Deepwater jacket platform digital twin structure health management method

Abstract

The invention discloses a deepwater jacket platform digital twinning structure health management method, and belongs to the field of ocean platform structure integrity management and digital twinning operation and maintenance. The invention provides a digital twinning grading implementation strategy, and the integration and fusion of data are realized by constructing a deepwater jacket physical sensing end and a virtual mapping end and combining network transmission and a twinning algorithm. The core idea of the method is to construct a twin mapping inversion model of the deepwater jacket platform on the basis of state sensing data. Under the guidance of twinborn data and a twinborn model, intelligent production, management, operation and maintenance, decision making and full-life-cycle health management of the deepwater jacket platform are achieved. The method has a wide application prospect in the fields of fixed ocean platform structural integrity management and digital twin operation and maintenance.

Description

Technical Field

[0001] The present invention relates to a digital twin structural health management method for a deepwater jacket platform, and belongs to the field of marine platform structural integrity management and digital twin operation and maintenance. Background Art

[0002] As a tall fixed offshore platform, the deepwater jacket is characterized by large size, heavy weight, high flexibility and complex structure. It serves in deep sea areas with harsh environments.

[0003] Due to the unique structure and complex loads, deepwater jacket platforms often face multiple failure risks, such as strength damage and fatigue failure, posing significant challenges to structural safety. Research into digital structural health and operation and maintenance management is crucial.

[0004] Digital twin technology, an information-based, visual, and interactive digital structural health and operation management model, plays a vital role in the preventative maintenance of deepwater jacket platforms. Guided by twin data and twin models, it enables intelligent production, management, operation, maintenance, decision-making, and lifecycle health management of deepwater jackets. By building a digital twin model based on the on-site sensing system and establishing a virtual-to-real mapping relationship for the deepwater jacket, it provides long-term, reliable status prediction and safety warning services for the platform.

[0005] With the continuous advancement of sensor technology, monitoring techniques for the marine environment and the dynamic response of offshore platforms have matured. Furthermore, the rapid development of big data and artificial intelligence technologies has made real-time inversion and prediction of platform status possible. Therefore, this paper proposes a digital twin structural health management method for deepwater jacket platforms based on monitoring information from marine environmental loads and platform responses, combined with big data and artificial intelligence technologies. This method effectively provides platform status prediction and safety warning services. Summary of the Invention

[0006] To address these issues, the purpose of this patent is to provide a digital twin structural health management method for deepwater jacket platforms. This digital twin structural health management method, based on an on-site sensing system, structural dynamics simulation, and big data artificial intelligence technologies, establishes a hierarchical digital twin model and constructs a virtual-to-real mapping relationship for the deepwater jacket, thereby providing long-term and reliable status prediction and safety warning services for the platform.

[0007] The technical solution employed in this invention is a deepwater jacket platform digital twin structural health management system. The system comprises a monitoring terminal and a central processing unit (CPU). The monitoring terminal utilizes an Internet of Things (IoT) transmission node and a CPU. The central controller utilizes an industrial computer (IPC) and is electrically connected to a signal receiver. The signal receiver is electrically connected to the IoT transmission node within the monitoring terminal. The virtual mapping terminal includes data and network management, twin and twin model algorithms, and an IPC. The CPU includes a data receiver and an IPC, each electrically connected to the IoT transmission node. The IPC displays and stores the operating status of the twin structure through a built-in digital twin deepwater jacket platform structural health management system.

[0008] The deepwater jacket platform digital twin system is designed as a hierarchical technical approach, with a defined digital twin technology roadmap. The first level, model correction and precise mapping, aims to correct the model based on measured data and establish an accurate virtual mapping model. The second level, dynamics and hydrodynamics simulation and data fusion, conducts massive dynamics simulation analysis and further modifies the simulation results based on monitoring data to form a twin driven by the dynamics model. The third level, intelligent perception and predictive assessment, leverages the accumulated data from the second level and field prototype monitoring to achieve data-driven intelligent perception and response prediction for the platform, employing multiple deep learning models. This provides platform status prediction and safety warning services.

[0009] The digital twin structural health management system for a deepwater jacket platform comprises the following steps:

[0010] A. Install the monitoring terminal on the offshore platform. It is recommended that the anemometer and wave radar be installed at the edge of the deck, the acceleration sensor be installed near the center of gravity of the platform, and the fiber grating sensor be installed on the platform using a fixture to measure the force behavior of the rod above and below the water. It is recommended that the inclination sensor be placed at the four corners of the platform to measure the angle change of the platform. The current sensor can be installed underwater (such as acoustic current measurement) or at the edge of the platform deck (such as radar current measurement) according to its own measurement behavior.

[0011] B. The signal conditioning module formats the sensor data and passes it to the signal transmission module. The signal storage module stores and backs up the standardized monitoring information. The standardized monitoring information is then transmitted to the signal receiver, which then transmits it to the industrial computer.

[0012] C. Based on the measured data of the installed sensors, the parameters of the platform simulation model are modified.

[0013] D. Based on statistical analysis of years of measured sea conditions in the Liuhua Sea area, we extracted finite element simulation joint working conditions. Using dynamic simulation software, we performed extensive platform load-dynamic response calculations, creating a twin database of massive simulation data.

[0014] E. Construct a nonlinear interpolation algorithm for the twin database based on machine learning to realize the intelligent interpolation platform based on the platform's measured sea condition data and focus on the inversion results of the rod response.

[0015] F. Based on the on-site load and structural response monitoring sensors, the real-time status data of the platform is obtained, and the intelligent prediction of the platform's overall indicators based on deep learning, the intelligent prediction of the mechanical response of the focus rods, and the intelligent mapping model of the sensor network are constructed.

[0016] G. Based on the measured data input into the deepwater jacket platform digital twin structural health management system in the industrial computer, the platform intelligent mapping and response prediction are carried out in real time, providing status prediction and safety warning services for the platform. At the same time, the industrial computer will display and save the calculation results in real time.

[0017] The present invention has the following advantages:

[0018] 1. The present invention utilizes monitoring data derivation, intelligent algorithm drive and other functions to realize the digital twin of the deepwater jacket platform. Traditional structural health operation and maintenance methods mostly rely on manual inspections or simple sensor monitoring. The data acquisition efficiency is low and not comprehensive enough. Only obvious problems on the surface of the structure can be obtained through monitoring and detection, and it is difficult to obtain deep-level structural status information. The present invention continuously collects multi-dimensional parameters including stress, acceleration, waves, vibration, wind speed, wind direction, inclination, ocean current, etc. through a variety of sensors. Through the digital twin simulation model, the big data driven model conducts in-depth analysis and calculation of the real-time monitoring data collected by the sensors, which can capture subtle parameter changes that may threaten the structural safety of the jacket platform and issue early warnings accordingly. At the same time, the system can be continuously improved according to the accumulation of monitoring data, iterate the twin database, and continuously improve the intelligent level of real-time analysis and status perception of the jacket, providing a more scientific and accurate basis for operation and maintenance decisions, and helping to formulate reasonable maintenance plans in advance.

[0019] 2. The deepwater jacket platform digital twin system is built on a foundation of dynamic simulation and field monitoring. Considering the limited availability of initial monitoring data, a three-phase digital twin technology roadmap was developed, employing a hierarchical digital twin implementation approach. This tiered approach enables the deepwater jacket platform digital twin system to more efficiently manage data, equipment, and O&M tasks, improving system reliability and overall performance throughout each stage of the platform's lifecycle.

[0020] 3. The constructed digital twin system, combining structural health monitoring, simulation, and big data-driven technologies, effectively improves the robustness and reliability of the monitoring system and enhances its adaptability to extreme environmental conditions. Even if some sensors experience problems, the model's self-correction and data fusion ensure stable system operation and accurate monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a structural schematic diagram of the digital twin structural health management system for a deepwater jacket platform.

[0022] Figure 2 This is a schematic diagram of the physical sensing end of the digital twin structural health management system for a deepwater jacket platform.

[0023] Figure 3 It is a schematic diagram of the virtual mapping end of the digital twin structural health management system of a deepwater jacket platform. DETAILED DESCRIPTION

[0024] The present invention will be described in detail below in conjunction with the accompanying drawings. However, it should be understood that the accompanying drawings are only provided for a better understanding of the present invention and should not be understood as limiting the present invention.

[0025] Based on the service environment and structural characteristics of deepwater jacket platforms, simulation analysis of deepwater jacket stress behavior was conducted. Finite element numerical simulation methods were used to construct a platform model that more closely resembled actual service. The overall mass was adjusted based on the construction process and the weight of the monitored topsides. The pile-soil parameter curves, such as py and tz, were also modified to align the jacket structural model with the first three frequencies of field monitoring. This resulted in a high-fidelity simulation model, achieving first-level twinning.

[0026] Based on field-measured data on environmental loads such as wind, waves, and currents, the distribution characteristics of environmental parameters in the waters where deepwater jacket platforms are located were extracted, and marginal and joint probability models of these parameters were established, providing a data foundation and theoretical basis for the selection of load conditions. Outliers in the measured data were studied, and the performance of the outlier removal method was verified. Calculation and verification conditions were constructed based on the joint distribution of environmental loads, and extreme conditions were constructed based on the extreme value distribution model used in the platform design. This resulted in an environmental load database covering a wide range of sea conditions, from routine to extreme. Based on the constructed high-fidelity model and a large number of joint environmental load conditions obtained through statistical analysis, finite element simulation software was used to calculate the jacket structure response under different sea conditions, ultimately forming a twin database of platform responses covering a wide range of sea conditions, from routine to extreme. A nonlinear interpolation algorithm for the twin database was constructed based on machine learning, which intelligently interpolated the member response inversion results based on the platform's measured sea condition data, achieving secondary twinning.

[0027] Based on deep learning, an intelligent prediction algorithm for the overall response of the platform and the stress behavior of key members is constructed. The complex nonlinear mapping relationship between deepwater jacket loads and overall indicators is learned to improve the accuracy of member stress predictions, forming a high-precision overall indicator prediction model to further ensure structural safety. Furthermore, the problem of underwater sensors in deepwater jackets being unable to operate in place for a long time is addressed, which easily leads to the loss of measurement point information, seriously affecting the reliability and accuracy of deepwater jacket safety warnings. A deep learning algorithm is used to train and establish an intelligent mapping model for the sensor network. This allows the data of the faulty measurement point to be mapped and restored through other in-place sensors when some sensors fail, thereby improving the long-term reliability of the on-site monitoring system, ensuring the safe and stable operation of the platform, and achieving three-level twinning.

[0028] The twin system is encapsulated in the working condition machine. The signal receiver receives monitoring information from the signal transmission module and transmits it to the industrial computer. The deepwater jacket platform digital twin structural health management system performs real-time intelligent mapping and response prediction for the platform, providing status prediction and safety warning services. The industrial computer also displays and saves the calculation results in real time.

Claims

1. A deepwater jacket platform structural health management method based on digital twin, characterized by: The system used in the method includes a physical sensing end and a virtual mapping end; the physical sensing end includes a monitoring end and a central processor, the central controller is an industrial computer electrically connected to a signal receiver; the signal receiver is electrically connected to an Internet of Things transmission node in the monitoring end; The virtual mapping end includes data and network management, twin and twin model algorithms, and an industrial computer. The industrial computer also includes a deepwater jacket platform structural health management system with a built-in digital twin. The specific steps of this method are: A. Install a monitoring terminal on the offshore platform; the monitoring terminal uses an IoT transmission node to electrically connect to the sensor system; B. The sensor starts working according to the preset sampling frequency and acquisition mode; C. The signal conditioning module unifies the format and passes the information to the signal transmission module, while the signal storage module performs data backup; D. The signal transmission module transmits the monitoring information in a unified format to the signal receiver, which then transmits it to the industrial computer; E. The industrial computer uses its built-in system to process sensor data, combines it with computing to achieve virtual mapping, assess platform status, and provide support for operational and maintenance decisions. It also displays and saves the calculation results in real time.

2. The method for structural health management of a deepwater jacket platform based on digital twin according to claim 1, characterized in that: The sensor system includes marine environment load sensors such as wind, wave, and current, acceleration sensors, fiber grating sensors, and inclination sensors.

3. The deepwater jacket platform digital twin structural health management method according to claim 1, characterized in that: The method includes the process of sensor signal acquisition and processing, data set preparation, twin model establishment, simulation model correction, algorithm model hyperparameter selection, model training and error analysis; The specific steps include: (1) Based on the platform's service environment and structural characteristics, a simulation analysis of the deepwater jacket's stress behavior was conducted, and a high-precision platform model was constructed using the finite element numerical method; (2) Based on the construction process and the weight of the monitored upper assembly, the platform overall mass, pile-soil PY, TZ and other curves are modified to be consistent with the on-site monitoring frequency, and a high-fidelity simulation model is established; (3) Combine environmental load data and environmental parameters of the sea area where the platform is located to study its distribution characteristics, establish a probability model and conduct simulation; (4) Construct a nonlinear interpolation algorithm through machine learning and perform intelligent interpolation based on measured sea state data; (5) Using deep learning to build an intelligent prediction model for the overall response of the platform and the stress behavior of key members, and learn the nonlinear relationship between load and response; (6) Integrate the twin system into the working machine to perform intelligent mapping and response prediction in real time.

Images