An undersea cable data management method and system based on the Internet of Things

Abstract

This invention relates to the field of data management, and more particularly to a method and system for managing submarine optical cable data based on the Internet of Things (IoT). The method includes: acquiring monitoring signals from multiple monitoring points deployed along the submarine optical cable; calculating the fluctuation intensity and disturbance accumulation of the monitoring signals at each moment; wherein the fluctuation intensity is the energy difference between the monitoring signal within a window of a set length and the energy of the monitoring signal within the previous window; calculating the spatial disturbance factor for each monitoring point; calculating an adaptive time window length based on the disturbance accumulation and spatial disturbance factor; wherein the adaptive time window length is negatively correlated with the disturbance accumulation and spatial disturbance factor; and extracting a portion of the monitoring signal equal to the adaptive time window length, inputting the extracted monitoring signal into an LSTM network model to obtain the health status of the submarine optical cable. This invention improves the detection performance of the LSTM model.

Description

Technical Field

[0001] This invention relates to the field of data management, and in particular to a method and system for managing submarine optical cable data based on the Internet of Things. Background Technology

[0002] Submarine optical cables are a crucial component of communication data, and their security and reliability are of paramount importance. Traditional management of submarine optical cables relies on periodic manual inspections, which suffers from problems such as delayed response, high costs, and lack of early warning. With the development of IoT technology, submarine optical cable management systems deploy distributed fiber optic sensor networks (such as distributed acoustic sensors (DAS) and distributed vibration sensors (DVS)) along the cables and integrate environmental monitoring sensors along the route, such as those for temperature, pressure, and ocean currents, to construct an underwater IoT system. This IoT system generates multi-source, heterogeneous, and multi-dimensional time-series data streams every moment.

[0003] Submarine fiber optic cables are susceptible to external environmental influences, such as: the instantaneous impact of ship anchoring, minor submarine landslides, continuous scouring by ocean currents causing the cable to gradually become suspended, long-term minor disturbances from third-party construction activities, and continuous stress changes caused by geological subsidence. The external environment significantly impacts the health of fiber optic cables, thus necessitating monitoring. Long Short-Term Memory (LSTM) networks can learn data patterns of fiber optic cables under normal conditions and identify abnormal disturbances. Existing data analysis methods based on LSTM network models typically employ a fixed, pre-defined time window to capture and analyze the data stream.

[0004] However, the fixed-length time window makes it unable to adapt to the diverse events sensed by the Internet of Things. If the window is set too short, for slow-evolving anomalies with long durations (such as the continuous scouring of ocean currents causing optical cables to gradually become suspended, or the continuous weak strain caused by geological subsidence), the model will be unable to observe the complete anomaly pattern due to the limited field of view, resulting in missed detections. Conversely, if the window is set too long, when a sudden impact occurs, the window will contain a large amount of irrelevant normal data. The normal data dilutes and submerges the transient anomaly signal features, causing the model's sensitivity to decrease and resulting in misjudgments. Summary of the Invention

[0005] To address the technical problem of poor detection performance caused by the fixed-length temporal receptive field of LSTM models, this invention provides a method and system for managing submarine optical cable data based on the Internet of Things.

[0006] In a first aspect, the present invention provides a method for managing submarine optical cable data based on the Internet of Things, employing the following technical solution:

[0007] Acquire monitoring signals from multiple monitoring points deployed along the coastal fiber optic cable; calculate the fluctuation intensity and disturbance accumulation of the monitoring signal at each moment, wherein the disturbance accumulation is the weighted sum of the fluctuation intensity at the current moment and the disturbance accumulation at the previous moment; the fluctuation intensity is the energy difference between the monitoring signal within a window of a set length and the monitoring signal within the previous window;

[0008] Calculate the spatial disturbance factor for each monitoring point. The spatial disturbance factor is negatively correlated with the mean correlation between the monitoring signals of the monitoring point and its neighboring monitoring points. Calculate the adaptive time window length based on the disturbance accumulation and the spatial disturbance factor. The adaptive time window length is negatively correlated with the disturbance accumulation and the spatial disturbance factor.

[0009] The monitoring signal with a length equal to the adaptive time window length is extracted, and the extracted monitoring signal is input into the LSTM network model to obtain the health status of the submarine optical cable.

[0010] By comprehensively calculating the perturbation accumulation degree, which reflects the temporal variation of the signal, and the spatial perturbation factor, which reflects the spatial range of the event, the adaptive time window length is jointly determined. For instantaneous impacts, the window can automatically shorten to focus on anomalous features and avoid being diluted by normal data; for slow deformations, the window can automatically lengthen to capture the complete process and avoid missed detections due to limited field of view. This improves the accuracy and flexibility of the LSTM model in detecting different types of anomalous events.

[0011] Preferably, the method for calculating the fluctuation intensity is as follows: taking the corresponding time as the endpoint, two segments of signal with equal length are intercepted along the historical direction, the cumulative sum of the amplitude at each time point within each monitoring signal segment is calculated, and the difference between the two cumulative sums is taken as the fluctuation intensity at the corresponding time.

[0012] The fluctuation intensity is quantified by calculating the difference in energy of the monitored signal within two adjacent time windows. This method is more sensitive to changes in the signal and can more effectively highlight instantaneous energy surges caused by events such as ship anchoring or sand and gravel impacts. It provides a more accurate input for subsequent calculation of disturbance accumulation, thereby improving the response speed and identification capability for sudden events.

[0013] Preferably, the expression for the perturbation cumulancy is:

[0014]

[0015] In the formula, This represents the cumulative degree of disturbance at time point t. This represents the cumulative degree of disturbance at time point t-1. This indicates the intensity of the fluctuation in the monitored signal at time point t. This represents the preset forgetting factor at time point t.

[0016] The exponentially weighted moving average algorithm with a forgetting factor is adopted to make the system pay more attention to recent disturbances and smooth out occasional noise fluctuations. This allows the disturbance accumulation to reflect the cumulative effect of persistent anomalies (such as continuous friction) and to gradually decay after a rapid response to instantaneous impacts, thus more accurately characterizing the current overall disturbance status of the optical cable.

[0017] Preferably, the forgetting factor is calculated as follows: taking time point t as the endpoint, extend the time length k to the left to obtain the time period (tk, t), and obtain the fluctuation intensity sequence of the monitoring signal corresponding to the time period (tk, t). ,calculate percentile rank The percentile rank is used as the exponent of the preset baseline forgetting factor to adjust the baseline forgetting factor, thus obtaining the forgetting factor at time point t.

[0018] The forgetting factor is dynamically adjusted based on the percentile rank of the current fluctuation intensity in historical data. When there is a sharp fluctuation, the percentile rank is high and the forgetting factor decreases, allowing the system to quickly forget the past and focus on the current impact. When the system is stable, the forgetting factor is close to the baseline value, accumulating more historical information, making the calculation of disturbance accumulation more intelligent and accurate, and improving the model's adaptability to events of different intensities.

[0019] Preferably, the spatial disturbance factor is calculated as follows: For any monitoring point i on the submarine optical cable, multiple neighboring monitoring points are obtained by extending a certain distance on both sides of monitoring point i as the center. The absolute value of the Pearson correlation coefficient between the signal of monitoring point i and the neighboring monitoring point j is calculated. The mean of the absolute values ​​of the multiple Pearson correlation coefficients is further obtained. The difference between the value 1 and the mean of the absolute values ​​of the Pearson correlation coefficients is obtained. The difference is normalized to obtain the spatial disturbance factor.

[0020] By calculating the Pearson correlation between the central monitoring point and its two adjacent points, it is possible to determine whether a disturbance is local noise or a large-scale, persistent event. This introduces spatial dimension information. When the correlation is high (indicating a large-scale event), the spatial disturbance factor decreases, which helps the system make decisions to use a longer time window for observation, thereby effectively distinguishing between local anomalies and potential risks with a wider impact.

[0021] Preferably, the spatial disturbance factor is calculated as follows: for any monitoring point i on the submarine optical cable, multiple neighboring monitoring points are obtained by extending a certain distance along one side from monitoring point i as the endpoint, and a hysteresis window is constructed. p represents the lag time length; calculate the cross-correlation coefficient between monitoring point i and neighboring monitoring point j under different lag time lengths, and take the lag time length when the cross-correlation coefficient reaches its maximum value as the response time length. Calculate the mean of the absolute values ​​of multiple response time lengths. , the mean The normalized result is used as the spatial perturbation factor.

[0022] By calculating the cross-correlation and response time of signals between the monitoring point and neighboring points, it is suitable for detecting events propagating along the direction of the optical cable (such as ocean currents scouring parallel to the optical cable), providing another dimension of evidence for judging the physical causes of abnormal events and enhancing the system's ability to identify and adapt to specific types of events.

[0023] Preferably, the expression for the adaptive time window length is:

[0024] ;

[0025] In the formula, This represents the length of the adaptive time window corresponding to time point t. This represents the cumulative degree of disturbance at time point t. This indicates the preset minimum time window length. This indicates the preset maximum time window length. This represents the spatial perturbation factor. The symbol represents the floor function, and sigmoid represents the activation function.

[0026] By combining the cumulative perturbation degree in the time dimension and the perturbation factor in the spatial dimension, it is ensured that the window length smoothly approaches the maximum value when the perturbation is small and quickly converges to the minimum value when the perturbation is large. This makes the adjustment of the window length both sensitive and stable, and achieves precise control over the receptive field of the analysis window.

[0027] Preferably, the health status of the submarine optical cable includes: normal status, instantaneous impact, continuous friction, and slow deformation.

[0028] Preferably, the monitoring signal is preprocessed. The preprocessing method is as follows: the monitoring signal is filtered using a bandpass filter, and then the filtered monitoring signal is normalized.

[0029] By employing bandpass filtering and normalization, environmental noise and high-frequency interference unrelated to the optical cable's condition can be effectively filtered out, and the impact of signal amplitude differences at different monitoring points or time periods can be eliminated. This ensures the quality of data input into subsequent analysis models, improves the signal-to-noise ratio, and lays a solid foundation for the accuracy and stability of subsequent feature calculations and condition assessments.

[0030] Secondly, the present invention provides a submarine optical cable data management system based on the Internet of Things, which adopts the following technical solution:

[0031] A submarine optical cable data management system based on the Internet of Things (IoT) includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the aforementioned submarine optical cable data management method based on the Internet of Things.

[0032] The aforementioned IoT-based submarine optical cable data management method generates a computer program, which is stored in a memory for loading and execution by a processor. Thus, a system is created based on the memory and processor for convenient use.

[0033] The present invention has the following technical effects:

[0034] This invention dynamically adjusts the data window length input to the LSTM network model by calculating the cumulative temporal disturbance and spatial correlation of the monitored signal in real time. When the disturbance is severe, the window automatically shortens to focus; when the disturbance is mild and widespread, the window automatically lengthens for comprehensive observation. This improves the accuracy and flexibility of detecting diverse abnormal events such as instantaneous impacts and continuous friction, and further enhances the detection performance of the LSTM model. Attached Figure Description

[0035] Figure 1 This is a flowchart of a submarine optical cable data management method based on the Internet of Things according to the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] This invention discloses a method for managing submarine optical cable data based on the Internet of Things (IoT), referring to... Figure 1 The process includes steps S1-S5, as detailed below:

[0038] S1: Acquire monitoring signals and perform preprocessing.

[0039] By deploying an Internet of Things (IoT) sensing system along the submarine optical cable, multi-source monitoring signals are acquired. These signals include high-frequency vibration data and acoustic time-series data. The monitoring signals are collected by distributed optical fiber sensing devices (such as DAS), with each device serving as a monitoring point. Multiple monitoring points are evenly distributed along the length of the submarine optical cable. The monitoring signals are pre-processed using a bandpass filter to remove environmental noise and high-frequency white noise. Then, a linear normalization algorithm is used to normalize the filtered monitoring signals, resulting in the processed monitoring signals.

[0040] S2: Calculate the fluctuation intensity of the monitoring signal at each time point.

[0041] Submarine optical cables are subjected to severe physical impacts, such as anchor dragging or sand and gravel impacts, which cause significant vibrations in the cables. This is reflected in the monitoring signal as a sudden surge in energy. Under stable conditions, the energy changes in the monitoring signal are relatively gradual. Therefore, by calculating the intensity of the fluctuations in the monitoring signal, we can understand the changes in the energy of the monitoring signal.

[0042] The expression for wave intensity is:

[0043]

[0044] In the formula, This indicates the intensity of the fluctuation in the monitored signal at time point t. This indicates the amplitude of the monitored signal at time point k, and w represents the preset number of time points. For example, the number of time points is 100.

[0045] Calculate the absolute value of the energy difference between two adjacent time windows of length w. When the monitoring signal changes drastically due to external environmental influences (such as the appearance of a sharp pulse), the energy of adjacent windows will produce a large difference, leading to... The value of suddenly becomes very large; conversely, when the monitoring signal changes steadily or slowly, The value will be in a lower range.

[0046] In conclusion, when A larger value indicates that the submarine cable is more likely to be affected by the external environment at time point t, or that the health status of the submarine cable has changed at time point t; when When the value is small, it indicates that the submarine cable is less affected by the external environment at time point t, and the submarine cable is in a good health condition.

[0047] S3: Calculate the cumulative disturbance of the monitoring signal at each time step.

[0048]

[0049] In the formula, This represents the cumulative degree of disturbance at time point t. This represents the cumulative degree of disturbance at time point t-1. This indicates the intensity of the fluctuation in the monitored signal at time point t. This represents the forgetting factor at time point t.

[0050] Understandably, each calculation point corresponds to a disturbance accumulation degree. The disturbance accumulation degrees of multiple consecutive time points constitute a disturbance sequence, which is used to construct a disturbance curve. When the submarine cable is subjected to a brief impact at time point t, the fluctuation intensity of the monitoring signal will increase instantaneously, further leading to an instantaneous increase in the disturbance accumulation degree. Subsequently, due to the effect of the forgetting factor α, the disturbance accumulation degree gradually decays over the following time, which is represented by a peak at time point t in the disturbance curve, followed by a curve that tends to flatten out. When the submarine cable is subjected to a sustained low-intensity impact, the fluctuation intensity of the monitoring signal will increase slowly, further leading to a slow increase in the value of the disturbance accumulation degree. Under the effect of the forgetting factor α, the disturbance accumulation degree remains at a high level over the following time, which is represented by a curve that gradually and steadily rises and then approaches a straight line in the disturbance curve.

[0051] In one embodiment, the forgetting factor is calculated as follows: taking time point t as the endpoint, extend the time length k to the left to obtain the time period (tk, t), and obtain the fluctuation intensity sequence of the monitoring signal corresponding to the time period (tk, t). ,calculate percentile rank The expression for the forgetting factor is: In the formula, This represents the forgetting factor at time point t. This represents a preset baseline forgetting factor, the value of which is manually set according to actual conditions, and its value is less than 1. For example... The value is 0.95. This represents the percentile rank of the monitored signal fluctuation intensity at time point t. The percentile rank indicates the position of the signal fluctuation intensity below a certain threshold in the entire fluctuation intensity sequence. The percentage of data represented is calculated using existing technology, and the specific calculation steps will not be elaborated here.

[0052] When a submarine optical cable is suddenly subjected to a violent physical impact, the fluctuation intensity of the monitoring signal increases abruptly, leading to an increase in the percentile rank at the corresponding time point. At this time, the forgetting factor decreases, indicating rapid forgetting of historical events and a greater focus on the current time point. When the submarine optical cable is in a relatively stable environment, the fluctuation intensity of the monitoring signal is smaller, and the forgetting factor at each time point approaches the baseline forgetting factor, indicating a greater focus on the accumulation of historical disturbances. This method yields highly accurate results and is suitable for scenarios with stringent requirements for cable condition monitoring.

[0053] In another embodiment, the forgetting factor at each time point is set to a fixed value. This method has a low computational load and is suitable for scenarios where the cable condition monitoring results are relatively lenient.

[0054] S4: Calculate the adaptive time window length.

[0055] When using LSTM network models to analyze the state of submarine optical cables, the interference generated by different types of events (such as anchor impact, sand and gravel impact, cable suspension, ocean current scouring, and geological subsidence) requires input data of different time lengths. Therefore, it is necessary to dynamically adjust the receptive field corresponding to the event to improve the classification results of the LSTM network model. For example, for disturbances with high cumulative intensity (usually corresponding to violent or ongoing events), a shorter time window is needed for focused analysis to avoid information being diluted by normal signals; for disturbances with low cumulative intensity (usually corresponding to stable states or the early stages of slow events), a longer time window is needed for comprehensive observation to capture complete periodic changes.

[0056] S41: Calculate the spatial disturbance factor of the monitoring point.

[0057] In one embodiment, for any monitoring point i on the submarine optical cable, multiple neighboring monitoring points are obtained by extending a certain distance on both sides of monitoring point i as the center. For example, for the 10th monitoring point, two monitoring points are extended to one side and two monitoring points are extended to the other side to obtain four neighboring monitoring points, namely the 8th, 9th, 11th, and 12th monitoring points. The absolute value of the Pearson correlation coefficient between the signal of monitoring point i and the neighboring monitoring point j is calculated, and the mean of the absolute values ​​of multiple Pearson correlation coefficients is further obtained. The expression for the spatial disturbance factor is: ;in, This represents the spatial perturbation factor. The absolute mean of the Pearson correlation coefficient between the monitoring point and the signals of neighboring monitoring points is represented by , and sigmoid represents the activation function used for normalization.

[0058] The Pearson correlation coefficient represents the correlation between the monitoring signal of monitoring point i and its neighboring monitoring point j. A value closer to 1 indicates a similarity between the monitoring signal of monitoring point i and that of neighboring monitoring point j, suggesting a greater likelihood that the submarine cable is affected by a large-scale, sustained event (such as ocean current erosion or geological subsidence). Therefore, it is necessary to increase the length of the time window to observe the complete development pattern of the event. In summary, a smaller spatial disturbance factor indicates a greater correlation between the monitoring point and its neighboring monitoring points, and vice versa. This calculation method is suitable for scenarios where the direction of ocean current erosion is approximately perpendicular to the direction of optical cable deployment.

[0059] In another embodiment, for any monitoring point i on the submarine optical cable, multiple neighboring monitoring points are obtained by extending a certain distance along one side from monitoring point i as the endpoint, and a hysteresis window is constructed. Let p be the lag time length, defined as follows: a negative value indicates advance, and a positive value indicates delay. Calculate the cross-correlation coefficient between monitoring point i and its neighboring monitoring point j under different lag time lengths. The lag time length at which the cross-correlation coefficient reaches its maximum value is taken as the response time length. This means that monitoring point i and any neighboring monitoring point have a response time length. Calculate the mean of the absolute values ​​of multiple response time lengths. The expression for the spatial perturbation factor is: ;in, The sigmoid function represents the spatial perturbation factor and is used for normalization.

[0060] The cross-correlation coefficient measures the similarity between two signals at different lags p. A value closer to 1 indicates that the signal from monitoring point i is similar to that of its neighboring monitoring point j. The lag length indicates the greater the likelihood that monitoring point i is affected by large-scale, sustained events (such as ocean current erosion or geological subsidence) earlier (or later) than its neighboring monitoring point j. Therefore, it is necessary to increase the length of the time window based on the lag length to observe the complete development pattern of the event. This calculation method is suitable for scenarios where the direction of ocean current erosion is approximately parallel to the direction of optical cable deployment.

[0061] S42: Calculate the adaptive time window length based on the spatial perturbation factor and perturbation cumulative degree.

[0062] The expression for the adaptive time window length is:

[0063]

[0064] In the formula, This represents the length of the adaptive time window corresponding to time point t. This represents the cumulative degree of disturbance at time point t. This indicates the preset minimum time window length. This indicates the preset maximum time window length. This represents the spatial disturbance factor, used to adjust the sensitivity of the time window length to the cumulative disturbance. Its value is set manually according to the actual situation. The symbol represents the floor function, and sigmoid represents the activation function used for normalization.

[0065] when When the value is small (approaching 0, indicating a very stable system), the denominator is close to 1. As it approaches its maximum value, the LSTM network model will monitor with a longer field of view. When When the value is large (indicating a violent event), the denominator increases. As it approaches its minimum value, the LSTM network model switches to a short field of view for focused analysis.

[0066] Furthermore, when the correlation between a monitoring point and its neighboring monitoring points is high, it indicates a large-scale, continuous event. In this case, the adaptive time window length is increased, allowing the LSTM network model to monitor with a longer field of view. When the correlation between a monitoring point and its neighboring monitoring points is low, it indicates a localized event. In this case, the adaptive time window length is decreased, allowing the LSTM network model to focus on monitoring with a shorter field of view.

[0067] S5: Predict the status of submarine optical cables based on the adaptive time window length.

[0068] Monitoring signals within an adaptive time window are extracted and input into an LSTM network model. The LSTM network model outputs a classification result of the submarine optical cable's status, including: normal status, instantaneous impact, continuous friction, and slow deformation. For example, if the adaptive time window length at time point t is 100, then inputting time points within the range (t-100, t) into the LSTM network model yields the health status of the submarine optical cable at time point t.

[0069] This invention also discloses an IoT-based submarine optical cable data management system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an IoT-based submarine optical cable data management method according to the present invention.

[0070] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0071] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for managing submarine optical cable data based on the Internet of Things, characterized in that, Including the following steps: Acquire monitoring signals from multiple monitoring points deployed along the coastal fiber optic cable; calculate the fluctuation intensity and disturbance accumulation of the monitoring signal at each moment, where the disturbance accumulation is the weighted sum of the fluctuation intensity at the current moment and the disturbance accumulation at the previous moment; the fluctuation intensity is the energy difference between the monitoring signal within a window of a set length and the monitoring signal within the previous window. Calculate the spatial disturbance factor for each monitoring point: For any monitoring point i on the submarine optical cable, extend a certain distance on both sides of monitoring point i to obtain multiple neighboring monitoring points. Calculate the absolute value of the Pearson correlation coefficient between the signal of monitoring point i and the neighboring monitoring point j. Further obtain the mean of the absolute values ​​of multiple Pearson correlation coefficients. Subtract the value 1 from the mean absolute value of the Pearson correlation coefficients to obtain the difference. Normalize the difference to obtain the spatial disturbance factor. The spatial disturbance factor is negatively correlated with the mean correlation between the monitoring signals of the monitoring point and its neighboring monitoring points. Based on the disturbance accumulation degree and the spatial disturbance factor, calculate the adaptive time window length. In the formula, This represents the length of the adaptive time window corresponding to time point t. This represents the cumulative degree of disturbance at time point t. This indicates the preset minimum time window length. This indicates the preset maximum time window length. This represents the spatial perturbation factor. The symbol indicates the rounding up sign, and sigmoid represents the activation function; the adaptive time window length is negatively correlated with the perturbation accumulation and the spatial perturbation factor; The monitoring signal with a length equal to the adaptive time window is extracted and input into the LSTM network model to obtain the health status of the submarine optical cable.

2. The method for managing submarine optical cable data based on the Internet of Things according to claim 1, characterized in that, The method for calculating the fluctuation intensity is as follows: taking the corresponding time as the endpoint, two segments of the signal of equal length are intercepted along the historical direction, the cumulative sum of the amplitudes at each time point within each monitoring signal segment is calculated, and the difference between the two cumulative sums is taken as the fluctuation intensity at the corresponding time.

3. The method for managing submarine optical cable data based on the Internet of Things according to claim 1, characterized in that, The expression for the perturbation accumulation is: ； In the formula, This represents the cumulative degree of disturbance at time point t. This represents the cumulative degree of disturbance at time point t-1. This indicates the intensity of the fluctuation in the monitored signal at time point t. This represents the preset forgetting factor at time point t.

4. The method for managing submarine optical cable data based on the Internet of Things according to claim 3, characterized in that, The forgetting factor is calculated as follows: taking time point t as the endpoint, extend the time length k to the left to obtain the time interval (tk, t), and then obtain the fluctuation intensity sequence of the monitoring signal corresponding to the time interval (tk, t). ,calculate The percentile rank is used as the exponent of the preset baseline forgetting factor to adjust the baseline forgetting factor, thus obtaining the forgetting factor at time point t.

5. The method for managing submarine optical cable data based on the Internet of Things according to claim 1, characterized in that, The health status of the submarine optical cable includes: normal status, instantaneous impact, continuous friction, and slow deformation.

6. The method for managing submarine optical cable data based on the Internet of Things according to claim 1, characterized in that, The monitoring signal is preprocessed by using a bandpass filter to filter the monitoring signal, and then normalizing the filtered monitoring signal.

7. A submarine optical cable data management system based on the Internet of Things, characterized in that, include: A processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement the Internet of Things-based submarine optical cable data management method according to any one of claims 1-6.

Images