CN-121993749-A - Multi-source joint judgment method for pipeline leakage

CN121993749ACN 121993749 ACN121993749 ACN 121993749ACN-121993749-A

Abstract

The invention discloses a multi-source joint judging method for pipeline leakage, which relates to the technical field of fluid tightness test of structures, and comprises the steps of collecting an original sequence of an acoustic channel and a smoke temperature measuring point sequence, obtaining a sound velocity grading graph in a grading mode, intercepting noise baseline segments from the acoustic sequence, determining an ascending threshold value, recognizing a first arrival time sequence, intercepting time window segments to form a first arrival waveform segment set, extracting features to generate a path fingerprint table, determining a position lattice table by combining a thermal face arrangement graph, calculating propagation time to obtain an expected arrival ranking table, comparing the first arrival time sequence, screening out candidate position tables, combining the structure position table to generate an expected reverberation band table, comparing the expected reverberation band table with the path fingerprint table, and combining the ranking difference table to calculate candidate scores to give a positioning result and uncertainty.

Inventors

WU ZHEN
ZHANG HANG
Gong Shuxun
SHI ZHENGBIN
Yuan Cixian

Assignees

国能常州发电有限公司
国能常州第二发电有限公司

Dates

Publication Date: 20260508
Application Date: 20260403

Claims (9)

1. The multi-source joint judging method for the pipeline leakage is characterized by comprising the following steps of: Collecting an original sequence of an acoustic channel and a smoke temperature measuring point sequence, and grading the smoke temperature measuring point sequence to obtain a sound velocity grading diagram; Segment interception is carried out in an original sequence of the acoustic channel to obtain a noise baseline segment, an ascending threshold value is determined according to the noise baseline segment, the moment when the ascending threshold value is exceeded for the first time is taken as a first time sequence, and a first time waveform segment set is obtained by intercepting time window segments around the first time sequence; The method comprises the steps of taking a noise baseline segment as a reference, extracting features, generating a path fingerprint table, determining a position lattice point table according to a pre-acquired thermal surface arrangement chart, calculating propagation time for the position lattice point table according to a sound velocity grading chart and the pre-acquired sensor position table, sorting to obtain an expected arrival sorting table, comparing the expected arrival sorting table with a first arrival time sequence to obtain a sorting difference table, and screening by the sorting difference table to obtain a candidate position table; Generating an expected reverberation band table based on the candidate position table and a pre-acquired structure position table, comparing the expected reverberation band table with the path fingerprint table, calculating by combining the sorting difference table to obtain candidate scores, generating a positioning distribution table according to the candidate scores, and giving a positioning result and uncertainty by the positioning distribution table.
2. The multi-source joint judgment method for pipeline leakage according to claim 1, wherein the step of grading the sequence of the flue gas temperature measuring points to obtain a sound velocity grading graph comprises the following steps: setting temperature grading intervals to divide a flue gas temperature measuring point sequence into a plurality of temperature intervals, endowing sound velocity grades, and setting a sound velocity meter to determine sound velocity values of the sound velocity grades according to the central temperature of the temperature intervals; And calculating the spatial distance between each grid point of the position grid point table and each flue gas temperature measuring point, setting a first adjacent radius, screening an adjacent measuring point set, determining weights according to the spatial distances, weighting and rounding the sound velocity grades to obtain grid point sound velocity grades, and executing graded updating according to graded updating intervals to obtain a sound velocity graded graph.
3. The method for multi-source joint judgment of pipeline leakage according to claim 1, wherein the method for obtaining the noise baseline segment comprises the following steps: Defining a search time length at the front section of an original sequence of an acoustic channel by taking a unified time scale as an index, calculating an absolute value sequence of a waveform sampling value for each channel in the search time length, and averaging the absolute value sequence of the waveform sampling value according to the unified time scale to obtain a mean value amplitude sequence; Determining a low-energy threshold through a preset low-energy position dividing in the average amplitude sequence, screening continuous time periods when the average amplitude sequence does not exceed the low-energy threshold to obtain a baseline candidate time period set, determining a pre-event fragment according to the length of a first time window, and intercepting to obtain a noise baseline fragment.
4. The method of claim 1, wherein the step of determining the rise threshold comprises: Depolarizing the noise baseline segment and extracting the maximum amplitude; Setting an excess coefficient, wherein the excess coefficient is used for representing the excess proportion of the rising threshold value relative to the maximum amplitude; Determining an excess according to the maximum amplitude and the excess coefficient, wherein the excess is the amplitude excess obtained by amplifying the maximum amplitude according to the excess coefficient; the rise threshold is the level of superposition of the maximum amplitude and the excess.
5. The method for multi-source joint judgment of pipeline leakage according to claim 1, wherein the method for obtaining the first-reaching waveform segment set comprises the following steps: for each acoustic channel, taking the first arrival time of the acoustic channel in the first arrival time sequence as the center, and carrying out segment interception in the original sequence of the acoustic channel; Setting a second time window length, wherein the segment interception adopts a mode of front-back symmetry at the first time, the front section length and the rear section length are half of the second time window length, and the intercepted continuous sampling segment is recorded as a time window segment; And repeatedly intercepting the segments of each acoustic channel, and recording time window segments according to the channel number and the first arrival time in an associated mode to obtain a first arrival waveform segment set.
6. The method of claim 3, wherein if the baseline candidate period set has a continuous period with a length not less than the first time window length, the continuous period that satisfies the length condition at the earliest is taken as a pre-event segment, and if the baseline candidate period set has no continuous period with a length not less than the first time window length, the longest continuous period in the baseline candidate period set is taken as a pre-event segment, and the first time window length is taken as the length of the longest continuous period.
7. The multi-source joint judgment method of pipeline leakage according to claim 3, wherein the maximum amplitude of a base line is extracted from a base line segment of noise, a preset amplitude upper limit is determined based on the maximum amplitude of the base line, a unified time mark of ending of a segment before an event is taken as a starting point, the base line segment is searched forward along an original sequence of an acoustic channel, a unified time mark position where the absolute value of a waveform sampling value of any channel exceeds the preset amplitude upper limit for the first time is found and is recorded as an event reference time mark, each channel segment with the length of a first time window before the event reference time mark is intercepted again as the segment before the event, and the noise base line segment is updated.
8. The method for multi-source joint judgment of pipeline leakage according to claim 5, wherein the method for obtaining the path fingerprint table comprises the following steps: for each time window segment in the first-time waveform segment set, intercepting a time window segment starting segment, wherein the starting segment length is limited by an early segment length; Calculating average energy from the absolute value of the waveform sampling value of the noise baseline segment, and normalizing the average energy according to the unified time scale length of the noise baseline segment to obtain the average energy of the noise baseline; Calculating average energy from the absolute value of the waveform sampling value of the initial segment, and normalizing the average energy according to the unified time scale length of the initial segment to obtain the average energy of the initial segment; Recording peak time marks by taking the first arrival time in the first arrival time sequence as a zero point, and obtaining a peak interval sequence according to adjacent peak time mark differences; Extracting a frequency band energy column for each time window segment in the first-time waveform segment set; The path fingerprint table includes channel number, early energy ratio, peak interval sequence, band energy column.
9. The method of claim 8, wherein the method for extracting the band energy array comprises: Setting a frequency band boundary, wherein the lower boundary of the frequency band boundary is 0 hertz, and the upper boundary is half of the sampling rate of the original sequence of the acoustic channel; Dividing 0 Hz to an upper bound into a plurality of first frequency bands according to the frequency band boundaries; calculating discrete frequency components of the time window segments according to the sampling point sequences corresponding to the second time window length; Selecting discrete frequency components falling into the boundary of each frequency band, and accumulating the square of the amplitude of the discrete frequency components to obtain the frequency band energy of the frequency band; the band energies are recorded in the order of the bands, and a band energy train is obtained.

Description

Multi-source joint judgment method for pipeline leakage Technical Field The invention relates to the technical field of fluid tightness test of structures, in particular to a multi-source joint judgment method for pipeline leakage. Background In the operation and maintenance process of a boiler of a thermal power plant, real-time monitoring and positioning of pipeline leakage in a heating surface area are important links for guaranteeing stable operation of a unit. The regional pipeline bundles are densely distributed, the periphery surrounds various components such as furnace walls, steel structures and the like, acoustic signals generated by leakage are easily influenced by the surrounding environment and the structures in the propagation process, and complex challenges are brought to signal receiving and positioning analysis. At present, the leakage positioning of the area usually adopts a scheme of combining an acoustic sensor array with an arrival time difference positioning technology, and the technology is also a mainstream application mode of acoustic positioning in industrial scenes. The core principle of the arrival time difference positioning technology is that acoustic signals generated by a leakage source are synchronously received through a plurality of acoustic sensors arranged in a monitoring area, the time difference value of the signals reaching each sensor is calculated, a positioning equation set is constructed by combining the spatial position information of the sensors and preset sound velocity field distribution parameters, and the spatial coordinates of the leakage source are determined by solving the equation set. The positioning accuracy of the technology depends on the stability of a signal propagation path, the accuracy of time difference calculation and the uniqueness of equation set solving, and can realize more reliable positioning output under the scene of single signal propagation path and less interference of peripheral structures. However, in the practical application scenario that the boiler heating surface tube bundles are dense, the positioning reliability requirement is difficult to fully meet by the technology, and the core problem is caused by multipath propagation and superposition effects of acoustic signals. The acoustic signal generated by leakage is not only transmitted to the sensor through a single direct path, but also contains direct waves, diffracted waves formed by diffraction through a tube row gap, and reflected waves formed by reflection of a furnace wall and a steel structure, and various waveforms are mutually overlapped in the transmission process and finally synchronously received by the sensor. Because of the complexity of the spatial structure of the dense region of the tube bundle, the leakage sources at different spatial positions can generate nearly identical arrival time difference combinations in the current sensor layout mode and the sound velocity field environment under the current working condition, namely, the equal time difference pseudo solution is formed. Meanwhile, the change of the unit operation load can cause the fluctuation of a flue gas temperature field, the temperature change can directly change the sound velocity distribution, and further the signal propagation path and the time difference calculation result are influenced, so that the space position of an equal time difference false solution drifts along with the change of the operation working condition, different positioning results of the same leakage source are calculated in different operation time periods, and the positioning results are often deviated to adjacent tube rows or heating surface areas. The influence of the multipath superposition and the pseudo solution drift is not fully considered in the prior art, on the premise of not increasing the number of sensors, an effective direct wave and multipath wave separation means is lacked, equal time difference pseudo solutions cannot be accurately identified and removed, the multi-peak property and uncertainty of a positioning result are difficult to quantitatively characterize, a single equation set solving result can be only relied on as a positioning basis, and reliability evaluation references of the positioning result cannot be provided for a joint judgment process. Due to the existence of the problems, the leakage positioning result is deviated from the actual leakage position, and the positioning accuracy cannot be effectively ensured. The same leakage source is positioned to different areas in different time periods, so that judgment of the leakage position by monitoring staff can be interfered, and difficulty and time consumption of leakage investigation are increased. The method has the advantages that the positioning result with false solutions is directly used as an effective basis in the joint judgment process due to the lack of quantitative output of uncertainty of the positioning result, so that unneces