Dynamic evaluation method for safety margin of steam turbine blade

Abstract

The invention belongs to the field of steam turbine safety assessment, and discloses a steam turbine blade safety margin dynamic assessment method, which can obtain the unsteady aerodynamic load distribution of the blade surface under the real geometric and dynamic load conditions by establishing a parameterized three-dimensional finite element model and carrying out transient CFD simulation, and has the advantages that compared with the traditional static working condition hypothesis, the dynamic assessment method has the advantages that the efficiency is high; the influence of dynamic load changes such as power grid AGC adjustment, deep peak regulation and quick start and stop on blade stress and temperature fields can be accurately reflected, so that the authenticity and timeliness of load input are improved. After the aerodynamic load is mapped to a structural mechanical model, the distribution rule of the multi-axis complex stress state of the blade can be obtained by combining centrifugal stress and thermal mechanical coupling stress calculation, and an accurate basis is provided for subsequent safety margin calculation.

Description

Technical Field

[0001] This invention belongs to the field of steam turbine safety assessment, and specifically relates to a method for dynamic assessment of the safety margin of steam turbine blades. Background Technology

[0002] As the core power equipment of modern energy systems, the operational safety of steam turbines directly affects the stability and economic operation of power systems. Blades are key components in steam turbines that convert thermal energy into mechanical energy, operating under harsh environments of high temperature, high pressure, and high-speed rotation for extended periods: their operating temperature can reach as high as 565℃, under which conditions materials are prone to creep and structural degradation; their operating pressure exceeds 16MPa, posing a potential risk of stress corrosion cracking; and when the rotational speed reaches 3000rpm, the centrifugal acceleration experienced by the blades can exceed 10000g, easily inducing the initiation and propagation of high-cycle fatigue cracks.

[0003] Existing blade safety assessments mostly employ static design methods based on ASME or ISO standards, using a deterministic safety factor (typically 1.5–2.0) to verify blade strength. While these methods are simple and feasible in the early design stages, they have significant limitations under complex dynamic operating conditions: First, the load assumptions are too idealistic, failing to consider ±10% rated load fluctuations caused by grid AGC regulation and frequent thermomechanical cycles caused by peak-shaving start-ups and shutdowns; second, they still use a linear cumulative damage model (Miner's rule) to handle the interaction between creep and fatigue, neglecting the coupling behavior of micro-dislocation movement and grain boundary slip, resulting in large damage assessment errors; third, they rely on periodic manual inspections during shutdowns (typically every 18–24 months), making it impossible to timely detect and predict sudden blade damage (such as foreign object impacts or cracks induced by localized notches) during operation. Summary of the Invention

[0004] The purpose of this invention is to overcome the problem that existing technologies cannot achieve real-time, accurate and predictive assessment of the safety status of turbine blades under complex dynamic operating conditions, and to provide a dynamic assessment method for the safety margin of turbine blades.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for dynamic evaluation of the safety margin of steam turbine blades, comprising the following steps: Based on the turbine blade parameters, a parametric three-dimensional finite element model of the turbine blade is established. Transient CFD simulation is performed on the parametric three-dimensional finite element model of the turbine blade to obtain the unsteady aerodynamic load distribution on the blade surface. The aerodynamic load is mapped into the structural mechanics model to calculate the centrifugal stress and thermomechanical fatigue stress field distribution of the blade under multiaxial stress. Real-time signal data of turbine blades is collected, and the real-time signal data is preprocessed to obtain the required dynamic signal; The required dynamic signal is compared with the theoretical stress-strain model, and the theoretical stress-strain model is updated accordingly. The updated theoretical stress-strain model is used to process the real-time signal data of the turbine blades, calculate the equivalent stress of the blades, and compare the equivalent stress with the preset multi-level early warning standards to obtain the evaluation results.

[0006] A further improvement of this invention lies in establishing a parametric three-dimensional finite element model of the turbine blade based on the turbine blade parameters, performing transient CFD simulation on the parametric three-dimensional finite element model of the turbine blade to obtain the unsteady aerodynamic load distribution on the blade surface, and mapping the aerodynamic loads to the structural mechanics model. The specific method for calculating the centrifugal stress and thermomechanical fatigue stress field distribution of the blade under multiaxial stress is as follows: Based on the geometric design parameters and material performance parameters of the turbine blades, a parametric three-dimensional finite element model of the turbine blades is established. Transient CFD simulations were performed on the established model using SST. The turbulence model solves the airflow distribution characteristics on the blade surface to obtain the unsteady aerodynamic load distribution of the blade under actual operating conditions; the aerodynamic pressure and shear force information obtained from the simulation results are mapped onto the structural mechanics model, and together with the centrifugal load and temperature field boundary conditions of the blade, a thermo-mechanical coupling analysis model is established. By using structural finite element analysis, the centrifugal stress field, thermal stress field, and fatigue stress distribution of the blade under multiaxial stress conditions were obtained.

[0007] A further improvement of this invention lies in the method of acquiring real-time signal data from turbine blades, preprocessing the real-time signal data, and obtaining the required dynamic signal, as follows: During the operation of the steam turbine, fiber optic grating sensor arrays and phase-sensitive optical time-domain reflectometry systems are deployed at key locations on the blades to collect the strain, vibration, and temperature signals of the blades in real time. Kalman filtering is used to estimate the state of the dynamic signal and remove random noise. An improved particle filtering algorithm is used to suppress the periodic interference caused by steam excitation and to compensate for the thermal strain caused by temperature changes, so as to obtain the true mechanical strain and vibration response of each monitoring section of the blade. The actual mechanical strain and vibration response of each monitored section of the blade are converted into equivalent stress values, which are consistent with the stress dimensions in the simulation model, and used as the required dynamic signal.

[0008] A further improvement of this invention lies in comparing the required dynamic signal with the theoretical stress-strain model, and the specific method for updating the theoretical stress-strain model is as follows: The measured dynamic signal data is compared and analyzed with the theoretical stress-strain model. By matching the simulated stress and the measured stress at the same cross section and the same time point, the difference between the two is calculated. The Bayesian update algorithm is used to adaptively correct the model parameters, so that the theoretical model gradually converges to a state consistent with the actual working conditions, thus completing the update of the theoretical stress-strain model.

[0009] A further improvement of this invention lies in using an updated theoretical stress-strain model to process real-time signal data of turbine blades, calculate the equivalent stress of the blades, and compare the equivalent stress with preset multi-level early warning standards to obtain the evaluation results. The specific method is as follows: The updated theoretical stress-strain model is used, combined with real-time signal data at the current moment, to calculate the equivalent stress of the turbine blades; The preset multi-level early warning standard is a dynamic safety margin, which is divided into three levels; When the equivalent stress exceeds the preset multi-level warning standard, a graded warning signal will be automatically generated.

[0010] Secondly, the present invention provides a dynamic assessment system for the safety margin of steam turbine blades, comprising the following steps: The stress field acquisition module is used to establish a parametric three-dimensional finite element model of the turbine blade based on the turbine blade parameters, perform transient CFD simulation on the parametric three-dimensional finite element model of the turbine blade, obtain the unsteady aerodynamic load distribution on the blade surface, map the aerodynamic load into the structural mechanics model, and calculate the centrifugal stress and thermomechanical fatigue stress field distribution of the blade under multiaxial stress state. The preprocessing module is used to collect real-time signal data from the turbine blades, preprocess the real-time signal data, and obtain the required dynamic signal. The model update module is used to compare the required dynamic signals with the theoretical stress-strain model and update the theoretical stress-strain model accordingly. The data processing module is used to process the real-time signal data of the turbine blades using the updated theoretical stress-strain model, calculate the equivalent stress of the blades, and compare the equivalent stress with the preset multi-level early warning standards to obtain the evaluation results.

[0011] A further improvement of this invention is that the function of the stress field acquisition module is implemented through the following method: Based on the geometric design parameters and material performance parameters of the turbine blades, a parametric three-dimensional finite element model of the turbine blades is established. Transient CFD simulations were performed on the established model using SST. The turbulence model solves the airflow distribution characteristics on the blade surface to obtain the unsteady aerodynamic load distribution of the blade under actual operating conditions; the aerodynamic pressure and shear force information obtained from the simulation results are mapped onto the structural mechanics model, and together with the centrifugal load and temperature field boundary conditions of the blade, a thermo-mechanical coupling analysis model is established. By using structural finite element analysis, the centrifugal stress field, thermal stress field, and fatigue stress distribution of the blade under multiaxial stress conditions were obtained.

[0012] A further improvement of this invention is that the function of the preprocessing module is implemented through the following method: During the operation of the steam turbine, fiber optic grating sensor arrays and phase-sensitive optical time-domain reflectometry systems are deployed at key locations on the blades to collect the strain, vibration, and temperature signals of the blades in real time. Kalman filtering is used to estimate the state of the dynamic signal and remove random noise. An improved particle filtering algorithm is used to suppress the periodic interference caused by steam excitation and to compensate for the thermal strain caused by temperature changes, so as to obtain the true mechanical strain and vibration response of each monitoring section of the blade. The actual mechanical strain and vibration response of each monitored section of the blade are converted into equivalent stress values, which are consistent with the stress dimensions in the simulation model, and used as the required dynamic signal.

[0013] A further improvement of this invention is that the function of the model update module is implemented through the following method: The measured dynamic signal data is compared and analyzed with the theoretical stress-strain model. By matching the simulated stress and the measured stress at the same cross section and the same time point, the difference between the two is calculated. The Bayesian update algorithm is used to adaptively correct the model parameters, so that the theoretical model gradually converges to a state consistent with the actual working conditions, thus completing the update of the theoretical stress-strain model.

[0014] A further improvement of this invention is that the function of the data processing module is implemented through the following method: The updated theoretical stress-strain model is used, combined with real-time signal data at the current moment, to calculate the equivalent stress of the turbine blades; The preset multi-level early warning standard is a dynamic safety margin, which is divided into three levels; When the equivalent stress exceeds the preset multi-level warning standard, a graded warning signal will be automatically generated.

[0015] Compared with the prior art, the present invention has the following beneficial effects: This invention, by establishing a parametric three-dimensional finite element model and performing transient CFD simulation, can obtain the unsteady aerodynamic load distribution on the blade surface under realistic geometric and dynamic load conditions. Compared with traditional static operating condition assumptions, it can accurately reflect the impact of dynamic load changes such as power grid AGC regulation, deep peak shaving, and rapid start-stop on the blade's stress and temperature field, thereby improving the realism and timeliness of load input. After mapping the aerodynamic load to the structural mechanics model, combined with the calculation of centrifugal stress and thermomechanical coupling stress, the distribution law of the multiaxial complex stress state of the blade can be obtained, providing an accurate basis for subsequent safety margin calculation. This invention, by collecting real-time strain, vibration, and temperature signals of the blade during operation and using filtering and signal fusion algorithms to preprocess the raw data, can effectively eliminate measurement noise and environmental interference, making the dynamic signals more stable and reliable. This step enables the timely capture of the blade's real operating response under high-speed and high-temperature conditions, providing highly reliable monitoring data for dynamic safety assessment. This invention compares real-time monitoring signals with the theoretical stress-strain model and performs online updates. Through methods such as Bayesian updates or recursive least squares algorithms, it achieves adaptive calibration of model parameters, compensating for model deviations caused by material performance degradation and boundary condition fluctuations. The updated theoretical model continuously conforms to the actual operating conditions of the blade, ensuring long-term accuracy and robustness of the calculation results. This invention, through improved stress assessment and dynamic safety margin calculation methods, compares real-time equivalent stress with allowable stress considering temperature, strain aging, and material degradation factors, and introduces multi-level early warning standards, achieving an automated closed loop from safety status quantification to early warning decision-making. The system can issue yellow, orange, or red warnings in advance based on the trend of safety margin changes, providing maintenance personnel with a clear risk level reference. In summary, this method realizes the transformation from static assessment based on design assumptions to dynamic prediction based on real-time data and model fusion. It can achieve high-precision, low-latency safety status assessment during turbine blade operation, identify potential failure risks such as fatigue, creep, or cracks in advance, and effectively improve the safety, reliability, and life management level of equipment operation. Attached Figure Description

[0016] Figure 1 This is a flowchart of the present invention; Figure 2 This is a system diagram of the present invention. Detailed Implementation

[0017] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0018] Example 1: See Figure 1A method for dynamic assessment of the safety margin of steam turbine blades includes the following steps: S1. Based on the turbine blade parameters, establish a parametric three-dimensional finite element model of the turbine blade. Perform transient CFD simulation on the parametric three-dimensional finite element model of the turbine blade to obtain the unsteady aerodynamic load distribution on the blade surface. Map the aerodynamic load to the structural mechanics model and calculate the centrifugal stress and thermomechanical fatigue stress field distribution of the blade under multiaxial stress state.

[0019] S2 collects real-time signal data from the turbine blades, preprocesses the real-time signal data, and obtains the required dynamic signal.

[0020] S3 compares the required dynamic signal with the theoretical stress-strain model and updates the theoretical stress-strain model.

[0021] S4 uses an updated theoretical stress-strain model to process real-time signal data of turbine blades, calculates equivalent stress of blades, and compares equivalent stress with preset multi-level early warning standards to obtain evaluation results.

[0022] Example 2: See Figure 2 A dynamic assessment system for the safety margin of steam turbine blades, characterized by comprising the following steps: The stress field acquisition module is used to establish a parametric three-dimensional finite element model of the turbine blade based on the turbine blade parameters, perform transient CFD simulation on the parametric three-dimensional finite element model of the turbine blade to obtain the unsteady aerodynamic load distribution on the blade surface, map the aerodynamic load into the structural mechanics model, and calculate the centrifugal stress and thermomechanical fatigue stress field distribution of the blade under multiaxial stress state.

[0023] The preprocessing module is used to collect real-time signal data from the turbine blades, preprocess the real-time signal data, and obtain the required dynamic signal.

[0024] The model update module is used to compare the required dynamic signals with the theoretical stress-strain model and update the theoretical stress-strain model.

[0025] The data processing module is used to process the real-time signal data of the turbine blades using the updated theoretical stress-strain model, calculate the equivalent stress of the blades, and compare the equivalent stress with the preset multi-level early warning standards to obtain the evaluation results.

[0026] Example 3: In this embodiment, a parametric three-dimensional finite element model of the turbine blade is first established based on the geometric design parameters (including blade height, chord length, thickness distribution, and torsion angle) and material property parameters (elastic modulus, coefficient of thermal expansion, and density). This model can be automatically generated on platforms such as ANSYS Workbench, enabling rapid modification and reconstruction of different blade structures.

[0027] Subsequently, transient CFD (Computational Fluid Dynamics) simulations were performed on the established model using SST. A turbulence model is used to solve for the airflow distribution characteristics on the blade surface, obtaining the unsteady aerodynamic load distribution of the blade under actual operating conditions. The aerodynamic pressure and shear force information obtained from the simulation results are mapped onto the structural mechanics model, and together with the centrifugal load and temperature field boundary conditions of the blade, a thermo-mechanical coupled analysis model is established.

[0028] By using structural finite element analysis, the centrifugal stress field, thermal stress field, and fatigue stress distribution of the blade under multiaxial stress can be obtained, providing theoretical benchmark data for subsequent dynamic safety margin analysis. The output of this step includes: equivalent stress, temperature field distribution, and time-series load data for each key section. These data will serve as inputs for subsequent sensor data comparison and model calibration.

[0029] This embodiment starts from geometric and operating parameters to obtain the multiaxial stress field of the blade under time-varying aerodynamic / thermal / centrifugal combined effects, which serves as a theoretical prior for subsequent online calibration and evaluation.

[0030] Blade parameters include spanwise radius range chord length Installation corner Thickness distribution Torsion Law Leading edge radius Material parameters wait.

[0031] Operating conditions include speed Inlet total pressure / total temperature, outlet static pressure / flow rate, external heat exchange boundary .

[0032] First, parametric 3D finite element model (CAD / ANSYS):

[0033] Generate parameter vector Blade solid; surface boundary layer mesh Total number of units > .

[0034] Then perform transient CFD (SST) Time step Take the Courant number Co Output the force density on the blade surface for each step:

[0035] Perform load mapping, and use least squares / projection to maintain torque conservation. Integral as surface element force And mapped to structural nodes of equivalent effect :

[0036] Ensure that the global force and moment conservation error is less than 1%.

[0037] Defined temperature field and thermal boundary:

[0038] Perform centrifugal body force and thermomechanical constitutive modeling (using elastoplastic / Chaboche methods):

[0039] (If only flexibility is assessed, this can be ignored) ) Obtaining equivalent stress (multi-axis):

[0040] Example 4: In this embodiment, during turbine operation, various types of sensors, such as fiber optic grating (FBG) sensors or accelerometers, are deployed at key locations on the blades (e.g., blade root, leading edge, pressure surface) to collect operational signals in real time. The collected raw data typically contains high-frequency noise and operational disturbances, thus requiring signal preprocessing. Preprocessing includes signal denoising (e.g., Kalman filtering, particle filtering algorithms), drift correction, and data synchronization, unifying multi-source signals onto the same time reference. After processing, a dynamic signal set reflecting the actual operating state of the blades is obtained, providing accurate input for subsequent model calibration and stress calculation.

[0041] Input sensor raw data: FBG / Φ-OTDR strain ,temperature Vibration acceleration / displacement The timestamp in step one is used to align with the work status record.

[0042] Perform time alignment and use cross-correlation to determine lag. :

[0043] Perform temperature compensation (FBG):

[0044] Perform noise reduction filtering: state ,

[0045] Recursion:

[0046] Frequency band selection and band-stop / band-pass filter suppression. The narrowband noise is normalized and resampled to obtain a unified frequency grid.

[0047] Example 5: This embodiment compares and analyzes the preprocessed dynamic signal data with the theoretical stress-strain model obtained in step one. By comparing the differences between the measured strain and temperature response and the model predictions, deviations in model parameters (such as material stiffness, coefficient of thermal expansion, or load distribution weights) are identified. An adaptive update mechanism (such as Bayesian update or least squares iteration) is used to correct the parameters of the theoretical model, making the model output more consistent with the measured signals. The updated theoretical model becomes the online calibration model under the current operating conditions, which can reflect the real stress and fatigue state of the blade under actual working conditions in real time.

[0048] Input the reference field / load timing sequence from step one. The dynamic signal in step two .

[0049] Calculate the observation operator (project the field quantity onto the sensor position / orientation):

[0050] in, For sensor coordinates, This is a sensitive area.

[0051] Error definition and objective function:

[0052] in, For parameters to be updated (materials) Heat exchange Damping ratio (load amplitude coefficient, etc.)

[0053] Perform Bayesian update / recursive least squares:

[0054] When approximating Gaussian:

[0055] After obtaining online calibration So that the (strain / displacement) residuals satisfy root mean square (For example, 3%).

[0056] Example 6: This embodiment utilizes an updated theoretical stress-strain model to dynamically calculate the equivalent stress (e.g., Von Mises stress or principal stress components) of the blade during operation from the real-time acquired signal data. The calculated equivalent stress is compared with the allowable stress of the material to calculate the dynamic safety margin (DSM). Based on preset multi-level safety thresholds (e.g., DSM > 0.8 for a yellow warning, DSM > 1.0 for an orange warning, and DSM > 1.2 for a red warning), the safety status of the blade is determined. When the safety margin exceeds the threshold, the system automatically sends a warning signal to the turbine's DCS control system via a communication protocol (e.g., OPC-UA), achieving real-time monitoring of the operating status and risk alerts.

[0057] Dynamic Safety Margin (DSM) and Reliability Mapping:

[0058] If we consider uncertainty, let... Given the mean and standard deviation, the reliability index (first-order second-order approximation) is...

[0059] Multi-level early warning criteria: yellow:

[0060] orange color:

[0061] red: or

[0062] When the threshold is triggered, the alarm topic and load are written to the DCS via OPC-UA.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method of dynamic assessment of safety margins for a turbine blade, characterized in that, The method comprises the following steps: According to the parameters of the steam turbine blade, a parameterized three-dimensional finite element model of the steam turbine blade is established, transient CFD simulation is performed on the parameterized three-dimensional finite element model of the steam turbine blade, the unsteady aerodynamic load distribution on the surface of the blade is obtained, the aerodynamic load is mapped to the structural mechanics model, and the centrifugal stress and thermal mechanical fatigue stress field distribution of the blade under the multi-axial stress state are calculated. Real-time signal data of the steam turbine blade are collected, and the real-time signal data are preprocessed to obtain the required dynamic signal. The required dynamic signal is compared with the theoretical stress-strain model, and the theoretical stress-strain model is updated. The updated theoretical stress-strain model is used to process the real-time signal data of the steam turbine blade, the equivalent stress of the blade is calculated, and the equivalent stress is compared with the preset multi-level early warning standard to obtain the evaluation result.

2. A method of dynamic assessment of safety margins of a turbine blade according to claim 1, characterized in that, The specific method of establishing a parameterized three-dimensional finite element model of the steam turbine blade according to the parameters of the steam turbine blade, performing transient CFD simulation on the parameterized three-dimensional finite element model of the steam turbine blade, obtaining the unsteady aerodynamic load distribution on the surface of the blade, mapping the aerodynamic load to the structural mechanics model, and calculating the centrifugal stress and thermal mechanical fatigue stress field distribution of the blade under the multi-axial stress state is as follows: According to the geometric design parameters and material performance parameters of the steam turbine blade, a parameterized three-dimensional finite element model of the steam turbine blade is established. The established model is simulated by transient CFD, and the SST The turbulent flow model is used to solve the air flow distribution characteristics of the blade surface, and the unsteady aerodynamic load distribution of the blade under the actual operating condition is obtained. The aerodynamic pressure and shear force information obtained in the simulation results are mapped to the structural mechanics model, and the centrifugal load and temperature field boundary conditions of the blade are used to establish a thermal-mechanical coupling analysis model. Through structural finite element calculation, the centrifugal stress field, thermal stress field and fatigue stress distribution of the blade under the multi-axial stress state are obtained.

3. A method of dynamic assessment of safety margins of a turbine blade according to claim 1, characterized in that, The specific method of collecting real-time signal data of the steam turbine blade, preprocessing the real-time signal data, and obtaining the required dynamic signal is as follows: During the operation of the steam turbine, an array of fiber Bragg grating sensors and a phase-sensitive optical time domain reflectometry system are arranged at key positions of the blade to collect strain, vibration and temperature signals of the blade in real time. Kallman filtering is used for state estimation of the dynamic signal to remove random noise, an improved particle filtering algorithm is used to suppress periodic interference caused by steam excitation, and thermal strain caused by temperature change is compensated to obtain the real mechanical strain and vibration response of each monitoring section of the blade. The real mechanical strain and vibration response of each monitoring section of the blade are converted into equivalent stress values, which are consistent with the stress dimension in the simulation model, and are used as the required dynamic signal.

4. A method of dynamic assessment of safety margins of a turbine blade according to claim 1, characterized in that, The specific method of comparing the required dynamic signal with the theoretical stress-strain model and updating the theoretical stress-strain model is as follows: The measured dynamic signal data are compared and analyzed with the theoretical stress-strain model, the simulation stress and the measured stress at the same section and the same time point are matched, the difference between the two is calculated, the model parameters are adaptively corrected by using the Bayesian updating algorithm, the theoretical model is gradually converged to the state consistent with the actual working condition, and the updating of the theoretical stress-strain model is completed.

5. A method of dynamic assessment of safety margins of a turbine blade according to claim 1, characterized in that, The specific method of using the updated theoretical stress-strain model to process the real-time signal data of the steam turbine blade, calculating the equivalent stress of the blade, comparing the equivalent stress with the preset multi-level early warning standard, and obtaining the evaluation result is as follows: The equivalent stress of the steam turbine blade is calculated by using the updated theoretical stress-strain model and the real-time signal data at the current moment; The preset multi-level early warning standard is dynamic safety margin, and the dynamic safety margin is divided into three levels; When the equivalent stress exceeds the preset multi-level early warning standard, a graded early warning signal is automatically generated.

6. A system for dynamic assessment of safety margins of a turbine blade, characterized in that The method comprises the following steps: A stress field acquisition module is configured to establish a parameterized three-dimensional finite element model of the steam turbine blade according to parameters of the steam turbine blade, perform transient CFD simulation on the parameterized three-dimensional finite element model of the steam turbine blade, obtain the distribution of unsteady aerodynamic load on the surface of the blade, map the aerodynamic load to a structural mechanics model, and calculate the centrifugal stress and thermal mechanical fatigue stress field distribution of the blade under a multi-axial stress state; A preprocessing module is configured to collect real-time signal data of the steam turbine blade, pre-process the real-time signal data, and obtain required dynamic signals; A model updating module is configured to compare the required dynamic signals with a theoretical stress-strain model, and update the theoretical stress-strain model; A data processing module is configured to process the real-time signal data of the steam turbine blade by using the updated theoretical stress-strain model, calculate the equivalent stress of the blade, compare the equivalent stress with the preset multi-level early warning standard, and obtain an evaluation result.

7. A system for dynamic assessment of safety margins of a turbine blade according to claim 6, characterized in that The function of the stress field acquisition module is implemented by the following method: A parameterized three-dimensional finite element model of the steam turbine blade is established according to geometric design parameters and material performance parameters of the steam turbine blade; The established model was simulated by transient CFD, and SST The turbulent model was used to solve the air flow distribution characteristics of the blade surface, and the unsteady aerodynamic load distribution of the blade under the actual operating condition was obtained. The aerodynamic pressure and shear force information obtained in the simulation results were mapped to the structural mechanics model, and acted together with the centrifugal load and temperature field boundary conditions of the blade to establish a thermal-mechanical coupling analysis model. The centrifugal stress field, thermal stress field and fatigue stress distribution of the blade under a multi-axial stress state are obtained through structural finite element calculation.

8. A system for dynamic assessment of safety margins of a turbine blade according to claim 6, characterized in that The function of the preprocessing module is implemented by the following method: During the operation of the steam turbine, a fiber grating sensor array and a phase-sensitive optical time domain reflectometry system are arranged at key positions of the blade to collect strain, vibration and temperature signals of the blade in real time; Kalman filtering is used to estimate the state of the dynamic signals, remove random noise, use an improved particle filtering algorithm to suppress periodic interference caused by steam excitation, and compensate for thermal strain caused by temperature changes to obtain real mechanical strain and vibration response of each monitoring section of the blade; The real mechanical strain and vibration response of each monitoring section of the blade are converted into equivalent stress values, which are consistent with the stress dimension in the simulation model, and are used as required dynamic signals.

9. A system for dynamic assessment of safety margins of a turbine blade according to claim 6, characterized in that The function of the model updating module is implemented by the following method: The measured dynamic signal data is compared and analyzed with the theoretical stress-strain model, the difference between the simulation stress and the measured stress at the same section and the same time point is calculated, the model parameters are adaptively corrected by using a Bayesian updating algorithm, the theoretical model is gradually converged to a state consistent with the actual working condition, and the updating of the theoretical stress-strain model is completed.

10. A system for dynamic assessment of safety margins of a turbine blade according to claim 6, characterized in that The function of the data processing module is implemented by the following method: The equivalent stress of the steam turbine blade is calculated by using the updated theoretical stress-strain model and the real-time signal data at the current moment; The preset multi-level early warning standard is dynamic safety margin, and the dynamic safety margin is divided into three levels; When the equivalent stress exceeds the preset multi-level early warning standard, a graded early warning signal is automatically generated.

Images