Rotor fatigue life analysis method and device

Abstract

The invention relates to a rotor fatigue life analysis method and device, and relates to the technical field of steam turbine rotor performance analysis, and the method comprises the steps: determining a steam turbine rotor geometric model, and determining a stress field analysis result of a steam turbine rotor according to the steam turbine rotor geometric model; determining a rotor stress monitoring point of the steam turbine rotor and the actual stress of the monitoring point according to the stress field analysis result, and determining an initial stress point parameter of the rotor stress monitoring point; based on the monitoring point actual stress of the rotor stress monitoring point and the initial stress point parameters, performing parameter adjustment on the to-be-trained neural network model to obtain a rotor stress prediction model; and performing stress prediction on the to-be-predicted steam turbine rotor based on the rotor stress prediction model to obtain an equivalent stress prediction result, and analyzing the fatigue life of the to-be-predicted steam turbine rotor according to the equivalent stress prediction result. The analysis efficiency is improved.

Description

Technical Field

[0001] This disclosure relates to the field of performance analysis technology for steam turbine rotors, and more specifically, to a method and apparatus for analyzing rotor fatigue life. Background Technology

[0002] In existing methods for analyzing rotor fatigue life, finite element simulation can be used to analyze the fatigue life of turbine rotors. However, this method has the following limitations: on the one hand, the calculation time is too long, resulting in low analysis efficiency, which makes it difficult to meet the real-time requirements of online monitoring and rapid early warning; on the other hand, the adaptability of the simulation model is poor, which leads to low accuracy of the obtained analysis results.

[0003] It should be noted that the information in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this disclosure is to provide a method and apparatus for analyzing rotor fatigue life, thereby overcoming, to some extent, the problems of low analysis efficiency and low accuracy of analysis results caused by the limitations and defects of related technologies.

[0005] According to one aspect of this disclosure, a method for analyzing rotor fatigue life is provided, comprising: determining a turbine rotor geometric model, and determining stress field analysis results of the turbine rotor based on the turbine rotor geometric model; determining rotor stress monitoring points and actual stresses at the monitoring points of the turbine rotor based on the stress field analysis results, and determining initial stress point parameters of the rotor stress monitoring points; adjusting the parameters of a neural network model to be trained based on the actual stresses at the monitoring points and the initial stress point parameters to obtain a rotor stress prediction model; performing stress prediction on the turbine rotor to be predicted based on the rotor stress prediction model to obtain equivalent stress prediction results, and analyzing the fatigue life of the turbine rotor to be predicted based on the equivalent stress prediction results.

[0006] In one exemplary embodiment of this disclosure, determining the turbine rotor geometric model includes: acquiring a two-dimensional rotor model of a solar thermal turbine, and splitting the two-dimensional rotor model to obtain multiple vector elements to be assembled; simplifying each of the vector elements to be assembled to obtain simplified vector elements, and generating an initial rotor geometric model based on the simplified vector elements; performing mesh generation on the initial rotor geometric model to obtain an initial mesh generation result, and performing mesh refinement processing on the initial mesh generation result to obtain the turbine rotor geometric model of the solar thermal turbine.

[0007] In one exemplary embodiment of this disclosure, a turbine rotor geometric model of a solar thermal turbine is obtained by mesh refinement processing of an initial mesh division result. This includes: analyzing the initial mesh division result to obtain regions requiring mesh refinement processing; wherein the regions requiring refinement processing include at least one of the following: rounded corner transition regions, regions with discontinuous geometric shapes, and regions where the temperature difference between the rotor surface and the rotor interior exceeds a preset temperature difference threshold due to transient heat transfer processes; performing mesh refinement processing on the regions requiring refinement to obtain a refined mesh division result, and calculating the heat transfer coefficient of each different region in the refined mesh division result based on preset steam parameters; configuring the heat transfer coefficient for each different region in the refined mesh division result to obtain the turbine rotor geometric model of the solar thermal turbine.

[0008] In one exemplary embodiment of this disclosure, determining the stress field analysis result of the turbine rotor based on the turbine rotor geometric model includes: performing a temperature field analysis on the turbine rotor geometric model based on preset boundary conditions to obtain a first temperature field analysis result of the turbine rotor; and performing a stress analysis on the turbine rotor geometric model based on the first temperature field analysis result to obtain a stress field analysis result of the turbine rotor.

[0009] In an exemplary embodiment of this disclosure, temperature field analysis is performed on the turbine rotor geometric model based on preset boundary conditions to obtain a first temperature field analysis result of the turbine rotor. This includes: configuring preset boundary conditions for the turbine rotor geometric model; wherein the preset boundary conditions include at least one of transient thermal boundary conditions, mechanical load conditions, and displacement boundary conditions; obtaining a turbine start-up characteristic curve corresponding to the turbine rotor, and determining initial steady-state thermal boundary conditions for steady-state analysis of the turbine rotor based on the turbine start-up characteristic curve; loading the initial steady-state thermal boundary conditions into the turbine rotor geometric model to obtain a rotor steady-state temperature field, and determining initial transient conditions for transient analysis of the turbine rotor based on the rotor steady-state temperature field; and performing transient analysis on the turbine rotor based on the initial transient conditions and the transient thermal boundary conditions to obtain a first temperature field analysis result of the turbine rotor.

[0010] In an exemplary embodiment of this disclosure, determining the rotor stress monitoring point and the actual stress at the monitoring point of the steam turbine rotor based on the stress field analysis results, and determining the initial stress point parameters of the rotor stress monitoring point, includes: determining the stress concentration region of the steam turbine rotor based on the stress field analysis results, and determining the rotor stress monitoring point of the steam turbine rotor based on the stress concentration region; determining the actual stress at the monitoring point of the rotor stress monitoring point based on the stress field analysis results, and determining the initial temperature, heating rate, initial heating time, and temperature rise rate of the rotor stress monitoring point; and generating the initial stress point parameters based on the initial temperature, heating rate, initial heating time, and temperature rise rate of the stress monitoring point.

[0011] In one exemplary embodiment of this disclosure, a rotor stress prediction model is obtained by adjusting the parameters of a neural network model to be trained based on the actual stress at the rotor stress monitoring point and the initial stress point parameters. This includes: inputting the initial stress point parameters of the rotor stress monitoring point into the neural network model to be trained to obtain the predicted stress at the monitoring point; constructing a target loss function based on the predicted stress at the monitoring point and the actual stress at the monitoring point; and adjusting the parameters of the neural network model to be trained based on the target loss function to obtain the rotor stress prediction model.

[0012] In an exemplary embodiment of this disclosure, the neural network model to be trained includes a feature expansion layer, an initial convolutional layer, a downsampling path, and an upsampling path. The process of inputting the initial stress point parameters of the rotor stress monitoring point into the neural network model to obtain the predicted stress of the monitoring point includes: expanding the initial temperature based on the initial stress point parameters using the feature expansion layer according to the rate of increase, initial heating time, and temperature rise rate to obtain expanded temperatures at multiple different time points; performing convolution processing on the expanded temperatures at the multiple different time points based on the initial convolutional layer to obtain convolution processing results at multiple different time points; performing downsampling processing on the convolution processing results at the multiple different time points based on the downsampling path to obtain local temperature sampling results at different time points; and performing upsampling processing on the local temperature sampling results at the multiple different time points based on the upsampling path to obtain the predicted stress of the monitoring point.

[0013] In one exemplary embodiment of this disclosure, stress prediction of the turbine rotor to be predicted is performed based on the rotor stress prediction model to obtain equivalent stress prediction results, and fatigue life of the turbine rotor to be predicted is analyzed based on the equivalent stress prediction results. This includes: determining the rotor stress parameters of the turbine rotor to be predicted, and inputting the rotor stress parameters into the rotor stress prediction model to perform stress prediction on the turbine rotor to obtain equivalent stress prediction results of the turbine rotor at different time points; obtaining the rotor material parameters of the turbine rotor to be predicted, and inputting the rotor material parameters and the equivalent stress prediction results of the turbine rotor to be predicted at different time points into a preset fatigue life analysis model to obtain fatigue life analysis results of the turbine rotor to be predicted.

[0014] According to one aspect of this disclosure, a rotor fatigue life analysis apparatus is provided, comprising: a stress field analysis result determination module, used to determine a turbine rotor geometric model and determine the stress field analysis result of the turbine rotor based on the turbine rotor geometric model; a rotor stress monitoring point determination module, used to determine the rotor stress monitoring points and actual stresses at the monitoring points of the turbine rotor based on the stress field analysis result, and determine the initial stress point parameters of the rotor stress monitoring points; a rotor stress prediction model determination module, used to adjust the parameters of a neural network model to be trained based on the actual stresses at the monitoring points and the initial stress point parameters to obtain a rotor stress prediction model; and a rotor fatigue life analysis module, used to perform stress prediction on the turbine rotor to be predicted based on the rotor stress prediction model to obtain equivalent stress prediction results, and analyze the fatigue life of the turbine rotor to be predicted based on the equivalent stress prediction results.

[0015] This disclosure provides a method for analyzing rotor fatigue life. Firstly, it determines a turbine rotor geometric model and, based on this model, determines the stress field analysis results of the turbine rotor. Then, based on the stress field analysis results, it determines the rotor stress monitoring points and the actual stress at each monitoring point, and determines the initial stress point parameters. Next, based on the actual stress at the monitoring points and the initial stress point parameters, it adjusts the parameters of the neural network model to be trained to obtain a rotor stress prediction model. Finally, based on the rotor stress prediction model, it predicts the stress of the turbine rotor to be predicted to obtain equivalent stress prediction results, and then, based on the equivalent stress… The force prediction results can be used to analyze the fatigue life of the turbine rotor under test without relying on finite element simulation, thus solving the problem of low analysis efficiency caused by the need for finite element simulation in the existing technology. On the other hand, since the equivalent stress prediction results can be obtained by predicting the stress of the turbine rotor under test based on the rotor stress prediction model, and the fatigue life of the turbine rotor under test can be analyzed based on the equivalent stress prediction results, the problem of low accuracy of analysis results caused by poor adaptability of simulation models is solved, thereby improving the accuracy of the obtained fatigue life analysis results.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0018] Figure 1 The flowchart illustrates a method for analyzing rotor fatigue life according to an exemplary embodiment of the present disclosure.

[0019] Figure 2 The diagram schematically illustrates a structural example of a rotor stress prediction model obtained according to an exemplary embodiment of the present disclosure.

[0020] Figure 3 The diagram schematically illustrates a structural example of a two-dimensional rotor model of a solar thermal turbine according to an exemplary embodiment of the present disclosure.

[0021] Figure 4 The diagram schematically illustrates an example of thermal boundary conditions for a low-pressure rotor and a high-pressure rotor according to an exemplary embodiment of the present disclosure.

[0022] Figure 5 An example diagram illustrating a cold start-first start curve of a medium-low pressure rotor according to an exemplary embodiment of the present disclosure is shown.

[0023] Figure 6 An example diagram illustrating the first temperature field analysis results of a medium-low pressure rotor obtained according to an exemplary embodiment of the present disclosure is shown.

[0024] Figure 7 An example diagram illustrating the stress field analysis results of a medium- and low-pressure rotor obtained according to an exemplary embodiment of the present disclosure is shown.

[0025] Figure 8 An example diagram illustrating the rotor stress monitoring points of a medium- or low-pressure rotor obtained according to an exemplary embodiment of the present disclosure is shown.

[0026] Figure 9 The diagram schematically illustrates a structural example of a rotor fatigue life analysis apparatus according to an exemplary embodiment of the present disclosure.

[0027] Figure 10 The diagram schematically illustrates an example structure of an electronic device for implementing a rotor fatigue life analysis method according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0028] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of these specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0029] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0030] As the core component of a solar thermal turbine (CST) unit system, the turbine rotor plays a crucial role in converting high-temperature thermal energy into mechanical energy. In practical applications, the turbine rotor operates in extremely harsh environments, enduring prolonged high temperatures, frequent start-ups and shutdowns, and severe load fluctuations. This presents it with multiple challenges, including thermo-mechanical coupling stress, high-temperature creep, and low-cycle fatigue. Therefore, accurate thermo-mechanical coupling analysis of the rotor is essential for ensuring the safety, reliability, and long-term operation of the solar thermal turbine unit.

[0031] Furthermore, due to the influence of day and night and weather on solar energy resources, solar thermal turbine units must adapt to frequent start-ups and load changes, which places extremely high demands on the turbine rotor's rapid start-up capability and fatigue resistance. However, rapid load changes lead to a sharp increase in the surface steam temperature rise rate of the turbine rotor, creating a significant transient temperature gradient inside the rotor and inducing substantial thermal stress. Consequently, under the combined action of alternating mechanical and thermal stresses, the turbine rotor material is prone to low-cycle fatigue damage, which can lead to crack initiation and severely limit the unit's lifespan, even causing equipment failure. Therefore, accurately analyzing the fatigue life of the turbine rotor has become a problem that needs to be solved.

[0032] To address the aforementioned issues, some related technical solutions rely on finite element method (FEM) simulation to analyze the fatigue life of turbine rotors. However, this method has several limitations: firstly, the computation time is too long, making it difficult to meet the real-time requirements of online monitoring and rapid early warning; secondly, the simulation model lacks adaptability and computational efficiency when facing the unsteady and random fluctuation conditions unique to concentrated solar power (CSP); and thirdly, the accuracy of FEM analysis heavily depends on the accuracy of boundary condition settings, while heat transfer boundaries in actual operation are often uncertain. Therefore, these technical bottlenecks make it difficult for traditional methods to effectively support rotor condition assessment and predictive maintenance, highlighting the urgent need for efficient and accurate new analysis tools. Against this backdrop, with the rapid development of artificial intelligence and deep learning technologies, the fusion strategy of surrogate models and heuristic algorithms has made significant research progress in the field of turbine rotor stress prediction and optimization. However, related technical solutions do not document a method for accurately and efficiently predicting the stress of turbine rotors based on stress prediction models.

[0033] Based on this, this exemplary embodiment first provides a method for analyzing rotor fatigue life, which can run on terminal devices, servers, server clusters, or cloud servers, etc. Of course, those skilled in the art can also run the method disclosed herein on other platforms as needed, and this exemplary embodiment does not impose any special limitations on this. Specifically, refer to... Figure 1 As shown, the rotor fatigue life analysis method described in the exemplary embodiments of this disclosure may include the following steps: Step S110. Determine the turbine rotor geometric model, and determine the stress field analysis results of the turbine rotor based on the turbine rotor geometric model; Step S120. Based on the stress field analysis results, determine the rotor stress monitoring points and actual stresses at the monitoring points of the turbine rotor, and determine the initial stress point parameters of the rotor stress monitoring points; Step S130. Based on the actual stress at the monitoring point of the rotor stress monitoring point and the initial stress point parameters, adjust the parameters of the neural network model to be trained to obtain the rotor stress prediction model. Step S140. Based on the rotor stress prediction model, perform stress prediction on the turbine rotor to be predicted to obtain the equivalent stress prediction result, and analyze the fatigue life of the turbine rotor to be predicted based on the equivalent stress prediction result.

[0034] In the rotor fatigue life analysis method described above, on the one hand, the geometric model of the turbine rotor is determined, and the stress field analysis results of the turbine rotor are determined based on the turbine rotor geometric model; then, based on the stress field analysis results, the rotor stress monitoring points and the actual stress at the monitoring points are determined, and the initial stress point parameters of the rotor stress monitoring points are determined; then, based on the actual stress at the monitoring points and the initial stress point parameters, the parameters of the neural network model to be trained are adjusted to obtain the rotor stress prediction model; finally, based on the rotor stress prediction model, the stress of the turbine rotor to be predicted is predicted to obtain the equivalent stress prediction result, and based on the equivalent stress... The prediction results analyze the fatigue life of the turbine rotor to be predicted without relying on finite element simulation, thus solving the problem of low analysis efficiency caused by the need for finite element simulation in existing technologies. On the other hand, since the equivalent stress prediction results can be obtained by predicting the stress of the turbine rotor to be predicted based on the rotor stress prediction model, and the fatigue life of the turbine rotor to be predicted can be analyzed based on the equivalent stress prediction results, the problem of low accuracy of analysis results caused by poor adaptability of simulation models is solved, thereby improving the accuracy of the obtained fatigue life analysis results.

[0035] The following will provide a detailed explanation and description of the rotor fatigue life analysis method described in the exemplary embodiments of this disclosure, in conjunction with the accompanying drawings.

[0036] First, the technical implementation principle of the exemplary embodiments of this disclosure will be explained and described. Specifically, addressing the numerous limitations in the existing stress assessment process for turbine rotors of solar thermal turbines, the exemplary embodiments of this disclosure provide a method for analyzing rotor fatigue life. The aim is to provide a novel HRENet model for stress prediction of solar thermal turbine rotors under different starting parameters. The HRENet model predicts the stress at critical points under different starting parameters based on the stress results at critical points during the cold start-up process of the rotor. Based on this, high-precision and high-efficiency stress prediction can be provided for turbine rotors under different starting parameters.

[0037] Based on this, to achieve the above objectives, the present disclosure provides the following technical solution in an example embodiment: First, a geometric model of a solar thermal turbine rotor is constructed; specifically, the process involves: obtaining the assembly drawing of the solar thermal turbine rotor; constructing an initial geometric model of the rotor based on the assembly drawing using a finite element analysis tool; refining the mesh based on the initial rotor geometric model to generate a turbine rotor geometric model; second, determining the monitoring points in the rotor geometric model using a finite element analysis tool; specifically, generating the rotor temperature field results based on preset boundary conditions using finite element analysis software; generating the rotor stress field results based on the temperature field results; determining the monitoring points based on the rotor stress field results; further, training a pre-constructed rotor stress prediction model based on the temperature and stress of the monitoring points, and predicting the stress trend of the monitoring points; specifically, constructing a model dataset based on the monitoring points using a finite element analysis tool; wherein, the obtained The obtained dataset can include training and testing datasets. Based on this, HRENet is trained using the model dataset to construct a rotor stress field prediction model. Simultaneously, the model training process can be achieved as follows: HRENet is trained using the training dataset to generate an initial prediction model; using a preset loss function and optimizer, the initial prediction model is optimized using the testing dataset to generate a rotor stress field prediction model; finally, using the trained rotor stress prediction model, the stress changes at monitoring points under different conditions are determined based on changes in input parameters. Specifically, this can be achieved as follows: by setting different initial temperatures, heating rates, initial heating times, and temperature rise rates, the corresponding rotor thermal boundary conditions and mechanical load histories can be derived; based on this, thermo-mechanical coupling analysis methods can be further used to determine the transient and steady-state stress responses of key rotor monitoring points under different operating conditions.

[0038] Furthermore, with the rapid development of artificial intelligence technology, intelligent algorithms such as neural networks have demonstrated significant advantages in the industrial field due to their powerful learning capabilities, thus providing a new technical path for the real-time prediction and monitoring of the stress state of steam turbine units. The Hybrid Residual Encoding Network (HRENet) is a deep learning model for accurately predicting the stress concentration points of the rotor of a solar-thermal steam turbine under different starting parameters. It achieves an excellent balance between feature extraction depth, spatial information preservation, computational efficiency, and engineering generalization ability, making it an efficient and reliable solution in the field of structural mechanics response prediction. Based on this, the technical solution disclosed in the example embodiments of this disclosure can drive the prediction of rotor stress field data based on Hybrid Residual Coding Network (HRENet). In practical applications, firstly, for the cold start condition of the turbine rotor, high-fidelity finite element simulation is used to systematically calculate and obtain the full-field data of the temperature field and stress field of the rotor during the transient process. On this basis, a high-quality dataset for model training is constructed to establish a rotor stress prediction model. The rotor stress prediction model established in this way can predict the full-field stress distribution trend of the rotor in the future time series with high accuracy and efficiency based on the real-time temperature monitoring data of a limited number of measuring points during the rotor start-up and shutdown process. Furthermore, compared with traditional mechanism models, the data-driven deep learning model proposed in the example embodiments of this disclosure has significant advantages such as high prediction accuracy, fast response speed, and strong generalization stability, and can provide key real-time stress state information and decision support for rotor life assessment and operation optimization of turbines under different start-up parameters.

[0039] The Hybrid Residual Coding Network (HRENet) involved in the exemplary embodiments of this disclosure will be explained and described below. Specifically, refer to... Figure 2 As shown, the Hybrid Residual Coding Network (HRENet) may include a main input layer 201, a feature expansion layer 202, an initial convolutional layer 203, a downsampling path 204, an upsampling path 205, and an output layer 206. The feature expansion layer described herein may consist of two fully connected layers, the initial convolutional layer described herein may consist of a convolutional layer and a normalization layer, the downsampling path described herein may consist of multiple residual layers, and the upsampling path described herein consists of multiple concatenation layers and residual layers.

[0040] The following will combine Figure 2 right Figure 1 The analysis method for rotor fatigue life shown will be further explained and illustrated. Specifically: In step S110, the turbine rotor geometric model is determined, and the stress field analysis results of the turbine rotor are determined based on the turbine rotor geometric model.

[0041] In this example embodiment, firstly, the rotor geometric model of the steam turbine is determined; specifically, this can be achieved as follows: obtaining a two-dimensional rotor model of the solar thermal steam turbine, and splitting the two-dimensional rotor model to obtain multiple vector elements to be assembled; simplifying each of the vector elements to be assembled to obtain simplified vector elements, and generating an initial rotor geometric model based on the simplified vector elements; meshing the initial rotor geometric model to obtain an initial meshing result, and refining the initial meshing result to obtain the turbine rotor geometric model of the solar thermal steam turbine.

[0042] In one example embodiment, simplifying each vector element to be assembled to obtain simplified vector elements can be achieved as follows: selecting blade vector elements corresponding to the blade structure and blade root groove vector elements corresponding to the blade root groove structure from the vector elements to be assembled; simplifying the blade vector elements and blade root groove vector elements respectively to obtain simplified blade vector elements and simplified blade root groove vector elements. It should be noted that the reason for simplifying the blade vector elements and blade root groove vector elements is to ensure that the obtained initial rotor geometric model constitutes a complex continuous geometric structure to facilitate meshing; furthermore, the simplification process described here may include, but is not limited to, scaling or reducing lines and surfaces, etc.

[0043] In one example embodiment, obtaining an initial meshing result by meshing an initial rotor geometric model can be achieved as follows: The initial rotor geometric model is imported into a preset meshing software, and the imported initial rotor geometric model is simplified based on the preset meshing software to obtain a simplified initial rotor geometric model; the meshing parameters required for meshing the simplified initial rotor geometric model are determined; wherein, the meshing parameters include the mesh type and the mesh element size; the simplified initial rotor geometric model is meshed according to the meshing parameters using the preset meshing software to obtain the initial meshing result.

[0044] In an exemplary embodiment, obtaining the turbine rotor geometric model of a solar thermal turbine by refining the initial mesh division result can be achieved as follows: Analyzing the initial mesh division result to obtain the region to be refined; wherein the region to be refined includes at least one of the following: a rounded corner transition region, a region with discontinuous geometry, and a region where the temperature difference between the rotor surface and the rotor interior is greater than a preset temperature difference threshold due to a transient heat transfer process; refining the mesh in the region to be refined to obtain a refined mesh division result, and calculating the heat transfer coefficient of each different region in the refined mesh division result according to preset steam parameters; configuring the heat transfer coefficient for each different region in the refined mesh division result to obtain the turbine rotor geometric model of the solar thermal turbine.

[0045] The following section will further explain and illustrate the specific process of determining the geometric model of the turbine rotor of a solar thermal turbine. Specifically, in practical applications, firstly, a two-dimensional rotor model of the turbine rotor of the solar thermal turbine is obtained; here, the two-dimensional rotor model is a two-dimensional finite element rotor model; this two-dimensional finite element rotor model refers to a model in which the planar region corresponding to the model is discretized into a finite number of elements using the finite element method and then numerically analyzed. In practical applications, the two-dimensional rotor model of medium and low-pressure rotors can be referenced. Figure 3 As shown in 301, the two-dimensional rotor model of the high-pressure rotor can be referenced. Figure 3 As shown in 302; it should also be noted that the rotor fatigue life analysis method described in the example embodiments of this disclosure is based on a medium- and low-pressure rotor; in practical applications, the medium- and low-pressure rotor is the core and weak link of the steam turbine in terms of structural load-bearing, energy conversion, safe operation and life management; during the operation of the steam turbine, the turbine rotor is not only a key component for ultimately converting thermal energy into electrical energy, but its complex mechanical-thermal combined load state also makes it the focus of the entire unit's life prediction and operation strategy optimization; therefore, the condition monitoring, accurate stress prediction and life assessment of the medium- and low-pressure rotor are of paramount importance to ensuring the safe, efficient and flexible operation of large steam turbine units; based on this, constructing an accurate steam turbine rotor geometric model is very important for the specific stress prediction process. Furthermore, the high-pressure rotor and the medium- and low-pressure rotor involved in the example embodiments of this disclosure are both made of 30Cr1Mo1V; the specific material properties of 30Cr1Mo1V material parameters can be referred to in Table 1 below.

[0046] Table 1

[0047] Secondly, an initial rotor geometric model is constructed based on the turbine rotor assembly drawing using a finite element analysis (FEM) tool. The FEM tool described here is a computer simulation technique based on the principles of "discretization" and "numerical approximation." The basic idea of ​​FEM is to divide a complex continuous geometric structure into a large number of simple, finite-sized micro-elements. These elements are interconnected through nodes to form a "mesh." Based on this, physical equations are established for each simple element, and then assembled into a large overall set of equations and solved to approximate the behavior of the original complex structure under various loads (such as temperature and stress). The FEM tool used here can be selected according to actual needs; this example does not impose any special restrictions.

[0048] Then, the initial rotor geometric model is meshed to generate the turbine rotor geometric model. Specifically, this can be achieved as follows: During mesh generation, significant stress concentration is easily induced in geometrically discontinuous regions of the rotor (such as fillet transition zones between stages and abrupt shape changes). Furthermore, during variable operating conditions, the complex surface heat transfer conditions in these regions lead to highly nonlinear thermal boundary conditions, resulting in severe temperature gradients and thermal stresses. Numerical calculations show that the maximum equivalent stress in these regions, under the coupled effects of the above factors, may exceed the material's yield strength, easily inducing local plastic strain. Moreover, this cyclic plastic deformation is the main mechanism for low-cycle fatigue damage accumulation in the rotor, directly limiting its service life. Therefore, to achieve high-precision elastoplastic analysis and fatigue life prediction, a local mesh refinement strategy must be implemented in these key areas to accurately capture stress gradients and plastic strain distributions. The stress concentration areas requiring special attention include, but are not limited to: fillet transition zones, geometrically discontinuous areas, and locations where significant temperature differences between the surface and interior occur due to transient heat transfer processes. Furthermore, based on steam parameters, empirical formulas are used to calculate the heat transfer coefficients of various parts of the rotor. These steam parameters include, but are not limited to, temperature, pressure, and flow rate. The empirical formulas include, but are not limited to, the Dittus-Boelter formula and the turbine manufacturer's formula. In addition, empirical formulas can be selected based on the required heat transfer coefficient and actual conditions; this embodiment of the invention does not limit this selection. The reason for determining the heat transfer coefficients of each part is that the heat transfer coefficient determines the rate of heat exchange between the rotor surface and the working fluid (such as steam). A higher value indicates faster heat transfer and a stronger work capacity of the unit. For example, optimizing the heat transfer coefficient of the moving blades can reduce heat loss of steam on the blade surface and improve the efficiency of steam expansion work. Based on this, the accuracy of the obtained stress field analysis results can be further improved.

[0049] Secondly, the stress field analysis results of the turbine rotor are determined based on the turbine rotor geometric model. Specifically, this can be achieved as follows: temperature field analysis is performed on the turbine rotor geometric model based on preset boundary conditions to obtain the first temperature field analysis result of the turbine rotor; stress analysis is performed on the turbine rotor geometric model based on the first temperature field analysis result to obtain the stress field analysis result of the turbine rotor.

[0050] In an exemplary embodiment, a temperature field analysis is performed on the turbine rotor geometric model based on preset boundary conditions to obtain a first temperature field analysis result for the turbine rotor. This can be achieved by: configuring preset boundary conditions for the turbine rotor geometric model; wherein the preset boundary conditions include at least one of transient thermal boundary conditions, mechanical load conditions, and displacement boundary conditions; obtaining the turbine start-up characteristic curve corresponding to the turbine rotor, and determining the initial steady-state thermal boundary conditions for steady-state analysis of the turbine rotor based on the turbine start-up characteristic curve; loading the initial steady-state thermal boundary conditions into the turbine rotor geometric model to obtain the rotor steady-state temperature field, and determining the initial transient conditions for transient analysis of the turbine rotor based on the rotor steady-state temperature field; performing transient analysis on the turbine rotor based on the initial transient conditions and the transient thermal boundary conditions to obtain a first temperature field analysis result for the turbine rotor.

[0051] The following will further explain and illustrate the specific process for determining the stress field analysis results. Specifically, the determination of the stress field analysis results can be achieved using finite element analysis tools combined with preset boundary conditions. These preset boundary conditions can include thermal boundary conditions, mechanical load conditions, and displacement boundary conditions, etc. Of course, the required boundary conditions can also be set according to actual needs and experience; this embodiment of the invention does not limit this. Taking thermal boundary conditions as an example, the thermal boundary conditions for the low-pressure rotor can be referenced... Figure 4 As shown in 401; the thermal boundary conditions of the high-pressure rotor can be referenced. Figure 4 As shown in 402; furthermore, in determining the results of the first temperature field analysis, this can be achieved as follows: First, determine the initial thermal boundary conditions of the rotor based on the start-up curve (i.e., the turbine start-up characteristic curve); taking cold start-first start as an example, the cold start-first start curve can include the changes of parameters such as main steam temperature, reheat steam temperature, speed, main steam pressure, reheat steam pressure, main steam flow rate, reheat flow rate, and load over time. Specific cold start-first start curves can be found by referring to... Figure 5As shown; secondly, the initial thermal boundary conditions of the rotor are loaded into the rotor geometric model, and the steady-state temperature field of the rotor is calculated using finite element analysis tools; then, the steady-state temperature field is used as the initial condition for transient analysis, and transient temperature field analysis is performed using finite element analysis tools based on the transient thermal analysis boundary conditions to obtain the first temperature field analysis result; the obtained first temperature field analysis result of the medium and low pressure rotor can be referenced. Figure 6 As shown; where 601 is the rotor temperature field at time P0 (1560 s); 602 is the rotor temperature field at time P1 (2298 s); and 603 is the rotor temperature field at the moment of maximum stress (6198.1 s). Based on this, the transient temperature field results (i.e., the first temperature field analysis results) are imported into the structural analysis module to calculate the stress distribution of the rotor under different operating conditions, generating rotor stress field results (i.e., stress field analysis results). The obtained stress field analysis results for medium and low pressure rotors can be used as a reference. Figure 7 As shown; where 701 is the rotor stress cloud diagram at time P0 (1560 s); 702 is the rotor stress cloud diagram at time P1 (2298 s); and 703 is the rotor stress cloud diagram at the time of maximum stress (6198.1 s).

[0052] In step S120, the rotor stress monitoring point and the actual stress at the monitoring point of the turbine rotor are determined based on the stress field analysis results, and the initial stress point parameters of the rotor stress monitoring point are determined.

[0053] Specifically, this can be achieved as follows: Determine the stress concentration region of the turbine rotor based on the stress field analysis results, and determine the rotor stress monitoring points of the turbine rotor based on the stress concentration region; determine the actual stress at the monitoring points of the rotor stress monitoring points based on the stress field analysis results, and determine the initial temperature, rate of increase, initial heating time, and temperature rise rate of the rotor stress monitoring points; generate the initial stress point parameters based on the initial temperature, rate of increase, initial heating time, and temperature rise rate of the stress monitoring points. That is, in practical applications, rotor monitoring points can be determined based on the rotor stress field results (i.e., stress field analysis results). In determining the rotor monitoring points, given that a high level of thermal stress may appear on the outer surface of the rotor when the main steam temperature rises, especially on the inlet side, exhaust side, and at the positions of each stage of the turbine disk and steam seal, stress concentration phenomena are prone to occur. Therefore, multiple key monitoring points can be selected for detection. Thus, corresponding key points need to be selected as rotor stress monitoring points in the stress concentration region. The rotor stress monitoring points obtained using a medium- and low-pressure rotor as an example can be used as a reference. Figure 8 As shown; in Figure 8In the example diagram shown, assuming the low-pressure rotor geometry is a two-dimensional axisymmetric finite element model, three monitoring points can be selected on the rotor's outer surface, as shown in A, B, and C. Furthermore, by conducting in-depth analysis of the stress values ​​at these monitoring points, the rotor's safety performance under actual operating conditions can be more accurately detected, providing a basis for optimized design. Moreover, the actual stress at the monitoring points mentioned above can be determined from the stress field analysis results. Finally, the initial temperature, heating rate, initial heating time, and temperature rise rate mentioned above can be determined according to actual needs. Simultaneously, to improve the accuracy of the obtained rotor stress prediction model, the initial temperature, heating rate, initial heating time, and temperature rise rate can be determined based on the temperature corresponding to the actual stress at the monitoring points, or they can be determined in other ways; this example does not impose any special restrictions on this.

[0054] In step S130, based on the actual stress at the monitoring point of the rotor stress monitoring point and the initial stress point parameters, the parameters of the neural network model to be trained are adjusted to obtain the rotor stress prediction model.

[0055] Specifically, the model training process can be implemented as follows: The initial stress point parameters of the rotor stress monitoring point are input into the neural network model to be trained to obtain the predicted stress of the monitoring point; a target loss function is constructed based on the predicted stress and the actual stress of the monitoring point; and the parameters of the neural network model to be trained are adjusted based on the target loss function to obtain the rotor stress prediction model. The specific determination process of the predicted stress of the monitoring point can be determined as follows: Based on the feature expansion layer, the initial temperature is feature-expanded according to the rise rate, initial heating time, and temperature rise rate in the initial stress point parameters to obtain expanded temperatures at multiple different time nodes; based on the initial convolutional layer, the expanded temperatures at multiple different time nodes are convolved to obtain convolution processing results at multiple different time nodes; based on the downsampling path, the convolution processing results at multiple different time nodes are downsampled to obtain local temperature sampling results at different time nodes; based on the upsampling path, the local temperature sampling results at multiple different time nodes are upsampled to obtain the predicted stress of the monitoring point.

[0056] The following will further explain and illustrate the specific training process of the model. Specifically, in practical applications, it is first necessary to construct a model dataset based on the rotor stress monitoring point data; specifically, in practical applications, since the temperature and stress of the monitoring points will change with time during the startup process, there is a certain correlation between the sample data of adjacent time points; therefore, temperature and stress prediction can be regarded as a time series modeling problem, and the stress and the temperature corresponding to the stress obtained by the finite element cold start simulation calculation are used as the original data to normalize the data, thereby improving the convergence of the neural network. The normalization formula can be shown in the following formula (1): ; Formula (1) in, For the original data that needs to be normalized, For the normalized data, and These represent the maximum and minimum values ​​for each feature in the entire training dataset; furthermore, during actual training, the initial temperature can be... rate of increase Initial heating time and temperature rise rate Four key physical parameters are used as input data, and the corresponding actual stress values ​​are used as the corresponding label data. The combination of input data and corresponding label data is used to determine the model dataset. Then, HRENet is trained using the model dataset to construct a rotor stress field prediction model. The model dataset includes a training dataset and a test dataset. Specifically, the normalized model dataset is divided into a training dataset and a test dataset according to a preset ratio. The preset ratio can be set according to actual needs. As an optional scheme, the ratio between the training dataset and the test dataset is 8:2.

[0057] Furthermore, a rotor stress field prediction model based on monitoring points was established using HRENet. The feature width was reduced and the feature depth was increased by using a downsampling path, which transformed the low-dimensional features of the input data into high-dimensional features. Then, the high-dimensional features of the extracted input data were transformed into low-dimensional features of the output data by using an upsampling path, thereby realizing the stress prediction of the turbine rotor. On this basis, the initial model was optimized based on the test dataset using a preset loss function. Specifically, the loss was calculated on the predicted data and the label data in the test dataset using the loss function to generate the loss result. Then, the model parameters of the initial model were adjusted according to the loss result to achieve model optimization and generate the rotor stress prediction model. HRENet uses mean squared error (MSE) as the loss function. This function has the mathematical property of being continuously differentiable and is highly compatible with the gradient descent optimization algorithm. At the same time, MSE is more sensitive to large errors. This property allows the model to correct significant deviations first, thereby effectively improving the prediction accuracy. The expression of the loss function can be shown in the following formula (2): ; Formula (2) in, n The number of samples (i.e., the number of data sets consisting of the actual stress at the monitoring point and the temperature corresponding to that actual stress). For the first The true value of each sample (i.e., the actual stress at the monitoring point); For the first The predicted value of each sample (i.e., the monitored predicted stress).

[0058] In step S140, the stress prediction of the turbine rotor to be predicted is performed based on the rotor stress prediction model to obtain the equivalent stress prediction result, and the fatigue life of the turbine rotor to be predicted is analyzed based on the equivalent stress prediction result.

[0059] Specifically, this can be achieved as follows: Determine the rotor stress parameters of the turbine rotor to be predicted, and input these parameters into the rotor stress prediction model to perform stress prediction on the turbine rotor and obtain the equivalent stress prediction results of the turbine rotor at different time points; obtain the rotor material parameters of the turbine rotor to be predicted, and input these parameters and the equivalent stress prediction results of the turbine rotor at different time points into a preset fatigue life analysis model to obtain the fatigue life analysis results of the turbine rotor to be predicted. That is, in practical applications, by using the trained HRENet model to perform stress prediction based on different startup parameters, the equivalent stress of the generated rotor can be obtained; where the input parameters (i.e., the rotor stress parameters to be predicted) may include the initial temperature. rate of increase Initial heating time and temperature rise rate The input parameters are fed into the rotor stress prediction model to generate the rotor equivalent stress. The turbine rotor to be predicted described here can be the same turbine rotor involved in the model training process, or other turbine rotors; this example does not impose any special restrictions on this. Furthermore, the preset fatigue life analysis model used here can be a large language model or a BERT model; this example does not impose any special restrictions on this.

[0060] Thus, the rotor fatigue life analysis method described in the exemplary embodiments of this disclosure has been fully implemented. Based on the foregoing description, the rotor fatigue life analysis method described in the exemplary embodiments of this disclosure has at least the following advantages: Firstly, it avoids the problem of excessively long calculation times, which makes it difficult to meet the real-time requirements of online monitoring and rapid early warning; secondly, it avoids the problem of insufficient adaptability and computational efficiency of simulation models when facing the non-steady-state and random fluctuation conditions unique to concentrated solar power generation; thirdly, it avoids the problem that the analysis accuracy of finite element simulation heavily depends on the accuracy of boundary condition settings, while the heat transfer boundaries in actual operation are often uncertain; furthermore, the rotor fatigue life analysis method described in the exemplary embodiments of this disclosure can effectively support rotor condition assessment and predictive maintenance.

[0061] The following are embodiments of the apparatus disclosed herein, which can be used to execute embodiments of the method disclosed herein. For details not disclosed in the apparatus embodiments of this disclosure, please refer to the embodiments of the method disclosed herein.

[0062] This disclosure also provides an example embodiment of a rotor fatigue life analysis apparatus. Specifically, refer to... Figure 9 As shown, the rotor fatigue life analysis device may include a stress field analysis result determination module 910, a rotor stress monitoring point determination module 920, a rotor stress prediction model determination module 930, and a rotor fatigue life analysis module 930. Wherein: The stress field analysis result determination module 910 can be used to determine the turbine rotor geometric model and determine the stress field analysis results of the turbine rotor based on the turbine rotor geometric model; the rotor stress monitoring point determination module 920 can be used to determine the rotor stress monitoring points and actual stress at the monitoring points of the turbine rotor based on the stress field analysis results, and determine the initial stress point parameters of the rotor stress monitoring points; the rotor stress prediction model determination module 930 can be used to adjust the parameters of the neural network model to be trained based on the actual stress at the monitoring points and the initial stress point parameters to obtain the rotor stress prediction model; the rotor fatigue life analysis module 940 can be used to perform stress prediction on the turbine rotor to be predicted based on the rotor stress prediction model to obtain the equivalent stress prediction result, and analyze the fatigue life of the turbine rotor to be predicted based on the equivalent stress prediction result.

[0063] In one exemplary embodiment of this disclosure, determining the turbine rotor geometric model includes: acquiring a two-dimensional rotor model of a solar thermal turbine, and splitting the two-dimensional rotor model to obtain multiple vector elements to be assembled; simplifying each of the vector elements to be assembled to obtain simplified vector elements, and generating an initial rotor geometric model based on the simplified vector elements; performing mesh generation on the initial rotor geometric model to obtain an initial mesh generation result, and performing mesh refinement processing on the initial mesh generation result to obtain the turbine rotor geometric model of the solar thermal turbine.

[0064] In one exemplary embodiment of this disclosure, a turbine rotor geometric model of a solar thermal turbine is obtained by mesh refinement processing of an initial mesh division result. This includes: analyzing the initial mesh division result to obtain regions requiring mesh refinement processing; wherein the regions requiring refinement processing include at least one of the following: rounded corner transition regions, regions with discontinuous geometric shapes, and regions where the temperature difference between the rotor surface and the rotor interior exceeds a preset temperature difference threshold due to transient heat transfer processes; performing mesh refinement processing on the regions requiring refinement to obtain a refined mesh division result, and calculating the heat transfer coefficient of each different region in the refined mesh division result based on preset steam parameters; configuring the heat transfer coefficient for each different region in the refined mesh division result to obtain the turbine rotor geometric model of the solar thermal turbine.

[0065] In one exemplary embodiment of this disclosure, determining the stress field analysis result of the turbine rotor based on the turbine rotor geometric model includes: performing a temperature field analysis on the turbine rotor geometric model based on preset boundary conditions to obtain a first temperature field analysis result of the turbine rotor; and performing a stress analysis on the turbine rotor geometric model based on the first temperature field analysis result to obtain a stress field analysis result of the turbine rotor.

[0066] In an exemplary embodiment of this disclosure, temperature field analysis is performed on the turbine rotor geometric model based on preset boundary conditions to obtain a first temperature field analysis result of the turbine rotor. This includes: configuring preset boundary conditions for the turbine rotor geometric model; wherein the preset boundary conditions include at least one of transient thermal boundary conditions, mechanical load conditions, and displacement boundary conditions; obtaining a turbine start-up characteristic curve corresponding to the turbine rotor, and determining initial steady-state thermal boundary conditions for steady-state analysis of the turbine rotor based on the turbine start-up characteristic curve; loading the initial steady-state thermal boundary conditions into the turbine rotor geometric model to obtain a rotor steady-state temperature field, and determining initial transient conditions for transient analysis of the turbine rotor based on the rotor steady-state temperature field; and performing transient analysis on the turbine rotor based on the initial transient conditions and the transient thermal boundary conditions to obtain a first temperature field analysis result of the turbine rotor.

[0067] In an exemplary embodiment of this disclosure, determining the rotor stress monitoring point and the actual stress at the monitoring point of the steam turbine rotor based on the stress field analysis results, and determining the initial stress point parameters of the rotor stress monitoring point, includes: determining the stress concentration region of the steam turbine rotor based on the stress field analysis results, and determining the rotor stress monitoring point of the steam turbine rotor based on the stress concentration region; determining the actual stress at the monitoring point of the rotor stress monitoring point based on the stress field analysis results, and determining the initial temperature, heating rate, initial heating time, and temperature rise rate of the rotor stress monitoring point; and generating the initial stress point parameters based on the initial temperature, heating rate, initial heating time, and temperature rise rate of the stress monitoring point.

[0068] In one exemplary embodiment of this disclosure, a rotor stress prediction model is obtained by adjusting the parameters of a neural network model to be trained based on the actual stress at the rotor stress monitoring point and the initial stress point parameters. This includes: inputting the initial stress point parameters of the rotor stress monitoring point into the neural network model to be trained to obtain the predicted stress at the monitoring point; constructing a target loss function based on the predicted stress at the monitoring point and the actual stress at the monitoring point; and adjusting the parameters of the neural network model to be trained based on the target loss function to obtain the rotor stress prediction model.

[0069] In an exemplary embodiment of this disclosure, the neural network model to be trained includes a feature expansion layer, an initial convolutional layer, a downsampling path, and an upsampling path. The process of inputting the initial stress point parameters of the rotor stress monitoring point into the neural network model to obtain the predicted stress of the monitoring point includes: expanding the initial temperature based on the initial stress point parameters using the feature expansion layer according to the rate of increase, initial heating time, and temperature rise rate to obtain expanded temperatures at multiple different time points; performing convolution processing on the expanded temperatures at the multiple different time points based on the initial convolutional layer to obtain convolution processing results at multiple different time points; performing downsampling processing on the convolution processing results at the multiple different time points based on the downsampling path to obtain local temperature sampling results at different time points; and performing upsampling processing on the local temperature sampling results at the multiple different time points based on the upsampling path to obtain the predicted stress of the monitoring point.

[0070] In one exemplary embodiment of this disclosure, stress prediction of the turbine rotor to be predicted is performed based on the rotor stress prediction model to obtain equivalent stress prediction results, and fatigue life of the turbine rotor to be predicted is analyzed based on the equivalent stress prediction results. This includes: determining the rotor stress parameters of the turbine rotor to be predicted, and inputting the rotor stress parameters into the rotor stress prediction model to perform stress prediction on the turbine rotor to obtain equivalent stress prediction results of the turbine rotor at different time points; obtaining the rotor material parameters of the turbine rotor to be predicted, and inputting the rotor material parameters and the equivalent stress prediction results of the turbine rotor to be predicted at different time points into a preset fatigue life analysis model to obtain fatigue life analysis results of the turbine rotor to be predicted.

[0071] The specific details of each module in the above-mentioned rotor fatigue life analysis device have been described in detail in the corresponding rotor fatigue life analysis method, so they will not be repeated here.

[0072] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0073] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0074] In an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method is also provided.

[0075] Those skilled in the art will understand that various aspects of this disclosure can be implemented as a system, method, or program product. Therefore, various aspects of this disclosure can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software aspects, collectively referred to herein as a "circuit," "module," or "system."

[0076] The following reference Figure 10 To describe an electronic device 1000 according to such an embodiment of the present disclosure. Figure 10 The electronic device 1000 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0077] like Figure 10 As shown, the electronic device 1000 is manifested in the form of a general-purpose computing device. The components of the electronic device 1000 may include, but are not limited to: at least one processing unit 1010, at least one storage unit 1020, a bus 1030 connecting different system components (including storage unit 1020 and processing unit 1010), and a display unit 1040.

[0078] The storage unit stores program code that can be executed by the processing unit 1010, causing the processing unit 1010 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. For example, the processing unit 1010 can perform actions such as... Figure 1The steps shown are as follows: Step S110: Determine the turbine rotor geometric model and determine the stress field analysis results of the turbine rotor based on the turbine rotor geometric model; Step S120: Determine the rotor stress monitoring points and actual stress at the monitoring points of the turbine rotor based on the stress field analysis results, and determine the initial stress point parameters of the rotor stress monitoring points; Step S130: Based on the actual stress at the monitoring points and the initial stress point parameters, adjust the parameters of the neural network model to be trained to obtain a rotor stress prediction model; Step S140: Based on the rotor stress prediction model, perform stress prediction on the turbine rotor to be predicted to obtain equivalent stress prediction results, and analyze the fatigue life of the turbine rotor to be predicted based on the equivalent stress prediction results.

[0079] Storage unit 1020 may include readable media in the form of volatile storage units, such as random access memory (RAM) 10201 and / or cache memory 10202, and may further include read-only memory (ROM) 10203.

[0080] Storage unit 1020 may also include a program / utility 10204 having a set (at least one) program module 10205, such program module 10205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0081] Bus 1030 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0082] Electronic device 1000 can also communicate with one or more external devices 1100 (e.g., keyboard, pointing device, Bluetooth device, etc.), one or more devices that enable a user to interact with electronic device 1000, and / or any device that enables electronic device 1000 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 1050. Furthermore, electronic device 1000 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1060. As shown, network adapter 1060 communicates with other modules of electronic device 1000 via bus 1030. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 1000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0083] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0084] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible implementations, various aspects of this disclosure may also be implemented as a program product including program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps of the various exemplary embodiments of this disclosure described in the "Exemplary Methods" section above.

[0085] The program product for implementing the above-described method according to embodiments of the present disclosure may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, the readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0086] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0087] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0088] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0089] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0090] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0091] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention described herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not invented by this disclosure. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

Claims

1. A method for analyzing rotor fatigue life, characterized in that, include: Determine the geometric model of the steam turbine rotor, and determine the stress field analysis results of the steam turbine rotor based on the geometric model of the steam turbine rotor; Based on the stress field analysis results, the rotor stress monitoring points and actual stresses at the monitoring points of the turbine rotor are determined, and the initial stress point parameters of the rotor stress monitoring points are determined. Based on the actual stress at the monitoring point and the initial stress point parameters of the rotor stress monitoring point, the parameters of the neural network model to be trained are adjusted to obtain the rotor stress prediction model. Based on the rotor stress prediction model, the stress of the turbine rotor to be predicted is predicted to obtain the equivalent stress prediction result, and the fatigue life of the turbine rotor to be predicted is analyzed based on the equivalent stress prediction result.

2. The method for analyzing rotor fatigue life according to claim 1, characterized in that, Determine the turbine rotor geometry model, including: A two-dimensional rotor model of a solar thermal turbine is obtained, and the two-dimensional rotor model is split into multiple vector elements to be assembled. The vector elements to be assembled are simplified to obtain simplified vector elements, and an initial rotor geometric model is generated based on the simplified vector elements. The initial rotor geometric model is meshed to obtain the initial meshing result, and the initial meshing result is refined to obtain the turbine rotor geometric model of the solar thermal turbine.

3. The method for analyzing rotor fatigue life according to claim 2, characterized in that, The initial mesh generation results are refined to obtain the turbine rotor geometric model of the solar thermal turbine, including: The initial mesh division results are analyzed to obtain the regions to be refined that require mesh refinement processing; wherein, the regions to be refined include at least one of the following: rounded corner transition regions, regions with discontinuous geometry, and regions where the temperature difference between the rotor surface and the rotor interior is greater than a preset temperature difference threshold due to transient heat transfer processes; The area to be encrypted is subjected to mesh encryption processing to obtain the encrypted mesh division result, and the heat transfer coefficient of each different region in the encrypted mesh division result is calculated according to the preset steam parameters. The heat transfer coefficient is configured for each different region in the encrypted mesh division result to obtain the turbine rotor geometric model of the solar thermal turbine.

4. The method for analyzing rotor fatigue life according to claim 1, characterized in that, The stress field analysis results of the turbine rotor, determined based on the turbine rotor geometric model, include: Temperature field analysis is performed on the geometric model of the steam turbine rotor based on preset boundary conditions to obtain the first temperature field analysis result of the steam turbine rotor; Based on the first temperature field analysis results, stress analysis is performed on the turbine rotor geometric model to obtain the stress field analysis results of the turbine rotor.

5. The method for analyzing rotor fatigue life according to claim 4, characterized in that, Temperature field analysis is performed on the turbine rotor geometric model based on preset boundary conditions to obtain the first temperature field analysis result of the turbine rotor, including: Configure preset boundary conditions for the turbine rotor geometric model; wherein the preset boundary conditions include at least one of transient thermal boundary conditions, mechanical load conditions, and displacement boundary conditions; Obtain the turbine start-up characteristic curve corresponding to the turbine rotor, and determine the initial steady-state thermal boundary conditions for the turbine rotor to perform steady-state analysis based on the turbine start-up characteristic curve; The initial steady-state thermal boundary conditions are applied to the turbine rotor geometric model to obtain the rotor steady-state temperature field, and the initial transient conditions for transient analysis of the turbine rotor are determined based on the rotor steady-state temperature field. Transient analysis of the turbine rotor is performed based on the initial transient conditions and transient thermal boundary conditions to obtain the first temperature field analysis results of the turbine rotor.

6. The method for analyzing rotor fatigue life according to claim 1, characterized in that, Based on the stress field analysis results, the rotor stress monitoring points and actual stresses at the monitoring points of the turbine rotor are determined, and the initial stress point parameters of the rotor stress monitoring points are determined, including: The stress concentration region of the turbine rotor is determined based on the stress field analysis results, and the rotor stress monitoring point of the turbine rotor is determined based on the stress concentration region. Based on the stress field analysis results, the actual stress at the rotor stress monitoring point is determined, and the initial temperature, heating rate, initial heating time, and temperature rise rate at the rotor stress monitoring point are also determined. The initial stress point parameters are generated based on the initial temperature, heating rate, initial heating time, and temperature rise rate of the stress monitoring point.

7. The method for analyzing rotor fatigue life according to claim 1, characterized in that, Based on the actual stress at the monitored rotor stress points and the initial stress point parameters, the parameters of the neural network model to be trained are adjusted to obtain the rotor stress prediction model, including: The initial stress point parameters of the rotor stress monitoring point are input into the neural network model to be trained to obtain the predicted stress of the monitoring point; A target loss function is constructed based on the predicted stress at the monitoring point and the actual stress at the monitoring point. The parameters of the neural network model to be trained are then adjusted based on the target loss function to obtain the rotor stress prediction model.

8. The method for analyzing rotor fatigue life according to claim 7, characterized in that, The neural network model to be trained includes a feature expansion layer, an initial convolutional layer, a downsampling path, and an upsampling path; The process of inputting the initial stress point parameters of the rotor stress monitoring point into the neural network model to be trained to obtain the predicted stress of the monitoring point includes: Based on the feature extension layer, the initial temperature is extended according to the rate of rise, initial heating time and temperature rise rate in the initial stress point parameters to obtain extended temperatures at multiple different time points. Based on the initial convolutional layer, convolution processing is performed on the extended temperatures at multiple different time points to obtain convolution processing results at multiple different time points; Based on the downsampling path, the convolution processing results at multiple different time nodes are downsampled to obtain local temperature sampling results at different time nodes; Upsampling processing is performed on the local temperature sampling results at multiple different time points along the upsampling path to obtain the predicted stress at the monitoring point.

9. The method for analyzing rotor fatigue life according to claim 1, characterized in that, Based on the rotor stress prediction model, the equivalent stress prediction result is obtained by predicting the stress of the turbine rotor to be predicted. Then, the fatigue life of the turbine rotor to be predicted is analyzed based on the equivalent stress prediction result, including: The stress parameters of the turbine rotor to be predicted are determined and input into the rotor stress prediction model to perform stress prediction on the turbine rotor to be predicted and obtain the equivalent stress prediction results of the turbine rotor at different time nodes. Obtain the rotor material parameters of the turbine rotor to be predicted, and input the rotor material parameters and the equivalent stress prediction results of the turbine rotor to be predicted at different time nodes into the preset fatigue life analysis model to obtain the fatigue life analysis results of the turbine rotor to be predicted.

10. An analytical apparatus for rotor fatigue life, characterized in that, include: The stress field analysis result determination module is used to determine the turbine rotor geometric model and determine the stress field analysis results of the turbine rotor based on the turbine rotor geometric model. The rotor stress monitoring point determination module is used to determine the rotor stress monitoring point and the actual stress at the monitoring point of the turbine rotor based on the stress field analysis results, and to determine the initial stress point parameters of the rotor stress monitoring point. The rotor stress prediction model determination module is used to adjust the parameters of the neural network model to be trained based on the actual stress at the monitoring point and the initial stress point parameters of the rotor stress monitoring point to obtain the rotor stress prediction model. The rotor fatigue life analysis module is used to predict the stress of the turbine rotor to be predicted based on the rotor stress prediction model to obtain the equivalent stress prediction result, and to analyze the fatigue life of the turbine rotor to be predicted based on the equivalent stress prediction result.

Images