Leak detector data intelligent analysis system and method

Abstract

The invention discloses a leak detector data intelligent analysis system and method, and the system carries out the time sequence mode feature extraction of a vibration signal at a first monitoring position through employing a neural network model based on deep learning, so as to excavate the time sequence implicit features of the vibration signal, and preliminarily recognizes the potential leakage abnormality based on the time sequence implicit features. For the abnormal signal, the system further obtains a vibration signal of an adjacent monitoring point, similarly, extracts a time sequence mode coding feature of the vibration signal of the adjacent monitoring point, then calculates a vibration mode space-time correlation degree between the two, and further realizes multi-dimensional verification of the leakage abnormal event through comparison between the correlation degree and a preset threshold value. Through the mode, the limitation of traditional single-point static threshold detection is broken through, the influence of environmental noise and non-leakage transient interference is effectively reduced, and the identification accuracy of the leakage signal is improved, so that the reliability of operation and maintenance decision is enhanced.

Description

Technical Field

[0001] The present application relates to the field of intelligent analysis, and more specifically, to a leak detector data intelligent analysis system and method. Background Art

[0002] In the field of pipeline leakage monitoring, traditional leak detection technology mainly relies on single-point vibration signal analysis and threshold alarm mechanism, but there are significant limitations in practical applications under complex working conditions. Existing systems usually use fixed-band filtering methods, which are difficult to dynamically adapt to signal spectrum characteristics, resulting in the frequency domain overlap problem of interference sources such as mechanical vibration and water flow noise and tiny leakage signals that cannot be effectively separated; especially in strong environmental noise scenarios (such as close-range construction or equipment start-up and shutdown), the weak energy of the leakage signal is easily submerged by noise, and the conventional energy threshold detection algorithm often leads to missed detection or false alarms due to the lack of in-depth analysis of the signal timing pattern; in addition, the transient waveforms generated by non-leakage events (such as rapid valve action or heavy load impact) have similar burst characteristics to the leakage signal, and the existing technology relies on the static model of a single monitoring point and is difficult to achieve accurate distinction.

[0003] Therefore, an optimized intelligent analysis method of leak detector data is expected. Summary of the invention

[0004] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiment of the present application provides a leak detector data intelligent analysis system and method, which uses a neural network model based on deep learning to extract the time series pattern features of the vibration signal of the first monitoring position, so as to dig out the implicit time series features of the vibration signal, and preliminarily identify potential leakage anomalies based on this. In response to the abnormal signal, the system further obtains the vibration signal of the adjacent monitoring point, and similarly extracts the time series pattern coding features of the vibration signal of the adjacent monitoring point, and then calculates the spatiotemporal correlation of the vibration patterns between the two, and then, by comparing the correlation with the preset threshold, a multi-dimensional verification of the leakage anomaly event is achieved. In this way, the limitations of the traditional reliance on single-point static threshold detection are broken through, the influence of environmental noise and non-leakage transient interference is effectively reduced, the recognition accuracy of the leakage signal is improved, and the reliability of operation and maintenance decisions is enhanced.

[0005] According to one aspect of the present application, a method for intelligently analyzing leak detector data is provided, which includes: It extracts the time-series pattern features of the vibration signal at the first monitoring position by using a neural network model based on deep learning to mine the time-series implicit features of the vibration signal, and preliminarily identifies potential leakage anomalies based on this. For this abnormal signal, the system further obtains the vibration signals of adjacent monitoring points. Similarly, it extracts the time-series pattern coding features of the vibration signals of adjacent monitoring points, and then calculates the spatio-temporal correlation degree of the vibration patterns between the two. Furthermore, through the comparison of the correlation degree with a preset threshold, the multi-dimensional verification of the leakage abnormal event is realized. In this way, it breaks through the limitation of traditional single-point static threshold detection, effectively reduces the influence of environmental noise and non-leakage transient interference, improves the recognition accuracy of leakage signals, and thus enhances the reliability of operation and maintenance decisions.

[0006] According to another aspect of the present application, a leak detector data intelligent analysis system is provided, which includes: A first leakage vibration monitoring signal acquisition module, configured to acquire a first leakage vibration monitoring signal collected by a leak detector deployed at a first monitoring position of a pipeline; A filtering processing module, configured to perform filtering processing on the first leakage vibration monitoring signal by applying a band-pass filter to obtain a filtered first leakage vibration monitoring signal; A first vibration time-series pattern feature extraction module, configured to extract the vibration time-series pattern features of the filtered first leakage vibration monitoring signal to obtain a first vibration time-series pattern feature coding vector; An anomaly detection module, configured to input the first vibration time-series pattern feature coding vector into an anomaly detection model to obtain an anomaly detection result; A second leakage vibration monitoring signal acquisition module, configured to, in response to the anomaly detection result being a potential anomaly, acquire a second leakage vibration monitoring signal collected by a leak detector deployed at a second monitoring position of the pipeline, where the first monitoring position and the second monitoring position are adjacent; A second vibration time-series pattern feature extraction module, configured to extract the vibration time-series pattern features of the second leakage vibration monitoring signal to obtain a second vibration time-series pattern feature coding vector; A vibration pattern spatio-temporal correlation coding module, configured to calculate the spatio-temporal correlation degree of the vibration patterns between the second vibration time-series pattern feature coding vector and the first vibration time-series pattern feature coding vector; A leakage result analysis module, configured to determine whether to adjust the anomaly detection result to a high-confidence leakage event based on the comparison between the spatio-temporal correlation degree of the vibration pattern and a preset threshold.

[0007] Compared with the prior art, the present application provides a leak detector data intelligent analysis system and method, which uses a neural network model based on deep learning to extract the time series pattern features of the vibration signal at the first monitoring position, so as to mine the implicit time series features of the vibration signal, and preliminarily identify potential leakage anomalies based on this. In response to the abnormal signal, the system further obtains the vibration signal of the adjacent monitoring point, and similarly extracts the time series pattern coding features of the vibration signal of the adjacent monitoring point, and then calculates the spatiotemporal correlation of the vibration patterns between the two, and then, by comparing the correlation with the preset threshold, realizes multi-dimensional verification of leakage anomaly events. In this way, the limitations of traditional reliance on single-point static threshold detection are broken through, the influence of environmental noise and non-leakage transient interference is effectively reduced, the recognition accuracy of leakage signals is improved, and the reliability of operation and maintenance decisions is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 is a flow chart of a method for intelligently analyzing leak detector data according to an embodiment of the present application; Figure 2 A data flow diagram of a leak detector data intelligent analysis method according to an embodiment of the present application; Figure 3 is a flowchart of sub-step S3 of the leak detector data intelligent analysis method according to an embodiment of the present application; Figure 4 4 is a block diagram of a leak detector data intelligent analysis system according to an embodiment of the present application. DETAILED DESCRIPTION

[0010] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0011] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0012] Although this application makes various references to certain modules in the system according to embodiments of this application, any number of different modules can be used and run on a user terminal and / or a server. The modules are merely illustrative, and different aspects of the system and method can use different modules.

[0013] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the operations before or below do not necessarily have to be executed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more steps can be removed from these processes.

[0014] Next, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the exemplary embodiments described herein.

[0015] In the technical solution of this application, an intelligent analysis method for leak detector data is proposed. Figure 1 It is a flowchart of the intelligent analysis method for leak detector data according to an embodiment of this application. Figure 2 It is a schematic diagram of data flow of the intelligent analysis method for leak detector data according to an embodiment of this application. As Figure 1 and Figure 2 shown, the intelligent analysis method for leak detector data according to an embodiment of this application includes the steps of: S1, obtaining a first leakage vibration monitoring signal collected by a leak detector deployed at a first monitoring position of a pipeline; S2, applying a band-pass filter to the first leakage vibration monitoring signal for filtering to obtain a filtered first leakage vibration monitoring signal; S3, extracting the vibration time series pattern features of the filtered first leakage vibration monitoring signal to obtain a first vibration time series pattern feature coding vector; S4, inputting the first vibration time series pattern feature coding vector into an anomaly detection model to obtain an anomaly detection result; S5, in response to the anomaly detection result being a potential anomaly, obtaining a second leakage vibration monitoring signal collected by a leak detector deployed at a second monitoring position of the pipeline, where the first monitoring position and the second monitoring position are adjacent; S6, extracting the vibration time series pattern features of the second leakage vibration monitoring signal to obtain a second vibration time series pattern feature coding vector; S7, calculating the vibration pattern spatio-temporal correlation degree between the second vibration time series pattern feature coding vector and the first vibration time series pattern feature coding vector; S8, based on the comparison between the vibration pattern spatio-temporal correlation degree and a preset threshold, determining whether to adjust the anomaly detection result to a high-confidence leakage event.

[0016] Specifically, for S1, obtain the first leakage vibration monitoring signal collected by the leak detector deployed at the first monitoring position of the pipeline. It should be understood that the leakage vibration monitoring signal reflects the weak mechanical vibration characteristics inside the pipeline caused by leakage. These signals contain direct physical evidence of the occurrence of leakage and are an important basis for judging the pipeline status and locating the leakage. It is worth mentioning that during the process of obtaining the first leakage vibration monitoring signal, when a pipeline leakage event occurs, mechanical vibration signals in a specific frequency band are generated by the interaction between the fluid pressure wave and the pipe wall. These weak vibration waves propagate along the pipe wall at a specific speed, forming physical characteristics with spatio-temporal correlation. The leak detector is directly installed on the outer surface of the pipeline based on high-sensitivity sensors (such as accelerometers or piezoelectric vibration sensors). By collecting the tiny vibration signals generated by the pipeline, the physical vibration energy is converted into electrical signals. These electrical signals are preliminarily amplified and converted, which constitute the original data collected by the leak detector and reflect the vibration performance of leakage or abnormal events on the pipeline structure. The leak detector has high-frequency response and anti-interference capabilities, can capture the subtle vibration waveforms caused by leakage, and effectively supports subsequent intelligent filtering and pattern recognition algorithms. By deploying the leak detector at the first monitoring position, the system can obtain the real-time vibration information of the pipeline at this key node, providing a reliable data basis for subsequent time-series feature extraction and anomaly discrimination, and ensuring the overall accuracy and stable operation of the intelligent analysis system.

[0017] Specifically, for S2, apply a band-pass filter to the first leakage vibration monitoring signal for filtering to obtain the filtered first leakage vibration monitoring signal. It should be understood that in a complex pipeline leakage monitoring environment, the first leakage vibration monitoring signal collected often contains a large amount of noise and interference signals in non-target frequency bands, such as low-frequency components generated by mechanical vibration, water flow turbulence noise, and other environmental noises. These interference signals not only mask the weak leakage vibration characteristics but also greatly affect the accuracy of subsequent feature extraction and anomaly detection. To improve the effectiveness of the signal, it is necessary to filter the original vibration signal to remove the interference components in irrelevant frequency bands, thereby highlighting the leakage signal characteristics. Therefore, in the technical solution of this application, a band-pass filter is applied to the first leakage vibration monitoring signal, that is, by setting the upper and lower cut-off frequency ranges, only the signal components within this frequency band are retained, and the low-frequency interference below the cut-off frequency and the high-frequency noise above the cut-off frequency are effectively filtered out.

[0018] It is worth mentioning that as an electronic filter, the core working principle of the bandpass filter is to combine the functions of the low-pass filter and the high-pass filter, while allowing the vibration signal in the center frequency band to pass through, while blocking other frequency bands, thereby achieving dynamic focusing on the spectrum of the first leakage vibration monitoring signal. The bandpass filter accurately selects the common frequency band range of the leakage vibration signal, which not only protects the integrity of the effective leakage signal, but also significantly reduces the confusion caused by mechanical vibration and environmental noise. Among them, the filtering process usually includes signal sampling, the application of digital filtering algorithms, and the output of the filtered signal, so that the subsequent vibration time series pattern feature extraction can be analyzed based on purer and more representative signal data. In this way, the signal-to-noise ratio of the first leakage vibration monitoring signal is improved, and the leakage feature representation of the first leakage vibration monitoring signal is significantly enhanced, thereby improving the detection model's recognition accuracy for leakage anomalies and enhancing the reliability and practicality of the entire leakage detection system under complex working conditions.

[0019] In particular, the step S3 extracts the vibration time series pattern characteristics of the first leakage vibration monitoring signal after filtering to obtain a first vibration time series pattern characteristic encoding vector. In a specific example of the present application, Figure 3 As shown, the S3 includes: S31, extracting the vibration local timing pattern characteristics of the first leakage vibration monitoring signal after filtering to obtain the time series of the first vibration local timing pattern feature coding vector; S32, performing vibration pattern timing context coding on the time series of the first vibration local timing pattern feature coding vector to obtain the first vibration timing pattern feature coding vector.

[0020] Specifically, the S31 extracts the vibration local time series pattern feature of the first leakage vibration monitoring signal after filtering to obtain the time series of the first vibration local time series pattern feature coding vector. In the technical solution of the present application, the first leakage vibration monitoring signal after filtering is subjected to the vibration local time series pattern feature extraction based on one-dimensional convolution coding to obtain the time series of the first vibration local time series pattern feature coding vector. It should be understood that although the first leakage vibration monitoring signal after filtering has effectively filtered out many irrelevant noises, the signal itself still contains a large number of complex and weak time series features, which reflect the dynamic change law of the leakage event in the time dimension. In order to deeply mine the key leakage information in the signal, the system adopts the vibration local time series pattern feature extraction method based on one-dimensional convolution coding to convert the continuous vibration signal into a feature representation with more discriminative power and expression ability. Among them, the one-dimensional convolution coding slides along the time axis by designing multiple convolution kernels to locally perceive and feature-respond to the input vibration time series signal to capture the local pattern of the vibration signal in different time windows, such as amplitude change, frequency component and burst signal characteristics.

[0021] Specifically, in the process of extracting the vibration local time series pattern features based on one-dimensional convolutional coding for the first leakage vibration monitoring signal after filtering, first, a one-dimensional convolution operation is performed on the filtered signal to extract the features related to the local vibration in the time series. Subsequently, through the processing of the non-linear activation function and the pooling layer, the compression and abstraction of the vibration local time series signal are realized, and the time series of the first vibration local time series pattern feature coding vector is obtained. In the technical solution of this application, the powerful local feature extraction ability of the convolutional neural network is utilized to effectively capture the subtle changes of the leakage signal in the time dimension and suppress the interference caused by environmental noise or non-leakage events, thereby forming a high-dimensional and information-rich feature representation, providing accurate and strongly expressive inputs for the subsequent structured recursive coding and anomaly detection models. Through one-dimensional convolutional coding, the system can realize the transformation from the original vibration monitoring signal to high-level features, significantly improve the recognition ability of weak leakage features and the discrimination effect of abnormal events, further enhance the sensitivity and reliability of the leak detection system under complex working conditions, and promote the precision and practicality of intelligent pipeline condition monitoring.

[0022] Specifically, in S32, a vibration mode time series context coding is performed on the time series of the first vibration local time series pattern feature coding vector to obtain the first vibration time series pattern feature coding vector. It should be understood that the real leakage vibration signal will form a decaying wavefront with spatio-temporal continuity along the pipe wall, and there are dynamic correlations such as phase delay and energy diffusion between the waveforms in adjacent time windows, while interference signals such as local mechanical shocks often show isolated transient waveforms. Although the traditional one-dimensional convolutional coding can capture local features within a short time window, it cannot model the unique cross-time-step dynamic evolution law of the leakage signal. For example, the vibration wave caused by leakage will show quasi-periodic fluctuations due to changes in fluid pressure during propagation, and its spectral energy distribution shows an exponential decay trend over time. This cross-time-step context dependence requires more refined time series modeling capabilities. Therefore, in order to construct a feature representation system with spatio-temporal collaborative perception ability to realize the joint modeling of the hidden vibration local time series dependence and spatial correlation in the time series of the first vibration local time series pattern feature coding vector, in the technical solution of this application, a vibration mode time series context coding is performed on the time series of the first vibration local time series pattern feature coding vector to obtain the first vibration time series pattern feature coding vector.

[0023] That is, by decoupling time series encoding and structured message passing, joint modeling of the deep spatio-temporal patterns of vibration signals is achieved. Here, a recurrent neural network (RNN) is used to perform preliminary temporal encoding on the time series of the input vibration local temporal pattern feature encoding vectors, giving full play to its inherent temporal memory ability to capture the dynamic features of the signal changing over time, thereby generating the initial encoding vectors for sequence transmission. This step enhances the model's understanding of the signal continuity and context relationship in the time dimension and lays a foundation for subsequent multi-dimensional information fusion. In addition, by introducing a spatio-temporal confidence constraint factor based on the attention mechanism, the time intervals with significant leakage signal features (such as the persistent low-frequency vibration segment caused by leakage) are adaptively focused on, while suppressing the interference of transient interference events (such as isolated pulses caused by valve actions) on the model judgment. At the same time, the introduction of the spatial confidence constraint factor incorporates the topological structure knowledge of the pipeline network into the model, quantifying the information propagation potential between adjacent monitoring nodes and laying a foundation for subsequent multi-node collaborative verification. In particular, the spatio-temporal collaborative constraint factor realizes the non-linear fusion of the time and space dimensions through the representation of a constant curvature space, avoiding the representation bias caused by the weight assignment in a single dimension and ensuring the geometric rationality of the fusion space through the curvature compensation mechanism, making the evolution trajectory of the vibration pattern in the spatio-temporal joint space closer to the propagation characteristics of the real leakage signal. The finally generated first vibration temporal pattern feature encoding vector integrates local temporal details and global context associations, enabling the model to accurately distinguish the subtle differences between real leakage events and interference sources such as environmental noise and equipment start-stop in the time-frequency domain. In this way, the system can fully explore and fuse the complex relationships of pipeline leakage vibration signals in the temporal dynamic development and multi-node spatial structure, enabling the system to carefully distinguish leakage signals from non-leakage transient interferences in a complex noise environment, improving the confidence and robustness of anomaly recognition, thus effectively reducing the false alarm rate and missed detection risk, significantly enhancing the safety guarantee and operation and maintenance efficiency of the pipeline network, and promoting the development of intelligent leak detection technology to a higher level.

[0024] Specifically, first, the time series of the first vibration local time series pattern feature encoding vector is input into a sequence encoder based on a recurrent neural network to obtain the time series of the first vibration sequence transfer initial encoding vector. It should be understood that although one-dimensional convolutional encoding can extract local vibration features within a short-time window, the limited perception field of its convolutional kernel is difficult to capture the persistent time series evolution pattern unique to the leakage signal (such as the attenuation characteristics of the pressure wave propagating along the pipe wall or the gradual process of low-frequency oscillation caused by the expansion of the leakage hole). In the technical solution of this application, the time series of the first vibration local time series pattern feature encoding vector is input into a sequence encoder based on a recurrent neural network to obtain the time series of the first vibration sequence transfer initial encoding vector. Among them, relying on its internal feedback mechanism, the recurrent neural network can break through the limitation of the fixed sliding window and continuously accumulate the hidden state information of the historical time steps during the transfer process, so as to establish a correlation chain between short-time local features and long-term time series context, so as to reweave discrete local feature segments into a dynamically represented sequence with time continuity, making the time series difference between the self-similar vibration pattern unique to the leakage signal during the continuous leakage stage (such as quasi-periodic pressure fluctuations), the transient oscillation caused by sudden interference events, and the system noise be explicitly encoded.

[0025] Specifically, mechanical vibration noise may present a short-time repeated impact waveform, while the vibration caused by real leakage usually forms a low-frequency vibration envelope with phase continuity due to the continuous leakage of fluid. The RNN gradually constructs the dynamic evolution trajectory of the vibration signal by performing recurrent iterative processing on the time series of the feature encoding vector. During this process, the first vibration sequence transfer initial encoding vector output at each time step not only contains the local time series features of that moment, but also implicitly integrates the cumulative vibration state information of the previous time steps. For example, when a small leakage occurs in the pipeline, its vibration signal shows low-energy broadband noise in the initial stage, and as the leakage volume accumulates, an energy enhancement trend in a specific frequency band gradually appears. Through the information transfer between time steps, the RNN can capture this gradual change process of the energy distribution and form a sequence of encoding vectors representing continuous changes in the state, thus laying a high-quality time series feature foundation for the downstream spatio-temporal collaborative analysis. In this way, the sensitivity and discrimination ability of the anomaly detection model to the leakage vibration pattern are significantly improved, the false alarm and missed alarm rates are reduced, and it is ensured that the leak detection system can still stably and accurately monitor pipeline leakage events under noise superposition and variable working conditions, thereby enhancing the operation safety and maintenance efficiency of the pipe network.

[0026] In a specific example of this application, the time series of the first vibration local time series pattern feature encoding vector is input into a sequence encoder based on a recurrent neural network by the following formula to obtain the time series of the first vibration sequence transfer initial encoding vector; where the formula is:

[0027] Where is the time series of the first vibration local timing pattern feature coding vector, are respectively the 1st, 2nd, th, and th first vibration local timing pattern feature coding vectors in the time series of the first vibration local timing pattern feature coding vector, represents a recurrent neural network, are respectively the 1st, 2nd, th, and th first vibration sequence transfer initial coding vectors in the time series of the first vibration sequence transfer initial coding vector.

[0028] Next, calculate the first vibration timing confidence constraint factor of each first vibration sequence transfer initial coding vector in the time series of the first vibration sequence transfer initial coding vector. It should be understood that the vibration caused by a real leak usually has the characteristic of continuous phase coherence (such as the attenuation law when a pressure wave propagates along the pipe wall), while the vibration generated by transient events such as rapid valve operation often exhibits isolated pulse characteristics. However, in the hidden state transfer process of a conventional RNN, the coding vectors at all time steps are given the same important weight, resulting in the model being difficult to suppress the information interference during the sudden noise period. Therefore, in the technical solution of this application, calculate the first vibration timing confidence constraint factor of each first vibration sequence transfer initial coding vector in the time series of the first vibration sequence transfer initial coding vector.

[0029] Here, by calculating the first vibration timing confidence constraint factor, the system can adaptively focus on the continuous vibration pattern period with high information value (such as the quasi-periodic fluctuation section unique to the leak signal), while weakening the influence of the coding vectors corresponding to isolated pulses or random noise. For example, in the initial stage of a small leak in a pipeline, its vibration signal may be intermittently submerged in the equipment operation noise. The timing confidence constraint factor can enhance the information contribution of the weak but continuous leak feature period through dynamic weight adjustment, and suppress the interference of the time steps dominated by random noise on the model decision. This dynamic focus shift ability enables the model to maintain sensitivity to the real leak signal in the face of non-stationary interference, significantly reducing the misjudgment risk caused by local period data contamination, and providing a high signal-to-noise ratio timing feature basis for subsequent spatio-temporal collaborative analysis.

[0030] In a specific example of this application, calculate the first vibration timing confidence constraint factor of each first vibration sequence transfer initial coding vector in the time series of the first vibration sequence transfer initial coding vector with the following formula; where the formula is:

[0031] where, represents function and respectively represent trainable weighted hyperparameters and is a weight matrix represents vector multiplication is the key node weight vector of the first vibration local time series pattern is the time queue of the first vibration time series confidence measurement factor represents function is the first vibration time series confidence constraint factor

[0032] Subsequently, the first vibration space confidence constraint factor of each first vibration sequence transfer initial coding vector in the time series of the first vibration sequence transfer initial coding vector is calculated. It should be understood that the traditional single-point detection model cannot distinguish whether the vibration signal is caused by local noise interference or the propagation of a distal real leakage event because it ignores the spatial topological relationship of the pipe network layout. Therefore, in the technical solution of this application, the first vibration space confidence constraint factor of each first vibration sequence transfer initial coding vector in the time series of the first vibration sequence transfer initial coding vector is calculated. That is, by modeling the structured spatial attributes of the pipeline network, the geographical weights and topological influences of different monitoring nodes during information transmission are quantified. For example, in a pipe network with densely arranged sensors, the monitoring points at the elbows of the main pipelines carry more fluid pressure wave reflection signals, and the initial coding vectors of their vibration sequence transfers are more likely to contain key leakage features compared to the nodes on the straight sections. The calculation of the space confidence constraint factor is essentially an explicit expression of these implicit spatial structure knowledge, enabling the model to automatically identify the key node data with high information propagation potential.

[0033] Specifically, by extracting the feature distribution spatial structure of the first vibration sequence transfer initial coding vector (such as modeling the physical distance between sensor nodes and the connection weights formed by the pipeline orientation through a graph neural network), the system can dynamically evaluate the structural centrality of each node in the message propagation network. For example, when the vibration wave propagates along the pipeline, the nodes with higher centrality are more likely to capture the attenuation propagation path characteristics of the leakage signal, and higher weights will be assigned to the space confidence constraint factors of the corresponding coding vectors. This dynamic weight assignment mechanism effectively strengthens the vibration mode representation ability that conforms to the physical propagation model of the pipeline network, while suppressing the interference of abnormal fluctuations in the coding vectors caused by local random vibrations (such as transient impacts of construction machinery) on the model decision-making.

[0034] It is worth mentioning that since the spatial propagation of the real leakage signal follows a specific acoustic attenuation law (such as exponential energy attenuation with the increase of distance), there will be a characteristic correlation with consistent direction between adjacent monitoring nodes. The spatial confidence constraint factor can guide the model to focus on the vibration feature combinations that conform to the physical law of spatial propagation by quantifying the structural correlation strength between nodes (such as the propagation loss coefficient calculated based on the measurement point spacing and pipeline material). At the same time, for the isolated vibration signals generated by non-leakage events (such as instantaneous pressure shocks near a single sensor), due to the lack of spatial propagation consistency across nodes, the corresponding spatial confidence constraint factor will automatically reduce the propagation weight of the relevant coding vectors, fundamentally avoiding the system-level false alarm problem caused by false triggering of a single sensor. This dynamic modulation mechanism of the spatial dimension confidence enables the system to achieve robust suppression of local noise interference and precise focusing on real leakage events in complex pipeline network scenarios.

[0035] In a specific example of the present application, the first vibration spatial confidence constraint factor of each first vibration sequence transfer initial coding vector in the time series of the first vibration sequence transfer initial coding vector is calculated by the following formula; where the formula is:

[0036] where represents the exponential operation, represents the square of the norm, represents the corresponding first vibration spatial confidence score value, is the first vibration spatial confidence constraint factor.

[0037] Then, the first vibration space confidence constraint factor and the first vibration time series confidence constraint factor are fused to obtain the initial first vibration mode transfer spatio-temporal collaborative constraint factor. It should be understood that real leakage events and interference signals often exhibit similar characteristics in a single dimension (time or space), but there are essential differences between the two in the spatio-temporal joint action domain. For example, although the transient vibration generated by the rapid action of the valve has a sudden characteristic in the time dimension (high time series confidence), its energy propagation is restricted by the pipeline structure and is difficult to spread to adjacent nodes (low space confidence); while the vibration wave caused by real leakage not only has a continuous time evolution law (high time series confidence), but also forms an attenuation characteristic that conforms to the acoustic propagation model along the pipeline topology structure (high space confidence). At this time, simply relying on the independent action of the time series or space confidence constraint factor will cause the model to fall into the single-dimension misjudgment trap. If only the time dimension is concerned, local transient interference may be misjudged as leakage, while simply relying on the space dimension is likely to ignore the early weak leakage signals of isolated nodes. Therefore, in order to construct a collaborative action mechanism for spatio-temporal double-domain confidence, the first vibration space confidence constraint factor and the first vibration time series confidence constraint factor are fused to obtain the initial first vibration mode transfer spatio-temporal collaborative constraint factor.

[0038] That is, by non-linearly fusing the first vibration time series confidence constraint factor (representing the persistence intensity of the vibration mode on the time axis) and the space confidence constraint factor (reflecting the information propagation value of the monitoring node in the pipeline network topology), the system can generate a physically interpretable joint weight factor. For example, a gated fusion mechanism is adopted to dynamically adjust the contribution ratio of spatio-temporal constraints: when the monitoring node is located at a key hub position of the pipeline, the space confidence factor dominates the fusion process, strengthening the vibration characteristics that conform to the pipeline network propagation law; while in the initial weak signal stage of leakage, the weight of the time series confidence factor increases, focusing on the continuously existing low-energy vibration mode. This dynamic collaborative mechanism effectively solves the problem of feature matching misalignment caused by the separation of time series analysis and space analysis in traditional methods, enabling the model to capture the implicit coupling correlation of vibration signals in the spatio-temporal double domain.

[0039] In a specific example of the present application, the first vibration space confidence constraint factor and the first vibration time series confidence constraint factor are fused by the following formula to obtain the initial first vibration mode transfer spatio-temporal collaborative constraint factor; where the formula is:

[0040] Where, represents function, and are fusion weight parameters, is the initial first vibration mode transfer spatio-temporal collaborative constraint factor.

[0041] Furthermore, the initial spatio-temporal collaborative constraint factor for the first vibration mode transmission is optimized by spatio-temporal curvature compensation to obtain the spatio-temporal collaborative constraint factor for the first vibration mode transmission. In particular, it should be understood that when the vibration wave generated by a real leakage event propagates along the pipe wall, it will be affected by factors such as fluid pressure and pipe damping, and its spatio-temporal characteristics show a non-linear attenuation law, while interference signals such as mechanical shocks often have a localized spatio-temporal distribution. When the traditional method directly linearly fuses the temporal confidence constraint factor and the spatial confidence constraint factor, non-Euclidean curvature may occur in the fused spatio-temporal characteristics in the mathematical space due to the independent modeling of the time dimension (such as the quasi-periodicity of the leakage signal) and the spatial dimension (such as the matching of sensor spacing and wave speed). For example, when there are uneven welds in a certain section of the pipeline, the propagation speed of the vibration wave will show local anomalies. At this time, if the spatio-temporal curvature is not compensated, the fused feature vector may mis-characterize the leakage direction or intensity, resulting in the failure of adjacent sensor collaborative verification. Therefore, in a preferred example of the present application, the initial spatio-temporal collaborative constraint factor for the first vibration mode transmission is optimized by spatio-temporal curvature compensation to obtain the spatio-temporal collaborative constraint factor for the first vibration mode transmission.

[0042] That is, a constant curvature space representation and a spherical coordinate approximation mapping are constructed, and the initial spatio-temporal collaborative constraint factor is projected into the manifold space that conforms to the pipeline acoustic propagation model. During this process, by introducing a curvature compensation weight parameter system, the curvature contribution ratio of the time and space dimensions is dynamically adjusted, so that the fused collaborative constraint factor approaches the characteristics of the Euclidean plane space. For example, when vibration signals that conform to the pipeline attenuation model are detected at adjacent monitoring nodes, the spatio-temporal curvature compensation will reduce the curvature weight of the time dimension and enhance the geometric constraint of the spatial propagation path, so that such signals are mapped as straight trajectories with high confidence in the fused space, forming a significant distinction from the distorted trajectories of local noise. After curvature compensation, the system can identify the time synchronization and spatial propagation non-uniformity of such events through the reconstructed plane space, and correct their characteristic trajectories into isolated scattered points. On the contrary, although the low-frequency vibration generated by a real small leakage has weak energy, its cross-node phase delay characteristics are strengthened as a coherent propagation path in the compensated flat space. This geometric correction mechanism enables the model to break through the limitations of traditional threshold algorithms, accurately capture the spatio-temporal evolution pattern of leakage signals that conform to physical laws in a strong noise background, significantly reduce the false alarm rate caused by geometric representation distortion, and improve the reliability of multi-node collaborative verification.

[0043] In this preferred example, the initial spatio-temporal collaborative constraint factor for the first vibration mode transmission is optimized by spatio-temporal curvature compensation to obtain the spatio-temporal collaborative constraint factor for the first vibration mode transmission according to the following formula; where the formula is:

[0044] Where, is a representation of a space with constant curvature, is a representation of the approach in spherical coordinates, is the spatio-temporal co - constraint factor for the transmission of the first vibration mode.

[0045] Specifically, first, based on each first vibration sequence to transmit the initial encoding vector the corresponding first vibration time - series confidence constraint factor and the first vibration space confidence constraint factor , to construct the representation of the space with constant curvature and the representation of the approach in spherical coordinates :

[0046] Then, take the fused initial spatio - temporal co - constraint factor for the transmission of the first vibration mode as the fused space - metric benchmark representation, and determine the weight parameters in the fusion formula by calculating the following formula, for example and :

[0047] That is, when , it can make the single - mode correlation representations and in the time and space dimensions approach the flat space - time coupling attraction, thus reflecting the approach to Euclidean - like in the fused space. That is, by compensating for the generation of single - mode negative curvature to achieve the plane - keeping in the fused space, thereby enhancing the fused expression effect of the initial spatio - temporal co - constraint factor for the transmission of the first vibration mode .

[0048] Furthermore, based on the spatio - temporal co - constraint factor for the transmission of the first vibration mode, perform first - vibration - sequence message - passing structure modulation on each first vibration sequence to transmit the initial encoding vector to obtain a time - series of first - vibration - sequence transmitted structurally - modulated encoding vectors. Considering that when the leakage vibration wave propagates along the pipeline axis, it will form spatio - temporal continuous characteristics with directionality and attenuation law, while local interferences such as mechanical shocks are manifested as spatially isolated and temporally transient waveforms. For example, when there is periodic construction vibration near the pipeline, although its energy may cover the frequency band of the leakage signal, the propagation time delay of its vibration wave between adjacent sensors does not conform to the hydrodynamic model, and there is a lack of quasi - periodic attenuation mode in the time dimension. Traditional methods directly process the features of all time windows with equal weights, resulting in the over - amplification of noise features. Therefore, to achieve the dynamic purification and enhancement of the feature space, in the technical solution of this application, based on the spatio - temporal co - constraint factor for the transmission of the first vibration mode, perform first - vibration - sequence message - passing structure modulation on each first vibration sequence to transmit the initial encoding vector to obtain a time - series of first - vibration - sequence transmitted structurally - modulated encoding vectors.

[0049] Here, the spatio-temporal co-domain collaborative control of the information flow is realized by performing an element-wise multiplication of the spatio-temporal co-constraint factor for the first vibration mode transfer and the initial coding vector for the first vibration sequence transfer. For example, the vibration signals captured by the monitoring nodes at the pipe bends may exhibit high-frequency oscillation characteristics due to the fluid impact effect. However, the spatio-temporal co-constraint factor optimized by curvature compensation can identify the short-term suddenness (low temporal confidence) and locality (low spatial confidence) of such signals, and suppress the transmission of invalid information by reducing its weight. On the contrary, for the low-frequency vibration signals that conform to the pipe sound speed propagation model, their high spatio-temporal co-weights will drive the feature coding to be strengthened along the pipe network topology direction, forming a cross-node phase delay correlation chain. This modulation mechanism essentially constructs a vibration feature propagation path based on physical laws, making the information transmission path form a spatial mapping with the diffusion trajectory of the real leakage signal. This organic propagation mechanism enables the system to achieve deep purification and directional enhancement of the value of vibration feature information while retaining the details of the original vibration signal.

[0050] Subsequently, the position-wise sum of the time series of the structural modulation coding vectors for the first vibration sequence transfer is calculated to obtain the first vibration time-sequence mode feature coding vector. Considering that when the structural modulation coding vectors modulated by the spatio-temporal co-constraint factor still maintain the time series form, the leakage features contained therein may be scattered in different time steps and are difficult to be accurately captured by the subsequent anomaly detection model. Therefore, to overcome the information dilution defect of the traditional sliding window averaging method and concentrate the high-order features optimized in the spatio-temporal co-domain of the vibration signal, in the technical solution of this application, the position-wise sum of the time series of the structural modulation coding vectors for the first vibration sequence transfer is calculated to obtain the first vibration time-sequence mode feature coding vector. In this way, the complexity of the subsequent model processing is reduced, while the overall characteristics and the expression of important information of the vibration signal on the time axis are maintained and strengthened, effectively avoiding the interference of redundancy and noise, and improving the stability and robustness of the features through the integration in the time dimension. In addition, the position-wise sum operation can mitigate the influence of short-term anomalies or occasional noises in the time series data, ensuring that the finally generated vibration time-sequence coding features have strong generalization ability while representing the overall vibration mode. This provides an efficient, compact and information-rich input for the anomaly detection model that integrates information from multiple time points, covers dynamic evolution and structural associations, greatly improving the recognition accuracy and confidence of the model for leakage anomaly events.

[0051] In a specific example of this application, the position-wise sum of the time series of the structural modulation coding vectors for the first vibration sequence transfer is calculated by the following formula to obtain the first vibration time-sequence mode feature coding vector; where the formula is:

[0052] Where, Scale the time series of the structural modulation coding vector transmitted for the first vibration sequence Is the first vibration time series pattern feature coding vector

[0053] Specifically, in S4, the first vibration time series pattern feature coding vector is input into an anomaly detection model to obtain an anomaly detection result. In the technical solution of this application, the first vibration time series pattern feature coding vector is input into an anomaly detection model based on a classifier to obtain an anomaly detection result, and the anomaly detection result is used to indicate whether there is an anomaly in the first leakage vibration monitoring signal. It should be understood that after multi-level signal processing and feature extraction, the obtained first vibration time series pattern feature coding vector represents the deep and comprehensive spatio-temporal dynamic information in the pipeline vibration signal. However, relying solely on this high-dimensional feature vector itself is not sufficient to directly determine whether there is a leakage anomaly in the pipeline. Therefore, in the technical solution of this application, the first vibration time series pattern feature coding vector is input into an anomaly detection model to obtain an anomaly detection result. Among them, the anomaly detection model refers to an intelligent prediction system based on machine learning or deep learning technology, which can learn the distribution law of signal features in the normal vibration state, and then distinguish abnormal vibration behaviors that deviate from the normal mode. This model classifies or scores anomalies for the input vibration time series feature coding vector, and outputs an anomaly detection result for the discovered abnormal mode, such as whether there is a potential leakage, the confidence level of the anomaly degree, etc. In a specific example, an anomaly detection model based on a classifier can be adopted to learn the boundary between normal and abnormal vibration samples to achieve the classification and recognition of abnormal vibration signals.

[0054] Specifically, by inputting the first vibration time series pattern feature coding vector into an anomaly detection model based on a classifier, the system can map complex multi-source spatio-temporal features to clear anomaly discrimination results, not only distinguishing leakage signals from non-leakage events such as environmental noise and equipment operation interference, but also evaluating the confidence level and severity of the anomaly, greatly improving the accuracy and practicality of the monitoring system response. In this way, intelligent and automatic anomaly recognition of vibration characteristics can be achieved, providing a solid basis for the rapid early warning of pipeline leakage events.

[0055] Specifically, in S5, in response to the abnormal detection result being a potential anomaly, the second leakage vibration monitoring signal collected by the leak detector deployed at the second monitoring position of the pipeline is acquired. The first monitoring position is adjacent to the second monitoring position. It should be understood that relying on a single-point vibration signal for anomaly judgment is often interfered by environmental noise and non-leakage events, resulting in frequent false alarms and missed alarms, especially in complex working conditions or strong noise environments. In the technical solution of this application, to address this limitation, a response mechanism is designed. That is, when the anomaly detection model determines that the first vibration time-series pattern feature encoding vector collected at the first monitoring position is a potential anomaly, the system immediately acquires the leakage vibration monitoring signal of the second monitoring point deployed at the adjacent position. That is, by means of the second leakage vibration monitoring signal collected by the adjacent leak detector, multi-point spatio-temporal correlation analysis is carried out, thereby improving the accuracy and credibility of anomaly judgment.

[0056] Specifically, acquiring the vibration monitoring signal collected by the leak detector deployed at the second adjacent monitoring position can provide the system with independent but related leakage information from different spatial positions. Since the vibration signal generated by leakage has certain propagation characteristics and spatial continuity, a true leakage event usually generates a correlated vibration pattern at adjacent monitoring points synchronously or within a short time. On the contrary, local interference or transient non-leakage events are likely to affect only a single point and are unlikely to generate consistent abnormal signals at multiple adjacent monitoring positions. Therefore, by introducing the vibration data of the second monitoring point, the system can effectively distinguish based on the spatio-temporal correlation of the vibration patterns at the two points and play a role in collaborative verification. In this way, not only the deficiency of the single-point anomaly detection model is made up for, making the anomaly determination no longer isolated from the local signal, but also the risk of false alarms and missed alarms is significantly reduced through the multi-node cross-verification mechanism, thus achieving a more accurate identification of abnormal events and a leakage alarm output with a higher confidence level. This not only improves the stability and reliability of the leak detection system in the complex industrial field environment, but also provides more reliable decision-making support for operation and maintenance personnel, greatly optimizing the intelligent level and operation and maintenance cost-effectiveness of on-site monitoring.

[0057] Specifically, in step S6, the vibration time-series pattern features of the second leakage vibration monitoring signal are extracted to obtain the second vibration time-series pattern feature coding vector. That is, similar to the processing method of the first leakage vibration monitoring signal, in the technical solution of this application, first, a band-pass filter is applied to filter the second leakage vibration monitoring signal to obtain the filtered second leakage vibration monitoring signal; then, vibration local time-series pattern feature extraction based on one-dimensional convolutional coding is performed on the filtered second leakage vibration monitoring signal to obtain the time series of the second vibration local time-series pattern feature coding vector; furthermore, vibration pattern time-series context coding is performed on the time series of the second vibration local time-series pattern feature coding vector to obtain the second vibration time-series pattern feature coding vector. In this way, the original second leakage vibration monitoring signal that may contain noise can be converted into a coding vector with important features representing leakage events, ensuring that multi-point signals are expressed in a unified and efficient representation space, facilitating subsequent comparison and spatio-temporal correlation analysis with the signal features at the first monitoring location, and laying an accurate and reliable data foundation for spatio-temporal correlation calculation. It is worth mentioning that the generation of the second vibration time-series pattern feature coding vector not only improves the system's ability to capture signal details but also achieves feature consistency and comparability across monitoring points, providing accurate and stable input for subsequent calculation of the vibration pattern spatio-temporal correlation degree, ultimately helping the system significantly improve the recognition rate of real leakage events and the confidence level of alarms, reducing the false alarm rate, and providing strong technical support for pipeline safety assurance in complex environments.

[0058] Specifically, in step S7, the vibration pattern spatio-temporal correlation degree between the second vibration time-series pattern feature coding vector and the first vibration time-series pattern feature coding vector is calculated. It should be understood that in a complex pipeline monitoring environment, the vibration signals collected by a single-point sensor may be affected by various interference factors, such as mechanical vibration, water flow noise, and transient waveforms generated by non-leakage events. These interferences often highly overlap with real leakage signals in both the frequency domain and the time domain. It is difficult to effectively distinguish true and false anomalies by relying solely on anomaly detection at a single monitoring location. This leads to the traditional leakage detection method being prone to missed detections or false alarms, seriously affecting the reliability and practicality of the system. To address this problem, in the technical solution of this application, the vibration pattern spatio-temporal correlation degree between the second vibration time-series pattern feature coding vector and the first vibration time-series pattern feature coding vector is calculated.

[0059] That is, by comparing the vibration signal feature codes collected at adjacent monitoring points, the similarity and synchronization in the timing pattern between the two signals are revealed. Utilizing the fact that the vibration signals caused by leakage events have spatial propagation continuity, the confidence level of anomaly determination is enhanced. Among them, compared with potential anomaly samples, the signals caused by real leakage often exhibit highly correlated waveform features and timing structures between adjacent nodes, while local environmental interference or random noise is difficult to synchronously appear at multiple spatial positions. Therefore, by calculating and quantifying the spatio-temporal correlation degree between the second vibration timing pattern feature code vector and the first vibration timing pattern feature code vector, the system can effectively filter out isolated anomalies and false signals, and achieve precise identification of real leakage events. In a specific example of this application, the spatio-temporal correlation degree of vibration modes can be obtained by performing a spatio-temporal correlation difference calculation on the second vibration timing pattern feature code vector and the first vibration timing pattern feature code vector. In this way, the internal connection between the signals at the first monitoring position and the second monitoring position is established, realizing in-depth excavation and intelligent identification of weak leakage signals in a complex environment, and laying a technical foundation for the practicality and reliability of the leak detector data intelligent analysis system.

[0060] Specifically, in step S8, based on the comparison between the spatio-temporal correlation degree of vibration modes and a preset threshold, it is determined whether to adjust the anomaly detection result to a high-confidence leakage event. In a specific example of this application, when the spatio-temporal correlation degree exceeds the threshold, the system will automatically adjust the "potential anomaly" result originally judged based on single-point data to a "high-confidence leakage event", that is, confirm that the anomaly signal has continuity and consistency in space, greatly excluding the possibility of single-point accidental noise or interference; conversely, if the correlation degree does not reach the preset standard, the cautious judgment of the anomaly is maintained to prevent false alarms. This threshold comparison and result adjustment mechanism based on spatio-temporal correlation degree effectively guarantees the rapid and accurate response of the leak detector data intelligent analysis system to pipeline leakage events, improves the accuracy of anomaly detection and the credibility of alarms, and thus solves the misjudgment problem of traditional leak detection systems due to the lack of multi-node spatio-temporal correlation verification.

[0061] In summary, the intelligent analysis method for leak detector data according to the embodiments of the present application is elucidated. By using a neural network model based on deep learning, it extracts the time-series pattern features of the vibration signal at the first monitoring position to discover the latent time-series features of the vibration signal, and based on this, preliminarily identifies potential leakage anomalies. For this abnormal signal, the system further obtains the vibration signals of adjacent monitoring points. Similarly, it extracts the time-series pattern coding features of the vibration signals of adjacent monitoring points, then calculates the spatio-temporal correlation degree of vibration modes between the two, and further, by comparing the correlation degree with a preset threshold, realizes the multi-dimensional verification of leakage abnormal events. In this way, it breaks through the limitation of traditional single-point static threshold detection, effectively reduces the influence of environmental noise and non-leakage transient interference, improves the recognition accuracy of leakage signals, and thus enhances the reliability of operation and maintenance decisions.

[0062] Furthermore, an intelligent analysis system for leak detector data is also provided.

[0063] Figure 4 FIG. is a block diagram of an intelligent analysis system for leak detector data according to the embodiments of the present application. As Figure 4 shown, the intelligent analysis system 300 for leak detector data according to the embodiments of the present application includes: a first leakage vibration monitoring signal acquisition module 310, configured to acquire a first leakage vibration monitoring signal collected by a leak detector deployed at a first monitoring position of a pipeline; a filtering processing module 320, configured to filter the first leakage vibration monitoring signal by applying a band-pass filter to obtain a filtered first leakage vibration monitoring signal; a first vibration time-series pattern feature extraction module 330, configured to extract the vibration time-series pattern features of the filtered first leakage vibration monitoring signal to obtain a first vibration time-series pattern feature coding vector; an anomaly detection module 340, configured to input the first vibration time-series pattern feature coding vector into an anomaly detection model to obtain an anomaly detection result; a second leakage vibration monitoring signal acquisition module 350, configured to, in response to the anomaly detection result being a potential anomaly, acquire a second leakage vibration monitoring signal collected by a leak detector deployed at a second monitoring position of the pipeline, where the first monitoring position and the second monitoring position are adjacent; a second vibration time-series pattern feature extraction module 360, configured to extract the vibration time-series pattern features of the second leakage vibration monitoring signal to obtain a second vibration time-series pattern feature coding vector; a vibration mode spatio-temporal correlation coding module 370, configured to calculate the spatio-temporal correlation degree of vibration modes between the second vibration time-series pattern feature coding vector and the first vibration time-series pattern feature coding vector; and a leakage result analysis module 380, configured to determine whether to adjust the anomaly detection result to a high-confidence leakage event based on the comparison between the spatio-temporal correlation degree of vibration modes and a preset threshold.

[0064] As described above, the leak detector data intelligent analysis system 300 according to the embodiments of the present application can be implemented in various wireless terminals, such as a server with a leak detector data intelligent analysis algorithm. In a possible implementation manner, the leak detector data intelligent analysis system 300 according to the embodiments of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the leak detector data intelligent analysis system 300 can be a software module in the operating system of the wireless terminal, or can be an application program developed for the wireless terminal; of course, the leak detector data intelligent analysis system 300 can also be one of the many hardware modules of the wireless terminal.

[0065] Alternatively, in another example, the leak detector data intelligent analysis system 300 and the wireless terminal can also be separate devices, and the leak detector data intelligent analysis system 300 can be connected to the wireless terminal through a wired and / or wireless network and transmit interaction information in accordance with a predefined data format.

[0066] The embodiments of the present disclosure have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, the practical application, or the improvement of the technology in the market, or to enable other ordinary skill in the art in the technical field to understand the embodiments disclosed herein.

Claims

1. A leak detector data intelligent analysis method, characterized in that: include: Acquire a first leakage vibration monitoring signal collected by a leak detector deployed at a first monitoring position of the pipeline; Applying a bandpass filter to filter the first leakage vibration monitoring signal to obtain a filtered first leakage vibration monitoring signal; Extracting the vibration time series pattern feature of the filtered first leakage vibration monitoring signal to obtain a first vibration time series pattern feature encoding vector; Inputting the first vibration time series pattern feature encoding vector into an anomaly detection model to obtain an anomaly detection result; In response to the abnormality detection result being a potential abnormality, obtaining a second leakage vibration monitoring signal collected by a leak detector deployed at a second monitoring position of the pipeline, the first monitoring position being adjacent to the second monitoring position; Extracting the vibration time series pattern feature of the second leakage vibration monitoring signal to obtain a second vibration time series pattern feature encoding vector; Calculating the spatiotemporal correlation of the vibration pattern between the second vibration time series pattern feature coding vector and the first vibration time series pattern feature coding vector; Based on the comparison between the spatiotemporal correlation of the vibration pattern and a preset threshold, it is determined whether to adjust the abnormal detection result to a high-confidence leakage event.

2. The method for intelligent analysis of leak detector data according to claim 1, characterized in that: Extracting the vibration time series pattern feature of the filtered first leakage vibration monitoring signal to obtain a first vibration time series pattern feature encoding vector includes: Extracting the vibration local time series pattern feature of the filtered first leakage vibration monitoring signal to obtain a time series of the first vibration local time series pattern feature encoding vector; Vibration pattern temporal context encoding is performed on the time series of the first vibration local temporal pattern feature encoding vector to obtain a first vibration temporal pattern feature encoding vector.

3. The method for intelligent analysis of leak detector data according to claim 2, characterized in that: Extracting the vibration local time series pattern feature of the filtered first leakage vibration monitoring signal to obtain a time series of the first vibration local time series pattern feature encoding vector, including: The first leakage vibration monitoring signal after filtering is subjected to vibration local time series pattern feature extraction based on one-dimensional convolution coding to obtain a time series of the first vibration local time series pattern feature coding vector.

4. The method for intelligent analysis of leak detector data according to claim 2, characterized in that: Performing vibration pattern temporal context encoding on the time series of the first vibration local temporal pattern feature encoding vector to obtain the first vibration temporal pattern feature encoding vector, comprising: Calculating the first vibration mode spatiotemporal synergy confidence constraint factor of the time series of the first vibration local temporal pattern feature encoding vector to construct the first vibration mode transmission spatiotemporal synergy constraint factor; Based on the first vibration mode spatiotemporal synergy constraint factor, a first vibration sequence transfer aggregation analysis is performed on the time series of the first vibration local time sequence mode feature coding vector to obtain the first vibration time sequence mode feature coding vector.

5. The method for intelligent analysis of leak detector data according to claim 4, characterized in that: Calculating the first vibration mode spatiotemporal synergy confidence constraint factor of the time series of the first vibration local temporal pattern feature encoding vector to construct the first vibration mode spatiotemporal synergy constraint factor, including: Inputting the time sequence of the first vibration local temporal pattern feature encoding vector into a sequence encoder based on a recurrent neural network to obtain the time sequence of the first vibration sequence transmission initial encoding vector; Calculating the first vibration timing confidence constraint factor of each first vibration sequence transmitting the initial coding vector in the time series of the first vibration sequence transmitting the initial coding vector; Calculating the first vibration space confidence constraint factor of each first vibration sequence transmitting the initial coding vector in the time series of the first vibration sequence transmitting the initial coding vector; Based on the first vibration space confidence constraint factor and the first vibration time sequence confidence constraint factor of each first vibration sequence transmitting initial coding vector, the first vibration mode transmission space-time coordination constraint factor of each first vibration sequence transmitting initial coding vector is constructed.

6. The method for intelligent analysis of leak detector data according to claim 5, characterized in that: Based on the first vibration space confidence constraint factor and the first vibration time sequence confidence constraint factor of each first vibration sequence transmitting initial coding vector, the first vibration mode transmission time and space coordination constraint factor of each first vibration sequence transmitting initial coding vector is constructed, including: The first vibration space confidence constraint factor and the first vibration time sequence confidence constraint factor are merged to obtain the initial first vibration mode transmission time and space coordination constraint factor; The initial first vibration mode transmission space-time coordination constraint factor is optimized by space-time curvature compensation to obtain the first vibration mode transmission space-time coordination constraint factor.

7. The method for intelligent analysis of leak detector data according to claim 6, characterized in that: Based on the first vibration mode spatiotemporal synergy constraint factor, a first vibration sequence transfer aggregation analysis is performed on the time series of the first vibration local time sequence mode feature coding vector to obtain the first vibration time sequence mode feature coding vector, including: Based on the first vibration mode transmission spatiotemporal synergy constraint factor, each first vibration sequence transmission initial coding vector is subjected to first vibration sequence message transmission structural modulation to obtain a time sequence of the first vibration sequence transmission structural modulation coding vector; The position-wise sum of the time series of the structural modulation coding vector transmitted by the first vibration sequence is calculated to obtain the first vibration timing pattern feature coding vector.

8. The method for intelligent analysis of leak detector data according to claim 7, characterized in that: Inputting the first vibration time series pattern feature encoding vector into the anomaly detection model to obtain an anomaly detection result includes: The first vibration time series pattern feature encoding vector is input into a classifier-based anomaly detection model to obtain an anomaly detection result, and the anomaly detection result is used to indicate whether the first leakage vibration monitoring signal has an anomaly.

9. The method for intelligent analysis of leak detector data according to claim 8, characterized in that: Calculating the spatiotemporal correlation of the vibration mode between the second vibration time sequence pattern feature coding vector and the first vibration time sequence pattern feature coding vector, comprising: The second vibration time sequence pattern feature coding vector and the first vibration time sequence pattern feature coding vector are subjected to vibration mode spatiotemporal correlation difference calculation to obtain the vibration mode spatiotemporal correlation degree.

10. A leak detector data intelligent analysis system, characterized in that: include: A first leakage vibration monitoring signal acquisition module, used to acquire a first leakage vibration monitoring signal collected by a leak detector deployed at a first monitoring position of the pipeline; A filtering processing module, used for applying a bandpass filter to filter the first leakage vibration monitoring signal to obtain a filtered first leakage vibration monitoring signal; A first vibration time series pattern feature extraction module, used to extract the vibration time series pattern feature of the filtered first leakage vibration monitoring signal to obtain a first vibration time series pattern feature encoding vector; An anomaly detection module, used for inputting the first vibration time series pattern feature encoding vector into an anomaly detection model to obtain an anomaly detection result; A second leakage vibration monitoring signal acquisition module is used to acquire a second leakage vibration monitoring signal collected by a leak detector deployed at a second monitoring position of the pipeline in response to the abnormality detection result being a potential abnormality, the first monitoring position being adjacent to the second monitoring position; A second vibration time series pattern feature extraction module, used to extract the vibration time series pattern feature of the second leakage vibration monitoring signal to obtain a second vibration time series pattern feature encoding vector; A vibration pattern spatiotemporal correlation coding module, used to calculate the spatiotemporal correlation degree of the vibration pattern between the second vibration time sequence pattern feature coding vector and the first vibration time sequence pattern feature coding vector; The leakage result analysis module is used to determine whether to adjust the abnormal detection result to a high-confidence leakage event based on the comparison between the spatiotemporal correlation of the vibration mode and a preset threshold.

Images