Oil and gas pipeline valve internal leakage online diagnosis method based on voiceprint analysis

Abstract

The present application relates to the technical field of oil and gas pipeline valve internal leakage online diagnosis, and discloses an oil and gas pipeline valve internal leakage online diagnosis method based on voiceprint analysis. The method acquires valve operation voiceprint data through a distributed interface to construct a feature database, designs a diagnosis framework according to the database, and determines a multi-source voiceprint collection scheme. A voiceprint collection terminal acquires multi-dimensional voiceprint signals according to the scheme, inputs a model after feature extraction to complete internal leakage level determination and abnormal pattern recognition, and outputs diagnosis results. The diagnosis results are transmitted to a management optimization module to realize diagnosis strategy adjustment through timestamp calibration, mapping table construction and the like. A three-level evaluation sub-framework is also constructed to assist diagnosis. The method realizes accurate diagnosis and strategy optimization of oil and gas pipeline valve internal leakage, and improves diagnosis management level.

Description

Technical Field

[0001] This invention relates to the field of online diagnosis technology for internal leakage of oil and gas pipeline valves, specifically to an online diagnosis method for internal leakage of oil and gas pipeline valves based on acoustic fingerprint analysis. Background Technology

[0002] In the oil and gas transportation industry, the normal operation of oil and gas pipeline valves plays a decisive role in the safety and efficiency of the entire transportation system. Internal leakage in valves not only wastes oil and gas resources but can also trigger serious safety accidents such as fires and explosions, threatening human lives and the ecological environment. Therefore, accurate and timely diagnosis of valve leakage is crucial.

[0003] Traditional methods for diagnosing internal leaks in oil and gas pipeline valves have several limitations. Some methods rely on inspectors periodically visiting the site to assess valve condition through listening, visual inspection, and touch. This approach is heavily influenced by the inspectors' experience and subjective judgment, making it difficult to detect early, minute leaks. Furthermore, manual inspections are inefficient, costly, and cannot achieve real-time monitoring. Other methods employ physical parameter monitoring, such as pressure and flow rate monitoring, to determine the presence of leaks by analyzing changes in pipeline pressure and flow. However, during oil and gas transportation, these physical parameters are easily affected by variations in transport volume, ambient temperature fluctuations, and other factors, leading to inaccurate results and instances of misdiagnosis or missed detection.

[0004] With technological advancements, sensor-based monitoring technologies have been gradually applied, but most focus on single-parameter monitoring, failing to comprehensively reflect the internal leakage status of valves. Furthermore, these technologies have limited data processing and analysis capabilities, hindering the full extraction of data value and making accurate diagnosis and effective management of valve internal leakage difficult. Simultaneously, existing diagnostic methods lack dynamic adjustment mechanisms for diagnostic strategies, failing to optimize diagnostic processes and resource allocation in a timely manner based on actual diagnostic conditions, thus failing to meet the growing demand for internal leakage diagnosis in oil and gas pipeline valves. Therefore, a more efficient, accurate, and intelligent online diagnostic method for internal leakage in oil and gas pipeline valves is urgently needed. Summary of the Invention

[0005] The purpose of this invention is to provide an online diagnostic method for internal leakage in oil and gas pipeline valves based on acoustic signature analysis, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this invention provides an online diagnostic method for internal leakage in oil and gas pipeline valves based on acoustic signature analysis, the method comprising:

[0007] Acquire acoustic fingerprint data of oil and gas pipeline valves under operating conditions, and construct an acoustic fingerprint feature database based on the acoustic fingerprint data;

[0008] Based on the aforementioned voiceprint feature database, an internal leakage diagnostic framework is designed, and a multi-source voiceprint acquisition scheme is determined by combining the diagnostic framework with the voiceprint feature database.

[0009] The acoustic signature acquisition terminal acquires the valve's multidimensional acoustic signature signal according to the acquisition scheme, performs internal leakage level determination and abnormal pattern recognition based on the multidimensional acoustic signature signal, and outputs the diagnostic results.

[0010] The diagnostic results are transmitted to the management optimization module for adjustment of diagnostic strategies.

[0011] Preferably, the step of acquiring the acoustic fingerprint data of the oil and gas pipeline valves under operating conditions and constructing an acoustic fingerprint feature database based on the acoustic fingerprint data specifically involves:

[0012] By connecting the valve body sensor and the pipeline auxiliary microphone array through a distributed interface, the time-domain waveform information of the valve operation sound pattern and the frequency domain distribution information of the steady-state sound pattern are obtained.

[0013] The time-domain waveform information of the valve operation soundprint is processed by frame-by-frame windowing using time-frequency conversion technology, and the background noise is filtered by the frequency domain distribution information of the steady-state soundprint using the energy threshold method to obtain the denoised soundprint feature unit.

[0014] A voiceprint feature mapping network is constructed based on feature association rules. The denoised voiceprint feature units are input into the mapping network for temporal association analysis to extract core feature dimensions that are strongly correlated with the leakage state.

[0015] The core feature dimensions are standardized and labeled with corresponding internal leakage status tags. Features are aggregated according to valve type and operating period to construct a voiceprint feature database containing time domain features, frequency domain features and equipment features.

[0016] Preferably, the step of designing an internal leakage diagnostic framework based on the voiceprint feature database, and determining a multi-source voiceprint acquisition scheme by combining the diagnostic framework with the voiceprint feature database, specifically includes:

[0017] The voiceprint feature database is decomposed into dimensions to extract three core evaluation dimensions: voiceprint energy distribution, frequency characteristic offset, and time-varying fluctuation. Initial weight parameters are assigned to each dimension based on expert experience.

[0018] The core evaluation dimensions are topologically mapped to the spatial structure hierarchy of oil and gas pipelines, and a three-level evaluation sub-framework is constructed, including the valve body layer, pipe section layer, and system layer. The synergistic influence weights between the sub-frameworks are calculated through feature cross-validation.

[0019] The core feature dimensions in the voiceprint feature database are matched with the three-level evaluation sub-framework to check the degree of matching, and the feature coverage missing items corresponding to each sub-framework are screened out.

[0020] Based on the missing feature coverage items, the types of acoustic signals that need to be collected are determined. Taking into account the sensor deployment density and data transmission bandwidth limitations in the pipeline environment, the collection frequency and deployment location of each type of signal are calculated.

[0021] Based on the decision tree algorithm, the acquisition frequency, deployment location and feature coverage missing items are comprehensively judged to generate a multi-source voiceprint acquisition scheme that includes signal type, acquisition frequency and deployment location.

[0022] Preferably, the step of using a decision tree algorithm to comprehensively determine the acquisition frequency, deployment location, and missing feature coverage items to generate a multi-source speaker acquisition scheme that includes signal type, acquisition frequency, and deployment location is as follows:

[0023] The hardware performance indicators of each acoustic signature acquisition terminal in the pipeline environment are obtained, including signal sensitivity, sampling rate, and battery life.

[0024] Construct a set of data acquisition constraints, which includes constraints on the maximum sampling rate of a single terminal, the maximum storage capacity of a single terminal, and the total transmission bandwidth of the system.

[0025] Using the collection frequency, deployment location, and missing feature coverage as decision factors, a correlation function between each decision factor and the set of collection constraints is established.

[0026] By fusing the importance scores of each decision factor through the decision tree splitting rule, the feasibility scores of each signal type at different acquisition frequencies are obtained.

[0027] The sampling frequency combinations with a feasibility score higher than a set threshold are selected, and the sensor resources for each signal type are dynamically configured according to the deployment location to generate a multi-source acoustic signature acquisition scheme.

[0028] Preferably, the acoustic signature acquisition terminal acquires the valve's multi-dimensional acoustic signature signal according to the acquisition scheme, performs internal leakage level determination and abnormal pattern recognition based on the multi-dimensional acoustic signature signal, and outputs diagnostic results, specifically as follows:

[0029] According to the acquisition frequency and signal type requirements of the acquisition scheme, the acoustic fingerprint acquisition terminal synchronously acquires the vibration acoustic fingerprint data of the valve body, the flow acoustic fingerprint data of the pipeline fluid, and the interference acoustic fingerprint data of the environmental background.

[0030] The vibration acoustic data is extracted in the time domain using the short-time zero-crossing rate algorithm, the flow acoustic data is extracted in the frequency domain using the Mel-Cepstral Coefficient method, and the noise features of the interference acoustic data are extracted using the envelope analysis method to obtain a fused feature vector.

[0031] The fused feature vector is input into a pre-trained voiceprint classification model, which is based on a feedforward neural network to construct an internal leakage level determination layer and a recurrent neural network to construct an abnormal pattern recognition layer.

[0032] The internal leakage level determination layer outputs the internal leakage level classification result, and the abnormal feature matching result is output through the abnormal pattern recognition layer. The classification result and the matching result are then logically ANDed to output the diagnostic result.

[0033] Preferably, the step of transmitting the diagnostic results to the management optimization module for diagnostic strategy adjustment specifically involves:

[0034] The diagnostic results are encapsulated based on the data transmission protocol to generate structured feedback information including internal leak level, abnormal mode, and location of occurrence.

[0035] The structured feedback information is transmitted to the storage database of the management and optimization module through a data transmission middleware, and the historical strategy data and current feedback information in the storage database are timestamped.

[0036] Construct a strategy adjustment mapping table, which includes the correspondence between internal leakage level and inspection priority, the correspondence between abnormal mode and repair process, and the correspondence between occurrence location and resource allocation;

[0037] Based on the mapping table, the parameters of the historical strategy data are adjusted to generate an updated diagnostic strategy.

[0038] Preferably, the step of transmitting the structured feedback information to the storage database of the management optimization module via a data transmission middleware, and performing a timestamp calibration operation on the historical strategy data and current feedback information in the storage database, specifically involves:

[0039] The timestamp distribution characteristics of historical strategy data stored in the management optimization module's database are obtained, and the average interval of the timestamps is calculated as the benchmark calibration period.

[0040] The timestamp of the structured feedback information is rounded down according to the reference calibration period to obtain the calibration timestamp;

[0041] The time calibration value of the historical strategy data is calculated by performing an exponentially weighted average on the data whose timestamps are within one period before and after the calibration timestamp.

[0042] The numerical data of the structured feedback information is nonlinearly interpolated with the time calibration value to obtain the time-calibrated feedback data sequence.

[0043] Preferably, the construction strategy adjustment mapping table includes the correspondence between internal leakage level and inspection priority, the correspondence between anomaly mode and repair process, and the correspondence between occurrence location and resource allocation, specifically:

[0044] Correlation analysis was performed on the feedback data sequence after time calibration to calculate the correlation coefficients between internal leak level and inspection priority, abnormal mode and repair process, and occurrence location and resource allocation.

[0045] Select variable pairs whose absolute correlation coefficient is greater than a set threshold as the core correlation items in the mapping table;

[0046] The core related items are divided into intervals for statistical analysis to determine the range of values ​​of another variable corresponding to each variable interval, and a strategy adjustment mapping table containing discrete interval mapping relationships is constructed.

[0047] Preferably, the step of adjusting the parameters of historical strategy data according to the mapping table to generate an updated diagnostic strategy specifically involves:

[0048] Adjust the inspection priority parameter in the historical strategy data according to the mapping relationship between internal leakage level and inspection priority;

[0049] For the repair process parameters in the historical strategy data, replace the matching standard process according to the mapping relationship between the anomaly pattern and the repair process;

[0050] Based on the mapping relationship between the location of occurrence and resource allocation, the resource allocation ratio is adjusted for the resource allocation parameters in the historical strategy data.

[0051] The adjusted inspection priorities, repair processes, and resource allocation parameters are verified for consistency, and an updated diagnostic strategy is generated.

[0052] Preferably, the construction of the three-level evaluation sub-framework comprising the valve body layer, pipe section layer, and system layer is specifically as follows:

[0053] The valve body layer evaluation subframe includes three evaluation metrics for a single valve: acoustic energy value, frequency offset, and duration.

[0054] The pipe segment evaluation subframe includes three evaluation indicators: acoustic attenuation of associated pipe segments, impact length value, and number of affected nodes.

[0055] The system-level evaluation sub-framework includes three evaluation metrics: global voiceprint coverage, redundancy detection rate, and recovery timeliness.

[0056] The comprehensive score of each level of the evaluation subframe is calculated based on factor analysis, and a three-level evaluation subframe including valve body layer, pipe section layer and system layer is constructed by linear weighted summation.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] This invention presents an online diagnostic method for internal leaks in oil and gas pipeline valves based on acoustic signature analysis. It constructs an acoustic signature feature database by acquiring acoustic signature data during valve operation, and utilizes a distributed interface to connect with various devices to obtain time-domain and frequency-domain information. Through denoising and feature extraction, the method comprehensively and accurately integrates the valve operation acoustic signature data, providing a rich data foundation for subsequent diagnostics. When determining the multi-source acoustic signature acquisition scheme, it comprehensively considers hardware performance, acquisition constraints, and feature requirements, and dynamically configures acquisition parameters based on a decision tree algorithm. This ensures that the acquisition scheme closely matches the actual pipeline environment, guaranteeing the representativeness and effectiveness of the acquired acoustic signature data.

[0059] The voiceprint acquisition terminal synchronously acquires multi-dimensional voiceprint signals, employs various algorithms to extract features, and inputs these features into a pre-trained model for internal leak level determination and abnormal pattern recognition. Analyzing the voiceprint signals from multiple perspectives improves the accuracy and comprehensiveness of the diagnosis. After the diagnostic results are transmitted to the management optimization module, the diagnostic strategy is dynamically adjusted through operations such as timestamp calibration and the construction of a strategy adjustment mapping table. Inspection priorities are adjusted based on the internal leak level, repair processes are matched according to abnormal patterns, and resources are allocated according to the location of occurrence. This makes the diagnostic strategy more aligned with actual needs, improving the efficiency and management level of internal leak diagnosis in oil and gas pipeline valves, and reducing resource waste and safety hazards.

[0060] In addition, the constructed three-level assessment sub-framework conducts multi-level assessments of valve internal leakage from the valve body layer, pipe section layer, and system layer. Different assessment indicators are set for each level to comprehensively reflect the impact of valve internal leakage at different levels, providing a more comprehensive reference for diagnosis and decision-making, and helping to achieve systematic and scientific management of valve internal leakage diagnosis in oil and gas pipelines. Attached Figure Description

[0061] Figure 1 This is a schematic diagram illustrating the working principle of the online diagnostic method for internal leakage in oil and gas pipeline valves based on acoustic signature analysis as described in this invention.

[0062] Figure 2 A schematic diagram illustrating the working principle of building a voiceprint feature database.

[0063] Figure 3 A flowchart for generating a multi-source voiceprint acquisition scheme based on the decision tree algorithm;

[0064] Figure 4 A flowchart outlining the workflow for transmitting diagnostic results and adjusting diagnostic strategies. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Please see Figures 1-4 The online diagnostic method for internal leakage of oil and gas pipeline valves based on acoustic fingerprint analysis of the present invention is specifically implemented as follows:

[0067] The process involves acquiring acoustic fingerprint data from oil and gas pipeline valves during operation and constructing an acoustic fingerprint feature database based on this data. A distributed interface is used to connect with valve body sensors and pipeline-attached microphone arrays to acquire the time-domain waveform information of valve operation acoustic fingerprints and the frequency-domain distribution information of steady-state acoustic fingerprints. Time-frequency conversion technology is used to perform frame-by-frame windowing processing on the time-domain waveform information, and the frequency-domain distribution information is filtered for background noise using an energy thresholding method to obtain denoised acoustic fingerprint feature units. An acoustic fingerprint feature mapping network is constructed based on feature association rules. The denoised acoustic fingerprint feature units are input into the network for temporal correlation analysis to extract core feature dimensions strongly correlated with internal leakage status. These core feature dimensions are then standardized and labeled with corresponding internal leakage status tags. Feature aggregation is performed according to valve type and operating period to complete the construction of the acoustic fingerprint feature database.

[0068] Based on the constructed voiceprint feature database, an internal leakage diagnosis framework was designed. A multi-source voiceprint acquisition scheme was determined by combining the diagnosis framework with the voiceprint feature database. The voiceprint feature database was dimensionally decomposed, extracting three core evaluation dimensions: voiceprint energy distribution, frequency characteristic shift, and time-varying fluctuation. Initial weight parameters were assigned to each dimension based on expert experience. The core evaluation dimensions were topologically mapped to the spatial structure hierarchy of oil and gas pipelines, constructing a three-level evaluation sub-framework including valve body layer, pipe section layer, and system layer. The synergistic influence weights between each sub-framework were calculated through feature cross-validation. The matching degree of the core feature dimensions in the voiceprint feature database with the three-level evaluation sub-framework was checked, identifying missing feature coverage items for each sub-framework. Based on the missing items, the types of voiceprint signals that need to be supplemented for acquisition were determined. Considering the sensor deployment density and data transmission bandwidth limitations in the pipeline environment, the acquisition frequency and deployment location of each type of signal were calculated. Finally, a decision tree algorithm was used to comprehensively determine the acquisition frequency, deployment location, and missing feature coverage items, generating a multi-source voiceprint acquisition scheme.

[0069] Then, the acoustic signature acquisition terminal acquires multi-dimensional acoustic signature signals from the valve according to the acquisition scheme. Based on these signals, it performs internal leakage level determination and abnormal pattern recognition, outputting diagnostic results. The terminal simultaneously acquires vibration acoustic signature data of the valve body, flow acoustic signature data of the pipeline fluid, and interference acoustic signature data from the environmental background, as required by the acquisition scheme. The short-time zero-crossing rate algorithm is used to extract temporal features from the vibration acoustic signature data, the Mel-frequency cepstral coefficient method is used to extract frequency features from the flow acoustic signature data, and envelope analysis is used to extract noise features from the interference acoustic signature data, resulting in a fused feature vector. This fused feature vector is input into a pre-trained acoustic signature classification model. This model uses a feedforward neural network to construct an internal leakage level determination layer and a recurrent neural network to construct an abnormal pattern recognition layer. The internal leakage level determination layer outputs the internal leakage level classification result, and the abnormal feature recognition layer outputs the abnormal feature matching result. The classification result and the matching result are then logically ANDed to output the diagnostic result.

[0070] Finally, the diagnostic results are transmitted to the management and optimization module for diagnostic strategy adjustment. Based on the data transmission protocol, the diagnostic results are encapsulated to generate structured feedback information containing internal leak level, anomaly mode, and location of occurrence. This structured feedback information is transmitted to the management and optimization module's storage database via data transmission middleware. The historical strategy data in the storage database is then timestamped with the current feedback information. A strategy adjustment mapping table is constructed, containing the correspondence between internal leak level and inspection priority, anomaly mode and remediation process, and location of occurrence and resource allocation. Based on this mapping table, parameters are adjusted for historical strategy data to generate an updated diagnostic strategy.

[0071] Example 1:

[0072] In determining the multi-source acoustic signature acquisition scheme, the hardware performance indicators of each acoustic signature acquisition terminal in the pipeline environment were first obtained. These indicators cover aspects such as signal sensitivity, sampling rate, and battery life. Signal sensitivity determines the acquisition terminal's ability to capture weak acoustic signature signals, and different models of acquisition terminals vary in this indicator. The sampling rate affects the amount of data acquired per unit time; a higher sampling rate can acquire richer data, but it also places higher demands on the terminal's computing and storage capabilities. Battery life relates to the continuous working time of the acquisition terminal, affecting the continuity of data acquisition.

[0073] Subsequently, a set of acquisition constraints was constructed. The maximum sampling rate constraint for a single terminal is set based on the computing and data processing capabilities of the acquisition terminal hardware. If the sampling rate is too high, the acquisition terminal may be unable to process and store the acquired data in a timely manner, leading to data loss or terminal lag. The maximum storage capacity constraint for a single terminal takes into account the capacity limitations of the acquisition terminal's own storage device, preventing the acquired data from exceeding its storage capacity and avoiding data acquisition interruptions due to insufficient storage space. The total system transmission bandwidth constraint ensures the stable operation of the entire data transmission system. In a pipelined environment, multiple acquisition terminals transmit data simultaneously; if the total transmission bandwidth is insufficient, congestion, delays, or even data loss may occur during data transmission.

[0074] Using acquisition frequency, deployment location, and missing feature coverage as decision factors, correlation functions were established between each decision factor and the set of acquisition constraints. Acquisition frequency is closely related to the maximum sampling rate constraint of a single terminal; a higher acquisition frequency means an increase in the amount of data acquired per unit time. If the maximum sampling rate of a single terminal is exceeded, this constraint will be violated. Simultaneously, acquisition frequency also affects the amount of data transmitted, which is related to the total system bandwidth constraint. Determining the deployment location affects the quality and integrity of the voiceprint signal acquired by the acquisition terminal. Different deployment locations will result in different levels of interference for the acquisition terminal, affecting signal sensitivity. Furthermore, the deployment location is related to the power supply and data transmission of the acquisition terminal, and is associated with battery life and the total system bandwidth. Missing feature coverage reflects key voiceprint features not acquired under the current acquisition scheme, determining the types of voiceprint signals that need to be supplemented for acquisition. This decision factor interacts with acquisition frequency and deployment location, as different types of voiceprint signals may require different acquisition frequencies and suitable deployment locations for effective acquisition.

[0075] The importance scores of each decision factor are fused using decision tree splitting rules. These rules assess and categorize the importance of three decision factors—acquisition frequency, deployment location, and missing feature coverage—according to specific logic and standards. The assessment process comprehensively considers the constraints of the acquisition limit set, as well as the requirements of different oil and gas pipeline environments and valve characteristics on each decision factor. For example, in some complex pipeline environments, missing feature coverage may be more important because obtaining comprehensive and accurate acoustic signature features is crucial for internal leak diagnosis; while under limited hardware performance, the correlation between acquisition frequency and the maximum sampling rate constraint per terminal may be higher, requiring priority consideration of the impact of acquisition frequency on hardware. This method yields the feasibility score for each signal type at different acquisition frequencies. The feasibility score is derived from the comprehensive relationship between each decision factor and the acquisition limit set, reflecting the feasibility and rationality of acquiring a certain signal type at a specific acquisition frequency.

[0076] Finally, sampling frequency combinations with feasibility scores higher than a set threshold were selected, and sensor resources for each signal type were dynamically configured based on deployment location. The set threshold is a standard value determined based on actual conditions and experience; sampling frequency combinations exceeding this threshold are considered more feasible solutions under current conditions. Dynamic configuration based on deployment location involves rationally allocating sensor resources according to the signal acquisition needs and hardware conditions of different locations. For example, in critical locations such as near valves, it may be necessary to increase the sensor deployment density to obtain more accurate acoustic fingerprint signals; for locations that are far away and have difficulty in signal transmission, it is necessary to select appropriate sensor types and deployment methods to ensure that the required acoustic fingerprint signals can be effectively acquired. Ultimately, a multi-source acoustic fingerprint acquisition scheme is generated, including signal type, sampling frequency, and deployment location, so that the acquisition scheme can acquire acoustic fingerprint signals related to oil and gas pipeline valves as comprehensively and accurately as possible while meeting hardware performance and transmission condition limitations.

[0077] Example 2:

[0078] The acoustic signature acquisition terminal adheres to the acquisition frequency and signal type requirements specified in the acquisition scheme during the process of acquiring and diagnosing multi-dimensional acoustic signature signals from valves. The acquisition scheme is formulated based on the analysis of the operating characteristics and diagnostic needs of oil and gas pipeline valves, aiming to comprehensively acquire acoustic signature information related to valve internal leakage. The acquisition terminal simultaneously acquires vibration acoustic signature data of the valve body, flow acoustic signature data of the pipeline fluid, and interference acoustic signature data from the environmental background.

[0079] After acquiring the vibration acoustic data, a short-time zero-crossing rate algorithm is used to process it. This algorithm extracts time-domain features by calculating the number of times the vibration acoustic signal crosses the zero axis per unit time. The vibration characteristics of a valve change during normal operation and when abnormal conditions such as internal leakage occur, and these changes are reflected in the time-domain characteristics of the vibration acoustic signal. By calculating the short-time zero-crossing rate of the vibration acoustic data, these subtle changes can be captured, thereby obtaining effective feature information related to valve vibration.

[0080] For flow acoustic fingerprint data of pipeline fluids, the Mel-Cepstral Coefficient (MCC) method is used for frequency domain feature extraction. The MCC method is based on the human ear's perception of sound frequency, converting the actual frequency of sound to the Mel frequency scale. In oil and gas pipelines, normal fluid flow and abnormal flow caused by factors such as valve leakage will generate acoustic fingerprint signals with different frequency distributions. Processing the flow acoustic fingerprint data using the MCC method allows for analysis of these signals from a frequency domain perspective, extracting features related to the pipeline fluid flow state, and revealing the characteristics of fluid acoustic fingerprints in a way that more closely matches human auditory perception.

[0081] Envelope analysis is used to extract noise features from interference acoustic fingerprint data in the context of environmental background. Envelope analysis involves demodulation and other processing operations on the interference acoustic fingerprint signal. In actual oil and gas pipeline environments, various interference factors generate noise, which can interfere with the effective acoustic fingerprint signal required for valve leakage diagnosis. Envelope analysis can extract noise features from complex interference acoustic fingerprint signals, distinguishing noise from the effective acoustic fingerprint signal.

[0082] The temporal features of vibration acoustic data, the frequency features of flow acoustic data, and the noise features of interference acoustic data extracted using the different methods described above are fused to form a fused feature vector. This fused feature vector integrates acoustic feature information from various sources, comprehensively reflecting the acoustic characteristics generated during valve operation.

[0083] The pre-trained voiceprint classification model plays a crucial role in the diagnostic process. This model comprises an internal leakage level determination layer based on a feedforward neural network and an abnormal pattern recognition layer based on a recurrent neural network. After the fused feature vector is input into the voiceprint classification model, the feedforward neural network of the internal leakage level determination layer performs multi-layer processing and calculation on the input feature vector. Each layer of the feedforward neural network performs specific transformations and processing on the input information. Through layer-by-layer transmission and calculation, features related to the internal leakage level are gradually extracted, and the internal leakage level classification result is output, classifying the valve's internal leakage into different levels.

[0084] The recurrent neural network in the anomaly pattern recognition layer leverages its ability to process sequential data to analyze the fused feature vectors. During valve operation, different anomaly patterns generate acoustic signature signals with specific sequential characteristics. The recurrent neural network can capture these sequential features, match them with pre-trained anomaly pattern features, output anomaly feature matching results, and identify the specific anomaly pattern occurring in the valve.

[0085] Finally, a logical AND operation is performed between the internal leakage level classification result output by the internal leakage level determination layer and the anomaly feature matching result output by the anomaly pattern recognition layer. The logical AND operation comprehensively evaluates the two results, and the final diagnostic result is output only when both the internal leakage level classification result and the anomaly feature matching result meet certain conditions. This diagnostic result comprehensively reflects the level of valve internal leakage and the anomaly pattern that has occurred, providing detailed and accurate information for subsequent handling of internal leakage in oil and gas pipeline valves.

[0086] Example 3:

[0087] When transmitting diagnostic results to the management and optimization module for adjustment of diagnostic strategies, the first step is to encapsulate the diagnostic results based on the data transmission protocol. The diagnostic results include information such as internal leak level, anomaly mode, and location of occurrence. The data transmission protocol specifies the encapsulation format of this information, forming structured feedback information that is easy to transmit and process within the system.

[0088] Structured feedback information is transmitted to the management optimization module's storage database via a data transmission middleware. During this process, timestamp calibration is performed on the historical policy data and current feedback information in the storage database. First, the timestamp distribution characteristics of the historical policy data in the management optimization module's storage database are obtained. Timestamps are information that marks the time when data is generated or an event occurs. By analyzing the timestamps of the historical policy data, the average interval of the timestamps is calculated, and this average interval is set as the baseline calibration period, using a symbol... express.

[0089] The timestamps of the structured feedback information are processed. The timestamps of the structured feedback information are then adjusted according to the baseline calibration period. Perform a round-down operation to obtain the calibration timestamp. Let the original timestamp of the structured feedback information be... The calibration timestamp is The calculation method for rounding down can be expressed as:

[0090] in, This represents the floor function, which takes the largest integer not greater than the value in parentheses.

[0091] Then, an exponentially weighted average is calculated for the historical strategy data whose timestamps fall within one period before and after the calibration timestamp, to determine the time calibration value of the historical strategy data. Let the historical strategy data be at the timestamp... The value at that location is The weighting coefficient of the exponentially weighted average is ( ), calculate time calibration value The process is as follows: Determine the time range as Historical strategy data within this range is used in the calculation. For each timestamp Data at the location Assign weights The weight varies with the calibration timestamp. The time decays exponentially with increasing time intervals. Time calibration value. The calculation formula is:

[0092] The formula means that the historical strategy data within a period before and after the calibration timestamp are summed according to the time interval with the calibration timestamp, and then divided by the sum of all weights to obtain a time calibration value that comprehensively considers the time correlation.

[0093] Finally, the numerical data of the structured feedback information is nonlinearly interpolated with the time calibration values. The numerical data in the structured feedback information, such as the quantized values ​​corresponding to the internal leakage level and the encoded values ​​of the abnormal patterns, are processed with the calculated time calibration values. The nonlinear interpolation method establishes a nonlinear correspondence between the two based on the data distribution characteristics, obtaining the time-calibrated feedback data sequence through this relationship. The specific nonlinear interpolation method can be selected according to the actual data characteristics, such as polynomial interpolation or spline interpolation. Through this series of operations, the structured feedback information is kept consistent with historical strategy data in the time dimension, allowing data generated at different times to be compared and analyzed on the same time scale, facilitating subsequent adjustments to the diagnostic strategy based on this data.

[0094] Example 4:

[0095] When constructing the strategy adjustment mapping table, the time-calibrated feedback data sequence is processed first. Assume there exists a set of time-calibrated feedback data containing multiple data records, each covering information such as internal leakage level, anomaly mode, and location. For example, one record shows an internal leakage level of "moderate" for a valve in an oil and gas pipeline, an anomaly mode of "seal wear," and a location of "valve V-101 in pipeline A"; another record shows an internal leakage level of "slight," an anomaly mode of "pipeline vibration causing loose connection," and a location of "valve V-202 in pipeline B."

[0096] The correlation coefficients between internal leak level and inspection priority, anomaly mode and remediation process, and occurrence location and resource allocation are calculated separately. Taking the calculation of the correlation coefficient between internal leak level and inspection priority as an example, all time-calibrated feedback data records are traversed. For internal leak levels, they are divided into different categories such as "minor," "moderate," and "critical." For inspection priorities, corresponding levels are also set, such as "low," "medium," and "high." In the data records, the records with an internal leak level of "minor" are examined, and the distribution of the number of records with inspection priorities of "low," "medium," and "high" is statistically analyzed. Similarly, similar statistics are performed on the records with internal leak levels of "moderate" and "critical." By analyzing the correspondence between these data, the correlation coefficients are calculated. The calculation methods for the correlation coefficients between anomaly mode and remediation process, and occurrence location and resource allocation are similar, all based on the analysis and calculation of the correspondence between different variables in the feedback data records.

[0097] Next, a threshold is set, determined based on actual conditions and experience. Assuming a threshold of 0.5, variable pairs with correlation coefficients greater than this threshold are selected as core correlation items in the mapping table. For example, calculations show that the correlation coefficient between an internal leak level of "severe" and an inspection priority of "high" is 0.6, greater than the threshold of 0.5. Therefore, the variable pair "internal leak level - severe" and "inspection priority - high" will be identified as a core correlation item. Similarly, the correlation coefficient between the anomaly pattern "seal wear" and the repair process "replace seal" is 0.7, and this will also be included as a core correlation item.

[0098] The core correlation items are statistically segmented into intervals. Taking the core correlation item between internal leakage level and inspection priority as an example, for internal leakage level, the intervals are further subdivided. For example, the "minor" internal leakage level is further subdivided into "minor-early" and "minor-late", and the "moderate" internal leakage level is subdivided into "moderate-mild" and "moderate-severe", etc. For each subdivided internal leakage level interval, the corresponding reasonable inspection priority value range is statistically analyzed. For example, after statistical analysis of a large amount of data, it was found that when the internal leakage level is in "minor-early", the inspection priority is usually in the "low-lower" range; while when the internal leakage level is "moderate-severe", the inspection priority is mostly in the "moderate-higher" range. For the core correlation item between abnormal mode and repair process, the abnormal mode is further classified according to the specific fault manifestation, such as "seal wear" is divided into "surface wear" and "deep wear", and then the standard repair process corresponding to each subdivided interval of abnormal mode is determined. For the core correlation between the location of the incident and resource allocation, the location of the incident is divided into intervals based on factors such as the geographical location and importance of the pipeline. For example, the pipeline is divided into "critical transport section" and "secondary transport section". Then, the resource allocation ratio corresponding to each interval is determined. For example, the allocation ratio of human and material resources is relatively high when allocating resources to the "critical transport section".

[0099] Through the above series of operations, a strategy adjustment mapping table containing discrete interval mapping relationships is constructed. This mapping table records in detail the relationship between each interval of internal leakage level and the corresponding interval of inspection priority, such as "minor-early" internal leakage level corresponding to "low-lower" inspection priority; the correspondence between each interval of abnormal mode and repair process, such as "seal-surface wear" abnormal mode corresponding to "clean the sealing surface and apply sealant" repair process; and the correspondence between each interval of occurrence location and resource allocation, such as "critical transport section" occurrence location corresponding to a higher proportion of human and material resources allocation. This strategy adjustment mapping table provides a specific reference for subsequent adjustments to the diagnostic strategy based on the diagnostic results, enabling the diagnostic strategy to be adjusted specifically according to the actual internal leakage situation.

[0100] Example 5:

[0101] A three-level evaluation sub-framework is constructed, comprising valve body layer, pipe section layer, and system layer, to evaluate the internal leakage of valves in oil and gas pipelines from different levels and dimensions.

[0102] The valve body evaluation sub-framework revolves around a single valve, setting three evaluation indicators: acoustic energy value, frequency offset, and duration. The acoustic energy value characterizes the amount of energy contained in the acoustic signal during valve operation. Under normal valve operation, the acoustic energy value remains within a relatively stable range; however, when the valve experiences abnormal conditions such as internal leakage, the vibration and flow changes caused by fluid leakage will alter the acoustic energy value, potentially manifesting as an increase in energy or abnormal fluctuations. The frequency offset reflects the degree of difference between the valve's acoustic frequency and its frequency under normal operation. A normally operating valve's acoustic frequency conforms to a specific pattern and range; once the valve's internal structure is damaged, the seal fails, or there is foreign object interference, the acoustic frequency will shift. Monitoring the frequency offset allows for the detection of these subtle changes. The duration records the length of time abnormal acoustic signals occur. Some brief acoustic anomalies may be caused by accidental interference, while abnormal acoustic signals lasting for a certain period are more likely to indicate substantial problems such as internal leakage. This indicator helps determine the severity and persistence of the abnormal situation.

[0103] The pipe segment assessment sub-framework focuses on associated pipe segments and encompasses three assessment metrics: acoustic signature attenuation, impact length, and number of affected nodes. Acoustic signature attenuation measures the energy attenuation of acoustic signature signals as they propagate through the pipe segment. In oil and gas pipeline systems, acoustic signature signals attenuate along the pipe segment due to factors such as pipe material, diameter variations, and curvature. When valve leakage occurs, the attenuation characteristics of the acoustic signature signal generated by the leakage during propagation within the pipe segment differ from those under normal conditions. By analyzing the acoustic signature attenuation, the location of the leakage and its impact on surrounding pipe segments can be inferred. Impact length indicates the length of the pipe segment affected by valve leakage. Leakage may cause changes in fluid pressure and velocity within the pipe segment, and this impact propagates along the pipe segment. The impact length directly reflects the spatial extent of the leakage's influence. The number of affected nodes reflects the number of pipe nodes affected by the leakage. Pipeline systems contain numerous connection nodes and monitoring nodes. The occurrence of leakage may cause abnormal parameters at multiple nodes. By statistically counting the number of affected nodes, the degree of impact of the leakage on the overall structure and operating status of the pipe segment can be assessed.

[0104] The system-level evaluation sub-framework takes a global perspective and includes three evaluation indicators: global acoustic signature coverage, redundancy detection rate, and recovery time. Global acoustic signature coverage reflects the coverage of the entire oil and gas pipeline system by the acoustic signature acquisition equipment, i.e., the proportion of the system where acoustic signature signals can be effectively acquired. A higher global acoustic signature coverage means more comprehensive acquisition of acoustic signature information from all parts of the system, enabling timely detection of potential internal leaks; conversely, the existence of acoustic signature acquisition blind spots may prevent internal leaks from being detected in a timely manner. Redundancy detection rate indicates the proportion of the system with redundant detection capabilities. By setting multiple acoustic signature acquisition points or using multiple detection methods to achieve redundant detection, the accuracy and reliability of internal leak detection can be improved. The redundancy detection rate reflects the system's fault tolerance and backup capabilities in detection. Recovery time is used to evaluate the time required for the system to recover from a fault state to normal operation after detecting an internal leak. It comprehensively considers factors such as maintenance procedures, resource allocation, and repair technologies, and is an important indicator for measuring the system's ability to recover from internal leaks.

[0105] After obtaining the evaluation index data for each level of the evaluation sub-framework, the comprehensive score of each level of the evaluation sub-framework is calculated based on factor analysis. Factor analysis extracts a few comprehensive factors by dimensionality reduction of multiple index data. These comprehensive factors can reflect most of the information of the original indexes. Each comprehensive factor is assigned a corresponding weight according to its explanatory power for the original index. The comprehensive score of each level of the evaluation sub-framework is obtained by weighted summation of the scores of each index data on the comprehensive factors. Finally, the comprehensive scores of the three evaluation sub-frameworks—valve body layer, pipe section layer, and system layer—are integrated by linear weighted summation. Each sub-framework is assigned an appropriate weight according to its importance in the overall evaluation. The final evaluation result of the three-level evaluation sub-framework, including the valve body layer, pipe section layer, and system layer, is calculated, realizing a multi-level and comprehensive evaluation of internal leakage of oil and gas pipeline valves from individual valves and related pipe sections to the entire system.

[0106] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0107] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis, characterized in that, Includes the following steps: Acquire acoustic fingerprint data of oil and gas pipeline valves under operating conditions, and construct an acoustic fingerprint feature database based on the acoustic fingerprint data; Based on the aforementioned voiceprint feature database, an internal leakage diagnostic framework is designed, and a multi-source voiceprint acquisition scheme is determined by combining the diagnostic framework with the voiceprint feature database. The acoustic signature acquisition terminal acquires the valve's multidimensional acoustic signature signal according to the acquisition scheme, performs internal leakage level determination and abnormal pattern recognition based on the multidimensional acoustic signature signal, and outputs the diagnostic results. The diagnostic results are transmitted to the management optimization module for adjustment of the diagnostic strategy. The process involves designing an internal leakage diagnostic framework based on the voiceprint feature database, and determining a multi-source voiceprint acquisition scheme by combining the diagnostic framework with the voiceprint feature database. Specifically: The voiceprint feature database is decomposed into dimensions to extract three core evaluation dimensions: voiceprint energy distribution, frequency characteristic offset, and time-varying fluctuation. Initial weight parameters are assigned to each dimension based on expert experience. The core evaluation dimensions are topologically mapped to the spatial structure hierarchy of oil and gas pipelines, and a three-level evaluation sub-framework is constructed, including the valve body layer, pipe section layer, and system layer. The synergistic influence weights between the sub-frameworks are calculated through feature cross-validation. The core feature dimensions in the voiceprint feature database are matched with the three-level evaluation sub-framework to check the degree of matching, and the feature coverage missing items corresponding to each sub-framework are screened out. Based on the missing feature coverage items, the types of acoustic signals that need to be collected are determined. Taking into account the sensor deployment density and data transmission bandwidth limitations in the pipeline environment, the collection frequency and deployment location of each type of signal are calculated. Based on the decision tree algorithm, the acquisition frequency, deployment location and feature coverage missing items are comprehensively judged to generate a multi-source acoustic fingerprint acquisition scheme that includes signal type, acquisition frequency and deployment location. The acoustic signature acquisition terminal acquires multi-dimensional acoustic signature signals of the valve according to the acquisition scheme, performs internal leakage level determination and abnormal pattern identification based on the multi-dimensional acoustic signature signals, and outputs diagnostic results, specifically: According to the acquisition frequency and signal type requirements of the acquisition scheme, the acoustic fingerprint acquisition terminal synchronously acquires the vibration acoustic fingerprint data of the valve body, the flow acoustic fingerprint data of the pipeline fluid, and the interference acoustic fingerprint data of the environmental background. The vibration acoustic data is extracted in the time domain using the short-time zero-crossing rate algorithm, the flow acoustic data is extracted in the frequency domain using the Mel-Cepstral Coefficient method, and the noise features of the interference acoustic data are extracted using the envelope analysis method to obtain a fused feature vector. The fused feature vector is input into a pre-trained voiceprint classification model, which is based on a feedforward neural network to construct an internal leakage level determination layer and a recurrent neural network to construct an abnormal pattern recognition layer. The internal leakage level determination layer outputs the internal leakage level classification result, and the abnormal feature matching result is output through the abnormal pattern recognition layer. The classification result and the matching result are then logically ANDed to output the diagnostic result.

2. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 1, characterized in that, The acquisition of acoustic fingerprint data under the operating status of oil and gas pipeline valves, and the construction of an acoustic fingerprint feature database based on the acoustic fingerprint data, specifically involves: By connecting the valve body sensor and the pipeline auxiliary microphone array through a distributed interface, the time-domain waveform information of the valve operation sound pattern and the frequency domain distribution information of the steady-state sound pattern are obtained. The time-domain waveform information of the valve operation soundprint is processed by frame-by-frame windowing using time-frequency conversion technology, and the background noise is filtered by the frequency domain distribution information of the steady-state soundprint using the energy threshold method to obtain the denoised soundprint feature unit. A voiceprint feature mapping network is constructed based on feature association rules. The denoised voiceprint feature units are input into the mapping network for temporal association analysis to extract core feature dimensions that are strongly correlated with the leakage state. The core feature dimensions are standardized and labeled with corresponding internal leakage status tags. Features are aggregated according to valve type and operating period to construct a voiceprint feature database containing time domain features, frequency domain features and equipment features.

3. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 1, characterized in that, The decision tree algorithm is used to comprehensively determine the acquisition frequency, deployment location, and missing feature coverage, generating a multi-source speaker acquisition scheme that includes signal type, acquisition frequency, and deployment location. Specifically: The hardware performance indicators of each acoustic signature acquisition terminal in the pipeline environment are obtained, including signal sensitivity, sampling rate, and battery life. Construct a set of data acquisition constraints, which includes constraints on the maximum sampling rate of a single terminal, the maximum storage capacity of a single terminal, and the total transmission bandwidth of the system. Using the collection frequency, deployment location, and missing feature coverage as decision factors, a correlation function between each decision factor and the set of collection constraints is established. By fusing the importance scores of each decision factor through the decision tree splitting rule, the feasibility scores of each signal type at different acquisition frequencies are obtained. The sampling frequency combinations with a feasibility score higher than a set threshold are selected, and the sensor resources for each signal type are dynamically configured according to the deployment location to generate a multi-source acoustic signature acquisition scheme.

4. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 1, characterized in that, The step of transmitting the diagnostic results to the management optimization module for diagnostic strategy adjustment specifically involves: The diagnostic results are encapsulated based on the data transmission protocol to generate structured feedback information including internal leak level, abnormal mode, and location of occurrence. The structured feedback information is transmitted to the storage database of the management and optimization module through a data transmission middleware, and the historical strategy data and current feedback information in the storage database are timestamped. Construct a strategy adjustment mapping table, which includes the correspondence between internal leakage level and inspection priority, the correspondence between abnormal mode and repair process, and the correspondence between occurrence location and resource allocation; Based on the mapping table, the parameters of the historical strategy data are adjusted to generate an updated diagnostic strategy.

5. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 4, characterized in that, The process of transmitting the structured feedback information to the storage database of the management optimization module via a data transmission middleware, and performing a timestamp calibration operation on the historical strategy data and current feedback information in the storage database, specifically involves: The timestamp distribution characteristics of historical strategy data stored in the management optimization module's database are obtained, and the average interval of the timestamps is calculated as the benchmark calibration period. The timestamp of the structured feedback information is rounded down according to the reference calibration period to obtain the calibration timestamp; The time calibration value of the historical strategy data is calculated by performing an exponentially weighted average on the data whose timestamps are within one period before and after the calibration timestamp. The numerical data of the structured feedback information is nonlinearly interpolated with the time calibration value to obtain the time-calibrated feedback data sequence.

6. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 5, characterized in that, The constructed strategy adjustment mapping table includes the correspondence between internal leak levels and inspection priorities, the correspondence between anomaly modes and remediation processes, and the correspondence between occurrence locations and resource allocation, specifically: Correlation analysis was performed on the feedback data sequence after time calibration to calculate the correlation coefficients between internal leak level and inspection priority, abnormal mode and repair process, and occurrence location and resource allocation. Select variable pairs whose absolute correlation coefficient is greater than a set threshold as the core correlation items in the mapping table; The core related items are divided into intervals for statistical analysis to determine the range of values ​​of another variable corresponding to each variable interval, and a strategy adjustment mapping table containing discrete interval mapping relationships is constructed.

7. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 6, characterized in that, The step of adjusting parameters of historical strategy data according to the mapping table to generate an updated diagnostic strategy specifically involves: Adjust the inspection priority parameter in the historical strategy data according to the mapping relationship between internal leakage level and inspection priority; For the repair process parameters in the historical strategy data, replace the matching standard process according to the mapping relationship between the anomaly pattern and the repair process; Based on the mapping relationship between the location of occurrence and resource allocation, the resource allocation ratio is adjusted for the resource allocation parameters in the historical strategy data. The adjusted inspection priorities, repair processes, and resource allocation parameters are verified for consistency, and an updated diagnostic strategy is generated.

8. The method for online diagnosis of internal leakage in oil and gas pipeline valves based on acoustic signature analysis according to claim 1, characterized in that, The construction of the three-level evaluation sub-framework, comprising the valve body layer, pipe section layer, and system layer, is as follows: The valve body layer evaluation subframe includes three evaluation metrics for a single valve: acoustic energy value, frequency offset, and duration. The pipe segment evaluation subframe includes three evaluation indicators: acoustic attenuation of associated pipe segments, impact length value, and number of affected nodes. The system-level evaluation sub-framework includes three evaluation metrics: global voiceprint coverage, redundancy detection rate, and recovery timeliness. The comprehensive score of each level of the evaluation subframe is calculated based on factor analysis, and a three-level evaluation subframe including valve body layer, pipe section layer and system layer is constructed by linear weighted summation.

Images