An electrical spark non-destructive testing evaluation method for cable trench waterproof layer integrity

Abstract

The application discloses a kind of for cable trench waterproof layer integrity electric spark nondestructive testing evaluation method, belong to nondestructive testing technical field, it includes by preset mobile electrode applies multi-frequency electric signal;Real-time acquisition mobile electrode and the initial contact pressure of waterproof layer surface, adjust the pressure of mobile electrode contact position, capture the electric spark signal generated by the multi-frequency electric signal applied by mobile electrode at the defect of cable trench waterproof layer, extract the characteristic parameter of electric spark signal to generate current electric spark characteristic data;Current electric spark characteristic data is used to distinguish defect type, to generate defect evaluation result;According to defect evaluation result, output defect distribution information and trigger sound light alarm.The application adopts multi-frequency electric signal excitation, electric spark signal feature extraction and analysis combined detection mode, can distinguish waterproof layer defect type, evaluate its severity, improve the accuracy of detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of non-destructive testing, in particular to a spark non-destructive testing evaluation method for the integrity of a waterproof layer of a cable trench. BACKGROUND

[0002] A cable trench is an important carrier of urban power and communication lines, and the stability of its internal environment is crucial to the safe operation of the cable. To prevent groundwater infiltration and moisture erosion, a waterproof layer is usually laid on the inner wall of the cable trench. The waterproof layer may have defects such as pinholes, cracks, and material aging and thinning during construction or long-term use, which can damage its integrity. Therefore, it needs to be regularly tested for non-destructive testing, and the spark testing method is one of the common techniques for evaluating the integrity of such insulating coatings.

[0003] The existing spark testing technology, when applied to the detection of the waterproof layer of a cable trench, usually uses a single high-voltage DC power supply, and the tester holds the probe to scan the surface of the waterproof layer. When the probe passes through the defect, the high voltage will break through the air or medium at the defect to form a spark, and the instrument will issue an audible and visual alarm to indicate the presence of a defect. The tester manually records the approximate location of the defect according to the alarm signal.

[0004] However, the detection results of the existing detection method are relatively single, and only the presence or absence of defects can be determined, without providing more in-depth information about the type, size, or severity of the defects, resulting in a lack of targetedness in maintenance decisions. Secondly, the detection process is greatly affected by human and environmental factors, and the manually controlled probe pressure is unstable, which may cause poor contact and lead to missed detection, or excessive pressure that damages the waterproof layer; the fixed detection voltage also cannot adapt to changes in temperature and humidity on site, which may cause false positives or false negatives. Finally, the recording and sorting of detection results rely on manual work, which is inefficient and prone to errors, and cannot form an intuitive defect distribution map, which is not conducive to the macro evaluation of the overall condition of the waterproof layer. SUMMARY

[0005] To solve the above problems, the present application provides a spark non-destructive testing evaluation method for the integrity of a waterproof layer of a cable trench, which uses a detection method combining multi-frequency electrical signal excitation, spark signal feature extraction and analysis, and can distinguish the type of defects in the waterproof layer and evaluate their severity, improving the accuracy of detection.

[0006] The above objectives can be achieved by the following solutions:

[0007] An electrical discharge non-destructive testing evaluation method for integrity of a waterproof layer of a cable trench, comprising generating a multi-frequency electrical signal containing a plurality of frequency components, and applying the multi-frequency electrical signal through a preset mobile electrode; acquiring an initial contact pressure of the mobile electrode and a surface of the waterproof layer in real time, and adjusting the pressure of the contact position of the mobile electrode based on the initial contact pressure; based on the adjusted pressure, capturing an electrical discharge signal generated by the multi-frequency electrical signal applied by the mobile electrode at a defect of the waterproof layer of the cable trench, extracting feature parameters of the electrical discharge signal to generate current electrical discharge feature data; distinguishing defect types using the current electrical discharge feature data to generate a defect evaluation result; and outputting defect distribution information and triggering an audible and visual alarm according to the defect evaluation result.

[0008] Optionally, the generating a multi-frequency electrical signal containing a plurality of frequency components comprises: acquiring a frequency range parameter for defining a signal frequency range, and determining a plurality of different frequency components of the electrical signal based on the frequency range parameter; acquiring real-time environmental parameter data, and adjusting the output power of the electrical signal based on the environmental parameter data; and generating the multi-frequency electrical signal based on the plurality of different frequency components and the output power.

[0009] Optionally, the acquiring real-time environmental parameter data, and adjusting the output power of the electrical signal based on the environmental parameter data comprises: separating environmental humidity parameters and surface temperature parameters from the real-time environmental parameter data; calculating an equivalent environmental impedance based on the environmental humidity parameters and the surface temperature parameters; determining a power adjustment parameter according to the equivalent environmental impedance; and adjusting the output power of the electrical signal based on the power adjustment parameter.

[0010] Optionally, the acquiring an initial contact pressure of the mobile electrode and a surface of the waterproof layer in real time, and adjusting the pressure of the contact position of the mobile electrode based on the initial contact pressure comprises: acquiring the initial contact pressure of the mobile electrode and the surface of the waterproof layer in real time to generate target pressure data; determining a pressure range for maintaining stable electrical contact without damaging the waterproof layer; comparing the target pressure data with the pressure range in real time to generate a pressure adjustment instruction; and adjusting the pressure of the contact position of the mobile electrode according to the pressure adjustment instruction.

[0011] Optionally, the capturing an electrical discharge signal generated by the multi-frequency electrical signal applied by the mobile electrode at a defect of the waterproof layer of the cable trench based on the adjusted pressure, and extracting feature parameters of the electrical discharge signal to generate current electrical discharge feature data comprises: capturing the electrical discharge signal generated by the multi-frequency electrical signal applied by the mobile electrode at the defect of the waterproof layer of the cable trench based on the adjusted pressure; performing feature processing on the electrical discharge signal to obtain a signal amplitude; determining whether the signal amplitude is greater than a preset amplitude threshold; and if not, determining that there is no defect.

[0012] Optionally, the capturing, based on the adjusted pressure, the electric spark signal generated by the multi-frequency electric signal applied by the mobile electrode at the waterproof layer defect of the cable trench, extracting the characteristic parameters of the electric spark signal to generate the current electric spark characteristic data further comprises: if the signal amplitude is greater than the amplitude threshold, analyzing the waveform of the electric spark signal to obtain a spark intensity parameter; determining the start and end time points of the waveform of the electric spark signal to obtain a spark duration parameter; identifying the dominant frequency component when the electric spark signal occurs; and combining the spark intensity parameter, the spark duration parameter and the dominant frequency component to generate the current electric spark characteristic data.

[0013] Optionally, the identifying the dominant frequency component when the electric spark signal occurs comprises: locating an electric signal segment within the electric spark signal occurrence time window; performing frequency spectrum analysis on the electric signal segment to generate frequency spectrum data; obtaining the energy values of each frequency point in the frequency spectrum data and sorting the energy values of each frequency point from large to small; and selecting the frequency component corresponding to the frequency point with the largest energy value as the dominant frequency component.

[0014] Optionally, the distinguishing the defect type using the current electric spark characteristic data to generate a defect evaluation result comprises: collecting historical defect types and historical electric spark characteristic data corresponding to the historical defect types to obtain a historical data set; using the historical electric spark characteristic data as input and the defect type labels corresponding to the historical defect types as output, establishing and training a neural network model using the historical data set to obtain a defect classification model; inputting the current electric spark characteristic data into the defect classification model to output a current defect type label; calculating a current defect severity score based on the current electric spark characteristic data; and combining the current defect type label and the current defect severity score to generate a defect evaluation result.

[0015] Optionally, the outputting defect distribution information and triggering sound-light alarm according to the defect evaluation result comprises: extracting defect location information based on the position of the mobile electrode; generating defect distribution information based on the defect location information and the defect evaluation result; and triggering differential sound-light alarm according to the defect distribution information.

[0016] Based on the same inventive concept, the application also provides an electric spark non-destructive detection evaluation system for the integrity of a cable trench waterproof layer, comprising: a signal generation module for generating a multi-frequency electric signal containing multiple frequency components and applying the multi-frequency electric signal through a preset moving electrode; a pressure control module for acquiring the initial contact pressure of the moving electrode and the surface of the waterproof layer in real time and adjusting the pressure of the contact position of the moving electrode based on the initial contact pressure; a feature extraction module for capturing the electric spark signal generated by the multi-frequency electric signal applied by the moving electrode at the defect of the cable trench waterproof layer based on the adjusted pressure, extracting the characteristic parameters of the electric spark signal to generate current electric spark feature data; a defect analysis module for distinguishing the defect type by using the current electric spark feature data to generate a defect evaluation result; and an alarm module for outputting defect distribution information and triggering sound and light alarms according to the defect evaluation result.

[0017] Compared with the prior art, the application has the following advantages:

[0018] 1. The application realizes accurate distinction of defect types and quantitative evaluation of severity by introducing multi-frequency electric signal excitation and combining analysis of electric spark characteristic parameters, can provide users with detailed information about defect nature and harmfulness, provides a scientific basis for subsequent maintenance decisions, and improves the fine level of detection and the accuracy of evaluation;

[0019] 2. The application dynamically adjusts the contact pressure of the moving electrode, ensures stable application of the detection signal on uneven surfaces and avoids physical damage to the waterproof layer; at the same time, adjusts the signal output power according to the environmental temperature and humidity, effectively eliminates the interference of environmental factors, and thus improves the reliability, stability and repeatability of the detection results under different working conditions;

[0020] 3. The application realizes automatic classification of defects by using a neural network model, reduces the dependence on the experience of operators; at the same time, binds the defect evaluation result with the position information, generates visual defect distribution information and triggers differentiated sound and light alarms, realizes full-process automation from data acquisition to result presentation and risk warning, and improves the on-site detection efficiency, safety and user experience.

[0021] Other features and advantages of the application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the application. The objects and other advantages of the application can be realized and obtained by the structure indicated in the specification, claims and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings described below are only some embodiments of the present application, and the drawings can also be obtained by those skilled in the art without creative effort.

[0023] Figure 1 is a flowchart of a method for evaluating the integrity of the waterproof layer of the cable trench according to an embodiment of the present application.

[0024] Figure 2 is a time-domain waveform diagram of a multi-frequency electrical signal according to an embodiment of the present application.

[0025] Figure 3 is a spectrum diagram of a multi-frequency electrical signal according to an embodiment of the present application.

[0026] Figure 4 is a time-domain fragment diagram of an electrical spark signal according to an embodiment of the present application.

[0027] Figure 5 is a diagram of the dominant frequency component of an electrical spark signal according to an embodiment of the present application.

[0028] Figure 6 is a schematic diagram of the principle of classifying defect types using a neural network model according to an embodiment of the present application.

[0029] Figure 7 is a structural diagram of a system for evaluating the integrity of the waterproof layer of the cable trench according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0031] With reference to Figure 1 An embodiment of the present application provides a method for evaluating the integrity of the waterproof layer of the cable trench. The method adopts a detection mode combining multi-frequency electrical signal excitation, electrical spark signal feature extraction and analysis, and can distinguish the defect type of the waterproof layer, evaluate the severity thereof, and improve the accuracy of detection.

[0032] The method according to the embodiment specifically includes the following steps.

[0033] generate a multi-frequency electrical signal containing multiple frequency components, and apply the multi-frequency electrical signal through a preset mobile electrode;

[0034] acquire an initial contact pressure of the mobile electrode and the waterproof layer surface in real time, and adjust the pressure of the mobile electrode contact position based on the initial contact pressure;

[0035] based on the adjusted pressure, capture an electric spark signal generated by the multi-frequency electrical signal applied by the mobile electrode at the defect of the cable trench waterproof layer, extract feature parameters of the electric spark signal to generate current electric spark feature data;

[0036] distinguish the defect type using the current electric spark feature data to generate a defect evaluation result;

[0037] According to the defect evaluation result, output defect distribution information and trigger sound and light alarm.

[0038] The present application first uses a composite electrical signal containing multiple frequency components as an excitation source, which is applied to the surface of the cable trench waterproof layer through a mobile electrode, aiming to use the response difference of different frequency signals to different types of defects to improve the detection range. Through real-time pressure feedback control, the contact pressure of the electrode and the waterproof layer surface is dynamically adjusted to ensure the stability and non-destructiveness of signal application. When the mobile electrode passes through the defect, the electric spark signal generated is captured by the system, and its features are extracted to generate a set of current electric spark feature data that can describe the nature of the discharge event. Then, the system uses this set of feature data to distinguish the type of defect, thereby obtaining a defect evaluation result containing property judgment. Finally, the evaluation result is combined with the position information of the mobile electrode to generate the macroscopic distribution of the defect, and according to the severity of the evaluation, the corresponding sound and light alarm is triggered to realize a complete technical closed loop from detection, analysis to early warning.

[0039] The present application uses multi-frequency electrical signal detection and combines deep analysis and classification of electric spark feature parameters, so that detection is no longer limited to simply judging whether there is a defect, but can distinguish the specific type of defect, improving the fineness and accuracy of evaluation. Secondly, real-time adaptive pressure control effectively overcomes measurement errors caused by uneven surface or improper human operation, ensuring the stability of the detection process and the repeatability of the results, while ensuring the non-destructiveness of the detection, avoiding secondary damage to the waterproof layer. Finally, the evaluation result is combined with the position information to generate intuitive defect distribution information and trigger differentiated sound and light alarms, providing immediate and clear guidance for on-site operators, improving detection efficiency and the scientific nature of subsequent maintenance decisions, and improving the intelligent level and engineering practical value of cable trench waterproof layer integrity detection.

[0040] Optionally, the generation of the multi-frequency electrical signal containing multiple frequency components comprises:

[0041] obtaining a frequency range parameter for defining a signal frequency range, and determining a plurality of different frequency components of the electrical signal based on the frequency range parameter;

[0042] obtaining real-time environmental parameter data, and adjusting an output power of the electrical signal based on the environmental parameter data;

[0043] generating a multi-frequency electrical signal based on the plurality of different frequency components and the output power.

[0044] Specifically, first, the system pre-sets a frequency range parameter for defining a signal frequency range, which defines the lowest frequency and the highest frequency of the detection signal, aiming to cover the response characteristics of various types of defects that may occur in the waterproof layer of the cable trench. Based on this frequency range parameter, the system internally determines one or more sets of discrete frequency components, which constitute the basis of the multi-frequency electrical signal. Next, the system obtains real-time environmental parameter data through integrated environmental sensors, including key data such as environmental humidity and surface temperature. The system dynamically adjusts the output power of the signal generator based on these environmental parameter data, especially the influence of humidity and temperature on the surface electrical properties of the waterproof layer. This adjustment process ensures that the electric field intensity applied to the surface of the waterproof layer maintains an effective and safe level in both dry and humid environments, avoiding missed detection or false positives due to environmental changes. Finally, the signal generator superimposes and synthesizes the aforementioned plurality of different frequency components, and generates the final multi-frequency electrical signal for detection with the output power adjusted by the environment compensation. This signal is output to the mobile electrode, providing a stable and highly sensitive excitation source for subsequent electrical spark detection. As shown in Figure 2 , the time-domain waveform of the composite signal superimposed by a plurality of different frequency sine waves is shown, and as shown in Figure 3 , the corresponding frequency spectrum diagram contains frequency components of 50Hz, 200Hz, and 500Hz.

[0045] Optionally, the obtaining real-time environmental parameter data, and adjusting an output power of the electrical signal based on the environmental parameter data comprises:

[0046] separating environmental humidity parameters and surface temperature parameters from the real-time environmental parameter data;

[0047] calculating an equivalent environmental impedance based on the environmental humidity parameters and the surface temperature parameters;

[0048] determining a power adjustment parameter according to the equivalent environmental impedance;

[0049] adjusting the output power of the electrical signal based on the power adjustment parameter.

[0050] In particular, the system first continuously collects environmental parameter data on-site through integrated sensors deployed on the detection device. The system separates environmental humidity parameters and surface temperature parameters from these composite data streams. Environmental humidity parameters reflect the water vapor content in the air, which directly affects the conductivity of the waterproof layer surface; surface temperature parameters affect the dielectric constant and resistivity of the waterproof layer material itself. Subsequently, the system uses a pre-set empirical model to calculate an integrated indicator, i.e. the equivalent environmental impedance, based on the real-time acquired environmental humidity parameters and surface temperature parameters. The equivalent environmental impedance aims to quantify the comprehensive impact of the current environmental conditions on the propagation and leakage of electrical signals on the waterproof layer surface. The calculation can follow the following model:

[0051] ,

[0052] where, represents the calculated equivalent environmental impedance; is the reference environmental impedance measured under standard test conditions, serving as a fixed reference value; is the environmental humidity parameter acquired in real time through the humidity sensor; is the surface temperature parameter acquired in real time through the temperature sensor; is a dimensionless function representing the comprehensive impact of humidity and temperature changes on impedance, and the form and coefficients of this function are usually obtained by fitting and calibrating a large amount of experimental data. After obtaining the equivalent environmental impedance, the system will determine a power adjustment parameter based on this value. This parameter is used to dynamically adjust the output power of the electrical signal to compensate for the impact of environmental changes, with the core goal of maintaining the effective electric field strength applied to the waterproof layer constant. The determination of the power adjustment parameter usually follows a positive correlation with the equivalent environmental impedance, i.e. when the environmental impedance increases, the output power is appropriately increased to ensure that the air gap at the defect can be effectively broken down, and vice versa when the environmental impedance decreases, the output power is reduced to avoid false alarms caused by the enhancement of surface conductivity. Finally, the system applies this power adjustment parameter to the signal generator to adjust the final output power of the multi-frequency electrical signal in real time, completing a closed-loop power adaptive adjustment.

[0053] Optionally, the real-time acquisition of the initial contact pressure of the mobile electrode and the waterproof layer surface, and the adjustment of the pressure of the mobile electrode contact position based on the initial contact pressure include:

[0054] Real-time acquisition of the initial contact pressure of the mobile electrode and the waterproof layer surface to generate target pressure data;

[0055] Determination of a pressure range for maintaining stable electrical contact without damaging the waterproof layer;

[0056] Real-time comparison of the target pressure data with the pressure range to generate pressure adjustment instructions;

[0057] adjusting the pressure of the mobile electrode contact position according to the pressure adjustment instruction.

[0058] Specifically, first, a pressure sensor is integrated in the probe assembly of the mobile electrode. The sensor can monitor the normal force generated when the electrode contacts the surface of the waterproof layer of the cable trench in real time, and convert the force signal into an electrical signal. After processing, continuous target pressure data is generated. At the same time, a key parameter is pre-set in the system, that is, a pressure range for maintaining stable electrical contact and not damaging the waterproof layer. The pressure range includes a minimum pressure threshold and a maximum pressure threshold, which is determined according to the mechanical and electrical properties of the waterproof layer material to be tested through preliminary experiments or according to the material standard. The minimum pressure threshold ensures that a continuous and effective electrical path is formed between the electrode and the surface of the waterproof layer, avoiding signal interruption or attenuation due to poor contact; the maximum pressure threshold ensures that the applied pressure will not cause indentation, scratches or structural damage to the waterproof layer, ensuring the non-destructive nature of the detection. During the detection process, the system continuously compares the real-time target pressure data with the pre-set pressure range. If the target pressure data is lower than the minimum pressure threshold, the control unit generates a pressure adjustment instruction to increase the pressure; if the target pressure data is higher than the maximum pressure threshold, a pressure adjustment instruction to reduce the pressure is generated; if the pressure is within the pre-set range, the instruction maintains the current state. Finally, the pressure adjustment instruction is sent to a servo drive or pneumatic actuator connected to the mobile electrode. The mechanism accurately adjusts the vertical position or applied thrust of the mobile electrode according to the instruction, thereby dynamically adjusting the pressure of the mobile electrode contact position, forming a closed-loop adaptive pressure control system.

[0059] Optionally, based on the adjusted pressure, capturing the electric spark signal generated by the multi-frequency electric signal applied by the mobile electrode at the defect of the waterproof layer of the cable trench, and extracting the characteristic parameters of the electric spark signal to generate current electric spark characteristic data includes:

[0060] capturing the electric spark signal generated by the multi-frequency electric signal applied by the mobile electrode at the defect of the waterproof layer of the cable trench based on the adjusted pressure;

[0061] characteristic processing of the electric spark signal to obtain a signal amplitude;

[0062] determining whether the signal amplitude is greater than a pre-set amplitude threshold;

[0063] If not, it is determined that there is no defect.

[0064] Specifically, after the contact pressure between the mobile electrode and the surface of the waterproof layer is precisely adjusted and stabilized within a preset range, the mobile electrode continuously applies a multi-frequency electrical signal to the waterproof layer. The system monitors the electrical signal in the electrode loop in real time at a high sampling rate. When the mobile electrode sweeps through the defect area of the waterproof layer of the cable trench, such as a pinhole or a crack, due to the destruction of the insulation performance at the defect, the multi-frequency electrical signal will break through the air gap at that location, generating a transient and high-energy discharge pulse, which is the electrical spark signal. This signal is completely captured by the data acquisition unit. Next, the system processes the captured electrical spark signal for features. The first extracted feature is the signal amplitude. The signal amplitude, usually referring to the peak voltage or peak current of the electrical spark signal in the time waveform, is the most direct parameter representing the strength of the discharge energy. The system compares the extracted signal amplitude with a preset amplitude threshold. The amplitude threshold is determined by testing standard samples of the waterproof layer without defects, analyzing the background noise level, and leaving a sufficient safety margin. If the currently measured signal amplitude is not greater than the amplitude threshold, the system determines that the signal is background noise or meaningless weak fluctuations, rather than a discharge caused by a real defect. In this case, the system determines that the current detection position is defect-free and continues to scan and detect subsequent areas without further analyzing the signal.

[0065] Optionally, based on the adjusted pressure, capturing the electrical spark signal generated by the multi-frequency electrical signal applied by the mobile electrode at the defect of the waterproof layer of the cable trench, and extracting the feature parameters of the electrical spark signal to generate current electrical spark feature data, further comprises:

[0066] If the signal amplitude is greater than the amplitude threshold, analyzing the waveform of the electrical spark signal to obtain a spark intensity parameter;

[0067] Determining the start and end time points of the waveform of the electrical spark signal to obtain a spark duration parameter;

[0068] Identifying the dominant frequency component when the electrical spark signal occurs;

[0069] Combining the spark intensity parameter, the spark duration parameter, and the dominant frequency component to generate current electrical spark feature data.

[0070] Specifically, first, the system analyzes the complete waveform of the valid electrical spark signal to obtain a spark intensity parameter. This parameter is usually obtained by calculating the integral of the envelope line of the signal waveform on the time axis, and it reflects the total energy released by a single discharge event, which is an important indicator for measuring the severity of defects. Its calculation model can be represented as:

[0071] ,

[0072] wherein, a spark intensity parameter; is the voltage function of the electrical spark signal over time; the integral interval is the entire duration of the electrical spark signal. At the same time, the system accurately determines the start and end time points of the electrical spark signal waveform through an algorithm, and the time difference between the two time points is defined as the spark duration parameter. The start time point is usually defined as the time when the signal amplitude first stabilizes above the background noise level, and the end time point is the time when the signal amplitude last decays below the background noise level. This parameter reflects the persistence of the discharge process and is related to the geometric morphology of the defect. Then, the system performs spectral analysis on the time segment where the electrical spark signal occurs to identify the dominant frequency component that plays a decisive role in this discharge event. This process involves performing a fast Fourier transform on the signal segment to generate spectral data, and then finding the frequency with the highest energy proportion from the spectral data, which is the dominant frequency component. Finally, the system structurally combines the calculated spark intensity parameter, spark duration parameter and identified dominant frequency component to form a multi-dimensional feature vector, which is the current electrical spark feature data for subsequent defect type and severity assessment.

[0073] Optionally, the identifying the dominant frequency component when the electrical spark signal occurs comprises:

[0074] locating an electrical signal segment within the time window of the electrical spark signal occurrence;

[0075] performing spectral analysis on the electrical signal segment to generate spectral data;

[0076] obtaining the energy values of each frequency point in the spectral data, and sorting the energy values of each frequency point from large to small;

[0077] selecting the frequency component corresponding to the frequency point with the first energy value as the dominant frequency component.

[0078] Specifically, first, the captured original electrical signal is positioned in time, and the system uses the determined start and end time points of the electrical spark signal to extract the electrical signal segment containing the complete discharge process from the continuous data stream. This time window ensures that the analysis object is limited to the signal part directly related to the defect discharge, excluding irrelevant background noise before and after. Then, the system performs spectral analysis on the electrical signal segment, which is usually achieved by performing a fast Fourier transform algorithm to convert the signal from time domain representation to frequency domain representation, thereby generating spectral data. The spectral data intuitively shows the frequency components and their corresponding intensities that make up the electrical signal segment. In order to quantify these intensities, the system further obtains the energy value of each discrete frequency point in the spectral data. A common way to calculate is to take the modulus square of the Fourier transform result of the frequency point to obtain the power spectral density:

[0079] ,

[0080] wherein, represents a frequency point corresponding energy value; is a complex value at the frequency point obtained by fast Fourier transform; is the total number of data points in the electric signal segment, used for normalization processing. After calculating the energy values of all frequency points, the system sorts these energy values from large to small. Finally, the system selects the frequency point ranked first in the energy value sorting result, and the frequency component corresponding to the frequency point is determined as the dominant frequency component of this time electric spark event and is output as one of the key features. As shown in Figure 4 , an electric spark signal segment positioned and cut out from the original signal stream is shown. As shown in Figure 5 , a frequency spectrum obtained after fast Fourier transform (FFT) is performed on the segment is shown, wherein the frequency point with the highest energy is identified as the dominant frequency component, and the dominant frequency is 2500 Hz.

[0081] Optionally, the distinguishing of the defect type by using the current electric spark feature data to generate a defect evaluation result comprises:

[0082] collecting historical defect types and historical electric spark feature data corresponding to the historical defect types to obtain a historical data set;

[0083] using the historical data set to establish and train a neural network model by taking the historical electric spark feature data as input and taking a defect type label corresponding to the historical defect type as output, to obtain a defect classification model;

[0084] inputting the current electric spark feature data into the defect classification model to output a current defect type label;

[0085] calculating a current defect severity score based on the current electric spark feature data;

[0086] combining the current defect type label and the current defect severity score to generate a defect evaluation result.

[0087] Specifically, first, a large number of historical defect samples are collected through experimental means, which includes artificially manufacturing various known historical defect types on standard waterproof layer materials, such as pinholes, cracks, material thinning, bubbles, etc., and labeling each type with a clear defect type label. Subsequently, the detection system is used to scan these samples, capture their corresponding spark signals, and extract historical spark feature data. In this way, a historical data set containing a large amount of feature data and its corresponding defect type label is constructed. Based on this data set, the system uses a supervised learning method to establish and train a neural network model. In this process, the historical spark feature data is used as the input of the model, and the corresponding defect type label is used as the expected output. The network weights are continuously adjusted through algorithms such as backpropagation until the model can accurately identify the defect type from the feature data. Finally, a trained defect classification model is obtained. In the online real-time evaluation stage, when a new defect is captured and current spark feature data is generated, the data is first input into the pre-trained defect classification model. The model will analyze and calculate the current feature data based on the knowledge it has learned from the historical data set, and output a most likely defect type label, such as "pinhole" or "crack". At the same time, the system also calculates a current defect severity score based on the current spark feature data. The score is usually a comprehensive quantitative indicator, which can be obtained by weighted calculation of key parameters in the feature data, for example:

[0088] ,

[0089] where, represents the current defect severity score; is the spark intensity parameter extracted from the current spark feature data; is the spark duration parameter extracted from the current spark feature data; and are preset weight coefficients, which are calibrated according to a large amount of experimental data to reflect the contribution of different parameters to defect severity. Finally, the system combines the current defect type label output by the neural network model with the calculated current defect severity score to generate a detailed defect evaluation result that not only indicates the nature of the defect but also quantifies its severity. As shown in Figure 6 , a two-dimensional feature space is shown, composed of spark intensity and spark duration, and different types of historical defect data form different clusters in the space. When a current detected defect data point, i.e. current spark feature data, is input, the model can accurately determine its defect type based on its position in the space.

[0090] Optionally, the outputting defect distribution information and triggering sound-light alarm according to the defect evaluation result comprises:

[0091] Based on the position of the mobile electrode, the defect position information is extracted;

[0092] Based on the defect position information and the defect evaluation result, the defect distribution information is generated;

[0093] According to the defect distribution information, the differentiated sound-light alarm is triggered.

[0094] Specifically, first, the system tracks the position of the mobile electrode in the cable trench in real time through the positioning unit integrated on the mobile electrode device, such as a mileage encoder or an inertial measurement unit. Whenever the system generates a defect evaluation result based on the analysis of the electric spark signal, it immediately binds the result with the position data output by the positioning unit at the same time, thereby extracting accurate defect position information, which is usually expressed in coordinates relative to a reference point or mileage. Then, the system stores this complete data point containing defect position information, defect type label and defect severity score. As the detection work continues, all identified defect data points are aggregated into a database. Based on this database, the system can generate intuitive defect distribution information, which is usually represented as a digital map or diagram, marking the position, type and severity of each defect on a two-dimensional or three-dimensional model of the cable trench with different colors or symbols, thereby providing the operator with a macroscopic view of the overall health of the waterproof layer. Finally, the system uses real-time defect evaluation results to trigger differentiated sound-light alarms. The system has a set of preset alarm rules, which correspond different defect severity score levels to specific sound-light alarm modes. For example, when a low-severity defect is detected, the system may trigger a yellow warning light to flash and a slow prompt sound; when a high-severity defect is detected, it will immediately trigger a red high-frequency flashing light and a loud, urgent alarm sound.

[0095] Based on the same inventive concept, as Figure 7 shown, the present application also provides an electric spark non-destructive detection and evaluation system for the integrity of the waterproof layer of a cable trench, which comprises:

[0096] A signal generation module for generating a multi-frequency electric signal containing multiple frequency components and applying the multi-frequency electric signal through a pre-set mobile electrode;

[0097] A pressure control module for acquiring the initial contact pressure of the mobile electrode and the surface of the waterproof layer in real time and adjusting the pressure of the contact position of the mobile electrode based on the initial contact pressure;

[0098] a feature extraction module configured to capture, based on the adjusted pressure, an electric spark signal generated by the multi-frequency electric signal applied by the mobile electrode at the defect of the waterproof layer of the cable trench, and extract a feature parameter of the electric spark signal to generate current electric spark feature data;

[0099] a defect analysis module configured to distinguish defect types by using the current electric spark feature data to generate a defect evaluation result;

[0100] an alarm module configured to output defect distribution information and trigger audible and visual alarms according to the defect evaluation result.

[0101] To verify the feasibility of the application in implementation, the application is applied to a 5-kilometer-long cable trench waterproof layer integrity acceptance detection project in a certain project. The cable trench adopts polyurethane waterproof coating, and the acceptance standard is extremely high. The traditional manual observation and single-frequency electric spark detection method has the problems of low efficiency, easy to be affected by environmental humidity to produce false alarm, unable to distinguish defect types, and possible secondary damage to the waterproof layer due to improper operation.

[0102] In the embodiment, the detection device scans along the cable trench at a preset speed. The system first generates a multi-frequency electric signal containing multiple frequency components (1 kHz-20 kHz). During the detection process, the system obtains environmental parameters in real time through the integrated sensor. For example, at 10 o'clock on a certain day, the field environment is dry, the temperature is 28℃, and the humidity is 50%. The system calculates a relatively high equivalent environmental impedance based on this and automatically adjusts the output power to 110% of the preset reference value; at 15 o'clock in the afternoon, after the rain, the field humidity rises to 90% and the temperature drops to 24℃. The system calculates a relatively low equivalent environmental impedance in real time and quickly reduces the output power to 95% of the reference value to avoid false alarms caused by enhanced surface conductivity and ensure the consistency of all-weather detection results.

[0103] During the movement of the mobile electrode, the pressure control module monitors the contact pressure in real time. There are uneven areas on the surface of the cable trench. When moving to K2+350 meters, a small concrete protrusion makes the initial contact pressure of the mobile electrode reach 6.2N instantaneously, which exceeds the preset safe pressure range of 2.0N to 5.0N. The pressure control module generates a pressure reduction instruction within 50 milliseconds, adjusts the electrode height through the servo mechanism, and quickly stabilizes the pressure at 4.5N, effectively avoiding scratching the waterproof layer.

[0104] When the detection device travels to K1+120 meters, the system captures a transient electric spark signal with a signal amplitude of 5.5V, which is greater than the preset amplitude threshold of 1.0V, and is determined as an effective defect signal. The feature extraction module immediately performs in-depth analysis on the signal, and calculates the spark intensity parameter as 3.8V·μs and the spark duration parameter as 2.5μs. At the same time, through spectral analysis on the signal segment, it is identified that the 15kHz with the highest energy proportion is the dominant frequency component. These parameters are combined into the current electric spark feature data.

[0105] The feature data is input into the neural network defect classification model pre-trained by the historical data set. The defect type label output by the model after analysis is "pinhole". At the same time, the system calculates the current defect severity score as 85 points based on the spark intensity and duration, with a full score of 100. Combined with the defect type and severity score, the system generates the final defect evaluation result: "position K1+120, type: pinhole, severity: high". Since the severity score is higher than 80 points, the alarm module immediately triggers a high-frequency red audible and light alarm.

[0106] Subsequently, at K3+500 meters, the system detects another effective defect signal with an amplitude of 3.2V. The feature extraction module analyzes and obtains the spark intensity as 2.1V·μs, the spark duration as 6.8μs, and the dominant frequency component as 8kHz. After analysis of this set of feature data by the model, the output defect type label is "crack", and the calculated defect severity score is 60 points. The system generates the evaluation result: "position K3+500, type: crack, severity: medium". Since the severity score is medium, i.e. 50-80 points, the system triggers a lower-pitched yellow audible and light alarm.

[0107] After the entire detection is completed, the system automatically generates a defect distribution map of the 5-kilometer cable trench waterproof layer based on all recorded defect position information and evaluation results, which directly indicates the accurate position, type, and severity level of each defect point, providing guidance for subsequent point repair work.

[0108] It should be noted that the above formulas can be translated into unitless standard values or same-dimension superimposable parameters by using the principle of dimensional consistency and mathematical standardization methods (such as normalization processing, dimensionless parameter conversion, or unit system unification), so as to eliminate the interference of different dimensions on the operation logic, make the formula retain the original data distribution characteristics, and have mathematical operation rationality and objective law adaptability. It is a conventional technical means, which will not be described here. The electrical connection between the above-mentioned units does not necessarily mean direct connection or indirect connection, as long as the purpose of the present application is achieved, it can be applied to the embodiments of the present application. The above-described is only an exemplary embodiment of the present application, which cannot limit the scope of the present application.

[0109] That is, any equivalents of the above described subject matter, as well as other conventional equivalents, are intended to be covered herein. Various embodiments can be implemented in hardware, software, or a combination thereof. The various embodiments can be implemented in one or more computer systems or other processing systems.

Claims

1. A non-destructive testing method for evaluating the integrity of waterproof layers in cable trenches using electrical spark testing, characterized in that, The method includes: A multi-frequency electrical signal containing multiple frequency components is generated and applied through a preset movable electrode; the process includes: acquiring a frequency range parameter to define the signal frequency range, and determining multiple different frequency components of the electrical signal based on the frequency range parameter; acquiring real-time environmental parameter data, and separating environmental humidity and surface temperature parameters from the real-time environmental parameter data; calculating an equivalent environmental impedance based on the environmental humidity and surface temperature parameters; determining a power adjustment parameter based on the equivalent environmental impedance; adjusting the output power of the electrical signal based on the power adjustment parameter; and generating the multi-frequency electrical signal based on the multiple different frequency components and the output power. The initial contact pressure between the moving electrode and the surface of the waterproof layer is acquired in real time, and the pressure at the contact position of the moving electrode is adjusted based on the initial contact pressure. Based on the adjusted pressure, the system captures the electric spark signal generated at the defect in the waterproof layer of the cable trench by the multi-frequency electrical signal applied by the moving electrode. It then extracts the characteristic parameters of the electric spark signal to generate current electric spark characteristic data. This includes: capturing the electric spark signal generated at the defect in the waterproof layer of the cable trench by the multi-frequency electrical signal applied by the moving electrode based on the adjusted pressure; performing feature processing on the electric spark signal to obtain the signal amplitude; determining whether the signal amplitude is greater than a preset amplitude threshold; if not, determining that there is no defect; if the signal amplitude is greater than the amplitude threshold, analyzing the waveform of the electric spark signal to obtain a spark intensity parameter; determining the start and end times of the waveform of the electric spark signal to obtain a spark duration parameter; identifying the dominant frequency component when the electric spark signal occurs; and combining the spark intensity parameter, the spark duration parameter, and the dominant frequency component to generate current electric spark characteristic data. The current electrical spark characteristic data is used to distinguish defect types in order to generate defect assessment results; Based on the defect assessment results, output defect distribution information and trigger an audible and visual alarm.

2. The method for non-destructive testing and evaluation of the integrity of waterproof layers in cable trenches according to claim 1, characterized in that, The step of acquiring the initial contact pressure between the movable electrode and the surface of the waterproof layer in real time, and adjusting the contact position of the movable electrode based on the initial contact pressure, includes: The initial contact pressure between the moving electrode and the surface of the waterproof layer is acquired in real time to generate target pressure data; Determine the pressure range required to maintain stable electrical contact without damaging the waterproofing layer; The target pressure data is compared with the pressure range in real time to generate a pressure adjustment command; Adjust the pressure at the contact position of the moving electrode according to the pressure adjustment command.

3. The method for non-destructive testing and evaluation of the integrity of waterproof layers in cable trenches according to claim 1, characterized in that, The identification of the dominant frequency component during the generation of the electrical spark signal includes: Locate the electrical signal segment within the time window of the electrical spark signal occurrence; Perform spectral analysis on the electrical signal segment to generate spectral data; Obtain the energy value of each frequency point in the spectrum data, and sort the energy values ​​of each frequency point from largest to smallest; The frequency component corresponding to the frequency point with the highest energy value is selected as the dominant frequency component.

4. The method for non-destructive testing and evaluation of the integrity of waterproof layers in cable trenches according to claim 3, characterized in that, The step of using the current electrical spark feature data to distinguish defect types and generate defect assessment results includes: Collect historical defect types and corresponding historical electrical spark feature data to obtain a historical dataset; Using historical electrical spark feature data as input and the defect type labels corresponding to historical defect types as output, a neural network model is established and trained using the historical dataset to obtain a defect classification model. The current electrical spark feature data is input into the defect classification model to output the current defect type label; The current defect severity score is calculated based on the current electrical spark characteristic data. By combining the current defect type label with the current defect severity score, a defect assessment result is generated.

5. The method for non-destructive testing and evaluation of the integrity of waterproof layers in cable trenches according to claim 4, characterized in that, The step of outputting defect distribution information and triggering an audible and visual alarm based on the defect assessment results includes: Based on the position of the moving electrode, the defect location information is extracted; Based on the defect location information and the defect assessment results, defect distribution information is generated; Differentiated audible and visual alarms are triggered based on the defect distribution information.

Images