Cable temperature detection method, device and system and storage medium

Abstract

The invention relates to a cable temperature detection method, device and system and a storage medium, and relates to the technical field of cable detection. The method comprises the following steps: respectively injecting a plurality of groups of Gaussian white noise signals into an original temperature signal to obtain a plurality of groups of noise adding signals; according to the intrinsic mode function, performing signal decomposition operation and mean value processing operation on each group of noise adding signals to obtain an initial mode signal and an initial residual signal; performing signal decomposition operation on each group of Gaussian white noise signals by circularly adopting an intrinsic mode function, after multiple groups of noise decomposition signals are obtained each time, injecting each group of noise decomposition signals into a residual signal output last time, and performing signal decomposition operation and mean value processing operation on each group of noise adding residual signals to obtain a residual signal output last time; and obtaining a modal signal output at this time and a residual signal output at this time, and performing signal fusion on the multiple modal signals and the final residual signal to judge whether the target temperature signal is abnormal or not.

Description

Technical Field

[0001] This application relates to the field of cable testing technology, and in particular to cable temperature testing methods, devices, systems and storage media. Background Technology

[0002] Power cables have become the core carrier of urban power supply networks. During operation, internal defects in these cables can easily lead to localized overheating, and because they are often laid in concealed environments, temperature anomalies are difficult to detect in a timely manner, posing significant safety hazards. Currently, external fiber optic temperature measurement is commonly used to sense temperature in real time and identify anomalies based on this real-time temperature reading. However, the collected temperature signals are mixed with a large amount of environmental interference, and these noises are not effectively separated, resulting in data distortion and an inability to accurately reflect the true temperature. Summary of the Invention

[0003] This invention provides a cable temperature detection method, apparatus, system, and storage medium to at least solve the problem of excessive interference noise in the detected temperature signal. The technical solution of this invention is as follows: According to a first aspect of the present invention, a cable temperature detection method is provided. The method includes: injecting multiple sets of Gaussian white noise signals into an original temperature signal to obtain multiple sets of noise-added signals; performing signal decomposition and averaging operations on each set of noise-added signals according to intrinsic mode functions to obtain an initial mode signal and an initial residual signal; using the initial residual signal and the initial mode signal as initial outputs, sequentially applying intrinsic mode functions to perform signal decomposition operations on each set of Gaussian white noise signals; and after obtaining multiple sets of noise-decomposed signals each time, processing each set of noise-decomposed signals separately. The signal is injected into the residual signal of the previous output to obtain multiple sets of noisy residual signals. Based on the intrinsic mode function (IMF), each set of noisy residual signals is subjected to signal decomposition and averaging operations to obtain the modal signal and residual signal of the current output. The call termination condition is determined, and multiple modal signals output according to the IMF and the final residual signal output by the last call to the IMF are obtained. The multiple modal signals and the final residual signal are fused to obtain the target temperature signal. A detection and recognition model is used to determine whether the target temperature signal is abnormal.

[0004] In one implementation, multiple sets of Gaussian white noise signals are injected into the original temperature signal to obtain multiple sets of noise-added signals, including: randomly generating multiple sets of Gaussian white noise signals; adjusting the signal intensity of the multiple sets of Gaussian white noise signals according to the initial noise amplitude to obtain multiple sets of noise-adjusted signals; and fusing each of the multiple sets of noise-adjusted signals with the original temperature signal to obtain multiple sets of noise-added signals.

[0005] In another implementation, based on the intrinsic mode function, signal decomposition and averaging operations are performed on each of the multiple sets of noisy signals to obtain the initial mode signal and the initial residual signal. This includes: using the intrinsic mode function to perform signal decomposition on each set of noisy signals to obtain multiple sets of initial candidate signals; determining the average value of the multiple sets of initial candidate signals as the initial mode signal; and determining the signal difference between the original temperature signal and the initial mode signal as the initial residual signal.

[0006] In another implementation, each set of noise decomposition signals is injected into the residual signal of the previous output to obtain multiple sets of noise-added residual signals. This includes: adjusting the signal strength of multiple sets of noise decomposition signals according to the corresponding noise amplitude to obtain multiple sets of adjusted noise decomposition signals; and fusing each set of adjusted noise decomposition signals with the residual signal of the previous output to obtain multiple sets of noise-added signals.

[0007] In another implementation, based on the intrinsic mode function, signal decomposition and averaging operations are performed on each of the multiple sets of noisy residual signals to obtain the current output modal signal and the current output residual signal. This includes: using the intrinsic mode function to perform signal decomposition on each set of noisy residual signals to obtain multiple sets of current candidate signals; determining the average value of the multiple sets of current candidate signals as the current output modal signal; and determining the signal difference between the previous output residual signal and the current output modal signal as the current output residual signal.

[0008] In another implementation, the detection and recognition model is a convolutional neural network model. The detection and recognition model is trained based on abnormal sample data with temperature anomaly labels and normal sample data with temperature normal labels. The method also includes: training the model parameters of the initial recognition model based on the abnormal sample data and normal sample data until the difference loss is less than or equal to the preset loss, and then determining the initial recognition model under the corresponding model parameters as the detection and recognition model. The difference loss includes the difference loss between the abnormal features output by the model and the preset abnormal features of the labels, as well as the difference loss between the normal features output by the model and the preset normal features of the corresponding labels.

[0009] In another implementation, a detection and identification model is used to determine whether the target temperature signal is abnormal. This includes: inputting the target temperature signal into the detection and identification model to obtain the identification result; if the identification result indicates that the target temperature signal is abnormal, the cable temperature is determined to be abnormal; if the identification result indicates that the target temperature signal is normal, the cable temperature is determined to be normal.

[0010] According to a second aspect of the present invention, a cable temperature detection device is provided, comprising: an injection unit for injecting multiple sets of Gaussian white noise signals into an original temperature signal to obtain multiple sets of noise-added signals; a first processing unit for performing signal decomposition and mean processing operations on each set of noise-added signals according to the intrinsic mode function to obtain an initial mode signal and an initial residual signal; and a second processing unit for performing signal decomposition operations on each set of Gaussian white noise signals sequentially using the initial residual signal and the initial mode signal as initial outputs; and after obtaining multiple sets of noise-decomposed signals each time, processing each set... The noise decomposition signals are injected into the residual signals of the previous output, resulting in multiple sets of noisy residual signals. Based on the intrinsic mode function (IMF), each set of noisy residual signals undergoes signal decomposition and averaging operations to obtain the current output modal signal and the current output residual signal. The fusion unit determines the call termination condition, acquires multiple modal signals output according to the IMF and the final residual signal output from the last IMF call, and fuses the multiple modal signals with the final residual signal to obtain the target temperature signal. The detection unit uses a detection and identification model to determine whether the target temperature signal is abnormal.

[0011] According to a third aspect of the present invention, a cable temperature detection system is provided. The system includes: a sensing layer deployed on the cable side for acquiring temperature optical signals from the cable surface; the sensing layer includes a temperature-measuring optical fiber, a linear laser, an optical fiber coupler, a photoelectric converter, and an analog-to-digital converter; a transmission layer for transmitting the digital signals acquired by the sensing layer to a processing layer; the transmission layer includes a cable-side data switch and a control-side data switch; a processing layer including a data processing module and a status monitoring module; the data processing module includes intrinsic mode functions for performing signal decomposition and averaging operations; the status monitoring module includes a detection and identification model; and an application layer configured to perform fault warnings based on the judgment results of the processing layer and to integrate digital twin technology to construct a cable operating status visualization platform; the system is configured to execute a cable temperature detection method as described in the first aspect and any possible implementation thereof.

[0012] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which instructions are stored, such that when the instructions in the computer-readable storage medium are executed by a processor of a cable temperature detection device, the cable temperature detection device is able to perform a cable temperature detection method as described in the first aspect and any possible implementation thereof.

[0013] According to a fifth aspect of the present disclosure, a computer program product is provided, the computer program product including computer instructions, which, when executed on a cable temperature detection device, cause the cable temperature detection device to perform the cable temperature detection method of the first aspect and any possible implementation thereof.

[0014] The technical solution provided by the embodiments of the present invention brings at least the following beneficial effects: Through adaptive noise injection and multiple decomposition averaging, the original temperature signal can be accurately decomposed into a series of modal signals composed of an Intrinsic Mode Function (IMF) component at different time scales and a residual signal composed of a residual trend term. This allows for signal reconstruction by removing the high-frequency IMF component representing random noise and high-frequency interference, while retaining the mid-to-low-frequency IMF component representing the true temperature change, thereby obtaining highly pure target temperature data. Furthermore, after the initial signal decomposition operation, subsequent signal decomposition processes do not involve re-noising the original signal and performing a complete decomposition. Instead, the residual signal from the previous order is also noise-added and processed. The subsequent noise-added signal is based on the signal after IMF decomposition processing, thus adding adaptive noise of a specific scale, filtered by EMD, to the current residual. This results in higher noise utilization, targeted extraction of components at different scales, and higher decomposition accuracy. Moreover, the sum of the aforementioned modal signals and the final residual is strictly equal to the original temperature signal, thereby achieving complete reconstruction.

[0015] The temperature detection method described above, through its unique "sequence decomposition" and "adaptive noise injection" mechanisms, not only fundamentally solves the modal aliasing problem, but also overcomes the shortcomings of large computational load and incomplete reconstruction in related technologies, thus laying a solid technical foundation for extracting real and pure temperature change characteristics from complex cable temperature signals.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0018] Figure 1 This is a schematic diagram of a cable temperature detection system according to an exemplary embodiment; Figure 2 This is a flowchart illustrating a cable temperature detection method according to an exemplary embodiment; Figure 3This is a cable temperature detection device illustrated according to an exemplary embodiment; Figure 4 This is a schematic diagram of a cable temperature detection device according to an exemplary embodiment. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0020] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0021] Before providing a detailed description of the cable temperature detection method provided in this application embodiment, let's briefly introduce the application scenarios and implementation environment involved in this application embodiment.

[0022] With the acceleration of urbanization, the conflict between urban space and power transmission channels is becoming increasingly prominent. Power cables, due to their advantages such as strong spatial concealment, good aging resistance, and stable power supply, are gradually replacing overhead lines and becoming the core carrier of urban power transmission. However, affected by factors such as manufacturing processes, construction quality, and operating environment, the cable body is prone to typical defects such as metal sharp points, air gap residue, and damage to the aluminum sheath. During long-term operation, these defects can cause temperature field distortion, induce partial discharge, and form a vicious cycle between partial discharge and insulation deterioration. Furthermore, since cables are often laid in concealed environments, temperature rises are difficult to detect in a timely manner, and there is a lack of effective early warning methods. Therefore, real-time online temperature monitoring has become a key technology to ensure the safe operation of cables.

[0023] Related detection methods, such as inspection robots and infrared temperature imaging, still suffer from temporal and spatial lag, making it difficult to capture temperature anomalies in a timely and comprehensive manner. However, external fiber optic temperature sensors offer a feasible solution for monitoring cable temperature anomalies. Mainstream methods for power cable temperature detection can be divided into two categories: direct measurement and indirect measurement. Direct measurement includes infrared thermometry, wireless thermometry, and external temperature-measuring fiber optics, while indirect measurement derives cable temperature by detecting current carrying capacity.

[0024] While these methods can identify abnormal cable temperature operation to some extent, they generally have significant limitations. Meanwhile, the external fiber optic temperature sensor detection method proposed by researchers provides a feasible path for monitoring abnormal cable temperature.

[0025] The principle is as follows: When the temperature of the copper core inside the cable changes, the heat is conducted and diffused, causing a corresponding change in the surface temperature of the cable after a period of time. At this time, photons incident on the optical fiber undergo inelastic collisions with the molecules of the medium, resulting in energy transfer between the molecules and the photons. This process follows the Boltzmann distribution theorem. This leads to the derivation of the Raman scattering intensity ratio formula.

[0026]

[0027] In the formula: E I Let g be the energy of the i-th state. i Energy E I The degeneracy of the energy levels, k is the Boltzmann constant, and T is the thermodynamic temperature. For Boltzmann factor, I AS (T) represents the intensity of the anti-Stokes light at temperature T, I S Let C be the Stokes light intensity at temperature T, C be the correlation coefficient between the optical fiber and the measuring cable, v0 be the frequency of the incident light, and v AS v is the frequency of anti-Stokes light. S For the frequency of Stokes light, This is a Raman frequency shift.

[0028] Among them, the frequency of anti-Stokes light is higher than that of incident light, and its intensity is extremely sensitive to temperature. This formula can be used to accurately infer the temperature of the conductor inside the cable.

[0029] Cable temperature signals are complex, non-stationary signals composed of a superposition of various physical sources, including the actual heat generated by the cable conductor (low frequency, slowly varying), ambient temperature fluctuations (medium to low frequency), and electromagnetic interference, mechanical vibration, etc. (high frequency). Related technologies cannot effectively separate these components across different time scales, leading to their entanglement and interference, a phenomenon known as "modal aliasing." This makes extracting a pure, true temperature trend extremely difficult.

[0030] The lack of a supporting high-precision demodulation system and intelligent data processing platform for external fiber optic temperature sensor detection methods hinders the construction of an unmanned, real-time digital monitoring system, making it difficult to achieve rapid response and accurate early warning for cable temperature anomalies. The following technical issues need to be addressed: First, eliminate fluctuations in cable temperature data caused by external interference to improve data accuracy; second, construct an efficient real-time monitoring and data processing system to achieve real-time capture of cable temperature changes and accurate identification of abnormal states; third, build a complete hardware and software collaborative system to form a full-process digital solution from temperature signal acquisition, transmission, processing to early warning, meeting the actual needs of power systems for monitoring cable temperature anomalies.

[0031] To address the aforementioned issues, this application proposes a cable temperature detection method. Through adaptive noise injection and multiple decomposition averaging, the original temperature signal can be accurately decomposed into a series of modal signals composed of an intrinsic mode function component at different time scales and a residual signal composed of a residual trend term. By removing the high-frequency IMF component representing random noise and high-frequency interference, and retaining the mid-to-low-frequency IMF component representing the actual temperature change, the signal can be reconstructed to obtain highly pure target temperature data.

[0032] Secondly, the implementation architecture involved in this application will be briefly introduced below.

[0033] Figure 1 This is a schematic diagram of a cable temperature detection system provided in this application. Figure 1 As shown, the cable temperature detection system includes a sensing layer 11, a transmission layer 12, a processing layer 13, and an application layer 14.

[0034] Sensing layer 11, deployed on the cable side, is used to collect temperature optical signals from the cable surface.

[0035] The sensing layer 11 includes a temperature-sensing optical fiber, a linear laser, an optical fiber coupler, a photoelectric converter, and an analog-to-digital converter.

[0036] The transmission layer 12 is used to transmit the digital signals collected by the perception layer to the processing layer.

[0037] The transport layer 12 includes a cable-side data switch and a control-side data switch.

[0038] Processing layer 13 includes a data processing module and a status monitoring module.

[0039] The data processing module includes intrinsic mode functions, which are used to perform signal decomposition and mean processing operations.

[0040] The status monitoring module includes a detection and recognition model.

[0041] Application layer 14 is configured to provide fault warnings based on the judgment results of the processing layer, and integrates digital twin technology to build a cable operation status visualization platform.

[0042] In some implementations, the interaction details of the above devices are as follows: the temperature-measuring optical fiber collects optical signals related to the surface temperature of the cable, which are then modulated by a fiber optic coupler and a linear laser. The optical signal is converted into an electrical signal by a photoelectric converter, and then into a digital signal by an analog-to-digital converter. The digital signal is transmitted to the PC via a data switch on the cable side and a data switch on the control side. The PC performs data processing and status judgment through a built-in algorithm module.

[0043] The specific steps are as follows.

[0044] First, on-site installation and deployment: After the cable is installed, install the temperature measuring fiber on the cable surface and connect the temperature measuring fiber to a 2×2 coupler, and at the same time connect it to a linear light source.

[0045] Secondly, on-site signal acquisition and conversion: The linear laser operates according to the set parameters to monitor the cable operating temperature in real time. The photoelectric converter converts the optical signal into an electrical signal, and then the analog-to-digital converter converts it into a digital signal, which is then transmitted to the data acquisition card.

[0046] Third, on-site and remote collaborative data transmission: After the data acquisition card acquires digital signals, it transmits them to the PC through the cable-side industrial data switch and the control-side data switch.

[0047] Fourth, the remote PC performs data processing: After receiving the temperature signal, the PC decomposes the received temperature signal through CEEMDAN to eliminate fluctuations in cable temperature caused by other factors.

[0048] Fifth, remote PC execution status monitoring and judgment: CNN uses its training experience to monitor the signals processed by CEEMDAN in real time, and judges whether there is an abnormal operating state by comparing them with the characteristics of the normal operating state.

[0049] Sixth, remote PC terminal performs early warning and feedback: if an abnormal state is detected, the system triggers an early warning and simultaneously feeds back the abnormal information to the operation and maintenance management platform to provide a basis for fault handling.

[0050] In some embodiments, firstly, after the 66kV cable is installed, temperature-sensing optical fibers are evenly laid on the cable surface to ensure that the optical fibers are in close contact with the cable surface. The temperature-sensing optical fibers are then connected to the input end of a 2×2 coupler, and the output end of the coupler is connected to a linear light source.

[0051] Furthermore, the linear laser was set to a 50MW operating mode with a wavelength of 1550nm and directly powered by a 220V AC power supply module to ensure stable light source output.

[0052] Furthermore, the optoelectronic converter and coupler are connected via SC / SC connectors, the analog-to-digital converter communicates with the optoelectronic converter via a serial SPI interface, and the data acquisition card and analog-to-digital converter are connected in eight-channel configuration to ensure that no signal is missed during transmission.

[0053] Furthermore, the data acquisition card is connected to the cable-side data switch via a network cable. The cable-side data switch is connected to the control-side data switch via optical fiber. The control-side data switch is connected to the PC's network card via a network cable, thus completing the transmission link setup. The switch input voltage is set to 220V and 50Hz to ensure that the switching capacity and packet forwarding rate meet the data transmission requirements.

[0054] Furthermore, data processing and monitoring software was installed on the PC, with built-in CEEMDAN and CNN algorithm modules. The CNN algorithm has been trained using a large amount of temperature data from normal and abnormal states of 66kV cables, establishing a comprehensive feature database.

[0055] Furthermore, after the system is started, the linear laser continuously emits 1550nm wavelength light. The temperature-sensing fiber senses the temperature change on the cable surface and converts it into an optical signal, which is transmitted to the photoelectric converter via a coupler. The photoelectric converter converts the optical signal into an electrical signal (voltage range ±10V). The analog-to-digital converter converts the electrical signal into a 16-bit digital signal. The data acquisition card acquires the digital signal in real time at a 200K acquisition rate and transmits it to the PC via a switch.

[0056] Furthermore, after receiving the data, the PC software decomposes the digital signal using CEEMDAN to separate the intrinsic mode functions and residual components, removing interference fluctuations. The processed signal is then input into the CNN module, which determines the current cable operating status by comparing it with normal state features in the feature database.

[0057] Furthermore, experimental tests showed that when the temperature of the 66kV cable is within the normal operating range (0-90℃) and the abnormal critical temperature range (90-110℃), the correlation coefficient between the temperature detection signal output by the system and the actual operating temperature of the cable is ≥0.95, which can accurately infer the internal conductor temperature of the cable. When an abnormal temperature is detected (≥90℃), the system immediately triggers a "cable temperature abnormality" warning and displays the abnormal location, temperature value and trend on the software interface, while storing the abnormal information in the database for subsequent analysis.

[0058] For ease of understanding, the cable temperature detection method provided in this application will be described in detail below with reference to the accompanying drawings.

[0059] Figure 2 This is a flowchart illustrating a cable temperature detection method according to an exemplary embodiment, such as... Figure 2 As shown, the cable temperature detection method includes the following steps.

[0060] S21, inject multiple sets of Gaussian white noise signals into the original temperature signal to obtain multiple sets of noise-added signals.

[0061] S22, based on the intrinsic mode function, perform signal decomposition and mean processing operations on each of the multiple sets of noise signals to obtain the initial mode signal and the initial residual signal.

[0062] The initial mode signal is also called the first IMF component.

[0063] The initial residual signal is also called the first residual sequence.

[0064] S23 uses the initial residual signal and the initial mode signal as the initial output, and sequentially applies the intrinsic mode function to perform signal decomposition operations on each group of Gaussian white noise signals.

[0065] S24. After obtaining multiple sets of noise decomposition signals each time, each set of noise decomposition signals is injected into the residual signal output in the previous time to obtain multiple sets of noisy residual signals. Based on the intrinsic mode function, signal decomposition operation and mean processing operation are performed on each set of noisy residual signals to obtain the modal signal and residual signal output in this time.

[0066] S25, determine the call termination condition, and obtain multiple modal signals output by the intrinsic modal function and the final residual signal output by the last call to the intrinsic modal function.

[0067] The stopping condition can be either a monotonic function or the residual signal of the current output.

[0068] S26, multiple modal signals are fused with the final residual signal to obtain the target temperature signal.

[0069] S27 uses a detection and identification model to determine whether the target temperature signal is abnormal.

[0070] The aforementioned detection and recognition models include LSTM (i.e., Long Short-Term Memory Network) models or CNN (Convolutional Neural Network) models.

[0071] The cable temperature detection method described above, through adaptive noise injection and multiple decomposition averaging, can accurately decompose the original temperature signal into a series of modal signals composed of an Intrinsic Mode Function (IMF) component at different time scales and a residual signal composed of a residual trend term. This allows for signal reconstruction by removing the high-frequency IMF component representing random noise and high-frequency interference, while retaining the mid-to-low-frequency IMF components representing the true temperature change, thus obtaining highly pure target temperature data. Furthermore, after the initial signal decomposition, subsequent decompositions do not involve re-noising the original signal and performing a complete decomposition. Instead, the residual signal from the previous stage is also noise-added and processed. The subsequent noise-added signal is based on the signal after IMF decomposition processing, adding adaptive noise of a specific scale, filtered by EMD, to the current residual. This results in higher noise utilization, targeted extraction of components at different scales, and higher decomposition accuracy. Moreover, the sum of all the modal signals and the final residual is strictly equal to the original temperature signal, achieving complete reconstruction.

[0072] In one noise-adding method, the noise-adding process in step S21 above is as follows.

[0073] First, multiple sets of Gaussian white noise signals are randomly generated to ensure adaptability.

[0074] Secondly, based on the initial noise amplitude, the signal strength of multiple sets of Gaussian white noise signals is adjusted to obtain multiple sets of noise adjustment signals.

[0075] Fourth, each of the multiple noise adjustment signals is fused with the original temperature signal to obtain multiple noise-added signals.

[0076] In one initial signal decomposition method, the signal decomposition operation and mean processing operation in step S22 above are as follows.

[0077] Firstly, the intrinsic mode function is used to decompose each group of noisy signals to obtain multiple initial candidate signals.

[0078] Secondly, the average value of multiple initial candidate signals is determined as the initial modal signal.

[0079] Third, the signal difference between the original temperature signal and the initial modal signal is determined as the initial residual signal.

[0080] In one residual noise-adding method, the specific process of generating multiple sets of noisy residual signals is as follows.

[0081] First, based on the corresponding noise amplitude, the signal strength of multiple sets of noise decomposition signals is adjusted to obtain multiple sets of adjusted noise decomposition signals.

[0082] Secondly, each of the adjusted noise decomposition signals in the multiple sets of adjusted noise decomposition signals is fused with the residual signal of the previous output to obtain multiple sets of noise-added signals.

[0083] In another subsequent signal decomposition method, after the initial signal decomposition operation and mean processing operation, and before calling the stop condition, the specific process of other steps is as follows.

[0084] First, the intrinsic mode function is used to decompose each group of noisy residual signals to obtain multiple groups of candidate signals for this test.

[0085] Secondly, the average value of multiple candidate signals is determined as the modal signal for this output.

[0086] Third, the signal difference between the previous residual signal and the current modal signal is determined as the current residual signal.

[0087] In one implementation, the detection and recognition model is a convolutional neural network model; the detection and recognition model is trained based on abnormal sample data with temperature anomaly labels and normal sample data with temperature normal labels.

[0088] Based on this implementation method, the training process for the detection and recognition model is as follows.

[0089] Based on abnormal and normal sample data, the model parameters of the initial recognition model are trained until the difference loss is less than or equal to the preset loss. The initial recognition model under the corresponding model parameters is then determined as the detection and recognition model. The difference loss includes the difference loss between the abnormal features output by the model and the preset abnormal features of the label, as well as the difference loss between the normal features output by the model and the preset normal features of the corresponding label.

[0090] In some implementations, a detection and identification model is used to determine whether the target temperature signal is abnormal, including: inputting the target temperature signal into the detection and identification model to obtain the identification result; if the identification result indicates that the target temperature signal is abnormal, determining that the cable temperature is abnormal; if the identification result indicates that the target temperature signal is normal, determining that the cable temperature is normal.

[0091] In some specific embodiments, the specific steps for processing the cable temperature signal are as follows.

[0092] First, N Gaussian white noise is injected into the input raw temperature signal y(t) to construct N preprocessed sequences y. n (t), where n=1, 2, ..., N.

[0093]

[0094] in, Constants are typically used to control the intensity of white noise, and their commonly used range is generally 0.01 to 0.2. The specific value depends on the characteristics of the original signal. The first round of nth injection is Gaussian white noise.

[0095] Furthermore, for all preprocessed sequences Empirical Mode Decomposition (EMD, a type of signal decomposition) is performed to obtain the first IMF component. The average of these N IMF components is then calculated to obtain the first IMF component of CEEMDAN. At the same time, the first residual sequence is obtained as shown in the following formula. .

[0096]

[0097]

[0098] Furthermore, the residual sequence is also... Adding Gaussian white noise constructs N new sequences For each sequence, perform EMD decomposition and extract the first IMF component. Calculate the mean of the first IMF component to obtain the second IMF component. .

[0099]

[0100] in, This is an E1-based noise scaling process. Specifically, the amplitude is adjusted using ε1 to obtain a second round of noise signal that matches the current residual sequence scale.

[0101] Furthermore, the subsequent residual sequence satisfies the following formula.

[0102] (m≥2) Furthermore, regarding Perform N EMD decompositions to obtain the (m+1)th IMF sequence after CEEMDAN decomposition.

[0103]

[0104] Furthermore, repeat the above noise scaling and EMD decomposition steps until the decomposition stops, leaving only the residual sequence.

[0105] From the perspective of reconstructing the original signal, the formula is as follows.

[0106]

[0107] Based on this formula, the recursive relationship is as follows.

[0108]

[0109] Furthermore, the termination condition (i.e., the call stop condition) is when the remaining amount... When there are no longer enough extreme points (extreme points ≤ 2), CEEMDAN considers the signal to be beyond further decomposition, and at this point, the final trend term is obtained: That is, the expression after the CEEMDAN decomposition of the signal sequence.

[0110]

[0111] in, The noise amplitude, The white noise sequence is injected for the nth time.

[0112] Finally, the sequences decomposed from CEEMDAN are input into the CNN model.

[0113] The LSTM (Long Short-Term Memory) model can be used instead of the CNN model. LSTM has advantages in time series data processing and can achieve anomaly identification by learning the time series features of temperature changes, which can also achieve the purpose of fault early warning.

[0114] To achieve the above functions, the cable temperature detection device includes corresponding hardware structures and / or software modules for performing each function. Those skilled in the art will readily recognize that, based on the algorithmic steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is implemented in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0115] This application embodiment also provides a method such as Figure 3 The cable temperature detection device shown includes: an injection unit 31, a first processing unit 32, a second processing unit 33, a fusion unit 34, and a detection unit 35.

[0116] The injection unit 31 is used to inject multiple sets of Gaussian white noise signals into the original temperature signal to obtain multiple sets of noise-added signals.

[0117] The first processing unit 32 is used to perform signal decomposition and mean processing operations on each of the multiple sets of noisy signals according to the intrinsic mode function, so as to obtain the initial mode signal and the initial residual signal.

[0118] The second processing unit 33 is used to take the initial residual signal and the initial modal signal as the initial output, and sequentially use the intrinsic mode function to perform signal decomposition operations on each group of Gaussian white noise signals. After obtaining multiple groups of noise decomposition signals each time, the noise decomposition signals are injected into the residual signal of the previous output to obtain multiple groups of noisy residual signals. According to the intrinsic mode function, the signal decomposition operation and mean processing operation are performed on each group of noisy residual signals to obtain the modal signal and the residual signal of the current output.

[0119] The fusion unit 34 is used to determine the call stop condition, acquire multiple modal signals output by the intrinsic mode function and the final residual signal output by the last call to the intrinsic mode function; and fuse the multiple modal signals with the final residual signal to obtain the target temperature signal.

[0120] The detection unit 35 is used to determine whether the target temperature signal is abnormal by employing a detection and identification model.

[0121] In one embodiment, the injection unit 31 is specifically used to: randomly generate multiple sets of Gaussian white noise signals; adjust the signal intensity of the multiple sets of Gaussian white noise signals according to the initial noise amplitude to obtain multiple sets of noise adjustment signals; and fuse each set of noise adjustment signals with the original temperature signal to obtain multiple sets of noise-added signals.

[0122] In another embodiment, the first processing unit 32 is specifically used to: use intrinsic mode functions to decompose each group of noisy signals to obtain multiple initial candidate signals; determine the average value of the multiple initial candidate signals as the initial mode signal; and determine the signal difference between the original temperature signal and the initial mode signal as the initial residual signal.

[0123] In another embodiment, the first processing unit 32 is specifically used to: adjust the signal strength of multiple noise decomposition signals according to the corresponding noise amplitude to obtain multiple adjusted noise decomposition signals; and fuse each of the adjusted noise decomposition signals with the residual signal output from the previous time to obtain multiple noise-added signals.

[0124] In another embodiment, the second processing unit 33 is specifically used to: use the intrinsic mode function to decompose each group of noisy residual signals to obtain multiple groups of current candidate signals; determine the average value of the multiple groups of current candidate signals as the current output mode signal; and determine the signal difference between the previous output residual signal and the current output mode signal as the current output residual signal.

[0125] In another implementation, the detection and recognition model is a convolutional neural network model; the detection and recognition model is trained based on abnormal sample data with temperature anomaly labels and normal sample data with temperature normal labels; the detection unit 35 is specifically used to: train the model parameters of the initial recognition model based on the abnormal sample data and normal sample data until the difference loss is less than or equal to the preset loss, and then determine the initial recognition model under the corresponding model parameters as the detection and recognition model; the difference loss includes the difference loss between the abnormal features output by the model and the preset abnormal features of the label, and the difference loss between the normal features output by the model and the preset normal features of the corresponding label.

[0126] In another embodiment, the detection unit 35 is specifically used to: input the target temperature signal into the detection and recognition model to obtain the recognition result; if the recognition result indicates that the target temperature signal is abnormal, determine that the cable temperature is abnormal; if the recognition result indicates that the target temperature signal is normal, determine that the cable temperature is normal.

[0127] Regarding the apparatus in the above embodiments, the specific manner in which each unit module performs its operations has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0128] Figure 4 This is a schematic diagram of a cable temperature detection device provided in this application. Figure 4 The cable temperature detection device 50 includes: a first processor 501, a communication bus 502, a memory 503, a communication interface 504, an output device 505, an input device 506, and a second processor 507.

[0129] The cable temperature detection device 50 may include at least one first processor 501 and a memory 503 for storing processor-executable instructions. The first processor 501 is configured to execute the instructions in the memory 503 to implement the cable temperature detection method in the following embodiments.

[0130] In addition, the cable temperature detection device 50 may also include a communication bus 502, at least one communication interface 504, an input device 506, and an output device 505.

[0131] The first processor 501 may be a processor (central processing unit, CPU), a microprocessor unit, an ASIC, or one or more integrated circuits for controlling the execution of programs according to the present application.

[0132] The communication bus 502 may include a path for transmitting information between the aforementioned components.

[0133] Communication interface 504 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0134] Input device 506 is used to receive input signals and output device 505 is used to output signals.

[0135] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed discs, laser discs, optical discs, digital universal discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory may exist independently and be connected to the processing unit via a bus. Memory may also be integrated with the processing unit.

[0136] The memory 503 stores instructions for executing the scheme of this application, and the execution is controlled by the first processor 501. The first processor 501 executes the instructions stored in the memory 503 to realize the functions of the method of this application.

[0137] In a specific implementation, as one example, the first processor 501 may include one or more CPUs, for example... Figure 4 CPU0 and CPU1 in the CPU.

[0138] In a specific implementation, as one example, the cable temperature detection device 50 may include multiple processors, such as... Figure 4 The first processor 501 and the second processor 507 are described. Each of these processors can be a single-core processor or a multi-core processor. A processor here can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0139] The cable temperature detection device, such as Figure 4 The diagram includes a first processor 501 and a memory 503 for storing executable instructions of the first processor 501; wherein the first processor 501 is configured to execute executable instructions to implement the cable temperature detection method as described in any of the possible embodiments above. And it can achieve the same technical effect, so to avoid repetition, it will not be described again here.

[0140] This application also provides a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by the processor of a cable temperature detection device, the cable temperature detection device is able to perform the cable temperature detection method as described in any of the possible embodiments above. And it can achieve the same technical effect; to avoid repetition, it will not be described again here.

[0141] This application also provides a computer program product, including a computer program or instructions, which are executed by a processor as a cable temperature detection method according to any of the possible embodiments described above. It achieves the same technical effects, and to avoid repetition, will not be described again here.

[0142] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0143] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for detecting cable temperature, characterized in that, The method includes: Multiple sets of Gaussian white noise signals are injected into the original temperature signal to obtain multiple sets of noise-added signals; Based on the intrinsic mode function, each of the multiple groups of noise-added signals is subjected to signal decomposition and mean processing operations to obtain the initial mode signal and the initial residual signal. Using the initial residual signal and the initial modal signal as the initial output, the intrinsic mode function is sequentially applied to perform signal decomposition operations on each group of Gaussian white noise signals. After obtaining multiple groups of noise decomposition signals each time, each group of noise decomposition signals is injected into the residual signal of the previous output to obtain multiple groups of noisy residual signals. According to the intrinsic mode function, signal decomposition and mean processing operations are performed on each group of noisy residual signals to obtain the modal signal and the residual signal of the current output. Determine the call termination condition, and obtain multiple modal signals output by the intrinsic mode function and the final residual signal output by the last call to the intrinsic mode function; The multiple modal signals are fused with the final residual signal to obtain the target temperature signal; A detection and identification model is used to determine whether the target temperature signal is abnormal.

2. The method according to claim 1, characterized in that, The process involves injecting multiple sets of Gaussian white noise signals into the original temperature signal to obtain multiple sets of noisy signals, including: The multiple sets of Gaussian white noise signals are generated randomly; Based on the initial noise amplitude, the signal strength of the multiple sets of Gaussian white noise signals is adjusted to obtain multiple sets of noise adjustment signals; Each of the multiple noise adjustment signals is fused with the original temperature signal to obtain the multiple noise-added signals.

3. The method according to claim 1, characterized in that, The step of performing signal decomposition and mean processing operations on each of the multiple sets of noisy signals according to the intrinsic mode function to obtain the initial mode signal and the initial residual signal includes: Using the intrinsic mode function, each group of noisy signals is decomposed to obtain multiple initial candidate signals; The average value of the multiple initial candidate signals is determined as the initial modal signal; The signal difference between the original temperature signal and the initial mode signal is determined as the initial residual signal.

4. The method according to claim 1, characterized in that, The step of injecting each set of noise decomposition signals into the residual signal of the previous output to obtain multiple sets of noisy residual signals includes: Based on the corresponding noise amplitude, the signal strength of the multiple sets of noise decomposition signals is adjusted to obtain multiple sets of adjusted noise decomposition signals. Each of the adjusted noise decomposition signals in the plurality of adjusted noise decomposition signals is fused with the residual signal output from the previous time to obtain the plurality of noise-added signals.

5. The method according to claim 1, characterized in that, The step of performing signal decomposition and averaging operations on each of the multiple sets of noisy residual signals based on the intrinsic mode function to obtain the current output modal signal and the current output residual signal includes: Using the intrinsic mode function, each group of noisy residual signals is decomposed to obtain multiple groups of candidate signals for this test. The average value of the multiple sets of current candidate signals is determined as the modal signal output this time; The signal difference between the previous residual signal and the current modal signal is determined as the residual signal of the current output.

6. The method according to claim 2, characterized in that, The detection and recognition model is a convolutional neural network model; the detection and recognition model is trained based on abnormal sample data with temperature anomaly labels and normal sample data with temperature normal labels; the method further includes: Based on the abnormal sample data and the normal sample data, the model parameters of the initial identification model are trained until the difference loss is less than or equal to the preset loss. Then, the initial identification model under the corresponding model parameters is determined as the detection identification model. The difference loss includes the difference loss between the abnormal features output by the model and the preset abnormal features of the identifier, and the difference loss between the normal features output by the model and the preset normal features of the corresponding identifier.

7. The method according to any one of claims 1 to 6, characterized in that, The call stopping conditions include: the residual signal output in this call is a monotonic function; the step of using a detection and identification model to determine whether the target temperature signal is abnormal includes: The target temperature signal is input into the detection and recognition model to obtain the recognition result; If the identification result indicates that the target temperature signal is abnormal, the cable temperature is determined to be abnormal. If the identification result indicates that the target temperature signal is normal, the cable temperature is determined to be normal.

8. A cable temperature detection device, characterized in that, The device includes: The injection unit is used to inject multiple sets of Gaussian white noise signals into the original temperature signal to obtain multiple sets of noise-added signals. The first processing unit is used to perform signal decomposition and mean processing operations on each of the multiple groups of noisy signals according to the intrinsic mode function, so as to obtain the initial mode signal and the initial residual signal. The second processing unit is used to take the initial residual signal and the initial modal signal as initial outputs, and sequentially apply the intrinsic mode function to perform signal decomposition operations on each group of Gaussian white noise signals; after obtaining multiple groups of noise decomposition signals each time, each group of noise decomposition signals is injected into the residual signal output in the previous time to obtain multiple groups of noisy residual signals, and according to the intrinsic mode function, each group of noisy residual signals is subjected to signal decomposition operation and mean processing operation to obtain the modal signal output in this time and the residual signal output in this time. The fusion unit is used to determine the call termination condition, acquire multiple modal signals output according to the intrinsic mode function and the final residual signal output by the last call to the intrinsic mode function; and fuse the multiple modal signals with the final residual signal to obtain the target temperature signal. The detection unit is used to determine whether the target temperature signal is abnormal by employing a detection and identification model.

9. A cable temperature detection system, characterized in that, The system includes: A sensing layer, deployed on the cable side, is used to collect temperature optical signals from the cable surface; the sensing layer includes a temperature-sensing optical fiber, a linear laser, an optical fiber coupler, a photoelectric converter, and an analog-to-digital converter; A transmission layer is used to transmit the digital signals collected by the sensing layer to the processing layer; the transmission layer includes a cable-side data switch and a control-side data switch. The processing layer includes a data processing module and a status monitoring module; The data processing module includes intrinsic mode functions, used to perform signal decomposition and mean processing operations; The status monitoring module includes a detection and recognition model; The application layer is configured to provide fault warnings based on the judgment results of the processing layer, and integrates digital twin technology to build a cable operation status visualization platform; The system is configured to perform the cable temperature detection method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the cable temperature detection method as described in any one of claims 1-7.

Images