Efficient Automatic Determination Method for Buried Cables for Cable Condition Monitoring

Abstract

Distributed fiber optic sensing (DFOS) systems and methods autonomously collect DFOS data using an optical fiber sensor cable and use artificial intelligence / machine learning (AI / ML) to identify above-ground (aerial), below-ground (buried), and sections that are buried but sometimes above ground of the optical fiber sensor cable, and further identify changes occurring with respect to the optical fiber sensor cable in such sections.

Description

Technical Field

[0001] This application generally relates to distributed fiber optic sensing (DFOS) systems, methods, structures, and machine learning (ML) techniques. More particularly, this application relates to an efficient method for automatically determining buried cables for cable condition monitoring.

Background Art

[0002] For telecommunications service providers, it is extremely important to have the ability to quickly and reliably identify the location (above or below ground) of each section of telecommunications facilities. Unfortunately, there is no suitable method for determining the location of underground (buried) cables. Therefore, telecommunications service providers need to dispatch service personnel to make direct determinations, which is a time-consuming and costly procedure.

Summary of the Invention

[0003] According to aspects of the present disclosure, advancements in technologies related to distributed fiber optic sensing (DFOS) systems and methods are achieved, adopting artificial intelligence / machine learning (AI / ML) methodologies to provide real-time monitoring of the entire optical fiber cable route and an integrated system and method for automatically and instantaneously (in less than one minute for a 25 km route) identifying buried cables and aerial cables.

[0004] In contrast to the prior art, the systems and methods according to aspects of the present disclosure autonomously determine the location of optical fiber cables using AI / ML methodologies, and advantageously determine the buried portions of the optical fiber cable route and detect changes over time in the condition of the optical fiber cable route. Automated operations are provided by the AI / ML methodologies, which are unsupervised (no manual operation is required when monitoring a new optical fiber cable route), do not require pre-training, do not require human annotation of data collection and classifiers trained for each route individually, and are not affected by on-site environmental conditions.

Brief Description of the Drawings

[0005]

Figure 1(A)

Figure 1(B)

[0006]

Figure 2

[0007]

Figure 3

[0008]

Figure 4

[0009]

Figure 5

[0010]

Figure 6

[0011]

Figure 7

[0012]

Figure 8

[0013]

Figure 9

[0014]

Figure 10

DETAILED DESCRIPTION OF THE INVENTION

[0015] The following merely illustrates the principles of the present disclosure. Thus, it will be understood by those skilled in the art that although not explicitly described or illustrated herein, various configurations that embody the principles of the present disclosure and are within its spirit and scope can be devised.

[0016] Furthermore, all examples and conditional terms described herein are intended solely for the educational purpose of assisting the reader in understanding the concepts contributed by the inventors to facilitate the principles and techniques of the present disclosure, and should not be construed as being limited to such specifically recited examples and conditions.

[0017] Furthermore, all descriptions in this specification that describe the principles, aspects, and embodiments of the present disclosure, as well as specific examples thereof, are intended to encompass both their structural and functional equivalents. Furthermore, such equivalents are intended to include both currently known equivalents and equivalents developed in the future, i.e., developed elements that perform the same function regardless of structure.

[0018] Thus, for example, it will be understood by those skilled in the art that any block diagram in this specification represents a conceptual diagram of an exemplary circuit that implements the principles of the present disclosure.

[0019] Unless otherwise specified herein, the figures constituting the drawings are not drawn to scale.

[0020] As some additional background, note that a distributed fiber optic sensing system interconnects an optoelectronic interrogator to an optical fiber (or cable) and converts the fiber into an array of sensors distributed along the fiber. In practice, the fiber becomes the sensor, and the interrogator generates / injects laser light energy into the fiber and senses / detects events along the fiber.

[0021] As will be understood and appreciated by those skilled in the art, DFOS technology can be deployed to continuously monitor vehicle movement, human traffic, excavation activities, seismic activity, temperature, structural integrity, liquid and gas leaks, and many other conditions and activities. This is used worldwide to monitor power plants, communication networks, railways, roads, bridges, borders, critical infrastructure, onshore and offshore power lines and pipelines, and downhole applications in oil, gas and enhanced geothermal power. Advantageously, distributed fiber optic sensing is not constrained by line of sight or remote power access, and depending on the system configuration, can be deployed in continuous lengths exceeding 30 miles, with sensing / detection possible at all points along that length. Thus, the cost per sensing point over long distances is typically not comparable to competing technologies.

[0022] Distributed fiber optic sensing measures changes in the "backscattering" of light that occur within an optical sensing fiber when the optical sensing fiber encounters environmental changes including events of vibration, strain, or temperature change. As described above, the optical sensing fiber functions as a sensor over its entire length, providing real-time information regarding the physical / environmental surroundings and the integrity / security of the fiber. Further, distributed fiber optic sensing data identifies the exact location of events and conditions occurring at or near the sensing fiber.

[0023] A schematic diagram illustrating a generalized arrangement and operation of a distributed optical fiber sensing system that can advantageously include artificial intelligence / machine learning (AI / ML) analysis is exemplarily shown in FIG. 1(A). Referring to FIG. 1(A), it can be seen that an optical sensing fiber is connected to an interrogator. Although not shown in detail, the interrogator can include an encoded DFOS system that can employ a coherent receiver arrangement known in the art as shown in FIG. 1(B).

[0024] As is well known, a modern interrogator is a system that generates an input signal to an optical sensing fiber, reflects / backscatters it, and then detects / analyzes the received signal. The received signal is analyzed to generate an output indicating the environmental conditions encountered along the fiber. The received backscattered signal may be due to reflections within the fiber, such as Raman backscattering, Rayleigh backscattering, Brillouin backscattering, etc.

[0025] As is understood, modern DFOS systems include an interrogator that periodically generates optical pulses (or any encoded signal) and directs them into an optical sensing fiber. The incident optical pulse signal is transmitted along the optical fiber.

[0026] At positions along the fiber, a small portion of the signal is backscattered / reflected and returned to the interrogator where it is received. The backscattered / reflected signal conveys information that the interrogator uses to detect, for example, changes in power levels indicating mechanical vibrations.

[0027] The received backscattered signal is converted to the electrical domain and processed within the interrogator. Based on the pulse incidence time and the time when the received signal is detected, the interrogator can determine from which position along the optical sensing fiber the received signal has returned, and as a result, sense the activity at each position along the optical detection fiber. Classification methods may further be used to detect events or other environmental conditions, including acoustic and / or vibration and / or heat, along the optical sensing fiber to identify the location.

[0028] Figure 2 is a schematic diagram showing an exemplary operation of automatic buried cable determination according to an aspect of the present disclosure. Referring to this figure, the system and method of the present invention according to an aspect of the present disclosure uses a DFOS system in combination with an AI / ML methodology to provide an integrated system and method for real-time monitoring of the entire path of an optical fiber cable, and a set of functions including the following.

[0029] First, a real-time algorithm (less than 1 minute for a 25 km path) that examines the state of the optical fiber cable and automatically and instantaneously distinguishes between buried cables and aerial cables.

[0030] High-speed allocation AI modules for various applications such as road traffic monitoring AI for buried cables and pole and transformer integrity monitoring AI for aerial cables.

[0031] Resistance to external influences such as environmental noise, signal strength and signal-to-noise ratio (SNR), burial depth of the optical fiber cable, and weather conditions.

[0032] Cable self-monitoring that does not require additional sensors.

[0033] Fast response to cable status that reduces service downtime.

[0034] Figure 3 is a schematic diagram showing an example of a cable condition monitoring application (normal and abnormal) according to an aspect of the present disclosure. As schematically shown in Figure 3, the automatic buried cable position determination has aspects lacking in the provisioning and offering of fiber-optic-based services. Although position determination is very important, it should be noted that the condition monitoring of fiber-optic cables is another important application of the system and method of the present invention. However, according to the approach of the present invention, 1) the buried section can be automatically determined from the cable route based on the sensing data received from the DFOS system, 2) an operation-based baseline can be established from long-term data including the accumulated evidence, and 3) an abnormal condition can be identified by comparing the detection results with the historical baseline using moving windows of different lengths.

[0035] Figure 4 is a schematic flow diagram showing the prior art workflow for manual inspection and determination of the existing fiber-optic cable condition. However, such manual methods are accompanied by many known problems.

[0036] Unknown positions from OTDR traces: The optical time domain reflectometer (OTDR) is the most common technology used to check the loss and length of fiber-optic cables. However, the OTDR trace does not contain information about individual cable sections, and there is no way to identify buried or aerial cables.

[0037] Blind spots in on-site inspections: Due to geographical constraints, some cable spans, such as under rivers, lakes, seas, or in forests, may be inaccessible.

[0038] Inaccurate baseline maps: The baseline map may not be accurate because the information is not up-to-date.

[0039] Efficiency: It is time-consuming and laborious for engineers to conduct on-site inspections manually, resulting in a significant delay in reporting.

[0040] FIG. 5 is a schematic flow diagram showing an exemplary workflow for automatic buried optical fiber cable detection and determination according to an aspect of the present invention.

[0041] FIG. 6 is a schematic diagram showing an exemplary system configuration according to an aspect of the present disclosure. Referring to this figure, it is shown that a distributed fiber optic sensing system (DFOS) (which can be a distributed acoustic sensing (DAS) and / or a distributed vibration sensing (DVS)) is installed in a central control station / central office (CO) for remote monitoring of the entire optical fiber cable path. The DFOS system is shown to be integrated with an optical switch and connected to an optical sensing fiber to provide a sensing function in a plurality of optical fiber paths. Advantageously, the sensing fiber can be a dark fiber or an operating fiber that transmits live communication traffic of a service provider.

[0042] During operation, the AI engine of the present invention employs special filtering and temporal smoothing procedures based on signal characteristics across the spatial, temporal, and frequency domains. In certain applications, in addition to buried cable sections, there may be aerial optical fiber cable sections, buried optical fiber cables that are sometimes exposed on the ground, optical fiber cables arranged in a central office (where there are often generators and air conditioners in a building), or the ends of optical fiber cables. According to an aspect of the present disclosure, these various types of optical fiber sections can be identified as noise by a density-based spatial clustering algorithm. The algorithm learns through self-adjustment regardless of how the overall sensitivity of the optical fiber sensor varies (due to weather, ground conditions, day-night differences, etc.). An important underlying assumption of the method of the present invention is that the intensity from the signal of the buried optical fiber cable section forms a compact cluster in a one-dimensional intensity space that is invariant across different paths.

[0043] The automatic buried cable determination (ABCD) method of the present invention includes the following steps.

[0044] For each chunk of fiber sensing data collected:

[0045] 1: Frequency domain: A special high-pass filter with a cut-point frequency is applied to the sensing signal and the aggregated intensity is calculated at each location every 120 ms.

[0046] 2: Time domain: For each location, a number of summary statistics are calculated based on the intensity over a period of time (e.g., 48 seconds).

[0047] The choice of quantile statistic is related to the parameters of the high-pass filter, depending on the application of interest. Such choices are exemplarily shown in Table 1.

[0048] 3: Spatial domain: Apply the designed median filter to the vector of quantile statistics.

[0049] 4: Unsupervised learning: Run DBSCAN spatial clustering on the filtered vectors and specify the minimum length of the buried interval as a parameter.

[0050] FIG. 7 is a series of plot traces illustrating median filter designs for spatial smoothing of optical fiber data along an optical fiber cable path according to an embodiment of the present invention. The spatial data can be original waterfall data or processed quantile statistics. As shown in plot trace (a), the cable path includes (1) a fiber in a DFOS system that can generate large vibrations continuously, (2) two buried sections, (3) one aerial section, and (4) a section after the fiber end. In a distributed sensor configuration, a maximum sensing monitoring range (e.g., 50 km for a typical commercially available DAS) is specified. In practice, the field cable can be any length that is shorter than the maximum sensing monitoring range.

[0051] At each time point, the sense data can be represented as a vector of length L, as shown in plot trace (b). The goal of the algorithm is to identify buried sites in an unsupervised manner.

[0052] Due to the influence of external factors such as traffic and wind, the intensity along the fiber optic path fluctuates. As shown in plot trace (c), without applying a median filter, the algorithm may generate suboptimal results. To remove these artifacts, a spatial smoothing procedure is applied. However, at both ends of the fiber path, special processing of mirroring is required. This is because waterfall noise within the DFOS is generated at the start point of the data (caused by the vibration of the fan within the DFOS system). It is necessary to mask the noise at the start of the waterfall.

[0053] As shown in plot trace (b), at each end, a vector of length W is concatenated, and the values are mirrored symmetrically with respect to the values of the first W (or last W) that were sensed. After mirroring, the length of the vector becomes L + 2W. The median filter generates the correct result as shown in plot trace (e).

[0054] After spatio - temporal processing of the distributed sensor data, each position is represented as a numerical value representing the average level of vibration intensity. There is no fixed threshold value that can distinguish between the buried cable section and the aerial cable section. There is no time to label the data and train the classifier on the same path. When the classifier is trained on different paths with different sensor configurations, there is no guarantee that the classifier can generalize to a new path. Therefore, the supervised learning approach is not appropriate.

[0055] The number of aerial or buried sections along each path is also unknown. Therefore, many clustering techniques that require specifying the number of clusters in advance cannot be applied.

[0056] Spatial clustering methods such as DBSCAN density - based spatial clustering for applications with noise can be applied to solve this problem because they do not require specifying the number of clusters.

[0057] Parameter Setting I (Number of Samples): The minimum length of the buried section is specified as a parameter of DBSCAN (e.g., 100 meters). This restricts the minimum number of samples (or total weight) in the vicinity of a point considered as a core point.

[0058] Parameter Setting II (Distance between Samples): After frequency domain processing and spatio-temporal smoothing, the maximum variation in the statistics of the buried section in particular is significantly reduced. This parameter specifies the range of intensities between the strongest buried point and the weakest buried point that are considered to be within the same cluster.

[0059] Interpretation of Output: After this step, the vector of the cable route is divided into K clusters and noise (indicated by the number -1), i.e., "-1, 0, 1,..., K". Table 2 summarizes the interpretation of the unique output of DBSCAN. DBSCAN can cover various route conditions regardless of the presence or absence of the aerial section.

[0060] In addition to the fact that the number -1 always indicates noise, it is not yet known which numbers indicate the buried section and which numbers indicate the aerial section.

[0061] 5: Post-Processing (Label Switching): Identify low-intensity clusters as buried sections. To address the label switching problem of the clustering algorithm, a post-processing procedure is added. The pseudo-code of the algorithm is described as follows.

Number

[0062] The cluster with the lowest intensity as a "buried" section is labeled "0". The probability that a certain position is "buried" is defined by the ratio of "0" in the entire CSV file.

[0063] Note that the same numerical value 0 may be assigned to multiple embedded sections. To obtain the number of embedded sections in the path, as shown in FIG. 8, a continuity detection step can be applied along the path. FIG. 8 is a schematic diagram showing an exemplary continuity detection for calculating the total number of embedded sections according to an aspect of the present invention.

[0064] 6: Accumulate evidence. The detection results can be accumulated over multiple periods. At each position a

Number

[0065] The flowchart of the embedded cable determination algorithm is shown in detail in FIG. 9. FIG. 9 is a schematic flowchart showing an exemplary data processing workflow for automatic embedded optical fiber cable detection and determination according to an aspect of the present invention. When processing long-term data (e.g., 24×7 hours) offline, a huge number of data files may be created. Using multiprocessing, they can be processed in parallel.

[0066] In the case of a long-term cable condition monitoring application, an embedded cable determination method can be applied to analyze the data stream within a moving window. Compare the latest data with a history-based baseline to detect changes in the state. When two periods are selected, the short-term state of the cable position can be compared with the state over a longer period at the same position to check for changes.

[0067] This method is shown in detail in FIG. 10. FIG. 10 is a schematic diagram showing an exemplary long-term monitoring of the state change of an optical fiber cable for automatic detection and determination of an embedded optical fiber cable according to an aspect of the present invention.

[0068] As shown in FIG. 10, a subsampling procedure is used to reduce the computational cost. At each position a, the algorithm of the present invention is executed once every period T. The results are used to update the short-term average result and the long-term average result. A change point detection method such as the short-term average / long-term average (STA / LTA) method is used to determine whether the state at position a has changed.

[0069] In evaluating the method of the present invention, in the test, a waterfall image of about 1.5 minutes was generated for a 17 km route including an optical fiber cable in a cross-connect box, a buried section, an aerial section, a central office section, and a section beyond the far end (fiberless section). From the data, it was found that the intensity and distribution characteristics of the noise were different for each section. When the buried cable determination algorithm was executed, a result for one line was generated. The results for multiple periods were accumulated, and the ratio of the time assigned as a buried cable for each position was calculated. It was found that there are three buried cable sections of 117 - 200 m, 2 - 6.8 km, and 6.9 - 16.5 km in this route, thereby verifying the method of the present invention.

[0070] So far, the present disclosure has been presented using several specific examples, but those skilled in the art will recognize that the present teachings are not so limited. Therefore, the present disclosure should be limited only by the scope of the claims appended hereto.

Table 1

Table 2

Claims

1. An automated method for determining the location of a buried optical fiber cable, comprising: an optical sensor fiber; an optical interrogator that optically communicates with the optical sensor fiber, the optical interrogator being configured to generate optical pulses from a laser beam, capture the pulses in the optical sensor fiber, and receive a backscattered signal from the optical sensor fiber; an analyzer for analyzing the received backscattered signal, the received backscattered signal being generated at a plurality of positions along the optical sensor fiber; providing a distributed fiber optic sensing (DFOS) system having; continuously operating the DFOS system to collect received backscattered signals from the plurality of positions along the optical sensing fiber, analyzing the received backscattered signals by applying a high-pass filter of a cut-off frequency to the received backscattered signals to determine an aggregated intensity at each of the plurality of positions; for each of the plurality of positions, determining a summary statistic based on the signal intensity over a certain period; applying a median filter to the determined summary statistic; performing spatial clustering on the median-filtered and determined summary statistic; identifying low intensity as a buried location; a method including reporting the buried location.

2. The method according to claim 1, further comprising adding an additional optical sensor fiber to the optical sensor fiber, identifying an additional buried location, and reporting the additional buried location.

3. The method according to claim 1, including generating a baseline for the buried location.

4. The method according to claim 3, including comparing the generated baseline with a newly identified buried location and reporting the newly identified buried location if the newly identified location is not included in the baseline.

5. The method according to claim 3, further comprising identifying a buried location exposed on the ground surface and an aerial cable in contact with the ground surface.

Images