch to a 6th-order elliptic filter (stopband rejection > 65dB).

[0064] 3. Frequency band processing:

[0065] The filter bank is divided into four sub-bands (0-5kHz, 5-10kHz, 10-15kHz, 15-20kHz), and the 15-20kHz channel is disabled by energy detection (energy is below a threshold). ).

[0066] Activate sub-bands at sub-sampling rate (0-5kHz): 5-10kHz: f 10-15kHz: ).

[0067] 4. Fault Response:

[0068] Spectrum matching revealed a sudden energy surge at 3.2 kHz, with 85% similarity to the characteristics of a bearing inner ring fault, triggering emergency mode.

[0069] Instantly increase the sampling rate to 2×8kHz=16kS / s.

[0070] Enable the digital notch filter (Q=30) at 3.2kHz±200Hz.

[0071] 5. Effect Verification:

[0072] The aliasing noise power is reduced to 18% of the traditional solution, the signal-to-noise ratio of bearing defect features is increased to 22dB, and the false alarm rate is reduced by 45%.

[0073] Example 2: Leak Detection of Urban Underground Water Supply Pipelines

[0074] Scene description:

[0075] Deploying a DAS system under the water supply pipeline beneath the main urban road requires suppressing traffic vibration noise (frequency band 0-15kHz) and detecting leakage acoustic emission signals (0.1-2kHz).

[0076] Implementation steps of this invention:

[0077] 1. Noise perception:

[0078] STFT analysis shows the upper limit of the noise bandwidth. (When a heavy truck passes by), transient noise is labeled as continuous broadband vibration.

[0079] 2. Dynamic filtering:

[0080] set up A 6th-order elliptic filter was selected.

[0081] 3. Frequency band processing:

[0082] It is divided into four channels: 0-4kHz, 4-8kHz, 8-12kHz, and 12-16kHz. The 12-16kHz channel is turned off (energy percentage <5%).

[0083] Subsampling rate setting: 0-4kHz ( ), 4-8kHz ( ), 8-12kHz ( ).

[0084] 4. Signal reconstruction and leakage detection:

[0085] After digital interpolation, the signal-to-noise ratio in the 0.1-2kHz band was improved to 18dB, a 0.8kHz leakage characteristic signal was detected, and the positioning accuracy reached ±5m.

[0086] 5. Resource optimization:

[0087] Data volume was reduced by 42% (1 / 4 channel disabled + compression algorithm), and storage requirements decreased from 1.2TB / day to 0.7TB / day.

[0088] In summary, the core of this invention lies in constructing a closed-loop control system encompassing noise perception, dynamic filtering, resource optimization, and fault response, specifically including:

[0089] 1. Real-time noise spectrum feedback mechanism:

[0090] Noise frequency band boundaries are extracted in real time using Short Time Fourier Transform (STFT) and wavelet analysis. Based on energy distribution and transient characteristics, a dynamic noise profile is generated.

[0091] 2. Adaptive anti-aliasing filtering:

[0092] according to Dynamically adjust the cutoff frequency of the analog filter ( ) and type (Butterworth / elliptic filter), balancing transition band smoothness and stopband rejection ratio.

[0093] 3. Dynamic optimization of frequency band resources:

[0094] The signal is divided into multiple sub-bands, invalid channels are shut down by energy detection, and active sub-bands are sampled at a sub-sampling rate. This reduces the amount of data by 30%-50%.

[0095] 4. Fault diagnosis and collaborative anti-aliasing:

[0096] A pre-stored database of equipment fault noise features (such as bearing wear spectrum) is used to match and trigger an emergency mode (oversampling + digital notch filtering) in real time to avoid aliasing masking fault features.

[0097] The beneficial effects of this invention are: it can effectively suppress aliasing noise; theoretical calculations show that the dynamic cutoff frequency and filter switching strategy can reduce aliasing noise power to less than 20% of the traditional solution, and improve the signal-to-noise ratio by ≥12dB; it can improve resource efficiency; frequency band dynamic start-stop and data compression (differential coding + Huffman) reduce storage requirements by 40%-60%, and is suitable for long-distance DAS systems; it can improve the success rate of fault detection; in the case of sudden noise, the emergency mode can reduce the false alarm rate by ≥35% (Monte Carlo simulation verification).

[0098] 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 principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for anti-aliasing filtering in a distributed fiber optic acoustic sensing system, characterized in that: Including the following: S1. Perform short-time high-speed sampling on the raw signal acquired by the distributed fiber optic acoustic sensing system, and extract the frequency band boundary of the current ambient noise through time-frequency analysis. , and transient noise markers; S2, based on the noise frequency band boundary Dynamically calculate the cutoff frequency of the anti-aliasing filter and based on Select the filter type based on the frequency band range; S3. The signal is divided into multiple sub-bands by analog filter bank, energy detection is performed on the signal of each sub-band, the sampling channel of the sub-band with energy below the preset threshold is closed, and the active sub-band is converted from analog to digital at sub-sampling rate to obtain the digital signal of the active sub-band. S4. Perform digital interpolation and time-domain synthesis on the digital signal of the active sub-band to generate a full-band reconstructed signal; S5. Perform similarity matching between the full-band reconstructed signal and the pre-stored equipment fault feature library. If the matching degree exceeds the threshold, trigger the anti-aliasing emergency mode, increase the sampling rate and enable the digital notch filter to prevent aliasing. The anti-aliasing emergency mode is as follows: Instantly increase the sampling rate to The duration is 1.2-2 times the length of the burst noise window; In the fault characteristic frequency band Internally enabled digital notch filter to suppress bandwidth The characteristic frequency is ±5%-10%.

2. The anti-aliasing filtering method for a distributed fiber optic acoustic sensing system as described in claim 1, characterized in that: The analog filter bank mentioned in step S3 is a multi-channel bandpass filter, which divides the signal into... Sub-band The bandwidth of each sub-band is .

3. The anti-aliasing filtering method for a distributed fiber optic acoustic sensing system as described in claim 1, characterized in that: The subsampling rate described in step S3 satisfies Furthermore, the sampling rates of different sub-bands are independent of each other.

4. The anti-aliasing filtering method for a distributed fiber optic acoustic sensing system as described in claim 1, characterized in that: The energy detection threshold in step S3 is calculated as follows: ,in For background noise energy, k These are preset coefficients.

5. The anti-aliasing filtering method for a distributed fiber optic acoustic sensing system as described in claim 4, characterized in that: Step S5 is as follows: S51. The equipment fault feature database stores at least the typical fault spectrum features of bearing wear, gear breakage, and pipeline leakage. S52. The similarity matching adopts a normalized cross-correlation algorithm or cosine similarity calculation. S53. If the similarity exceeds the threshold, the anti-aliasing emergency mode is triggered.

6. The anti-aliasing filtering method for a distributed fiber optic acoustic sensing system as described in claim 1, characterized in that: In step S2: when When choosing a filter, select a 4th-order Butterworth filter; when When the filter is switched to a 6th-order elliptic filter.

7. An anti-aliasing filter device for a distributed fiber optic acoustic sensing system, characterized in that: include: Noise sensing module: Includes an oversampling ADC unit, a short-time Fourier transform processor, and a transient noise detection unit. It is used to perform short-time high-speed sampling of the raw signals acquired by the distributed fiber optic acoustic sensing system, and extract the frequency band boundaries of the current ambient noise through time-frequency analysis. , and transient noise markers; Dynamic filter module: Includes an analog low-pass filter with adjustable cutoff frequency and a filter type switching submodule, used to adjust the filter type according to the noise frequency band boundary. Dynamically calculate the cutoff frequency of the anti-aliasing filter and based on Select the filter type based on the frequency band range; Frequency band processing module: includes a multi-channel bandpass filter bank, a sub-band energy detection unit and an independent ADC array, used to divide the signal into multiple sub-bands through the analog filter bank, perform energy detection on the signal of each sub-band, close the sampling channel of the sub-band with energy below a preset threshold, and perform analog-to-digital conversion on the active sub-band at a sub-sampling rate to obtain the digital signal of the active sub-band; Signal reconstruction module: Includes a digital interpolator and a time-domain synthesizer, used to output a full-band reconstructed signal, and to perform digital interpolation and time-domain synthesis on the digital signals of the active sub-bands to generate a full-band reconstructed signal; Fault Response Module: Includes a device fault feature library memory, a similarity matching processor, and an emergency mode controller. It is used to perform similarity matching between the full-band reconstructed signal and the pre-stored device fault feature library. If the matching degree exceeds the threshold, the anti-aliasing emergency mode is triggered, the sampling rate is increased, and a digital notch filter is enabled to prevent aliasing. The anti-aliasing emergency mode is as follows: Instantly increase the sampling rate to The duration is 1.2-2 times the length of the burst noise window; In the fault characteristic frequency band Internally enabled digital notch filter to suppress bandwidth The characteristic frequency is ±5%-10%.

Images