Turbine disk life and failure risk assessment method

Abstract

The invention discloses a turbine disc service life and failure risk assessment method, and belongs to the technical field of turbine discs, and the turbine disc service life and failure risk assessment method comprises the steps: carrying out the cross-scale multi-field coupling simulation modeling of a turbine disc, and extracting key features, including millimeter-scale overall structure features, micron-scale local features, and nano-scale microstructure features; a multi-mode degradation model based on a turbine disc failure mechanism is constructed, and the weight of each degradation mode is determined; performing multi-source data fusion and calibration on the multi-mode degradation model, and correcting model parameters; and deploying the calibrated multi-mode degradation model in an edge computing terminal, and performing residual life prediction and multi-mode failure risk grading on the turbine disc. In the implementation process of the technical scheme, through cross-scale data fusion and model iterative optimization, the state of the turbine disc is monitored in real time, the evaluation strategy is dynamically adjusted, it is ensured that the prediction result is highly matched with the actual working condition, and design optimization and operation and maintenance decision making of the turbine disc are effectively guided.

Description

Technical Field

[0001] This application relates to the field of turbine disk technology, specifically a method for assessing turbine disk lifespan and failure risk. Background Technology

[0002] A turbine disk is a core rotating component that connects turbine blades to the main shaft. It is commonly used in aero engines and features high temperature resistance, high pressure resistance, and high speed resistance. It can withstand multiple coupled effects such as thermal stress, aerodynamic loads, vibration loads, and high-temperature oxidation.

[0003] The performance of turbine disks directly affects the safety and lifespan of engines. Their failure modes are complex, including low-cycle fatigue, high-cycle fatigue, creep, thermomechanical fatigue, and oxidative thermal corrosion. These failure modes are easily coupled and mutually reinforcing, making turbine disk life prediction and failure risk assessment extremely complex. In existing technologies, assessments can usually only be performed using a single failure mode, which is insufficient to fully reflect the combined effects under actual operating conditions. This results in significant biases in the assessment results and fails to accurately guide practical applications.

[0004] Existing methods for assessing the lifespan and failure risk of turbine disks have some drawbacks. For example, the complex structure leads to insufficient simulation accuracy, data acquisition difficulties make it hard to build assessment models, the failure mechanism under the coupling effect of multiple factors is unclear, and there is a lack of effective comprehensive assessment methods, resulting in a large deviation between the assessment results and actual operating conditions, which cannot provide a reliable basis for the design and maintenance of turbine disks.

[0005] Therefore, it is necessary to provide a method for assessing the lifespan and failure risk of turbine disks to address the above issues.

[0006] It should be noted that the information disclosed in this background section is only for understanding the background technology of this application concept, and therefore may include information that does not constitute prior art. Summary of the Invention

[0007] Based on the aforementioned problems in the existing technology, the problem to be solved by this application is to provide a method for assessing the life and failure risk of turbine disks, which accurately captures the failure mechanism and improves the assessment accuracy by combining a multi-field coupling model with data-driven approaches.

[0008] The technical solution adopted by this application to solve its technical problem is: a method for assessing the life and failure risk of a turbine disk, including: A multi-scale, multi-field coupled simulation model of the turbine disk was performed, and key features were extracted, including millimeter-level overall structural features, micrometer-level local features, and nanometer-level microstructure features. A multi-mode degradation model based on the failure mechanism of turbine disks was constructed, and the weights of each degradation mode were determined by combining cross-scale simulation results. Multi-source data fusion and calibration are performed on the multi-mode degradation model, and the model parameters are corrected to enhance data features; The calibrated multi-mode degradation model is deployed in an edge computing terminal to predict the remaining lifetime of the turbine disk and classify the multi-mode failure risk.

[0009] In the implementation of the technical solution of this application, the turbine disk status is monitored in real time through cross-scale data fusion and model iterative optimization, and the evaluation strategy is dynamically adjusted to ensure that the prediction results are highly consistent with the actual working conditions, effectively guiding the design optimization and operation and maintenance decisions of the turbine disk.

[0010] Furthermore, the millimeter-level overall structural features are extracted based on the three-dimensional geometric model of the turbine disk, and a thermo-mechanical coupling model is established using finite element software. This thermo-mechanical coupling model includes thermal load analysis and mechanical load analysis. The thermal load analysis is based on the heat transfer analysis of the engine's hot-end components to obtain the radial and axial temperature distribution of the turbine disk. The mechanical load combines aerodynamic load and centrifugal force to apply cyclic loads. In the aerodynamic load, the aerodynamic force will generate circumferential bending stress, and the centrifugal force is used to calculate the radial stress from the rotational speed. The cyclic load is for start-up and shutdown conditions, and stress and strain are calculated using finite element analysis software.

[0011] Furthermore, the micron-level local features are obtained by refining the mesh at the rim root and the surface of the central hole with a unit size of less than 0.1 mm, and by combining the crystal plasticity theory to describe the material deformation behavior.

[0012] Furthermore, by solving the local feature scale, the local equivalent strain, microcrack initiation location, crack propagation path, and porosity distribution are output. Among them, the local equivalent strain reflects the local stress state of the material, the microcrack initiation location is 0.1 to 0.5 mm away from the surface, the crack propagation path is along the direction of the maximum principal stress, and the porosity distribution affects the local stress concentration.

[0013] Furthermore, a growth model of the oxide layer was established based on the oxidation theory to simulate the selective oxidation process of elements in the oxide layer under high temperature conditions, where the high temperature environment is a working condition greater than 1000 degrees Celsius.

[0014] Furthermore, based on the nanoscale microstructure characteristics, the curve of oxide layer thickness changing with time and the location of oxide layer cracking are calculated. When the microcrack initiation point coincides with the oxide layer cracking location, it is determined that the crack propagation rate is accelerated, the degree of material damage is aggravated, and the risk of turbine disk failure is increased.

[0015] Furthermore, the multi-mode degradation model includes low-cycle fatigue degradation model, high-temperature creep degradation model, thermomechanical fatigue degradation model, and multi-mode competitive failure model. These models correspond to the failure mechanisms of turbine disks under different stress and temperature environments.

[0016] Furthermore, the process of multi-source data fusion and calibration for the multi-mode degradation model includes: acquiring multi-source data and preprocessing the acquired multi-source data, which includes simulation data, bench test data, and field monitoring data; cleaning and feature enhancement of the multi-source data, including noise suppression, feature extraction, and data enhancement; calibrating the model and performing transfer learning, and minimizing the error between the simulation prediction value and the actual measurement value based on the simulation data.

[0017] Furthermore, noise suppression employs wavelet soft thresholding to remove high-frequency noise from the vibration signal. Based on the cleaned data, time-frequency analysis is performed, and the mean stress cycle value, stress amplitude, temperature gradient, and kurtosis of vibration energy at the wheel flange root are extracted. The kurtosis of vibration energy reflects the transient characteristics of the vibration signal. Combined with the coupling effect of stress and temperature, a feature vector is constructed.

[0018] Furthermore, in the process of predicting the remaining lifetime, random variables in the multi-field coupled model are sampled based on Monte Carlo simulation to calculate the probability distribution of the remaining lifetime, and the prediction results are corrected by using an LSTM model.

[0019] The beneficial effects of this application are as follows: The turbine disk life and failure risk assessment method provided by this application monitors the turbine disk status in real time through cross-scale data fusion and model iterative optimization, dynamically adjusts the assessment strategy, ensures that the prediction results are highly consistent with the actual working conditions, effectively guides the design optimization and operation and maintenance decisions of the turbine disk, and realizes real-time data processing and analysis through edge deployment, reduces latency, improves response speed, and ensures the safe operation of the turbine disk.

[0020] In addition to the purposes, features, and advantages described above, this application has other purposes, features, and advantages. A further detailed description of this application will be provided below with reference to the figures. Attached Figure Description

[0021] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of an overall method for assessing the lifespan and failure risk of a turbine disk according to this application. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] like Figure 1 As shown, this application provides a method for assessing the lifespan and failure risk of turbine disks. This method is applied to the entire process of turbine disk design, manufacturing, and maintenance. Through real-time monitoring of multiple parameters and data analysis, it accurately predicts lifespan and failure risk, ensuring operational safety and efficiency. The method includes the following steps: Step 01: Perform cross-scale multi-field coupled simulation modeling on the turbine disk and extract key features, including millimeter-level overall structural features, micrometer-level local features, and nanometer-level microstructure features. Turbine disks have complex structures and come in many types, making it difficult for traditional evaluation methods to fully cover them. Therefore, in this embodiment, multi-scale simulation technology is introduced to establish a multi-physics coupled simulation model, thereby enabling a comprehensive evaluation of the entire life cycle of the turbine disk. Key features include millimeter-level overall structural features, micrometer-level local features, and nanometer-level microstructure features. The millimeter-level overall structural features are extracted based on the three-dimensional geometric model of the turbine disk, such as hub, rim, mounting edge, and center hole. A thermo-mechanical coupling model is established using finite element software. This thermo-mechanical coupling model includes thermal load analysis and mechanical load analysis. The thermal load analysis is based on the heat transfer analysis of the engine's hot-end components to obtain the radial and axial temperature distribution of the turbine disk, such as turbine blade radiation heat transfer and axial heat conduction. The mechanical load combines aerodynamic load and centrifugal force to apply cyclic loads. The aerodynamic force in the aerodynamic load generates circumferential bending stress, and the centrifugal force is used to calculate the radial stress from the rotational speed. The cyclic loads are applied under start-up and shutdown conditions. Stress and strain calculations are performed using finite element analysis software to determine the stress distribution and deformation of the turbine disk under different operating conditions. Based on multi-scale simulation technology, the overall stress cloud map, temperature gradient, and deformation caused by centrifugal force are solved. The overall stress cloud map is mainly the stress concentration area at the root of the wheel rim, with a stress concentration factor of 3 to 5. The temperature gradient is the temperature difference between the wheel rim and the hub. The deformation is evaluated by combining radial and axial displacement to ensure accuracy. Micrometer-level local features are achieved by refining the mesh at the rim root and the surface of the central hole with a unit size of less than 0.1 mm, and by combining it with crystal plasticity theory to describe the material deformation behavior; In addition to the three-dimensional geometric model of the turbine disk, material properties also need to be considered. These material properties are not only difficult to describe accurately using traditional modeling methods, but also lose a lot of details due to size. Therefore, nanoscale microstructure features are used for modeling, and the mesh size of the rim root and the surface of the central hole is refined to less than 0.1 mm. The rim root is mainly a stress concentration area, and the surface of the central hole is mainly a stress-sensitive area. In the process of refining the mesh, the changes in microstructure can be captured more accurately, and the deformation behavior of the material can be described by combining the crystal plasticity theory to ensure the accuracy of the model. For example, for high-temperature alloys, it is necessary to consider their phase transformation characteristics and creep behavior. For powder metallurgy turbine disks, it is also necessary to simulate the influence of porosity on stress concentration and crack propagation. Among them, the crack source is porosity, the crack propagation path is determined by finite element analysis, porosity is positively correlated with crack propagation rate, and creep behavior is aggravated under high temperature environment. It is necessary to combine aging treatment to simulate material aging and ensure the comprehensiveness and accuracy of the evaluation model. The local equivalent strain, microcrack initiation location, crack propagation path, and porosity distribution are output by solving the local characteristic scale. The local equivalent strain reflects the local stress state of the material. The microcrack initiation location is 0.1 to 0.5 mm away from the surface. The crack propagation path is along the direction of the maximum principal stress. The porosity distribution affects the local stress concentration. The nanoscale microstructure characteristics were established based on the oxidation theory to build an oxide layer growth model, simulating the selective oxidation process of elements in the oxide layer under high temperature conditions, where the high temperature environment is a working condition greater than 1000 degrees Celsius. During the use of turbine disks, some microscopic features may exist, such as the selective oxidation process of aluminum and chromium, which can lead to an increase in oxide layer thickness and oxide layer cracking, thereby affecting the material's oxidation resistance and fatigue life. Cracks in the oxide layer are prone to become crack initiation points, exacerbating material damage. The oxide layer thickness is calculated based on the oxidation rate constant, activation energy, gas constant, and absolute temperature. The specific formula is d = k * t * exp(-Ea / RT), where d is the oxide layer thickness, k is the oxidation rate constant, t is time, Ea is the activation energy, R is the gas constant, and T is the absolute temperature. This formula can accurately predict the change in oxide layer thickness. Furthermore, when the thermal stress caused by the difference in thermal expansion coefficients between the oxide layer and the substrate exceeds the bonding strength of the oxide layer, the oxide layer is prone to peeling, forming new crack sources, which further aggravates material damage and reduces the service life of the turbine disk. Based on the nanoscale microstructure characteristics, the curve of oxide layer thickness changing with time and the location of oxide layer cracking are obtained. When the microcrack initiation point coincides with the location of oxide layer cracking, the crack propagation rate will be significantly accelerated, the degree of material damage will be aggravated, and the risk of turbine disk failure will be greatly increased. Step 02: Construct a multi-mode degradation model based on the failure mechanism of the turbine disk, and determine the weight of each degradation mode by combining the cross-scale simulation results; Typical failure models of turbine disks include, but are not limited to, fatigue, creep, thermal damage, and oxidation corrosion. The interaction of each failure mode needs to be comprehensively considered, and their contribution to the overall failure should be reflected by weight allocation to ensure that the model can accurately predict the degradation behavior of turbine disks under different operating conditions, thereby optimizing the design and extending service life. Specifically, the multi-mode degradation model includes low-cycle fatigue degradation model, high-temperature creep degradation model, thermomechanical fatigue degradation model, and multi-mode competitive failure model. These models correspond to the failure mechanisms of turbine disks under different stress and temperature environments. By coupling the parameters of each model, the material degradation process under actual working conditions is simulated, and finally, the accurate prediction of turbine disk life is achieved. The coupling of parameters in each model needs to consider the interaction of multiple factors such as temperature, stress, and time to ensure that the simulation results match the actual working conditions. For the low-cycle fatigue degradation model, the cumulative fatigue damage can be calculated based on the cyclic stress amplitude, the number of cycles, and the material fatigue life curve. For the high-temperature creep degradation model, creep stress, temperature, and creep time need to be considered to evaluate the creep deformation. The thermomechanical fatigue model needs to combine temperature cycling and mechanical stress cycling to analyze the impact of thermal stress on the material fatigue life. By integrating the results of each model, the degradation behavior of the turbine disk under different working conditions can be accurately predicted, providing a reliable basis for optimized design and extended service life. Step 03: Perform multi-source data fusion and calibration on the multi-mode degradation model, and correct the model parameters to enhance data features; After determining the multi-mode degradation model, it is necessary to perform multi-source data fusion and calibration to improve the model's generalization ability. This process includes: Multi-source data acquisition is performed, and the acquired multi-source data is preprocessed. The multi-source data includes simulation data, bench test data, and field monitoring data. Simulation data includes parameters for multiple operating conditions, such as temperature, stress, and frequency, covering both design and extreme conditions. Bench test data refers to thermo-cyclic tests conducted on a high-temperature fatigue testing machine, recording crack propagation rate, creep strain, and oxide layer thickness. Field monitoring data is collected through the engine health management system, including turbine disk root strain, center hole temperature, and vibration acceleration time-series signals. Operating parameters such as running time, number of starts, and number of stops are recorded simultaneously to ensure data consistency. Perform data cleaning and feature enhancement on multi-source data, including noise suppression, feature extraction, and data augmentation; Noise suppression employs wavelet soft thresholding to remove high-frequency noise from the vibration signal. Time-frequency analysis is performed on the cleaned data, and the mean stress cycle, stress amplitude, temperature gradient, and kurtosis of vibration energy at the wheel flange root are extracted. The kurtosis of vibration energy reflects the transient characteristics of the vibration signal. Combined with the coupling effect of stress and temperature, a feature vector is constructed. Model calibration and transfer learning are performed, and the error between simulation predictions and actual measurements is minimized based on simulation data. During model calibration and transfer, bench test data is used as a benchmark. The particle swarm optimization algorithm is used to minimize the error between the simulation prediction and the experimental value (e.g., when the crack propagation rate error is greater than 10%, the C and m parameters in the Paris formula are adjusted). An LSTM neural network is used to train the time-series monitoring data. The input is a multi-sensitive feature sequence of length 100, and the output is the remaining lifetime probability distribution. During the transfer learning process, the calibration model of other turbine disk models is used as the source domain and transferred to the new turbine disk model in the same series to achieve cross-domain adaptation of model parameters and improve prediction accuracy. Step 04: Deploy the calibrated multi-mode degradation model in the edge computing terminal to predict the remaining lifetime of the turbine disk and classify the multi-mode failure risk.

[0025] An edge computing terminal is a distributed computing device with real-time data processing capabilities and features low latency and high reliability. It can quickly respond to changes in the operating status of the turbine disk, update the remaining life prediction results in real time, and issue early warnings based on the failure risk level to ensure the safe operation of the equipment. In the process of predicting remaining useful life, random variables in the multi-field coupled model are sampled based on Monte Carlo simulation to calculate the probability distribution of RUL (remaining useful life). For example, 10,000 samples are taken for material parameter fluctuations to simulate the degradation process under different operating conditions. The mean RUL is calculated to be 500 hours, and the RUL distribution within the 95% confidence interval is between 400 and 600 hours. After outputting the probability distribution of the remaining lifespan, the prediction results need to be corrected using an LSTM model to further optimize the prediction accuracy and ensure that the results are closer to the actual operating conditions. Multimode failure risk classification dynamically assesses the failure mode risk of turbine disks based on the remaining lifetime probability distribution and real-time monitoring data. It also uses fuzzy comprehensive evaluation method to quantify and classify failure modes, and constructs a failure mode identification matrix by combining historical failure data to update the risk level in real time. Multi-mode failure risk assessment includes failure probability quantification. Based on the multi-mode degradation model, the failure probability of each mode under the current working condition is calculated. Based on the functional importance and detection difficulty of the turbine disk, a comprehensive assessment is conducted to determine the priority. Then, the risk index is calculated according to the fuzzy comprehensive evaluation method, and the risk index is divided into three levels: high, medium and low, which is fed back to the operation and maintenance personnel in real time. In one possible scenario, a Monte Carlo simulation was performed 10,000 times to calculate the RUL probability distribution (mean RUL = 650 hours, 95% confidence interval [580, 720] hours). The probabilities of each mode were 0.75, 0.15, and 0.1, respectively. The consequence level of high-pressure turbine disk failure was 10, and the detection difficulty of wheel rim root sensor failure was 5. Therefore, the risk index was 0.75*10 / 5=1.5, which is considered low-risk operation. Maintenance personnel can adjust the maintenance plan based on this result, extend the detection cycle, and reduce maintenance costs. At the same time, continuous monitoring of high-risk factors should be maintained to ensure that the equipment operates stably within a safe range.

[0026] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for assessing the lifespan and failure risk of a turbine disk, characterized in that: include: A multi-scale, multi-field coupled simulation model of the turbine disk was performed, and key features were extracted, including millimeter-level overall structural features, micrometer-level local features, and nanometer-level microstructure features. A multi-mode degradation model based on the failure mechanism of turbine disks was constructed, and the weights of each degradation mode were determined by combining cross-scale simulation results. Multi-source data fusion and calibration are performed on the multi-mode degradation model, and the model parameters are corrected to enhance data features; The calibrated multi-mode degradation model is deployed in an edge computing terminal to predict the remaining lifetime of the turbine disk and classify the multi-mode failure risk.

2. The method for assessing the lifespan and failure risk of a turbine disk according to claim 1, characterized in that: The millimeter-level overall structural features were extracted based on the three-dimensional geometric model of the turbine disk, and a thermo-mechanical coupling model was established using finite element software. This thermo-mechanical coupling model includes thermal load analysis and mechanical load analysis. The thermal load analysis is based on the heat transfer analysis of the hot end components of the engine to obtain the radial and axial temperature distribution of the turbine disk. The mechanical load combines aerodynamic load and centrifugal force to apply cyclic load. The aerodynamic force in the aerodynamic load will generate circumferential bending stress, and the centrifugal force is calculated from the rotational speed to obtain the radial stress. The cyclic load is for start-up and shutdown conditions, and stress and strain are calculated using finite element analysis software.

3. The method for assessing the lifespan and failure risk of a turbine disk according to claim 1, characterized in that: Micrometer-level local features are achieved by refining the mesh at the rim root and the surface of the central hole with a unit size of less than 0.1 mm, and by combining it with crystal plasticity theory to describe the material deformation behavior.

4. The method for assessing the lifespan and failure risk of a turbine disk according to claim 3, characterized in that: The local equivalent strain, microcrack initiation location, crack propagation path, and porosity distribution are output by solving the local characteristic scale. The local equivalent strain reflects the local stress state of the material. The microcrack initiation location is 0.1 to 0.5 mm away from the surface. The crack propagation path is along the direction of the maximum principal stress. The porosity distribution affects the local stress concentration.

5. The method for assessing the lifespan and failure risk of a turbine disk according to claim 1, characterized in that: The nanoscale microstructure characteristics were established based on the oxidation theory to build an oxide layer growth model, simulating the selective oxidation process of elements in the oxide layer under high temperature conditions, where the high temperature environment is a working condition greater than 1000 degrees Celsius.

6. The method for assessing the lifespan and failure risk of a turbine disk according to claim 5, characterized in that: Based on the nanoscale microstructure characteristics, the curve of oxide layer thickness changing with time and the location of oxide layer cracks are calculated. When the microcrack initiation point coincides with the oxide layer crack location, it is determined that the crack propagation rate is accelerated, the degree of material damage is aggravated, and the risk of turbine disk failure is increased.

7. The method for assessing the lifespan and failure risk of a turbine disk according to claim 1, characterized in that: The multi-mode degradation model includes low-cycle fatigue degradation model, high-temperature creep degradation model, thermomechanical fatigue degradation model and multi-mode competitive failure model. These models correspond to the failure mechanisms of turbine disks under different stress and temperature environments.

8. The method for assessing the lifespan and failure risk of a turbine disk according to claim 1, characterized in that: The process of multi-source data fusion and calibration for multi-mode degradation models includes: acquiring multi-source data and preprocessing the acquired multi-source data, which includes simulation data, bench test data, and field monitoring data; cleaning and feature enhancement of the multi-source data, including noise suppression, feature extraction, and data augmentation; calibrating the model and performing transfer learning, and minimizing the error between the simulation prediction and the actual measurement based on the simulation data.

9. The method for assessing the lifespan and failure risk of a turbine disk according to claim 8, characterized in that: Noise suppression employs wavelet soft thresholding to remove high-frequency noise from the vibration signal. Time-frequency analysis is then performed on the cleaned data, and the mean stress cycle, stress amplitude, temperature gradient, and kurtosis of vibration energy at the wheel flange root are extracted. The kurtosis of vibration energy reflects the transient characteristics of the vibration signal. Combined with the coupling effect of stress and temperature, a feature vector is constructed.

10. The method for assessing the lifespan and failure risk of a turbine disk according to claim 1, characterized in that: In the process of predicting remaining lifetime, random variables in the multi-field coupled model are sampled based on Monte Carlo simulation to calculate the probability distribution of remaining lifetime, and the prediction results are corrected by LSTM model.

Images