Gas turbine power generation system three-dimensional model construction method based on digital twinning

Abstract

The invention provides a gas turbine power generation system three-dimensional model construction method based on digital twinning, and belongs to the technical field of digital twinning. And the correctness of the new logic is ensured by updating the simulation calculation core logic and performing compatibility verification. And performing parallel operation on the simulation calculation core logic before and after updating in the digital twin model, performing simulation aiming at the typical operation condition of the gas turbine, and performing difference evaluation on the simulation calculation core logic before and after updating based on a simulation output result. If the difference evaluation result is larger than a preset difference threshold value, the updated simulation calculation core logic is adjusted and evaluation is conducted again till the difference evaluation result is smaller than the preset difference threshold value, and finally the updated simulation calculation core logic is executed, or when the requirement cannot be met, the simulation calculation core logic is rolled back to the simulation calculation core logic before updating. According to the method, the problem that a digital twin model is inconsistent with a physical entity in the prior art is effectively solved, and the operation risk and economic loss possibly caused by model deviation are overcome.

Description

Technical Field

[0001] This disclosure relates to the field of digital twin technology, and in particular to a method for constructing a three-dimensional model of a gas turbine power generation system based on digital twins. Background Technology

[0002] In modern industry, gas turbine power generation systems are crucial energy equipment. To ensure their efficient and stable operation and enable predictive maintenance, digital twin technology is widely used. Digital twins create virtual mappings and analyses of equipment status by constructing three-dimensional digital models of physical entities. However, in practical applications, due to factors such as long-term equipment operation, environmental changes, and software updates, inconsistencies may arise between the digital twin model and the physical entity, posing challenges to accurate monitoring and fault diagnosis. Especially in complex systems like gas turbines, which demand extremely high precision and reliability, even minute model deviations can accumulate and lead to serious operational risks and economic losses.

[0003] For example, during service, a gas turbine casing inevitably experiences numerous start-stop cycles and significant power load fluctuations. These operations cause localized areas of the casing to repeatedly experience stress cycles due to thermal expansion and contraction. Over time, this continuous physical stress can lead to minute but persistent relative displacements of several physical reference markers used for spatial registration on the casing. These displacements are typically very subtle, imperceptible to the naked eye, and the displacement of a single marker is often within the error range of conventional testing tools. Therefore, these cumulative, gradual displacements often go undetected during routine equipment inspections and maintenance. This hidden physical change foreshadows the distortion of subsequent digital twin models.

[0004] When the "spatial registration program for 3D point cloud data" processes new point cloud data, it aligns the data based on physical reference points that are no longer perfectly accurate and have undergone minor displacement. Although the program's alignment algorithm is sophisticated and can handle a certain degree of data noise and local bias, the input references—the physical reference points themselves—involve systematic and cumulative biases, leading to systematic geometric distortions in local areas of the newly generated digital twin 3D model. This distortion is usually small enough not to trigger the "abnormality" warning thresholds set by the model building system, as these thresholds are typically designed to capture more significant geometric defects rather than these minor deviations caused by the cumulative displacement of the reference points. Therefore, the problem remains hidden within the model and is not automatically identified by the system.

[0005] Meanwhile, critical components inside the gas turbine, such as the "high-temperature turbine blades," may also experience progressive physical degradation under extreme operating conditions. When the digital twin system simulates the thermodynamic behavior and structural stress of these blades, the boundary conditions and internal flow field estimates it calculates are affected by the aforementioned erroneous geometric information. This bias prevents the digital twin system from accurately capturing the accelerated creep process of the blades, leading to a deviation between the assessment of the blade's health and the actual situation.

[0006] During this period, the digital twin platform underwent a routine software maintenance upgrade. This upgrade included optimization of the mesh generation logic for finite element analysis, primarily aimed at improving computational efficiency. However, this logic optimization, in its design, did not fully consider the potential systematic geometric distortions in the underlying 3D model caused by the displacement of the aforementioned physical reference points. When the new mesh generation logic was applied to this model with subtle distortions, stress analysis of the "high-temperature turbine blade" resulted in unexpected instability or deviations in the simulation results under specific operating conditions. These unstable simulation results further masked the true degradation trend of the blade, amplifying the risk of failure, as the simulation data now provided by the system is not only inaccurate but also potentially misleading.

[0007] Ultimately, considering the geometric distortion of the aforementioned digital twin 3D model, the inaccurate simulation of accelerated creep of the turbine high-temperature blades, the engineers' experience-based judgments, and the instability of simulation results due to the mismatch between software logic updates and the underlying model, the digital twin system as a whole failed to timely and accurately identify that the actual structural health of the "turbine high-temperature blades" had reached a critical point. The system's predicted maintenance schedule was severely out of sync with the actual situation. This resulted in a significant weakening of the structural integrity of critical components, yet the system still considered them to be in a safe operating state. Summary of the Invention

[0008] To address this issue, this application discloses a method for constructing a three-dimensional model of a gas turbine power generation system based on digital twins, aiming to solve the problem that inconsistencies between digital twin models and physical entities lead to challenges in accurate monitoring and fault diagnosis.

[0009] This invention proposes a method for constructing a three-dimensional model of a gas turbine power generation system based on digital twins, comprising: S1: Obtain the digital twin model of the gas turbine power generation system, perform topology calibration and parameter updates on the digital twin model of the gas turbine power generation system, and determine whether the core logic of the simulation calculation of the digital twin model of the gas turbine power generation system needs to be updated. S2: If it is determined that the digital twin model of the gas turbine power generation system needs to update the core logic of simulation calculation, update the core logic of simulation calculation and verify the compatibility of the core logic of simulation calculation before and after the update; among them, updating the core logic of simulation calculation includes at least updating the mesh generation algorithm and adjusting the physical mediator parameters; S3: After passing compatibility verification, the core logic of simulation calculation before and after the update is run in parallel in the digital twin model of the gas turbine power generation system, and simulation is performed for typical operating conditions of the gas turbine. S4: Based on the simulation output results, evaluate the differences in the core simulation calculation logic before and after the update; S5: If the difference assessment result is greater than the preset difference threshold, adjust the updated simulation calculation core logic, execute steps S2-S4, and perform the difference assessment again. S6: If the updated simulation calculation core logic is adjusted to make the difference evaluation result less than the preset difference threshold, then the updated simulation calculation core logic is executed; otherwise, the simulation calculation core logic is rolled back to the original logic.

[0010] This application provides a systematic method for ensuring the accuracy and reliability of the new logic when updating the core logic of the simulation calculation in the digital twin model of a gas turbine power generation system. This effectively solves the problem of inconsistency between the model and the physical entity, thereby improving the accuracy monitoring and fault diagnosis capabilities of the digital twin model.

[0011] Topology calibration and parameter updates are performed on the digital twin model of the gas turbine power generation system, including: Dynamic reference topology self-calibration is performed on the digital twin model of the gas turbine power generation system to correct the geometric distortion of the physical reference points on the gas turbine casing; The digital twin model of the gas turbine power generation system is used to perform reverse inference of the operating status and update the material property parameters to reflect the progressive physical degradation of key components of the gas turbine.

[0012] This application's solution, through dynamic benchmark topology self-calibration and reverse inference of operating status and material property parameter updates, can more accurately correct model geometric distortion and reflect component degradation, thereby further improving the accuracy and real-time performance of the digital twin model.

[0013] Based on this, dynamic reference topology self-calibration is performed on the digital twin model of the gas turbine power generation system to correct the geometric distortion of the physical reference points on the gas turbine casing, including: Continuously collect micro-geometric data of all physical reference points on the gas turbine casing used for spatial registration and their surrounding areas; Establish a dynamic reference point topology network and continuously calculate the relative position changes between physical reference points in the reference point topology network; When a cumulative change exceeding a preset micro-deformation threshold is detected relative to any physical reference point with respect to its initial position or adjacent physical reference points, the coordinate transformation matrix and the weight parameters of the registration algorithm used to align point cloud data with the model within the digital twin model of the gas turbine power generation system are adjusted. By adjusting the settings, the subsequent point cloud registration process automatically compensates for the actual micro-displacement of the physical reference point that has undergone cumulative changes, thereby eliminating geometric distortion caused by reference point drift at the global level.

[0014] This application's solution continuously monitors the changes in the microscopic geometric data and topological relationships of the physical reference point and dynamically adjusts the registration algorithm parameters, which can effectively eliminate geometric distortion caused by reference point drift and further improve the accuracy and robustness of model topology calibration.

[0015] Furthermore, the digital twin model of the gas turbine power generation system is used for reverse inference of operating status and updating of material property parameters to reflect the progressive physical degradation of key gas turbine components, including: Receive real-time operating data from key components of the gas turbine, including at least the turbine blade area; The real-time operating data from key components of the gas turbine are compared with the simulation results of the digital twin model of the gas turbine power generation system under the current operating conditions. When a deviation is detected between the two, the changes in the physical parameters that caused the deviation are deduced, and the material properties of the key components of the gas turbine are updated to reflect the actual deterioration state of the key components of the gas turbine.

[0016] This application's solution compares real-time running data with simulation results and inversely infers changes in physical parameters, which can accurately reflect the progressive physical degradation of key components, thereby enabling the digital twin model to more realistically simulate the operating state of the physical entity.

[0017] Furthermore, the core simulation logic before and after the update is run in parallel within the digital twin model of the gas turbine power generation system, and simulations are performed for typical operating conditions of the gas turbine, including: Before starting the core logic of simulation calculation before and after parallel operation update, obtain the internal state snapshot of the digital twin model of the gas turbine power generation system. The internal state snapshot includes at least all variable internal state information, including material properties and geometric parameters. The internal state snapshot is used as the initial model state input when running the core logic of simulation calculation before and after the update in the digital twin model of the gas turbine power generation system. During the parallel simulation process, it is ensured that the core logic of simulation calculation before and after the update is calculated based on the internal state snapshot in their respective simulation processes. After the simulation is completed, the output results of the simulation calculation core logic before and after the update are compared in the digital twin model of the gas turbine power generation system.

[0018] This application's solution improves the accuracy and reliability of difference assessment by obtaining an internal state snapshot and using it as the initial input for parallel simulation, ensuring that the logic before and after the update is compared under the same benchmark.

[0019] Based on the above, the simulation test scenarios in the digital twin model of the gas turbine power generation system, which involves running the core logic of the simulation calculations before and after the update in parallel, and simulating typical operating conditions of the gas turbine, include: Steady-state full-load operation: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system, respectively simulating the steady-state operation state of the gas turbine at rated power output; Rapid start-stop cycle: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system, respectively cyclically simulates the complete process from cold start to full load operation and then to shutdown; Transient load change: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system simulates the operation of the gas turbine power generation system when the load suddenly increases from the first load to the second load, or when the load suddenly drops from the second load to the first load within a preset first time interval. The first load is less than the second load.

[0020] This application comprehensively verifies the performance of the core simulation logic before and after the update under different complex scenarios by covering typical operating conditions such as steady-state full-load operation, rapid start-stop cycle, and transient load change, thereby ensuring the comprehensiveness and reliability of the model update.

[0021] Preferably, based on the simulation output results, a difference assessment is performed on the core simulation calculation logic before and after the update, including: Obtain the first key output parameter of the core logic output of the simulation calculation before the digital twin model of the gas turbine power generation system is updated, and the second key output parameter of the core logic output of the simulation calculation after the update; among them, Key output parameters include turbine blade stress data, temperature gradient distribution data, and vibration mode data; Based on the first and second key output parameters, the differences in the core simulation calculation logic before and after the update are evaluated.

[0022] This application proposes a scheme that selects key output parameters such as turbine blade stress, temperature gradient distribution, and vibration modes for difference evaluation. This approach can comprehensively and accurately reflect the performance differences of the core simulation calculation logic before and after the update, thus providing a strong basis for subsequent adjustments.

[0023] Furthermore, based on the first and second key output parameters, the differences in the core simulation logic before and after the update are evaluated, including: When turbine blade stress data is used as a key output parameter, the difference between the first turbine blade stress data output by the simulation calculation core logic before the digital twin model of the gas turbine power generation system is updated and the second turbine blade stress data output by the simulation calculation core logic after the update is compared with the ratio of the first turbine blade stress data to the preset turbine blade stress data deviation threshold. The comparison result is used as the blade turbine stress data deviation evaluation index. When temperature gradient distribution data is used as a key output parameter, the root mean square of the first temperature gradient distribution data when running the core logic of the simulation calculation before the update of the digital twin model of the gas turbine power generation system and the root mean square of the second temperature gradient distribution data when running the core logic of the simulation calculation after the update are compared with the preset temperature gradient distribution data deviation threshold. The comparison result is used as the temperature gradient distribution deviation evaluation index. When vibration modal data is used as a key output parameter, the difference between the natural frequency in the first vibration modal data when the core logic of the simulation calculation before the update of the digital twin model of the gas turbine power generation system is run and the natural frequency in the second vibration modal data when the core logic of the simulation calculation after the update is run is compared with a preset vibration modal frequency deviation threshold. The comparison result is used as a vibration modal deviation evaluation index.

[0024] This application proposes specific quantitative evaluation indicators and threshold comparison methods for different key output parameters (turbine blade stress, temperature gradient distribution, vibration modes), which makes the difference evaluation more refined and objective, thereby improving the accuracy and guidance of the evaluation results.

[0025] Based on the above, the updated core logic of the simulation calculation has been adjusted, including: Adjust mesh density parameters: Adjust the parameters of local mesh density in the updated simulation calculation core logic to increase or decrease the number of meshes in a specified area until the turbine blade stress calculation results are consistent with the old logic or physical verification results; Adjust the solver convergence parameters: If the bias is related to numerical stability, adjust the number of iterations, relaxation factor, or convergence criterion in the new physics solver to ensure the stability of the computation process and the accuracy of the results. Update material model coefficients: If the updated simulation calculation core logic interprets material behavior differently, adjust the material model coefficients used in the digital twin model to compensate for the systematic deviations caused by the updated simulation calculation core logic.

[0026] This application's solution effectively addresses computational deviations caused by model updates by finely adjusting mesh density, solver convergence parameters, and material model coefficients, thereby ensuring the accuracy and reliability of the updated simulation calculation core logic.

[0027] Furthermore, if the updated simulation calculation core logic is adjusted to make the difference evaluation result less than the preset difference threshold, then the updated simulation calculation core logic is executed; otherwise, it is rolled back to the simulation calculation core logic before the update, including: If the adjusted and updated simulation calculation core logic fails to make the difference assessment result less than the preset difference threshold, a deviation report will be generated. The deviation report includes the nature of the difference, the magnitude of the difference, and the affected components or performance parameters; Based on the deviation report, determine the handling method; the handling methods include: manually adjusting the new logic parameters so that the difference assessment result is less than the preset difference threshold; rolling back to the old logic version, executing the simulation calculation core logic before the update, and conducting physical experiments for verification.

[0028] This application provides a flexible and comprehensive solution for situations where automatic adjustment cannot reach the preset threshold by introducing deviation reports and multiple handling methods, thereby ensuring the robustness and security of the model update process.

[0029] Unlike existing technologies, the present invention provides a method for constructing a 3D model of a gas turbine power generation system based on digital twins. By acquiring a digital twin model of the gas turbine power generation system and performing topology calibration and parameter updates, it can reflect the actual state of the physical entity in a timely manner. When it is determined that the core simulation logic of the digital twin model needs to be updated, this application ensures the correctness of the new logic by updating the core simulation logic and performing compatibility verification. Subsequently, the core simulation logic before and after the update is run in parallel in the digital twin model, and simulations are performed for typical operating conditions of the gas turbine. The difference between the core simulation logic before and after the update is evaluated based on the simulation output results. If the difference evaluation result is greater than a preset difference threshold, the updated core simulation logic is adjusted and evaluated again until the difference evaluation result is less than the preset difference threshold. Finally, the updated core simulation logic is executed, or if the requirements cannot be met, the system rolls back to the original core simulation logic. This application effectively solves the problem of inconsistency between the digital twin model and the physical entity in existing technologies, overcoming the operational risks and economic losses that may be caused by model deviations.

[0030] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0031] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a method for constructing a three-dimensional model of a gas turbine power generation system based on digital twins, provided by the present invention. Detailed Implementation

[0032] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0033] In modern industry, gas turbine power generation systems are crucial energy equipment. To ensure their efficient and stable operation and enable predictive maintenance, digital twin technology is widely used. Digital twins create virtual mappings and analyses of equipment status by constructing three-dimensional digital models of physical entities. However, in practical applications, due to factors such as long-term equipment operation, environmental changes, and software updates, inconsistencies may arise between the digital twin model and the physical entity, posing challenges to accurate monitoring and fault diagnosis. Especially in complex systems like gas turbines, which demand extremely high precision and reliability, even minute model deviations can accumulate and lead to serious operational risks and economic losses.

[0034] like Figure 1 As shown, this application proposes a method for constructing a three-dimensional model of a gas turbine power generation system based on digital twins, including: S1: Obtain the digital twin model of the gas turbine power generation system, perform topology calibration and parameter updates on the digital twin model of the gas turbine power generation system, and determine whether the core logic of the simulation calculation of the digital twin model of the gas turbine power generation system needs to be updated.

[0035] After acquiring the digital twin model of the gas turbine power generation system, topology calibration and parameter updates are required. Topology calibration can be achieved by manually comparing the 3D scan data with the model and adjusting geometric deviations in the model. Parameter updates can be based on historical operating data, using expert judgment or simple statistical analysis methods to periodically adjust material properties, efficiency parameters, and other parameters in the model. After calibration and updates are completed, based on a preset strategy or user input, it is determined whether the current digital twin model of the gas turbine power generation system needs an update to its core simulation calculation logic. For example, it can be set to check for new simulation algorithms every six months, or trigger an update request when the deviation between the model simulation results and actual operating data exceeds a certain threshold.

[0036] S2: If it is determined that the digital twin model of the gas turbine power generation system needs to update the core logic of simulation calculation, update the core logic of simulation calculation and verify the compatibility of the core logic of simulation calculation before and after the update; wherein, updating the core logic of simulation calculation includes at least updating the mesh generation algorithm and adjusting the physical mediator parameters.

[0037] When the system determines that the core simulation logic needs updating, it will perform the corresponding update operation. Updating the mesh generation algorithm can be done by manually modifying the mesh generation rules or replacing it with a new mesh generation tool. Adjusting the physics mediator parameters can be done by manually inputting new parameter values ​​or selecting a predefined parameter configuration file. After the update is complete, compatibility verification of the core simulation logic before and after the update is required. For example, a series of preset simple test cases can be run to check whether the old and new logic produce the same or similar outputs under the same input, to ensure that the basic functions have not been broken.

[0038] S3: After passing compatibility verification, the core logic of simulation calculation before and after the update is run in parallel in the digital twin model of the gas turbine power generation system, and simulation is performed for typical operating conditions of the gas turbine.

[0039] After compatibility verification, to comprehensively evaluate the update's effectiveness, the system will simultaneously run the core simulation logic before and after the update within the digital twin model of the gas turbine power generation system. Parallel operation can be achieved by loading the old and new logic separately in different computation threads or processes. Simulations will be performed for typical gas turbine operating conditions; for example, the steady-state operating condition of the gas turbine under rated load can be selected, or a simple start-stop cycle can be simulated.

[0040] S4: Based on the simulation output results, evaluate the differences in the core simulation calculation logic before and after the update.

[0041] After the simulation is complete, the system collects the output results of the core simulation logic before and after the update and performs a difference assessment. The difference assessment can be done by directly comparing the numerical differences of key output parameters, such as comparing the temperature or stress values ​​at a specific point on a turbine blade. Alternatively, visualization tools can be used to overlay the simulation results of the old and new logics, allowing the differences to be observed visually.

[0042] S5: If the difference assessment result is greater than the preset difference threshold, adjust the updated simulation calculation core logic, execute steps S2-S4, and perform the difference assessment again.

[0043] If the difference assessment results show a significant difference between the updated simulation core logic and the original logic, and this difference exceeds a preset difference threshold, then the updated simulation core logic needs to be adjusted. Adjustments may include manually modifying the parameters of the mesh generation algorithm or fine-tuning certain coefficients of the physics mediator. After adjustment, the system will re-execute steps S2 to S4, i.e., perform compatibility verification, parallel simulation, and difference assessment again, until the difference meets the requirements.

[0044] S6: If the updated simulation calculation core logic is adjusted to make the difference evaluation result less than the preset difference threshold, then the updated simulation calculation core logic is executed; otherwise, the simulation calculation core logic is rolled back to the original logic.

[0045] After multiple adjustments and evaluations, if the difference assessment result of the updated simulation calculation core logic is ultimately less than the preset difference threshold, the update is considered successful, and the updated simulation calculation core logic is enabled. Conversely, if multiple adjustments still fail to meet the difference requirements, to avoid introducing potential risks, the system will roll back to the simulation calculation core logic before the update, continue using the old logic for simulation, and may trigger further analysis or manual intervention.

[0046] The overall working principle of this application lies in ensuring a high degree of consistency between the digital twin model and the physical entity through a continuous calibration and update mechanism. When the core logic of the simulation calculation needs to be updated, this application does not directly replace it, but introduces a rigorous verification process. First, compatibility verification ensures that the basic functions of the old and new logics are not compromised. Then, the old and new logics are run in parallel in the digital twin model, and simulations are performed for typical operating conditions. This allows for a comprehensive and detailed comparison of the performance of the old and new logics without affecting the operation of the existing system. The difference assessment based on the simulation output results provides a quantitative basis for judging the effectiveness of the new logic. If the difference exceeds a preset threshold, the new logic is iteratively adjusted and re-evaluated until the accuracy requirements are met. This iterative optimization process ensures the reliability and accuracy of the new logic. Finally, based on the evaluation results, a decision is made on whether to enable the new logic or roll back to the old logic, thus forming a closed-loop, adaptive digital twin model maintenance and update system. Each step is closely linked, jointly ensuring the high fidelity and reliability of the gas turbine power generation system digital twin model during long-term operation.

[0047] The core innovation of this application lies in its introduction of compatibility verification, parallel simulation, and iterative adjustment and evaluation mechanisms for the core simulation logic. Traditional methods, when updating the simulation logic of a digital twin model, often involve direct replacement and limited testing. This can lead to the introduction of unknown errors or incompatibility with existing systems, affecting the model's accuracy and reliability. In contrast, this application ensures the stability of basic functions by performing compatibility verification between the core simulation logic before and after the update. More importantly, this application runs the core simulation logic before and after the update in parallel within the digital twin model and simulates typical operating conditions of a gas turbine. This allows for a comprehensive and detailed comparison of the performance differences between the old and new logic before practical application. This parallel simulation mechanism not only improves verification efficiency but also reduces the risks caused by problems introduced by the new logic. Furthermore, the difference evaluation and iterative adjustment mechanism based on simulation output results allows the new logic to be finely optimized until it meets the preset accuracy requirements. This rigorous verification and optimization process significantly improves the accuracy, reliability, and robustness of the digital twin model of the gas turbine power generation system, providing stronger technical support for predictive maintenance and fault diagnosis of gas turbines.

[0048] In some embodiments of this application, the steps of topology calibration and parameter updating of the digital twin model of the gas turbine power generation system may specifically include the following: Dynamic reference topology self-calibration is performed on the digital twin model of the gas turbine power generation system to correct the geometric distortion of the physical reference points on the gas turbine casing; The digital twin model of the gas turbine power generation system is used to perform reverse inference of the operating status and update the material property parameters to reflect the progressive physical degradation of key components of the gas turbine.

[0049] Specifically, dynamic reference topology self-calibration involves continuously monitoring and adjusting the spatial relationships of physical reference points on the gas turbine casing to ensure a high degree of geometric consistency between the digital twin model and the actual physical entity. Its purpose is to correct for minute displacements or deformations of the physical reference points caused by long-term operation, thermal stress, vibration, and other factors, thereby eliminating the impact of these geometric distortions on model accuracy.

[0050] The reverse inference of operating status and updating of material property parameters involve analyzing real-time operating data of key components of the gas turbine to deduce internal physical changes, such as progressive degradation like material fatigue, wear, and corrosion. Based on these inferences, the corresponding material property parameters in the digital twin model are updated in real time to ensure that the model accurately reflects the current actual state and performance degradation trend of the physical components. The aim is to synchronize the digital twin model not only geometrically but also physically with the actual gas turbine, providing more realistic and reliable foundational data for subsequent simulation calculations.

[0051] This application introduces dynamic reference topology self-calibration and operational state inverse inference with material property parameter updates, enabling the digital twin model of a gas turbine power generation system to more accurately reflect the real-time state of the physical entity. The dynamic reference topology self-calibration mechanism ensures the accuracy of the model's geometry, continuously correcting geometric distortions at physical reference points and avoiding discrepancies between the model and reality caused by physical deformation. Simultaneously, the operational state inverse inference and material property parameter update mechanism focuses on the intrinsic degradation of physical components. Through real-time data analysis and parameter adjustments, the digital twin model can capture and reflect the progressive physical degradation of key components. Therefore, the digital twin model maintains a high degree of consistency with the actual gas turbine not only at the macroscopic geometric level but also at the microscopic material property level, providing a more solid and accurate data foundation for subsequent simulation calculations.

[0052] In order to correct the geometric distortion of the physical reference points on the gas turbine casing when performing topology calibration and parameter updates on the digital twin model of the gas turbine power generation system, this application proposes a more detailed implementation method, including: Continuously collect micro-geometric data of all physical reference points on the gas turbine casing used for spatial registration and their surrounding areas; Establish a dynamic reference point topology network and continuously calculate the relative position changes between physical reference points in the reference point topology network; When a cumulative change exceeding a preset micro-deformation threshold is detected relative to any physical reference point with respect to its initial position or adjacent physical reference points, the coordinate transformation matrix and the weight parameters of the registration algorithm used to align point cloud data with the model within the digital twin model of the gas turbine power generation system are adjusted. By adjusting the settings, the subsequent point cloud registration process automatically compensates for the actual micro-displacement of the physical reference point that has undergone cumulative changes, thereby eliminating geometric distortion caused by reference point drift at the global level.

[0053] Specifically, continuously acquiring microscopic geometric data of all physical reference points on the gas turbine casing used for spatial registration and their surrounding areas refers to periodically or continuously acquiring data from pre-defined physical reference points on the gas turbine casing and their adjacent areas using high-precision sensors (such as laser scanners and structured light sensors) for three-dimensional spatial positioning and model alignment. This microscopic geometric data includes, but is not limited to, the three-dimensional coordinates of points, surface normals, and texture information, with the aim of providing foundational data for subsequent topological network establishment and deformation detection.

[0054] The establishment of a dynamic reference point topology network, and the continuous calculation of the relative positional changes between physical reference points within this network, can be understood as constructing a mathematical model or data structure representing the geometric connections and relative positional relationships between physical reference points based on collected microscopic geometric data. This network can track the changes in the historical position of each physical reference point relative to other reference points or itself in real-time or near real-time, for example, by calculating Euclidean distance, angle, or relative displacement vectors. Its purpose is to quantify and monitor the minute deformations that may occur in the gas turbine casing during operation.

[0055] In practical applications, when a cumulative change exceeding a preset micro-deformation threshold is detected relative to any physical reference point's initial position or adjacent physical reference points, the coordinate transformation matrix and the weight parameters of the registration algorithm used to align point cloud data with the model within the digital twin model of the gas turbine power generation system are adjusted. This means that once the position drift or cumulative deformation of a physical reference point exceeds the system's preset tolerance range, the system will automatically trigger a correction to the registration mechanism within the digital twin model. The adjustment of the coordinate transformation matrix aims to correct global or local coordinate system deviations between the point cloud data and the digital twin model, while the adjustment of the registration algorithm's weight parameters optimizes the trust level for different reference points or regions during the registration process to adapt to changes in the actual physical structure. The goal is to ensure that the point cloud data can be accurately aligned with the deformed physical entity model.

[0056] Therefore, by adjusting the data, the subsequent point cloud registration process automatically compensates for the actual micro-displacements of the physical reference points that have undergone cumulative changes, eliminating geometric distortion caused by reference point drift at the global level. This means that after the above adjustments, when the digital twin model receives new point cloud data for registration, it can intelligently identify and correct deviations caused by deformation of the physical reference points, thereby eliminating geometric inaccuracies caused by the instability of these reference points at the entire model level, ensuring that the digital twin model always maintains a high degree of consistency with the latest geometric state of the gas turbine physical entity.

[0057] The proposed solution continuously collects microscopic geometric data of physical reference points on the gas turbine casing and establishes a dynamic topological network to monitor the relative positional changes between these reference points. Once cumulative deformation exceeding a preset threshold is detected, the system intelligently adjusts the coordinate transformation matrix and the weight parameters of the registration algorithm within the digital twin model. This dynamic adjustment mechanism enables the subsequent point cloud registration process to automatically compensate for the actual micro-displacements of the physical reference points, thereby effectively eliminating geometric distortion caused by reference point drift at the global level and ensuring that the digital twin model accurately reflects the latest geometric state of the gas turbine physical entity.

[0058] To more accurately reflect the progressive physical degradation of key components of a gas turbine, this application further proposes a specific method for inverse inference of operating status and updating of material property parameters of a digital twin model of a gas turbine power generation system. This includes: First, real-time operational data from key components of the gas turbine is received. These key components include, at least, the turbine blade area. Real-time operational data may include, but is not limited to, sensor data on temperature, pressure, vibration, stress, speed, and flow rate. This data can directly or indirectly reflect the operating status and health of key components. As a core component of the gas turbine that withstands high temperature, high pressure, and high-speed airflow, the turbine blade area's deterioration has a decisive impact on the overall performance and lifespan of the gas turbine; therefore, data collection from this area is particularly important.

[0059] Secondly, real-time operating data from key components of the gas turbine are compared with the simulation results of the digital twin model of the gas turbine power generation system under current operating conditions. This comparison aims to identify discrepancies between the physical entity and the digital twin model. The comparison can be achieved through various data analysis techniques, such as statistical analysis, machine learning algorithms, or deviation detection methods based on the physical model.

[0060] Furthermore, when a deviation is detected between the two, the changes in the physical parameters causing the deviation are deduced, and the material properties of key gas turbine components are updated to reflect the actual degradation state of these components. The occurrence of a deviation usually indicates some degree of degradation or performance decline in the physical entity. Through reverse derivation, the specific changes in physical parameters causing these deviations can be identified, such as reduced material strength, changes in elastic modulus, and accumulation of fatigue damage. Once these changes are derived, the material property parameters of the corresponding key components in the digital twin model can be updated accordingly. This allows the digital twin model to more accurately simulate the current performance and behavior of the physical entity, truly reflecting its progressive physical degradation.

[0061] This application's solution constructs a closed-loop feedback mechanism by continuously monitoring the real-time operating data of key gas turbine components and comparing it in real-time with the simulation results of a digital twin model. When discrepancies arise between the physical entity and the digital twin model, these discrepancies are considered signals of degradation in the physical entity. By performing reverse analysis and derivation on these discrepancies, the root cause of the discrepancies—namely, changes in the material properties of key components—can be identified. Consequently, the material property parameters in the digital twin model are dynamically updated, enabling the model to accurately reflect the actual health status and degree of degradation of the physical entity in real time. This mechanism ensures that the digital twin model maintains a high degree of consistency with the physical entity, thus providing a reliable foundation for subsequent simulation, prediction, and decision-making.

[0062] In the embodiments of this application, detailed steps are provided regarding the specific implementation of running the core logic of simulation calculations before and after the update in parallel in the digital twin model of the gas turbine power generation system and simulating typical operating conditions of the gas turbine.

[0063] Specifically, the core simulation logic before and after the update is run in parallel within the digital twin model of the gas turbine power generation system, and simulations are performed for typical operating conditions of the gas turbine, including: Before starting the core logic of simulation calculation before and after parallel operation update, obtain the internal state snapshot of the digital twin model of the gas turbine power generation system. The internal state snapshot includes at least all variable internal state information, including material properties and geometric parameters. The internal state snapshot is used as the initial model state input when running the core logic of simulation calculation before and after the update in the digital twin model of the gas turbine power generation system. During the parallel simulation process, it is ensured that the core logic of simulation calculation before and after the update is calculated based on the internal state snapshot in their respective simulation processes. After the simulation is completed, the output results of the simulation calculation core logic before and after the update are compared in the digital twin model of the gas turbine power generation system.

[0064] An internal state snapshot is a complete record of all key variable parameters of the digital twin model of a gas turbine power generation system at a specific point in time. These parameters may include, but are not limited to, material properties (e.g., elastic modulus, coefficient of thermal expansion, fatigue life, etc. of components), geometric parameters (e.g., blade wear, casing deformation, etc.), and other internal configurations or state variables that may affect the simulation results. The purpose of obtaining an internal state snapshot is to provide a unified and consistent starting condition for subsequent parallel simulations, ensuring that the core logic of the simulation calculations before and after the update is compared under the same physical state.

[0065] Using internal state snapshots as initial model state inputs means that when running the simulation core logic before and after the update in parallel, both will start from the exact same model state. This is crucial for ensuring the fairness and accuracy of the comparison. During parallel simulation, by ensuring that the simulation core logic before and after the update is calculated based on internal state snapshots in its respective simulation process, deviations in results caused by differences in initial conditions or state drift during simulation can be avoided.

[0066] After the simulation, comparing the output results of the core simulation logic before and after the update in the digital twin model of the gas turbine power generation system is a crucial step in evaluating the performance and behavioral consistency of the old and new logic. This comparison will provide basic data for subsequent difference assessment.

[0067] The proposed solution obtains a snapshot of the internal state of the digital twin model of the gas turbine power generation system before initiating parallel simulation and uses this snapshot as a unified initial input for the core simulation logic before and after the update. This ensures that both logic versions are calculated based on the same, defined model state during parallel simulation. This avoids errors introduced by inconsistent initial conditions, allowing subsequent comparisons of simulation outputs to accurately reflect the differences in the core simulation logic itself before and after the update, rather than the influence of other external factors. Through this controlled parallel simulation environment, the consistency of behavior and performance of the old and new logics in handling typical gas turbine operating conditions can be accurately evaluated.

[0068] Through the above technical solution, this application can ensure a fair and accurate comparison of the core simulation calculation logic before and after the update. By using a unified internal state snapshot as the initial input, the deviation in simulation results that may be caused by differences in initial conditions is eliminated, thereby improving the reliability of the difference assessment. This allows for a more accurate judgment based on more precise data when deciding whether to adopt the updated core simulation calculation logic, effectively reducing the risk introduced by misjudgment and ensuring the accuracy and stability of the digital twin model in long-term operation.

[0069] In the embodiments of this application, when running the simulation calculation core logic before and after the update in parallel within the digital twin model of the gas turbine power generation system, simulations are required for typical operating conditions of the gas turbine. However, if simulations rely solely on a general understanding of typical operating conditions, they may not fully cover all critical or extreme scenarios that the gas turbine might encounter in actual operation. This could lead to insufficient verification of the updated simulation calculation core logic, posing a risk that potential compatibility or accuracy issues might not be detected in a timely manner. If these problems are not addressed, the updated core logic may exhibit instability or inaccuracy in practical applications, affecting the reliability of the digital twin model. Therefore, this application further proposes a specific configuration of the simulation test scenarios in the steps of running the simulation calculation core logic before and after the update in parallel within the aforementioned digital twin model of the gas turbine power generation system and simulating typical operating conditions of the gas turbine, to ensure comprehensive and rigorous verification of the updated core logic.

[0070] Specifically, in the steps of running the core simulation calculation logic before and after the update in parallel within the digital twin model of the gas turbine power generation system, and simulating typical operating conditions of the gas turbine, the simulation test scenarios include: Steady-state full-load operation: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system, respectively simulating the steady-state operation state of the gas turbine at rated power output; Rapid start-stop cycle: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system, respectively cyclically simulates the complete process from cold start to full load operation and then to shutdown; Transient load change: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system simulates the operation of the gas turbine power generation system when the load suddenly increases from the first load to the second load, or when the load suddenly drops from the second load to the first load within a preset first time interval. The first load is less than the second load.

[0071] Steady-state full-load operation is the condition under which the gas turbine outputs power stably at or near its maximum power under design parameters. This scenario aims to verify the accuracy and stability of the updated simulation core logic under long-term, high-load stable operation conditions. Rapid start-stop cycle simulates the entire process of a gas turbine starting from a completely stopped (cold) state, gradually accelerating to full-load operation, and then decelerating to a complete shutdown. This scenario is mainly used to evaluate the performance of the updated simulation core logic under drastic changes in temperature, pressure, and mechanical stress, with particular attention to transient response and component fatigue accumulation during start-up and shutdown. Transient load change is when the gas turbine experiences a rapid increase or decrease in load within a short period. For example, within a preset first time interval, the load suddenly increases from a lower first load to a higher second load, or decreases abruptly from a higher second load to a lower first load. This scenario aims to test the updated simulation core logic's responsiveness to rapid changes in external load, the stability of the control system, and the stress distribution of components under extreme transient conditions.

[0072] This application's solution addresses the issue of incomplete verification that may result from relying solely on typical operating conditions by introducing specific and representative test scenarios. The steady-state full-load operation scenario ensures the computational accuracy and convergence of the updated simulation core logic under long-term stable conditions, verifying its reliability under normal operating conditions. The rapid start-stop cycle scenario focuses on evaluating the core logic's performance under extreme transient changes. For example, during startup and shutdown, the temperature, pressure, and stress inside the gas turbine undergo drastic changes. This scenario effectively exposes potential defects in the core logic's handling of these complex transient processes, such as mesh deformation, solver instability, or inaccurate physical models. The transient load change scenario further focuses on the core logic's rapid response and stability to external disturbances, which is crucial for the flexible operation of the gas turbine in the power grid. It reveals the accuracy of the core logic's simulation of component stress, temperature distribution, and control strategies when responding to sudden load changes. Through these multi-dimensional test scenarios covering both steady-state and transient conditions, the compatibility and differences between the updated and unupdated simulation core logic can be evaluated more comprehensively and deeply, thus ensuring the reliability of the update.

[0073] In some preferred embodiments, this application is implemented as follows: Suppose that the digital twin model of a gas turbine power generation system needs to update the core logic for hydrodynamic simulation of its turbine blade region. After the update, in order to verify the compatibility and accuracy of the old and new core logic, parallel simulations will be performed according to the test scenario described above.

[0074] First, a steady-state full-load operation simulation is performed. In the digital twin model, the old and new core logics respectively simulate the steady-state operation of the gas turbine at rated power (e.g., 200MW) for 8 hours. During this process, key parameters such as the temperature distribution on the turbine blade surface, stress concentration areas, and flow velocity distribution are closely monitored to ensure that the new logic remains consistent with the old logic or actual data under steady-state conditions.

[0075] Secondly, a rapid start-stop cycle simulation was conducted. The old and new core logics respectively simulated the complete cycle of a gas turbine starting from ambient temperature (cold state), reaching full load operation within 30 minutes, operating for 2 hours, and then completely shutting down within 20 minutes. This cycle was repeated 5 times. During this process, special attention was paid to the transient temperature gradient, thermal stress changes, and vibration response of the turbine blades and combustion chamber during the start-up and shutdown phases to evaluate the performance of the new logic under drastic dynamic changes.

[0076] Finally, a transient load change simulation was conducted. The old and new core logics simulated a gas turbine operating stably at 50% load (first load), then experiencing a sudden load increase to 90% (second load) within 10 seconds and maintaining this level for 5 minutes; subsequently, the load abruptly dropped back to 50% within 10 seconds and remained there for 5 minutes. This process was repeated three times. In this scenario, the focus was on analyzing the gas turbine control system's response speed to load changes, the fluctuation of turbine outlet temperature, and the peak transient stress of key components to verify the accuracy and stability of the new logic in handling rapid load changes.

[0077] By evaluating the differences in simulation results across these three test scenarios, we can comprehensively determine whether the updated simulation calculation core logic meets the requirements, thereby deciding whether to execute the new logic or roll back to the old logic.

[0078] Specifically, in the above-mentioned step of evaluating the differences between the core simulation calculation logic before and after the update based on the simulation output results, this application proposes a more detailed and targeted evaluation method, including: Obtain the first key output parameter of the core logic output of the simulation calculation before the digital twin model of the gas turbine power generation system is updated, and the second key output parameter of the core logic output of the simulation calculation after the update; among them, Key output parameters include turbine blade stress data, temperature gradient distribution data, and vibration mode data; Based on the first and second key output parameters, the differences in the core simulation calculation logic before and after the update are evaluated.

[0079] Specifically, key output parameters refer to physical quantities that have a decisive impact on the performance, reliability, and lifespan of a gas turbine during operation. The accuracy of these parameters directly affects the predictive ability of the digital twin model to depict the behavior of the physical entity. Turbine blade stress data refers to the distribution of mechanical stress experienced by turbine blades under different operating conditions, reflecting fatigue damage and potential failure risks. Temperature gradient distribution data refers to the temperature field distribution and rate of change within the gas turbine, particularly in high-temperature components (such as the combustion chamber and turbine), which is crucial for assessing thermal stress, material creep, and thermal fatigue. Vibration modal data refers to the natural frequencies and mode shapes exhibited by the gas turbine structure under specific excitations, which are closely related to the dynamic stability, resonance risk, and component wear of the equipment.

[0080] The first key output parameter refers to the key physical quantity data extracted from the digital twin model of the gas turbine power generation system after the simulation core logic was run before the update. The second key output parameter refers to the corresponding key physical quantity data extracted from the same digital twin model after the simulation core logic was run after the update. By acquiring and comparing these two sets of parameters, the differences in key performance indicators between the simulation core logic before and after the update can be quantitatively evaluated.

[0081] This application's solution identifies and obtains the first and second key output parameters from the core simulation logic of a gas turbine power generation system before and after an update, and then evaluates the differences based on these parameters. The working principle of this method is that the operating state and health of a gas turbine power generation system can be characterized by a series of key physical quantities. When the core simulation logic is updated, its internal physical model, numerical algorithm, or parameter settings may change, and these changes are directly reflected in the key parameters of the simulation output. For example, an update to the mesh generation algorithm may affect the accuracy of stress calculations, and adjustments to the physical modulator parameters may change the predicted temperature field. By focusing on parameters crucial to the gas turbine, such as turbine blade stress data, temperature gradient distribution data, and vibration mode data, the impact of the update on the model's predictive capability can be directly and effectively reflected. Evaluating the differences in these key parameters reveals whether the updated core simulation logic accurately simulates the physical behavior of the gas turbine and whether there are significant deviations between it and the logic before the update.

[0082] Specifically, the above-mentioned evaluation of the differences in the core simulation calculation logic before and after the update based on the simulation output results can be further refined into the following methods.

[0083] When turbine blade stress data is used as a key output parameter, the ratio of the difference between the first turbine blade stress data output by the core logic of the simulation calculation before the update and the second turbine blade stress data output by the core logic of the simulation calculation after the update, and the ratio of this difference to the first turbine blade stress data, is compared with a preset turbine blade stress data deviation threshold. The comparison result serves as an evaluation index for turbine blade stress data deviation. The first and second turbine blade stress data represent the mechanical stress distribution experienced by the turbine blade under the simulation calculation core logic before and after the update, respectively. By calculating the ratio of the difference between the two data to the data before the update, the relative impact of the update on the turbine blade stress calculation can be quantified.

[0084] When temperature gradient distribution data is used as a key output parameter, the difference between the root mean square (RMS) of the first temperature gradient distribution data when running the core simulation logic of the gas turbine power generation system before the update and the RMS of the second temperature gradient distribution data when running the updated simulation logic is compared with a preset temperature gradient distribution data deviation threshold. This comparison result serves as an evaluation index for temperature gradient distribution deviation. The RMS of the first and second temperature gradient distribution data respectively reflect the overall characteristics of the simulation logic before and after the update in terms of temperature distribution uniformity or thermal stress concentration. By comparing the differences in RMS, the impact of the update on system thermal management and thermal fatigue prediction can be assessed.

[0085] When vibration modal data is used as a key output parameter, the difference between the natural frequencies in the first vibration modal data (before running the updated simulation core logic) and the natural frequencies in the second vibration modal data (after running the updated simulation core logic) of the gas turbine power generation system digital twin model is compared with a preset vibration modal frequency deviation threshold. The comparison result serves as a vibration modal deviation evaluation index. The natural frequencies in the first and second vibration modal data represent the vibration characteristics of key gas turbine components under different simulation logics. The difference in natural frequencies directly relates to the resonance risk and structural integrity of the components; therefore, accurate evaluation of these frequencies is crucial.

[0086] This application's solution employs a specific and quantitative difference assessment of three key output parameters: turbine blade stress data, temperature gradient distribution data, and vibration modal data. This allows for a comprehensive and in-depth analysis of the impact of updates to the core logic of simulation calculations. Specifically, by calculating the relative deviation of turbine blade stress data, the potential impact of the update on component structural integrity and fatigue life prediction can be accurately captured; by comparing the root mean square differences in temperature gradient distribution data, the changes in thermal management strategies and thermal stress distribution caused by the update can be effectively assessed; and by comparing the natural frequency differences in vibration modal data, the potential impact of the update on system dynamic response and resonance risk can be promptly identified. This multi-dimensional and refined evaluation mechanism ensures a comprehensive understanding of the effects of updates to the core logic of simulation calculations.

[0087] In some of the embodiments described above in this application, it is proposed that when the difference assessment result is greater than a preset difference threshold, the updated simulation calculation core logic needs to be adjusted. However, in its implementation, if there is a lack of specific adjustment strategies, the adjustment process may be inefficient, making it difficult to quickly converge to the required difference range, and it may even fail to effectively eliminate the deviation between the simulation calculation core logic before and after the update, thereby affecting the accuracy and reliability of the digital twin model.

[0088] In response, this application further proposes an adjusted and updated core logic for simulation calculations, including: Adjust mesh density parameters: Adjust the parameters of local mesh density in the updated simulation calculation core logic to increase or decrease the number of meshes in a specified area until the turbine blade stress calculation results are consistent with the old logic or physical verification results; Adjust the solver convergence parameters: If the bias is related to numerical stability, adjust the number of iterations, relaxation factor, or convergence criterion in the new physics solver to ensure the stability of the computation process and the accuracy of the results. Update material model coefficients: If the updated simulation calculation core logic interprets material behavior differently, adjust the material model coefficients used in the digital twin model to compensate for the systematic deviations caused by the updated simulation calculation core logic.

[0089] Specifically, adjusting the mesh density parameter refers to optimizing simulation results by modifying the fineness of the mesh used to discretize the physical domain in the core simulation logic. For example, in critical component areas such as turbine blades, due to their complex geometry and exposure to high stress and temperature, extremely high simulation accuracy is required. Therefore, the number of meshes in this region can be increased, i.e., the local mesh density can be improved to capture physical phenomena more precisely. Conversely, in regions where accuracy requirements are not high, the number of meshes can be appropriately reduced to improve computational efficiency. The goal is to optimize the mesh distribution so that the simulation results, especially the turbine blade stress calculation results, are consistent with the previous logic or actual physical verification results, thereby eliminating deviations caused by improper mesh settings.

[0090] Adjusting the solver convergence parameters can be understood as optimizing the control strategy of the numerical solution algorithm used in the simulation. When simulation results deviate and are related to numerical stability—for example, large fluctuations in the calculation results, non-convergence, or slow convergence—these can be improved by adjusting parameters such as the number of iterations, relaxation factors, or convergence criteria. The number of iterations refers to the maximum number of computational loops the solver performs before reaching the convergence condition; the relaxation factor controls the magnitude of variable updates in each iteration, and an appropriate relaxation factor can accelerate convergence or prevent oscillations; the convergence criterion defines the standard for judging whether the calculation results have reached a stable state, such as the absolute or relative value of the residuals. The purpose is to ensure the stability of the simulation process and the accuracy of the results, avoiding deviations caused by numerical errors or instability.

[0091] In practical applications, updating material model coefficients specifically refers to correcting the material property parameters used in the digital twin model based on the updated interpretation of material behavior by the core simulation logic. For example, if the new simulation logic employs a more advanced material constitutive model or provides a more precise description of material behavior under extreme conditions such as high temperature and high pressure, the original material model coefficients may no longer be applicable. In this case, it is necessary to adjust material property coefficients such as elastic modulus, Poisson's ratio, yield strength, and creep parameters to ensure that the digital twin model accurately reflects the material response of key gas turbine components during actual operation. The purpose is to compensate for any systematic biases that may be introduced by the updated core simulation logic, ensuring that the digital twin model can realistically simulate the performance and degradation of physical components.

[0092] The proposed solution effectively addresses the problem of unclear adjustment strategies in the aforementioned basic solution by making multi-dimensional and refined adjustments to the updated simulation calculation core logic. Specifically, when the difference assessment result exceeds a preset difference threshold, firstly, by adjusting the mesh density parameters, the simulation accuracy of key areas (such as turbine blades) can be specifically improved, ensuring that the simulation results more accurately reflect physical reality in these areas that have the greatest impact on performance and lifespan, and remain consistent with the old logic or physical verification results. Secondly, by adjusting the solver convergence parameters, the stability and efficiency of the numerical calculation process can be optimized, avoiding deviations caused by numerical non-convergence or calculation errors, thereby ensuring the reliability of the simulation results. Finally, by updating the material model coefficients, the digital twin model can better adapt to the updated simulation calculation core logic's interpretation of material behavior, compensating for systematic deviations caused by model updates, and ensuring that the model can accurately predict the physical degradation of components. It is precisely because of these targeted adjustment measures that the updated simulation calculation core logic can quickly and effectively converge to a state consistent with the old logic or physical reality, thus ensuring the accuracy and practicality of the digital twin model.

[0093] In some preferred embodiments, a specific example is given below. Suppose that after updating the core logic of the simulation calculation of the digital twin model of the gas turbine power generation system, the stress simulation results for the turbine blade region are significantly different from the logic or physical verification results before the update, and this difference exceeds a preset difference threshold.

[0094] First, the system initiates a mechanism to adjust the mesh density parameters. Specifically, since turbine blades are high-stress areas, to improve simulation accuracy, the system automatically identifies the turbine blades and their surrounding regions and increases the local mesh density in these areas. For example, the mesh size on the blade surface is reduced from 2 mm to 0.5 mm to more precisely capture stress concentration and gradient changes. Through multiple iterations of adjustment and simulation, the deviation between the calculated turbine blade stress results and the old logical or physical verification results is reduced to an acceptable range.

[0095] Secondly, if numerical instability or slow convergence is detected in the simulation results, the system will further adjust the solver convergence parameters. For example, the upper limit of the number of iterations will be increased from 1000 to 2000, and the residual convergence criterion will be adjusted from e -4 Adjusted to e -5 This ensures that the calculation process has sufficient opportunity to achieve higher accuracy and stability. Additionally, the relaxation factor can be adjusted, for example, from 0.7 to 0.5, to prevent oscillations during the calculation process and accelerate convergence.

[0096] In some of the embodiments described above in this application, when adjusting the updated simulation calculation core logic fails to reduce the difference evaluation result to a preset difference threshold, the solution simply instructs a rollback to the simulation calculation core logic before the update. However, this simple rollback mechanism may not provide sufficient information to diagnose the root cause of the problem, nor does it offer a more flexible and targeted solution, especially when facing complex or persistent deviations. This can lead to repeated adjustments and rollbacks, reducing the efficiency and reliability of system updates.

[0097] In response, this application further proposes the following steps: if the updated simulation calculation core logic is adjusted to make the difference evaluation result less than a preset difference threshold, then the updated simulation calculation core logic is executed; otherwise, the simulation calculation core logic is rolled back to the previous one. If the adjusted and updated simulation calculation core logic fails to make the difference assessment result less than the preset difference threshold, a deviation report will be generated. The deviation report includes the nature of the difference, the magnitude of the difference, and the affected components or performance parameters; Based on the deviation report, determine the handling method; the handling methods include: manually adjusting the new logic parameters so that the difference assessment result is less than the preset difference threshold; rolling back to the old logic version, executing the simulation calculation core logic before the update, and conducting physical experiments for verification.

[0098] Specifically, if, after one or more adjustments, the updated simulation calculation core logic still fails to meet the preset difference threshold, the system will be triggered to generate a detailed deviation report. This deviation report aims to comprehensively record and analyze the current deviations. Its content is designed to cover the nature of the difference, such as numerical calculation errors, physical model mismatches, or improper boundary condition handling; the magnitude of the difference, i.e., the quantified value of the deviation from the preset threshold; and the affected components or performance parameters, such as turbine blade stress data, temperature gradient distribution data, or vibration modal data. This report allows for a preliminary diagnosis of the specific manifestations and potential causes of the problem.

[0099] Furthermore, after generating a deviation report, the system determines the appropriate handling methods based on the report's content. These handling methods offer a variety of strategies to address deviations of different natures and severities. For example, if the deviation report indicates the problem is localized and controllable, manual adjustment of the new logic parameters can be considered to ensure that the deviation assessment result ultimately falls below a preset deviation threshold through refined intervention. If the deviation is systemic or difficult to resolve through fine-tuning, a rollback to the old logic version can be implemented to ensure stable system operation. Additionally, for deviations that are highly critical, have far-reaching impacts, and are difficult to fully determine through simulation, physical experiments can be conducted to obtain the most realistic feedback data, thereby guiding subsequent logic optimization or decision-making.

[0100] Finally, if discrepancies persist after the above adjustments, and analysis indicates this may be related to a different interpretation of material behavior by the updated simulation logic, the system will update the material model coefficients. For example, if the new simulation logic employs a more accurate high-temperature creep model, the original creep coefficients in the digital twin model may need to be corrected. The system will adjust the creep rate constant and stress exponent of the turbine blade material according to the new physical model requirements to compensate for systematic biases introduced by the updated simulation logic, ensuring the model can accurately predict blade deformation and lifespan under high-temperature operation.

[0101] Through the above series of targeted adjustments, the simulation results of the updated simulation calculation core logic were finally made to match the preset difference threshold, thus successfully completing the update and verification of the digital twin model.

[0102] Through the above technical solution, this application provides a more refined and intelligent digital twin model update management process. When the adjustment of the updated simulation calculation core logic fails to achieve the expected results, instead of simply rolling back, a detailed deviation report is generated, making the problem diagnosis more accurate and avoiding blind rollbacks or repeated ineffective adjustments. This mechanism significantly improves the efficiency and accuracy of problem solving and reduces the potential risks caused by improper model updates. Furthermore, diverse handling methods, including manually adjusting new logic parameters, rolling back to the old logic version, and conducting physical experimental verification, provide technicians with flexible and powerful tools to choose the most appropriate solution based on the specific deviation, thereby ensuring the continuous accuracy and reliability of the digital twin model and ultimately improving the safety and economy of the gas turbine power generation system.

[0103] Although embodiments of the present invention have been shown and described above, it is understood that the embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the embodiments within the scope of the present invention.

Claims

1. A method for constructing a three-dimensional model of a gas turbine power generation system based on digital twinning, characterized by, Comprise: S1: Obtain a digital twin model of a gas turbine power generation system, topologically calibrate and update parameters of the digital twin model, and determine whether the digital twin model needs to update the simulation calculation core logic; S2: If it is determined that the digital twin model needs to update the simulation calculation core logic, update the simulation calculation core logic and perform compatibility verification on the updated and un-updated simulation calculation core logic; wherein updating the simulation calculation core logic at least includes updating the grid generation algorithm and adjusting the physical regulator parameters; S3: After passing the compatibility verification, run the updated and un-updated simulation calculation core logic in parallel in the digital twin model of the gas turbine power generation system, and perform simulation for typical operating conditions of the gas turbine; S4: Based on the simulation output results, difference evaluation is performed on the updated and un-updated simulation calculation core logic; S5: If the difference evaluation result is greater than the preset difference threshold, adjust the updated simulation calculation core logic, execute steps S2-S4, and perform difference evaluation again; S6: If the adjusted updated simulation calculation core logic makes the difference evaluation result less than the preset difference threshold, execute the updated simulation calculation core logic, otherwise, roll back to the un-updated simulation calculation core logic.

2. The method according to claim 1, wherein, The topological calibration and parameter updating of the digital twin model of the gas turbine power generation system comprises: Performing dynamic reference topology self-calibration on the digital twin model of the gas turbine power generation system to correct the geometric distortion of the physical reference points on the gas turbine casing; Performing running state reverse inference and material attribute parameter updating on the digital twin model of the gas turbine power generation system to reflect the progressive physical degradation of the key components of the gas turbine.

3. The method of claim 2, wherein the method further comprises: The dynamic reference topology self-calibration on the digital twin model of the gas turbine power generation system to correct the geometric distortion of the physical reference points on the gas turbine casing comprises: Continuously collecting micro-geometric data of all physical reference points for spatial registration on the gas turbine casing and their surrounding areas; Establishing a dynamic reference point topology relationship network and continuously calculating the relative position changes between each physical reference point in the reference point topology relationship network; When detecting that any physical reference point has a cumulative change exceeding a preset micro-deformation threshold relative to its initial position or adjacent physical reference points, adjusting the coordinate transformation matrix and registration algorithm weight parameters used in the digital twin model for aligning point cloud data with the model; By adjusting, the actual micro-displacement of the physical reference point that has accumulated changes is automatically compensated in the subsequent point cloud registration process, and the geometric distortion caused by the reference point drift is eliminated on a global level.

4. The method of claim 2, wherein the method further comprises: The running state reverse inference and material attribute parameter updating on the digital twin model of the gas turbine power generation system to reflect the progressive physical degradation of the key components of the gas turbine comprises: Receiving real-time running data from the key components of the gas turbine, wherein the key components of the gas turbine at least include turbine blade regions; The real-time operating data from the key components of the gas turbine are compared with the simulation results of the digital twin model of the gas turbine power generation system under the current operating conditions. When a deviation is detected between the two, the changes in the physical parameters that caused the deviation are deduced, and the material properties of the key components of the gas turbine are updated to reflect the actual deterioration state of the key components of the gas turbine.

5. The method of claim 1, wherein the method further comprises: The core logic for simulation calculations before and after the update is run in parallel within the digital twin model of the gas turbine power generation system, and simulations are performed for typical operating conditions of the gas turbine, including: Before initiating the core logic of simulation calculation before and after parallel operation of the update, an internal state snapshot of the digital twin model of the gas turbine power generation system is obtained. The internal state snapshot includes at least all variable internal state information, including material properties and geometric parameters. The internal state snapshot is used as the initial model state input when running the core logic of simulation calculation before and after the update in the digital twin model of the gas turbine power generation system. During the parallel simulation process, it is ensured that the core logic of simulation calculation before and after the update is calculated based on the internal state snapshot in its respective simulation process. After the simulation is completed, the output results of the simulation calculation core logic before and after running the update in the digital twin model of the gas turbine power generation system are compared.

6. The method of claim 1, wherein the method further comprises: In the process of running the core simulation calculation logic before and after the update in parallel within the digital twin model of the gas turbine power generation system, and simulating typical operating conditions of the gas turbine, the simulation test scenarios include: Steady-state full-load operation: The core logic of the simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system simulates the steady-state operation state of the gas turbine at rated power output. Rapid start-stop cycle: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system, respectively simulates the complete process from cold start to full load operation and then to shutdown. Transient load change: The core logic of simulation calculation before and after the parallel operation and update of the digital twin model of the gas turbine power generation system simulates the operation of the gas turbine power generation system when the load suddenly increases from the first load to the second load, or when the load suddenly drops from the second load to the first load within a preset first time interval; Wherein, the first load is less than the second load.

7. The method of claim 1, wherein the method further comprises: Based on the simulation output results, the differences in the core simulation calculation logic before and after the update are evaluated, including: Obtain the first key output parameter of the core logic output of the simulation calculation before the digital twin model of the gas turbine power generation system is updated, and the second key output parameter of the core logic output of the simulation calculation after the update; wherein, Key output parameters include turbine blade stress data, temperature gradient distribution data, and vibration mode data; Based on the first key output parameter and the second key output parameter, the differences in the core simulation calculation logic before and after the update are evaluated.

8. The method according to claim 7, wherein, Based on the first key output parameter and the second key output parameter, the differences in the core simulation calculation logic before and after the update are evaluated, including: When the turbine blade stress data is used as the key output parameter, the difference between the first turbine blade stress data output by the simulation calculation core logic before the digital twin model of the gas turbine power generation system is updated and the second turbine blade stress data output by the simulation calculation core logic after the update is compared with the ratio of the first turbine blade stress data to the preset turbine blade stress data deviation threshold. The comparison result is used as the blade turbine stress data deviation evaluation index. When the temperature gradient distribution data is used as the key output parameter, the root mean square of the first temperature gradient distribution data when the digital twin model of the gas turbine power generation system runs the simulation calculation core logic before the update, and the root mean square of the second temperature gradient distribution data when running the updated simulation calculation core logic, are compared with the preset temperature gradient distribution data deviation threshold. The comparison result is used as the temperature gradient distribution deviation evaluation index. When the vibration modal data is used as the key output parameter, the difference between the natural frequency in the first vibration modal data when the digital twin model of the gas turbine power generation system runs the simulation calculation core logic before the update and the natural frequency in the second vibration modal data when running the updated simulation calculation core logic is compared with a preset vibration modal frequency deviation threshold, and the comparison result is used as a vibration modal deviation evaluation index.

9. The method of claim 1, wherein the method further comprises: receiving a plurality of data sets from a plurality of sensors; and determining a plurality of operating parameters of the gas turbine power generation system based on the plurality of data sets. The updated simulation calculation core logic has been adjusted, including: Adjust mesh density parameters: Adjust the parameters of local mesh density in the updated simulation calculation core logic to increase or decrease the number of meshes in a specified area until the turbine blade stress calculation results are consistent with the old logic or physical verification results; Adjust the solver convergence parameters: If the bias is related to numerical stability, adjust the number of iterations, relaxation factor, or convergence criterion in the new physics solver to ensure the stability of the computation process and the accuracy of the results. Update material model coefficients: If the updated simulation calculation core logic interprets material behavior differently, adjust the material model coefficients used in the digital twin model to compensate for the systematic deviations caused by the updated simulation calculation core logic.

10. The method of claim 1, wherein the method is a method of constructing a three-dimensional model of a gas turbine power generation system based on digital twinning. If the updated simulation calculation core logic is adjusted to make the difference evaluation result less than the preset difference threshold, then the updated simulation calculation core logic is executed; otherwise, the simulation calculation core logic is rolled back to the original logic. This includes: If adjusting the updated simulation calculation core logic fails to make the difference assessment result less than the preset difference threshold, a deviation report will be generated. The deviation report includes the nature of the difference, the magnitude of the difference, and the affected components or performance parameters; Based on the deviation report, a handling method is determined; wherein, the handling method includes: manually adjusting the new logic parameters so that the difference assessment result is less than the preset difference threshold; rolling back to the old logic version, executing the simulation calculation core logic before the update, and conducting physical experimental verification.

Images