Intelligent monitoring method and system for sealing performance of hydraulic drive compressor

Abstract

The invention provides an intelligent monitoring method and system for the sealing performance of a liquid-driven compressor, and the method comprises the steps: collecting the high-frequency pressure data and circulating liquid flow data of a sealing cavity of the liquid-driven compressor, and a friction sound signal of a sealing interface; based on the high-frequency pressure data and the circulating liquid flow data, constructing a pressure flow covariant relationship; performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information; adopting a DIANA algorithm to carry out adaptive clustering processing on the sound energy intensity distribution information to generate a sealing abnormity index; and the pressure flow covariant relation and the sealing abnormity index are analyzed in a combined mode, a monitoring result representing the sealing performance of the hydraulic drive compressor is generated, and when the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered. According to the invention, the timeliness and reliability of sealing performance monitoring are improved.

Description

Technical Field

[0001] The present application relates to the technical field of compressor sealing performance monitoring, and in particular to an intelligent monitoring method and system for the sealing performance of a liquid-driven compressor. Background Art

[0002] Liquid-driven compressors are widely used in the petrochemical industry, gas transportation, and other fields, and their sealing performance directly impacts the equipment's operational efficiency and safety. Because the compressor piston rod is subjected to long-term, high-speed reciprocating motion, micro-leaks are prone to occur at the sealing interface. Traditional monitoring methods struggle to detect early failures. Therefore, an intelligent monitoring method is urgently needed that can capture dynamic sealing performance changes in real time and accurately warn of leakage risks.

[0003] Currently, a monitoring solution based on vibration signal analysis is being used to evaluate the seal performance of hydraulic compressors. This solution utilizes a vibration sensor placed outside the seal chamber to capture the vibration signal caused by piston rod movement. It then uses frequency domain feature extraction to identify abnormal vibration patterns and then determines the seal status through threshold comparison. Some improved solutions also incorporate machine learning algorithms, using historical data to train classification models and improve the accuracy of anomaly detection.

[0004] Vibration signals are susceptible to interference from mechanical noise, especially under complex operating conditions. Micro-leakage signatures can be masked by background vibration, leading to false alarms or missed detections. Furthermore, vibration analysis struggles to distinguish seal failure from other mechanical faults (such as bearing wear), reducing the relevance of monitoring results. Furthermore, reliance on a single vibration parameter and the lack of collaborative analysis of multiple physical quantities limits the ability to detect even weak leaks in the early stages. Summary of the Invention

[0005] The present application provides an intelligent monitoring method and system for the sealing performance of a liquid-driven compressor, which is used to solve the problems of poor timeliness and low reliability of sealing performance monitoring in the prior art.

[0006] In a first aspect, the present application provides an intelligent monitoring method for the sealing performance of a liquid-driven compressor, comprising: Collect high-frequency pressure data, circulating liquid flow data, and friction sound signals of the sealing interface of the liquid-driven compressor seal cavity; Constructing a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data; performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information; Adopting the DIANA algorithm to perform adaptive clustering processing on the sound energy intensity distribution information to generate a sealing abnormality index; The pressure-flow covariance relationship and the sealing abnormality index are jointly analyzed to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor. When the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

[0007] Optionally, the adaptive clustering process of the sound energy intensity distribution information using the DIANA algorithm to generate a sealing abnormality index includes: The spatial grid of the sealing interface is used as the sample point to construct the initial cluster; Calculating the degree of dispersion of the sound energy intensity values ​​within the initial cluster, and when the degree of dispersion of the sound energy intensity values ​​within the initial cluster is greater than a preset split threshold, dividing the initial cluster into two subclusters; Taking each of the subclusters as a corresponding intermediate cluster, iteratively performing calculation operations and splitting operations until the discrete degree of the sound energy intensity values ​​in all the intermediate clusters is less than or equal to a preset splitting threshold, thereby obtaining a target cluster; The number of spatial grids in the target cluster whose sound energy intensity values ​​exceed the preset abnormal boundary is counted to generate a sealing abnormality index.

[0008] Optionally, the calculating the discrete degree of the sound energy intensity values ​​within the initial clusters includes: Based on the sound energy intensity values ​​corresponding to all spatial grids in the initial cluster, forming a sound energy intensity value set; performing outlier correction processing on the sound energy intensity value set; Based on the corrected sound energy intensity value set, a standard deviation calculation formula is used to determine the degree of dispersion of the sound energy intensity values ​​within the initial cluster.

[0009] Optionally, dividing the initial cluster into two subclusters includes: Constructing a spatial gradient field of sound energy intensity values ​​of the initial clusters; Determining a gradient vector with a maximum modulus value in the acoustic energy intensity value spatial gradient field; Establishing a segmentation normal plane along the direction of the gradient vector; The initial cluster is divided into two subclusters using the segmentation normal plane as a boundary.

[0010] Optionally, constructing a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data includes: Dividing the high-frequency pressure data into a plurality of discrete pressure segments according to a single reciprocating motion cycle of a piston rod of a liquid-driven compressor; Extracting a pressure peak time series sequence from each of the pressure segments; Synchronously dividing the circulating fluid flow data into a plurality of flow fluctuation segments; Calculate the instantaneous flow rate change rate of each flow fluctuation segment; Establishing a dynamic correlation between the pressure peak time series and the instantaneous flow rate change rate within the same motion cycle; Based on the dynamic correlation relationship of N consecutive motion cycles, a pressure-flow covariance relationship is constructed, where N is greater than or equal to 3.

[0011] Optionally, performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information includes: spatially selectively enhancing the acoustic energy component of the friction acoustic signal by acoustic wave guidance; Perform frequency domain analysis on the enhanced acoustic energy components to identify specific frequency bands associated with high-frequency leakage; obtaining sound energy intensity data within the specific frequency band; Mapping the acoustic energy intensity data to a sealing interface spatial coordinate system to form a two-dimensional acoustic energy intensity distribution map; The two-dimensional sound energy intensity distribution diagram is divided into a plurality of equally spaced spatial grids, the cumulative value of the sound energy intensity in each spatial grid is calculated, and the sound energy intensity distribution information is generated based on the cumulative value of the sound energy intensity in each spatial grid.

[0012] Optionally, the jointly analyzing the pressure-flow covariance relationship and the sealing abnormality index to generate a monitoring result characterizing the sealing performance of the liquid drive compressor includes: Calculating the pressure-flow state scalar based on the pressure-flow covariance relationship; converting the sealing anomaly indicator into an acoustic anomaly scalar; A weighted fusion calculation is performed on the pressure flow state scalar and the acoustic anomaly scalar, and the weighted fusion calculation result is used as a monitoring result characterizing the sealing performance of the liquid drive compressor.

[0013] In a second aspect, the present application provides an intelligent monitoring system for the sealing performance of a liquid-driven compressor, comprising: The acquisition module is used to collect high-frequency pressure data of the liquid-driven compressor sealing chamber, circulating liquid flow data, and friction sound signals of the sealing interface; A construction module, configured to construct a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data; an enhancement module, configured to perform directional enhancement processing on the friction sound signal to form sound energy intensity distribution information; A generation module, configured to perform adaptive clustering processing on the sound energy intensity distribution information using a DIANA algorithm to generate a sealing abnormality index; The trigger module is used to jointly analyze the pressure-flow covariance relationship and the sealing abnormality index to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor. When the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

[0014] In a third aspect, the present application provides a computing device comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute an intelligent monitoring method for the sealing performance of a liquid-driven compressor as described in any one of the first aspects.

[0015] In a fourth aspect, the present application provides a computer storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement an intelligent monitoring method for the sealing performance of a liquid-driven compressor as described in any one of the first aspects.

[0016] In the present application, an intelligent monitoring method for the sealing performance of a liquid-driven compressor is provided, the method comprising: collecting high-frequency pressure data, circulating liquid flow data, and a friction sound signal of a sealing interface of a sealing chamber of the liquid-driven compressor; constructing a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating liquid flow data; performing directionally enhanced processing on the friction sound signal to form acoustic energy intensity distribution information; performing adaptive clustering processing on the acoustic energy intensity distribution information using the DIANA algorithm to generate a sealing anomaly index; jointly analyzing the pressure-flow covariance relationship and the sealing anomaly index to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor, and when the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

[0017] The technical solution provided by this application has the following beneficial effects: This application achieves simultaneous multi-source data acquisition, providing fundamental data support for comprehensive sealing performance evaluation. A dynamic correlation model between pressure and flow is established to accurately reflect the changing characteristics of the fluid state within the sealing cavity. It effectively extracts characteristic signals of micro-leakage at the sealing interface, improving the detection sensitivity of weak acoustic signals. It also achieves adaptive zoning of the acoustic energy intensity distribution, accurately identifying areas of abnormal energy accumulation. By integrating pressure and flow parameters with acoustic characteristics, a multi-dimensional collaborative assessment of the sealing status is achieved.

[0018] Furthermore, the present application also divides the sealing interface into spatial grid units as sample points, first constructs an initial cluster containing all grids; calculates the discrete degree of the sound energy intensity value within the cluster, and when it exceeds the set threshold, divides the initial cluster into two sub-clusters along the direction of maximum change; it iteratively executes the splitting process until the discrete degree of all sub-clusters meets the convergence condition; finally, counts the number of grids in each cluster that exceed the abnormal boundary to generate a quantitative sealing anomaly index.

[0019] In addition, this solution realizes intelligent zoning detection of acoustic energy distribution at the sealing interface, accurately identifies abnormal areas through an adaptive hierarchical splitting algorithm, overcomes the limitations of traditional fixed threshold methods, and improves the accuracy of micro-leak detection and early warning capabilities.

[0020] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 A flow chart of an intelligent monitoring method for sealing performance of a liquid-driven compressor provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of an intelligent monitoring system for the sealing performance of a liquid-driven compressor provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0024] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0025] Existing solutions based on vibration signal analysis for monitoring seal performance in hydraulic compressors have limitations: vibration signals are susceptible to mechanical noise interference, causing micro-leakage characteristics to be masked by background vibration, resulting in unreliable detection results. Furthermore, this method struggles to distinguish seal failures from other mechanical faults, resulting in poor fault identification specificity. Furthermore, single-vibration parameter analysis lacks the synergy of multiple physical quantities, resulting in insufficient sensitivity for detecting early, weak leaks. These issues stem from the limited nature of the monitoring dimension and the crude nature of the signal processing methods, making them difficult to meet the demands of high-precision seal status monitoring.

[0026] In response to the above-mentioned defects, this application proposes an intelligent monitoring method for the sealing performance of liquid-driven compressors. By synchronously collecting high-frequency pressure data, circulating liquid flow data and friction sound signals, a pressure-flow dynamic covariance model is constructed, and combined with the adaptive clustering analysis of the sound energy intensity distribution, a comprehensive evaluation of the sealing status is achieved. Specifically, the acoustic wave guidance technology is used to enhance the micro-leakage characteristics in the friction sound signal, and the DIANA algorithm is used to perform hierarchical splitting and clustering of the sound energy distribution to accurately identify abnormal energy accumulation areas; at the same time, the pressure-flow covariance parameters and acoustic anomaly indicators are integrated to establish a multi-dimensional sealing performance evaluation system. This method breaks through the limitations of single vibration monitoring, and through the collaborative analysis of multi-source signals and feature fusion, it effectively suppresses noise interference, improves the sensitivity of micro-leakage detection and the accuracy of fault identification, and provides a reliable early warning for the degradation of compressor sealing performance.

[0027] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0028] Figure 1 This is a flow chart of an intelligent monitoring method for the sealing performance of a liquid-driven compressor provided in an embodiment of the present application, such as Figure 1 As shown, the method includes: Step 101: collecting high-frequency pressure data of the sealed cavity of the liquid-driven compressor, circulating liquid flow data, and friction sound signals of the sealed interface.

[0029] In step 101, high-frequency pressure data represents the rapid changes in fluid pressure within the seal chamber, reflecting the dynamic pressure fluctuations within the seal chamber during piston rod movement. Circulating fluid flow data represents the flow rate of the working medium in the compressor's circulation system, characterizing the fluid exchange between the seal chamber's inlet and outlet. The frictional acoustic signal represents the acoustic wave signal generated by the relative motion of the seal interface, containing microscopic characteristic information about the seal contact state.

[0030] In the embodiments of this application, a pressure sensor installed in the sealed chamber collects pressure fluctuation signals in real time. A flow meter simultaneously records the liquid flow rate in the circulation pipeline. An acoustic sensor array is placed on the outside of the sealing ring to capture friction sound waves. The pressure sensor continuously records the sealed chamber pressure at very short intervals, while the flow meter measures the flow rate of the liquid through a specific cross-section. The acoustic sensor array utilizes a specific structure to enhance its ability to receive sound waves in the direction normal to the sealing interface. These three data acquisition devices achieve data synchronization through a unified clock signal.

[0031] For example, in testing a certain type of liquid-driven compressor, a Brand A pressure sensor collects seal chamber pressure at a frequency of 10,000 times per second, a Type B flowmeter records flow changes in the circulation pipeline, and a Series C acoustic sensor array is arranged around the seal ring. The pressure sensor is installed in the center of the seal chamber, the flowmeter is installed in the straight section of the circulation pipeline, and the acoustic sensors are evenly spaced. A synchronization controller ensures that the three devices are time-aligned in data acquisition. The pressure data reflects the periodic pressure fluctuations caused by the reciprocating motion of the piston rod, the flow data displays the corresponding relationship with pressure changes, and the acoustic signal reveals the friction characteristics of the seal interface.

[0032] Step 102: Constructing a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data.

[0033] In step 102, the pressure-flow covariance relationship represents a dynamic correlation model between pressure and flow parameters, and characterizes the cooperative characteristics of the fluid state changes in the sealing system.

[0034] In this embodiment, the collected high-frequency pressure data is segmented into several time periods based on the piston rod motion cycle, and the sequence of pressure peaks within each time period is extracted. The flow rate data for these time periods is then processed simultaneously, and the rate of change of flow rate is calculated. A corresponding relationship between pressure peaks and flow rate changes is established. By statistically analyzing the synergistic changes between the two over multiple motion cycles, a covariance model reflecting the dynamic characteristics of the sealing system fluid is constructed.

[0035] For example, the pressure data obtained from the test is segmented according to the 120-millisecond movement cycle of the piston rod, and 12 extreme pressure points in each segment are extracted to form a sequence. The flow data of the corresponding period is used to calculate the instantaneous change in each millisecond interval. By analyzing the three sets of complete movement cycle data, it is found that the flow change begins to increase at a specific time after the pressure peak occurs, and the correlation coefficient reaches 0.82. A dynamic correlation model between the two is established. The correlation coefficient calculation formula is: ,in is the peak pressure, is the corresponding flow rate change rate, and are the mean values ​​respectively.

[0036] Step 103: performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information.

[0037] In step 103 , the acoustic energy intensity distribution information represents spatial distribution characteristic data of acoustic wave energy at each position of the sealing interface.

[0038] In this embodiment, spatial filtering is performed on the collected friction sound signal to enhance the acoustic wave component normal to the sealing interface. Spectral analysis is performed on the enhanced signal to identify the active components within a specific frequency band. The intensity characteristics of the signal within this frequency band are extracted and mapped onto a two-dimensional coordinate system of the sealing ring according to the acoustic sensor position, forming a distribution map reflecting the acoustic wave energy at each location.

[0039] For example, in the test, a sound guide structure was used to enhance the acoustic signal in the normal direction of the sealing ring. Spectral analysis determined that the characteristic frequency band was 8500 to 11500 Hz. The root mean square value of the signal in this frequency band was extracted as the intensity index and mapped to the sealing ring surface coordinate system according to the sensor position to form a 40 mm square acoustic energy distribution diagram. The intensity calculation formula is: ,in is a discrete sampling value, is the number of sampling points.

[0040] Step 104: Adopting the DIANA algorithm to perform adaptive clustering processing on the sound energy intensity distribution information to generate a sealing abnormality index.

[0041] In step 104 , the sealing abnormality index represents a comprehensive parameter that quantitatively reflects the degree of abnormality in the sealing performance.

[0042] In this embodiment, the acoustic energy distribution map is divided into several grid cells as sample points, and all grid cells are initially grouped into a cluster. The degree of dispersion of the acoustic energy intensity at each point in the cluster is calculated. When the dispersion exceeds a set threshold, the cluster is split into two subclusters along the direction of the acoustic energy variation. This splitting process is iterated until the degree of dispersion of all subclusters meets the required level. Finally, the proportion of grid cells in each cluster that exceed the abnormality threshold is calculated to generate a sealing anomaly index.

[0043] For example, the acoustic energy distribution map was divided into 400 grids. The initial clustering calculated had a standard deviation of 0.85. When the threshold of 0.8 was exceeded, the clusters were split into two subclusters along an axial deflection of 15 degrees. After iterative splitting, seven stable clusters were obtained. 23% of the grids had an intensity exceeding 5, which was used as an anomaly indicator. The threshold of 0.8 was determined through historical data analysis.

[0044] Step 105: jointly analyzing the pressure-flow covariance relationship and the sealing abnormality index to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor. When the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

[0045] In step 105 , the monitoring results represent quantitative parameters for comprehensively evaluating the sealing performance status.

[0046] In this embodiment, the pressure-flow covariance parameter is converted into a state scalar, and the acoustic anomaly indicator is converted into an anomaly scalar. A weighted fusion calculation is performed on these two scalars to obtain a monitoring value that comprehensively reflects the sealing performance. When the monitoring value continuously exceeds the warning line, an early warning signal is triggered.

[0047] For example, if the pressure-flow correlation coefficient is 0.82 and the abnormality index is 23%, it will be converted to 76.7 points. The monitoring value is fused with weights of 0.6 and 0.4, resulting in a monitoring value of 0.79. An alarm is triggered if the value exceeds 0.75 three times in a row. The warning value of 0.75 is determined based on historical normal data statistics.

[0048] This method achieves comprehensive monitoring of the seal performance of liquid-driven compressors through the simultaneous acquisition and fusion analysis of multi-source data. The pressure-flow covariance accurately reflects the fluid state of the sealing system, while acoustic energy distribution analysis effectively identifies micro-leakage characteristics. An adaptive clustering algorithm improves anomaly detection accuracy, while multi-parameter fusion evaluation ensures early warning reliability, providing an effective technical approach for compressor seal maintenance.

[0049] In order to solve the problem of insufficient accuracy in identifying micro-leakage features in monitoring the sealing performance of liquid-driven compressors, in some embodiments, step 104: using the DIANA algorithm to adaptively cluster the acoustic energy intensity distribution information to generate a sealing anomaly index includes: Step 201: constructing initial clusters using the spatial grid of the sealing interface as sample points.

[0050] In step 201, the initial cluster refers to a set of all grid cells as initial analysis objects.

[0051] In this embodiment, the sealing ring surface is first divided into a regularly arranged square grid based on its geometric dimensions, ensuring that each grid cell has the same area. The acoustic energy intensity value collected from each grid cell is used as the sample point attribute. All grid cells together form an initial cluster, providing a basic data set for subsequent analysis.

[0052] Step 202: Calculate the degree of dispersion of the sound energy intensity values ​​within the initial cluster. When the degree of dispersion of the sound energy intensity values ​​within the initial cluster is greater than a preset split threshold, split the initial cluster into two subclusters.

[0053] In step 202, the degree of dispersion refers to the degree of dispersion of the acoustic energy intensity values ​​within each grid cell within a cluster, reflecting the uniformity of the acoustic energy distribution. The split threshold is the critical value used to determine whether a cluster needs to be split. It is determined based on statistical analysis of historical normal data. Subclusters are new cluster units generated through the split operation.

[0054] In this embodiment, the standard deviation of the acoustic energy intensity values ​​for all grids within the initial cluster is calculated as a dispersion indicator. When this value exceeds a preset splitting threshold, the spatial variation trend of the acoustic energy intensity at each grid point is analyzed to determine the direction of the greatest acoustic energy variation. The initial cluster is then split into two new subclusters along this direction, ensuring relatively uniform acoustic energy distribution within each subcluster.

[0055] Step 203: taking each of the subclusters as a corresponding intermediate cluster, iteratively performing calculation operations and splitting operations until the discrete degree of the sound energy intensity values ​​in all intermediate clusters is less than or equal to a preset splitting threshold, thereby obtaining a target cluster.

[0056] In step 203, the calculation operation is followed by calculating the degree of dispersion of the spatial distribution of acoustic energy within the intermediate clusters. The intermediate clusters refer to the set of subclusters to be analyzed during the iterative process. The target cluster refers to the stable cluster unit that ultimately meets the degree of dispersion requirement.

[0057] In this embodiment of the present application, the subclusters obtained in step 202 are used as new analysis objects, and the process of calculating the degree of discreteness and determining the splitting conditions is repeated. Subclusters that do not meet the discreteness requirements are further split along the direction of maximum change until the discreteness of all subclusters is below the splitting threshold, thereby obtaining the final target cluster set.

[0058] Step 204: Count the number of spatial grids in the target cluster whose sound energy intensity values ​​exceed the preset abnormal boundary to generate a sealing abnormality index.

[0059] In step 204 , the abnormal boundary refers to a critical value for determining whether the acoustic energy intensity of a grid unit is abnormal, and is determined based on the statistical distribution of acoustic energy intensity under normal operating conditions.

[0060] In the embodiment of the present application, the number of grid cells whose acoustic energy intensity exceeds the abnormal boundary in all target clusters is counted, and the ratio of the number of grid cells to the total number of grid cells is calculated as the sealing abnormality index. This index quantitatively reflects the degree of abnormality of the sealing interface.

[0061] Here's a specific example: In the sealing performance test of the liquid-driven compressor, based on the collected pressure, flow and acoustic signal data, the acoustic energy intensity distribution information is adaptively clustered. First, the 40 mm square acoustic energy distribution map of the sealing ring surface is divided into 400 square grid units with a side length of 0.2 mm. The acoustic energy intensity value E of each grid is calculated by the formula Calculated, where M = 2000 is the number of sampling points in a 20 millisecond period; all grids are used as initial clusters, and the standard deviation of 400 intensity values ​​is calculated as σ = 0.85, which is obtained by the formula Find, among them is the intensity value of the i-th grid, μ is the average value, and N=400; when σ exceeds the preset splitting threshold of 0.8, the spatial variation gradient of the acoustic energy intensity at each grid point is calculated, and the direction of maximum variation along the axial deflection of the piston rod of 15 degrees is determined. The initial cluster is divided into two subclusters along this direction, containing 158 and 242 grid units respectively; the above calculation process is repeated for these two subclusters, and the splitting is stopped when the standard deviation of all subclusters is ≤0.8, and finally 5 stable clusters are obtained; the number of grids with an intensity exceeding 5 is counted as 89, accounting for 22.25% of the total number of grids, which are used as sealing anomaly indicators. The splitting threshold of 0.8 and the anomaly boundary of 5 are determined by analyzing 100 sets of normal operating data of this model compressor, and the 95% and 99% percentile values ​​are taken respectively to ensure the reliability of the detection results.

[0062] In the embodiment of the present application, the method realizes intelligent zoning detection of acoustic energy distribution at the sealing interface through adaptive clustering analysis, accurately identifies areas of abnormal energy accumulation, overcomes the limitations of the fixed threshold method, improves the accuracy of micro-leak detection and early warning capabilities, and provides a reliable basis for compressor sealing status assessment.

[0063] In order to further improve the accuracy of calculating the dispersion degree of acoustic energy intensity, in some embodiments, step 202: calculating the dispersion degree of acoustic energy intensity values ​​within the initial clusters includes: Step 301: forming a sound energy intensity value set based on the sound energy intensity values ​​corresponding to all spatial grids in the initial cluster.

[0064] In step 301 , the sound energy intensity value set refers to a set of sound energy intensity data corresponding to all spatial grid cells in the initial cluster, and each data point represents a sound energy intensity measurement value of a grid cell.

[0065] In an embodiment of the present application, all spatial grid cells contained in the initial cluster are first traversed, the sound energy intensity values ​​recorded by each cell are extracted, and these values ​​are arranged in order of grid numbers to form a complete sound energy intensity value data set, providing basic data for subsequent analysis.

[0066] Step 302: performing outlier correction processing on the sound energy intensity value set.

[0067] In step 302 , outlier correction processing refers to the process of identifying and adjusting extreme values ​​in the sound energy intensity value set that are significantly deviated from the normal range, in order to eliminate the influence of measurement errors or local interference.

[0068] In the embodiment of the present application, the box plot method is used to identify outliers. First, the quartiles and interquartile ranges of the sound energy intensity value set are calculated. Values ​​that exceed the upper quartile plus 1.5 times the interquartile range or are lower than the lower quartile minus 1.5 times the interquartile range are determined to be outliers, and these outliers are replaced by the median values ​​of adjacent grid cells to ensure data quality.

[0069] Step 303: Based on the corrected sound energy intensity value set, a standard deviation calculation formula is used to determine the degree of dispersion of the sound energy intensity values ​​within the initial cluster.

[0070] In step 303 , the standard deviation calculation formula refers to a mathematical expression used to quantify the degree of dispersion of the sound energy intensity values, reflecting the distribution of data points around the average value.

[0071] In this embodiment, the average of all corrected sound energy intensity values ​​is calculated. The sum of the squares of the differences between each value and the average is then divided by the total number of data points and the square root is taken. This sum is used to determine the degree of dispersion. This standard deviation is used to determine whether the clusters need to be further split.

[0072] Here's a specific example: In the sealing performance test of the liquid-driven compressor, based on the set of 400 grid unit sound energy intensity values ​​obtained in Example 2, the box plot method is first used to correct the outliers. The specific process is to calculate the upper quartile Q3 of the data set as 3.8 and the lower quartile Q1 as 2.4, and obtain the interquartile range IQR=Q3-Q1=1.4. The values ​​greater than Q3+1.5IQR, i.e. 5.9, or less than Q1-1.5IQR, i.e. 0.3, are judged as outliers. A total of 8 abnormal grid points are identified, and these outliers are replaced by the median values ​​of the adjacent 8 grids, where the replacement value is determined by sorting and taking the middle value; after the correction is completed, the degree of dispersion is calculated based on the corrected sound energy intensity value set. The average value μ=3.2 of all grid intensity values ​​is first obtained, and then the standard deviation formula is used. The calculated value is σ=0.83, where N=400 is the total number of grids.

[0073] In the embodiment of the present application, this method effectively improves the accuracy of the assessment of the discrete degree of sound energy intensity by combining outlier correction and standard deviation calculation, eliminates the influence of local interference on cluster analysis, lays a data foundation for the reliable generation of subsequent sealing abnormality indicators, and enhances the stability and anti-interference ability of the entire monitoring system.

[0074] In order to further improve the accuracy of cluster division, in some embodiments, step 202: dividing the initial cluster into two sub-clusters includes: Step 401: Constructing a spatial gradient field of sound energy intensity values ​​of the initial clusters.

[0075] In step 401, the acoustic energy intensity spatial gradient field refers to a vector field reflecting the spatial change rate of the acoustic energy intensity at each position of the sealing ring surface. Each grid point corresponds to a gradient vector, which represents the change direction and amplitude of the acoustic energy intensity at that point in space.

[0076] In an embodiment of the present application, first, for all grid cells contained in the initial cluster, the sound energy intensity change rate between each cell and the eight adjacent cells is calculated, and the axial and radial intensity change components of the grid point are determined by the central difference method. These two components are combined to form the gradient vector of the point, and the gradient vectors of all grid points together constitute the spatial gradient field.

[0077] Step 402: Determine the gradient vector with the largest modulus in the acoustic energy intensity value spatial gradient field.

[0078] In step 402, the maximum modulus refers to the maximum value obtained by comparing the lengths (modulos) of all gradient vectors in the acoustic energy intensity density gradient field. The direction of this gradient vector represents the spatial location where the acoustic energy intensity density changes most dramatically. Its modulus is calculated by taking the square root of the sum of the partial derivatives of the acoustic energy intensity density with respect to the spatial coordinates. A gradient vector is the longest vector in the gradient field. Its direction represents the spatial location where the acoustic energy intensity changes most dramatically, and its modulus reflects the magnitude of that change.

[0079] In an embodiment of the present application, the gradient vectors of all grid points in the gradient field are traversed, the length of each vector is calculated, the vector with the largest length is found by comparison, and the direction angle and modulus value of the vector are recorded as the basis for subsequent segmentation.

[0080] Step 403: Establish a segmentation normal plane along the direction of the gradient vector.

[0081] In step 403, the segmentation normal plane refers to a virtual plane perpendicular to the direction of the maximum gradient vector, which is used to divide the cluster into two parts. The position of the plane is determined by an optimization algorithm.

[0082] In the embodiment of the present application, an infinitely extending virtual plane is established along the direction of the maximum gradient vector, and the position of the plane is adjusted to minimize the difference in sound energy intensity within the subclusters on both sides of the plane, ensuring that the interiors of the two divided subclusters are as uniform as possible.

[0083] Step 404: Divide the initial cluster into two subclusters using the segmentation normal plane as a boundary.

[0084] In the embodiment of the present application, all grid cells are divided into two sets on both sides of the plane according to the determined position of the segmentation plane, and the statistical characteristics of the sound energy intensity of each set are calculated to verify whether its discreteness meets the requirements.

[0085] Here's a specific example: In the sealing performance test of the liquid drive compressor, based on the acoustic energy intensity data of 400 grid cells, when the calculated initial cluster standard deviation σ = 0.85 exceeds the threshold value of 0.8, the following division operation is performed: First, the acoustic energy intensity gradient field is constructed, and the central difference formula is used for each grid point. and Calculate the gradient component, where Δx=Δy=0.2 mm is the grid spacing, and E(x,y) represents the grid acoustic energy intensity value with coordinates (x,y); traverse all the grid point gradient vectors and find that the gradient modulus of the grid point with coordinates (12,18) is the largest, reaching 14.3, with a direction of axial deflection of 16 degrees; establish a segmentation normal plane along this direction, and determine through the optimization algorithm that the plane position should be 6.8 mm away from the center line of the sealing ring. At this time, the standard deviations of the acoustic energy intensity of the subclusters on both sides of the plane are 0.61 and 0.59 respectively; finally, divide the initial cluster into two subclusters along this plane, containing 165 and 235 grid units respectively. The segmentation position optimization adopts the criterion of minimizing the sum of the internal variances of the subclusters. The specific expression is: ,in and Represent the sound energy intensity values ​​of the two subclusters respectively, is the average value of the corresponding sub-cluster.

[0086] In the embodiment of the present application, the method achieves accurate division of clusters by constructing an acoustic energy intensity gradient field and optimizing the segmentation plane, ensuring the uniformity of acoustic energy distribution within the sub-clusters, providing a reliable basis for subsequent abnormal area identification, and effectively improving the accuracy of sealing performance detection.

[0087] To further improve the accuracy of establishing the pressure-flow covariance relationship, in some embodiments, step 102: establishing the pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data, includes: Step 501: Segment the high-frequency pressure data into a plurality of discrete pressure segments according to a single reciprocating motion cycle of a piston rod of a liquid-driven compressor.

[0088] In step 501, a single reciprocating motion cycle of the piston rod refers to the duration of the entire extension and retraction process of the piston rod of the liquid-driven compressor. This cycle is determined by the mechanical motion signal measured by the piston rod displacement sensor or the crankshaft angle encoder, and its duration is directly determined by the compressor speed. The relationship between the piston rod and the liquid-driven compressor: The piston rod is the core moving component of the liquid-driven compressor. One end of the piston rod is connected to the hydraulic drive system, and the other end drives the compressor piston to reciprocate in the sealed chamber, achieving gas compression through mechanical displacement. Its motion state directly determines the volume change and pressure fluctuation of the sealed chamber. Discrete pressure segments refer to time period data that are intercepted from the continuously collected high-frequency pressure data according to the complete reciprocating motion cycle of the piston rod. Each segment contains the pressure change characteristics of a complete motion cycle.

[0089] In an embodiment of the present application, the start and end time points of a single reciprocating motion are first determined by a piston rod displacement sensor, and the continuously collected high-frequency pressure data is divided into multiple data segments of equal length according to the time point. Each data segment corresponds to a complete piston motion cycle, providing basic time period data for subsequent feature extraction.

[0090] Step 502: extracting a pressure peak time series from each of the pressure segments.

[0091] In step 502 , the pressure peak time series refers to a set of extreme value points arranged in time order in each pressure segment, reflecting the key features of the pressure fluctuations within the cycle.

[0092] In an embodiment of the present application, for each segmented pressure segment data, all peak points are identified through an extreme value detection algorithm, and are arranged in chronological order to form a peak sequence, retaining the temporal characteristics of the pressure change.

[0093] Step 503: Synchronously dividing the circulating fluid flow data into a plurality of flow fluctuation segments.

[0094] In step 503, the flow fluctuation segment refers to a flow data segment segmented synchronously with the pressure segment, which represents the dynamic flow state of the fluid in the same motion cycle.

[0095] In an embodiment of the present application, the flow data collected by the flow sensor is synchronously intercepted according to the time nodes of the pressure data segmentation to ensure that the time range of each flow segment is completely consistent with the corresponding pressure segment.

[0096] Step 504: Calculate the instantaneous flow rate change rate of each flow fluctuation segment.

[0097] In step 504, the instantaneous flow rate change rate refers to the flow rate change speed between adjacent sampling points in the flow fluctuation segment, reflecting the dynamic response characteristics of the flow.

[0098] In an embodiment of the present application, for each continuous sampling point of a traffic segment, the difference between the traffic values ​​of two adjacent points is calculated and divided by the sampling time interval to obtain the instantaneous change rate at that time point, thereby forming a traffic change rate curve.

[0099] Step 505: establishing a dynamic correlation between the pressure peak time series and the instantaneous flow rate change rate within the same motion cycle.

[0100] In step 505, the same motion cycle means that the pressure data segments and flow data segments strictly correspond to the same mechanical motion phase of the piston rod (for example, both are extension cycles). Data segment matching is achieved by aligning the crankshaft angle signal or piston rod displacement signal in the time domain. The dynamic correlation relationship refers to the coordinated temporal variation of the pressure peak and the flow rate change rate, which represents the coupling characteristics of the fluid state in the sealed chamber.

[0101] In an embodiment of the present application, the pressure peak sequence time points within the same motion cycle are aligned with the flow rate change curve, the response pattern and delay time of the flow rate change rate after the peak appears are analyzed, and a timing correlation model between the two is established.

[0102] Step 506: Construct a pressure-flow covariance relationship based on the dynamic correlation relationship of N consecutive motion cycles, where N is greater than or equal to 3.

[0103] In step 506, the continuous determination must meet two conditions: (1) the cycle interval is equal to an integer multiple of the piston rod movement period; and (2) there is no missing data in adjacent cycles. The minimum number of cycles is determined by the compressor characteristics, and generally greater than or equal to 3 cycles is required to establish a valid statistical pattern.

[0104] In an embodiment of the present application, the pressure-flow dynamic correlation data of multiple motion cycles are continuously analyzed, and the consistency index of the correlation characteristics of each cycle is calculated. When the fluctuation of the index for three consecutive cycles is less than the set threshold, it is confirmed that the covariance relationship is stably established.

[0105] Here's a specific example: In the sealing performance test of a certain type of liquid-driven compressor, the specific implementation process of constructing the pressure-flow covariance relationship based on the collected synchronous data is as follows: First, the high-frequency pressure data collected 10,000 times per second is divided into 120-millisecond movement cycles of the piston rod. In each cycle segment, 12 pressure peak points are extracted using the extreme value detection algorithm to form a time series sequence; the flow data is processed synchronously, and the instantaneous change rate of the flow value in the same time period is calculated at millisecond intervals. The formula is: ,in It represents the flow value at time t, and Δt=1 millisecond is the sampling interval; select the data of three consecutive movement cycles, match the pressure peak time in each cycle with the flow change rate at the corresponding time, and use the correlation coefficient formula The correlation between the two was calculated; the correlation coefficients of the three cycles were 0.81, 0.83 and 0.82, respectively, with an average of 0.82, indicating that the flow rate change rate began to rise synchronously about 2 milliseconds after the pressure peak appeared, and a stable pressure-flow covariance relationship model was established.

[0106] In the embodiment of the present application, the method accurately captures the coordinated change law of pressure and flow through time alignment and dynamic correlation analysis. The established covariation relationship can effectively reflect the changes in the fluid state of the sealing chamber, and provides a reliable basis for fluid dynamic characteristics for sealing performance evaluation.

[0107] To further improve the accuracy of friction sound signal processing, in some embodiments, step 103: performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information includes: Step 601: spatially and selectively enhance the acoustic energy component of the friction acoustic signal through acoustic wave guidance.

[0108] In step 601, the acoustic wave guide refers to an acoustic guiding device with a specific geometric structure that selectively enhances the acoustic wave signal component in the direction normal to the sealing interface and suppresses interfering noise in other directions. The acoustic energy component is derived from the friction sound signal, which is spatially selectively enhanced by the acoustic wave guide structure. This energy component is obtained by directional filtering the original acoustic signal using the physical guiding structure.

[0109] In an embodiment of the present application, a sound-conducting structure arranged in a ring array is installed around the sealing ring, and its internal cavity design focuses and enhances the sound waves in the vertical direction of the sealing interface, while the obliquely incident sound waves are attenuated, thereby improving the signal-to-noise ratio of the effective signal.

[0110] Step 602: Perform frequency domain analysis on the enhanced acoustic energy component to identify a specific frequency band range associated with high-frequency leakage.

[0111] In step 602, frequency domain analysis involves converting the time-domain acoustic signal into a frequency distribution representation, identifying characteristic frequency bands containing leakage information through spectral features. High-frequency leakage refers to acoustic signal characteristics above 8 kHz, generated when a hydraulic compressor seal fails. The distinction is made as follows: In the context of seal performance monitoring, "high frequency" specifically refers to the 8 kHz-12 kHz frequency range (experimental measurements show that the energy in this frequency range increases by more than 6 dB compared to normal operating conditions during leakage), while "low frequency" refers to the frequency range below 1 kHz (primarily consisting of mechanical vibration noise). High-frequency leakage signals are generated by high-frequency acoustic waves generated by turbulence in micron-scale seal gaps, while low-frequency signals primarily originate from mechanical vibrations of components such as the piston rod. After Fourier transform separation of the two, only the high-frequency component is used for leak diagnosis. A specific frequency band refers to a continuous frequency range determined through frequency domain analysis to be highly correlated with seal leakage characteristics. Within this frequency band, the acoustic signal energy exceeds background noise and exhibits a specific spectral pattern, which is used to extract effective leakage signal components.

[0112] In an embodiment of the present application, the enhanced acoustic signal is subjected to a fast Fourier transform to generate a spectrum diagram, and the energy distribution of each frequency band is analyzed. Continuous frequency bands whose energy is higher than the background noise and matches the leakage characteristics are selected as the effective analysis range.

[0113] Step 603: Obtain sound energy intensity data within the specific frequency band.

[0114] In step 603, the sound energy intensity data refers to the quantitative representation of the sound wave energy in the characteristic frequency band, reflecting the intensity characteristics of the signal in a specific frequency range.

[0115] In an embodiment of the present application, the time domain waveform of the characteristic frequency band signal is extracted, and its root mean square value is calculated as the intensity index, which comprehensively reflects the energy of the sound wave in the frequency band.

[0116] Step 604: Map the acoustic energy intensity data to the sealing interface space coordinate system to form a two-dimensional acoustic energy intensity distribution map.

[0117] In step 604, the sealing interface spatial coordinates are established using the contact surface of the hydraulic compressor piston rod seal ring as a reference. The x-axis is along the circumference of the piston rod, the y-axis is along the axial direction, and the origin is fixed at the geometric center of the seal ring. This coordinate system is predefined based on the dimensions of the mechanical design drawings. The two-dimensional acoustic energy intensity distribution diagram is a visual representation of the energy distribution, created by mapping the acoustic energy intensity values ​​at each measurement point onto the two-dimensional plane of the seal ring according to the measurement location.

[0118] In the embodiment of the present application, according to the spatial position coordinates of each probe in the acoustic sensor array, the acoustic energy intensity values ​​measured at the corresponding points are marked on the sealing ring plane projection diagram, and a continuous two-dimensional distribution diagram is generated through an interpolation algorithm.

[0119] Step 605: Divide the two-dimensional sound energy intensity distribution map into a plurality of equally spaced spatial grids, calculate the cumulative value of the sound energy intensity in each spatial grid, and generate sound energy intensity distribution information based on the cumulative value of the sound energy intensity in each spatial grid.

[0120] In step 605, spatial gridding involves dividing the two-dimensional surface of the sealing ring into several regular small areas of equal area (such as squares or hexagons). Each grid cell corresponds to an independent data collection and analysis unit, which is used to quantify the spatial distribution characteristics of acoustic energy intensity. Acoustic energy intensity refers to the amount of acoustic energy passing through a unit area per unit time. Its value is derived from processing the square of the sound pressure signal in a specific frequency band (the high-frequency leakage characteristic band) and converted to a standard energy unit (μJ / mm²) using the acoustic sensor calibration coefficient. The cumulative value is the sum of the acoustic energy intensities at all sampling time points within each spatial grid cell, which is calculated by accumulating or integrating the total acoustic energy received by that grid area during the analysis period.

[0121] In the embodiment of the present application, the two-dimensional distribution map is divided into regular grids, and the intensity values ​​in each grid area are time-integrated and accumulated to obtain a gridded sound energy intensity data set, which is convenient for computer algorithm processing.

[0122] Here's a specific example: In the sealing performance test of a certain type of liquid-driven compressor, the specific implementation process of directional enhancement processing based on the collected acoustic signal data is as follows: first, the friction sound signal is spatially filtered through an 8-unit sound guide structure arranged in a ring. This structure increases the sound wave intensity in the normal direction of the sealing interface to 3 times that of the original signal, while attenuating the obliquely incident interference noise by more than 50%; the enhanced signal is spectrally analyzed, and the time domain signal is converted into frequency domain representation using fast Fourier transform. It is found that the signal energy in the frequency band of 8600 to 11400 Hz reaches 8 times that of the background noise. This frequency band range is determined by analyzing 100 sets of historical leakage data and can effectively characterize micro-leakage characteristics; the root mean square value of the signal in this frequency band is extracted as the intensity index, and the calculation formula is: , where M = 2000 is the number of sampling points within a 20 millisecond period; the intensity value of each sensor measurement point is mapped to the two-dimensional coordinate system of the sealing ring according to the installation position, and the sound energy intensity distribution map of the 40 mm × 40 mm area is generated by bilinear interpolation; the distribution map is divided into 6400 grid units of 0.5 mm × 0.5 mm, and the sound energy intensity of 60 milliseconds accumulated in three motion cycles in each unit is integrated and accumulated, where the accumulated value of the kth grid is , It represents the intensity value of the grid at the tth millisecond. The final generated sound energy intensity distribution information shows that there is a high-energy area at the 4 o'clock direction of the sealing ring, and its grid average cumulative value reaches 12.5 microjoules, which is 3.8 times that of other areas.

[0123] In the embodiment of the present application, the method effectively highlights the acoustic characteristics related to leakage through directional enhancement and characteristic frequency band extraction. The distribution information after gridding processing not only retains spatial details but also facilitates quantitative analysis, providing a high-quality data basis for accurately identifying sealing abnormal areas.

[0124] To further improve the comprehensiveness and accuracy of the sealing performance evaluation, in some embodiments, step 105: jointly analyzing the pressure-flow covariance relationship and the sealing anomaly index to generate a monitoring result characterizing the sealing performance of the liquid drive compressor includes: Step 701: Calculate the pressure-flow state scalar based on the pressure-flow covariance relationship.

[0125] In step 701, the pressure-flow state scalar refers to quantifying the dynamic covariance relationship between pressure and flow into a single numerical index, reflecting the comprehensive characteristics of the fluid state of the sealing system.

[0126] In an embodiment of the present application, key characteristic parameters in the pressure-flow covariance relationship, including correlation coefficient, response delay time, etc., are extracted. Through normalization processing and multi-parameter fusion algorithm, these characteristic parameters are converted into scalar values ​​ranging from 0 to 1. The larger the value, the higher the risk of abnormal fluid state.

[0127] Step 702: Convert the sealing anomaly index into an acoustic anomaly scalar.

[0128] In step 702, the acoustic anomaly scalar is to convert the sealing anomaly index into a numerical value of a unified dimension for collaborative analysis with the fluid state scalar.

[0129] In an embodiment of the present application, based on the statistical distribution of historical data, the sealing abnormality index is mapped to a scale of 0 to 100 points by percentile. The higher the score, the more serious the acoustic abnormality, which facilitates weighted calculation with fluid parameters.

[0130] Step 703: performing a weighted fusion calculation on the pressure flow state scalar and the acoustic anomaly scalar, and using the weighted fusion calculation result as a monitoring result characterizing the sealing performance of the liquid drive compressor.

[0131] In step 703, weighted fusion calculation refers to assigning different weights according to the reliability of different monitoring parameters and generating a comprehensive evaluation value through linear combination.

[0132] In an embodiment of the present application, the pressure flow state scalar and the acoustic anomaly scalar are linearly combined according to a preset weight coefficient, where the fluid state weight is slightly higher than the acoustic parameter, reflecting its higher sensitivity to early leakage. The final monitoring result value is used to trigger a graded warning.

[0133] Here's a specific example: In the sealing performance test of a certain type of liquid-driven compressor, based on the obtained analytical data, the specific implementation process of the joint evaluation is as follows: First, key parameters are extracted from the established pressure-flow covariance relationship, including the correlation coefficient of 0.82 and the peak response delay of 2 milliseconds, and then the formula is used to calculate the peak response delay. The pressure flow state scalar is calculated to get 78.4 points, where r is the correlation coefficient, Δt is the delay time, and τ=5 milliseconds is the maximum delay threshold allowed by the system; at the same time, the sealing abnormality index of 23% is linearly mapped through the formula Converted to an acoustic anomaly scalar of 73.5 points, where A is the current anomaly index value, and are the typical values ​​of historical normal and severe leakage respectively; finally, the two scalars are weightedly fused using the formula The calculated comprehensive monitoring value is 76.2 points, of which the weight coefficient is determined by analyzing 50 sets of experimental data, reflecting the characteristics that pressure and flow parameters are more sensitive to early leakage; when the monitoring values ​​of three consecutive movement cycles exceed the preset warning line of 75 points, an early warning signal is triggered. The warning value is determined based on the 99% confidence upper limit of the normal operation data of similar equipment for 1,000 hours.

[0134] In the embodiment of the present application, the method realizes the comprehensive evaluation of fluid dynamic characteristics and acoustic characteristics through multi-parameter weighted fusion, which not only utilizes the sensitivity of pressure flow parameters to early leakage, but also combines the ability of acoustic detection to locate tiny defects, thereby improving the accuracy and reliability of sealing status monitoring.

[0135] Figure 2 This is a schematic diagram of the structure of an intelligent monitoring system for the sealing performance of a liquid-driven compressor provided in an embodiment of the present application, such as Figure 2 As shown, the system includes: The acquisition module 21 is used to collect high-frequency pressure data of the sealed cavity of the liquid-driven compressor, circulating liquid flow data, and friction sound signals of the sealed interface.

[0136] The construction module 22 is used to construct a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data.

[0137] The enhancement module 23 is used to perform directional enhancement processing on the friction sound signal to form sound energy intensity distribution information.

[0138] The generating module 24 is configured to perform adaptive clustering processing on the acoustic energy intensity distribution information using the DIANA algorithm to generate a sealing abnormality index.

[0139] The trigger module 25 is used to jointly analyze the pressure-flow covariance relationship and the sealing abnormality index to generate a monitoring result representing the sealing performance of the liquid-driven compressor. When the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

[0140] Figure 2 The intelligent monitoring system for sealing performance of a liquid-driven compressor can be executed Figure 1 The implementation principle and technical effects of the intelligent monitoring method for the sealing performance of a liquid-driven compressor described in the illustrated embodiment will not be elaborated on here. The specific manner in which the various modules and units of the intelligent monitoring system for the sealing performance of a liquid-driven compressor in the above embodiment perform their operations has been described in detail in the embodiments of the method and will not be elaborated on here.

[0141] In one possible design, Figure 2 The intelligent monitoring system for sealing performance of a liquid-driven compressor of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0142] The processing component 32 is used to perform the above Figure 1 The embodiment provides an intelligent monitoring method for the sealing performance of a liquid-driven compressor.

[0143] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0144] The storage component 31 is configured to store various types of data to support operations on the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random-access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0145] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0146] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0147] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0148] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0149] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment provides an intelligent monitoring method for the sealing performance of a liquid-driven compressor.

[0150] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0151] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0152] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent monitoring method for the sealing performance of a liquid-driven compressor, characterized in that: include: Collect high-frequency pressure data, circulating liquid flow data, and friction sound signals of the sealing interface of the liquid-driven compressor seal cavity; Constructing a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data; performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information; Adopting the DIANA algorithm to perform adaptive clustering processing on the sound energy intensity distribution information to generate a sealing abnormality index; The pressure-flow covariance relationship and the sealing abnormality index are jointly analyzed to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor. When the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

2. The method according to claim 1, characterized in that The DIANA algorithm is used to perform adaptive clustering processing on the sound energy intensity distribution information to generate a sealing abnormality index, including: The spatial grid of the sealing interface is used as the sample point to construct the initial cluster; Calculating the degree of dispersion of the sound energy intensity values ​​within the initial cluster, and when the degree of dispersion of the sound energy intensity values ​​within the initial cluster is greater than a preset split threshold, dividing the initial cluster into two subclusters; Taking each of the subclusters as a corresponding intermediate cluster, iteratively performing calculation operations and splitting operations until the discrete degree of the sound energy intensity values ​​in all the intermediate clusters is less than or equal to a preset splitting threshold, thereby obtaining a target cluster; The number of spatial grids in the target cluster whose sound energy intensity values ​​exceed the preset abnormal boundary is counted to generate a sealing abnormality index.

3. The method according to claim 2, characterized in that The calculating the discrete degree of the sound energy intensity values ​​within the initial clusters includes: Based on the sound energy intensity values ​​corresponding to all spatial grids in the initial cluster, forming a sound energy intensity value set; performing outlier correction processing on the sound energy intensity value set; Based on the corrected sound energy intensity value set, a standard deviation calculation formula is used to determine the degree of dispersion of the sound energy intensity values ​​within the initial cluster.

4. The method according to claim 2, characterized in that The step of dividing the initial cluster into two subclusters comprises: Constructing a spatial gradient field of sound energy intensity values ​​of the initial clusters; Determining a gradient vector with a maximum modulus value in the acoustic energy intensity value spatial gradient field; Establishing a segmentation normal plane along the direction of the gradient vector; The initial cluster is divided into two subclusters using the segmentation normal plane as a boundary.

5. The method according to claim 1, characterized in that The constructing of a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data includes: Dividing the high-frequency pressure data into a plurality of discrete pressure segments according to a single reciprocating motion cycle of a piston rod of a liquid-driven compressor; Extracting a pressure peak time series sequence from each of the pressure segments; Synchronously dividing the circulating fluid flow data into a plurality of flow fluctuation segments; Calculate the instantaneous flow rate change rate of each flow fluctuation segment; Establishing a dynamic correlation between the pressure peak time series and the instantaneous flow rate change rate within the same motion cycle; Based on the dynamic correlation relationship of N consecutive motion cycles, a pressure-flow covariance relationship is constructed, where N is greater than or equal to 3.

6. The method according to claim 1, characterized in that The performing directional enhancement processing on the friction sound signal to form sound energy intensity distribution information includes: spatially selectively enhancing the acoustic energy component of the friction acoustic signal by acoustic wave guidance; Perform frequency domain analysis on the enhanced acoustic energy components to identify specific frequency bands associated with high-frequency leakage; obtaining sound energy intensity data within the specific frequency band; Mapping the acoustic energy intensity data to a sealing interface spatial coordinate system to form a two-dimensional acoustic energy intensity distribution map; The two-dimensional sound energy intensity distribution diagram is divided into a plurality of equally spaced spatial grids, the cumulative value of the sound energy intensity in each spatial grid is calculated, and the sound energy intensity distribution information is generated based on the cumulative value of the sound energy intensity in each spatial grid.

7. The method according to claim 1, characterized in that The combined analysis of the pressure-flow covariance relationship and the sealing abnormality index to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor includes: Calculating the pressure-flow state scalar based on the pressure-flow covariance relationship; converting the sealing anomaly indicator into an acoustic anomaly scalar; A weighted fusion calculation is performed on the pressure flow state scalar and the acoustic anomaly scalar, and the weighted fusion calculation result is used as a monitoring result characterizing the sealing performance of the liquid drive compressor.

8. An intelligent monitoring system for the sealing performance of a liquid-driven compressor, characterized in that: include: The acquisition module is used to collect high-frequency pressure data of the liquid-driven compressor sealing chamber, circulating liquid flow data, and friction sound signals of the sealing interface; A construction module, configured to construct a pressure-flow covariance relationship based on the high-frequency pressure data and the circulating fluid flow data; an enhancement module, configured to perform directional enhancement processing on the friction sound signal to form sound energy intensity distribution information; A generation module, configured to perform adaptive clustering processing on the sound energy intensity distribution information using a DIANA algorithm to generate a sealing abnormality index; The trigger module is used to jointly analyze the pressure-flow covariance relationship and the sealing abnormality index to generate a monitoring result characterizing the sealing performance of the liquid-driven compressor. When the monitoring result exceeds a preset warning value, a sealing failure monitoring alarm is triggered.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an intelligent monitoring method for the sealing performance of a liquid-driven compressor as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, an intelligent monitoring method for the sealing performance of a liquid-driven compressor as claimed in any one of claims 1 to 7 is implemented.

Images