Method and system for distributed acoustic sensing

Abstract

Systems and methods for generating a sensitivity map for a sensing fiber in relation to a monitored infrastructure are disclosed. One method includes: designing one or more event signals matching one

Description

METHOD AND SYSTEM FOR DISTRIBUTED ACOUSTIC SENSINGTECHNICAL FIELD

[0001] Aspects of the present disclosure generally relate to methods and systems of distributed acoustic sensing based on one or more optical fibers. More particularly, aspects of the present disclosure relate to methods and systems for determining the sensing sensitivity of a fiber optic network, and / or using the determined sensing sensitivity to identify acoustic events.BACKGROUND

[0002] Reference to any prior art in the specification is not an acknowledgment or suggestion that this prior art forms part of the common general knowledge in any jurisdiction or that this prior art could reasonably be expected to be understood, regarded as relevant, and / or combined with other pieces of prior art by a skilled person in the art.

[0003] Fiber-optic distributed acoustic sensing (DAS) can detect acoustic events in surrounding regions along an optical fiber, often termed the sensing fiber. An acoustic event can be caused by incidents such as underground digging near a gas pipe, water pipe or a power cable or leaks in such pipes due to accidents, or even just general wear and tear. Different types of incidents may cause different acoustic signatures in the acoustic event. Monitoring of acoustic events therefore allows for alerts to be generated for the prevention or identification of these incidents.

[0004] The method of deploying a dedicated sensing fiber for distributed acoustic sensing may make sense from a design perspective, such that the sensing fiber conditions and parameters (e.g., spatial uniformity along the optical fiber, trench depths, and levels of acoustic attenuation) are known or well-controlled upon installation. However, the installation of a dedicated sensing fiber for distributed acoustic sensing can be expensive and disruptive, particularly in and around an urban center.

[0005] Accordingly, dark fibers (i.e., spare optical fibers in existing optical networks) are often utilized for sensing purposes. For example, events such asdigging near other established infrastructure such as water pipes, gas pipes, or power cables, or leaks / breaks in such infrastructure can be detected using dark fibers of existing optical networks.

[0006] The ability of the dark fibers to detect such events is affected by a number of factors - these include the variability in horizontal and / or vertical offset between the dark fibers and the infrastructure being monitored, and / or the variability of the ground materials in the vicinity of the dark fibers and the detected infrastructure.SUMMARY

[0007] According to an aspect of the present disclosure, there is provided a method for generating a sensitivity map for a sensing fiber in relation to a monitored infrastructure. The method includes: designing one or more event signals matching one or more event types associated with the monitored infrastructure; conducting tests along the monitored infrastructure using the one or more designed event signals; determining sensitivity of one or more channels of the sensing fiber to detect one or more events in relation to the monitored infrastructure based on the tests; and generating a sensitivity map for the sensing fiber based on the determined sensitivity of the one or more channels, the sensitivity map indicating relative sensitivity of the one or more channels of the sensing fiber to detect the one or more events in relation to the monitored infrastructure.

[0008] According to another aspect of the present disclosure, there is provided a method for identifying a water leak in relation to a monitored water pipe network. The method comprising: storing sensitivity values and sensitivity mapping records for a sensing optical fiber that is part of an existing telecommunications network, the sensitivity values indicate a response of channels of the sensing optical fiber to water leaks of different levels and / or sizes in the water pipe network, and the sensitivity mapping records comprising power levels and frequency distribution values for the different levels and / or sizes of acoustic events detected by the sensing optical fiber; processing a return acoustic signal received by a distributed acoustic system from the sensing optical fiber in response to transmitting, by the distributed acoustic system, an interrogating signal to the sensing optical fiber, wherein processing the acoustic signal comprises: determining the power level of the acoustic signal received from thesensing fiber; comparing the power level of the acoustic signal with a reference power level that represents normal baseline conditions to determine a local rise in total acoustic power above the normal baseline; and upon determining the local rise in the total acoustic power, comparing the power level and / or frequency distribution of the acoustic signal with the power level and / or frequency distribution values in the sensitivity mapping records to identify the water leak.BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Fig. 1A illustrates an example of a system for distributed acoustic sensing.

[0010] Fig. IB illustrates an example of a density plot of electrical signals generated by the system of Fig. la over time.

[0011] Figs. 2A and 2B illustrate examples of horizontal offset between a sensing optical fiber and a monitored infrastructure network.

[0012] Fig. 3 is a schematic diagram of a cross section of a segment of a walkway showing vertical offsets between a sensing optical fiber and a monitored infrastructure network.

[0013] Fig. 4 is a flowchart illustrating an example method for generating a sensitivity map for a sensing optical fiber according to embodiments of the present disclosure.

[0014] Figs. 5A and 5B depict persistent spectrograms and processed 2-D PSDs of first and second designed leak signals and corresponding actual leak signals, respectively.

[0015] Fig. 5C is a power spectrum depicting a simulated leak signal and an actual leak signal at a 4L / s rate, where the simulated leak signal successfully mimics the actual leak signal.

[0016] Figs. 5D, 5E, and 5F illustrate spectrums of three example leak signals according to aspects of the present disclosure.

[0017] Fig. 6A is a block diagram of an example device operatively coupled to the monitored infrastructure network to simulate leaks.

[0018] Fig. 6B is a photograph of an example of the device depicted in Fig. 6A.

[0019] Figs. 7A and 7B depict spectrums and power spectrum density plots of the response of a leak signal input to a water pipe as measured by two vibrometers placed in the same location.

[0020] Figs. 8 A and 8B are spectrum and PSD plots, respectively of a first leak signal as measured at a first nearby sensing fiber channel.

[0021] Figs. 8C and 8D are spectrum and PSD plots, respectively for a second leak signal as measured at the first nearby sensing fiber channel.

[0022] Figs. 9A and 9B are spectrum and PSD plots, respectively, of a first leak signal as measured at a second nearby sensing fiber channel.

[0023] Figs. 9C and 9D are spectrum and PSD plots, respectively, of a second leak signal as measured at the second nearby sensing fiber channel.

[0024] Figs. 10A and 10B depict coherent output power (CoP) spectrum for example sensing fiber channels 125-141 centered around channel 133 and for example sensing channels 136-152 centered around channel 144 for a 4L / s leak, respectively.

[0025] Fig. 11 is an example sensitivity map generated at the end of the method of Fig. 4.

[0026] Fig. 12A-12B show leak test results for 0.5, 1.0, 2.0, 3.0 and 4.0L / s leaks induced along a first street and responses detected at sensing fiber channel 641.

[0027] Figs. 12C-12D show leak test results for 0.5, 1.0, 2.0, 3.0 and 4.0L / s leaks induced along the same street and responses detected at sensing fiber channel 662.

[0028] Fig. 13A depicts a coherent output power (CoP) spectrum for example sensing fiber channels 129-159 centered around channel 144 for a range of simulated leak signals over a 60 minute period.

[0029] Fig. 13B depicts a CoP spectrum also for sensing channels 129-159 centered around channel 144 for actual leak signals from 0.5L / s to 4L / s over a a 60 minute period.

[0030] Fig. 14 is a flowchart illustrating an example method for identifying one or more acoustic events based on the sensitivity map according to some aspects of the present disclosure.DETAILED DESCRIPTION

[0031] The principle of fiber-optic distributed acoustic sensing (DAS) relies on the occurrence of an acoustic event causing a corresponding localized perturbation of refractive index of an optical fiber. Due to the perturbed refractive index, an optical interrogation signal transmitted along an optical fiber and then back-scattered in a distributed manner (e.g. via Rayleigh scattering or other similar scattering phenomena) along the length of the fiber manifests in fluctuations (e.g., in intensity and / or phase) over time in the reflected light. The magnitude of the fluctuations relates to the severity or proximity of the acoustic event. The timing of the fluctuations along the distributed back-scattering time scale relates to the location of the acoustic event.

[0032] In one example, a unit 100 for use in distributed acoustic sensing (DAS) is illustrated in Fig. 1A. The DAS unit 100 includes an optical time-domain reflectometer (OTDR) 102. The OTDR 102 includes a light source 104 to emit an optical interrogation signal 106. The interrogation signal 106 is transmitted into an optical fiber 105 and the interrogation signal 106 may be in the form of a short optical pulse. The OTDR 102 includes a photodetector 108 configured to detect the reflected light 110 and produce a corresponding electrical signal 112 with an amplitude proportional to the reflected optical intensity.

[0033] The DAS unit 100 also includes a processing unit 114, within or separate from the OTDR 102, configured to measure the fluctuations in the electrical signal 112 for determining the acoustic event based on the measured fluctuations 116 in intensity as compared between two different times (ti and t2). These acoustic fluctuations are acoustic signals that contain several different acoustic frequencies at any one point and along a series of different spatial points that the processing unit converts to a digital representation of the nature and movement of the sound targets around the sensing fiber 104. These return acoustic signals include a significant number of frequency components and vector information, i.e., the amplitude information derived from the Fourier domain (of single channels) and the multichannel time domain (spatial information such as direction of the acoustic event).

[0034] The digitised electrical signal 112, any measured fluctuations 116 and / or processed data associated therewith may be stored in a storage unit (not shown). Thestorage unit 115 may include volatile memory, such as random access memory (RAM) for the processing unit 114 to execute instructions, calculate, compute or otherwise process data. The storage unit may include non-volatile memory, such as one or more hard disk drives for the processing unit 114 to store data before or after signal processing and / or for later retrieval. The processing unit 114 and storage unit may be distributed across numerous physical units and may include remote storage and potentially remote processing, such as cloud storage, and cloud processing, in which case the processing unit 114 and storage unit may be more generally defined as a cloud computing service.

[0035] Figure IB illustrates an example density plot combining electrical signals 112 generated by the DAS unit 100 over time. The horizontal axis (labelled “Channel”) represents position along the fiber, the vertical axis (labelled “Time”) represents time, and the color-coded amplitude of the plot represents reflected intensity. As referred to herein, the term “channel” refers to a specific spatial segment along an optical fiber that corresponds to a measurement point or sensing location. Each channel represents a discrete location along the optical fiber, typically spaced apart by a fixed distance. The spacing of the channels (i.e., channel resolution) is determined by the DAS unit’s configuration and can vary depending on the length of the fiber and the measurement requirements. Each fiber optic channel provides a discrete sensing segment or pathway for detecting, localizing, and analyzing perturbations, such as acoustic or vibrational events, along the monitored infrastructure.

[0036] If the OTDR 102 is phase -sensitive, phase fluctuations in the reflected light may be additionally or alternatively measured. Figure IB is also offset to remove the attenuation slope in the electrical signal 112 present in Fig. 1A. The acoustic event being determined may be indicative of specific events, such leaks and breaks in infrastructure such as wastewater pipes, gas pipes, or water pipes.

[0037] In case the acoustic event is a fluid leak (e.g., in a water pipe network or gas pipe network), the ability of the DAS unit 100 to detect the fluid leak depends on various factors, such as the degree of the leak (e.g., larger than 4L / s or smaller than 4L / s), the vertical and / or horizontal offset between the sensing optical fiber and the leak location in the monitored infrastructure network, and / or the variations in the material between the two.

[0038] The horizontal offset distance between the monitored pipe network and the optical fiber can vary markedly. Figs. 2A-2B depict examples of these various horizontal offsets. In particular, Figs. 2A-2B depict a water pipe network 202 and a sensing fiber 204 (e.g., which is part of an existing optical network). As seen from fig. 2A, in some zones, e.g., zone 206, the sensing fiber 204 is parallel to the water pipe 202 with a fixed horizontal offset 208 and in other zones, e.g., zone 210, the optical fiber 204 may be perpendicular to or at a different angle to the water pipe 202. Fig. 2B depicts another scenario where the water pipe 202 and optical fiber 204 cross perpendicularly or at some other angle at zone 220. As seen in these depictions, the horizontal offset distance between zones of the water pipe network 202 and the corresponding optical network 204 channels are not uniform. At some zones, e.g., at zone 206, the horizontal offset 208 is lesser than the horizontal offset 208 at zone 210. Further, the orientation of sensing fiber 204 in some segments is parallel to corresponding water pipe segments and in other segments the sensing fiber 204 may be at an angle to the corresponding water pipe segments.

[0039] Similarly, optical networks and other infrastructure networks such as water, gas, oil pipelines may be situated at different depths from the surface. For example, optical networks (that the sensing fiber is a part of) may be placed 200-300mm below surface pavement layers, whereas water pipelines may be typically situated at depths of between 800-1200mm below the surface and encapsulated in course granular material. Further, the vertical offset between these networks may not be uniform either and certain portions of the optical fibers may be more vertically offset from the water pipes than other portions.

[0040] For example, see Fig. 3 that shows a cross-section of a section of a street 300 that shows water pipes 304, 308 and optical fibers 302, 306. One set of optical fibers, i.e., fibers 302 are horizontally and vertically offset from water pipes 304. Further, the optical fibers 302 and the water pipes 304 are located in dissimilar ground material, which likely reduces acoustic transmission due to layer impedance changes. Additionally, in this case, a direct path for acoustic transmission is blocked by other infrastructure such as the wastewater pipes 310 and power cables 312. The other set of optical fibers, i.e., fibers 306 are vertically offset from the water pipes 308 by a smaller offset distance. Further, both the water pipes 308 and the optical fibers 306are located in the same ground material, which may be acoustically transmissive or less transmissive.

[0041] As such, all channels of a sensing fiber may not be able to detect all types of acoustic events (such as fluid leaks) equally.

[0042] To address one or more issues described above with respect to detecting acoustic event such as fluid leaks, described herein are methods and systems for generating a sensitivity map for sensing fibers in a fiber optic network in relation to an observed infrastructure network such as a water pipe network, wastewater pipe network, or a gas pipe network. The sensitivity map indicates the sensitivity level of different channels of the sensing fiber to detect events - for example, some channels may be able to detect even small / faint events (e.g., leaks that are less than 4L / s) in an observed infrastructure network whereas other channels may only be able to detect major events (e.g., leaks that are more than 4L / s) in the observed infrastructure. The sensitivity map indicates these sensitivity variations in the sensing fiber channels that may arise due to vertical and horizontal offsets between a sensing fiber channel and a corresponding zone / location in the observed network, and / or due to variations in ground material between the sensing fiber channel and the observed network zone / location.

[0043] Also described herein is a process and system for using the generated sensitivity map to identify events in relation to the observed infrastructure network. That is, using the sensitivity map, a determination can be made as to the exact location of a leak. Further, using the sensitivity map a transform function may be computed that correlates the sensitivity of a particular channel with an actual event.

[0044] The sensitivity map may also be utilized to determine whether the monitored infrastructure is sufficiently monitored. For example, if there are channels of the sensing fiber that are unable to detect events or only weakly detect events in nearby locations of the monitored infrastructure, the sensing fiber may be supplemented with additional sensing devices in those locations to be able to detect events. Additional sensing devices may be point devices or additional sensing fibers that are installed in closer proximity to those locations of the monitored infrastructure. As such, the sensitivity map can be utilized to improve the monitoring capabilities of a monitored infrastructure.Generating sensitivity may

[0045] Before the sensitivity map can be generated, testing is performed to determine the sensitivity to events at different channels of the optical fiber. However, this testing is not straightforward. This is because the source of a noise is often complex and multi-faceted. If the event is leaks, discharging fluids (such as water or gas) to test the sensitivity of a DAS unit 100 and the optical fiber may lead to acoustic surface noise, splashing, and water flowing in the stormwater drains. This can confound the testing results and may lead to false positive identification of leaks. For example, it may be difficult for a DAS unit 100 to determine whether acoustic noise being detected by the optical fiber is from a hydrant discharging water on to the road or whether it is from a leak in the pipes below the hydrant.

[0046] The presence of groundwater may also affect acoustic reflection and transmission between events occurring at the monitored infrastructure and the optical fiber. Further, the submergence of an event in a fluid (e.g., water) can significantly dampen acoustic and vibration energy coming from the event at its source. In addition to this, the direct measurement of the acoustic emissions from an event (e.g., a leak using a vibrometer) vary from the acoustic energy recorded at the optical fiber due to modification through the propagation media and the process of excitation of the optical signal and conversion of this to an acoustic signal at an interrogation unit.

[0047] The coupling between an optical fiber and its conduit can also vary, regardless of whether the optical fiber is situated in an acoustically transmissive surface layer, or beneath it, or whether the optical fiber is in a similar layer to the monitored infrastructure. Different types of fiber and conduits have more or less sensitivity due to the nature of the fibers themselves and fiber to conduit coupling behaviors.

[0048] There are numerous other physical parameters, in addition to those relating to the optical fiber, the horizontally and vertically varying materials they are embedded in, fiber types, fiber to conduit coupling, the type of event, the sounds the events make, groundwater effects and so on, which affect the sensitivity of event detection using a sensing fiber and the variability of this sensitivity as the fiber and infrastructure network are traversed.

[0049] One existing method for determining the sensing accuracy of a sensing fiber is to use the horizontal offset between the optical fiber and the monitored infrastructure as a proxy for all other variables and conduct extensive controlled event testing to try to calibrate the horizontal offsets in banded ranges to match with event size detection sensitivity. However, this approach is very effort intensive and may not even result in an accurate determination of the sensitivity of the sensing fiber.

[0050] An alternative approach, according to aspects of the present disclosure, is described with respect to method 400 in Fig. 4. The method commences at step 402, where one or more event signals that match or equate to different event types associated with a monitored infrastructure are designed.

[0051] At this step, a range of specifically designed waveforms and corresponding magnitude signals can be obtained from the acoustic / vibration signal input device and stored for various locations along the monitored infrastructure. The waveforms are designed so that, for each location, the transmission of event energy from the monitored infrastructure to the optical fiber varies from weak to strong, covering a range of target detectable event strengths. These designs are based on the outcomes of sufficient controlled testing.

[0052] Once the virtual event signals are determined (for the range of weak / small to strong / large events), the method 400 proceeds to step 404 where tests are conducted along the monitored infrastructure using the specifically designed event signals. When determining sensitivity of the sensing fibers to water leaks in water pipes, the tests may be run across either trunk mains, distribution mains, or both. These tests are used to evaluate the sensitivity of individual optical fiber channels along the sensing fiber to the occurrence of water leaks.

[0053] The output of this testing - that is the resulting signals measured at the various optical fiber channels for different virtual event signals applied to the monitored infrastructure can then be used at step 406 to identify channels of the sensing fiber where the sensitivity to monitored events is relatively lower and channels where the sensitivity to monitored event is relatively higher. The virtual signals have a standard / specific input magnitude and shape which allows relative sensitivities to be determined and the input signals can also be designed so that they also more closely simulate an actual event.

[0054] Once the sensitivities are determined for the sensing fiber along the monitored infrastructure the method proceeds to step 408 where the sensitivities can be mapped spatially to quantify the extent of the optical network (and more particularly the sensing fiber that is part of the optical network) that is sensitive to smaller and larger event detection. The result is a physically quantified and mapped relationship showing the extent of the monitored infrastructure able to be monitored for smaller events and the extent of the monitored infrastructure that is monitorable for larger events. The accuracy of the sensitivity map can be checked via limited controlled event testing afterwards.

[0055] The above process allows for the accurate quantification of the event detection threshold over relevant areas of the monitored infrastructure. Each of the method steps 402-408 will be described in detail in the following sections, with reference to an example monitored infrastructure - a water pipeline network. It will be appreciated that this is just an example of the type of infrastructure for which a sensing fiber’s sensitivity can be determined and that the same process can be applied for different types of monitored infrastructure with minor or no changes. In the case of a water pipeline network, the events are usually leaks of varying leak rates (e.g., small leaks of about 0.5-1 L / s to larger leaks of about 4L / s) at various different locations along the water pipeline network - e.g., behind meters, underground in the actual pipes, or at hydrants. Further, in this example, the design event signals are simulated leak signals and the sensitivity of the sensing fiber is determined along the length of the water pipeline network by applying simulated leak signals at different locations along the water pipeline network and then sensing the simulated leaks at nearby channels of the sensing fiber.Event signal design

[0056] Although actual water leaks can be created at various points along the water pipe network to determine the sensitivity of a corresponding sensing optical fiber, this is not desirable for a number of reasons. Instead, it is more desirable to simulate leaks of different degrees at various locations along the water pipe network to determine the sensitivity of the corresponding sensing optical fiber.

[0057] To do so, various different types of leak signals may be designed to mimic different types of water leaks. When designing the signals, it is important for thedesigned leak signals to be suitably representative of signal power and frequency characteristics of actual water leaks in different scenarios. In other words, the objective is to establish mechanical vibration inputs that provide an appropriate match with known manifestations of leak noises on the sensing optical fiber 105.

[0058] In certain embodiments, the leak signals may be designed based on a trial- and-error method, where different types of leak signals are applied to a water pipe to mimic different leak rates. The measurement of these signals is recorded at the water pipe, e.g., using a vibrometer. Similarly, actual water leaks are created at the same location at the water pipe and the actual leaks are also measured, e.g., using a vibrometer. The power spectral density (PSD) and spectrums of these simulated and actual leaks may then be compared and signals that generate similar PSDs and spectrums as actual leaks may be selected as the designed leak signals. In addition or alternatively, the actual and simulated leaks may be detected at the sensing optical fiber 105 and the PSDs generated by the DAS unit 100 may be compared to select appropriate leak signals.

[0059] One example method for designing the signals involves determining spectrograms for the designed leak signals and conducting persistent spectrogram analysis using, for example, Is measured acoustic data to generate 60 spectrograms over 1 minute and then determining the median 2D Power Spectrum Density (PSD) over the 1 minute using a standard persistent spectrogram approach. The total power, average power, peak power and power distribution over frequency for the designed leak signals and resultant median PSDs are compared with the 2D PSDs for actual pipe leaks, both derived from DAS acoustic data from a sensing fiber nearest the adjacent water network design leak signal or actual pipe leak location. The form of the designed leak signal is varied until the difference between the total power, average power, peak power, power distribution over frequency and / or any other parameters is minimized.

[0060] Fig. 5A depicts persistent spectrograms 500, 502 and processed 2-D PSDs 504, 506 of a designed leak signal and a corresponding actual leak signal, respectively. As can be seen from Fig. 5A, the persistent spectrograms 500 and 502 of the designed and actual leak signals, respectively, are very different, thereby making the designed leak signal unsuitable for mimicking the corresponding actual leak signal.

[0061] Fig. 5B depicts persistent spectrograms 510, 512 and processed 2-D PSDs 514, 516 of a second designed leak signal and a corresponding actual leak signal. As can be seen from Fig. 5B, the persistent spectrograms 510 and 512 of the second designed leak signal and the corresponding actual leak signal, respectively, are similar, thereby making the designed leak signal suitable for mimicking the corresponding actual leak signal.

[0062] Fig. 5C is a power spectrum 520 showing one example of the result of trial- and-error method where different types of simulated leak signals have been applied to a water pipe to successfully mimic a 4L / s leak rate. In this example, a sweeping 10 - 75Hz mechanical vibration source is applied to a water pipeline such that the power and frequency (amplitude and bandwidth) measured on the sensing fiber 105 at the nearest channel, when averaged over a 30s time window, closely resemble the power and frequency (amplitude and bandwidth) measured on the sensing fiber 105 at the nearest channel from an actual 4L / s leak.

[0063] In some other embodiments, fluid leaks of different types (e.g., small and large leaks of different fluid types at different locations) may be recorded and leak signals may be generated based on these recordings. These leak signal recordings can be input via a mechanical device at different points along the fluid pipe network for testing the sensitivity of the optical fiber network to these signals at the different points (e.g., at method step 404).

[0064] In another embodiment, the leak signals may be mechanical acoustic tapping signals that simulate water meter ticking noises due to behind the meter customer leaks. These tapping signals can be similarly input using a mechanical device at different points along the fluid pipe network (e.g., near meters) for testing the sensitivity of channels of the optical fiber network to these signals at the different points.

[0065] Spectral charts 530, 540, 550 of three example leak signals are shown in Figs. 5D-5F when the leak signals are applied to a water pipe either directly or via a fixture or fitting attached to the water pipe (e.g., a water connection, water meter, isolation valve, fire hydrant, control valve, pipe directly and / or other).

[0066] Fig. 5D shows the spectral chart 530 for a manual tapping signal, which was generated using an aluminum bar tapped at approximately 3-6Hz to simulate themagnitude and ticking rate and sound of a water meter with a downstream leak and continuous flow through it. Fig. 5E shows the spectral chart 540 of a 10-75Hz repeating sweep signal, and Fig. 5F shows the spectral chart 550 of varying short and longer mixed sweep leak signals input to pipe and / or pipe fittings.Applying leak signals

[0067] Once leak signals that mimic fluid leaks are designed, these signals of known power, pattern and frequency characteristics and ranges are applied to fluid pipes (such as water, wastewater or gas pipes) directly or via fixtures including fire hydrants, isolation valves, water meter connections, control valves and / or any other types of fixtures. These locations provide a set of locations for determining the sensitivity transfer level between the pipe generated noise and the nearby sensing fiber. Supplementary locations that are applicable to the signal input device are added as required to increase the number of locations and the accuracy of the sensitivity transfer level mapping.

[0068] The simulated leak signals are applied to the pipe network via a device. An example of this device 600 is depicted in Fig. 6A. The device 600 generally includes a transducer 602, an amplifier 604, and a controller 606. It may also include a power supply (not shown) in some examples. The transducer 602 may be coupled to the water pipe network 202 by some physical means (e.g., a magnet, a physical strap, or any suitable type of specialty housing). The controller 606 supplies leak signals to the amplifier 604, which amplifies the simulated leak signals and communicates the amplified leak signal to the transducer 602. The transducer 602 vibrates the water utility 202 in line with the simulated leak signal.

[0069] In one example, the device 600 includes a Fosi Audio M01-BT Mono Sub Amplifier, an EarthquakeSound MQB1 Shaker fitted with a magnetic attachment, and a Jackery Explorer 160 Portable Power Station (Lithium-ion battery, 11.6 Ah / 14.4V). This example device is depicted in Fig. 6B.

[0070] Illustrations of the response of the simulated leak signals input to a water pipe 202 as measured by two vibrometers placed in the same location are shown in Figs. 7A and 7B. In particular, Figs. 7A and 7B show the spectrograms 702, 712 and PSDs 704, 714 measured by the two vibrometers located at a simulated leak site in the water pipe network. A general 10s signal wave sweep is generated by the controller606 and transmitted to the transducer / amplifier arrangement. The power level is set to maximum (100%) for the device in this example and 3 x 10s sweeps over 30s are detected in the recorded responses which are shown as spectrograms 702, 712 and PSDs 704, 714, respectively.Sensitivity determination

[0071] The sensitivity of the optical fiber to the leak signals can then be determined using simulated leak tests.

[0072] In certain embodiments, acoustic vibrometers and / or accelerometers may be installed at the leak locations (real leaks or simulated leaks) to measure the real and / or simulated leaks at the source. These vibrometers / accelerometers measure a spectrogram and power spectral density (PSD) of the actual or simulated leak at the source. The same leaks are also measured at the nearest optical fiber channels using the DAS unit 100.

[0073] The spectrogram may be generated by standard methods. Generating the PSD includes first taking a fast Fourier transform of discrete time domain data to convert it into a frequency domain signal. This includes computing a fast Fourier transform of the discrete time-based data xn(n = 0, ... , N — 1) where N is the total length (data points) of the data which is defined by:where Pkk = 0, ... , N — 1) are the transformed results in the frequency domain and j is the imaginary unit.

[0074] The discrete fast Fourier transform (DFFT) is determined for the data xn(n = 0, ... , N — 1) which may also be a data series that is partitioned into K segments or batches with M data points in each segment (and optionally S data points for overlapping segments). For each segment k=l, , K the discrete fast Fourier transform (DFFT) is determined (eg, for non-overlapping segments) as:where . = i / M and u>m= a window function.

[0075] The power in each segment is then determined using the obtained DFFT as follows:

[0076] The average of the power level values in each segment is then determined to give the Welch’s method estimate of the PSD as:

[0077] The PSD is accordingly determined by squaring each FFT (or DFFT) value (squared value of each frequency data point) to obtain power values for each point for each frequency bin (window) used in the determination of the FFT and an average for any frame size applied that is greater than the frequency bin width.

[0078] The power values (squared FFT frequency data point values) are typically normalized by dividing by the signal recording rate in Hertz (sampling frequency Hz) such that the final representation of and units for the PSD values are in decibels (power) per Hertz (Hz) or dB / Hz and vary with frequency over the sampling frequency range for the data.

[0079] The raw data from the DAS unit 100 is similarly processed to generate spectrograms and PSD for optical fiber channels in the vicinity of the water pipe leak test location. The transformation of the leak signal as measured at the leak source in terms of power and frequency distribution in the spectrogram and / or PSD to the corresponding measurement at the nearest sensing fiber channel can be determined and quantified in terms of:• Maximum spectral power change (from leak source to optical fiber channel)• Average spectral power change (from leak source to optical fiber channel)• Total spectral power change (from leak source to optical fiber channel)• Distribution changes in spectral power in terms of power levels and frequencies at which they occur (from leak sources to optical fiber channel)• Other characteristics of the spectrograms and PSDs as they are transformed from the leak on the pipe measurement to the optical fiber channel recorded signal.

[0080] Numerous signal process techniques can be used to quantify the transformation function in both statistical and / or machine learning domains. For example, different types of actual leaks versus designed leak signals manifest with different levels of power distributed in frequency. As sensitivity testing is conducted in accordance with this disclosure, libraries of responses for different types of water pipe network leak faults, for different water pipe networks and different pressure and pipe types is accumulated. Similarly, as sensitivity testing is conducted in accordance with this disclosure, libraries of responses for different designed leak signals will be accumulated. Characteristics of both actual leak and designed leak responses can be categorized and used to identify optimal designed leak signals and for statistical classification of sensitivity levels. Machine learning models (e.g., convolution neural networks), trained to the libraries of responses, can also be used to identify optimal designed leak signals and for classification of sensitivity levels.

[0081] Once the transformation functions are determined they are available for setting the sensitivity level variations from one leak test location to the next and act as a calibration via which the efficacy of the sensing fiber to sense leaks is predetermined (for actual monitoring coverage levels).

[0082] Figs. 8A and 8B respectively depict the spectrogram 800 and PSD 810 for a 10s repeating leak signal applied to a water pipe with 100% power level. In particular, Fig. 8A shows the spectrogram 800 and Fig. 8B shows the PSD 810 of the repeating signal as measured by the DAS unit 100. Quantified measurements for the PSD were -• Total power above -85dB for 0-350Hz: 6165dBHz• Average power: 17.6dBHz• Maximum (peak) power: -55dB seconds sound recording of a 0.5L / s leak at a pipe location adjacent to a particular optical fiber channel.

[0083] Figs. 8C and 8D respectively depict the spectrogram 820 and PSD 830 of a 1.3 minute variable power and frequency signal applied to the same location in the water pipe network. In particular, Fig. 8C shows the spectrogram 820 and Fig. 8Dshows the PSD 830 of the repeating signal as measured by the DAS unit 100 at the sensing fiber channel. Quantified measurements for the PSD were -• Total power above -85dB for 0-350Hz: 5274dBHz• Average power: 15.1 dBHz• Maximum (peak) power: -65dB

[0084] As can be seen in these figures, the response of the sensing channel adjacent to the leak location that detected the leak is low in both instances. As a result of the above tests, the transmission of leak and / or mechanical device acoustic power between that water pipe location and the nearest sensing fiber channel was able to be quantified as “weak”.

[0085] Figs. 9A and 9B respectively depict the spectrogram 900 and PSD 910 response of a sensing fiber channel at another location along the water pipe network. The leak signal is a 10s repeating signal applied to the water pipe at the second location with 100% power level. In particular, Fig. 9A shows the spectrogram 900 and Fig. 9B 910 shows the PSD of the repeating signal as measured by the DAS unit 100 at the sensing fiber channel. Quantified measurements for the PSD were -• Total power above -85dB for 0-350Hz: 7527dBHz• Average power: 21.5 dBHz• Maximum (peak) power: -50dB

[0086] Figs. 9C and 9D respectively depict the spectrogram and PSD of a 1.3 minute variable power and frequency signal applied to the same water pipe location. In particular, Fig. 9C shows the spectrogram 920 and Fig. 9D shows the PSD 930 of the repeating signal as measured by the DAS unit 100 at the sensing fiber channel. Quantified measurements for the PSD were -• Total power above -85dB for 0-350Hz: 6572dBHzAverage power: 18.8dBHzMaximum (peak) power: -60dB

[0087] As a result of the above tests, the transmission of leak and / or mechanical device acoustic power between the water pipe and the nearest fiber channel was able to be quantified as “moderate to strong”.

[0088] Table 1 below shows the ratios of the total power, average power, and peak power for the relatively “moderate to strong” to relatively “weak” locations along the water pipe network for the 10s repeating versus 1.3 minute variable power and frequency signals, respectively.

[0089] The ratio of the total power for the 10s repeating and 1.3 minute variable power is 1.22 and 1.25, respectively. Similarly, the ratio of the average power is also 1.22 and 1.25, respectively, for the 10s repeating and 1.3 minute variable power and frequency signals. The ratio of the peak power, on the other hand, is 1.10 and 1.08, respectively, for the 10s repeating and 1.3 minute variable power and frequency signals. The peak power characteristic is less sensitive to the differences between the two locations and is an example of a less representative indicator due to the influence of unrepresentative peak power fluctuations (short duration). The total or average power indicators are more representative of the sensitivity of the sensing fiber. Indicators that take the distribution of the total or average power over frequency into account are even more representative of the relative sensitivity between locations.

[0090] In this manner, the sensitivity of each sensing fiber channel to leaks of different types along the water pipe network can be determined. That is, the mechanical / vibration device 600, with standardized leak signals, is affixed to water pipes and / or associate fixtures at different locations and the leak signals are induced and transmitted to the sensing fiber. The response of the sensing fiber can be determined at different channels and recorded. Based on the response, the sensitivity of the entire sensing fiber to detecting events in a nearby water pipe network can be determined.

[0091] It is generally accepted that fluid-filled pipes buried in the subsurface conduct sound waves, such as those generated by a leak. The manner in which this energy propagates is believed to couple acoustic wave propagation through the fluid with elastic wave propagation through the solid pipe wall. In this system, waves propagate in longitudinal, torsional, flexural modes (conceptualized as cylindrical harmonics). At low frequency (<5 kHz) and long range (>10 meters up to 1000 meters of linear pipeline), wave propagation is dominated by the fundamental longitudinal wave mode (L0, 1) that travels in the fluid, although other secondary modes pay a secondary role in certain cases. L0,l dispersion / attenuation characteristics depend on the pipe diameter, wall thickness, pipe elastic moduli, fluid properties and surrounding subsurface geology. Energy from this system will attenuate as it is shed elastically to the surrounding environment. As observed by a neighboring colinear distributed fiber optic sensor embedded in the subsurface near the pipe, these elastic waves will be detected as strain signals at a given frequency and amplitude that depends on the fiber-pipe offset. Distributed fiber sensing observations of a calibrated vibration source that is coupled to the water infrastructure can thus provide important information about attenuation patterns across the water network.

[0092] As the artificial leak signal propagates along the pipe and / or through coupled pipes and surrounding buried strata sensing fiber channels, upstream and downstream of the location where the mechanical / vibration device is located, these sensing fiber channels also record the transformed signal.

[0093] A typical range of rates of decrease of maximum power level, total power level and frequency distribution of power is identified for water pipes and networks with distance away from artificial leak signal source (accounting for all physical loss transformative processes along the pipe and sensing fiber paths). As well asatributing the sensitivity of a pipe to a sensing fiber response for a specific fiber location, the results along the sensing fiber moving away from the location at which the mechanical vibration device is applied are determined and used to extend the sensitivity mapping from point locations to general coverage.

[0094] If the rate of decrease of the input power and patern is greater than the standard rate then a relatively proportionate loss of sensing fiber sensitivity along the pipe can be assigned (given variations in acoustic transmission along it, its combination with surrounding strata and the separations between the pipe and sensing fiber). If the rate of decrease of the input power and patern is less than the standard rate, or even a net overall increase in the power, then a relatively proportionate loss (or gain) of sensing fiber sensitivity along the pipe can be assigned.

[0095] The spectra in Figs. 10A and 10B illustrate this. In particular, Fig. 10A depicts a coherent output power (CoP) spectrum 1000 for example sensing fiber channels 125-141 centered around channel 133. Fig. 10B depicts a CoP spectrum 1002 for example sensing channels 136-152 centered around channel 144 for a 4L / s leak.

[0096] The coherent output power (CoP) spectrum is obtained by selecting a number of sensing fiber side channels surrounding a target channel to identify propagating coherent noise, as quantified by the Maximum Coherent Output Power (MCOP).. For example, five side channels may be selected on either side of the target channel, where the left-hand side and right-hand side numbers indicate the number of channels used on each side of the nearest sensing fiber channel to a potential leak. For a sensing fiber channel that is 9 m long, this corresponds to approximately 50 m of sensing fiber on either side of the leak location. The methodology for computing the CoP, including the derivation of MCOP from the selected side channels, is described in detail below..

[0097] As can been seen from these spectra 1000, 1002, there is a high sensitivity extending from at least CH 128 to CHI 38 before a decrease to a moderate sensitivity from CH138 to CH145 before a drop to low sensitivity from CH145 and above. This drop from channel 145 accounts for the standardized rate of power drop as the distance from the sensing fiber channel back to the location of the leak on the water pipe increases.

[0098] FIG. 11 is a sensitivity map 1100 generated based on the simulated leak tests conducted around a particular area - in this example, the sensitivity map 1100 is generated for a sensing fiber that is part of an optical fiber network in Russells Ride, London. The sensing fiber 105 is depicted by dashed lines and the water network pipes 202 are depicted as solid lines. The sensitivity of the sensing fiber 105 was detected based on simulated leak tests conducted along the water network pipes 202 at the locations shown by dots 1102, squares 1104, and triangles 1106. The dots 1102 represent fire hydrants, the squares 1104 represent isolation valves and the purple triangles represent water meters.

[0099] The sensitivity map 1100 highlights different sensitivities of the sensing fiber. For example, zones A are highly sensitive zones of the sensing fiber 105 as these zones can sense leaks as small as 0.5 L / s in nearby water pipes 202. Zones B are high-to-moderate sensitivity zones as the sensing fiber channels in these zones can sense leaks as small as IL / s. Zones C on the other hand are moderate sensitivity zones as the smallest leaks they can effectively sense is a 2L / s leak. Finally, zones D and E are moderate to low and low sensitivity zones respectively as they can effectively only sense leaks greater than 3L / s and 4L / s, respectively.

[0100] Once a sensitivity map 1100 is generated in this manner, the accurate quantification of a leak detection threshold over relevant areas of the network can be determined.Experimental results

[0101] Actual leak testing was undertaken at Swanfield Street in London for water network pipe locations adjacent to sensing fiber channels 641 (relatively “weak” sensitivity) and 622 (relatively “moderate to strong” sensitivity) after the assessment of the sensitivity of each site to designed leak signals (as described previously). Actual Leak flows of 0.5, 1.0, 2.0, 3.0 and 4.0L / s were tested and processed in a number of ways, including autospectra, COP and MCOP methods, to determine responses and visualize the leak signals measured by the DAS unit 100 for the adjacent sensing fiber channels 641 and 662.

[0102] Fig. 12A and Fig. 12B show the leak test results for 0.5, 1.0, 2.0, 3.0 and 4.0L / s leaks induced along Swanfield St and responses detected at the sensing fiber channel 641. In particular, these figures show autospectrums and coherence outputpower (CoP) with frequency of the leak noise power across the horizontal axis, time along the vertical axis, and the periods within which the different leak flow rates occurred annotated. It is evident that the leaks do not manifest clearly for any flow rate in either the autospectra or CoP results with some visibility of the 4L / s leak apparent in the CoP results.

[0103] Fig 12C and Fig 12D show the leak test results for the 0.5, 1.0, 2.0, 3.0 and 4.0L / s leaks induced along Swanfield St and responses detected at the sensing fiber channel 662. In particular, these figures show autospectrums and coherence output power (CoP) with frequency of the leak noise power across the horizontal axis, time along the vertical axis, and the periods within which the different leak flow rates occurred annotated. It is evident that the leaks manifest progressively more clearly for flow rates above 2L / s for both the autospectra and CoP results.

[0104] The calculation of the coherent output power (CoP) spectrum is used to incorporate the contribution of multiple sensing fiber channels sensing a leak (or mechanically simulated leak) which is spatially closest to one of the sensing fiber channels. The CoP may be determined using standard methods.

[0105] The coherence of a linear system is the fraction of the output signal power that is correlated with the input. The ordinary coherence function between two time series x(t) and y(t) is given by:where Gxx(f) is the auto-spectral density of time series x(t), Gyy(f) is the auto- spectral density of time series y(t), and Gxy(f) is the cross-spectral density function between series x(t) and y(t)

[0106] We can use the above to estimate the acoustic power (spectral density) at one channel that is observed at another:COP = YxyGyywhere coherent output power (COP) above is the PSD (power-spectral density) at y which arises from (or correlated with) x.

[0107] The coherent output power for multiple time series (e.g., sensing fiber channel acoustic responses) can be categorized as per the following cases:1) All reference inputs and output(s) completely uncorrelated:2) Reference inputs uncorrelated with one another but each correlated with the output(s) with the use of the ordinary coherence function applicable:3) Reference inputs correlated with one another as well as correlated with the output(s) with the ordinary coherence function not able to be applied:

[0108] For case 2) above, with the reference inputs not correlated with one another, we can sum the ordinary coherence function outputs for each correlated input reference and output(s) to get the COP:

[0109] For case 3) above, we need to use a multiple time series (e.g., sensing fiber channel) coherent output power (MCOP) approach to avoid the sum of the correlated inputs and output(s), as well as the summed coherence between the inputs, exceeding unity. Where reference channels %i(t) and2(t) are partially correlated with y(t), and each other, the multiple coherence is given by:Where,There are numerous algorithms to permit calculation of the multiple coherence function. Bendat and Piersol have an iterative one (i.e., the use of estimated and then updated values of H* and H).

[0110] The spectra shown in Figs. 13A and 13B illustrate this. In particular, Fig. 13B depicts a coherent output power (CoP) spectrum for example sensing fiber channels 129-159 centered around channel 144 for a range of actual leaks from 0.5L / s to 4L / s over a 60 minute period (source at CH 144). Fig. 13A depicts a CoP spectrum also for sensing channels 129-159 centered around channel 144 but for the designed leak signals (set to various power and frequency patterns including 10s repeating sweeps and 1.3 minute variable power and frequency signals over a 60-minute period).

[0111] The relative sensitivities to the actual leaks and designed leak(s) are similar at CH 144 and for sensing fiber channels moving away from CH 144 in both directions. This confirms the use of the designed leak signals as proxies for actual leaks and also that rate of decrease of the designed leak signals along the sensing fiber (away from a source) is relatable to the rate of decrease for an actual leak and can be used to map the sensitivity of the sensing fiber without a direct designed leak signal being applied at all locations.Example method for detecting leaks

[0112] Fig. 14 is a flowchart illustrating an example method for detecting an acoustic event in a nearby infrastructure network using a sensing fiber in an existing optical fiber network in an area. Method 1400 will be described with reference to a water pipe network as the infrastructure network and water leaks as the acoustic event. However, as described previously, this is merely an example and the infrastructure network may be any other infrastructure network against which sensitivity of the sensing fiber is already determined.

[0113] At step 1402, the determined sensing sensitivity of the sensing fiber is stored by the processing unit 114 of the DAS 100. For example, for each channel of the sensing optical fiber, a corresponding sensitivity value may be stored - e.g., on a scale of 1-5, as an actual discrete sensitivity value, or as tags (e.g., high sensitivity, moderate sensitivity, etc.). In addition, for each sensitivity value, sensitivity mapping records may be stored, which include the response of the sensing fibre and / or theresponse of individual channels of the sensing fiber to various levels of leak noise, and leak sizes or extents. The response values include the power level (e.g., total, average and peal power levels) and frequency distribution detected at the fiber for each type / extent of leak. Further, the frequency distribution values include the spectrogram, autospectra and / or COP values. Furthermore, the variation in transfer to the nearest, next nearest and then increasingly distant fibre channels (as described with reference to Figs. 11A and 1 IB) is also stored along with the sensitivity values.

[0114] At step 1404, distributed acoustic sensing is performed using the DAS 100 - i.e., interrogating signals (sequence of temporally separated, coherent optical pulses) are transmitted in the sensing fiber 104. The interrogating signals may be transmitted continuously or periodically. The sensing fiber typically forms part of an existing optical fiber telecommunications network. The return signals (Rayleigh backscattered light) from the sensing fiber are received at the DAS 100 during observation periods following the transmission periods. This configuration permits determination of an acoustic signal (amplitude, frequency and phase) at every distance along the sensing fiber. In one embodiment, the photodetector / receiver records the arrival times of the pulses of reflected light to determine the location and therefore the channel where the reflected light was generated along the sensing fiber.

[0115] At step 1406 acoustic signals are demodulated from the return optical signals, and at step 1408, an acoustic filter is applied to the acoustic data to detect one or more acoustic events. In particular, when processing the acoustic data, the processing unit 114 may determine whether one or more acoustic events are detected at one or more physical channels of the sensing fiber using the acoustic filter. In one embodiment, a broad band Gaussian filter is utilized to identify water leaks. If an acoustic event is detected at one or more channels of the sensing fiber, the processing unit 114 retrieves the sensitivity information associated with those channels and uses the sensitivity information along with the return signals associated with the one more channels to determine whether a water leak is detected and what the severity of the leak is.

[0116] In one embodiment, step 1208 involves comparing the acoustic power level of the demodulated acoustic data with a reference or baseline level that represents normal background conditions and with the sensitivity mapping records. The baseline may correspond to the typical acoustic energy present in the absence ofleaks, such as background soil vibration, hydraulic noise, or ambient environmental sound.

[0117] When a leak occurs, the flow of pressurized fluid through a crack or defect in the pipe network generates turbulent flow and rapid pressure fluctuations that produce a characteristic increase in acoustic energy. This increase is manifested in the return signal as a local rise in the measured acoustic power level. The processing unit 114 identifies such events by monitoring for persistent or spatially localized increases in total acoustic power above the normal baseline.

[0118] To confirm that the detected increase originates from a water leak rather than an unrelated source (e.g., traffic vibration, pumps, or mechanical impacts), the system further analyses the frequency distribution of the elevated acoustic signal. The frequency content is determined by performing a spectral decomposition (for example, by a Fourier transform) of the time-domain signal measured for the relevant fiber channel or channels. The processing unit 114 then evaluates whether the resulting spectrum exhibits a broadband, statistically Gaussian distribution characteristic of fluid turbulence. In particular, leak-generated noise is typically continuous and spread across multiple frequency components rather than concentrated at a single tonal frequency.

[0119] In one implementation, a water leak is detected if the frequency spectrum displays at least one, and up to ten, distinct sub-bands of elevated energy, each extending across a minimum total bandwidth of approximately 10 Hz. These sub-bands may result from resonances or interference effects in the pipe wall, the surrounding soil, or the cavity formed by the leak. The Gaussian nature of the distribution is confirmed by statistical analysis of the amplitude histogram of the frequency components, verifying that the noise amplitude follows an approximately normal distribution rather than exhibiting periodic or deterministic characteristics.

[0120] Acoustic events meeting these criteria are classified as likely leak signatures and are compared with the sensitivity mapping records to determine the type of leak. For example, the power and frequency distribution values of the detected potential leak are compared with the calibrated response (i.e., power and frequency distribution values of the various severity and types of water leaksensitivity mapping records) for the particular channel and / or channels where the acoustic events are detected.

[0121] The relative magnitude, width, and persistence of the broadband response may be correlated with the size or severity of the leak: larger leaks typically generate higher acoustic power and broader frequency coverage due to stronger turbulent flow. Additional acoustic parameters, such as temporal coherence, amplitude modulation, or decay rate across neighbouring fiber segments, may also be analysed to estimate the spatial extent or pattern of the leak.

[0122] The time required for the processor 114 to identify a leak may vary depending on both the magnitude of the leak and the acoustic sensitivity of the optical fiber at the corresponding location. Smaller leaks, for example having flow rates below approximately 0.5 litres per second, typically produce low-level acoustic emissions that are close to the ambient background noise floor. In such cases, the system performs long-duration monitoring to allow statistical separation of true leak- related signals from random or transient noise.

[0123] This process involves repeated determination of the local acoustic power level at each spatial position along the fiber, tracking of those power levels over extended time intervals, and progressive enhancement of persistent signal components through techniques such as COP amplification. In coherent amplification, temporally correlated acoustic signals recorded from the same spatial segment are combined in phase across multiple time windows, thereby reinforcing consistent features associated with an ongoing leak while averaging down uncorrelated background noise. As a result, small or slowly developing leaks can be detected reliably overtimescales ranging from several days to several weeks.

[0124] For larger leaks, such as those exceeding approximately 0.5 litres per second, the acoustic energy transmitted to the fiber is substantially greater. The same signal processing steps are applied; however, in this case, the elevated acoustic power levels can be detected and confirmed in near real time. The system therefore provides rapid notification of major leaks or pipe bursts by continuously monitoring short-term increases in measured acoustic power.

[0125] It will be appreciated that the acoustic signals acquired from the optical fiber are processed by the processing unit 114. The interrogator transmits theacquired data to the processing unit 114, which may be located either locally (“EDGE” processing) or remotely (“CLOUD” processing). The processing unit 114 continuously extracts and analyses a range of acoustic parameters including, but not limited to, acoustic power level, frequency distribution, spectrograms, autospectra, coherent output power, cross-correlations, and other derived statistical features of the measured signal.

[0126] These data are fdtered and processed over configurable time intervals to accommodate both rapid event detection and long-term trend analysis. From the processed signals, waterfall-type visualizations (representing the variation of acoustic power as a function of time and fiber position) may be generated to facilitate visual interpretation and diagnostic validation.

[0127] The extracted information is further organized into detection modules or “detectors,” each of which is designed to respond to specific signal characteristics such as absolute acoustic power level, rate of change in power, or variations in spectral content. The detectors may operate independently or cooperatively, and may apply association, clustering, and zoning algorithms to group spatially or temporally related events. This clustering enables the system to distinguish between isolated transient disturbances and persistent, spatially coherent leak phenomena.

[0128] By weighting detections according to local sensitivity, the system improves confidence in identifying genuine leaks and minimizes false positives arising from environmental noise or low-sensitivity regions.

[0129] It will be understood that the invention disclosed and defined in this specification extends to all alternative combinations of two or more of the individual features mentioned or evident from the text or drawings. All of these different combinations constitute various alternative aspects of the invention.

[0130] As used herein, except where the context requires otherwise, the term "comprise" and variations of the term, such as "comprising", "comprises" and "comprised", are not intended to exclude further additives, components, integers or steps.

Claims

CLAIMS1. A method for generating a sensitivity map for a sensing fiber in relation to a monitored infrastructure, the method comprising: designing one or more event signals matching one or more event types associated with the monitored infrastructure; conducting tests along the monitored infrastructure using the one or more designed event signals; determining sensitivity of one or more channels of the sensing fiber to detect one or more events in relation to the monitored infrastructure based on the tests; generating a sensitivity map for the sensing fiber based on the determined sensitivity of the one or more channels, the sensitivity map indicating relative sensitivity of the one or more channels of the sensing fiber to detect the one or more events in relation to the monitored infrastructure.

2. The method of claim 1, wherein the monitored infrastructure is a water pipeline network.

3. The method of claim 2, wherein the one or more event signals are one or more leak signals that are representative of signal power and / or frequency characteristics of one or more actual leaks in the water pipeline network.

4. The method of claim 3, wherein designing the one or more leak signals comprises: causing one or more actual leaks in the water pipeline network; sensing the one or more actual leaks using the sensing fiber; generating power spectral density (PSD) and / or power spectrums of the sensed one or more actual leaks in the water pipeline network; generating a plurality of simulated leak signals; applying the plurality of simulated leak signals to the water pipeline network;sensing the plurality of leak signals using the sensing fiber; generating PSD and / or power spectrums of the sensed plurality of leak signals; comparing the PSD and / or power spectrums of the one or more sensed actual leaks with the PSD and / or power spectrums of the sensed plurality of leak signals; and selecting the one or more leak signals based on a substantial match between the PSD and / or power spectrum of at least one actual leak and the PSD and / or power spectrums of the sensed plurality of leak signals.

5. The method of claim 3, wherein the one or more leak signals represent actual leaks of one or more intensities.

6. The method of claim 3, wherein designing the one or more leak signals comprises recording an audio of one or more actual leaks and using the recorded audio as the one or more leak signals.

7. The method of claim 3, wherein the one or more leak signals comprise manual tapping signals of approximately 2-6 Hz frequency that simulate a magnitude and ticking rate of a water meter downstream of an actual leak.

8. The method of claim 3, further comprising applying the one or more leak signals to different locations along the water pipeline network.

9. The method of any one of claims 1-5, wherein determining the sensitivity of the one or more channels comprises: applying the one or more designed event signals to one or more locations along the monitored infrastructure; determining one or more of total power, average power or maximum power of the one or more designed event signals at the one or more locations;sensing the one or more designed event signals using the sensing fiber; computing one or more of total power, average power or maximum power of the sensed signals at one or more sensing fiber channels nearest to the one or more locations using a DAS unit; determining the sensitivity of the one or more channels of the sensing fiber based on the total power, average power and / or maximum power of the sensed signals and the corresponding total power, average power and / or maximum power of the one or more event signals at the one or more locations.

10. The method of any one of the preceding claims, further comprising using the sensitivity map to identify one or more events in relation to the monitored infrastructure and generating one or more alerts.

11. The method of claim 10, wherein identifying the one or more events in relation to the monitored infrastructure comprises: storing sensitivity values and sensitivity mapping records for the sensing fiber, the sensitivity mapping records comprising power levels and frequency distribution values for multiple levels and / or sizes of acoustic events detected by the sensing fiber; processing an acoustic signal received from the sensing fiber, wherein processing the acoustic signal comprises: determining the power level of the acoustic signal received from the sensing fiber; comparing the power level of the acoustic signal with a reference power level that represents normal baseline conditions to determine a local rise in total acoustic power above the normal baseline; and comparing the power level and / or frequency distribution of the acoustic signal with the frequency distribution values in the sensitivity mapping records to determine a level and / or size of the acoustic event.

12. The method of claim 11, further comprising: determining a likely water leak event if the acoustic signal has a substantially broadband Gaussian distribution.

13. The method of claim 12, further comprising: determining a likely water leak event if a frequency distribution of the acoustic signal includes between 1 and 10 distinct sub-bands of elevated energy, each sub-band extending across a minimum total bandwidth of approximately 10 Hz.

14. The method of any one of claims 10-13, wherein the one or more events are detected at one or more channels of the sensing fiber.

15. The method of claim 1, wherein the monitored infrastructure is any one of a gas pipeline network, a wastewater pipe network, a railway line, a fence, a building, or a geographical area.

16. The method of any one of the preceding claims, wherein the designed one or more event signals are applied to the monitored infrastructure via a mechanical device comprising a transducer, an amplifier and a controller.

17. The method of any of the preceding claims, wherein the sensitivity map indicates channels of the sensing fiber that are weakly sensitive to the one or more events associated with the monitored infrastructure and channels that are strongly sensitive to one or more events associated with the monitored infrastructure.

18. The method of claim 13, wherein weakly sensitive sensing fiber channels cannot detect events below a threshold intensity and strongly sensitive sensing fiber cannels can detect events below the threshold intensity.

19. A method for identifying a water leak in relation to a monitored water pipe network, the method comprising: storing sensitivity values and sensitivity mapping records for a sensing optical fiber that is part of an existing telecommunications network, the sensitivity values indicate a response of channels of the sensing optical fiber to water leaks of different levels and / or sizes in the water pipe network, and the sensitivity mapping records comprising power levels and frequency distribution values for the different levels and / or sizes of acoustic events detected by the sensing optical fiber; processing a return acoustic signal received by a distributed acoustic system from the sensing optical fiber in response to transmitting, by the distributed acoustic system, an interrogating signal to the sensing optical fiber, wherein processing the acoustic signal comprises: determining the power level of the acoustic signal received from the sensing fiber; comparing the power level of the acoustic signal with a reference power level that represents normal baseline conditions to determine a local rise in total acoustic power above the normal baseline; and upon determining the local rise in the total acoustic power, comparing the power level and / or frequency distribution of the acoustic signal with the power level and / or frequency distribution values in the sensitivity mapping records to identify the water leak.

20. The method of claim 19, further comprising: determining a level and / or size of the water leak based on the power level and / or frequency distribution value in the sensitivity mapping record the power level and / or frequency distribution of the acoustic signal matches the closest to.

21. The method of claim 19 or 20, wherein identifying the water leak event further comprises determining whether the acoustic signal has a substantially broadband Gaussian distribution.

22. The method of claim 21, wherein identifying the water leak event further comprises determining whether the frequency distribution of the acoustic signal includes between 1 and 10 distinct sub-bands of elevated energy, each sub-band extending across a minimum total bandwidth of approximately 10 Hz.

Images