RU2584721C1 - Passive-active acoustic method of detection and localisation of gas leaks in gas-liquid medium - Google Patents

Abstract

FIELD: measurement technology.

SUBSTANCE: invention relates to hydro acoustics, particularly to means of detecting leaks. Method comprises reception and registration of ambient signal acoustic noise in frequency range of appropriate intrinsic frequencies of pulsations of bubbles in liquid, breaking down signal to sub-bands, filtration, calculation of spectra of and plotting spectrograms. At that, spectral components are recovered with exponentially decays amplitude, determining frequency f_R, filtration considering frequency f_R filter, treatment time signal, amplifying and radiation repeated reception and band-pass filtering ambient signal acoustic noise with central frequency filter f_R and selection in it scattered bubble signals by calculating noise signal spectrum and plotting spectrograms. Detection of gas leaks are recorded at first appearance in spectrograms of spectral components of pulsed signals with symmetric exponentially increasing attenuation in time and amplitude with duration of periods of the sound field twice exceeding the duration of emission pulse, and filling frequency f_R, and localisation of point of gas outlet is made using triangulation method.

EFFECT: higher efficiency, reliability and accuracy of control.

1 cl, 1 dwg

Description

The invention relates to the acoustics of gas-liquid media, specifically to methods that can detect, identify and determine the position of the resonant sources of emission radiation of sound in a liquid and can be used in the oil and gas industry to control the tightness and detect gas leaks from gas pipelines and technical systems for the production of hydrocarbons under water, chemical industry - in relation to gas-liquid flows, in the power industry - for the diagnosis of processes occurring in the bubble modes boiling, in medicine - in the systems of visualization of bubbles in the body and physiotherapeutic methods of treatment, in ecology and environmental monitoring - for localization and monitoring of natural gas sources (vultures) at the bottom.

One of the main objects of emission radiation in a liquid medium is gas bubbles. The determination of the presence of bubbles and their number in the near-bottom region and in the thickness of the ocean is necessary for estimating gas reserves as a fuel and predicting the flow of methane from the seabed to the atmosphere [Judd A.G. The global importance and context of methane escape from the seabed // Geo-Mar. Lett. 23, 2003, p. 147-154]. The UN proposed program to reduce greenhouse gas emissions into the atmosphere provides an impetus to the development of methods for long-term monitoring of gas flows from the bottom of the oceans.

The importance of the prompt detection and elimination of gas leaks from underwater storage facilities and gas pipelines is that, undetected, such leaks can lead to huge financial losses and environmental disasters. This is relevant due to the fact that the infrastructure for gas production and transportation is rapidly aging in marine conditions. Gas leaks in shelf areas and even in the open ocean threaten marine ecosystems and lead to disruptions in the carbon and methane gas exchange cycles in the ocean [Leigthon T.G. and White P.R. Quantification of undersea gas leaks from carbon capture and storage facilities, from pipelines and from methane seeps, by their acoustic emissions // Proc. R. Soc. London, Ser. A 468, 2012, p. 485-510].

Sources of acoustic emission are present in various chemical technological processes, for example, gas-liquid mixing, chemical reactions with gas evolution, etc. Acoustic emission measurement methods are used to control environmental parameters, including aggressive ones, and allow for the non-contact monitoring of physical and chemical processes in real time. [Boyd J.W.R., Varley J. The uses of passive measurement of acoustic emissions from chemical engineering processes // Chemical Engineering Science, 56, 2001, p. 1749-1767].

Acoustic methods for detecting, identifying and determining the position of sound emission sources of sound in a liquid medium can be passive, active and combined (active-passive).

In active methods, acoustic signals are emitted into the studied liquid medium, signals are propagated in a medium with reflection and refraction processes, then reception and registration of acoustic signals reflected and (or) refracted in the medium are carried out. A model of the propagation of acoustic signals in the medium is formulated, and within the framework of the formulated model, the recorded acoustic signals are analyzed and processed. Based on the results of signal processing, emission sources are identified and their coordinates in space are calculated.

Passive acoustic methods include registration of emission radiation of bubbles coming from the places of their formation in case of leakage or destruction of technical systems or from the places of exits of natural vultures at the bottom of the sea. Within the framework of the existing models of acoustic emission, the processes mentioned above, the registered acoustic signals are analyzed and processed, and the emission objects are identified, the exit points of the bubbles are localized, and the required parameters of gas flows are determined.

A passive acoustic method for detecting leaks of gas-liquid flows from subsea pipelines includes receiving an acoustic signal from the area surrounding the subsea pipeline, transmitting the mentioned signal to a recording device, and subsequent processing of the received signal (see WO 02/025239 A1, IPC G01M 3/24). The signal is received by a hydrophone with a radiation pattern determined by the internal pressure in the pipeline. The received signal enters the signal reception and transmission unit, from which it is transmitted via cable to the signal processing system installed on the vessel. In the case of inspections of pipelines using a towed underwater vehicle, the cable is not used, and signals are transmitted to the surface using the standard cable of the device. At the processing stage, the registered signal is filtered in the frequency band 30 kHz - 70 kHz. The quadrature components of the signals are determined having operating frequencies in the range of sound (0-20 kHz) frequencies. An analog-to-digital signal conversion is performed and the operator performs audio monitoring of the process. Within the limits of the audio frequency band, calculation and graphical display of the spectrogram and signal amplitude on the monitor are performed every second. At the same time, a calculated graph of the signal power as a function of frequency and time is displayed on the monitor. Using the signals from the sound device and graphics on the monitor, the operator detects leak signals.

The disadvantages of this method are:

- low accuracy of localization of the acoustic emission source due to the movement of the receiver and the need for positioning in the space of the underwater measuring unit;

- low sensitivity of the method, depending on the level of ambient noise due to both the movement of the acoustic signal receiver itself and the operation of the escort vessel mechanisms;

- low efficiency and fundamental difficulties in the localization of pulsating and "sleeping" sources of emission signals due to the uncertainty of the duration, places and moments of the beginning of gas outflows at the bottom.

In addition, it should be noted that the known method does not allow monitoring over time and determining the amount of gas-liquid flow exiting the source.

Combined active-passive acoustic methods use combined or different sensors for receiving and emitting acoustic signals, which can be combined into antennas (linear flat or surround).

A combined active-passive acoustic method is proposed based on the measurement data of backscattering of pulsed signals and recording acoustic emission using a reversible multipath sonar with high spatial resolution SEABAT 7128 [Wendelboe G., Barchard SM, Maillard E., Bjørnø L. High-resolution multibeam sonar for subsea leakage detection // Proceedings of the 11th European Conference on Underwater Acoustics (ECUA 2012: Edinburgh UK, 2nd-6th July, 2012), p. 894-901]. The compact sonar antenna consists of 256 hydrophones and has a directivity pattern of 1280 horizontal and 280 vertical with a resolution of 0.5 °. The method and equipment made it possible to visualize small gas leaks in the active mode at a frequency of 400 kHz at distances up to 200 meters. In the mixed active-passive acoustic mode, identification, localization and determination of gas exit points were carried out under controlled conditions at distances up to 70 meters.

The disadvantage of the proposed method is the complexity of technical implementation and work only in the standard linear mode, using one (high) frequency when emitting and receiving acoustic signals, which limits the detection range of gas leaks due to reflection, scattering, and high attenuation of sound in inhomogeneous media.

A known method of acoustic research in the environment in order to detect objects whose acoustic impedance differs from the environmental impedance, based on the concept of time reversal [Fink M. Method and apparatus for acoustic examination using time reversal // US Patent 5428999, July 4, 1995]. The method includes emitting a diverging sound beam into the medium using at least one reversible transceiver. An echo reflected in the medium is received by an antenna from several reversible transducers. Using a given time window, echo signals from specific areas of the medium are extracted. Received echoes are stored, inverted over time and re-emitted to the medium. The signals newly reflected in the medium are again stored, and the time reversal operation and the emission of new signals are repeated. After the final conservation and emission of a time-reversed signal of rank 2m + 1 (where m is a positive non-zero integer), the nature of the wavefront for the entire time of research is determined in the form of the distribution of the time of the signal maximums on all transceivers. The indicated dependence can be approximated using the polynomial law.

The main disadvantage of using this method with the aim of detecting, identifying and localizing resonant sources of emission radiation in the form of resonant bubbles is the difficulty in determining the filling frequency and radiation times for backlight pulses at the first passive stage of the method implementation. This drawback is due to the fact that the processes of formation of gas bubbles are random in time, that is, it is a priori impossible to set the start time of the backlight. It is also impossible to set the fill frequency of the backlight pulses in advance, which should correspond to the resonant frequency of the natural oscillations of the bubble. This is due to the fact that the signal scattered by the bubble will have an amplitude above the level of ambient noise only at the frequency of the own resonance of the bubble, determined by its radius. The emission of backlight pulses with a high filling frequency and a high periodicity in time does not solve the problem because of the very small amplitude of non-resonant scattering of high-frequency pulse signals by bubbles.

Closest to the claimed method is a method for detecting and / or determining the position of a reflective sound source (US patent No. 6161434, G01N 29/14). This method allows you to detect reflective sound sources and determine the position of sound sources, including at sea, by receiving and recording ambient acoustic noise signals in a controlled area, reversing in time noise signals, amplification and emission of reversed signals, subsequent reception, recording and analysis of reflected acoustic signals in order to detect the possible presence of a reflective noise source.

The method for detecting gas leaks from gas pipelines passing along the bottom is implemented by receiving an ambient acoustic noise signal in the frequency range 3 Hz - 200 kHz, in which the presence of acoustic signals emitted by a cloud of bubbles emerging from the pipeline is assumed. Electrical signals corresponding to ambient acoustic noise signals are recorded, reversed in time, and emitted by combined transceivers. Propagating in a medium, signals are scattered both by a cloud of bubbles, and by other objects of the medium having impedance jumps at the water-object boundary, such as, for example, the bottom and objects at the bottom, a gas pipe, biological objects, etc. The reflected signals are received, recorded and analyzed to detect and isolate among all the registered signals reflected from the boundaries with jumps in impedance, the signal reflected by a cloud of bubbles. Subsequent analysis and classification of the received reflected signals is performed, in particular, by the presence of reflected pulse signals in the registered signals. If pulsed signals are absent or not visible among the noise in the received reflected signals, use the calculation of cross-correlation functions between the signals at different receivers. Additionally, if the signal from the emission source is not distinguished from the background noise in the reflected signal by the specified method and, therefore, is not detected, the process of inverting the received reflected signal and its radiation is repeated, adding one or more iterations of the process, increasing the sensitivity of the method in this way.

A prerequisite for implementing the method for identifying and localizing a reflective emission radiation source, which is a gas leak from a gas pipeline lying at the bottom, is that the reflected sound signals are in the same propagation conditions as the sound emission signals received during the initial passive stage. This is true if the time of each measurement cycle is less than the time during which defocusing of the inverted signal takes place, which may not be possible for pop-up bubbles in the cloud or for technical systems using moving receiving and emitting antennas.

In addition, the disadvantages of the known method in terms of detecting and determining the position of gas emission bubbles in a liquid medium are:

- the difficulty of identifying the event as a result of a low signal level of the total emission radiation of the bubbles and a high level of ambient noise in the operating frequency range (signal-to-noise ratio of about 1), which makes it difficult to detect a signal in the initial passive stage,

- due to the attenuation during propagation, a part of the energy of the direct emission signal and the signal components reflected from the boundaries are lost. In this case, the time reversal process becomes ineffective due to the low quality of the time reversed signal (a short useful signal with an amplitude above the noise level, i.e., a small part of the signal energy emitted by the bubble is focused), which leads to

- reducing the amplitude and duration of the signal reflected by the bubble and thereby reduce the probability of detection and the accuracy of localization of emission bubbles;

- a high probability of identification errors of the emission object at both the passive and active stages of detection due to the small number of signs for making a decision;

- the complexity of real-time measurements in view of the large volume and the necessary computation speed for detecting an emission signal for natural sipes or in anthropogenic gas leaks due to the random nature of the formation of bubbles;

- low efficiency and fundamental difficulties of localization of pulsating and "sleeping" sources of emission signals due to the uncertainty of the duration, places and moments of the beginning of gas outlets;

- it should also be noted that the known method does not allow long-term monitoring over time and to determine the amount of gas-liquid flow exiting the source.

This method, like the methods listed above, is not taken into account in the processes of identification and localization of emission radiation sources in the form of gas bubbles of the resonant nature of the radiation and sound scattering by bubbles coming out from technical leaks and in natural conditions.

The objective of the invention is the development of an effective remote method for passively active detection, identification and localization of natural gas outlets accidental in time or accidental gas leaks from technical production systems into the aquatic environment, as well as for monitoring natural gas outlets under water.

The technical result is efficiency, increasing the probability and reliability of identification of the emission object and reducing the number of false alarms in case of leakage or destruction in the control area, increasing the accuracy of determining the locations of gas-liquid flow outlets, as well as determining the quantitative parameters of gas flows in a wide range of bubble concentrations with the possibility of monitoring the studied processes in time.

The problem is solved by the proposed passive-active acoustic method for detecting and localizing gas leaks in a gas-liquid medium by real-time reception and recording in the controlled area of the signal of ambient acoustic noise in the frequency range from 0.1 kHz to 5 MHz, corresponding to the frequencies of the natural pulsations of the bubbles in liquid, dividing the frequency range into at least two frequency subbands and performing bandpass filtering of the signal in the selected frequency subbands, and spectra for signals received in each frequency subband and constructing the corresponding spectrograms allocation in the spectrogram of spectral components having an exponentially decaying with time the amplitude and duration of 5 to 30 periods, for determining the frequency f _R for the selected spectral component corresponding to the resonance frequency of the bubble ripple , and obtaining a working signal by bandpass filtering of the ambient acoustic noise signal with a center frequency f _R filter, time reversal The measured working signal, its amplification and emission of the corresponding acoustic signal, the repeated reception and band-pass filtering of the ambient acoustic noise signal with a central filter frequency f _R and the selection of pulse signals scattered by the bubble in it by calculating the spectrum of the noise signal and constructing spectrograms, highlighting the spectral components in the obtained spectrograms pulse signals having a symmetric exponentially increasing and decaying in time amplitude with a duration of the periods of the sound field, in two and times the duration of the emission pulse, and the filling frequency f _R and the subsequent identification of the gas leak by the first occurrence of the given signal in the spectrograms, and localization of the gas outlet is performed by the triangulation method by calculating the distances from the bubble to the receiving sensors from the measured propagation times of the pulses in the emitted and scattered signals and the known speed of sound in a liquid.

If necessary, the parameters of the outgoing gas stream are determined by calculating the number of outgoing bubbles in time and summing their volumes.

The inventive method is based on the resonant nature of the acoustic radiation of a bubble at its birth, which allows the beginning of gas output to be determined by the first bubble, which is separated from a solid surface and emits an emission signal whose properties are known. The detection of this signal, time reversal, re-emission, and reception of a signal resonantly scattered by the bubble makes it possible to identify the emission signal and determines the reliability of assigning the received signal to the emitted gas bubble, based on its uniqueness as a resonant system with high quality factor and natural resonance frequency determined by the radius of the bubble and parameters gas and surrounding liquid, the speed of detecting the onset of a leak or natural gas outlet, as well as high localization accuracy and mon Iteration of these processes in time.

It is known that the resonance properties of a bubble manifest themselves in the form of bubble pulsations with a natural resonance frequency in the frequency range from 0.1 kHz to 5 MHz with an exponentially decaying amplitude in time and a duration of 5 to 30 periods [Leighton, T.G. From seas to surgeries, from babbling brooks to baby scans: the acoustics of gas bubbles in liquids // Int. J. Mod. Phys. B, V. 18, p. 3267-3314, 2004].

A pulsating bubble emits an acoustic pulse signal, the amplitude and frequency of which corresponds to the type of pulsation of its surface. An emission pulse signal propagates in an inhomogeneous medium, being refracted and reflected at phase boundaries. The detection of pulse emission signals and the determination of their parameters in the received signals, in contrast to the prototype method, in the inventive method is performed at the first passive stage of the method. To detect a pulsed emission signal having a low level relative to the level of acoustic noise, in the inventive method, the frequency range of the measurements is divided into at least two equal subbands and band-pass filtering of the received signal for the frequencies of each subband is performed. This operation is performed in real time and is applied to all received and recorded signals. In parallel, the spectra and spectrograms are calculated for the signals obtained by filtering. If a spectrogram is detected by any suitable method, for example, by an operator or using automatic instrument discrimination and numerical analysis, a pulse signal with known (exponentially decaying in time amplitude and duration from 5 to 30 periods) properties of the emission pulse signal from a gas bubble at the moment of separation from surface, determine the time of arrival of the pulse and its filling frequency f _R. Select a processing time window for analysis of the emission signal corresponding to the propagation time of the acoustic signal over the distance (L) to the possible places of leakage and the known average statistical speed of sound (s) in the measurement area. The signal in the selected time window is filtered using a band-pass filter with a central frequency corresponding to the filling frequency f _{R of the} detected pulse signal, which allows to reduce the noise level, significantly reduce the frequency band of the signal and increase the ratio "amplitude of the emission signal" / "noise" at the pulse filling frequency , which in turn significantly increases the efficiency of the method at the active stage due to the use of a practically tonal inverse signal. The electrical signals received from a selected time window and filtered by a band-pass filter from each receiving sensor are reversed in time, amplified, and corresponding acoustic signals are emitted to the medium using emitting sensors. Due to the physical effect of time reversal, the emitted acoustic signals propagate along the same paths as the received emission pulses and focus on the emission bubble. Due to the narrow band of the emitted signal, the focusing effect and amplification of the reversed signal, the amplitude of the incident acoustic field in the region of the bubble is much greater than the level of ambient noise, and the acoustic field resonantly scattered by the bubble back into the medium has an amplitude (and signal-to-noise ratio) much larger than for primary emission radiation of the bubble.

Acoustic signals scattered on the bubble are received, recorded and filtered using a band-pass filter with a central frequency corresponding to the frequency f _{R of the} emitted signal, spectra and spectrograms are determined and displayed for all signals obtained by filtering from the receiving sensors. The analysis of spectra and spectrograms is carried out in order to identify the emission object (by the operator or using automatic instrument discrimination or by numerical analysis algorithms) by comparing with the known properties of the emission signal from a gas bubble, the propagation time for emitted and scattered pulses is determined and their filling frequency is refined. At the end of the identification, the place of exit of the gas bubble is localized according to the measured propagation times of the pulsed signals using angular and remote triangulation techniques and, if necessary, the parameters of the outgoing gas flow are determined by calculating the number of outgoing bubbles in time and summing their volumes.

Thus, at the active stage of the method implementation, two physical effects are used - focusing the signals when using the time reversal of the received signals and resonant scattering of the acoustic signals by gas bubbles.

In especially important cases, the identification and verification of the emission object can be additionally carried out by the exclusion method, namely, by emitting a time-reversed signal at a frequency different from the frequency of the emission signal. The absence of a scattered signal (echo) will confirm the presence of a source of emission radiation other than the emerging gas bubble, for example, biological, generated by the inhabitants of the sea (fish, crayfish, marine mammals, etc.) [Furduev A.V. Noises of the sea. In the book. under the editorship of L.M. Brekhovsky. Acoustics of the ocean. - M: Science, 1974].

A higher identification reliability and accuracy of the localization of the leakage site in the proposed method is achieved, firstly, due to the time synchronization of the events of the beginning of the gas leak, characterized by separation of the bubbles from the solid boundary and the beginning of the emission of acoustic radiation by the bubbles, and, secondly , the ability to accurately determine the parameters of the reflected signal from the bubble, used to determine the exit point and volume of the outgoing gas, due to the focus of the reverse wave and the effective scattered ia, as a result of the resonant nature of the pulsations of the bubbles.

Detection, localization and continuous monitoring over time in order to identify gas leaks and emissions of gas bubbles can be realized using active-passive stationary sonar stations like Dniester EM and MGK-607EM [Encyclopedia “Russia's Arms and Technologies. XXI Century ”, vol. 3, Armament of the Navy, 2001, 632 pp.]. For processing acoustic noise signals and generating radiation signals, one can use, for example, real-time control computers such as NVIDIA TESLA AND QUADRO GPUS [accelerations.html] with the possibility of implementing parallel multiprocessor calculations based on specialized NVIDIA devices (> 3000 processors), when processing signal information [time / supported / modular-real-time-target-machine.html].

In FIG. schematically shows the type of signals and the sequence of actions during the implementation of the invention, where (a) is the pulse (emission signal) generated by the gas bubble at its birth, (b) are the time-reversed signals emitted by the linear transmitting and receiving antenna, (c) is the pulse generated by a bubble under the action of a time-reversed wave, (d) - vertical receiving and transmitting antenna, (e) - gas bubble that is born at the leakage site, (f) - bottom of the reservoir, L - horizontal distance from the source to the antenna, h / 2 - distance between individual mi receiving and transmitting elements of the antenna.

At the first passive stage of the implementation of the method, acoustic signals are continuously received in a given operating frequency range, from a possible range of 0.1 kHz to 5 MHz, using receiving-emitting sensors (receivers) that form an antenna that is stationary located in the measurement area, for example, in near-bottom area for monitoring gas vultures, or in the area where mining or gas transmission systems are located. It is known that the type of emission signal (a, Fig.) Emitted by a bubble at its birth is described by the expression (1)

. Излученный пузырьком эмиссионный сигнал распространяется в среде, испытывая отражения от межфазных границ и преломления на неоднородностях импеданса. Прошедший через среду сигнал принимается акустическими датчиками, а соответствующие аналоговые электрические сигналы s_n(t), (n=1,2…N - число датчиков в антенне) преобразуют в цифровую форму и выполняют регистрацию цифровых сигналов в запоминающем устройстве (ЗУ). Определяют рабочий частотный диапазон для анализа сигналов, соответствующий рабочему частотному диапазону приемопередатчиков. Производят разбиение указанного рабочего частотного диапазона как минимум на два поддиапазона. Количество поддиапазонов может быть больше двух и определяется видом спектра шума в точке приема. При высоком уровне шума увеличение числа поддиапазонов уменьшает частотную полосу анализа и снижает уровень шума для указанной полосы. Определяют временной интервал Т_ов окна для анализа сигналов по формуле Т_ов=2×L/с, где L - требуемая максимальная дальность обнаружения выхода газа, С - скорость звука в воде в районе измерений. С помощью устройства выборки (УВ) в момент времени t₀ производят первую выборку цифровых сигналов из ЗУ для всех приемных датчиков на интервале анализа Т_ов и осуществляют полосовую фильтрацию сигналов с каждого приемного датчика в каждом частотном поддиапазоне для временного интервала Т_ов с помощью полосовых пропускающих фильтров (ПФ). Далее, с помощью вычислительного устройства (ВУ) выполняют спектральный анализ сигналов со всех приемных датчиков на временном интервале Т_ов в рабочих частотных поддиапазонах. Временное окно Т_са спектрального анализа сигналов выбирают с учетом длительности t_э эмиссионного излучения пузырьков t_э=2QT_*, где Q - величина добротности собственных колебаний пузырька имеет значения в диапазоне от 5 до 30, Т_* - период собственных колебаний пузырька, связанный с собственной (резонансной) частотой пульсаций пузырька Ω_*=2π/T_*, которая определяется известным способом с помощью приведенных в работе [Maksimov А.О., Burov В.A., Salomatin A.S., and Chernykh D.V. Sounds of marine seeps: A study of bubble activity near a rigid boundary // J. Acoust. Soc. Am. 136, 2014, pp. 1065-1076] выражений (2)here R ₀ is the radius of the bubble, Ω _* = 2πf _R is the circular frequency of natural vibrations of the bubble, β _tot = β _th + β _υ + β _r is the total attenuation coefficient due to heat and viscous losses, as well as radiation, c is the speed of sound , p ₀ - pressure amplitude, (in laboratory experiments, this pressure at a distance of 10 cm from the source is 20-40 Pa) [Leighton T.G. The Acoustic Bubble. Academic Press, London, 1994, 613 p]. The duration of the pulse of emission radiation during the exponential decay of the pulsations of the bubble is

. The emission signal emitted by the bubble propagates in the medium, experiencing reflections from interphase boundaries and refraction at impedance inhomogeneities. Passed through a medium signal is received by acoustic sensors and the corresponding analog electric signals _{s n (t), (n} = 1,2 ... N - the number of sensors at the antenna) is converted to digital form and perform the recording of digital signals in a storage device (memory). The operating frequency range for signal analysis corresponding to the operating frequency range of the transceivers is determined. The specified operating frequency range is divided into at least two sub-bands. The number of subbands may be more than two and is determined by the type of noise spectrum at the receiving point. At high noise levels, increasing the number of subbands reduces the analysis frequency band and reduces the noise level for the specified band. The time interval T _{s of the} window is determined for signal analysis according to the formula T _s = 2 × L / s, where L is the required maximum range for detecting a gas outlet, C is the speed of sound in water in the measurement region. Using the sampling device (HC) at time t ₀ , the first sampling of digital signals from the memory is made for all receiving sensors in the analysis interval T _s and band-pass filtering of signals from each receiving sensor in each frequency subband for the time interval T _{s is} performed using bandwidth transmitters filters (PF). Further, using a computing device (WU), spectral analysis of signals from all receiving sensors is performed on the time interval T _s in the working frequency subbands. The time window T _sa spectral analysis of signals selected in accordance with the duration t _e emission radiation bubbles t _E = 2QT _*, wherein Q - value Q of the natural oscillations of the bubble has a value in the range from 5 to 30, T _* - the natural period of oscillation of the bubble associated with its own the (resonant) bubble pulsation frequency Ω _* = 2π / T _* , which is determined in a known manner using [Maksimov A.O., Burov V.A., Salomatin AS, and Chernykh DV Sounds of marine seeps: A study of bubble activity near a rigid boundary // J. Acoust. Soc. Am. 136, 2014, pp. 1065-1076] expressions (2)

where Ω _{* 0} Ω and the natural frequencies of the pulsations bubbles distance h _b from the solid border and free of bubbles, respectively, γ - polytropic exponent, P ₀ - hydrostatic gas pressure in the _bubble, ρ _w, - density and H - depth of the water layer. When calculating the resonant frequency Ω _* in expression (2), the effects of the interaction of a pulsating bubble with solid boundaries (walls of gas pipelines, siph exit channels in the bottom, etc.) in the gas exit region were taken into account. The calculation of the coefficients β _υ and β _r is performed according to the formulas:

указанных сигналов, превышающих по амплитуде заданное пороговое значение A,

и амплитуда которых затухает во времени по экспоненциальному закону $$S_{n{\Omega }_{*}}\ast \exp (-{\beta }_{tot}t)$$

. Прием акустических сигналов, регистрация электрических сигналов и их спектральная обработка путем ЦОС, проводятся в режиме реального времени.where v is the coefficient of viscosity of the liquid, the length of the viscous wave, F _v (k) is the factor determining the effect of hard boundaries on the attenuation due to viscosity. The damping of pulsations of gas bubbles in water due to the thermal conductivity β _th under normal conditions can be neglected in comparison with the effects of β _υ and β _r . The calculated spectra for all signals are stored in the memory, and the spectral analysis procedure is repeated for all signals at time t ₀ +1, t ₀ +2, etc. and use the appropriate spectra to construct spectrograms of signals from all receiving sensors and for all frequency subbands. The detection of the emission signal in the passive measurement stage is carried out by analyzing the spectra and spectrograms from all receiving sensors. The criterion for detecting the bubble emission signal and setting the start of the active stage of identification and localization of the possible gas output will be the appearance and detection by the operator of the pulsed narrow-band signal with an exponentially decaying amplitude on the spectrogram (spectrograms), and the decay time of the amplitude of the discrete component corresponds to 5-30 periods of signal pulsation in time. In the case of the use of automatic instrument discrimination and numerical analysis to detect the emission signal of the bubble and set the start of the active stage of identification and localization of the leak or gas outlet, the digital signal processing (DSP) process is performed. DSP includes determining characteristics of a pulsed signal: frequency filling Ω _*, the pulse duration, β _t0t - total attenuation coefficient and the time of arrival τ _i pulse signals of the form (1), all the signals s _n (t) by determining the frequency components of the spectra

these signals exceeding the specified threshold value A in amplitude,

and whose amplitude decays exponentially in time $$S_{n{\Omega }_{*}}\ast \exp (-{\beta }_{tot}t)$$

. Acoustic signals reception, registration of electrical signals and their spectral processing by DSP are carried out in real time.

, представляют собой серию реплик исходного импульсного эмиссионного сигнала пузырька, так как прошедший через среду, зарегистрированный сигнал содержит отраженные от границ и преломленные в среде компоненты, которые могут накладываться друг на друга или быть разнесены во времени,

. Амплитуды частотных составляющих окружающего шума вне границ пропускания фильтра будут значительно (>60 dB) подавлены. Сигналы со всех приемных датчиков антенны после фильтрации обращаются во времени и усиливаются с коэффициентом, равным (в зависимости от требуемой дальности обнаружения утечки), например, K=10-1000 и каждый сигнал в момент времени

,

излучается соответствующим датчиком, в среду. Излучаемые, инвертированные во времени сигналы имеют вид (b, Фиг.). Результирующее поле обращенных во времени сигналов, в соответствии с базовым принципом обращения времени, фокусируется назад в точку положения источника. Резонансный характер пульсаций пузырька, как резонансной колебательной системы, приводит к резонансному рассеянию падающей на него обращенной волны. Проведенные авторами расчеты рассеяния обращенных во времени акустических сигналов (b, Фиг.) на резонансном пузырьке показали, что выражение для формы сигнала обратного рассеяния p_bs(t) пузырька может быть представлено в форме:. Upon detection of the received noise pulse signal corresponding to the mean emission signal emitted by the bubble (the expression (1) to perform filtering of the signals s _n (f) from all receive antennas of sensors in a time window T _s Signal filtering is performed by a bandpass pass filter - F (ω, Ω _* ) with a central frequency Ω _c = Ω _* . The signals obtained after filtering

, represent a series of replicas of the initial pulse emission signal of the bubble, since the signal passed through the medium contains the components reflected from the boundaries and refracted in the medium, which can overlap each other or be spaced in time,

. The amplitudes of the frequency components of the ambient noise outside the filter transmission boundaries will be significantly suppressed (> 60 dB). The signals from all the receiving sensors of the antenna after filtering are reversed in time and amplified with a coefficient equal to (depending on the required range of leak detection), for example, K = 10-1000 and each signal at a time

,

emitted by an appropriate sensor on Wednesday. Radiated, time-inverted signals have the form (b, Fig.). The resulting field of time-reversed signals, in accordance with the basic principle of time reversal, is focused back to the source position point. The resonant nature of the pulsations of the bubble, as a resonant oscillatory system, leads to resonance scattering of the incident wave incident on it. The authors performed calculations of the scattering of time-reversed acoustic signals (b, Fig.) By a resonant bubble showed that the expression for the backscattered signal waveform p _bs (t) of the bubble can be represented in the form:

(вид рассеянного сигнала (с), Фиг.).here φ = arctan (-Ω _* / β _tot ), R _t ~ 1 m is the parameter characterizing the emitted signal, and a = (h / 2) N is the effective aperture of the antenna. The amplitude of the signal scattered on the bubble linearly (with coefficient K) depends on the amplitude of the time-reversed and amplified signal, and the duration of the reflected pulse is twice as long as the duration of the emission pulse,

(view of the scattered signal (s), Fig.).

с датчиков будут, аналогично зарегистрированному эмиссионному сигналу, содержать реплики, но уже импульсного сигнала обратного рассеяния (4),

Зарегистрированные электрические сигналы

обрабатываются по методике, сходной с той, что была описана выше для сигналов s_n(t). Обработка включает в себя выделение рассеянных на пузырьке импульсных сигналов вида (с, Фиг. ) во всех полученных с приемных датчиков антенны электрических сигналах

путем определения частотных составляющих спектров

. Временной интервал (временное окно) T_cra спектрального анализа принимаемых датчиками рассеянного сигнала вида (4) выбирается в соответствии с длительностью рассеянного пузырьком обращенного эмиссионного сигнала (с, Фиг. ), которая в два раза больше длительности эмиссионного импульса (а, Фиг. ) T_cra=2×Т_са. Определение и уточнение характеристики отдельных реплик вида (4) (частота заполнения Ω_*, длительность импульсов, β_tot - коэффициенты затухания и времена приходов - t_pri) импульсных составляющих сигнала выполняются путем расчета и построения спектрограмм рассеянных сигналов для сигналов со всех приемных датчиков. Размер временного окна Т_р анализа принимаемого рассеянного сигнала теоретически должен соответствовать размеру Т_ов - окну анализа эмиссионного сигнала, но на практике может быть больше, Т_р≥Т_ов, за счет больших амплитуд рассеянных сигналов, лучшего соотношения сигнал/шум и большей «полезной» длительности каждого импульса, по сравнению с эмиссионными импульсами.The acoustic signal p _bs (t) scattered by the bubble propagates in the medium from the bubble to the antenna, experiencing reflections from the boundaries and refracting on the impedance inhomogeneities in the medium, and is received by the antenna sensors. Relevant electrical signals

from the sensors will, similarly to the registered emission signal, contain replicas, but already of a pulsed backscattering signal (4),

Registered Electrical Signals

processed according to a technique similar to that described above for signals s _n (t). The processing includes the selection of pulse signals of the form (s, Fig.) Scattered across the bubble in all electrical signals received from the antenna receiving sensors

by determining the frequency components of the spectra

. The time interval (time window) T _{cra of the} spectral analysis of the scattered signal received by the sensors of the form (4) is selected in accordance with the duration of the reversed emission signal scattered by the bubble (s, Fig.), Which is twice the duration of the emission pulse (a, Fig.) T _cra = 2 × T _sa. The determination and refinement of the characteristics of individual replicas of the form (4) (fill frequency Ω _* , pulse duration, β _tot — attenuation coefficients and arrival times — t _pri ) of the pulse components of the signal are performed by calculating and constructing spectrograms of the scattered signals for signals from all receiving sensors. The size of the time window T _{p of the} analysis of the received scattered signal should theoretically correspond to the size of T _s - the window of analysis of the emission signal, but in practice it can be larger, T _p ≥T _s , due to the large amplitudes of the scattered signals, a better signal to noise ratio and a larger "useful »The duration of each pulse, compared with emission pulses.

на всех приемных датчиках антенны, путем расчета дистанций L_i от эмиссионного пузырька до каждого элемента антенны,

, где С - скорость звука в среде распространения. Точность локализации в активном режиме выше, чем для пассивной локализации, за счет лучшего соотношения сигнал/шум и большей «полезной» длительности каждого импульса по сравнению с результатами пассивной локализации для эмиссионных сигналов.Localization of gas leaks is carried out according to well-known techniques for positioning “noisy” underwater objects in the passive mode by receiving the primary emission signal of the bubble, calculating the delays Δτ _ij = τ _i -τ _j from the antenna sensors and calculating the signal arrival angles to each sensor. The refinement localization is performed at the second, active stage, according to the time difference between the arrival time of the first pulses of the signal t _pri scattered on the bubble and the emitted inverted emission signal

on all receiving sensors of the antenna, by calculating the distances L _i from the emission bubble to each element of the antenna,

where C is the speed of sound in the propagation medium. The accuracy of localization in the active mode is higher than for passive localization, due to the better signal-to-noise ratio and longer “useful” duration of each pulse compared to the results of passive localization for emission signals.

, где р - это номера циклов измерений во времени, k - индекс суммирования по пузырькам разных радиусов в данный момент измерений. Радиусы пузырьков определяются по идентифицированным и измеренным резонансным частотам Ω_* с использованием выражений (2) и (3).Additionally, this method allows the gas flow to be calculated from the resulting gas bubbles over the time interval t≡ [t _n , t _k ] according to the well-known formula

, where p is the number of measurement cycles in time, k is the summation index over bubbles of different radii at the moment of measurement. The radii of the bubbles are determined by the identified and measured resonant frequencies Ω _* using expressions (2) and (3).

Thus, the proposed new method for the detection and localization of emission radiation sources in a gas-liquid medium provides an increase in the probability of identification of the emission object and a reduction in the number of false alarms by deciding on the presence of several criteria. In the passive stage, this is the presence of a pulse, monochromatic, exponentially decaying in time acoustic emission signal of the form described by expression (1), the dependence of the signal duration on frequency in the form of (2) and the exponent in the attenuation law in the form of (3). In the active stage, this is the resonance nature of the scattering of the emitted inverted signal, the type of pulses (4) and the doubled pulse duration in the received scattered signal. The resonance nature of the scattering of the identification object allows the third stage of identification and verification to be realized by emitting an inverted signal at a frequency of Ω ₁ other than the resonance frequency of Ω _* . If the emission object is not a resonance bubble, the reflected signal will have an amplitude two orders of magnitude smaller than the resonance scattering signal.

Claims (1)

Passive-active acoustic method for detecting and localizing gas leaks in a gas-liquid medium by real-time reception and recording in the controlled area of a signal of ambient acoustic noise in the frequency range from 0.1 kHz to 5 MHz, corresponding to the frequencies of natural pulsations of bubbles in a liquid dividing the frequency range into at least two frequency subbands and performing bandpass filtering of the signal in the selected frequency subbands, calculating the spectra for the signals received in each the frequency subband and constructing the corresponding spectrograms, isolating spectral components in the spectrograms having an exponentially decaying in time amplitude and duration from 5 to 30 periods, determining the frequency f _R for the selected spectral component corresponding to the resonant frequency of the pulsation of the bubble, and obtaining a working signal by band pass filtering ambient acoustic noise signal with a center frequency f _R of the filter, the time reference received operating signal, its gain and the radiation eniya corresponding acoustic signal, a second reception and band pass filtering of ambient acoustic noise signal with a center frequency of the filter f _R and excretion therein dispersed bubble pulse signals by calculating the spectrum of the noise signal, and constructing the spectrograms, the detection of gas leakage is monitored by the first appearance in the spectrograms spectral components pulse signals having a symmetric exponentially increasing and decaying in time amplitude with a duration of the periods of the sound field, twice the emission pulse duration and carrier frequency f _R, a gas outlet space localization triangulation method is performed by calculating distances from the bubble until receiving sensors from the measured pulse propagation times in the radiation and the scattered signals and the known velocity of sound in the fluid.

Images