Gas turbine blade life evaluation method and device, terminal equipment and storage medium

Abstract

The invention relates to the technical field of gas turbines, in particular to a gas turbine blade service life evaluation method and device, terminal equipment and a storage medium. The method comprises the following steps: acquiring boundary condition parameters of the current operation condition of the gas turbine in real time, inputting the boundary condition parameters into a trained full-connection neural network model, and outputting corresponding projection coefficients; reconstructing the fatigue life distribution of each node of the turbine blade under the current working condition according to the linear combination of the projection coefficient and the primary function; and calculating the life loss of each node of each turbine blade through linear cumulative damage in combination with the operation time and the start-stop times, and outputting the residual fatigue life of each turbine blade in combination with the fatigue life distribution. And the service life of the gas turbine blade can be quickly and accurately evaluated.

Description

Technical Field

[0001] This application relates to the field of gas turbine technology, and in particular to a method, apparatus, terminal equipment and storage medium for assessing the life of gas turbine blades. Background Technology

[0002] Gas turbines, as efficient and clean energy conversion devices, are widely used in power generation, aero propulsion, and industrial drives. Turbine blades, especially the first-stage moving blades of high-pressure turbines, are among the most critical hot-end components of gas turbines, operating for extended periods in extreme high-temperature (above 1200℃), high-pressure, high-speed, and corrosive gas environments. Under such complex load coupling, turbine blades are highly susceptible to failure modes such as creep, low-cycle fatigue, oxidation corrosion, and thermal shock damage, seriously threatening unit operational safety and incurring high maintenance and replacement costs. Therefore, accurate and efficient assessment of turbine blade fatigue life has become a core technical requirement for ensuring gas turbine reliability, optimizing maintenance strategies, and improving economic efficiency. Current technologies lack a method for assessing turbine blade fatigue life that can guarantee both high accuracy and real-time rapid response. In particular, with the increasing demands for flexible operation of gas turbines in new power systems (such as frequent start-stop and rapid load changes), traditional EOH methods are insufficient to meet the needs of refined operation and maintenance, while high-precision simulation methods are limited by computational efficiency bottlenecks. Therefore, there is an urgent need to develop a new generation of intelligent life prediction method that integrates high-fidelity physical models and data-driven technology to significantly improve computing speed while ensuring accuracy, so as to support the full life cycle health monitoring and real-time early warning of remaining life of turbine blades. Summary of the Invention

[0003] In view of this, embodiments of this application provide a method, apparatus, terminal equipment, and storage medium for assessing the lifespan of gas turbine blades, which can effectively solve the problem of low accuracy.

[0004] In a first aspect, embodiments of this application provide a method for assessing the lifespan of gas turbine blades, including: The boundary condition parameters of the gas turbine's current operating condition are acquired in real time, input into the trained fully connected neural network model, and the corresponding projection coefficients are output. The fatigue life distribution of each node of the turbine blade under the current operating condition is reconstructed based on the linear combination of the projection coefficients and basis functions. By combining the running time and the number of start-stop cycles, the life loss of each node of each turbine blade is calculated through linear cumulative damage, and the remaining fatigue life of each turbine blade is output based on the fatigue life distribution.

[0005] In some embodiments, the training process of the fully connected neural network model includes: Obtain boundary condition parameters of the gas turbine under various operating conditions; For each operating condition, the temperature and stress field distributions on and inside the turbine blade surface are calculated using a thermo-fluid-structure interaction simulation method. Based on the temperature field distribution and the stress field distribution, combined with the fatigue performance test data of the gas turbine blade material, the fatigue life value of the key nodes of the blade under each operating condition is calculated, forming a snapshot matrix of the fatigue life distribution under multiple operating conditions. The snapshot matrix is ​​subjected to intrinsic orthogonal decomposition to extract the first i dominant mode basis functions, where i satisfies that the sum of the first i eigenvalues ​​accounts for a proportion greater than a preset proportion and i is greater than 0; A fully connected neural network model is constructed, with the boundary condition parameters as input and the projection coefficients of fatigue life on each dominant modal basis function under the corresponding working condition as output, and the mapping relationship between the working condition parameters and the projection coefficients is obtained through training. In some embodiments, the boundary condition parameters include one or more of the following: air flow rate, natural gas flow rate, inlet temperature, inlet angle, cooling gas temperature, compressor outlet pressure, and turbine exhaust pressure.

[0006] In some embodiments, the fatigue performance test data is obtained in the following ways: Based on the temperature field and equivalent stress field distribution obtained from the simulation, the region where the stress, temperature or temperature gradient is greater than the preset value is selected as the sample cutting area; Tensile and fatigue specimens are obtained from the cut area; High-temperature low-cycle fatigue tests were conducted under simulated actual service conditions to obtain the fatigue life characteristics of materials at different temperatures and strain amplitudes.

[0007] In some embodiments, the key regions include one or more of the following: the leading edge of the blade, the trailing edge, the tenon transition area, the endwall connection, and the blade tip cavity area.

[0008] In some embodiments, the fully connected neural network includes an input layer, at least one hidden layer, and an output layer, wherein the number of neurons in the input layer is equal to the number of operating parameters, and the number of neurons in the output layer is equal to the number of selected dominant mode basis functions.

[0009] In some embodiments, the life loss of each node of each turbine blade includes start-stop damage; The maximum stress at each node of the turbine blade is calculated during a single start-up and shutdown process to obtain the corresponding fatigue life. Based on the fatigue life, the start-up and shutdown fatigue damage under standard operating conditions is obtained. Based on the magnitude of the peak power during start-stop, a damage coefficient is determined. Based on the damage coefficient and the start-stop fatigue damage, start-stop damage under different operating conditions is obtained.

[0010] Secondly, this application also provides a gas turbine blade life assessment device, comprising: The data acquisition module is used to acquire the boundary condition parameters of the current operating condition of the gas turbine in real time, input them into the trained fully connected neural network model, and output the corresponding projection coefficients. The data analysis module is used to reconstruct the fatigue life distribution of each node of the turbine blade under the current operating condition based on the linear combination of the projection coefficients and the basis functions. The data calculation module is used to calculate the life loss of each node of each turbine blade by combining the running time and the number of start-stop cycles through linear cumulative damage, and output the remaining fatigue life of each turbine blade by combining the fatigue life distribution.

[0011] Thirdly, this application also provides a terminal device, which includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the gas turbine blade life assessment method.

[0012] Fourthly, this application also provides a readable storage medium storing a computer program that, when executed on a processor, implements the gas turbine blade life assessment method.

[0013] The embodiments of this application have the following beneficial effects: The gas turbine blade life assessment method of this application acquires the boundary condition parameters of the current operating condition of the gas turbine in real time, inputs them into a trained fully connected neural network model, and outputs the corresponding projection coefficients. Based on the linear combination of the projection coefficients and the basis functions, the fatigue life distribution of each node of the turbine blade under the current operating condition is reconstructed. Combining the operating time and the number of start-stop cycles, the life loss of each node of each turbine blade is calculated through linear cumulative damage, and the remaining fatigue life of each turbine blade is output based on the fatigue life distribution. This enables rapid and accurate turbine blade life assessment, bridging the gap between accuracy and efficiency in existing methods, and providing key technical support for the intelligent health management of high-end power equipment. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This paper illustrates a flowchart of a gas turbine blade life assessment method according to an embodiment of this application. Figure 2 This paper illustrates a schematic diagram of a fully connected neural network model structure according to an embodiment of this application. Figure 3 This paper illustrates a schematic diagram of a fully connected neural network model training process according to an embodiment of this application. Figure 4 The image shown is a cloud map of blade temperature distribution under standard operating conditions, as illustrated in an embodiment of this application. Figure 5 This paper illustrates a schematic diagram of the damage calculation coefficient curve during a machine trip, according to an embodiment of this application. Figure 6 A schematic diagram of a gas turbine blade life assessment device according to an embodiment of this application is shown. Detailed Implementation

[0016] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0017] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0018] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0019] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0020] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0021] To address the problems of existing technologies, this application provides a method for assessing the lifespan of gas turbine blades. The method involves acquiring boundary condition parameters of the gas turbine's current operating condition in real time, inputting them into a trained fully connected neural network model, and outputting corresponding projection coefficients. The fatigue life distribution of each node of the turbine blade under the current operating condition is reconstructed based on a linear combination of the projection coefficients and the basis functions. Combining the operating time and number of start-stop cycles, the life loss of each node of each turbine blade is calculated through linear cumulative damage, and the remaining fatigue life of each turbine blade is output based on the fatigue life distribution. This enables rapid and accurate assessment of the turbine blade's lifespan.

[0022] The following examples illustrate the method for assessing the lifespan of gas turbine blades.

[0023] Figure 1 A flowchart illustrating a gas turbine blade life assessment method according to an embodiment of this application is shown. Exemplarily, the gas turbine blade life assessment method includes the following steps: Step S100: Obtain the boundary condition parameters of the current operating condition of the gas turbine in real time, input them into the trained fully connected neural network model, and output the corresponding projection coefficients.

[0024] When the gas turbine is working, boundary parameters are read in real time by sensors installed in the gas turbine. These boundary parameters include parameters such as air flow rate, natural gas flow rate, inlet temperature, inlet angle, cooling gas temperature, compressor outlet pressure, and turbine exhaust pressure. The specific parameters that need to be acquired can be adjusted according to the site conditions.

[0025] The acquired boundary condition parameters will be cleaned and normalized sequentially to adapt to the neural network input format.

[0026] This embodiment uses a pre-trained fully connected neural network model to process these boundary condition parameters. The fully connected neural network model processes these parameters and outputs the corresponding projection coefficients.

[0027] like Figure 2 The diagram shown illustrates the neural network structure and its output in this embodiment. The fully connected neural network model includes an input layer, at least one hidden layer, and an output layer. The number of neurons in the input layer is equal to the number of operating parameters, and the number of neurons in the output layer is equal to the number of selected dominant mode basis functions.

[0028] like Figure 3 The diagram shows the training process of this model, which includes: Step S110: Obtain the boundary condition parameters of the gas turbine under various typical operating conditions.

[0029] The method for obtaining the boundary conditions of this section is similar to the aforementioned method, and will not be repeated here.

[0030] Step S120: For each working condition, the temperature field distribution and stress field distribution on and inside the turbine blade surface are calculated using the thermo-fluid-structure interaction simulation method.

[0031] For each combination of operating parameters in the scheme, a complete coupled simulation calculation is performed. The CFD module calculates the aerodynamic pressure and temperature distribution on the blade surface. Subsequently, these thermal and aerodynamic loads are applied to the FEM model. Combined with centrifugal loads and the physical properties of the material, including density, thermal conductivity, coefficient of linear expansion, elastic properties, tensile curve, and Poisson's ratio, the steady-state stress field distribution and temperature field distribution inside the blade are finally calculated.

[0032] Figure 4 This is a cloud map showing the blade temperature distribution under standard operating conditions.

[0033] Figure 4 In the diagram, 'a' represents the temperature distribution contour map of the blade tip cavity. Figure 4 In the diagram, 'b' represents the temperature distribution contour map of the pressure surface. Figure 4 Figure 'c' shows the temperature distribution cloud map of the suction surface. It can be seen that the temperature of the blade tip cavity wall is relatively high, especially near the suction surface, where the highest temperature approaches 1069.8℃, making it the hottest area on the entire blade. This is because the mainstream flow velocity is high at this location, resulting in a high heat transfer coefficient. Extremely high heat flow is transferred from the mainstream to the blade, and the small contact area between the cavity wall and the blade limits its thermal conductivity, preventing the introduced heat flow from reaching the blade interior in time. The lower surface inside the blade tip cavity is cooler due to the influence of the cooling air, resulting in better cooling. The blade trailing edge also has a high temperature, approximately 1030℃. This is because the cooling holes at the trailing edge have a small diameter (2mm), a small cooling gas flow rate, and the trailing edge itself is long, narrow, and thin, making it difficult to cool and resulting in a higher temperature at the trailing edge. The blade leading edge is directly impacted by the incoming flow, resulting in a high temperature (approximately 940℃). The blade root is not directly exposed to the high-temperature exhaust gas and therefore has a lower temperature. The temperature in the middle region of the suction surface of the blade is below 940℃, while the temperature in the middle region of the pressure surface of the blade is slightly lower, below 910℃.

[0034] Step S130: Based on the temperature field distribution and the stress field distribution, and combined with the fatigue performance test data of the gas turbine blade material, calculate the fatigue life value of the grid nodes in the key area of ​​the blade under each operating condition, and form a snapshot matrix of the spatial distribution of fatigue life under multiple operating conditions.

[0035] The fatigue performance test data can be obtained by selecting high stress, high temperature or high temperature gradient regions as the specimen cutting positions based on the temperature field and equivalent stress field distribution obtained from the simulation; then tensile specimens and fatigue specimens are cut from the gas turbine blades; finally, high temperature and low cycle fatigue tests are carried out under simulated actual service environment conditions to obtain the fatigue life law of the material under different temperatures and strain amplitudes.

[0036] As an example, this embodiment first uses the intrinsic orthogonal decomposition method to extract modal features of the blade at different transition zone positions under different operating conditions. For example, the fatigue life calculation results of the blade under m operating conditions are used to form an n×m snapshot matrix S. m×n : .

[0037] Where m is the number of selected operating conditions, and n is the number of mesh nodes at the selected location on the blade. This represents the fatigue life of the nth grid node under the m-th working condition. Step S140: Perform eigenorthogonal decomposition on the snapshot matrix to extract the first i-th dominant mode basis functions.

[0038] For matrix nonzero eigenvalues Solve ( , Then, the eigenvectors of P are obtained through the eigenvalues. Furthermore, through the eigenorthogonal basis functions : ; Step S150: Construct a fully connected neural network model, taking the boundary condition parameters of the various working conditions as input and the projection coefficients of fatigue life on the basis functions of each dominant mode under the corresponding working conditions as output, and train it to obtain the mapping relationship between the working condition parameters and the projection coefficients. By summing the first i-th order eigenvalues ​​of the basis functions and calculating their proportion in the total sum of eigenvalues, for example, if the proportion of the sum of the first 7 eigenvalues ​​is much greater than 90%, then the first 7 basis functions can be considered dominant. Therefore, the first 7 basis functions can be used to reconstruct the fatigue life of the blade. ; in, These are the projection coefficients of the basis functions in the vector space, and these projection coefficients are the projection coefficients of the model output.

[0039] The input layer receives the boundary condition parameters, which are then processed by the intermediate hidden layer to output the corresponding projection coefficients. These coefficients are the projection coefficients of the fatigue life of each node of the blade under each working condition, used to characterize the coordinates of the spatial lifetime distribution in the linear basis function space.

[0040] It is understood that in the scenario of this embodiment, the high-dimensional fatigue life field obtained by the fluid-thermal-structure coupling simulation under each working condition is the life value on tens of thousands of finite element nodes on the blade surface. Directly performing machine learning modeling on this high-dimensional data is computationally expensive and difficult to converge. Therefore, this embodiment uses the projection coefficient to reduce the dimensionality of this high-dimensional feature in order to characterize the coordinate position of the current life in the low-dimensional space.

[0041] Step S200: Reconstruct the fatigue life distribution of each node of the turbine blade under the current operating condition based on the linear combination of the projection coefficients and the basis functions.

[0042] After obtaining the output projection coefficients, the fatigue life distribution of each node of the turbine blade under the previous operating condition can be reconstructed by combining the basis function.

[0043] Its calculation expression is as described above: ; The above formula can be used to reconstruct the blade and calculate its fatigue life. .

[0044] Step S300: Combining the running time and the number of start-stop cycles, calculate the life loss of each node of each turbine blade through linear cumulative damage, and output the remaining fatigue life of each turbine blade based on the fatigue life distribution.

[0045] This embodiment uses linear cumulative damage calculation to determine the life loss of each node on each turbine blade. This means that the damage to the material in each micro-time is considered independent, and the total damage can be linearly accumulated. The life loss of each finite element node on the blade is calculated based on the operating time under each operating condition, thus obtaining the remaining life of the blade. When the life of a blade node is exhausted, the blade material can be considered to have reached its fatigue limit.

[0046] The maximum stress at each node of the turbine blade is calculated during a single start-up and shutdown process to obtain the corresponding fatigue life. Based on the fatigue life, the start-up and shutdown fatigue damage under standard operating conditions is obtained. Based on the magnitude of the peak power during start-stop, a damage coefficient is determined. Based on the damage coefficient and the start-stop fatigue damage, start-stop damage under different operating conditions is obtained.

[0047] As an example, in fatigue accumulation calculations, the maximum stress σ at a node is quickly calculated by referring to a single start-stop process. max The fatigue life under this condition was calculated based on the test results. Start-stop damage is defined as 1 / To simplify practical applications, the start-stop fatigue damage under standard operating conditions is used as a parameter to represent the damage a / for each start-stop cycle. 'a' represents the damage factor. When the peak power during start-up and shutdown is between the rated power Pn and 0.6Pn, the damage factor 'a' is 1.0; when the peak power during start-up and shutdown is greater than Pn, the damage factor 'a' is 1.3; and when the peak power during start-up and shutdown is less than 0.6Pn, the damage factor 'a' is 0.5. In the event of a trip, the damage factor 'a' can be based on... Figure 5 To determine. Figure 5 The two curves represent the damage coefficient 'a' as a function of load rate in the event of a shutdown, under scenarios with and without intake air heating. The value of 'a' varies at different stages of the shutdown process. Figure 5 This is just an illustrative relationship curve; depending on the actual situation, similar graphs may exist. Figure 5 The details of the relationship curve may vary, but this graph is used here for illustration. In the event of a machine trip, the value of the damage coefficient 'a' can be determined by a similar relationship curve.

[0048] Based on the above method, fatigue life was calculated in this embodiment. The remaining fatigue life can be obtained by subtracting the fatigue life from the fatigue life loss.

[0049] The gas turbine blade life assessment method provided in this embodiment constructs a high-fidelity operating condition-life mapping relationship through a trained fully connected neural network, based on real physical field simulation and combined with measured material fatigue performance. This method can accurately quantify local stress, temperature, and fatigue damage under different combinations of operating parameters, solving the problems of existing methods neglecting dynamic changes in operating conditions and having low assessment accuracy. By extracting the dominant mode of the life distribution, the output is compressed from life values ​​of tens of thousands of grid nodes to 7 projection coefficients, significantly reducing the dimensionality of the neural network output and decreasing training difficulty and the risk of overfitting.

[0050] Figure 6 A schematic diagram of a gas turbine blade life assessment device according to an embodiment of this application is shown. Exemplarily, the gas turbine blade life assessment device includes: Data acquisition module 10 is used to acquire boundary condition parameters of the current operating condition of the gas turbine in real time, input them into the trained fully connected neural network model, and output the corresponding projection coefficients. Data analysis module 20 is used to reconstruct the fatigue life distribution of each node of the turbine blade under the current operating condition based on the linear combination of the projection coefficient and the basis function. The data calculation module 30 is used to calculate the life loss of each node of each turbine blade by combining the running time and the number of start-stop cycles through linear cumulative damage, and output the remaining fatigue life of each turbine blade by combining the fatigue life distribution.

[0051] It is understood that the device in this embodiment corresponds to the gas turbine blade life assessment method in the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described again here.

[0052] This application also provides a terminal device, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the terminal device to perform the functions of the various modules in the above-described gas turbine blade life assessment method or the above-described gas turbine blade life assessment device.

[0053] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0054] The memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory is used to store computer programs, and the processor can execute the computer programs accordingly after receiving execution instructions.

[0055] This application also provides a readable storage medium for storing the computer program used in the aforementioned terminal device.

[0056] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0057] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0058] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0059] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for assessing the lifespan of gas turbine blades, characterized in that, include: The boundary condition parameters of the gas turbine's current operating condition are acquired in real time, input into the trained fully connected neural network model, and the corresponding projection coefficients are output. The fatigue life distribution of each node of the turbine blade under the current operating condition is reconstructed based on the linear combination of the projection coefficients and basis functions. By combining the running time and the number of start-stop cycles, the life loss of each node of each turbine blade is calculated through linear cumulative damage, and the remaining fatigue life of each turbine blade is output based on the fatigue life distribution.

2. The gas turbine blade life assessment method according to claim 1, characterized in that, The training process of the fully connected neural network model includes: Obtain boundary condition parameters of the gas turbine under various operating conditions; For each operating condition, the temperature and stress field distributions on and inside the turbine blade surface are calculated using a thermo-fluid-structure interaction simulation method. Based on the temperature field distribution and the stress field distribution, combined with the fatigue performance test data of the gas turbine blade material, the fatigue life value of the key nodes of the blade under each operating condition is calculated, forming a snapshot matrix of the fatigue life distribution under multiple operating conditions. The snapshot matrix is ​​subjected to intrinsic orthogonal decomposition to extract the first i dominant mode basis functions, where i satisfies that the sum of the first i eigenvalues ​​accounts for a proportion greater than a preset proportion and i is greater than 0; A fully connected neural network model is constructed, with the boundary condition parameters as input and the projection coefficients of fatigue life on each dominant modal basis function under the corresponding working condition as output, and the mapping relationship between the working condition parameters and the projection coefficients is obtained through training.

3. The gas turbine blade life assessment method according to claim 2, characterized in that, The boundary condition parameters include one or more of the following: air flow rate, natural gas flow rate, inlet temperature, inlet angle, cooling gas temperature, compressor outlet pressure, and turbine exhaust pressure.

4. The method for assessing the lifespan of gas turbine blades according to claim 2, characterized in that, The fatigue performance test data is obtained through the following methods: Based on the temperature field and equivalent stress field distribution obtained from the simulation, the region where the stress, temperature or temperature gradient is greater than the preset value is selected as the sample cutting area; Tensile and fatigue specimens are obtained from the cut area; High-temperature low-cycle fatigue tests were conducted under simulated actual service conditions to obtain the fatigue life characteristics of materials at different temperatures and strain amplitudes.

5. The gas turbine blade life assessment method according to claim 2, characterized in that, The key areas include one or more of the following: the leading edge of the blade, the trailing edge, the tenon transition area, the end wall connection, and the blade tip cavity area.

6. The gas turbine blade life assessment method according to claim 1, characterized in that, The fully connected neural network includes an input layer, at least one hidden layer, and an output layer, wherein the number of neurons in the input layer is equal to the number of operating parameters, and the number of neurons in the output layer is equal to the number of selected dominant mode basis functions.

7. The method for assessing the life of gas turbine blades according to claim 1, characterized in that, The life loss of each node of each turbine blade includes start-stop damage. The maximum stress at each node of the turbine blade is calculated during a single start-up and shutdown process to obtain the corresponding fatigue life. Based on the fatigue life, the start-up and shutdown fatigue damage under standard operating conditions is obtained. Based on the magnitude of the peak power during start-stop, a damage coefficient is determined. Based on the damage coefficient and the start-stop fatigue damage, start-stop damage under different operating conditions is obtained.

8. A gas turbine blade life assessment device, characterized in that, include: The data acquisition module is used to acquire the boundary condition parameters of the current operating condition of the gas turbine in real time, input them into the trained fully connected neural network model, and output the corresponding projection coefficients. The data analysis module is used to reconstruct the fatigue life distribution of each node of the turbine blade under the current operating condition based on the linear combination of the projection coefficients and the basis functions. The data calculation module is used to calculate the life loss of each node of each turbine blade by combining the running time and the number of start-stop cycles through linear cumulative damage, and output the remaining fatigue life of each turbine blade by combining the fatigue life distribution.

9. A terminal device, characterized in that, The terminal device includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the gas turbine blade life assessment method according to any one of claims 1-7.

10. A readable storage medium, characterized in that, It stores a computer program that, when executed on a processor, implements the gas turbine blade life assessment method according to any one of claims 1-7.

Images