A method and system for online monitoring of submarine cables

Abstract

This invention belongs to the field of submarine cable monitoring technology, specifically relating to an online monitoring method and system for submarine cables, comprising the following steps: S1, collecting real-time vibration sequences, real-time temperature sequences, and real-time current load sequences at the location of submarine cable joints, and performing spatiotemporal alignment and preprocessing on the multi-source heterogeneous data to obtain a monitoring dataset with consistent time granularity; S2, based on the physical driving law of real-time current load on vibration amplitude under normal operating conditions, and the coupling correlation between temperature change and vibration response, constructing a thermal-vibration coupled breathing entropy using the monitoring dataset. This invention, by constructing physical coupling indicators and combining them with adaptive algorithms, solves the problem of false alarms caused by electrodynamic vibration under high load conditions in existing technologies, while effectively amplifying early weak fault characteristics, achieving accurate monitoring of the health status of submarine cable joints.

Description

Technical Field

[0001] This invention belongs to the field of submarine cable monitoring technology, specifically relating to an online monitoring method and system for submarine cables. Background Technology

[0002] Submarine cables are the core hubs for cross-sea power transmission and communication. Their operating environment is extremely harsh, typically buried deep in the seabed, and subjected to high voltage, low temperature, and complex mechanical stress. Among these components, the cable joint is the weakest link in the entire transmission chain, with the most complex insulation structure, and is also the most prone to failure. In actual operation, cable joints are subjected to thermal stress from high voltage and high current, as well as mechanical stress from ocean currents and tidal oscillations, resulting in a complex thermo-mechanical coupling state.

[0003] Current technologies primarily rely on distributed fiber optic temperature sensing (DTS) systems and distributed fiber optic acoustic sensing (DAS) systems for independent monitoring of connectors. However, two significant drawbacks exist in practical applications:

[0004] First, false alarms occur frequently under high loads. Existing DAS systems typically trigger alarms based on a single vibration energy threshold. When submarine cables transmit large currents, the enormous electrodynamic force causes normal, low-amplitude, high-frequency vibrations in the connector conductors. This physical phenomenon causes vibration data amplitudes to spike, which can easily be misinterpreted by existing algorithms as external damage or internal discharge, resulting in numerous false alarms. This forces maintenance personnel to frequently reduce the load for inspections, severely impacting production efficiency.

[0005] Second, early, gradual fault characteristics are often masked. Insulation degradation in submarine cable joints is often a slow process, such as water tree growth or loosening of minute structures. The anomalous signals generated by these faults are extremely weak and are often obscured by seabed environmental noise and normal operating condition fluctuations. Existing anomaly detection algorithms, such as Isolated Forest, typically treat temperature and vibration as independent features, ignoring the nonlinear physical coupling between the two under different loads, making it difficult to detect early degradation trends. Summary of the Invention

[0006] This invention provides a method and system for online monitoring of submarine cables to solve the technical problems of existing technologies that are prone to false alarms under high loads and have difficulty in detecting early and subtle faults.

[0007] In a first aspect, the present invention provides a method for online monitoring of submarine cables, comprising the following steps:

[0008] S1. Collect real-time vibration, temperature and current load sequences at the location of submarine cable joints, and perform spatiotemporal alignment and preprocessing on the multi-source heterogeneous data to obtain a monitoring dataset with consistent time granularity.

[0009] S2, based on the physical driving law of the real-time current load on the vibration amplitude under normal working conditions, and the coupling relationship between temperature change and vibration response, the thermal-vibration coupling breathing entropy is constructed using the monitoring dataset. The thermal-vibration coupling breathing entropy is used to evaluate the dynamic response state of submarine cable joints to environmental and electrodynamic excitation.

[0010] S3. Construct an isolated forest model based on load weights. Input the calculated thermal-vibration coupling respiratory entropy into the isolated forest model. Combine the matching degree between real-time current load and historical average load to calculate the corrected path length of the sample in the isolation tree, and then obtain the anomaly score.

[0011] S4. Based on the calculated anomaly score, the operating status of the submarine cable joint is evaluated at multiple levels. When the anomaly score exceeds the preset threshold, the corresponding early warning or alarm operation is triggered.

[0012] Furthermore, spatiotemporal alignment and preprocessing are performed on multi-source heterogeneous data, including:

[0013] Time synchronization of real-time vibration sequences, real-time temperature sequences, and real-time current load sequences based on GPS timestamps;

[0014] An interpolation algorithm is used to upsample the real-time temperature sequence and real-time current load sequence with low sampling rates to make them consistent with the time granularity of the real-time vibration sequence.

[0015] Wavelet packet decomposition was performed on the real-time vibration sequence to filter out low-frequency ocean wave background noise, retaining the characteristic frequency band energy representing the joint body structure response, and the vibration energy sequence was calculated. .

[0016] Furthermore, the thermal-vibration coupled respiratory entropy The calculation formula satisfies:

[0017]

[0018] In the formula, At the current sampling time, The length of the sliding time window. This is the offset within the window. For a moment The change in vibrational energy, For a moment Temperature change For a moment Current load, For electrodynamic influencing factors, As the reference stability constant, The coefficient of thermal inertia.

[0019] Furthermore, the change in vibrational energy and temperature change By calculating the vibration energy sequence respectively and temperature sequence First-order difference acquisition; reference stability constant A value of 1.0 is used to prevent current loads from being applied. When the denominator is 0, the fraction is 0.

[0020] Furthermore, an isolation forest model based on load weights is constructed, including:

[0021] Data sets from the historical normal operation period of submarine cable joints were selected as training samples.

[0022] During the construction of the isolation tree, for each leaf node in the tree, the average current load of all training samples falling into that leaf node is calculated and stored, denoted as . .

[0023] Furthermore, the corrected path length of the sample in the isolation tree is calculated. Its calculation formula satisfies:

[0024]

[0025] In the formula, Samples calculated by the original Isolation Forest algorithm Path length in the tree For the sample Real-time current load, For the sample The historical average load recorded by the falling leaf node, This represents the maximum design load of the submarine cable system. For the sample Thermo-vibrational coupling respiratory entropy, To adjust the weighting coefficients, It is a non-linear sensitivity index. It is a very small positive number.

[0026] Furthermore, the calculation logic for correcting the path length includes:

[0027] pass The assessment evaluates the degree of matching between the current operating conditions and historical operating conditions, and applies physical model correction only when the load matching degree is high;

[0028] Using logarithmic terms Modulate the corrected path length when When the value approaches 0, increase the correction path length to suppress false alarms; when When the value is large, reduce the length of the correction path to amplify the fault characteristics.

[0029] Furthermore, based on the calculated anomaly scores, the operational status of the submarine cable joints is assessed at multiple levels, including:

[0030] If the abnormal score is below the first-level threshold, it is considered to be in a healthy state;

[0031] If the abnormal score is between the first-level threshold and the second-level threshold, it is judged as a state of concern, indicating the existence of early potential risks;

[0032] If the abnormal score is higher than the secondary threshold, it is determined to be in a warning state, triggering an alarm and displaying the location of the fault.

[0033] Secondly, the present invention provides an online monitoring system for submarine cables, including a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned online monitoring method for submarine cables is implemented.

[0034] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described online monitoring method for submarine cables.

[0035] The beneficial effects are:

[0036] 1. Solves the problem of false alarms under high load: This invention utilizes physical laws to construct a thermal-vibration coupled breathing entropy, in which the load decoupling term uses current intensity to dynamically weight the change in vibration energy. In actual monitoring, even when the submarine cable is operating at full load and the vibration amplitude increases significantly, the calculated breathing entropy index can still remain stable and low, completely eliminating false alarms caused by electrodynamic vibration.

[0037] 2. Achieves rapid response to early and minor faults: This invention amplifies the nonlinear decoupling signal between temperature and vibration through the thermal-vibration coupling breathing entropy index. Once a slight change occurs in the internal structure of the joint, this index will show a vertical explosive growth, which can detect potential problems several hours or even days earlier than traditional methods that rely solely on temperature or vibration thresholds. Attached Figure Description

[0038] Figure 1 This is a flowchart of an online monitoring method for submarine cables.

[0039] Figure 2 This is a schematic diagram of the monitoring results of a single vibration feature in existing technologies.

[0040] Figure 3 This is a schematic diagram of the thermal-vibration coupling respiratory entropy monitoring results of the present invention. Detailed Implementation

[0041] 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.

[0042] An embodiment of the online monitoring method for submarine cables provided by this invention:

[0043] like Figure 1 As shown, the online monitoring method for submarine cables includes the following steps:

[0044] S1 collects real-time vibration, temperature, and current load sequences at the location of submarine cable joints, and performs spatiotemporal alignment and preprocessing on the multi-source heterogeneous data to obtain a monitoring dataset with consistent time granularity.

[0045] Specifically, multi-dimensional monitoring data of the submarine cable joint location is acquired in real time using composite fiber optic sensors coaxially deployed with the submarine cable. A distributed fiber optic acoustic sensing system is used to acquire acoustic vibration signals at the joint at a high sampling rate, for example, 1000Hz; a distributed fiber optic temperature measurement system is used to acquire real-time temperature sequences at the joint at a lower sampling rate, for example, 0.1Hz; and the real-time current load sequence of the submarine cable is synchronously acquired from the SCADA system of the onshore substation.

[0046] Due to differences in sampling clocks across systems and optical signal transmission delays, the three sets of data were first synchronized based on GPS timestamps. Subsequently, a cubic spline interpolation algorithm was used to synchronize the low-frequency temperature data. and current load Upsampling is performed to adjust its time granularity to match the vibration data. To maintain consistency, the vibration signal was finally subjected to wavelet packet decomposition to filter out known background noise frequencies from ocean waves, retaining the characteristic frequency bands representing the joint's structural response, and thus calculating the real-time vibration energy sequence. .

[0047] By synchronously acquiring, interpolating and aligning, and using wavelet packet denoising, the spatiotemporal differences and environmental noise interference of multi-source data were eliminated, laying a high-quality data foundation for the subsequent construction of accurate physical coupling indicators.

[0048] S2, based on the physical driving law of the real-time current load on the vibration amplitude under normal operating conditions, and the coupling relationship between temperature change and vibration response, uses the monitoring dataset to construct the thermal-vibration coupled breathing entropy, which is used to evaluate the dynamic response state of submarine cable joints to environmental and electrodynamic excitation.

[0049] Specifically, a normal submarine cable joint undergoes regular thermal expansion and contraction deformation when the current load changes. This deformation alters the joint's mechanical impedance, causing its vibration response to environmental or electrodynamic excitations to change. This dynamic correlation of "current-driven -> temperature change -> structural micro-change -> vibration change" is called the joint's breathing effect.

[0050] Therefore, this invention constructs a thermal-vibration coupled respiratory entropy, denoted as... Its calculation formula satisfies:

[0051]

[0052] In the formula, At the current sampling time, The length of the sliding time window, for example, is the number of data points in the most recent minute, used to smooth out short-term fluctuations; For a moment The change in vibration energy is calculated. The first-order difference is obtained, i.e. ; For a moment The amount of temperature change, calculated The first-order difference is obtained, i.e. ; For a moment The current load is directly read by the SCADA system; This is an electrodynamic influence factor, for example, 0.01, used to adjust the sensitivity of the current to vibration suppression; A reference stability constant, such as 1.0, is used to prevent damage when a current load is applied. When the denominator is 0, the formula calculation fails. The coefficient of thermal inertia, for example 10.0, is used to balance the hysteresis effect of temperature changes.

[0053] Scenario 1, Normal High Load: Assume that at a certain moment, the window length... Current A. Electrodynamic forces cause changes in vibration. Due to the normal coupling of thermal expansion and contraction, temperature changes The parameter is set to .

[0054] The denominator term is the load decoupling term: .

[0055] The exponential term is the state constraint: .

[0056] Final result .

[0057] If the formula is not used and the weighting is not adjusted, the vibration value will be 50, which is prone to false alarms. By using the formula, the value is significantly suppressed.

[0058] Scenario 2, Fault: Assuming the connector is loose, the current... A, but the vibration is intense. Furthermore, temperature changes are lagging or mismatched. .

[0059] Denominator term: .

[0060] Index item: .

[0061] Final result .

[0062] As can be seen from the comparison, although the vibration values ​​are the same, the high vibration under low load is a fault characteristic. The calculated thermal-vibration coupling breathing entropy is significantly higher than the normal value under high load, thus amplifying the fault characteristics.

[0063] By constructing a breathing entropy index that includes load decoupling terms and state constraint terms, the vibration data was dynamically weighted using physical laws, which completely eliminated false alarms under high loads and amplified the decoupling signals caused by faults.

[0064] S3. Construct an isolated forest model based on load weights. Input the calculated thermal-vibration coupling respiratory entropy into the isolated forest model. Combine the matching degree between real-time current load and historical average load to calculate the corrected path length of the sample in the isolation tree, and then obtain the anomaly score.

[0065] Specifically, the model is first trained by selecting a dataset from the historical normal operation period of submarine cable joints to construct feature vectors. An isolated forest model is established. During the construction of the isolated trees, for each leaf node, the average current load of all training samples falling into that node is statistically analyzed and stored, denoted as . .

[0066] During online detection, the corrected path length of the sample in the isolation tree is calculated. Its calculation formula satisfies:

[0067]

[0068] In the formula, This sample was calculated using the original Isolation Forest algorithm. The path length in the tree is the number of splits required to isolate a sample. The current sample to be tested Real-time current load; The historical average load recorded for the leaf node where the sample to be tested falls is obtained by querying the trained model; The maximum design load of the submarine cable system is a fixed constant. The current sample calculated in step S2 Thermo-vibrational coupling respiratory entropy; To adjust the weighting coefficients, which are used to control the influence of the correction term; It is a non-linear sensitivity index. The value is usually 2; To prevent extremely small positive numbers with a denominator of 0, for example .

[0069] Scenario A, Normal State: The connector is in healthy operation. It is very small, its value is close to 0, such as 0.1. , .

[0070] At this point, the correction term is positive, making In isolated forests, longer paths mean that samples are more difficult to isolate, thus making them more likely to be judged as normal, which further suppresses false alarms.

[0071] Situation B, Fault Status: The connector is abnormal. Very large, such as 50. , .

[0072] At this point, the correction term is negative, making The path becomes significantly shorter, indicating that the sample is easily isolated, which means it is judged to be more abnormal, thus amplifying the fault signal.

[0073] Load matching correction: This term is used to evaluate the similarity between the current operating condition and the historical operating condition represented by the leaf node. The above physical model correction logic is applied only when the load matching degree is high (i.e., the number in parentheses is positive). This prevents logical reversal caused by the failure of the physical model in the event of extreme overload or data anomalies.

[0074] By introducing load weights and breathing entropy to adaptively correct the path length of isolated forests, the algorithm achieves adaptive learning under different working conditions, significantly improving the ability to detect early and subtle faults in complex environments.

[0075] S4. Based on the calculated anomaly score, the operating status of the submarine cable joint is evaluated at multiple levels. When the anomaly score exceeds the preset threshold, the corresponding early warning or alarm operation is triggered.

[0076] Specifically, based on the final anomaly score calculated in step S3 and in accordance with the operational guidelines, the system automatically outputs a three-level status assessment:

[0077] Healthy state: When the abnormal score is below the first-level threshold, the interface displays green, indicating that the respiratory entropy is stable and the connector is in a healthy thermo-respiratory cycle.

[0078] Attention status, i.e. early hidden danger: the abnormal score is between the first and second level thresholds. At this time, although the temperature and vibration amplitude may not exceed the limits, the internal coupling relationship has been slightly changed, indicating that there may be water tree growth or slight loosening.

[0079] Warning state, i.e. fault outbreak: When the anomaly score exceeds the level 2 threshold, the system automatically triggers an audible and visual alarm and displays the location of the faulty connector through a 3D visualization interface to guide the maintenance vessel to perform targeted maintenance.

[0080] like Figure 2 and Figure 3 As shown, compared with existing technologies ( Figure 2 (green curve in the image) and the present invention ( Figure 3 The monitoring effect of the red curve in the figure shows that during the high-load period (t=20-45), existing technologies cause a significant spike in the index due to electrodynamic vibration, forming a false alarm zone for high load. However, the thermal-vibration coupling breathing entropy curve of this invention is tightly suppressed at the bottom by the algorithm, remaining stable and in the high-load suppression zone, demonstrating its ability to suppress interference under normal operating conditions. At the moment of the actual fault occurrence on the right side (t>75), the thermal-vibration coupling breathing entropy curve of this invention shows a vertical explosive growth, rapidly exceeding the threshold, demonstrating extremely high sensitivity to early faults.

[0081] Through a multi-level condition assessment mechanism, maintenance personnel can intuitively grasp the health status of submarine cable joints and intervene in a timely manner in the early stages of a fault, thus preventing the accident from escalating and ensuring the safe and stable transmission of power across the sea.

[0082] This invention also discloses an online monitoring system for submarine cables, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an online monitoring method for submarine cables according to the present invention.

[0083] This invention also discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements an online monitoring method for submarine cables according to this invention.

[0084] 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.

[0085] 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 online monitoring of submarine cables, characterized in that, Includes the following steps: S1. Collect real-time vibration, temperature and current load sequences at the location of submarine cable joints, and perform spatiotemporal alignment and preprocessing on the multi-source heterogeneous data to obtain a monitoring dataset with consistent time granularity. S2, based on the physical driving law of the real-time current load on the vibration amplitude under normal working conditions, and the coupling relationship between temperature change and vibration response, the thermal-vibration coupling breathing entropy is constructed using the monitoring dataset. The thermal-vibration coupling breathing entropy is used to evaluate the dynamic response state of submarine cable joints to environmental and electrodynamic excitation. Thermal-vibration coupled respiratory entropy The calculation formula satisfies: In the formula, At the current sampling time, The length of the sliding time window. The offset within the window. For a moment The change in vibrational energy, For a moment Temperature change For a moment Current load, For electrodynamic influencing factors, As the reference stability constant, The coefficient of thermal inertia; S3. Construct an isolation forest model based on load weights. Input the calculated thermal-vibration coupling respiration entropy into the isolation forest model, and combine the matching degree between real-time current load and historical average load to calculate the corrected path length of the sample in the isolation tree: , Samples calculated by the original Isolation Forest algorithm Path length in the tree For the sample Real-time current load, For the sample The historical average load recorded by the falling leaf node, This represents the maximum design load of the submarine cable system. To adjust the weighting coefficients, It is a non-linear sensitivity index. If the value is a very small positive number, then an abnormal fraction is obtained; S4. Based on the calculated anomaly score, the operating status of the submarine cable joint is evaluated at multiple levels. When the anomaly score exceeds the preset threshold, the corresponding early warning or alarm operation is triggered.

2. The online monitoring method for submarine cables according to claim 1, characterized in that, Spatiotemporal alignment and preprocessing of multi-source heterogeneous data, including: Time synchronization of real-time vibration sequence, real-time temperature sequence and real-time current load sequence based on GPS timestamp; An interpolation algorithm is used to upsample the real-time temperature sequence and real-time current load sequence with low sampling rates to make them consistent with the time granularity of the real-time vibration sequence. Wavelet packet decomposition was performed on the real-time vibration sequence to filter out low-frequency ocean wave background noise, retaining the characteristic frequency band energy representing the joint body structure response, and the vibration energy sequence was calculated. .

3. The online monitoring method for submarine cables according to claim 1, characterized in that, Change in vibrational energy and temperature change By calculating the vibration energy sequence respectively and temperature sequence First-order difference acquisition; reference stability constant A value of 1.0 is used to prevent current loads from being applied. When the denominator is 0, the fraction is 0.

4. The online monitoring method for submarine cables according to claim 1, characterized in that, Constructing an isolation forest model based on load weights includes: Data sets from the historical normal operation period of submarine cable joints were selected as training samples. During the construction of the isolated tree, for each leaf node in the tree, the average current load of all training samples falling into that leaf node is calculated and stored, denoted as . .

5. The online monitoring method for submarine cables according to claim 1, characterized in that, The calculation logic for correcting the path length includes: pass The assessment evaluates the degree of matching between the current operating conditions and historical operating conditions, and applies physical model correction only when the load matching degree is high; Using logarithmic terms Modulate the corrected path length when When the value approaches 0, increase the correction path length to suppress false alarms; when When the value is large, reduce the length of the correction path to amplify the fault characteristics.

6. The online monitoring method for submarine cables according to claim 1, characterized in that, Based on the calculated anomaly score, the operational status of the submarine cable joint is assessed at multiple levels, including: If the abnormal score is below the first-level threshold, it is considered to be in a healthy state; If the abnormal score is between the first-level threshold and the second-level threshold, it is judged as a state of concern, indicating the existence of early potential risks; If the abnormal score is higher than the secondary threshold, it is determined to be in a warning state, triggering an alarm and displaying the location of the fault.

7. An online monitoring system for submarine cables, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the online monitoring method for submarine cables as described in any one of claims 1-6 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements an online monitoring method for submarine cables according to any one of claims 1-6.

Images