Method and device for detecting cleanliness of long-distance refrigeration pipeline

Abstract

The application discloses a cleaning degree detection method and device suitable for long-distance refrigeration pipelines, relates to the related technical field of cleaning degree detection, and comprises the following steps: determining the injection gas type and injection pulse signal of the long-distance pipeline to be detected; connecting a liquid nitrogen vaporization device, controlling a pulse injection valve to inject gas into an inlet through the injection pulse signal, collecting the temperature of the outer surface by using an infrared thermal imaging array, and outputting real-time response data of the outer surface temperature; constructing a heat transfer abnormal area identification model through a three-dimensional transient heat transfer equation of the pipeline and marking the abnormal area to obtain the heat transfer abnormal area; and analyzing the heat transfer abnormal response data of the heat transfer abnormal area to generate a cleaning degree detection result. The technical problems that the long-distance refrigeration pipeline cleaning degree detection efficiency is low, real-time dynamic monitoring cannot be realized, and local pollution areas are difficult to accurately identify in the prior art are solved, and the technical effects of accurately positioning the heat transfer abnormal area and efficiently and accurately detecting the cleaning degree of the pipeline are achieved.

Description

Technical Field

[0001] This application relates to the technical field of cleanliness testing, specifically to a method and apparatus for cleanliness testing of long-distance refrigeration pipelines. Background Technology

[0002] Long-distance refrigeration pipelines are widely used in liquefied natural gas (LNG) transportation, cold chain logistics, large-scale central air conditioning systems, and chemical refrigeration. However, during long-term operation, these pipelines easily accumulate contaminants such as oil, scale, and particulate matter, leading to decreased heat transfer efficiency, increased energy consumption, and even potential pipeline blockage or corrosion, severely impacting system safety and economic efficiency. Traditional pipeline cleanliness testing mainly includes visual inspection, endoscopic inspection, pressure drop analysis, and sampling and testing. These methods are inefficient, have limited applicability, and cannot provide real-time monitoring. Especially for long-distance, large-diameter refrigeration pipelines, comprehensive and rapid cleanliness assessment is difficult. Existing non-contact detection methods based on infrared thermal imaging rely heavily on steady-state temperature field analysis, which struggles to effectively identify localized contamination or deposit distribution under dynamic gas flow conditions, thus hindering rapid and accurate assessment of the cleanliness of long-distance refrigeration pipelines.

[0003] Therefore, current technologies suffer from technical problems such as low efficiency in detecting the cleanliness of long-distance refrigeration pipelines, inability to perform real-time dynamic monitoring, and difficulty in accurately identifying localized contamination areas. Summary of the Invention

[0004] This application provides a method and apparatus for detecting the cleanliness of long-distance refrigeration pipelines, which solves the technical problems of low efficiency in detecting the cleanliness of long-distance refrigeration pipelines, inability to monitor in real time, and difficulty in accurately identifying local contamination areas in the prior art. It achieves the technical effect of accurately locating abnormal heat transfer areas and high-efficiency and high-precision cleanliness detection of pipelines.

[0005] This application provides a cleanliness detection method for long-distance refrigeration pipelines. The method includes: determining the type of injected gas and the injection pulse signal for the long-distance pipeline under test; connecting a liquid nitrogen vaporization device to the inlet of the long-distance pipeline under test, wherein the liquid nitrogen vaporization device controls a pulse injection valve to inject gas into the inlet of the long-distance pipeline under test through the injected gas type and the injection pulse signal; simultaneously using a deployed infrared thermal imaging array to collect the temperature of the outer surface of the long-distance pipeline under test and outputting real-time response data of the outer surface temperature; constructing a heat transfer anomaly region identification model through the three-dimensional transient heat transfer equation of the pipeline; marking the anomaly regions in the real-time response data of the outer surface temperature according to the heat transfer anomaly region identification model to obtain the heat transfer anomaly regions; analyzing the heat transfer anomaly response data of the heat transfer anomaly regions to generate the cleanliness detection result of the long-distance pipeline under test.

[0006] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further performs the following processing: finite element modeling of the long-distance pipeline under test based on its basic parameters and operating environment information to obtain a three-dimensional pipeline structure model; simulation of the three-dimensional transient heat transfer equation of the pipeline structure model according to the injected gas type and injection pulse signal, outputting simulated response data of the outer surface temperature; construction of multiple abnormal long-distance pipeline modeling samples containing known abnormal state labels; simulation of the three-dimensional transient heat transfer equation of the pipeline structure model according to the multiple abnormal long-distance pipeline modeling samples, outputting multiple simulated response data of abnormal outer surface temperature; and temperature response feature training of the known abnormal state labels based on the simulated response data of the outer surface temperature and the multiple simulated response data of abnormal outer surface temperature, constructing a heat transfer abnormal area identification model.

[0007] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further performs the following processing: establishing control groups for the simulated response data of the outer surface temperature and multiple simulated response data of abnormal outer surface temperatures, respectively, to obtain multiple sets of temperature simulation response training data; extracting temperature response feature vectors from the multiple sets of temperature simulation response training data, wherein the temperature response feature vectors include temperature drop value, temperature drop rate, and temperature residual distribution; performing multi-class classification training on a defined random forest based on the temperature drop value, temperature drop rate, temperature residual distribution, and the known abnormal state labels to obtain a heat transfer abnormal area identification model; wherein the definition parameters of the random forest include the number of random forest trees, tree depth, segmentation sample trees, and leaf node sample trees.

[0008] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further includes the following processing: acquiring the temperature of the outer surface of the long-distance pipeline under test using a deployed infrared thermal imaging array, wherein the infrared thermal imaging array includes multiple infrared thermal imaging units; acquiring the temperature of the outer surface of the long-distance pipeline under test using the deployed infrared thermal imaging array, and outputting the image time series corresponding to the multiple infrared thermal imaging units; performing temperature mapping based on the temperature calibration curve built into each infrared thermal imaging unit, and outputting multiple time-series temperature response matrices characterizing the temperature change at each pixel location; and fusing the multiple time-series temperature response matrices to obtain real-time response data of the outer surface temperature of the long-distance pipeline under test.

[0009] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further performs the following processing: extracting temperature response feature vectors from the real-time response data of the outer surface temperature according to the heat transfer anomaly region identification model, and outputting the real-time temperature drop value, the real-time temperature drop rate, and the real-time temperature residual distribution; performing random forest classification identification based on the real-time temperature drop value, the real-time temperature drop rate, and the real-time temperature residual distribution to obtain the heat transfer anomaly region and the heat transfer anomaly response data of the heat transfer anomaly region.

[0010] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further performs the following processing: quantifying the heat transfer anomaly response data of the heat transfer anomaly area and outputting a comprehensive anomaly index, wherein the heat transfer anomaly response data includes the location, area, and anomaly level of the heat transfer anomaly area; and performing cleanliness conversion according to the comprehensive anomaly index to obtain the cleanliness detection result of the long-distance pipeline under test.

[0011] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further includes the following process: determining the type of injected gas for the long-distance pipeline to be tested, wherein the injected gas type is a fluorocarbon or CO2 gas with a boiling point below -40°C.

[0012] In a possible implementation, the cleanliness detection method for long-distance refrigeration pipelines further performs the following processing: obtaining the pipeline length and thermal response time constant of the long-distance pipeline to be tested; scoring the pipeline response effectiveness based on the pipeline length and thermal response time constant of the long-distance pipeline to be tested; optimizing the pulse signal frequency, amplitude, and duration period under the constraint of satisfying a preset response effectiveness scoring threshold; and generating an injection pulse signal.

[0013] This application also provides a cleanliness detection device suitable for long-distance refrigeration pipelines. The device includes: an injection data determination module for determining the type of injected gas and the injection pulse signal for the long-distance pipeline under test; a response data output module for connecting a liquid nitrogen vaporization device to the inlet of the long-distance pipeline under test, wherein the liquid nitrogen vaporization device controls a pulse injection valve to inject gas into the inlet of the long-distance pipeline under test through the injected gas type and the injection pulse signal, and simultaneously uses a deployed infrared thermal imaging array to collect the temperature of the outer surface of the long-distance pipeline under test and outputs real-time response data of the outer surface temperature; an abnormal area marking module for constructing a heat transfer abnormal area identification model through the three-dimensional transient heat transfer equation of the pipeline, marking the abnormal areas of the real-time response data of the outer surface temperature according to the heat transfer abnormal area identification model, and obtaining the heat transfer abnormal area; and a cleanliness detection result generation module for analyzing the heat transfer abnormal response data of the heat transfer abnormal area and generating the cleanliness detection result of the long-distance pipeline under test.

[0014] This application proposes a method and apparatus for cleanliness detection of long-distance refrigeration pipelines. The method determines the type of injected gas and the injection pulse signal for the pipeline. A liquid nitrogen vaporization device is connected, and the injection pulse signal controls the pulse injection valve to inject gas into the inlet. An infrared thermal imaging array is used to collect the temperature of the outer surface, outputting real-time response data. A heat transfer anomaly region identification model is constructed using the pipeline's three-dimensional transient heat transfer equation, and anomaly regions are marked to obtain the heat transfer anomaly regions. The heat transfer anomaly response data of these regions is analyzed to generate cleanliness detection results. This method solves the technical problems of low efficiency in long-distance refrigeration pipeline cleanliness detection, inability to perform real-time dynamic monitoring, and difficulty in accurately identifying localized contamination areas in existing technologies. It achieves the technical effect of accurately locating heat transfer anomaly regions and high-efficiency, high-precision cleanliness detection of pipelines. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the apparatus according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0016] Figure 1 This is a schematic flowchart of a cleanliness testing method for long-distance refrigeration pipelines provided in an embodiment of this application.

[0017] Figure 2This is a schematic diagram of the experimental cooling rate of multiple abnormal long-distance pipeline modeling samples in the cleanliness detection method for long-distance refrigeration pipelines provided in the embodiments of this application.

[0018] Figure 3 A schematic diagram of the cleanliness detection device for long-distance refrigeration pipelines provided in this application embodiment.

[0019] Figure labeling: Injection data determination module 10, response data output module 20, abnormal area marking module 30, cleanliness detection result generation module 40. Detailed Implementation

[0020] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or apparatuses. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0023] This application provides a method for cleanliness testing of long-distance refrigeration pipelines, such as... Figure 1 As shown, the method includes:

[0024] Step S100: Determine the type of injected gas and the injection pulse signal for the long-distance pipeline to be tested.

[0025] Furthermore, step S100 also includes determining the type of injected gas for the long-distance pipeline to be tested, wherein the type of injected gas is a fluorocarbon or CO2 gas with a boiling point below -40°C.

[0026] Preferably, the long-distance pipeline under test refers to a closed-loop transport pipeline that is significantly longer than conventional pipelines in the refrigeration, energy, or industrial fields and requires cleanliness testing. It is used to transport cryogenic media such as liquefied natural gas, liquid ammonia, and chilled water. The inner wall is prone to accumulating contaminants such as oil, frost, and scale. The injection gas type and injection pulse signal are determined. The injection gas type refers to the type of cryogenic gas generated by a liquid nitrogen vaporization device and pulsedly injected into the pipeline during the testing process. Specifically, the injection gas type is a fluorocarbon or CO2 gas with a boiling point below -40℃. After injection into the long-distance pipeline under test, it can rapidly vaporize and form a significant temperature gradient, thereby generating a detectable thermal response signal within the pipeline. The pulse signal is a combination of parameters controlling the time, frequency, amplitude, and duration of gas injection. Periodic gas injection is achieved by adjusting the pulse valve to stimulate the dynamic thermal response of the pipeline.

[0027] Step S100 further includes: step S110, obtaining the pipe length and thermal response time constant of the long-distance pipe to be tested; step S120, scoring the pipe response effectiveness based on the pipe length and thermal response time constant of the long-distance pipe to be tested, optimizing the pulse signal frequency, amplitude and duration period under the constraint of satisfying the preset response effectiveness scoring threshold, and generating an injection pulse signal.

[0028] Preferably, the length of the long-distance pipeline under test is measured on-site using a laser rangefinder or ultrasonic rangefinder, or directly read from design drawings. This directly affects the propagation time of the pulse signal and the thermal response range. The thermal response time constant is calculated through experimental measurement, reflecting the pipeline material's response speed to temperature changes, i.e., the characteristic parameter of the time required for temperature to transfer to the outer surface. Where ρ is the density of the pipe material (kg / m³) 3 c pis the specific heat capacity J / (kg·K), k is the thermal conductivity W / (m·K), and L is the wall thickness m of the long-distance pipe to be tested. The pipe response effectiveness score is calculated based on the pipe length and thermal response time constant, i.e., assessing the detectability of the pulse signal under the current pipe parameters, ensuring the temperature response signal meets the signal-to-noise ratio requirements and covers the entire pipe. A scoring function is constructed based on the pipe length and thermal response time constant, i.e., a weighted calculation of the ratio of the pipe length and thermal response time constant to the corresponding reference pipe parameters. A preset response effectiveness score threshold ≥1.0 is set, and the pulse signal frequency, amplitude, and duration are optimized under the constraint of meeting the preset response effectiveness score threshold. Here, frequency refers to the number of gas pulse triggers per unit time, amplitude refers to the gas volume injected in a single pulse, and duration refers to the duration of each pulse. That is, the optimal pulse signal frequency, amplitude, and duration are calculated based on the pipe length and thermal response time constant, ultimately generating the injected pulse signal. For example, for long pipes, the frequency needs to be reduced and the pulse width extended to ensure the temperature wave propagates to the far end.

[0029] Step S200: A liquid nitrogen vaporization device is connected to the inlet of the long-distance pipeline to be tested. The liquid nitrogen vaporization device controls the pulse injection valve to inject gas into the inlet of the long-distance pipeline to be tested through the type of injected gas and the injection pulse signal. At the same time, an infrared thermal imaging array is deployed to collect the temperature of the outer surface of the long-distance pipeline to be tested and output the real-time response data of the outer surface temperature.

[0030] Preferably, a liquid nitrogen vaporization device is connected to the inlet of the long-distance pipeline to be tested. Its core components include a liquid nitrogen storage tank, a vaporizer, a pulse injection valve, and a control system. This device converts liquid nitrogen into cryogenic gaseous nitrogen, which is then injected into the long-distance pipeline according to set parameters via the pulse valve. Specifically, a quick-connect interface is installed at the pipeline inlet to connect to the output end of the liquid nitrogen vaporization device, ensuring a tight seal. A gas with a boiling point below -40°C is selected based on the pipeline characteristics to trigger a significant thermal response. The control system receives the optimized injection pulse signal and injects the gas through the pulse injection valve. The pulse injection valve opens and closes at a set rhythm to achieve intermittent cryogenic gas injection, creating controllable temperature fluctuations within the pipeline. This causes a characteristic delayed temperature drop in the fouled area due to differences in thermal resistance. For example, injection occurs every 2 seconds, lasting 100ms each time. The injected cryogenic gas rapidly absorbs heat within the pipeline, causing localized cooling of the inner wall and a corresponding change in the outer surface temperature. Simultaneously, an infrared thermal imaging array is deployed to collect the temperature of the outer surface of the long-distance pipeline under test. The infrared thermal imaging array consists of multiple infrared thermal imaging units deployed at equal intervals along the outer surface of the pipeline, covering the entire pipeline. The units are synchronized to ensure data time consistency. Each infrared thermal imaging unit captures thermal images of the outer surface of the pipeline at a fixed frequency, generates an image time series, and then converts it into actual temperature values ​​according to the built-in temperature calibration curve. Finally, the data from multiple infrared thermal imaging units are integrated into real-time response data of the outer surface temperature.

[0031] Furthermore, step S200 also includes step S210, which involves using a deployed infrared thermal imaging array to collect temperature data on the outer surface of the long-distance pipe under test, wherein the infrared thermal imaging array includes multiple infrared thermal imaging units; step S220, which involves using the deployed infrared thermal imaging array to collect temperature data on the outer surface of the long-distance pipe under test and outputting the image time series corresponding to the multiple infrared thermal imaging units; step S230, which involves performing temperature mapping based on the temperature measurement calibration curve built into each infrared thermal imaging unit and outputting multiple time-series temperature response matrices characterizing the temperature change at each pixel position; and step S240, which involves fusing the multiple time-series temperature response matrices to obtain the real-time response data of the outer surface temperature of the long-distance pipe under test.

[0032] Preferably, the infrared thermal imaging array comprises multiple infrared thermal imaging units, each of which is an independent infrared camera containing an infrared detector, optical lens, and signal processor. These units are installed at equal intervals along the outer surface of the pipe to ensure full coverage without blind spots. The deployed infrared thermal imaging array collects temperature data from the outer surface of the long-distance pipe under test; that is, each infrared thermal imaging unit captures thermal images at a fixed frequency, generating time-series thermal images, and then outputting the image time series corresponding to multiple infrared thermal imaging units. The temperature calibration curve refers to the calibration of each infrared thermal imaging unit before leaving the factory using a blackbody radiation source to establish a mapping relationship between pixel grayscale values ​​and temperature. Temperature mapping is performed based on the temperature calibration curve built into each infrared thermal imaging unit. The pixel values ​​of each frame of thermal image are converted into temperature values, generating a time-series temperature response matrix to characterize the temperature change at each pixel location. Finally, multiple time-series temperature response matrices are fused. That is, based on the installation position of the infrared thermal imaging unit, multiple time-series temperature response matrices are mapped to a unified three-dimensional coordinate system of the pipeline. The temperature values ​​of the overlapping areas are weighted and averaged. The weights are set based on the measurement distance. Finally, the real-time response data of the outer surface temperature of the long-distance pipeline under test is output, thus providing high-resolution and high-reliability temperature field data for pipeline cleanliness detection.

[0033] Step S300: Construct a heat transfer anomaly region identification model through the three-dimensional transient heat transfer equation of the pipeline, and mark the anomaly region of the real-time response data of the outer surface temperature according to the heat transfer anomaly region identification model to obtain the heat transfer anomaly region.

[0034] Step S300 further includes step S310, performing finite element modeling on the long-distance pipeline under test based on its basic parameters and operating environment information to obtain a three-dimensional pipeline structure model; step S320, simulating the three-dimensional transient heat transfer equation of the pipeline according to the injected gas type and injection pulse signal, and outputting simulated response data of the outer surface temperature; step S330, constructing multiple abnormal long-distance pipeline modeling samples containing known abnormal state labels; step S340, simulating the three-dimensional transient heat transfer equation of the pipeline according to the multiple abnormal long-distance pipeline modeling samples, and outputting multiple simulated response data of abnormal outer surface temperature; step S350, training the temperature response features of the known abnormal state labels based on the simulated response data of the outer surface temperature and the multiple simulated response data of abnormal outer surface temperature, and constructing a heat transfer abnormal area identification model.

[0035] Preferably, the basic parameters and operating environment information of the long-distance pipeline under test are obtained. The basic parameters include pipeline length, diameter, wall thickness, material density, and thermal conductivity, while the operating environment includes medium temperature, ambient temperature, and insulation layer thickness. Then, finite element modeling is performed on the long-distance pipeline under test using software such as ANSYS and COMSOL based on the basic parameters and operating environment information, outputting a three-dimensional pipeline structure model including material properties and boundary conditions. Next, the three-dimensional transient heat transfer equation of the pipeline is simulated according to the injected gas type and injection pulse signal. The expression of the three-dimensional transient heat transfer equation of the pipeline is as follows: Where ρ is the material density of the long-distance pipe to be tested, and c p ρ is the specific heat capacity, k is the thermal conductivity. Let T be the rate of change of temperature with time t, describing the transient change of temperature over time, where T is the temperature distribution. For heat conduction, It is a temperature gradient. It is a divergence operation, where q(x,y,z,t) is the heat flux term, representing the heat generated or consumed per unit volume per unit time by the pulsed gas injection. It can vary with spatial coordinates (x,y,z) and time t, and finally outputs simulated response data of the outer surface temperature, which may be simulated data of the outer surface temperature changing with time.

[0036] Preferably, based on historical pipeline fault data, multiple abnormal long-distance pipeline modeling samples are constructed, each containing known abnormal state labels. These known abnormal state labels include, but are not limited to, dirt or residue, scaling / deposition, lubricating oil film, oxide layer, and ice crystal layer. For example, dirt or residue reduces the thermal conductivity of the pipeline, scaling / deposition increases the pipeline wall thickness, lubricating oil film forms a nano-thin film that hinders heat transfer, oxide layer leads to a decrease in the thermal conductivity of the pipeline, and ice crystal layer causes the pipeline to exhibit latent heat of phase change (endothermic / exothermic). Then, the multiple abnormal long-distance pipeline modeling samples are mapped onto a three-dimensional pipeline structure model, and a three-dimensional transient heat transfer equation simulation of the pipeline is performed. That is, by solving the three-dimensional transient heat transfer equation of the pipeline, the dynamic temperature distribution of the outer surface of the pipeline after pulsed gas injection is simulated, and multiple simulated response data of abnormal outer surface temperatures are output, including temperature response data corresponding to various abnormalities (such as oil, scale, ice layer, etc.).

[0037] Preferably, the temperature response features of known abnormal state labels are trained based on simulated response data of external surface temperature and simulated response data of multiple external surface temperature anomalies. Specifically, using simulated temperature response data of normal and abnormal pipelines, features are extracted using machine learning algorithms, and a classification model is trained to automatically identify abnormal areas and their contamination types in the pipeline's external surface temperature data. Specifically, key features are extracted from the pipeline's external surface temperature response data, including temperature drop value, temperature drop rate, and temperature residual distribution. In abnormal areas, due to impaired heat transfer, the temperature drop value is usually smaller than in normal areas, and contaminants significantly reduce the cooling rate. The temperature residual distribution refers to the difference between actual data and simulated data of clean pipelines, reflecting local anomalies. Multiple simulated response data of external surface temperature anomalies are then mapped and associated with known abnormal state labels to form a supervised learning sample set. The supervised learning samples are used to train a random forest model for classification, constructing a heat transfer anomaly area identification model. This model can output the location of abnormal areas and the probability of anomaly types based on the pipeline's real-time infrared temperature data. An example of the experimental cooling rate is shown below. Figure 2 As shown, the horizontal axis represents time (0-10 seconds), i.e., the transient response stage after pulse injection, and the vertical axis represents the temperature drop rate. The clean pipe (Experimental clean) represents the ideal temperature drop state without dirt, the 50μm dirt layer (Experimental 50μm) simulates the temperature drop state of a slightly contaminated pipe, and the 200μm dirt layer (Experimental 200μm) simulates the temperature drop state of a severely contaminated pipe. Among them, the clean pipe has the highest heat transfer efficiency. The thermal resistance of the 50μm dirt layer increases by about 28%, and the thermal resistance of the 200μm dirt layer increases by up to 72%, reducing the cooling rate to 20% to 30% of that of the clean state.

[0038] Furthermore, step S350 also includes step S351, establishing control groups between the simulated response data of the outer surface temperature and multiple simulated response data of external surface temperature anomalies, to obtain multiple sets of temperature simulation response training data; step S352, extracting temperature response feature vectors from the multiple sets of temperature simulation response training data, wherein the temperature response feature vectors include temperature drop value, temperature drop rate, and temperature residual distribution; step S353, performing multi-class training on the defined random forest based on the temperature drop value, temperature drop rate, temperature residual distribution, and the known abnormal state labels, to obtain a heat transfer anomaly region identification model; wherein the definition parameters of the random forest include the number of random forest trees, tree depth, segmentation sample trees, and leaf node sample trees.

[0039] Preferably, control groups are established between the simulated response data of external surface temperature and multiple simulated response data of external surface temperature anomalies. That is, the simulated data of normal pipelines are aligned with the simulated data of corresponding abnormal pipelines at the same position and time point to form structured data pairs, thereby obtaining multiple sets of temperature simulation response training data. Then, the temperature response feature vectors of the multiple sets of temperature simulation response training data are extracted, including the temperature drop value, the temperature drop rate, and the temperature residual distribution. The temperature drop value is extracted by calculating the difference between the highest and lowest temperatures within the pulse period. The temperature drop rate is determined by taking the mean of the slope of the drop segment after performing the first derivative of the temperature curve. The temperature residual distribution is determined by subtracting the measured value from the predicted value of the clean pipeline model and statistically analyzing the standard deviation and extreme values.

[0040] Preferably, a multi-class classification training is performed on a defined random forest based on temperature drop value, temperature drop rate, temperature residual distribution, and known abnormal state labels. The random forest distinguishes multiple anomaly types through a voting mechanism, ultimately obtaining a heat transfer anomaly area identification model that can identify the anomaly types and corresponding probabilities of each region of the pipeline. The definition parameters of the random forest include the number of random forest trees, tree depth, split sample trees, and leaf node sample trees. The number of random forest trees is used to increase model robustness and reduce overfitting, such as 200; the tree depth is used to control model complexity and prevent overfitting, such as 15; the number of split sample trees refers to the minimum number of samples required for node splitting, avoiding overly fine division, such as 8; the number of leaf node sample trees refers to the minimum number of samples for leaf nodes, used to smooth the prediction results, such as 3.

[0041] Furthermore, step S300 also includes step S360, extracting temperature response feature vectors from the real-time response data of the outer surface temperature according to the heat transfer anomaly region identification model, and outputting real-time temperature drop value, real-time temperature drop rate, and real-time temperature residual distribution; step S370, performing random forest classification identification based on the real-time temperature drop value, real-time temperature drop rate, and real-time temperature residual distribution to obtain the heat transfer anomaly region and the heat transfer anomaly response data of the heat transfer anomaly region.

[0042] Preferably, the heat transfer anomaly area identification model extracts temperature response feature vectors from the real-time external surface temperature response data. Specifically, it calculates the difference between the highest and lowest temperatures on the pipe surface within the current pulse cycle to determine the real-time temperature drop value, such as 4.2℃ for a normal area and 2.5℃ for a contaminated area. The cooling rate is calculated using the temperature curve differentiation to determine the real-time temperature drop rate, for example, the rate drops to 40% of the normal value in an oil-contaminated area. The measured data is compared with the predicted values ​​from a pre-stored clean pipe heat transfer model to generate a residual matrix, determining the real-time temperature residual distribution; for example, anomaly areas typically exhibit local residuals >1.5℃. Then, the real-time temperature drop value, real-time temperature drop rate, and real-time temperature residual distribution are input into the heat transfer anomaly area identification model. A multi-decision tree voting mechanism outputs the anomaly area location, i.e., the pipe coordinate segments with abnormal temperature responses; anomaly type determination, i.e., the probability distribution of contaminant types; and anomaly response data, which may include location, area, temperature feature deviation, etc., ultimately obtaining the heat transfer anomaly area and its corresponding heat transfer anomaly response data.

[0043] Step S400: Analyze the heat transfer anomaly response data of the heat transfer anomaly area to generate the cleanliness test result of the long-distance pipeline to be tested.

[0044] Step S400 further includes step S410, which quantifies the heat transfer anomaly response data of the heat transfer anomaly area and outputs a comprehensive anomaly index, wherein the heat transfer anomaly response data includes the location, area and anomaly level of the heat transfer anomaly area; step S420, which converts the cleanliness according to the comprehensive anomaly index to obtain the cleanliness detection result of the long-distance pipeline to be tested.

[0045] Preferably, the heat transfer anomaly response data of the heat transfer anomaly area is analyzed and quantified, and converted into intuitive pipeline cleanliness assessment results. Specifically, the heat transfer anomaly response data of the heat transfer anomaly area is anomaly quantified, that is, a weighted score is performed based on the location, area, and anomaly level of the heat transfer anomaly area. This includes standardizing the distance from the pipeline inlet to 0-1 (inlet = 0, end = 1), with higher weight for anomalies near the end. The percentage of the anomaly area to the total pipeline surface area is calculated and multiplied by an amplification factor (e.g., oil stains × 1.0, scale × 1.5, ice crystals × 0.8). The anomaly level is determined by multiplying the probability value output by the model by the temperature deviation. Based on experimental data, the weights of the location, area, and anomaly level of the heat transfer anomaly area are set to 30%, 40%, and 30%, respectively, and a comprehensive anomaly index is output. Then, according to the comprehensive... The anomaly index is converted to a cleanliness level. This involves mapping the comprehensive anomaly index to a cleanliness level based on the comprehensive anomaly index-cleanliness mapping relationship, ultimately determining the cleanliness test result of the long-distance pipeline under test. For example, a comprehensive anomaly index of 0-30 indicates a contaminant coverage rate of <0.5% and a heat transfer efficiency loss of <3%, corresponding to an excellent cleanliness level requiring no treatment; a comprehensive anomaly index of 30-60 indicates localized micro-deposition (such as oil film) and an efficiency loss of 3-10%, corresponding to a good cleanliness level requiring planned inspection; a comprehensive anomaly index of 60-80 indicates significant scaling (such as limescale) and an efficiency loss of 10-25%, corresponding to a cleanliness level warning requiring cleaning within 3 months; a comprehensive anomaly index of 80-100 indicates large-area contamination or ice blockage and an efficiency loss of >25%, corresponding to a dangerous cleanliness level requiring immediate pipeline shutdown and cleaning.

[0046] In the above text, refer to Figure 1 A method for cleanliness testing of long-distance refrigeration pipelines according to embodiments of the present invention is described in detail. Next, reference will be made to... Figure 3 This invention describes a cleanliness testing device suitable for long-distance refrigeration pipelines according to an embodiment of the present invention.

[0047] The cleanliness detection device for long-distance refrigeration pipelines according to embodiments of the present invention solves the technical problems of low efficiency in cleanliness detection of long-distance refrigeration pipelines, inability to perform real-time dynamic monitoring, and difficulty in accurately identifying local contamination areas in the prior art. It achieves the technical effect of accurately locating abnormal heat transfer areas and high-efficiency, high-precision cleanliness detection of pipelines. Figure 3 As shown, the cleanliness detection device suitable for long-distance refrigeration pipelines includes: an injection data determination module 10, a response data output module 20, an abnormal area marking module 30, and a cleanliness detection result generation module 40.

[0048] The injection data determination module 10 is used to determine the type of injected gas and the injection pulse signal of the long-distance pipeline under test; the response data output module 20 is used to connect a liquid nitrogen vaporization device to the inlet of the long-distance pipeline under test, the liquid nitrogen vaporization device controls the pulse injection valve to inject gas into the inlet of the long-distance pipeline under test through the injected gas type and the injection pulse signal, and at the same time uses a deployed infrared thermal imaging array to collect the temperature of the outer surface of the long-distance pipeline under test and outputs the real-time response data of the outer surface temperature; the abnormal area marking module 30 is used to construct a heat transfer abnormal area identification model through the three-dimensional transient heat transfer equation of the pipeline, and mark the abnormal areas of the real-time response data of the outer surface temperature according to the heat transfer abnormal area identification model to obtain the heat transfer abnormal area; the cleanliness detection result generation module 40 is used to analyze the heat transfer abnormal response data of the heat transfer abnormal area and generate the cleanliness detection result of the long-distance pipeline under test.

[0049] The specific configuration of the anomaly region marking module 30 will be described in detail below. The anomaly region marking module 30 further includes: performing finite element modeling on the long-distance pipeline under test based on its basic parameters and operating environment information to obtain a three-dimensional pipeline structure model; simulating the three-dimensional transient heat transfer equation of the pipeline structure model according to the injected gas type and injection pulse signal, and outputting simulated response data of the outer surface temperature; constructing multiple modeling samples of the long-distance pipeline under test containing known anomaly state labels; simulating the three-dimensional transient heat transfer equation of the pipeline structure model according to the multiple modeling samples of the long-distance pipeline under test, and outputting multiple simulated response data of anomaly in the outer surface temperature; and training the temperature response features of the known anomaly state labels based on the simulated response data of the outer surface temperature and the multiple simulated response data of anomaly in the outer surface temperature to construct a heat transfer anomaly region identification model.

[0050] The specific configuration of the anomaly region marking module 30 will be described in detail below. The anomaly region marking module 30 further includes: establishing control groups for the simulated response data of the outer surface temperature and multiple simulated response data of anomalies in the outer surface temperature, obtaining multiple sets of temperature simulation response training data; extracting temperature response feature vectors from the multiple sets of temperature simulation response training data, wherein the temperature response feature vectors include temperature drop value, temperature drop rate, and temperature residual distribution; performing multi-class classification training on a defined random forest based on the temperature drop value, temperature drop rate, temperature residual distribution, and the known anomaly state labels to obtain a heat transfer anomaly region identification model; wherein the definition parameters of the random forest include the number of random forest trees, tree depth, segmentation sample trees, and leaf node sample trees.

[0051] The specific configuration of the response data output module 20 will be described in detail below. The response data output module 20 further includes: acquiring the temperature of the outer surface of the long-distance pipe under test using a deployed infrared thermal imaging array, wherein the infrared thermal imaging array includes multiple infrared thermal imaging units; acquiring the temperature of the outer surface of the long-distance pipe under test using the deployed infrared thermal imaging array, and outputting the image time series corresponding to the multiple infrared thermal imaging units; performing temperature mapping based on the temperature calibration curve built into each infrared thermal imaging unit, and outputting multiple time-series temperature response matrices characterizing the temperature change at each pixel location; and fusing the multiple time-series temperature response matrices to obtain real-time response data of the outer surface temperature of the long-distance pipe under test.

[0052] The specific configuration of the anomaly region marking module 30 will be described in detail below. The anomaly region marking module 30 further includes: extracting temperature response feature vectors from the real-time response data of the outer surface temperature based on the heat transfer anomaly region identification model, and outputting real-time temperature drop values, real-time temperature drop rates, and real-time temperature residual distributions; performing random forest classification based on the real-time temperature drop values, real-time temperature drop rates, and real-time temperature residual distributions to obtain heat transfer anomaly regions and their heat transfer anomaly response data.

[0053] The specific configuration of the cleanliness test result generation module 40 will be described in detail below. The cleanliness test result generation module 40 further includes: quantifying the heat transfer anomaly response data of the heat transfer anomaly region and outputting a comprehensive anomaly index, wherein the heat transfer anomaly response data includes the location, area, and anomaly level of the heat transfer anomaly region; and performing cleanliness conversion according to the comprehensive anomaly index to obtain the cleanliness test result of the long-distance pipeline under test.

[0054] The specific configuration of the injection data determination module 10 will be described in detail below. The injection data determination module 10 further includes: determining the type of injection gas for the long-distance pipeline to be tested, wherein the type of injection gas is a fluorocarbon or CO2 gas with a boiling point below -40°C.

[0055] The specific configuration of the injection data determination module 10 will be described in detail below. The injection data determination module 10 further includes: acquiring the pipe length and thermal response time constant of the long-distance pipe to be tested; scoring the pipe response effectiveness based on the pipe length and thermal response time constant of the long-distance pipe to be tested; optimizing the pulse signal frequency, amplitude, and duration period under the constraint of satisfying the preset response effectiveness score threshold; and generating an injection pulse signal.

[0056] The cleanliness testing device for long-distance refrigeration pipelines provided in this embodiment of the invention can execute the cleanliness testing method for long-distance refrigeration pipelines provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0057] Although this application makes various references to certain modules in the apparatus according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not intended to limit the scope of protection of this invention.

[0058] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for detecting the cleanliness of long-distance refrigeration pipelines, characterized in that, The method includes: Determine the type of gas to be injected and the injection pulse signal for the long-distance pipeline under test; A liquid nitrogen vaporization device is connected to the inlet of the long-distance pipeline under test. The liquid nitrogen vaporization device controls the pulse injection valve to inject gas into the inlet of the long-distance pipeline under test through the type of injected gas and the injection pulse signal. At the same time, an infrared thermal imaging array is deployed to collect the temperature of the outer surface of the long-distance pipeline under test and output the real-time response data of the outer surface temperature. A heat transfer anomaly region identification model is constructed by the three-dimensional transient heat transfer equation of the pipeline. Based on the heat transfer anomaly region identification model, the real-time response data of the outer surface temperature is marked as anomaly region to obtain the heat transfer anomaly region. Analyze the heat transfer anomaly response data of the heat transfer anomaly area to generate the cleanliness test result of the long-distance pipeline under test; Among them, a heat transfer anomaly region identification model is constructed by using the three-dimensional transient heat transfer equation of the pipeline. The method includes: Based on the basic parameters and operating environment information of the long-distance pipeline to be tested, a finite element model is performed on the long-distance pipeline to be tested to obtain a three-dimensional pipeline structure model; The three-dimensional transient heat transfer equation of the pipeline structure model is simulated according to the injected gas type and injection pulse signal, and the simulated response data of the outer surface temperature is output. Construct multiple abnormal long-distance pipeline modeling samples containing known abnormal state labels; Based on the modeling samples of multiple abnormal long-distance pipelines to be tested, the three-dimensional pipeline structure model is simulated using the three-dimensional transient heat transfer equation of the pipeline, and multiple simulated response data of abnormal external surface temperature are output. Based on the simulated response data of the outer surface temperature and the simulated response data of multiple outer surface temperature anomalies, the temperature response features of the known abnormal state labels are trained to construct a heat transfer anomaly region identification model.

2. The cleanliness detection method for long-distance refrigeration pipelines as described in claim 1, characterized in that, The method for training temperature response features on the known abnormal state labels based on the simulated response data of the outer surface temperature and multiple simulated response data of outer surface temperature anomalies includes: Establish control groups for the simulated response data of the outer surface temperature and multiple simulated response data of the outer surface temperature anomaly, and obtain multiple sets of temperature simulation response training data. Extract the temperature response feature vector from the multiple sets of temperature simulation response training data. The temperature response feature vector includes the temperature drop value, the temperature drop rate, and the temperature residual distribution. Based on the temperature drop value, temperature drop rate, temperature residual distribution, and known abnormal state labels, a multi-class training of the defined random forest is performed to obtain a heat transfer anomaly region identification model. The defining parameters of a random forest include the number of random forest trees, tree depth, split sample trees, and leaf node sample trees.

3. The cleanliness detection method for long-distance refrigeration pipelines as described in claim 1, characterized in that, The temperature of the outer surface of the long-distance pipe under test is collected by a deployed infrared thermal imaging array, wherein the infrared thermal imaging array includes multiple infrared thermal imaging units. The temperature of the outer surface of the long-distance pipe under test is collected by the deployed infrared thermal imaging array, and the image time series corresponding to the multiple infrared thermal imaging units is output. Temperature mapping is performed based on the temperature measurement calibration curve built into each infrared thermal imaging unit, and multiple time-series temperature response matrices characterizing the temperature change at each pixel location are output. The real-time response data of the outer surface temperature of the long-distance pipeline under test is obtained by fusing the multiple time-series temperature response matrices.

4. The cleanliness detection method for long-distance refrigeration pipelines as described in claim 1, characterized in that, The abnormal regions are identified by marking the real-time response data of the outer surface temperature according to the heat transfer anomaly identification model. The method includes: Based on the heat transfer anomaly region identification model, the temperature response feature vector of the real-time external surface temperature response data is extracted, and the real-time temperature drop value, real-time temperature drop rate and real-time temperature residual distribution are output. Random forest classification is performed based on the real-time temperature drop value, real-time temperature drop rate, and real-time temperature residual distribution to obtain heat transfer anomaly regions and heat transfer anomaly response data of the heat transfer anomaly regions.

5. The cleanliness detection method for long-distance refrigeration pipelines as described in claim 4, characterized in that, Analyzing the heat transfer anomaly response data in the heat transfer anomaly region to generate the cleanliness detection result of the long-distance pipeline under test, the method includes: The heat transfer anomaly response data of the heat transfer anomaly region is subjected to anomaly quantification, and a comprehensive anomaly index is output. The heat transfer anomaly response data includes the regional location, regional area, and anomaly level of the heat transfer anomaly region. The cleanliness test result of the long-distance pipeline under test is obtained by converting the cleanliness index according to the comprehensive anomaly index.

6. The cleanliness detection method for long-distance refrigeration pipelines as described in claim 1, characterized in that, The type of gas to be injected into the long-distance pipeline under test is determined. The type of gas is a fluorocarbon or other gas with a boiling point below -40°C. gas.

7. The cleanliness detection method for long-distance refrigeration pipelines as described in claim 1, characterized in that, Methods for determining the injection pulse signal of a long-distance pipeline under test include: Obtain the pipe length and thermal response time constant of the long-distance pipe to be tested; The pipeline response effectiveness score is calculated based on the pipeline length and thermal response time constant of the long-distance pipeline under test. The frequency, amplitude and duration of the pulse signal are optimized to generate the injected pulse signal, with the preset response effectiveness score threshold as a constraint.

8. A cleanliness testing device suitable for long-distance refrigeration pipelines, characterized in that, The apparatus is used to implement the cleanliness detection method for long-distance refrigeration pipelines as described in any one of claims 1 to 7, the apparatus comprising: The injection data determination module is used to determine the type of injected gas and the injection pulse signal for the long-distance pipeline under test; The response data output module is used to connect a liquid nitrogen vaporization device to the inlet of the long-distance pipeline under test. The liquid nitrogen vaporization device controls the pulse injection valve to inject gas into the inlet of the long-distance pipeline under test through the type of injected gas and the injection pulse signal. At the same time, the deployed infrared thermal imaging array collects the temperature of the outer surface of the long-distance pipeline under test and outputs real-time response data of the outer surface temperature. The abnormal region marking module is used to construct a heat transfer abnormal region identification model through the three-dimensional transient heat transfer equation of the pipeline, and mark the abnormal region of the real-time response data of the outer surface temperature according to the heat transfer abnormal region identification model to obtain the heat transfer abnormal region. The cleanliness test result generation module is used to analyze the heat transfer anomaly response data of the heat transfer anomaly area and generate the cleanliness test result of the long-distance pipeline under test.

Images