Digital-twin-based dynamic leakage monitoring method for high-purity gas pipeline

Abstract

The present application relates to the technical field of industrial process state monitoring and fault diagnosis, in particular to a high-purity gas pipeline dynamic leakage monitoring method based on digital twinning. The method fuses existing gas sensors to construct a digital twinning model of the gas pipeline, which includes a first virtual model corresponding to the physical pipeline and a first virtual sensor corresponding to the gas sensor; then, from historical maintenance records, historical pressure data related to the gas sensor and pipeline maintenance data related to the pipeline are obtained. The present application constructs a virtual gas pressure sensor through historical maintenance records, and associates the virtual gas pressure sensor with the maintenance state of the pipeline components, uses the maintenance state of the pipeline components as the basis for adjusting the monitoring data of the virtual gas pressure sensor, so that the gas pressure sensor can adjust the pressure data according to the pipeline state, thereby combining the virtual pressure data with the gas sensor to comprehensively monitor gas leakage.

Description

Technical Field

[0001] This invention relates to the field of industrial process condition monitoring and fault diagnosis technology, and more specifically, to a method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins. Background Technology

[0002] High-purity gases (such as oxygen) are indispensable raw materials in modern high-end industrial fields such as semiconductor manufacturing, biopharmaceuticals, and precision chemicals. The purity and integrity of their delivery pipelines are directly related to the quality of the final product and production safety.

[0003] In terms of monitoring technology, methods have evolved from primitive approaches such as applying soapy water to the outer walls of equipment to directly monitoring leaked gases into the environment using gas sensors. With the popularization of digital technology, newly built pipeline systems are usually pre-installed with various sensing devices, including gas, pressure, and flow sensors, laying the foundation for building high-fidelity digital twin models and realizing intelligent monitoring.

[0004] However, this digital twin monitoring model, which relies on multi-sensor fusion, faces significant challenges when applied to existing aging pipelines. Firstly, most aging pipelines were initially equipped with only basic gas sensors. Building a fully functional digital twin model requires large-scale modifications, including the addition of various types of sensors, making the engineering implementation difficult. Secondly, especially for pipelines with a relatively short remaining service life, the high cost of modification and monitoring makes this approach economically unfeasible. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic leakage monitoring method for high-purity gas pipelines based on digital twins. This method creates a virtual pressure sensor to work in conjunction with a gas sensor, thereby solving the problem mentioned in the background art, namely, that the digital twin monitoring mode, which relies on multi-sensor fusion, is difficult to apply to existing old pipelines.

[0006] To achieve the above objectives, the method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins includes the following steps: S1. Integrate existing gas sensors to construct a digital twin model of the gas pipeline. The digital twin model includes a first virtual model corresponding to the physical pipeline and a first virtual sensor corresponding to the gas sensor. S2. Obtain historical pressure data related to the gas sensor and pipeline maintenance data related to the pipeline from the historical maintenance records; S3. Based on historical pressure data, generate a second virtual sensor in the digital twin model according to the position of the first virtual sensor; S4. Based on pipeline maintenance data, verify and calibrate the pressure data trend reflected by the second virtual sensor; S5. By fusing and analyzing the data from the verified second virtual sensor and the first virtual sensor, gas leaks are monitored.

[0007] As a further improvement to this technical solution, the steps for constructing the digital twin model in S1 are as follows: S1.1 Collect basic attribute data of physical gas pipelines and deployment data of existing physical gas sensors; S1.2 Construct a three-dimensional geometric model of the gas pipeline in the simulation platform and embed the fluid dynamics mechanism to form the first virtual model; S1.3 In the digital twin model, a corresponding digital entity is created for each physical gas sensor to obtain the first virtual sensor, and a data communication link is established between the first virtual sensor and its corresponding physical gas sensor.

[0008] As a further improvement to this technical solution, the steps for obtaining historical pressure data and pipeline maintenance data in S2 are as follows: S2.1. Convert paper maintenance work orders into electronic images using a scanner, perform text recognition using optical character recognition technology, and convert them into searchable and processable text; S2.2 Establish a sensor keyword library and maintain an action keyword library; S2.3. Based on the sensor keyword library, obtain the starting pressure of the pipeline from the text. and end pressure Based on the maintenance action keyword library, retrieve the pipeline components that are under maintenance from the text.

[0009] As a further improvement to this technical solution, the sensor keyword library includes "pressure, pressure, pressure holding test, pressure test, set pressure".

[0010] As a further improvement to this technical solution, the maintenance action keyword library includes "replacement, disassembly, tightening, leakage, and sealing".

[0011] As a further improvement to this technical solution, the step of generating the second virtual sensor in S3 is as follows: S3.1 Obtain the three-dimensional coordinates of the gas sensor relative to a fixed reference point using a three-dimensional laser scanner; S3.2 Input the three-dimensional coordinates into the digital twin platform to obtain the position of the first virtual sensor, and set the second virtual sensor at the position of the first virtual sensor; S3.3 Obtain the total length of the pipeline Calculate the pressure data from the second virtual sensor. The calculation formula is as follows: ; In the formula, This represents the distance between the second virtual sensor and the beginning of the pipe. For pressure data from the second virtual sensor, , as well as The pressure is absolute pressure, which equals gauge pressure plus local atmospheric pressure.

[0012] As a further improvement to this technical solution, the steps for verifying and calibrating the second virtual sensor in S4 are as follows: S4.1 Obtain event data for pipeline component maintenance; S4.2. The time when the gas sensor detects a gas leak is taken as the start time of the pressure drop. The replacement time of maintenance components is taken as the end time of air pressure drop. And calculate the number of days of pressure drop. Number of days of pressure reduction Start time With end time The difference between them; S4.3. Combine industrial big data and historical replacement data to predict the replacement time of maintenance parts. And combined with the number of days of pressure reduction Calculate the pressure data from the second virtual sensor. The reduction time.

[0013] As a further improvement to this technical solution, the pressure data from the second virtual sensor... Reduction time = Replacement time - Number of days of pressure reduction .

[0014] As a further improvement to this technical solution, step S5 includes the following steps: When the gas sensor detects high-purity gas, it acquires the current date and the corresponding pressure data from the second virtual sensor. When the time for reduction is reached, a warning signal is issued; when the current date does not correspond to the pressure data of the second virtual sensor... No warning is issued when the time for reduction is reached.

[0015] As a further improvement to this technical solution, step S5 also includes the following steps: Based on the pressure drop days Tt and the pressure drop data, the expected gas concentration reference value that can be detected in the environment when a leak occurs at the corresponding part of the second virtual sensor is calculated; when the gas concentration detected by the gas sensor in real time continues to exceed the concentration reference value, it prompts to check other pipeline parts.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: In this digital twin-based dynamic leakage monitoring method for high-purity gas pipelines, a virtual pressure sensor is constructed using historical maintenance records. The virtual pressure sensor is then associated with the maintenance status of pipeline components. The maintenance status of the pipeline components is used as the basis for adjusting the monitoring data of the virtual pressure sensor, enabling the pressure sensor to adjust the pressure data according to the pipeline status. This allows the virtual pressure data to be combined with the gas sensor for comprehensive monitoring of gas leaks. Attached Figure Description

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

[0018] The technical solutions in 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] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] To address the problem that digital twin monitoring methods relying on multi-sensor fusion are difficult to apply to existing aging pipelines, this invention provides a dynamic leak monitoring method for high-purity gas pipelines based on digital twins. For example... Figure 1 As shown, the monitoring method includes the following steps: S1. Integrate existing gas sensors to construct a digital twin model of the gas pipeline. The digital twin model includes a first virtual model corresponding to the physical pipeline and a first virtual sensor corresponding to the gas sensor. S2. Obtain historical pressure data related to the gas sensor and pipeline maintenance data related to the pipeline from the historical maintenance records; S3. Based on historical pressure data, generate a second virtual sensor in the digital twin model according to the position of the first virtual sensor; S4. Based on pipeline maintenance data, verify and calibrate the pressure data trend reflected by the second virtual sensor; S5. Gas leaks are monitored by fusing and analyzing data from the verified second virtual sensor and the first virtual sensor.

[0021] Specifically, digital twin models can comprehensively determine the leakage status by analyzing monitoring data from multiple sensors. For example, if the pressure model shows everything is normal, but a gas sensor shows a brief abnormal reading, the digital twin will determine that this may be a false alarm or environmental interference, thus avoiding erroneous operation. Therefore, this comprehensive judgment method can not only provide early warning but also locate the leak when it occurs. However, some older pipelines are only equipped with gas sensors. To address this issue, this solution incorporates historical maintenance records into the data twin model and fuses them with the gas sensor data as a basis for gas leak monitoring. The construction method of the data twin model is as follows: The process involves collecting basic attribute data of the physical gas pipeline and deployment data of existing physical gas sensors. The basic attribute data of the physical gas pipeline includes, but is not limited to, the pipeline routing diagram (PID), pipe diameter, length, material, and the location and model of valves and connectors. The deployment data of the physical gas sensors includes unique identifiers (such as tag numbers), installation locations, measured gas types, measurement ranges, and communication protocols. Based on this data, a three-dimensional geometric model of the gas pipeline is constructed in a simulation platform, recreating its spatial layout and structure. Fluid dynamics mechanisms are embedded to build a computational model that can simulate the gas pressure and flow distribution within the pipeline, forming a digital twin model, i.e., the first virtual model. Next, for each physical gas sensor in the first virtual sensor, a corresponding digital entity (i.e., the first virtual sensor) is created within the digital twin model. A data communication link is established between the first virtual sensor and its corresponding physical gas sensor, enabling automatic synchronization and mapping of real-time measurement data from the physical sensors to the virtual sensors.

[0022] At this stage, the digital twin model already possesses the capability to monitor gas leaks. For example, when a gas sensor detects leaking gas, the detected data is automatically synchronized to the first virtual sensor, allowing the digital twin model to determine the leak location based on the location of the first virtual sensor. However, due to the lack of support from multi-source heterogeneous data, the current data twin model only triggers an alarm when the gas concentration reaches a fixed threshold, and it cannot determine whether the detected gas is from a pipeline leak or external environmental interference. Therefore, this invention chooses to introduce data from historical maintenance records to complete the missing data. Specifically: First, if the company uses a computerized maintenance management system, historical maintenance data text is directly retrieved from it. If not, paper maintenance work orders are collected, scanned into electronic images, and then optical character recognition (OCR) is used to recognize the text and convert it into searchable and processable text. Next, a sensor keyword library and a maintenance action keyword library are established. In this solution, since only gas sensors exist, pressure data related to the gas sensors must be selected as the basis for fusion with the gas sensor data. Therefore, the sensor keyword library includes terms such as "pressure," "pressure holding test," "pressure test," and "set pressure." The maintenance action keyword library mainly includes terms such as "replacement," "disassembly," "tightening," "leakage," and "sealing."

[0023] After establishing the sensor keyword library and maintenance action keyword library, historical pressure data and pipeline maintenance data are acquired. For historical pressure data, the starting pressure of the pipeline is obtained from the text based on the sensor keyword library. and end pressure For pipeline maintenance data, the main focus is on recording the replacement status of pipeline components. Specifically, this is done by retrieving pipeline components that are being maintained from the text based on a maintenance action keyword library.

[0024] After collecting historical pressure data, it can be used as the basis for deploying the second virtual sensor. During the deployment of the second virtual sensor, the pressure data at the corresponding location is calculated based on the location of the first virtual sensor and used as the monitoring data for the second virtual sensor. The specific method is as follows: First, fixed reference points around the gas sensor are identified. The 3D coordinates of the gas sensor relative to these reference points are obtained using a 3D laser scanner. These coordinates are then input into a digital twin platform to obtain the position of the first virtual sensor. A second virtual sensor is then positioned at the location of the first virtual sensor. Next, the total length of the pipeline is... Calculate the pressure data from the second virtual sensor. The calculation formula is as follows: ; In the formula, This represents the distance between the second virtual sensor and the beginning of the pipe. For pressure data from the second virtual sensor, , as well as The pressure is absolute pressure, which equals gauge pressure plus local atmospheric pressure. For example, record the pressure at the beginning of the pipeline. It is 0.9 MPa (gauge pressure). The local atmospheric pressure is about 0.1 MPa, so the absolute pressure is 1.0 MPa.

[0025] Taking a real-world scenario as an example, suppose the total length of a high-purity gas pipeline is... =50 meters. Initial pressure of the pipeline. The pressure is 0.5 MPa, and the terminal pressure is... The pressure is 0.45 MPa. The local atmospheric pressure is 0.1 MPa. The distance of the second virtual sensor from the beginning of the pipe... The length is 20 meters. At this point, first calculate the absolute pressure, that is, the pressure at the beginning of the pipe. The pressure is 0.5 MPa + 0.1 MPa = 0.6 MPa; terminal pressure The pressure is 0.45MPa + 0.1MPa = 0.55MPa. Substituting these values ​​into the formula yields the pressure data from the second virtual sensor. ≈0.5805MPa (absolute pressure), then subtract 0.1MPa of atmospheric pressure to get the gauge pressure data of approximately 0.48MPa.

[0026] Complete the pressure data from the second virtual sensor After calculation, since the pressure is calculated under leak-free conditions in the pipeline, and the second virtual sensor has no corresponding physical entity, it is impossible to update the internal air pressure of the pipeline in real time. Therefore, this invention uses the pressure data from the second virtual sensor during routine pipeline maintenance. An update is being performed. Details are as follows: First, acquire event data for pipeline component maintenance, including the component identifier (e.g., valve V-101), maintenance type (e.g., replacing a seal), maintenance execution timestamp, and post-maintenance test data (e.g., initial and stable pressures of a pressure holding test). Next, use the time when the gas sensor detects a gas leak as the start time of the pressure drop. The replacement time of maintenance components is taken as the end time of air pressure drop. And calculate the number of days of pressure drop. Number of days of pressure reduction Start time With end time The difference between them. For example, a gas sensor detects an external high-purity gas concentration of 10 ppm on January 1st, and the alarm threshold is 500 ppm; in this case, the gas sensor does not alarm. Suppose the maintenance component is replaced on January 31st, and this is the time when the gas pressure begins to drop. The end date is January 1st. January 31st, number of days of reduction It lasts for 30 days.

[0027] Subsequently, by combining industrial big data and historical replacement data, the replacement time of maintenance parts is predicted. And combined with the number of days of pressure reduction Calculate the pressure data from the second virtual sensor The reduction time. The specific calculation method is: replacement time. - Number of days of pressure reduction For example, predicting the replacement time of maintenance parts. 100 days later, the number of days of pressure reduction at this time. The pressure data from the second virtual sensor was obtained over 30 days. The reduction begins on day 70.

[0028] The predicted replacement time Td for the maintenance component can be estimated by analyzing the mean time between failures (MTBF) of that component model and combining it with the current operating time; or, it can be predicted based on the component's historical replacement records using time series analysis or machine learning models (such as regression models). Finally, the methods for monitoring gas leaks are as follows: When the gas sensor detects high-purity gas, it acquires the current date and the corresponding pressure data from the second virtual sensor. When the time for reduction is reached, a warning signal is issued; when the current date does not correspond to the pressure data of the second virtual sensor... No warning is issued when the pressure data from the second virtual sensor decreases. For example, if the gas sensor detects an external high-purity gas concentration of 100 ppm, and the current time is not within the range of the second virtual sensor's pressure data... If the pressure inside the pipeline does not decrease within the specified time, the detected high-purity gas is not considered a leak, and therefore no warning is issued.

[0029] Furthermore, based on the pressure drop days Tt and the pressure drop data, the expected gas concentration reference value that can be detected in the environment when a leak occurs at the corresponding part of the second virtual sensor can be calculated; when the gas concentration detected by the gas sensor in real time continuously exceeds the concentration reference value, it will prompt to check other pipeline parts.

[0030] In summary, this invention constructs a virtual pressure sensor through historical maintenance records and associates the virtual pressure sensor with the maintenance status of pipeline components. The maintenance status of pipeline components is used as the basis for adjusting the monitoring data of the virtual pressure sensor, enabling the pressure sensor to adjust the pressure data according to the pipeline status. Thus, the virtual pressure data is combined with the gas sensor to comprehensively monitor gas leaks.

[0031] It should be noted that the monitoring method provided by this invention can achieve early warning and comprehensive judgment of minor leaks in high-purity gas pipelines, and is applicable to various industrial scenarios with stringent requirements for pipeline integrity, such as high-purity gas delivery systems in semiconductor manufacturing, biopharmaceutical and other fields.

[0032] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins, characterized in that: Includes the following steps: S1. Integrate existing gas sensors to construct a digital twin model of the gas pipeline. The digital twin model includes a first virtual model corresponding to the physical pipeline and a first virtual sensor corresponding to the gas sensor. S2. Obtain historical pressure data related to the gas sensor and pipeline maintenance data related to the pipeline from the historical maintenance records; S3. Based on historical pressure data, generate a second virtual sensor in the digital twin model according to the position of the first virtual sensor; S4. Based on pipeline maintenance data, verify and calibrate the pressure data trend reflected by the second virtual sensor; S5. By fusing and analyzing the data from the verified second virtual sensor and the first virtual sensor, gas leaks are monitored.

2. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 1, characterized in that: The steps for constructing the digital twin model in S1 are as follows: S1.1 Collect basic attribute data of physical gas pipelines and deployment data of existing physical gas sensors; S1.2 Construct a three-dimensional geometric model of the gas pipeline in the simulation platform and embed the fluid dynamics mechanism to form the first virtual model; S1.3 In the digital twin model, a corresponding digital entity is created for each physical gas sensor to obtain the first virtual sensor, and a data communication link is established between the first virtual sensor and its corresponding physical gas sensor.

3. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 1, characterized in that: The steps for obtaining historical pressure data and pipeline maintenance data in S2 are as follows: S2.

1. Convert paper maintenance work orders into electronic images using a scanner, perform text recognition using optical character recognition technology, and convert them into searchable and processable text; S2.2 Establish a sensor keyword library and maintain an action keyword library; S2.

3. Based on the sensor keyword library, obtain the starting pressure of the pipeline from the text. and end pressure Based on the maintenance action keyword library, retrieve the pipeline components that are under maintenance from the text.

4. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 3, characterized in that: The sensor keyword library includes "pressure, pressure, pressure holding test, pressure test, set pressure".

5. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 3, characterized in that: The maintenance action keyword library includes "replacement, disassembly, tightening, leakage, and sealing".

6. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 3, characterized in that: The steps for generating the second virtual sensor in S3 are as follows: S3.1 Obtain the three-dimensional coordinates of the gas sensor relative to a fixed reference point using a three-dimensional laser scanner; S3.2 Input the three-dimensional coordinates into the digital twin platform to obtain the position of the first virtual sensor, and set the second virtual sensor at the position of the first virtual sensor; S3.3 Obtain the total length of the pipeline. Calculate the pressure data from the second virtual sensor. The calculation formula is as follows: ； In the formula, This represents the distance between the second virtual sensor and the beginning of the pipe. For pressure data from the second virtual sensor, , as well as The pressure is absolute pressure, which equals gauge pressure plus local atmospheric pressure.

7. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 6, characterized in that: The steps for verifying and calibrating the second virtual sensor in S4 are as follows: S4.1 Obtain event data for pipeline component maintenance; S4.

2. The time when the gas sensor detects a gas leak is taken as the start time of the pressure drop. The replacement time of maintenance components is taken as the end time of air pressure drop. And calculate the number of days of pressure drop. Number of days of pressure reduction Start time With end time The difference between them; S4.

3. Combine industrial big data and historical replacement data to predict the replacement time of maintenance parts. And combined with the number of days of pressure reduction Calculate the pressure data from the second virtual sensor. The reduction time.

8. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 7, characterized in that: Pressure data from the second virtual sensor Reduction time = Replacement time - Number of days of pressure reduction .

9. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 7, characterized in that: S5 includes the following steps: When the gas sensor detects high-purity gas, it acquires the current date and the corresponding pressure data from the second virtual sensor. When the time for reduction is reached, a warning signal is issued; when the current date does not correspond to the pressure data of the second virtual sensor... No warning is issued when the time for reduction is reached.

10. The method for dynamic leakage monitoring of high-purity gas pipelines based on digital twins according to claim 9, characterized in that: S5 further includes the following steps: Based on the pressure drop days Tt and the pressure drop data, the expected gas concentration reference value that can be detected in the environment when a leak occurs at the corresponding part of the second virtual sensor is calculated; when the gas concentration detected by the gas sensor in real time continues to exceed the concentration reference value, it prompts to check other pipeline parts.

Images