Digital twinning-based check valve full life cycle management system and method

Abstract

The invention discloses a one-way valve full life cycle management system and method based on digital twinning, and the system comprises a physical entity system which is an actual one-way valve physical system; the virtual twin model system is used for virtually duplicating the physical entity system and is used for dynamically mapping the running state of the one-way valve in real time; the information interaction system is used for mutual fusion of the entity signal and the twin simulation signal; the dynamic visual interface is used for displaying real-time operation parameter information of the physical entity system and the virtual twin model system; the health management system is used for recording all objects, models and data in the whole one-way valve operation cycle and realizing online fault diagnosis and early warning; and the predictive maintenance system is used for training the residual life transfer learning prediction model of the equipment by using the digital twin full-life-cycle simulation data of the equipment so as to realize cross-domain health management and prediction.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance and health management technology for industrial equipment, specifically to a digital twin-based full lifecycle management system and method for check valves. In particular, this invention is applicable to real-time status monitoring, fault diagnosis, and remaining life prediction of check valves in reciprocating high-pressure diaphragm pumps under varying operating conditions. Background Technology

[0002] Reciprocating high-pressure diaphragm pumps are core power equipment in slurry pipeline transportation systems in metallurgy, mining, and chemical industries. As the most frequently moving and easily damaged critical component within the pumping system, the health of the check valve directly determines the operational stability and reliability of the entire pumping system. A failure of the check valve will lead to unplanned downtime, causing significant losses to the company's production.

[0003] Currently, the health management of check valves mainly faces the following technical bottlenecks: Insufficient status awareness and diagnostic capabilities: Existing health management systems struggle to acquire the real-time and accurate operating status of check valves. Fault diagnosis often relies on single signal analysis or empirical models, which commonly leads to missed detections and misdiagnoses when faced with complex real-world operating conditions, resulting in equipment not being maintained in a timely manner.

[0004] The maintenance model is passive and outdated: the commonly used "repair after failure" or preventative maintenance model based on fixed cycles lacks foresight regarding the degradation process of equipment performance. This passive maintenance model has serious time lag and can no longer meet the requirements of modern smart factories for high equipment availability and zero unexpected downtime.

[0005] The model exhibits poor generalization ability under varying operating conditions: High-pressure diaphragm pumps often operate under non-stationary conditions with varying speeds and loads, especially between pump stations of different stages, where operating parameters differ significantly. These varying operating conditions cause the vibration, pressure, and other fault signal characteristics of check valves to drift, making it difficult to directly transfer and effectively apply fault diagnosis models and remaining life prediction models trained based on specific operating conditions to other operating scenarios. This severely restricts the large-scale deployment and application of health management systems.

[0006] Therefore, there is an urgent need for a health management system and method that can dynamically sense the status of one-way valves, perform accurate fault diagnosis online, and achieve high-precision prediction of remaining life under different operating conditions, in order to overcome the limitations of existing technologies. Summary of the Invention

[0007] The purpose of this invention is to address the aforementioned problems by providing a digital twin-based full lifecycle management system and method for check valves. This system dynamically monitors check valve operating parameters, performs online fault diagnosis and alarms, predicts and warns of remaining lifespan of check valves, enables cross-domain prediction of remaining lifespan of check valves, and provides visualized operation and maintenance throughout the entire lifecycle of check valves.

[0008] The technical solution of the present invention is as follows: A digital twin-based lifecycle management system for check valves includes: The physical entity system is the actual one-way valve system; A virtual twin model system, corresponding to a physical entity system, is used to dynamically map the operating status of a one-way valve in real time. An information interaction system is used to enable bidirectional data transmission and fusion between physical entity systems and virtual twin model systems; A dynamic visualization interface, connected to the virtual twin model system and information interaction system, is used to display the real-time operating parameters of the physical entity and the virtual twin model; The health management system, connected to the information interaction system and the virtual twin model system, is used for status monitoring and fault diagnosis of check valves based on fused data; Predictive maintenance systems, connected to health management systems and virtual twin models, are used to predict the remaining lifespan of check valves based on historical and real-time data.

[0009] Furthermore, the virtual twin model system includes: A physical model is a 1:1 geometric model of a physical entity system. The local linear model is constructed based on the operating mechanism of the one-way valve, including the fluid continuity equation, momentum equation and mathematical parameterization expression of liquid-solid two-phase flow; Behavioral models are parameterized representations of the control logic and external responses of physical entities, including electrical models, control models, and mechanical models.

[0010] Furthermore, the virtual twin model system uses a combination of data-driven and finite element model correction methods to update the model, dynamically correcting the material properties, dimensional parameters, or constraints of the model based on the system response data collected by the sensors.

[0011] Furthermore, the information interaction system includes: The intelligent sensing system is used to collect the operating parameters of a physical entity system throughout its entire life cycle. These operating parameters include vibration signals, pressure signals, and flow signals. The information fusion system is used to fuse multi-source heterogeneous data collected by the intelligent sensing system and to build a scene perception system and spatiotemporal database for one-way valve equipment.

[0012] Furthermore, the health management system is configured as follows: Record all objects, models, and data throughout the entire lifecycle of the check valve to create a digital archive; By integrating the Weibull proportional failure rate model, a dynamic evolution model of the performance degradation of the one-way valve is established to describe its degradation state and the changing patterns of its parameters over time.

[0013] Furthermore, the predictive maintenance system is configured to utilize full lifecycle simulation data generated by the equipment's digital twin to construct a database of one-way valve wear evolution trend models.

[0014] Furthermore, the predictive maintenance system further includes a transfer learning prediction model for the remaining life of equipment based on a deep residual shrinking network (DRSN), used to achieve cross-domain prediction of the remaining life of the check valve under different operating conditions.

[0015] Furthermore, an improved Parareal solution algorithm is used to solve the remaining lifetime transfer learning prediction model based on the Deep Residual Shrinking Network (DRSN).

[0016] Furthermore, the dynamic visualization interface combines augmented reality technology to overlay and display the parameter information of the physical entity and the virtual twin model.

[0017] This application also includes a digital twin-based method for full lifecycle health management of one-way valves, implemented based on a digital twin-based full lifecycle management system for one-way valves.

[0018] Compared with existing technologies, the advantages of this invention are: 1. This invention introduces digital twin technology, integrates multi-source heterogeneous information, and combines the failure mechanism of the equipment with domain knowledge to build a scene perception system for the digital twin of the check valve equipment. It also builds a high-fidelity virtual twin system for the check valve. Through high-performance accurate simulation and intelligent big data analysis, it realizes a high-fidelity digital description and intelligent application service for the entire life cycle of complex mechanical equipment from design, manufacturing, operation and maintenance to scrapping, providing effective simulation data for the full life cycle health management of the check valve. 2. This invention provides a one-way valve equipment status monitoring and health management solution based on digital twin technology. It fully utilizes the advantages of digital twin technology to construct a multi-source heterogeneous information fault diagnosis model, which solves the problems of single predictive maintenance methods in terms of model consistency, algorithm adaptability, and prediction accuracy. By constructing a multi-fault spatiotemporal database and combining machine learning technology to complete the training of a complete diagnostic model, it can achieve online, fast, and accurate diagnosis of the type and cause of faults. 3. This invention considers the real-time and transient nature of prediction, introduces deep transfer learning algorithms for big data analysis and mining, breaks through the technology of high-concurrency access and fast processing of diverse heterogeneous data, realizes non-shared collaborative feature extraction, non-direct reading monitoring data quality assessment and dataless multi-entity fault diagnosis knowledge-targeted transfer method, establishes a high-quality service status monitoring database for check valves, and constructs a distributed multi-entity collaborative fault diagnosis and prediction platform to predict the health status of check valves under varying operating conditions in actual industrial production. Attached Figure Description

[0019] Figure 1 This is a framework diagram of the system in this application.

[0020] Figure 2 This is the dynamic monitoring-level fault diagnosis architecture of this application.

[0021] Figure 3 This is the visual interface of the system in this application. Detailed Implementation

[0022] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0023] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0024] Please see Figure 1-3 A digital twin-based one-way valve lifecycle management system and method, such as Figure 1 As shown, it includes: a physical entity system, a virtual twin model system, an information interaction system, a dynamic visualization interface, a health management system, and a predictive maintenance system.

[0025] The physical entity system is the one-way valve entity, including valve box, valve seat, spring, limiter and valve core. The valve core includes pressure cap, rubber gasket, valve disc and guide claw. The valve core mainly relies on the pressure difference between the upper and lower liquids to open, and closes under the combined action of pressure difference force, spring resistance and the gravity of the valve core, to realize one-way transport of slurry.

[0026] The virtual twin model system is used to map the operating status of the one-way valve entity in real time. The virtual twin model is a mathematical description of the physical entity, constructing a digital twin of the one-way valve that integrates mechanism, data and knowledge, which can realize the reproduction and reconstruction of the attributes, characteristics and functions of the physical entity. Furthermore, the virtual twin model system includes a physical model, a local linear model, and a behavioral model. These three models represent the same device at different time scales and levels of precision, comprehensively reflecting the device's internal structure, real-time status, and control system. The physical model is a 1:1 geometric replica of the physical entity; the local linear model is a mathematical abstraction of the physical entity's operating mechanism, mainly including the mathematical model of the operating mechanism, fluid continuity equation, momentum equation, energy equation, and parameterized mathematical expressions for liquid-solid two-phase flow; the behavioral model is a parameterized expression of the physical entity's behavior, mainly including electrical models, control models, hydraulic models, mechanical models, and environmental parameter models. Furthermore, the physical entity system and the constructed virtual twin model system space can maintain the same operating state of the twin model and the physical one-way valve throughout their entire life cycle through the mutual transmission of measured data and virtual simulation data.

[0027] The information interaction system is used for the mutual fusion of the one-way valve physical operation signal and the twin simulation signal, and mainly includes an intelligent sensing system and an information fusion system. Furthermore, the intelligent sensing system is used to collect operational parameter information of the one-way valve physical entity system throughout its entire life cycle, mainly including vibration signals, pressure signals, and flow signals. The vibration signal acquisition system mainly includes a sensor detection module (accelerometer, sound pressure sensor), a data acquisition module (dynamic signal acquisition card), and a data acquisition, storage, and control module; the pressure and flow signal acquisition system mainly includes a pressure sensor and a flow sensor. Furthermore, the information fusion system is used for multi-source information fusion, combining the failure mechanism of the equipment and domain knowledge to build a scene perception system for the digital twin of the one-way valve equipment, and collecting multi-source, heterogeneous mechanical equipment scene data to provide a reliable data source for the one-way valve digital twin system.

[0028] The dynamic visualization interface is used to display real-time operating parameter information of the physical entity system and the virtual twin model system, providing a convenient terminal for human-computer interaction. It also provides a data interface, which can provide multi-source data for deep data mining.

[0029] The health management system records all objects, models, and data throughout the entire operating cycle of the check valve, serving as a digital archive of the check valve. It reflects the valve's characteristics, behaviors, processes, and states at each stage of its lifecycle, achieving unified, information-based, and visualized management of the entire high-pressure check valve process. Based on the check valve's actual real-time operating data, the digital twin model is corrected through feedback. A dynamic evolution model of check valve performance degradation is established by integrating the Weibull proportional failure rate model, describing the changes in degradation states and parameters over time, as well as the transmission process of model uncertainties.

[0030] The predictive maintenance system incorporates transfer learning to aid in deep learning model training. It establishes a residual lifetime (RUL) transfer learning prediction model based on a Deep Residual Shrinkage Network (DRSN), improving the generalization ability and robustness of the RUL and enabling cross-domain transfer of the RUL prediction model. Based on this, an improved parallel solution algorithm is used to solve the DRSN prediction transfer model. Finally, using full lifecycle simulation data from the equipment's digital twin, the DRSN cross-domain transfer model is trained and applied to predict the RUL of check valves in different stages of high-pressure diaphragm pumps. The prediction results are displayed through a visual interface, enabling real-time monitoring and control of the equipment throughout its entire lifecycle.

[0031] In another specific embodiment, such as the one-way valve health management system based on digital twins described above, Figure 1 As shown in this embodiment, dynamic monitoring and fault diagnosis of the entire lifecycle of a check valve are performed. First, a digital twin virtual system is constructed based on the physical entity of the check valve. Multi-source data from the actual operation of the check valve is used to provide feedback correction to the digital twin, ensuring the high fidelity of the constructed twin model. Second, sensors are deployed at multiple points to collect vibration, pressure, and flow signals during operation. Failure mechanisms and domain knowledge are integrated to build a scene perception system for the digital twin of the check valve equipment. Then, simulation data from the virtual system and actual operating data are integrated to build a real-time interactive system, achieving virtual-real linkage. The virtual-real operating data is stored in a spatiotemporal database, and the entire lifecycle virtual simulation data of the check valve is exported from the spatiotemporal database. A fault model is trained using deep learning methods. Finally, a visual interface is constructed to display the real-time operating parameters of the physical entity system and the virtual twin model system, as well as the fault diagnosis results. A data interface is also provided to offer a convenient terminal for human-computer interaction.

[0032] The system architecture for dynamic monitoring and fault diagnosis of the entire life cycle of a check valve is as follows: Figure 2As shown, the constructed digital twin spans all stages of the high-pressure check valve's entire lifecycle, recording all objects, models, and data throughout the entire cycle. It serves as a digital archive of the high-pressure diaphragm pump, reflecting its characteristics, behaviors, processes, and states at each stage of its lifecycle. By integrating machine learning to train fault models, it achieves unified, information-based, visualized, and intelligent management of the entire high-pressure check valve process. Figure 3 Simultaneously, enabling the transfer, exchange, and sharing of models and data at each stage, and allowing the invocation of objects, models, and data from past stages while remaining in the current stage, provides a basis and guarantee for full-process quality traceability and continuous improvement of the management level of high-pressure diaphragm pump systems.

[0033] In another specific embodiment, such as the one-way valve health management system based on digital twins described above, Figure 1 As shown, this embodiment demonstrates predictive maintenance of a one-way valve driven by the fusion of digital twin and deep transfer learning. First, a one-way valve digital twin system is constructed as described in Embodiment 1. Then, transfer learning is introduced to aid deep learning model training, establishing a device remaining life (RUL) transfer learning prediction model based on a Deep Residual Shrinkage Network (DRSN). This improves the generalization ability and robustness of RUL. Utilizing the shared model structure and parameters between the digital twin and the physical entity, various types of faults are injected into the digital twin model of complex mechanical equipment, generating multiple fault and performance degradation data to form the source domain of the DRSN predictive transfer learning algorithm. This is used to train the predictive maintenance model, enabling cross-domain transfer of the device remaining life prediction model. Finally, the twin data and actual operating data are fused and used as input to the predictive maintenance model to predict the remaining life of the one-way valve under different operating conditions. This prediction is then visually displayed through a graphical interface, achieving intelligent human-machine interaction. The visualization interface of the one-way valve full lifecycle health management system is shown below. Figure 3 As shown.

[0034] Matters not covered in this invention are common knowledge.

[0035] The embodiments described above merely illustrate specific implementation methods of this application, and while the descriptions are detailed and specific, they should not be construed as limiting the scope of protection of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the technical solution of this application, and these modifications and improvements all fall within the scope of protection of this application.

Claims

1. A one-way valve lifecycle management system based on digital twins, characterized in that, include: The physical entity system is the actual one-way valve system; A virtual twin model system, corresponding to a physical entity system, is used to dynamically map the operating status of a one-way valve in real time. An information interaction system is used to enable bidirectional data transmission and fusion between physical entity systems and virtual twin model systems; A dynamic visualization interface, connected to the virtual twin model system and information interaction system, is used to display the real-time operating parameters of the physical entity and the virtual twin model; The health management system, connected to the information interaction system and the virtual twin model system, is used for status monitoring and fault diagnosis of check valves based on fused data; Predictive maintenance systems, connected to health management systems and virtual twin models, are used to predict the remaining lifespan of check valves based on historical and real-time data.

2. The one-way valve lifecycle management system based on digital twin as described in claim 1, characterized in that, The virtual twin model system includes: A physical model is a 1:1 geometric model of a physical entity system. The local linear model is constructed based on the operating mechanism of the one-way valve, including the fluid continuity equation, momentum equation and mathematical parameterization expression of liquid-solid two-phase flow; Behavioral models are parameterized representations of the control logic and external responses of physical entities, including electrical models, control models, and mechanical models.

3. The one-way valve lifecycle management system based on digital twins according to claim 2, characterized in that, The virtual twin model system uses a combination of data-driven and finite element model correction methods to update the model. Based on the system response data collected by sensors, the material properties, dimensional parameters or constraints of the model are dynamically corrected.

4. The one-way valve lifecycle management system based on digital twin as described in claim 1, characterized in that, The information interaction system includes: The intelligent sensing system is used to collect the operating parameters of a physical entity system throughout its entire life cycle. These operating parameters include vibration signals, pressure signals, and flow signals. The information fusion system is used to fuse multi-source heterogeneous data collected by the intelligent sensing system and to build a scene perception system and spatiotemporal database for one-way valve equipment.

5. The one-way valve lifecycle management system based on digital twin as described in claim 1, characterized in that, The health management system is configured as follows: Record all objects, models, and data throughout the entire lifecycle of the check valve to create a digital archive; By integrating the Weibull proportional failure rate model, a dynamic evolution model of the performance degradation of the one-way valve is established to describe its degradation state and the changing patterns of its parameters over time.

6. The one-way valve lifecycle management system based on digital twin according to claim 1, characterized in that, The predictive maintenance system is configured to use full lifecycle simulation data generated by the equipment's digital twin to construct a database of one-way valve wear evolution trend models.

7. A one-way valve lifecycle management system based on digital twins according to claim 6, characterized in that, The predictive maintenance system further includes a transfer learning prediction model for the remaining life of equipment based on a deep residual shrinking network (DRSN), used to achieve cross-domain prediction of the remaining life of check valves under different operating conditions.

8. A one-way valve lifecycle management system based on digital twins according to claim 7, characterized in that, An improved Parareal solution algorithm is used to solve the remaining lifetime transfer learning prediction model based on the Deep Residual Shrinking Network (DRSN).

9. A one-way valve lifecycle management system based on digital twins according to claim 1, characterized in that, The dynamic visualization interface combines augmented reality technology to overlay and display parameter information of physical entities and virtual twin models.

10. A method for full lifecycle health management of a one-way valve based on digital twins, characterized in that, This is implemented based on the digital twin-based one-way valve lifecycle management system as described in any one of claims 1-9.

Images