A leak detection and location system for heat exchange stations based on pressure fluctuation patterns

Abstract

This invention discloses a leak detection and location system for heat exchange stations based on pressure fluctuation patterns, specifically relating to the field of monitoring and location. It includes a dynamic site selection module: determining the detection point location through multi-pipeline parameter collaborative matching; an adaptive operating condition module: dynamically adjusting the sampling strategy according to the heat exchange station's operating load; a leak screening module: establishing a leak judgment threshold by combining historical data and real-time operating conditions; a propagation velocity correction module: analyzing the pressure wave propagation velocity by comprehensively considering the characteristics of the medium and pipeline; a dual-constraint location module: achieving leak point location by combining time difference and eigenvalue dual constraints; and a closed-loop verification correction module: achieving digital and intelligent management by verifying the location results through theoretical verification and experimental verification. This invention, through three-dimensional coupled feature extraction and multi-operating-condition threshold dynamic correction methods, avoids the limitations of single-parameter judgment, improves the accuracy of leak judgment, and provides a scientific basis for operation and maintenance decisions.

Description

Technical Field

[0001] This invention relates to the field of monitoring and positioning technology, and more specifically, to a leak detection and positioning system for heat exchange stations based on pressure fluctuation patterns. Background Technology

[0002] Existing leak detection technologies for heat exchange stations mainly include static pressure monitoring, ultrasonic detection, and infrared thermal imaging. Their core ideas are mostly based on sudden changes in physical quantities caused by leaks, and they determine whether a leak has occurred by monitoring phenomena such as a drop in absolute pressure, abnormal sound signals, or temperature field distortion.

[0003] The advantages of existing technologies lie in their low operational threshold and the fact that some methods can achieve non-contact detection, making them suitable for rapid screening of large areas. However, they also have significant shortcomings: static pressure monitoring methods rely solely on changes in average pressure and cannot distinguish between fluctuations caused by leaks and normal fluctuations caused by medium flow and temperature changes, with positioning errors typically exceeding 10 meters; ultrasonic detection methods are easily affected by pipeline noise and ambient sound waves, and have a low recognition rate for small leaks; infrared thermal imaging methods are greatly affected by ambient temperature, with a significant decrease in detection sensitivity in low-temperature winter environments, and cannot penetrate insulation layers to detect internal leaks. Consequently, they are unable to provide accurate data support for heat exchange station operation and maintenance decisions, fail to meet the real-time, accurate, and decision-making support requirements of intelligent management for leak detection, and struggle to guarantee operation and maintenance costs, thus increasing the computational costs for relevant personnel.

[0004] To address the shortcomings of existing technologies, such as low positioning accuracy, weak anti-interference capabilities, and poor identification of minor leaks, a leak detection and location system for heat exchange stations based on pressure fluctuation patterns is needed. This system should combine multi-source data collaborative analysis and intelligent decision support technologies to achieve accurate leak detection and location while providing data support for optimizing the allocation of heat exchange station operation and maintenance resources and management decisions, thereby promoting the deep integration of leak detection technology and digital operation and maintenance management. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a leak detection and location system for heat exchange stations based on pressure fluctuation patterns, which solves the problems mentioned in the background art through the following solutions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a leak detection and location system for heat exchange stations based on pressure fluctuation patterns, comprising: Dynamic location module: Determines the location of detection points through collaborative matching of multiple pipeline parameters; Coupled calibration module: Establishes an associated benchmark library and constructs a coupled model; Operating condition adaptive module: dynamically adjusts the sampling strategy according to the operating load of the heat exchange station; Hierarchical suppression module: Identifies noise types and matches differentiated filtering strategies; Coupling Extraction Module: Extracts 3D coupling features and simultaneously constructs a feature synthesis model; Leakage screening module: Establishes leakage judgment thresholds by combining historical data and real-time operating conditions; Propagation speed correction module: Analyzes pressure wave propagation speed based on comprehensive medium and pipeline characteristics; Dual-constraint localization module: combines time difference and feature value constraints to locate the leak point; Closed-loop verification and correction module: Digital and intelligent management is achieved through theoretical verification and experimental verification of positioning results.

[0007] Preferably, the specific method for determining the location of the detection point is as follows: extracting the core parameters of the pipeline segment to be detected, establishing multi-parameter location constraints to screen candidate pipeline segments, dynamically allocating the spacing between detection points based on the characteristics of the candidate pipeline segments, and verifying the effectiveness of determining the final detection point; the core parameters include the pipeline inner diameter gradient. Turbulence risk coefficient and the uniformity of medium flow rate The location constraints are set according to the pipe material and the type of medium. , , The constraint threshold is used to filter candidate pipe segments that meet all thresholds; the method for allocating the detection point spacing is as follows: based on the length of the candidate pipe segment. Calculate the distance between adjacent detection points according to the detection point distance model. , where k represents the spacing adjustment coefficient that is positively correlated with the turbulence risk coefficient.

[0008] Preferably, the associated benchmark library selects commonly used media in heat exchange stations, including standard temperatures under standard environmental conditions. and standard pressure The dynamic viscosity of each medium was measured using a rotational viscometer. And the type of medium and dynamic viscosity Standard temperature Standard pressure Corresponding storage; the coupling model includes a viscosity correction model and a temperature effect model, wherein the viscosity correction model is used to correct the viscosity parameters of the medium; the method for correcting the viscosity parameters of the medium is: collecting real-time temperature data at each detection point. Dynamic viscosity and rated pressure , dynamic viscosity Rated pressure and standard pressure Importing the viscosity correction model yields the corrected dynamic viscosity. Where c represents the pressure correction coefficient matching the type of medium; the temperature effect model is used to calculate the temperature effect coefficient, and the corrected dynamic viscosity is... Dynamic viscosity Standard temperature Real-time temperature Import the temperature effect model to obtain the temperature effect coefficient. , where d represents the thermal sensitivity coefficient of the medium.

[0009] Preferably, the sampling strategy includes classifying load levels, determining an initial sampling scheme, implementing data segmentation, and verifying signal integrity. The load level classification method involves calculating the load change per unit time relative to the rated load to obtain the fluctuation degree, and classifying the load into low, medium, and high levels based on the fluctuation degree. The initial sampling scheme determination method involves calling the sampling frequency and duration from the associated benchmark library that match the current load level and fluctuation degree, where low load is matched with a low-frequency, long-duration sampling combination, medium load with a standard sampling combination, and high load with a high-frequency, short-duration sampling combination. The data segmentation method involves dividing the collected data into segments according to a fixed time window, calculating the fluctuation variance of each segment and setting a fluctuation variance judgment threshold, marking segments with fluctuation variance exceeding the threshold as high-fluctuation segments, extracting the pressure extremes and fluctuation frequencies of high-fluctuation segments separately, comparing them with historical data of normal fluctuation segments under the same operating conditions, determining whether the fluctuation is caused by leakage, and issuing a warning signal. The signal integrity verification method involves defining a pressure time series. For the i-th detection point, the continuous pressure data set within the acquisition time t is given. The preset threshold for the missing rate of each data segment is 0.5%. If the missing rate of a certain data segment exceeds the threshold, the resampling mechanism is triggered: for the time interval corresponding to the segment, the original sampling parameters are maintained, the supplementary sampling mode of the detection point sensor is started, and the missing segment is replaced after supplementary sampling. If the missing rate is still not met after 3 supplementary samplings, the sampling priority of the detection point is increased, the sampling interval is shortened, and the entire data segment is re-acquired.

[0010] Preferably, the method for identifying noise types is as follows: frequency domain decomposition of the acquired pressure signal to distinguish between pipeline vibration noise, sensor noise, and environmental interference noise in real time; the filtering strategy is as follows: adaptive notch filtering is used for pipeline vibration noise, low-pass filtering is used for sensor noise, and improved wavelet threshold filtering is used for environmental interference noise to obtain the filtered preprocessed pressure time series. ,in Represents the wavelet decomposition coefficients. This represents the dynamically calculated layer threshold. The adjustment factor is indicated to match the noise intensity. The method for verifying the filtering effect is to calculate the signal-to-noise ratio (SNR) of the filtered signal; otherwise, adjust the filtering parameters and reprocess. The method for verifying the filtering effect is to set a SNR threshold that matches the detection accuracy requirements, calculate the SNR of the filtered signal, and if the threshold is not reached, adjust the parameters of the corresponding filtering method by notch frequency and wavelet decomposition level, and re-filter until the target is met.

[0011] Preferably, the three-dimensional coupling features include fluctuation period, fluctuation amplitude, and pressure change rate; the fluctuation period The extraction method is as follows: preprocessing the pressure time series An FFT transform is performed to obtain the frequency domain amplitude spectrum. The frequencies corresponding to the top three amplitude peaks are identified, and the periods are calculated. Period stability is verified through period stability testing. If the period is below a stability threshold, the average period value is taken; otherwise, the period corresponding to the frequency with the largest amplitude value is taken. The fluctuation amplitude... The extraction method is as follows: calculate the preprocessed pressure time series. With rated pressure deviation sequence A sliding window is used to extract local maxima, and the average of multiple local maxima is taken as the fluctuation amplitude. The window length is a multiple of the period's dynamic coefficient, which is adjusted accordingly. The pressure change rate... The extraction method is as follows: the central difference method is used to calculate the preprocessed pressure time series. The first derivative of the signal is used to extract the maximum absolute value of the derivative as the rate of pressure change, and this is verified. and correlation If the correlation is less than 0.6, the sliding window length is adjusted and parameters are extracted again; the feature synthesis model is used to calculate the comprehensive value of pressure fluctuation features. Quantify the correlation of leakage.

[0012] Preferably, the method for establishing the leakage judgment threshold includes constructing a historical operating condition threshold database, correcting the leakage judgment threshold, optimizing the threshold parameters, and determining leakage and screening core detection points; the historical operating condition threshold database: collects normal operating data under low, medium, and high load conditions over the past year, and calculates the average value of the comprehensive pressure fluctuation characteristics of each detection point according to the operating condition classification. and standard deviation Establish basic thresholds Working condition type, , , A historical threshold database is formed by corresponding storage, where j=1, 2, 3 correspond to three types of load, and k represents a coefficient matching the risk level of the operating condition; the method for correcting the leakage judgment threshold is as follows: based on the current load level, the corresponding basic threshold is called, and then... Correction, among which This represents the comprehensive pressure fluctuation characteristic value at the i-th detection point under normal operating conditions, where Q represents the real-time load. For the rated load; the method for optimizing the threshold parameter is as follows: historical data is divided into training and testing sets through cross-validation, the threshold misjudgment rate is calculated, and if the misjudgment rate exceeds the preset value, the k value is adjusted until the requirements are met; the method for determining leakage and screening core detection points is as follows: when the comprehensive value of the pressure fluctuation characteristics of the detection point is... If a leakage risk is identified, the three largest detection points are selected as core detection points and marked as A, B, and C according to the pipeline route.

[0013] Preferably, the analysis of pressure wave propagation velocity includes basic velocity calculation, medium parameter correction, time difference weighted correction, and velocity verification; the basic velocity calculation method is as follows: calculating the basic velocity based on the pipe and medium characteristics: based on the pipe elastic modulus E and the medium density Calculate the foundation velocity of the pressure wave according to the elastic wave propagation theory. The aforementioned medium parameter correction is used to calculate the average viscosity of the medium at the core detection points, according to the formula... The base velocity is corrected, where c represents a coefficient that matches the viscosity characteristics of the medium. This represents the corrected base velocity; the time difference weighted correction is achieved by measuring the time difference between the peak pressure wave values ​​between core detection points, calculating the time difference variation coefficient, and if the variation coefficient exceeds a threshold, then a weighted average time difference is used, calculated according to the formula. Calculate the final velocity, where This indicates the total spacing between the core detection points. This represents the weighted average time difference; the speed verification method is to compare the deviation rate between the calculated final speed and the base speed, and if the deviation rate exceeds the accuracy threshold, it is corrected again.

[0014] Preferably, the leak point location includes leak area prediction, initial location calculation, feature value constraint correction, and region adaptation adjustment; the leak area prediction is used to calculate the ratio of the comprehensive value of pressure fluctuation features at the core detection point. = and ,in , , These represent the comprehensive pressure fluctuation characteristics of core monitoring points A, B, and C, respectively, and a ratio judgment threshold is set. Exceeding the threshold and If the value is below the threshold, the leak point is predicted to be between A and B. Exceeding the threshold and If the value is below the threshold, the leak is predicted to be between points B and C; otherwise, it is predicted to be near point B. The initial location calculation method is as follows: if the leak is predicted to be between points A and B, let the distance from the leak point to point A be x, and establish a system of time difference equations. , Solving the simultaneous equations yields the initial position. ,in This represents the absolute value of the difference between the time it takes for the pressure wave generated at the leak point to propagate to the core detection point A and the time it takes to propagate to the core detection point B. This represents the absolute value of the difference between the time it takes for the pressure wave generated at the leak point to propagate to core detection point A and the time it takes to propagate to core detection point C; the correction is used to calculate the feature weight coefficient. , , According to the formula Correcting the initial positioning yields the final position. ,in This represents the distance from the core monitoring point A to B. The method of regional adaptation adjustment is as follows: if it is predicted to be between B and C, the variable is replaced, the calculation is repeated, and the distance is converted; if it is predicted to be near point B, the positioning result is directly calculated based on the feature weight.

[0015] Preferably, the theoretical verification is based on the final location. Calculate the theoretical temperature influence coefficient at the leak point Then calculate the comprehensive value of theoretical pressure fluctuation characteristics. Finally, the deviation between the theoretical value and the actual value is calculated. If the theoretical verification threshold is exceeded, the sliding window length of feature extraction is adjusted and recalculated; the actual verification is used to adjust the sampling parameters to repeat sampling and positioning, obtain secondary positioning results, calculate the positioning deviation between the two, and if the deviation exceeds the actual verification threshold, optimize the propagation speed calculation parameters.

[0016] The technical effects and advantages of this invention are as follows: 1. This invention solves the problem of blind deployment of detection points in existing technologies by using a dynamic location method for detection points, ensuring the effectiveness of the collected signals and laying the foundation for digital operation and maintenance management; it also solves the problem of existing technologies being unable to distinguish between normal fluctuations and leakage fluctuations by using a multi-parameter coupling calibration and multi-scale noise hierarchical suppression method, achieving stable identification of minute leaks and meeting the precise requirements of intelligent management for leak detection. 2. This invention avoids the limitations of single-parameter judgment by using a three-dimensional coupled feature extraction and multi-condition threshold dynamic correction method, thereby improving the accuracy of leak detection and providing a scientific basis for operation and maintenance decisions. By combining multi-factor coupled correction of propagation speed with a dual-constraint positioning method, the positioning error is significantly reduced, solving the problem of low positioning accuracy in existing technologies. This provides accurate data support for operation and maintenance work order generation and emergency repair path planning, optimizes the allocation of operation and maintenance resources, significantly reduces operation and maintenance costs, improves the level of digital operation and maintenance management and safe operation capabilities of heat exchange stations, and also reduces the computational costs for relevant personnel. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0018] 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.

[0019] As attached Figure 1 The leak detection and location system for heat exchange stations based on pressure fluctuation patterns, as shown, includes: Dynamic location module: Determines the location of detection points through collaborative matching of multiple pipeline parameters; In this embodiment, the specific method for determining the location of the detection point is as follows: core parameters of the pipe segment to be detected are extracted, multi-parameter location constraints are established to screen candidate pipe segments, the spacing between detection points is dynamically allocated based on the characteristics of the candidate pipe segments, and the effectiveness of determining the final detection point is verified. The core parameters include the pipe inner diameter gradient. That is, the ratio of the difference in inner diameter of adjacent pipe sections to the distance, and the turbulence risk factor. This refers to the ratio of the number of valves / elbows to the pipe length, and the uniformity of the medium flow velocity. This refers to the difference between the maximum and minimum flow velocities within the pipe section; the location constraints are set based on the pipe material and media type. , , The constraint threshold is used to filter candidate pipe segments that meet all thresholds; the method for allocating the detection point spacing is as follows: based on the length of the candidate pipe segment. Calculate the distance between adjacent detection points according to the detection point distance model. Where k represents the spacing adjustment coefficient that is positively correlated with the turbulence risk coefficient; the validity is confirmed by calculating the interference shielding coefficient of each candidate detection point according to the interference shielding model. Where a and b represent weighting coefficients that match the purpose of the pipeline. If the interference shielding coefficient reaches the preset effectiveness threshold, it is determined as the final detection point; otherwise, the spacing is readjusted until the requirements are met.

[0020] Coupled calibration module: Establishes an associated benchmark library and constructs a coupled model; In this embodiment, it should be specifically noted that: the associated reference library: selects commonly used media in heat exchange stations (such as water, heat transfer oil, etc.), under standard environmental conditions including standard temperature. and standard pressure The dynamic viscosity of each medium was measured using a rotational viscometer. And the type of medium and dynamic viscosity Standard temperature Standard pressure Corresponding storage; the coupling model includes a viscosity correction model and a temperature effect model, wherein the viscosity correction model is used to correct the viscosity parameters of the medium; the method for correcting the viscosity parameters of the medium is: collecting real-time temperature data at each detection point. Dynamic viscosity and rated pressure , dynamic viscosity Rated pressure and standard pressure Importing the viscosity correction model yields the corrected dynamic viscosity. Where c represents the pressure correction coefficient matching the type of medium; the temperature effect model is used to calculate the temperature effect coefficient, and the corrected dynamic viscosity is... Dynamic viscosity Standard temperature Real-time temperature Import the temperature effect model to obtain the temperature effect coefficient. , where d represents the thermal sensitivity coefficient of the medium.

[0021] Operating condition adaptive module: dynamically adjusts the sampling strategy according to the operating load of the heat exchange station; In this embodiment, the sampling strategy specifically includes: classifying load levels, determining an initial sampling scheme, implementing data segmentation, and verifying signal integrity. The load level classification method involves calculating the load change per unit time relative to the rated load to obtain the fluctuation, and classifying the load into low, medium, and high levels based on the fluctuation magnitude. The initial sampling scheme determination method involves calling the sampling frequency and duration from the associated benchmark library that match the current load level and fluctuation. For low loads, a low-frequency, long-duration sampling combination is matched; for medium loads, a standard sampling combination is matched; and for high loads, a high-frequency, short-duration sampling combination is matched. The core of the initial sampling scheme's data collection revolves around the core data and auxiliary calibration data required for subsequent leak analysis. Specifically, it is divided into two categories: core data and auxiliary data, and all collection actions are aimed at constructing a high-quality pressure time series. This supports subsequent analysis, specifically as follows: Core Acquisition Data: This is the core of the initial sampling scheme, directly used to construct the basic dataset for leak analysis, namely the real-time pressure values ​​of the medium inside the pipeline. During the acquisition process, pressure data of the medium inside the pipeline is continuously captured at a sampling frequency adapted to the current operating conditions. After these data are arranged in chronological order of acquisition, they form the core foundation for all subsequent analyses—the pressure time series. The fluctuation characteristics caused by leakage will all be reflected in this sequence. Auxiliary data acquisition data, while not directly involved in pressure sequence construction, ensures sampling quality and the accuracy of subsequent data correction, serving as a necessary supplement to the initial sampling scheme: Real-time operating load of the heat exchange station: such as heat load and flow load, used to determine the fluctuation level of the current operating condition, thereby confirming the suitability of the sampling frequency and duration in the initial sampling scheme and avoiding invalid sampling data due to misjudgment of the operating condition; Basic physical parameters of the medium: mainly the real-time temperature and dynamic viscosity of the medium at each detection point. This data is used for subsequent correction of the influence of temperature and pressure on viscosity, eliminating cross-interference, and ensuring the accuracy of the pressure-time series-based data. The extracted fluctuation characteristics more closely resemble actual leakage conditions, rather than spurious fluctuations caused by changes in the physical properties of the medium. Simply put, the initial sampling scheme involves first collecting real-time pressure values ​​to construct a core pressure time series. Then, operating load and medium physical parameters are collected to adapt the sampling strategy and prepare for subsequent data calibration. The data segmentation method is as follows: the collected data is segmented according to a fixed time window, the fluctuation variance of each segment is calculated, and a fluctuation variance judgment threshold is set. Segments with fluctuation variance exceeding the threshold are marked as high-fluctuation segments. The pressure extreme values ​​and fluctuation frequencies of the high-fluctuation segments are extracted separately and compared with historical data of normal fluctuation segments under the same operating conditions to determine whether the fluctuation is caused by leakage and to issue an early warning signal. The method for verifying signal integrity is as follows: a pressure time series is defined. For the i-th detection point, the continuous pressure data set within the acquisition time t is given. The preset threshold for the missing rate of each data segment is 0.5%. If the missing rate of a certain data segment exceeds the threshold, the resampling mechanism is triggered: for the time interval corresponding to the segment, the original sampling parameters are maintained, the supplementary sampling mode of the detection point sensor is started, and the missing segment is replaced after supplementary sampling. If the missing rate is still not met after 3 supplementary samplings, the sampling priority of the detection point is increased, the sampling interval is shortened, and the entire data segment is re-acquired.

[0022] Hierarchical suppression module: Identifies noise types and matches differentiated filtering strategies; In this embodiment, the specific method for identifying noise types is as follows: The collected pressure signal is decomposed in the frequency domain, and combined with pipeline vibration characteristics, sensor characteristics, and environmental interference spectrum. Specifically, the first step is to extract a 30-second continuous pressure signal segment from each detection point; the second step is to convert the time-domain signal to a frequency-domain signal using a fast Fourier transform; the third step is to retrieve a preset spectrum feature library (storing typical spectra of pipeline vibration noise 50-100Hz, sensor noise >1kHz, and environmental interference noise <10Hz); and the fourth step is to compare the converted frequency-domain signal with the feature library to match the noise type of the corresponding frequency band. The filtering strategy is as follows: adaptive notch filtering (notch frequency tracks the vibration main frequency) is used for pipeline vibration noise; low-pass filtering (cutoff frequency matches the upper limit of the sensor noise main frequency) is used for sensor noise; and improved wavelet threshold filtering is used for environmental interference noise to obtain the filtered preprocessed pressure time series. ,in Represents the wavelet decomposition coefficients. This represents the dynamically calculated layer threshold. The adjustment factor is indicated to match the noise intensity. The method for verifying the filtering effect is to calculate the signal-to-noise ratio (SNR) of the filtered signal; otherwise, adjust the filtering parameters and reprocess. The method for verifying the filtering effect is to set a SNR threshold that matches the detection accuracy requirements, calculate the SNR of the filtered signal, and if the threshold is not reached, adjust the parameters of the corresponding filtering method by notch frequency and wavelet decomposition level, and re-filter until the target is met.

[0023] Coupling Extraction Module: Extracts 3D coupling features and simultaneously constructs a feature synthesis model; In this embodiment, it should be specifically noted that the three-dimensional coupling features include the fluctuation period, fluctuation amplitude, and pressure change rate; the fluctuation period The extraction method is as follows: preprocessing the pressure time series Perform an FFT transform to obtain the frequency domain amplitude spectrum, identify the frequencies corresponding to the first three peak values ​​of the amplitude, and calculate the period. , , Periodic stability is obtained through periodic stability verification. If the amplitude is below the stability threshold, the average value of the period is taken; otherwise, the period corresponding to the frequency of the maximum amplitude is taken. The extraction method is as follows: calculate the preprocessed pressure time series. With rated pressure deviation sequence A sliding window is used to extract local maxima, and the average of multiple local maxima is taken as the fluctuation amplitude. The window length is a multiple of the period's dynamic coefficient, which is adjusted accordingly. The pressure change rate... The extraction method is as follows: the central difference method is used to calculate the preprocessed pressure time series. The first derivative of the signal is used to extract the maximum absolute value of the derivative as the rate of pressure change, and this is verified. and correlation If the correlation is less than 0.6, the sliding window length is adjusted and parameters are extracted again; the feature synthesis model is used to calculate the comprehensive value of pressure fluctuation features. Quantify the correlation of leakage.

[0024] Leakage screening module: Establishes leakage judgment thresholds by combining historical data and real-time operating conditions; In this embodiment, it should be specifically explained that: the method for establishing the leakage judgment threshold includes constructing a historical operating condition threshold library, correcting the leakage judgment threshold, optimizing the threshold parameters, and determining leakage and screening core detection points; the historical operating condition threshold library: collects normal operating data under low, medium, and high load conditions over the past year, and calculates the average value of the comprehensive pressure fluctuation characteristics of each detection point according to the operating condition classification. and standard deviation Establish basic thresholds Working condition type, , , A historical threshold database is formed by corresponding storage, where j=1, 2, 3 correspond to three types of load, and k represents a coefficient matching the risk level of the operating condition; the method for correcting the leakage judgment threshold is as follows: based on the current load level, the corresponding basic threshold is called, and then... Correction, among which This represents the comprehensive pressure fluctuation characteristic value at the i-th detection point under normal operating conditions, where Q represents the real-time load. For the rated load; the method for optimizing the threshold parameter is as follows: historical data is divided into training and testing sets through cross-validation, the threshold misjudgment rate is calculated, and if the misjudgment rate exceeds the preset value, the k value is adjusted until the requirements are met; the method for determining leakage and screening core detection points is as follows: when the comprehensive value of the pressure fluctuation characteristics of the detection point is... If a leakage risk is identified, the three largest detection points are selected as core detection points and marked as A, B, and C according to the pipeline route.

[0025] Propagation speed correction module: Analyzes pressure wave propagation speed based on comprehensive medium and pipeline characteristics; In this embodiment, it should be specifically noted that the analysis of pressure wave propagation velocity includes basic velocity calculation, medium parameter correction, time difference weighted correction, and velocity verification; the basic velocity calculation method is as follows: the basic velocity is calculated based on the pipe and medium characteristics: according to the pipe elastic modulus E and the medium density Calculate the foundation velocity of the pressure wave according to the elastic wave propagation theory. The aforementioned medium parameter correction is used to calculate the average viscosity of the medium at the core detection points, according to the formula... The base velocity is corrected, where c represents a coefficient that matches the viscosity characteristics of the medium. This represents the corrected base velocity; the time difference weighted correction is achieved by measuring the time difference between the peak pressure wave values ​​between core detection points, calculating the time difference variation coefficient, and if the variation coefficient exceeds a threshold, then a weighted average time difference is used, calculated according to the formula. Calculate the final velocity, where This indicates the total spacing between the core detection points. This represents the weighted average time difference; the speed verification method is to compare the deviation rate between the calculated final speed and the base speed, and if the deviation rate exceeds the accuracy threshold, it is corrected again.

[0026] Dual-constraint localization module: combines time difference and feature value constraints to locate the leak point; In this embodiment, it should be specifically noted that: the leak point location includes leak area prediction, initial location calculation, feature value constraint correction, and region adaptation adjustment; the leak area prediction is used to calculate the comprehensive value ratio of pressure fluctuation features at the core detection point. = and ,in , , These represent the comprehensive pressure fluctuation characteristics of core monitoring points A, B, and C, respectively, and a ratio judgment threshold is set. Exceeding the threshold and If the value is below the threshold, the leak point is predicted to be between A and B. Exceeding the threshold and If the value is below the threshold, the leak is predicted to be between points B and C; otherwise, it is predicted to be near point B. The initial location calculation method is as follows: if the leak is predicted to be between points A and B, let the distance from the leak point to point A be x, and establish a system of time difference equations. , Solving the simultaneous equations yields the initial position. ,in This represents the absolute value of the difference between the time it takes for the pressure fluctuation generated at the leak point to propagate to core detection point A and the time it takes to propagate to core detection point B. This represents the absolute value of the difference between the time it takes for the pressure wave generated at the leak point to propagate to core detection point A and the time it takes to propagate to core detection point C; the correction is used to calculate the feature weight coefficient. , , According to the formula Correcting the initial positioning yields the final position. ,in This represents the distance from the core monitoring point A to B. The method of regional adaptation adjustment is as follows: if it is predicted to be between B and C, the variable is replaced, the calculation is repeated, and the distance is converted; if it is predicted to be near point B, the positioning result is directly calculated based on the feature weight.

[0027] Closed-loop verification and correction module: Digital and intelligent management is achieved through theoretical verification and experimental verification of positioning results.

[0028] In this embodiment, it should be specifically noted that the theoretical verification is based on the final location. Calculate the theoretical temperature influence coefficient at the leak point Then calculate the comprehensive value of theoretical pressure fluctuation characteristics. Finally, the deviation between the theoretical value and the actual value is calculated. If the deviation exceeds the theoretical verification threshold, the sliding window length for feature extraction is adjusted and recalculated. The actual verification is used to adjust the sampling parameters (adapted to the current working conditions), repeat sampling and positioning to obtain a secondary positioning result, calculate the deviation between the two positioning attempts, and if the deviation exceeds the actual verification threshold, optimize the propagation speed calculation parameters. Specifically, the sampling parameters adapted to the current working conditions are increased by one level in frequency, and data is re-collected to obtain a secondary positioning result. The deviation between the two positioning attempts is calculated, with a preset deviation threshold of 0.5 meters. If the deviation exceeds the threshold, the medium correction coefficient for the propagation speed is optimized (adjustment range 0.25-0.35).

[0029] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A leak detection and location system for heat exchange stations based on pressure fluctuation patterns, characterized in that, include: Dynamic location module: Determines the location of detection points through collaborative matching of multiple pipeline parameters; Coupled calibration module: Establishes an associated benchmark library and constructs a coupled model; Operating condition adaptive module: dynamically adjusts the sampling strategy according to the operating load of the heat exchange station; Hierarchical suppression module: Identifies noise types and matches differentiated filtering strategies; The noise type identification method is as follows: frequency domain decomposition is performed on the acquired pressure signal to distinguish between pipeline vibration noise, sensor noise, and environmental interference noise in real time; the filtering strategy is as follows: adaptive notch filtering is used for pipeline vibration noise, low-pass filtering is used for sensor noise, and improved wavelet threshold filtering is used for environmental interference noise to obtain the filtered preprocessed pressure time series. ,in Represents the wavelet decomposition coefficients. This represents the dynamically calculated layer threshold. This represents the adjustment factor that matches the noise level. Let be the set of continuous pressure data of the i-th detection point within the acquisition time t; the method to verify the filtering effect is as follows: set a signal-to-noise ratio threshold that matches the detection accuracy requirements, calculate the signal-to-noise ratio of the filtered signal, and if the threshold is not reached, adjust the parameters of the corresponding filtering method by notch frequency and wavelet decomposition level, and re-filter until the standard is met. Coupling Extraction Module: Extracts 3D coupling features and simultaneously constructs a feature synthesis model; Leakage screening module: Establishes leakage detection thresholds by combining historical data with real-time operating conditions; Propagation speed correction module: Analyzes pressure wave propagation speed based on comprehensive medium and pipeline characteristics; Dual-constraint localization module: combines time difference and feature value constraints to locate the leak point; Closed-loop verification and correction module: Digital and intelligent management is achieved through theoretical verification and experimental verification of positioning results.

2. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 1, characterized in that: The specific method for determining the location of the detection points is as follows: extracting the core parameters of the pipeline segment to be detected, establishing multi-parameter location constraints to screen candidate pipeline segments, dynamically allocating the spacing between detection points based on the characteristics of the candidate pipeline segments, and verifying the effectiveness of determining the final detection points; the core parameters include the pipeline inner diameter gradient. Turbulence risk coefficient and the uniformity of medium flow rate The location constraints are set according to the pipe material and the type of medium. , , The constraint threshold is used to filter candidate pipe segments that meet all thresholds; the method for allocating the detection point spacing is as follows: based on the length of the candidate pipe segment. Calculate the distance between adjacent detection points according to the detection point distance model. , where k represents the spacing adjustment coefficient that is positively correlated with the turbulence risk coefficient.

3. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 1, characterized in that: The associated benchmark library: selects commonly used media in heat exchange stations, under standard environmental conditions including standard temperatures. and standard pressure The dynamic viscosity of each medium was measured using a rotational viscometer. And the type of medium and dynamic viscosity Standard temperature Standard pressure Corresponding storage; the coupling model includes a viscosity correction model and a temperature effect model, wherein the viscosity correction model is used to correct the viscosity parameters of the medium; the method for correcting the viscosity parameters of the medium is: collecting real-time temperature data at each detection point. Dynamic viscosity and rated pressure , dynamic viscosity Rated pressure and standard pressure Importing the viscosity correction model yields the corrected dynamic viscosity. Where c represents the pressure correction coefficient matching the type of medium; the temperature effect model is used to calculate the temperature effect coefficient, and the corrected dynamic viscosity is... Dynamic viscosity Standard temperature Real-time temperature Import the temperature influence model to obtain the temperature influence coefficient. , where d represents the thermal sensitivity coefficient of the medium.

4. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 1, characterized in that: The sampling strategy includes load level classification, initial sampling scheme determination, data segmentation, and signal integrity verification. Load level classification involves calculating the load variation per unit time relative to the rated load to obtain the fluctuation level, and then classifying the load into low, medium, and high levels based on the fluctuation magnitude. Initial sampling scheme determination involves calling sampling frequencies and durations from a relevant benchmark library that match the current load level and fluctuation level. Low-load conditions are matched with low-frequency, long-duration sampling combinations; medium-load conditions with standard sampling combinations; and high-load conditions with high-frequency, short-duration sampling combinations. Data segmentation involves dividing the collected data into segments based on fixed time windows, calculating the fluctuation variance of each segment, setting a fluctuation variance threshold, marking segments with fluctuation variance exceeding the threshold as high-fluctuation segments, extracting the pressure extremes and fluctuation frequencies of high-fluctuation segments separately, comparing them with historical data from normal fluctuation segments under the same operating conditions, determining whether the fluctuation is caused by leakage, and issuing a warning signal. Signal integrity verification involves defining a pressure time series. For the i-th detection point, the continuous pressure data set within the acquisition time t is given. The preset threshold for the missing rate of each data segment is 0.5%. If the missing rate of a certain data segment exceeds the threshold, the resampling mechanism is triggered: for the time interval corresponding to the segment, the original sampling parameters are maintained, the supplementary sampling mode of the detection point sensor is started, and the missing segment is replaced after supplementary sampling. If the missing rate is still not met after 3 supplementary samplings, the sampling priority of the detection point is increased, the sampling interval is shortened, and the entire data segment is re-acquired.

5. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 3, characterized in that: The three-dimensional coupling features include fluctuation period, fluctuation amplitude, and pressure change rate; the fluctuation period The extraction method is as follows: preprocessing the pressure time series An FFT transform is performed to obtain the frequency domain amplitude spectrum. The frequencies corresponding to the top three amplitude peaks are identified, and the periods are calculated. Period stability is verified through period stability testing. If the period is below a stability threshold, the average period value is taken; otherwise, the period corresponding to the frequency with the largest amplitude value is taken. The fluctuation amplitude... The extraction method is as follows: calculate the preprocessed pressure time series. With rated pressure deviation sequence A sliding window is used to extract local maxima, and the average of multiple local maxima is taken as the fluctuation amplitude. The window length is a multiple of the period's dynamic coefficient, which is adjusted accordingly. The pressure change rate... The extraction method is as follows: the central difference method is used to calculate the preprocessed pressure time series. The first derivative of the signal is used to extract the maximum absolute value of the derivative as the rate of pressure change, and this is verified. and correlation If the correlation is less than 0.6, adjust the sliding window length and re-extract the parameters; The feature comprehensive model is used to calculate a pressure fluctuation feature comprehensive value Quantify the leakage correlation degree.

6. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 2, characterized in that: The method for establishing leakage judgment thresholds includes constructing a historical operating condition threshold database, revising the leakage judgment thresholds, optimizing threshold parameters, and determining leaks and selecting core detection points. The historical operating condition threshold database involves collecting normal operating data from the past year under low, medium, and high load conditions, and calculating the average of the comprehensive pressure fluctuation characteristics of each detection point according to operating condition classification. and standard deviation Establish basic thresholds Working condition type, , , A historical threshold database is formed by corresponding storage, where j=1, 2, 3 correspond to three types of load, and k represents a coefficient matching the risk level of the operating condition; the method for correcting the leakage judgment threshold is as follows: based on the current load level, the corresponding basic threshold is called, and then... Correction, among which This represents the comprehensive pressure fluctuation characteristic value at the i-th detection point under normal operating conditions, where Q represents the real-time load. For the rated load; the method for optimizing the threshold parameter is as follows: historical data is divided into training and testing sets through cross-validation, the threshold misjudgment rate is calculated, and if the misjudgment rate exceeds the preset value, the k value is adjusted until the requirements are met; the method for determining leakage and screening core detection points is as follows: when the comprehensive value of the pressure fluctuation characteristics of the detection point is... If a leakage risk is identified, the three largest detection points are selected as core detection points and marked as A, B, and C according to the pipeline route.

7. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 1, characterized in that: The analysis of pressure wave propagation velocity includes basic velocity calculation, medium parameter correction, time difference weighted correction, and velocity verification. The basic velocity is calculated based on the pipe and medium characteristics: the basic velocity is calculated according to the pipe's elastic modulus E and the medium density. Calculate the foundation velocity of the pressure wave according to the elastic wave propagation theory. ; The media parameter correction is used to calculate the average viscosity of the media at the core detection points, according to the formula... The base velocity is corrected, where c represents a coefficient that matches the viscosity characteristics of the medium. The corrected base speed is indicated by the time difference weighted correction, which calculates the time difference variation coefficient by measuring the peak time difference of pressure waves between core detection points. If the variation coefficient exceeds the threshold, a weighted average time difference is used to calculate the final speed. The speed verification method is to compare the deviation rate between the calculated final speed and the base speed. If the deviation rate exceeds the accuracy threshold, the speed is corrected again.

8. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 1, characterized in that: The leak point location includes leak area prediction, initial location calculation, feature value constraint correction, and region adaptation adjustment; the leak area prediction is used to calculate the ratio of the comprehensive value of pressure fluctuation features at the core detection point. = and ,in , , These represent the comprehensive pressure fluctuation characteristics of core monitoring points A, B, and C, respectively, and a ratio judgment threshold is set. Exceeding the threshold and If the value is below the threshold, the leak point is predicted to be between A and B. Exceeding the threshold and If the value is below the threshold, the leak is predicted to be between points B and C; otherwise, it is predicted to be near point B. The initial location calculation method is as follows: if the leak is predicted to be between points A and B, let the distance from the leak point to point A be x, and establish a system of time difference equations. , Solving the simultaneous equations yields the initial position. ,in This represents the absolute value of the difference between the time it takes for the pressure fluctuation generated at the leak point to propagate to core detection point A and the time it takes to propagate to core detection point B. This represents the absolute value of the difference between the time it takes for the pressure wave generated at the leak point to propagate to core detection point A and the time it takes to propagate to core detection point C; the correction is used to calculate the feature weight coefficient. , , According to the formula Correcting the initial positioning to obtain the final position ,in This represents the distance from the core monitoring point A to B. The method of regional adaptation adjustment is as follows: if it is predicted to be between B and C, the variable is replaced, the calculation is repeated, and the distance is converted; if it is predicted to be near point B, the positioning result is directly calculated based on the feature weight.

9. The heat exchange station leak positioning system based on pressure fluctuation law according to claim 1, characterized in that: The theoretical verification is based on the final location. Calculate the theoretical temperature influence coefficient at the leak point Then calculate the comprehensive value of theoretical pressure fluctuation characteristics. Finally, the deviation between the theoretical value and the actual value is calculated. If the theoretical verification threshold is exceeded, the sliding window length of feature extraction is adjusted and recalculated; the actual verification is used to adjust the sampling parameters to repeat sampling and positioning, obtain secondary positioning results, calculate the positioning deviation between the two, and if the deviation exceeds the actual verification threshold, optimize the propagation speed calculation parameters.

Images