Method for differentiating between a pipe ball passage and an external impact zone

By using energy statistics and physical spatial positioning algorithms from dual sensors, the false alarm problem of pigging events inside pipelines and external impact events is solved. It achieves low false alarm rate detection and quantitative location of vibration sources without the need for time synchronization, and is applicable to pigging positioning and external intrusion interference elimination in oil and gas transportation, urban pipeline networks and nuclear facility pipelines.

CN123066686APending Publication Date: 2026-11-10BEIJING XINCHI TIMES TECHNOLOGY CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202611247440.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-17
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing technologies are prone to false alarms when distinguishing between ball-passing events inside pipelines and external impact events. They also require high-precision time synchronization equipment and complex systems, and cannot effectively quantitatively locate the spatial position of the vibration source, thus having limited distinguishing capabilities.

Method used

By employing a sliding window energy statistics and vibration source spatial localization method, the vibration source location is calculated using the mean and variance of the vibration signal through the energy statistical characteristics between the two sensors and the physical spatial localization algorithm, achieving low false alarm rate detection without the need for time synchronization or modification of the sphere.

Benefits of technology

It enables accurate differentiation under different pipe diameters, media and noise levels, reduces system cost and deployment difficulty, provides a quantitative description of the spatial location of vibration sources, and improves the situational awareness capability of on-site operations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123066686A_ABST
    Figure CN123066686A_ABST
Patent Text Reader

Abstract

The application discloses a pipeline ball passing and impact distinguishing method based on sliding window energy statistics and vibration source spatial positioning. The pipeline wall vibration signals are synchronously collected by double sensors, the instantaneous energy difference coefficient is calculated, and the mean value and the standard deviation thereof are extracted in a continuous time window. Based on the vibration index attenuation model, the vibration source offset is inversely calculated from the mean value, and the vibration source stability is distinguished from the standard deviation. When the ball passes through the middle region, a small offset, small variance and stable equilibrium mode are presented, and the ball passing event is determined. When the external impact occurs, a large offset, large variance and transient imbalance mode are presented, and the external impact event is determined. The application distinguishes the ball passing and the interference from the spatial dimension, does not need time synchronization and additional devices, has strong anti-interference ability and accurate positioning.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of online monitoring technology for pipeline transport vehicles, specifically to a method for accurately distinguishing between pipeline pigging events and external impact events by utilizing the energy statistical characteristics and physical propagation model of dual-channel vibration signals. This method is particularly applicable to the positioning of pigging balls and the elimination of external intrusion interference in oil and gas transportation, urban pipeline networks, and nuclear facility pipelines. Background Technology

[0002] In pipeline cleaning and inspection operations, accurately determining whether the cleaning pig has passed through the inspection section (the pig-passing event) is a core technical challenge. Existing technologies are mainly divided into two categories: single-sensor intensity detection and dual-sensor time-difference positioning.

[0003] Single-sensor technology (such as US5549000) only senses vibration intensity, making it prone to false alarms when the pipeline is subjected to external impacts, as it cannot distinguish whether the vibration originates from the passage of a transport vehicle inside the pipeline or from external mechanical impact. While dual-sensor time-difference positioning technology (such as CN116774150A and CN110107816A) introduces a spatial dimension, its core logic remains the detection of the "time sequence of signal peak arrivals," locating the vibration source by determining the time difference between the arrival times of the vibration signal at the two sensors. However, this time-dimensional judgment has the following drawbacks: firstly, sudden strong external impacts can also generate vibration wavefronts with a clear time sequence, potentially deceiving the time-difference-based positioning logic; secondly, time-difference positioning methods often require high-precision time synchronization equipment, increasing system cost and on-site deployment difficulty.

[0004] In summary, the existing technologies generally have the following technical defects: (1) They rely on time difference or intensity threshold for judgment, which is easily deceived by external impacts and has a high false alarm rate; (2) They require high-precision time synchronization equipment or additional signal generation devices, which makes the system complex and costly; (3) They cannot quantitatively locate the spatial position of the vibration source and have limited distinguishing ability.

[0005] The technical problems to be solved by this invention are: how to distinguish between the sphere and interference from a spatial dimension rather than a temporal dimension; how to quantitatively calculate the position of the vibration source and eliminate interference ambiguity; and how to achieve low false alarm rate detection without the need for time synchronization and without modifying the sphere. Summary of the Invention

[0006] This invention provides a method for distinguishing between pipe ball passing and external impact based on sliding window energy statistics and vibration source spatial positioning. By introducing time-domain statistical features (mean and variance) and physical spatial positioning algorithm, it fundamentally solves the problem of distinguishing between ball passing and external impact, and achieves low false alarm rate detection without time synchronization or modification of the ball passing.

[0007] This invention provides the following technical solution:

[0008] 1. A method for distinguishing between ball passing through a pipeline and external impact, characterized by comprising the following steps:

[0009] S1: The first vibration sensor and the second vibration sensor are arranged at intervals along the axial direction of the pipeline. The distance L between the two sensors satisfies L=2×D×(1+β), where D is the inner diameter of the pipeline and β is the adjustment coefficient and 0.1≤β≤0.3. The vibration signal of the pipe wall is collected synchronously through the first vibration sensor and the second vibration sensor.

[0010] S2: Bandpass filtering is applied to the acquired vibration signal. The filtered signal is then truncately within a preset time window T. For each time window, the energies E1(t) and E2(t) of the vibration signals acquired by the first and second vibration sensors within that time window are calculated. The time window T satisfies the following conditions: Where v max To achieve the expected maximum running speed of the sphere;

[0011] S3: For each time window, calculate the instantaneous energy difference coefficient. Calculate the mean μ of δ(t) over N consecutive time windows. δ and standard deviation σ δ :

[0012] S4: Based on the mean μ δ Calculate the vibration source offset Where α is the pre-calibrated pipeline vibration attenuation coefficient;

[0013] S5: When |Δx| is less than the first preset distance threshold D th1 And σ δ Less than the first statistical threshold σ th1 When |Δx| is greater than the second preset distance threshold D, it is determined as a ball-passing event; th2 And σ δ Greater than the second statistical threshold σ th2 When the event occurs, it is determined to be an external impact event; where the first preset distance threshold D th1 =L / 6, the second preset distance threshold D th2 =L / 3, the first statistical threshold σ th1 With the second statistical threshold σ th2 Based on the standard deviation σ of pipeline background noise bg Perform adaptive calibration: σ th1 =λ I ·σ bs , where 3≤λ1≤8; σ th2 =λ2·σ bg Where 10≤λ2≤25; and σ th1 Negatively correlated with the number of sliding windows N:

[0014] 2. The method according to claim 1, wherein when the medium in the pipeline is gas, the bandpass filter frequency band in step S2 is 0.5kHz to 2kHz; and when the medium in the pipeline is liquid, the bandpass filter frequency band in step S2 is 1kHz to 5kHz.

[0015] 3. The method according to claim 1, wherein the calibration method of the attenuation coefficient α is as follows: apply a standard pulse excitation at a known position in the pipeline, record the energy ratio received by the two sensors, and then deduce the actual value of α.

[0016] 4. The method according to claim 1, wherein the value range of the continuous time window N in step S3 is 5≤N≤20, and the total observation time N×T is less than 1 / 2 of the time required for the sphere to completely pass through the distance between the two sensors.

[0017] 5. The method according to claim 1, characterized in that, when D th1 ≤|Δx|≤D th2 At this time, enter the fuzzy area to assist in the determination:

[0018] The total energy index E of continuously monitoring M sliding windows sum =E1+E2 and standard deviation σ δ The changing trend of M, where M is a preset positive integer and 2≤M≤5;

[0019] If E sum It monotonically increases over M consecutive time windows, and σ δ If the number of consecutive time windows decreases monotonically, it is considered a ball-passing event.

[0020] If E sum The increment within a single time window exceeds the preset energy mutation threshold ΔE th , and σ δ The increment within a single time window exceeds the preset variance mutation threshold Δσ th If so, it can be further determined to be an external shock event;

[0021] Wherein, the energy mutation threshold ΔE th and variance mutation threshold Δσ th Adaptive calibration based on pipeline background noise level:

[0022]

[0023] Δσ th =η σ ·σ bg

[0024] in Let η be the mean of the total energy under background noise. E The preset energy mutation coefficient is 3≤η. E ≤10, η σ The variance mutation coefficient is preset and 5≤η σ ≤20.

[0025] 6. The method according to claim 1, characterized in that, in step S2, before bandpass filtering the acquired vibration signal, the method further includes detrending term processing and amplitude normalization processing of the acquired vibration signal, wherein the detrending term processing adopts a polynomial least squares fitting method to subtract the fitted trend term from the original signal, and the amplitude normalization processing adopts a maximum-minimum value normalization method to linearly map each channel signal to the [0, 1] interval.

[0026] 7. The method according to claim 1, wherein the vibration sensor is a piezoelectric vibration sensor or a MEMS accelerometer, and the sampling frequency is not less than 20kHz. o

[0027] 8. The method according to claim 1, wherein the ball passing through the pipeline can be a pigging ball, a detection ball, a diameter measuring ball, or an intelligent detector.

[0028] 9. The method according to claim 1, characterized in that it further includes a confidence assessment step: based on the mean μ δ and standard deviation σ δ Calculate the signal-to-noise ratio (SNR). When the SNR is lower than a preset threshold, output the message "Signal is weak, cannot be determined".

[0029] The main beneficial effects of this invention are as follows: Based on the exponential decay law of pipe wall vibration wave propagation, this invention constructs a quantitative physical model between the energy received by dual sensors and the spatial position of the vibration source, directly mapping the energy difference coefficient to the vibration source offset, thus achieving quantitative inversion of the vibration source's spatial position. Furthermore, by combining the vibration source offset with the signal standard deviation for dual-dimensional discrimination, external impacts cannot be disguised as a pig-passing event, eliminating the possibility of false alarms from a physical standpoint. Simultaneously, this invention does not require GPS timing or high-precision clock synchronization, nor does it require installing magnets or transmitters on the pig-passing device; detection can be achieved with only two ordinary vibration signals, significantly reducing system costs and on-site deployment difficulty. In addition, this invention employs a sliding window within... The system uses the mean and standard deviation, rather than a single instantaneous value, for judgment. The mean ensures positioning accuracy, while the standard deviation ensures recognition stability and effectively filters out random noise spikes. The sensor spacing is coupled with the pipe diameter, the filtering frequency band follows the changes in the medium, the attenuation coefficient can be calibrated on-site, and the statistical threshold is adaptively generated based on the standard deviation of the background noise and dynamically adjusted with the number of sliding windows. This allows the invention to adapt to complex industrial environments with different pipe diameters, different media, and different noise levels. The invention can also output the specific value of the vibration source offset, realizing a quantitative description of the spatial position of the vibration source. This provides operators with auxiliary criteria for event type discrimination and position diagnosis information, which helps to further improve the situational awareness capability of on-site operations. Attached Figure Description

[0030] Figure 1 This is a flowchart of the overall method of the present invention.

[0031] Figure 2 This is a schematic diagram illustrating the sensor layout and the definition of the vibration source offset x in this invention.

[0032] Figure 3 The waveform and sliding window statistical characteristics (mean μ) of the instantaneous energy difference coefficient δ(t) under the spherical condition of this invention are presented. δ and standard deviation σ δ (Diagram)

[0033] Figure 4 This is a schematic diagram of the waveform and statistical characteristics of the instantaneous energy difference coefficient δ(t) under external impact conditions according to the present invention.

[0034] Figure 5 For different event types in this invention in (μ) δ , σ δ Cluster distribution diagram in two-dimensional feature space. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Experimental methods not specifying specific conditions in the embodiments are generally performed under conventional conditions or according to the conditions recommended by the manufacturer.

[0036] This invention achieves accurate differentiation between ball-passing events and external impact events based on the following core physical model and signal processing strategy.

[0037] (1) Exponential decay physical model

[0038] Vibration waves propagate through the pipe wall following a strict exponential decay law. Let the energy of the vibration source be E0, and the propagation distance be r, then the energy received by the sensor can be expressed as:

[0039] E = k·E0·e -ar

[0040] Where α is the pipe vibration attenuation coefficient (unit: Np / m), which depends on the pipe material, pipe diameter and internal medium type; k is the sensor coupling coefficient.

[0041] Assume the first and second sensors are spaced apart along the pipe axis with a spacing of L, and the offset of the vibration source relative to the midpoint of the two sensors is x (positive in the direction towards the first sensor). Then the distances from the vibration source to the first and second sensors are respectively... and The energies received by the two sensors are: E1 = kE0e -α(L / 2-x) E2=kE0e -α(L / 2+x) .

[0042] From the above two equations, we can see that the ratio of the energy received by the two sensors is... It depends only on the source offset x, and is independent of the source intensity E0 and coupling coefficient k. Therefore, the spatial location of the source can be uniquely determined by the ratio of the energy received by the two sensors, without relying on time synchronization equipment.

[0043] (2) Signed energy difference coefficient

[0044] Define the instantaneous energy difference coefficient:

[0045]

[0046] Its range is [-1, 1]. Substituting the double exponential decay model into the above definition, we obtain the energy expression: Take the natural logarithm of both sides: Substitute E1 and E2 We can obtain:

[0047] Therefore, the precise analytical relationship between the vibration source offset x and δ can be obtained:

[0048]

[0049] This formula is the core mapping relationship of this invention. It shows that by simply calculating the difference coefficient δ of the energy received by the two sensors, the spatial position x of the vibration source in the axial direction of the pipe can be quantitatively determined without measuring the signal arrival time difference, without high-precision clock synchronization, and without installing any additional signal generating device on the sphere.

[0050] (3) Statistical distinction between "steady-state equilibrium" and "transient imbalance"

[0051] The physical properties of a ball-passing event and an external shock are fundamentally different, and this difference can be distinguished by the statistical characteristics of δ(t).

[0052] Ball-passing event (steady-state equilibrium mode): The pigging ball is large and moves relatively slowly (typically 1–5 m / s). When it passes through the area between the two sensors, the vibration source position satisfies x≈0, and the position changes smoothly over a continuous time period. Therefore, δ(t) remains stable within the continuous time window, exhibiting a mean μ. δ Approaching 0, standard deviation σ δ Extremely small statistical characteristics. This characteristic is known as the "steady-state equilibrium" pattern.

[0053] External impact (transient imbalance mode): An external knock or impact occurs on one side of the pipe (near a sensor), the vibration source location satisfies a large |x|, and the impact is an instantaneous, sudden excitation with an extremely short duration (usually on the order of milliseconds), causing δ(t) to jump dramatically and exhibit a mean value |μ|. δ | Significantly large, standard deviation σ δ A dramatic increase in statistical characteristics. This characteristic is known as the "transient imbalance" pattern.

[0054] Based on the above physical principles, this invention uses μ δ Quantitatively locate the spatial position of the vibration source, using σ δ The stability of the vibration source is quantified, and the difference between ball impact and vibration is accurately distinguished through a two-dimensional joint discrimination.

[0055] Example 1

[0056] This embodiment provides a method for distinguishing between pipe ball passage and external impact based on sliding window energy statistics and vibration source spatial positioning.

[0057] (I) System Setup and Parameter Setting

[0058] Reference Figure 2 Two piezoelectric accelerometers are installed on a DN200 (inner diameter D = 0.2m) liquid transport pipeline. Taking an adjustment coefficient β = 0.2, calculate the sensor spacing.

[0059] L = 2 × 0.2 × (1 + 0.2) ≈ 0.48m

[0060] Rounded to 0.5m. Estimated maximum operating speed v of the pigging machine. max =5m / s, set sliding time window

[0061]

[0062] (i.e., 10ms). Set the number of sliding windows N=10, and the total observation time T. obs =N×T=100ms. Based on the characteristics of the liquid pipeline, the bandpass filter frequency is set to 1kHz~5kHz.

[0063] Before performing bandpass filtering in step S2, the acquired raw vibration signal is first preprocessed, including detrending term processing and amplitude normalization processing.

[0064] Detrend term removal: A first-order polynomial least squares fitting method is used to estimate the linear trend term of the original vibration signal. The fitted trend term is then subtracted from the original signal to obtain the signal after removing the trend term. This process effectively eliminates low-frequency interference introduced by sensor zero-point drift and temperature drift, and avoids systematic bias in subsequent energy calculations caused by the trend term.

[0065] Amplitude normalization: A maximum-minimum normalization method is used to linearly map each channel signal to the [0, 1] interval to eliminate amplitude inconsistencies caused by sensitivity differences between the two sensors. The specific calculation formula is as follows:

[0066]

[0067] Where x(i) is the signal sequence after detrending, x min and x max These are the minimum and maximum values ​​of the signal in the sliding time window, respectively. This process ensures that the amplitude scales of the two sensor channels are consistent, so that the subsequent calculation of the energy difference coefficient δ(t) is not affected by the channel sensitivity differences.

[0068] Attenuation coefficient calibration: Apply a standard pulse excitation at the midpoint of the pipe, record the energy ratio E1 / E2 received by the two sensors, and substitute it into the formula. The attenuation coefficient of the pipeline is calculated to be approximately 0.8 Np / m.

[0069] Background noise calibration: During a quiet period when there is no ball passing through the pipeline, no work, and no external interference, a dual-channel signal with a duration of more than 10 seconds is collected, and the δ during the quiet period is calculated. bg (t) and its standard deviation σ bg In this embodiment, σ was measured. bg =0.008. Taking λ1 = 5 and λ2 = 18, then:

[0070] σ th1 =λ1·σ bg =5 × 0.008 = 0.04

[0071] σ th2 =λ2·σ bg =18 × 0.008 = 0.144

[0072] Meanwhile, considering N=10, σ th1 Calibrate according to a relationship inversely proportional to N, and actually take σ. th1 =0.05, σ th2 =0.15. Set distance threshold:

[0073]

[0074] (II) Data Acquisition and Calculation for Ball-Passing Conditions

[0075] When the pigging ball passes through the area between the two sensors, a dual-channel vibration signal is collected. Following step S2, the signal is bandpass filtered, and the signal is captured in 10ms time windows. The energy E1(t) and E2(t) within each time window are calculated. Following step S3, the instantaneous energy difference coefficient δ(t) for each time window is calculated, and the mean and standard deviation of δ(t) are calculated over 10 consecutive time windows (100ms).

[0076] Under the sphere-passing condition, δ(t) fluctuates slightly between -0.1 and 0.1. The calculated mean μ is... δ =0.05, standard deviation σ δ =0.03.

[0077] Calculate the vibration source offset by substituting the values ​​into the positioning formula according to step S4:

[0078]

[0079] Since Δx > 0, it indicates that the vibration source is biased towards sensor 1 relative to the midpoint, which matches the actual position of the sphere just past the midpoint and biased towards sensor 1.

[0080] Perform pattern recognition according to step S5:

[0081] |Δx|=0.0625<D th1 (0.083), σ δ =0.03<σ th1 (0.05)

[0082] If both conditions are met simultaneously, the system is determined to be in steady-state energy balance mode, outputs "ball passing event", and indicates that the vibration source is deviated from the midpoint of the two sensors by about 6.25cm.

[0083] (III) Data Acquisition and Calculation for External Impact Conditions

[0084] When the flange upstream of the pipeline (near sensor 1) is struck by an external tool, sensor 1 receives significantly more energy than sensor 2. The calculated instantaneous δ(t) jumps to 0.85.

[0085] Within a 100ms observation window, the δ(t) data fluctuated dramatically. The mean μ was calculated to be... δ =0.82, standard deviation σ δ =0.35.

[0086] Substitute the values ​​into the positioning formula to calculate the vibration source offset:

[0087]

[0088] Perform pattern recognition:

[0089] |Δx|=1.625>D th2 (0.167), σ δ =0.35>σ th2 (0.15)

[0090] If both conditions are met simultaneously, the system is identified as a transient energy imbalance mode and outputs "external impact event".

[0091] (iv) Clustering effect in two-dimensional feature space

[0092] Reference Figure 5 Projecting multiple sets of experimental data onto a μ-scale δ σ is the horizontal axis. δ On a two-dimensional plane with the vertical axis as the ordinate, the data points for the ball-passing event are densely distributed near the origin (lower left corner region), while the data points for the external impact event are distributed in a region with high mean and high variance (upper right corner region). The two types of events have clear classification boundaries on the two-dimensional plane, verifying the effectiveness and high robustness of the dual-judgment condition of this invention.

[0093] This embodiment verifies the effectiveness of the basic technical solution of the present invention:

[0094] High detection accuracy: Under ball-passing conditions, the system accurately outputs the ball-passing event determination, |Δx| and σ.δ All conditions are met for steady-state equilibrium; under external shock conditions, the system accurately outputs the external shock event determination, |Δx| and σ. δ All satisfy the transient imbalance condition.

[0095] The physical model was validated: the source offset Δx derived from the exponential decay model matched the actual position of the source. When passing the ball, Δx≈0, and when impacted, |Δx| was significantly larger, which verified the correctness of the mapping relationship of the physical model.

[0096] Statistical characteristics that distinguish effectively: standard deviation σ δ The standard deviation is minimal (0.03) under ball-passing conditions and increases dramatically (0.35) under impact conditions, verifying the effectiveness of the standard deviation as a quantitative indicator of vibration source stability.

[0097] Dual-dimensional joint judgment enhances reliability: Single-dimensional (relying only on Δx or only on σ) δ Neither of these methods is sufficient to completely distinguish between the two types of events. The joint determination based on two dimensions forms a clear classification boundary and enhances the robustness of the determination.

[0098] Example 2

[0099] This embodiment, based on Embodiment 1, further introduces a statistical threshold based on the background noise standard deviation σ. bg The technical feature of adaptive calibration is used to verify the adaptability of the present invention under different environmental noise conditions.

[0100] (I) Threshold Calibration under Different Background Noise Scenarios

[0101] In industrial settings, pipeline background noise can vary significantly depending on the medium being transported, the flow rate, and the surrounding environment. This embodiment conducts experiments in both low-noise scenarios (such as nighttime shutdown) and high-noise scenarios (such as the operation of adjacent pump units).

[0102] Low-noise scenario: During the quiet period when there is no operation on the pipeline and no external interference, a 10-second dual-channel signal is collected, and σ is calculated. bg =0.006. Taking λ1 = 5 and λ2 = 18, then...

[0103] σ th1 =λ1·σ bg =5 × 0.006 = 0.03

[0104] σ th2 =λ2·σ bg =18 × 0.006 = 0.108

[0105] Meanwhile, considering N=10, σ th1 Calibrate according to a relationship inversely proportional to N, and actually take σ. th1=0.035, σ th2 =0.11.

[0106] High-noise scenario: Under the condition of nearby pump unit operation, a 10-second signal is collected, and σ is calculated. bg =0.025. Taking λ1 = 4 (appropriately reducing the factor to maintain detection sensitivity), λ2 = 15, then...

[0107] σ th1 =λ1·σ bg =4 × 0.025 = 0.10

[0108] σσ th2 =λ2·σ bg =15 × 0.025 = 0.375

[0109] Consider N = 10, σ th1 Calibrate according to a relationship inversely proportional to N, and actually take σ. th1 =0.09, σ th2 =0.35.

[0110] (II) Dynamic Threshold Verification

[0111] In the two scenarios described above, ball-handling and impact-handling tests were conducted respectively:

[0112] Low-noise scenario - pigging condition: When the pig passes through the middle area, the calculated μ is... δ =0.03, σ δ =0.025, substituting into the positioning formula, we get Δx≈0.037m. σ δ =0.025<σ th1 =0.035, and |Δx|≈0.037<D th1 =0.083m. Both conditions are met simultaneously, so it is correctly determined to be a ball-passing event.

[0113] Low-noise scenario - impact condition: When an external impact occurs, μ is calculated. δ =0.78, σ δ =0.32, substituting into the positioning formula, we get Δx≈1.08m. σ δ =0.32>σ tb2 =0.11, and |Δx|≈1.08>D th2 =0.167m. Both conditions are met simultaneously, and the event is correctly identified as an external impact event.

[0114] High-noise scenario – pigging operation: When the pig passes through the middle area, the background noise increases, and the fluctuation of δ(t) intensifies. The calculated μ... δ =0.06, σ δ=0.08, substituting into the positioning formula, we get Δx≈0.075m. Since the threshold has been adaptively adjusted to σ... th1 =0.09, σ δ =0.08<σ th1 =0.09, and |Δx|≈0.075<D th1 =0.083m. Both conditions are met simultaneously, and the event is still correctly identified as a ball-passing event, without any false alarm.

[0115] High-noise scenario - impact condition: When an external impact occurs, μ is calculated. δ =0.84, σ δ =0.40, substituting into the positioning formula, we get Δx≈1.55m. σ δ =0.40>σ th2 =0.35, and |Δx|≈1.55>D th2 =0.167m. Both conditions are met simultaneously, and the event is correctly identified as an external impact event.

[0116] (III) Verification of the dynamic threshold update mechanism

[0117] To ensure that the threshold always matches the current environmental state, this embodiment incorporates a timed update logic in the system: during system operation, background noise data is automatically collected every 24 hours during periods without events, and σ is recalculated. bg And update σ th1 and σ th2 .

[0118] To verify the effectiveness of this mechanism, this embodiment simulates a scenario where environmental noise gradually increases: initial σ bg =0.008, σ th1 =0.05; after 4 hours, the background noise increased to σ bg =0.018. If the threshold is not updated, σ under the ball-passing condition will be... δ A value of 0.06 would exceed the fixed threshold of 0.05, leading to a false alarm; however, with the dynamic update mechanism, the system automatically updates the threshold to σ during the most recent quiet period. th1 =0.07, σ for ball-passing condition δ =0.06<σ th1 =0.07, still able to correctly determine the value. Experiments show that this dynamic adaptive mechanism enables the present invention to maintain a recognition accuracy of over 98.5% under different operating conditions where the signal-to-noise ratio changes by more than 10 times, which is significantly better than the fixed threshold scheme (the false alarm rate of the fixed threshold is over 15% in high-noise scenarios).

[0119] This embodiment verifies the effectiveness of the technique based on adaptive calibration threshold for background noise:

[0120] Significantly enhanced environmental adaptability: The threshold automatically adjusts with the background noise level, maintaining stable detection performance in both low-noise and high-noise scenarios, avoiding false alarms or missed alarms in scenarios with fluctuating noise levels when the threshold is fixed.

[0121] Dynamic updates ensure long-term stability: The timed automatic update mechanism enables the system to track slow changes in environmental noise without the need for manual intervention in threshold adjustments, thus reducing operation and maintenance costs.

[0122] Significantly superior to the fixed threshold scheme: Under conditions where the signal-to-noise ratio changes by more than 10 times, the adaptive threshold scheme achieves an accuracy rate of over 98.5%, while the false alarm rate of the fixed threshold scheme is as high as 15%, verifying the necessity of adaptive calibration.

[0123] Example 3

[0124] Based on the previous embodiments, this embodiment further introduces a fuzzy region-assisted determination scheme to verify the discrimination effect when the vibration source position is in the intermediate transition region.

[0125] (I) Simulation of the fuzzy area scene and parameter settings

[0126] In practical engineering, there exists |Δx| between D th1 and D th2 In certain conditions, such as when a sphere has just entered the detection area but has not yet reached the midpoint, or when an external impact occurs just near the midpoint between the two sensors, the single Δx threshold method is insufficient to make an accurate judgment under these "fuzzy" conditions, necessitating the introduction of auxiliary discrimination logic.

[0127] In this embodiment, M=5, meaning five sliding windows are monitored continuously. Background noise calibration yields E. bg =0.10, σ bg =0.008. Take η E =5, η σ =10, then:

[0128]

[0129] Δσ th =η σ ·σ bg =10 × 0.008 = 0.08

[0130] This embodiment sets up two blurred area scenarios:

[0131] Scenario A: As the pig enters the detection area, |Δx|≈0.12m (between 0.083m and 0.167m), σ δ ≈0.06 (higher than σ) th1 =0.05 but lower than σ th2=0.15).

[0132] Scenario B: An external impact is applied near the midpoint between the two sensors, |Δx|≈0.10m (between the two thresholds), σ δ ≈0.18 (higher than σ) th1 =0.05 and slightly higher than σ th2 =0.15).

[0133] (II) Auxiliary Judgment Logic and Results

[0134] According to the fuzzy region determination method described in claim 5, the system enters the auxiliary determination process: continuously monitoring the total energy index E of 5 sliding windows. sum =E1+E2 and σ δ The changing trend of E. sum Five consecutive windows monotonically increasing and σ δ If five consecutive windows show a monotonically decreasing trend, it is determined that the ball-passing event is approaching the midpoint; if E sum Single window increment exceeds ΔE th And σ δ Single window increment exceeds Δσ th If so, it is determined to be an external shock event.

[0135] Scenario A discrimination result: E for 5 consecutive windows sum The values ​​are [0.30, 0.35, 0.40, 0.46, 0.52], showing a monotonically increasing trend; the σ values ​​for five consecutive windows... δ The values ​​are [0.075, 0.070, 0.065, 0.058, 0.052], showing a monotonically decreasing trend. When both conditions are met simultaneously, the system assists in determining that the "ball-crossing event is approaching the midpoint" and outputs a ball-crossing warning. Within the subsequent three observation windows, |Δx| continuously decreases until it falls below D. th1 The judgment has been confirmed to be correct.

[0136] Scenario B judgment result: Before the external knocking occurs, E sum Stable at around 0.12, σ δ It stabilizes at around 0.02; within a single time window of the tap, E sum The single-window increment ΔE jumped from 0.12 to 0.85. sum =0.73>ΔE th (0.50); ​​σ δ The single-window increment Δσ jumped from 0.02 to 0.20. δ =0.18>Δσ th (0.08). When both mutation conditions are met simultaneously, the system assists in determining it as an "external shock event" and outputs a shock alarm. In the subsequent observation window, δ(t) quickly recovers to a near-zero value, further confirming the transient characteristics of the shock event.

[0137] (III) Statistical Validation

[0138] In a comparative experiment involving 100 sets of fuzzy-zone conditions, the accuracy of using the Δx threshold alone was approximately 78.5%; introducing E... sum and σ δ Monotonicity judgment and ΔE th , Δσ th After the mutation threshold was used to assist in the judgment, the accuracy rate was improved to over 99.0%, which verified the effectiveness of the auxiliary judgment logic.

[0139] Beneficial effects of this embodiment

[0140] This embodiment verifies the technical effectiveness of fuzzy region-assisted determination:

[0141] Compensating for the inherent blind spot of the threshold method: When the vibration source location falls exactly between two distance thresholds, the single Δx threshold method cannot make a reliable judgment. The auxiliary discrimination logic analyzes E... sum monotonicity and σ δ The monotonicity of the trend effectively solves this problem. By adopting a judgment rule of monotonically increasing / decreasing over M consecutive windows, the trend judgment is made objective and quantifiable, avoiding ambiguity caused by subjective descriptions.

[0142] The mutation threshold provides an objective basis for judgment: by setting a preset energy mutation threshold ΔE th and variance mutation threshold Δσ th Furthermore, adaptive calibration is performed based on the background noise level, providing clear mathematical criteria for "instantaneous change" and "sharp increase," thereby improving the objectivity and reproducibility of the judgment rules.

[0143] Total energy and variance trends provide key discriminative information: E during ball passage sum It shows a monotonically increasing trend, σ δ It shows a monotonically decreasing trend; E during impact sum and σ δ All increments within a single window exceeded the preset mutation threshold. The two types of events differ fundamentally in both trend and mutation characteristics, and the auxiliary judgment logic accurately captured these differences.

[0144] Significantly improved accuracy: The auxiliary judgment improved the recognition accuracy of ambiguous areas from 78.5% to over 99.0%, verifying the necessity and effectiveness of the logic.

[0145] Example 4

[0146] This embodiment, based on the aforementioned embodiments, further introduces a confidence assessment step to verify the reliability of the determination under weak signal conditions.

[0147] (I) Experimental Setup

[0148] Under conditions where the medium inside the pipeline is high-pressure gas and the sensor coupling is poor, the signal attenuation is severe, and the energy received by both sensors approaches the background noise level. After signal acquisition, the mean μ is calculated according to steps S2 to S4. δ and standard deviation σ δ The signal-to-noise ratio is calculated based on the mean and standard deviation.

[0149]

[0150] When SNR < 3, the system determines that the signal quality is insufficient and outputs the prompt message "Signal is weak and cannot be determined" to avoid false alarms under low signal-to-noise ratio conditions.

[0151] (II) Verification Results

[0152] In 100 tests under low signal-to-noise ratio conditions, this confidence assessment step reduced the false alarm rate from 23% before implementation to 0%, and there were no cases of false alarms where valid signals were misjudged as "unable to determine" (i.e., when SNR≥3, the system was able to give the correct ball passing or impact judgment), verifying the effectiveness of this step in ensuring the reliability of detection.

[0153] (III) Beneficial Effects of This Embodiment

[0154] This embodiment verifies the technical effectiveness of the confidence assessment step:

[0155] Effectively avoids false alarms at low signal-to-noise ratios: When the signal is close to the background noise level, the system actively refuses to make a judgment instead of forcibly outputting a result, reducing the false alarm rate from 23% to 0%.

[0156] Without sacrificing the detection rate of effective signals: When the signal-to-noise ratio is higher than the threshold, the system can still make correct judgments without any missed detections, verifying the rationality of the threshold setting.

[0157] Enhancing overall system credibility: The confidence assessment step enables the system to provide clear prompts rather than erroneous results when signal quality is insufficient, thereby increasing user trust in the system's output.

[0158] The above description is merely 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 distinguishing between ball-passing and external impact in a pipeline, characterized in that, Includes the following steps: S1: The first vibration sensor and the second vibration sensor are arranged at intervals along the axial direction of the pipeline. The distance L between the two sensors satisfies L=2×D×(1+β), where D is the inner diameter of the pipeline and β is the adjustment coefficient and 0.1≤β≤0.

3. The vibration signal of the pipe wall is collected synchronously through the first vibration sensor and the second vibration sensor. S2: Bandpass filtering is applied to the acquired vibration signal. The filtered signal is then truncately within a preset time window T. For each time window, the energies E1(t) and E2(t) of the vibration signals acquired by the first and second vibration sensors within that time window are calculated. The time window T satisfies the following conditions: Where v max To achieve the expected maximum running speed of the sphere; S3: For each time window, calculate the instantaneous energy difference coefficient. Calculate the mean μ of δ(t) over N consecutive time windows. δ and standard deviation σ δ ; S4: Based on the mean μ δ Calculate the vibration source offset Where α is the pre-calibrated pipeline vibration attenuation coefficient; S5: When |Δx| is less than the first preset distance threshold D th1 And σ δ Less than the first statistical threshold σ th1 When |Δx| is greater than the second preset distance threshold D, it is determined as a ball-passing event; th2 And σ δ Greater than the second statistical threshold σ th2 When the event is determined to be an external impact event, the first preset distance threshold D is used. th1 =L / 6, the second preset distance threshold D th2 =L / 3, the first statistical threshold σ th1 With the second statistical threshold σ th2 Based on the standard deviation σ of pipeline background noise bg Perform adaptive calibration: σ th1 =λ1·σ bg , where 3≤λ1≤8: σ th2 =λ2·σ bg Where 10≤λ2≤25; and σ th1 Negatively correlated with the number of sliding windows N:

2. The method according to claim 1, characterized in that, When the medium inside the pipe is gas, the bandpass filter in step S2 has a frequency band of 0.5kHz to 2kHz; when the medium inside the pipe is liquid, the bandpass filter in step S2 has a frequency band of 1kHz to 5kHz.

3. The method according to claim 1, characterized in that, The calibration method for the attenuation coefficient α is as follows: apply a standard pulse excitation at a known location on the pipeline, record the energy ratio received by the two sensors, and then deduce the actual α value.

4. The method according to claim 1, characterized in that, In step S3, the value range of the continuous time window N is 5≤N≤20, and the total observation time N×T is less than 1 / 2 of the time required for the sphere to completely pass through the distance between the two sensors.

5. The method according to claim 1, characterized in that, When D thl ≤|Δx|≤D th2 At this time, enter the fuzzy area to assist in the determination: The total energy index E of continuously monitoring M sliding windows sum =E1+E2 and standard deviation σ δ The changing trend of M, where M is a preset positive integer and 2≤M≤5; If E sum It monotonically increases over M consecutive time windows, and σ δ If the number of consecutive time windows decreases monotonically, it is considered a ball-passing event. If E sum The increment within a single time window exceeds the preset energy mutation threshold ΔE th , and σ δ The increment within a single time window exceeds the preset variance mutation threshold Δσ th If so, it can be further determined to be an external shock event; Wherein, the energy mutation threshold ΔE th and variance mutation threshold Δσ th Adaptive calibration based on pipeline background noise level: Board th =the σ ·s bg in Let η be the mean of the total energy under background noise. E The preset energy mutation coefficient is 3≤η. E ≤10, η σ The variance mutation coefficient is preset and 5≤η σ ≤20.

6. The method according to claim 1, characterized in that, Before bandpass filtering the acquired vibration signal, step S2 further includes detrending and amplitude normalization of the acquired vibration signal. The detrending process uses a polynomial least squares fitting method to subtract the fitted trend term from the original signal. The amplitude normalization process uses a maximum-minimum normalization method to linearly map each channel signal to the [0, 1] interval.

7. The method according to claim 1, characterized in that, The vibration sensor is a piezoelectric vibration sensor or a MEMS accelerometer, with a sampling frequency of not less than 20kHz.

8. The method according to claim 1, characterized in that, The ball passing through the pipeline can be a pigging ball, a detection ball, a diameter measuring ball, or an intelligent detector.

9. The method according to claim 1, characterized in that, It also includes a confidence assessment step: based on the mean μ δ and standard deviation σ δ Calculate the signal-to-noise ratio (SNR). When the SNR is lower than a preset threshold, output the message "Signal is weak, cannot be determined".

Citation Information

Patent Citations

  • Oil and gas transmission pipe cleaner / detector infrasound real-time tracking and positioning system and method

    CN110107816A

  • Tracking and positioning system suitable for detector and pipe cleaner in oil and gas pipeline

    CN116774150A

  • Passive acoustic detection of pipeline pigs

    US5549000A